Mouton Grammar Library

A Grammar of Urarina

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
Mouton Grammar Library 37

Editors
Georg Bossong
Bernard Comrie
Matthew Dryer

Mouton de Gruyter
Berlin · New York
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
A Grammar of Urarina

by
Knut J. Olawsky

Mouton de Gruyter
Berlin · New York
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Mouton de Gruyter (formerly Mouton, The Hague)
is a Division of Walter de Gruyter G m b H & Co. KG, Berlin.

© Printed on acid-free paper which falls within the guidelines of the

ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data

Olawsky, Knut J.
A grammar of Urarina / by Knut J. Olawsky.
p. cm. - (Mouton grammar library ; 37)
Includes bibliographical references and index.
ISBN-13: 978-3-11-019020-5 (cloth : alk. paper)
ISBN-10: 3-11-019020-6 (cloth : alk. paper)
1. Urarina language - Grammar. I. Title.
PM7226.053 2006
498-dc22
2006032102

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

ISBN-13: 978-3-11-019020-5
ISBN-10: 3-11-019020-6
ISSN 0933-7636

© Copyright 2006 by Walter de Gruyter G m b H & Co. KG, D-10785 Berlin.

All rights reserved, including those of translation into foreign languages. No part of this book may
be reproduced or transmitted in any form or by any means, electronic or mechanical, including
photocopy, recording or any information storage and retrieval system, without permission in writing
from the publisher.
Printed in Germany.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
Contents

xv
Abbreviations
Map of Peru and Urarina territory

1 Introduction 1
1.1 Linguistic profile 1
1.2 Geographical location 2
1.3 Genetic affiliation 5
1.4 The current state of the language and its endangerment 10
1.5 Summary of social organisation 11
1.5.1 Means of subsistence 11
1.5.2 Material culture and daily life 12
1.5.3 Ceremonial life and spirituality 13
1.6 Database and language resources 16
1.7 Terminology and conventions 18
1.7.1 Transcription 18
1.7.2 Organisation of examples 23
1.8 Acknowledgements and personal notes 27
2 Phonology 30
2.1 Consonants 30
2.1.1 /t/ 32
2.1.2 Irl 33
2.1.3 /d/ 34
2.1.4 Ibl 35
2.1.5 Ik! 35
2.1.6 /kw/ 36
2.1.7 Is/ 37
2.1.8 /// 38
2.1.9 /tg/, [d3] 39
2.1.10 /h/ 40
2.1.11 /fw/, [hw] 42
2.1.12 /hj/ 44
2.1.13 /l/ 45
2.1.14 /n/ 46
2.1.15 /m/ 46
2.1.16 /jV 47
2.1.17 The status of [j] and [w] 48
2.2 Phonotactics 49
2.3 Vowels 51
2.4 Vowel length 55
2.5 Vowel sequences 58
2.6 Diphthongs 60
2.7 Nasalised vowels 63
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
vi Contents

2.8 Automatic phonological alternations 63

2.8.1 Insertion of transitional glides 64
2.8.2 Assimilation of /«/ 66
2.8.3 Vowel copying 67
2.8.4 Vowel raising 70
2.8.5 Simplification of /aW 70
2.8.6 Affricatisation /ri/ ~ [rj] ~ [CI3] 71
2.8.7 /aj/~[ej] 71
2.8.8 Insertion of glottal stop [?] 72
2.8.9 Palatalisation o f / n / after /i/ 72
2.8.10 Nasal spreading 74
2.9 Syllable structure 75
2.10 The notion of "Phonological Word" 77
2.10.1 Word length 77
2.10.2 Other properties of words 80
2.11 Clitics 80
3 Morpho-phonology 91
3.1 Alternations involving /k/ versus no onset on suffixes 91
3.1.1 Plural marker - a r u on nouns 91
3.1.2 Plural marker -«ru in verbs 95
3.1.3 Verbal alternation with [k], in the 3ps form 98
3.1.4 Verbal alternation with [k], [tp] in lps forms 99
3.1.5 Verbal alternation with [k], [tc] in the imperative form 102
3.1.6 Verbal alternation with [k] in 2ps forms 104
3.1.7 Stem extension with /he/ 106
3.2 Palatalisation of/r/ after Ν 107
3.3 Vowel lengthening 109
3.4 Merging between final root vowels and vowel-initial suffixes 110
3.5 Alternations with pre-root elements 114
3.5.1 Palatalisation o f / k / to [te] 114
3.5.2 HI [dj] 115
3.5.3 Vowel elision with object proclitics and reflexive prefix 116
3.6 Various rules of shortening or simplification 116
3.6.1 Simplification of /el/ 117
3.6.2 Denasalisation of final -1 117
3.6.3 /e/~[ej] 118
3.6.4 Omission or shortening of unaccented syllables 118
4 Tone 120
4.1 Tonal structure of nouns 123
4.2 Verb Phrase: effects of nouns on the tonal structure of the verb 127
4.3 Suffixes that cause deviations from the predicted tone patterns 130
4.3.1 lsg suffix / - a n « / 130
4.3.2 Imperative form 131
4.4 Noun Phrase: NounBrought
+ Adjective
to you by | Max-Planck-Gesellschaft - WIB6417 134
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Contents vii

4.5 Postpositional Phrase: Noun + Postposition 136

4.6 Overriding patterns 136
4.6.1 Possessive construction 137
4.6.2 Possessive construction with adjective 139
4.7 Sentences: More than one tone 141
4.7.1 Tonal patterns on verbs 141
4.7.2 VS and OVA constructions 143
4.7.3 Clauses with adverbs 145
4.7.4 Focus constructions 146
4.7.5 Serial Verb Constructions 148
4.8 Summary of tonal types 149
5 Word classes 151
5.1 Nouns 152
5.1.1 Proclitics and nouns 153
5.1.2 Nouns and the expression of location 154
5.1.3 Nouns as modifiers 161
5.2 Verbs 166
5.3 Adverbs 169
5.3.1 Adverbs related to time 170
5.3.2 Adverbs related to place 176
5.3.3 Adverbs related to manner 176
5.3.4 Adverbs related to degree 178
5.3.5 Epistemic adverbs 179
5.3.6 Other morphemes with adverbial function 182
5.4 Adjectives 187
5.4.1 Adjectival verbs 188
5.4.2 Adjectival nouns 193
5.4.3 Multifunctional word "kauatga" 201
5.4.4 Summary of word class assignment for adjective-like words 203
5.4.5 Sequences of noun modifiers 206
5.4.6 Comparison 207
5.5 Pronouns 213
5.6 Demonstratives 216
5.6.1 Function of demonstratives 217
5.6.2 Locational demonstratives 219
5.7 Postpositions 224
5.7.1 Locational postpositions 225
5.7.2 Postpositions with locational and temporal function 234
5.7.3 Postpositions with temporal function 236
5.7.4 Postpositions with other functions 238
5.8 Conjunctions 245
5.8.1 Temporal conjunctions 247
5.8.2 Other conjunctions 252
5.8.3 Coordinating conjunctions 256
5.9 Interrogatives 258
5.10 Clause introducers 259
5.10.1 "Strong" prohibitives "nihjauria", "hjauijte" 'don't' 260

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
viii Contents

5.10.2 "Weak" prohibitive "kwa" 'don't' 262

5.10.3 Negative "kwatia" 'not' 263
5.10.4 Hortative "kwane" 'let's' 264
5.10.5 Apprehensive "nabana" 'so that not' 265
5.10.6 Conditional "baana" ' i f 266
5.10.7 Negative question introducer "ta" 266
5.10.8 Suggestive "djatera" 267
5.11 Quantifiers 267
5.11.1 Quantifier "itulere" 'various (kinds of)' 268
5.11.2 Quantifier "laurilatiri" 'various' 269
5.11.3 Quantifier " t a b a w a " 'some' 270
5.11.4 Quantifier "tabal" 'much', 'enough' 270
5.11.5 Quantifier " k u r u a t a h e i " 'little', 'few' 271
5.11.6 Quantifier "lejhijiehei" 'not even one' 271
5.11.7 Quantifier "arahfl" 'various', 'several', 'many' 271
5.11.8 Quantifiers "hitarfi", "satfi" 'all' 272
5.12 Numerals 275
5.13 Interjections 282
5.14 Ideophones 283
5.15 Particles 284
5.15.1 Certainty emphasiser "tpa" 284
5.15.2 Emphasiser "hjä" 'just' 286
5.15.3 Contrastive emphasiser "läe" 287
5.15.4 Particle "tQHisi" 290
5.15.5 Pity marker "noane" 290
5.16 Word classes and inflection 291
5.17 Summary 296
6 Noun Phrase structure 298
6.1 Demonstratives 299
6.2 Numerals 300
6.3 Possessives 301
6.4 Quantifiers 303
6.5 Nominal and noun-like modifiers within the NP 308
6.5.1 Adjectives in the NP 309
6.5.2 Adjectival nouns that can function as nominal head 310
6.5.3 Deverbal adjectives 313
6.5.4 Other modifying nouns 315
6.5.5 More than one adjectival modifier 319
6.6 Relativisation 320
6.6.1 Nominalisation with "-i" 321
6.6.2 Nominalisation with "-era" 322
6.6.3 Nominalisation with "-naa" 323
6.6.4 Constructions with "-nakatira" 324
6.6.5 Relative clauses within the NP 325
6.7 Involvement of other word classes as modifiers 327

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Contents ix

6.8 Summary 328

7 Possession 331
7.1 Possession in the NP 331
7.1.1 Possessive construction without marker 333
7.1.2 Possessive marking with "raj" 333
7.1.3 Possessive marking with proclitics 337
7.1.4 Multiple possessive marking 339
7.1.5 Summary 340
7.2 Possession with verbs 341
7.2.1 Copula strategy 341
7.2.2 Possession with 'have' 342
7.3 Possessive derivation 343
7.3.1 Verb suffix "-oka" 343
7.3.2 Prefix b- 345
7.3.3 Associative prefix "ku-" 347
7.4 Inalienable possession 350
8 Number marking 355
8.1 Verb plural for 2ps 356
8.2 Plural marker for 3pl 357
8.3 Plural objects 360
8.4 Plural "-ana" on stative verbs 362
8.5 Distributive plural 362
8.6 Dual and lpl forms 364
8.7 Number agreement 365
8.8 Plural agreement in the NP 367
8.8.1 Numerals and quantifiers 367
8.8.2 Other noun modifiers 369
8.8.3 Plural agreement with relative constructions in modifier function . 371
8.9 Plural marking on other categories 373
9 Verb classes 375
9.1 Transitive verbs 376
9.2 Intransitive verbs 377
9.2.1 Active intransitive verbs 377
9.2.2 Stative intransitive verbs 378
9.2.3 Verbs of posture, shape, and colour 379
9.3 Reflexivity 385
9.4 Ambitransitivity 385
9.5 Ditransitivity 386
9.6 Copula 391
9.6.1 Identity or equation 392
9.6.2 Attribution 394
9.6.3 Locative 394
9.6.4 Existence or temporary presence 396
9.6.5 Possession 397
9.6.6 Auxiliary 397
9.7 Auxiliary "aja" 399
9.8 Verbless clauses 401

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
χ Contents

9.9 "Special" verbs 403

10 Word formation 404
10.1 Word-class changing derivation 404
10.1.1 Nominalisation 404
10.1.2 Verbalisation 406
10.2 Other types of verbal derivation 407
10.2.1 Size and number derivations 407
10.2.2 Derivations of degree 412
10.2.3 Derivations of manner 415
10.3 Reduplication 418
10.4 Compounding 425
11 Person inflection classes 431
11.1 D-form 434
11.2 E-form 435
11.3 A-form 441
11.4 Discussion 445
11.5 Conclusion 453
12 Verbal morphology 456
12.1 Prefixal positions 459
12.2 Suffixal positions 459
12.2.1 Suffixal slot 1: Causative-1 459
12.2.2 Suffixal slot 2: Causative-2 460
12.2.3 Suffixal slot 3: Impossibility 461
12.2.4 Suffixal slot 4: Aspect-1: Continuous 462
12.2.5 Suffixal slot 5: Impersonal passive 465
12.2.6 Suffixal slot 6: Aspect-2: Habitual 465
12.2.7 Suffixal "zone" 7: Distributive, Plural object 468
12.2.8 Suffixal slot 8: Velocity 471
12.2.9 Suffixal slot 9: Diminutive/Counterexpectation 474
12.2.10 Suffixal slot 10: Completive aspect 477
12.2.11 Suffixal slot 11: In-law talk 478
12.2.12 Suffixal slot 12: Plural-1 479
12.2.13 Suffixal slot 13: Epistemic modality: Probability 480
12.2.14 Suffixal slot 14: Mood/Future tense 481
12.2.15 Suffixal slot 15: Negation 484
12.2.16 Suffixal slot 16: Person 487
12.3 Enclitic positions 493
12.3.1 Slot 18: Politeness 493
12.3.2 Slot 19: Assertive 494
12.3.3 Slot 20: Witness evidential 496
12.3.4 "Zone" 21: Reportative evidential; remoteness 498
12.3.5 Slot 22: Reassurance 503
12.3.6 Slot 23: Negative, Interrogative 505
12.3.7 "Zone" 24: Attitude/Emotion, Rhetorical question 508
12.4 Other morphemes in postverbal position 516
12.5 Discussion 522
13 Politeness 530

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Contents xi

13.1 Kinship terms 530

13.2 Politeness with honorific enclitic 534
13.3 Politeness with in-law marker "-ana" 540
13.4 Alternative morphological politeness strategies 543
13.4.1 Polite function of diminutive suffix "-hee" 543
13.4.2 Polite function of velocity suffix "-uri" 546
13.5 Alternative lexical politeness strategies 547
13.5.1 Polite function of "karaj" 'for me' 547
13.5.2 Polite function of "atfl" 'nevertheless' 548
13.6 Summary 550
14 Negation 552
14.1 Clause negation 552
14.1.1 Negation and verb classes 553
14.1.2 Scope of negation 555
14.2 Adverbial function of negative copula 555
14.3 Interrogative pronouns and negation 556
14.4 Negation and the lexicon 558
14.4.1 Negative introducers 558
14.4.2 Negative interjections 559
14.4.3 Negative conjunction 562
14.4.4 Negation and idiomatic expressions 562
14.4.5 'Not yet' and 'Not any more' 563
14.4.6 'Without' 564
14.5 Negative questions 565
14.6 Multiple marking of negation 567
15 Imperative 569
15.1 Positive imperative 569
15.1.1 2ps imperative 569
15.1.2 Hortative ( l p s imperative) 573
15.1.3 Jussive (3ps imperative) 576
15.1.4 Lexical imperative with "d3ahaa" 578
15.2 Prohibitive 579
15.2.1 Prohibitives with " n i h j a u r i a " and " h j a u i j r e " 580
15.2.2 Prohibitive with "kwa" 581
15.3 Consecutive function of imperative 583
15.4 Imperative and future reference 585
15.5 Suggestive form 587
15.6 Other aspects regarding imperatives 588
15.7 Differences between imperative and declarative mood 589
16 Valency changing mechanisms 592
16.1 Passive 592
16.1.1 Modifier function of "-noi" 595
16.1.2 The role of the copular auxiliary 596
16.1.3 Argument structure 597
16.1.4 Further observations 600
16.2 Intransitiviser "ne-" 601

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
xii Contents

16.2.1 Reflexive function of "ne-" 603

16.2.2 Functions of "raaheniane" 'self 606
16.3 Reciprocal form with "ita" 607
16.4 Causative 609
16.4.1 Causative with "-erate" 609
16.4.2 Causative with "-a" 616
16.4.3 Differences between the two causative forms and their combination
618
16.4.4 Additional strategies to express causativity 621
16.5 Valency increase with "ke" 622
17 Serial verb constructions 629
17.1 Properties of VI 630
17.2 Properties of V2 633
17.3 Semantic structure of SVCs 638
17.4 Further properties of SVCs 643
17.4.1 Embedding in larger constructions 643
17.4.2 Contiguity 644
17.4.3 Order variation 645
17.4.4 Number of verbs involved in SVCs 646
17.4.5 Wordhood 647
17.4.6 Difference to other multi-verb constructions 648
17.5 Discussion and summary 650
18 Constituent order 652
18.1 Transitive clause 654
18.2 Intransitive clause 658
18.3 Deviating order types 660
18.4 Adjuncts 670
18.4.1 Position of postpositional phrases (PPs) 670
18.4.2 Position of adverbs 677
18.4.3 Position of locational adjuncts 681
18.4.4 Summary 681
18.5 Position of dependent clauses 682
19 Focus 692
19.1 Functions of focus markers 697
19.2 Focus marking and clause types 703
19.2.1 Imperative clauses 703
19.2.2 Questions 705
19.2.3 Focus and dependent clauses 709
19.2.4 Negative clauses 710
19.3 Scope of focus 712
19.3.1 Focus-marking of core arguments 712
19.3.2 Focus-marking of adjuncts 715
19.3.3 Focus marking in other environments 721
19.4 More than one constituent preceding the focus marker 724
19.5 More than one focus marker in a sentence 727
19.6 Summary 732
20 Multi-clause constructions 733

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Contents xiii

20.1 Subordination 734

20.1.1 Subordination without conjunction 735
20.1.2 Subordination with conjunction 743
20.1.3 Non-finite verb plus conjunction 754
20.1.4 Alternative strategies of subordination 756
20.2 Complement clauses 758
20.2.1 Same subject complement clauses 758
20.2.2 Different subject complement clauses 767
20.2.3 Special control verbs 776
20.2.4 Complementation strategies 779
20.2.5 Summary of complementation types 786
20.3 Participle clauses 788
20.3.1 Morphological features 789
20.3.2 Compatibility 791
20.3.3 Scope 792
20.3.4 Semantic functions 795
20.3.5 Syntactic properties 800
20.4 Other non-finite verbal suffixes 802
20.5 Clause coordination 802
20.5.1 "t£äe" 'also' 803
20.5.2 "nete" 'but' 806
20.6 Other strategies of coordination 808
20.7 Multi-clause constructions in comparison 811
21 Questions 814
21.1 Content questions 814
21.1.1 "d3a"'what','who' 815
21.1.2 "d3a" with other postpositions 818
21.1.3 "d3arihi"'which' 819
21.1.4 "d3anö'"why' 820
21.1.5 "d3atoanei" 'how' 821
21.1.6 "d3asfi'"how' 822
21.1.7 "d3abana"'when' 823
21.1.8 "d3«"'where' 823
21.1.9 "d3aelöo" 'what way' 824
21.1.10 'How much','how many' 825
21.1.11 Indefinite function 827
21.1.12 Indefinite emphasiser 830
21.2 Polar questions 831
21.3 Negative questions 834
21.4 Rhetorical questions 835
21.5 Indirect questions 836
21.6 Interrogative verbs 837
21.7 Summary 839
22 Discourse strategies and expressive elements 840
22.1 Phonological peculiarities 840

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
xiv Contents

22.2 Morphological features 842

22.3 Syntactic structures and discourse 844
22.3.1 Ellipsis 844
22.3.2 Iconic repetition 845
22.3.3 Long sentences 846
22.3.4 Clause chaining 847
22.4 Reference tracking 849
22.4.1 Overt NPs 850
22.4.2 Demonstrative "nii" 850
22.4.3 Verbal marking 851
22.4.4 No marking 852
22.4.5 Conjunctions "hana" and " h a a " 855
22.4.6 Introduction of quotations 857
22.5 Other features 860
22.6 Observations on non-narrative genres 864
23 Variation 866
23.1 Dialectal differences 866
23.1.1 Postvocalic aspiration 868
23.1.2 Vowel copying 873
23.1.3 Very long vowels 873
23.1.4 Diphthongisation 874
23.1.5 Monophthongisation and vowel elision 875
23.1.6 Lowering of [eeä] to [aaä] 876
23.1.7 Variations between /r/, /d/, and l\l 876
23.1.8 Consonant gemination 876
23.1.9 Plural "-kara" instead of "-tpara" 878
23.1.10 Pronunciation o f ' a little' 878
23.1.11 Summary of common dialectal features 879
23.1.12 Some dialect-specific features 881
23.2 Traditional and innovative language 885
23.2.1 Changes in phonology 886
23.2.2 Changes in morphology 886
23.2.3 Changes in the lexicon 888
23.2.4 Changes in syntax 892
23.3 Language endangerment and Hispanicisation 893

Appendix A: Texts 901

Appendix B: List of scientific names 929
Appendix C: Plates 933

References 935
Index 94 \

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
Abbreviations

Abbreviations in text and tables

DS Different subject (complement clause)

EL Elicitation
EL Elicited example
intr. Intransitive Verb
NP Noun Phrase
NT New Testament
PSC verbs Verbs of position, shape, or colour
$2, $3 Second/third syllable of a word
SS Same subject (complement clause)
ss Same subject (complement clause
SVC Serial Verb Construction
SVC Serial verb construction
tr. Transitive verb
VI, V2 Verb 1, Verb 2 (in SVC chapter)
VP Verb Phrase

Abbreviations in interlinearisation

[A] Subject of transitive clause [syntactic functions]

[ADJ] Adjective [syntactic functions]
[ADV] Adverb [syntactic functions]
[CC] Copula complement [syntactic functions]
[CNJ] Conjunction [syntactic functions]
[CS] Copula subject [syntactic functions]
[DEM] Demonstrative [syntactic functions]
[DEP] Dependent clause [syntactic functions]
[E] External argument (with postposition he) [syntactic functions]
[EXPR] Idiomatic expression [syntactic functions]
[INTR] Introducer [syntactic functions]
[MOD] Modifying noun [syntactic functions]
[N] Noun [syntactic functions]
[0] Object of transitive clause [syntactic functions]
[PART] Particle (syntax only) [syntactic functions]
[POSSR] Possessor [syntactic functions]
[PP] Postpositional phrase [syntactic functions]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
xvi Abbreviations

[QUOT] Quotation [syntactic functions]

[REL] Relative Construction (syntactic structure)
[S] Subject of intransitive clause [syntactic functions]
[V] Verb [syntactic functions]
-[V] Vowel lengthening
1 pi, 2pl, 3pl First/second/third person plural; verbal suffixes or free pronouns
lpl/du First person dual (verbal suffixes or free pronouns)
lpl/ex First person plural exclusive (verbal suffixes or free pronouns)
lpl/in First person plural inclusive (verbal suffixes or free pronouns)
lps, 2ps, 3ps First/second/third person unmarked for number (verbal suffixes;
sometimes the singular forms constitutes the base for plural
lsg, 2sg, 3sg First/second/third person singular; verbal suffixes or free pronouns
lsg/A, 3ps/E, ... Capital letters after person marking indicate the person inflection type
of the suffix (three classes: A/E/D type)
ADVRS Adversative (discourse marker niki)
AG Agentive Nominaliser/Relativiser -era
ASC Associative (noun prefix ku-)
ASCM Associative Modifyer (noun prefix b-)
ASS Assertive (verbal enclitic =ni, usually with Future tense)
AUX Intransitive Auxiliary (verbal root aj)
CAU1 Causative (verbal suffix -a)
CAU2 Causative (verbal suffix -erate)
CEXP Counterexpectation (particle tguisi)
CHNG Emphatic "Change" of situation (particle lae)'
CMP Comparative (verbal suffix -ni)
CND Conditional/Temporal (clause enclitic =ne)
CNT Continous Aspect (verbal suffix -ahe)
CPL Completive Aspect (verbal suffix -si)
CRTN Emphatic "Certainty" (particle tg,u)
DER Undefined Derivational suffix
DIM Diminutive (verbal suffix -heeka)
DIR Directional (noun suffix -heloo)
DSP Desperate Interrogative (verbal suffix -naare)
DSTL Distal (verbal suffix -ni)
DSTR Distributive Plural (verbal suffix -ahoa)
EMF Final Emphatic (clause enclitic =ra)
EMPH Emphasis (unspecified)
ENUM Enumerative verb (naaohwaa)
ERD Emphatic Respect Discourse (particle üra)
FEAR Fear (clause enclitic =naate)
FOC Focus (enclitics =ne, =na, =te)
FRS Frustrative/Negative Attitude (enclitic =ta)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Abbreviations xvii

GPF Gap filler (kaanii, also meaning 'sort o f )

H, L High/Low Tone
HAB1 Habitual-1 (verbal suffix -reheto)
HAB2 Habitual-2 (verbal suffix -nahaauka)
HORT Hortative (verbal suffix -e)
HRS Hearsay Evidential (particle hetau)
ILT In-law talk (verbal suffix -ana)
IMP Imperative (verbal suffix -u)
IMPS Impersonal passive (verbal suffix -era; usually with negation)
INF Infinitive/Same Subject Complement (verbal suffix -na)
INST Instrumental (postposition ke)
INT Interrogative (clause enclitic =na)
INTS Intensifier (verbal suffix -to)
IRN Irrealis Nominaliser (verbal suffix -ene)
IRR Irrealis (verbal suffix -re)
ITR Intransitiviser (verbal prefix ne-)
JUSS Jussive (verbal suffixes -iße, -mit)
LOC Locative suffix (on demonstratives: -ley, on nouns (mostly): -«)
LPP Locative Purposive (verbal suffixes -naha, -naanaha)
NDF Non-Distant Future (verbal enclitic =ni; identical to WIT)
NEG Negation (verbal suffixes)
NEGF Final Negative (clause enclitic =ne)
NEGQ Negative Question Introducer ta
NEW New/Innovative/Contemporary language, spoken by younger people
NOM Nominaliser (verbal suffix -naa)
NOMotj Nominaliser/Relativiser of object (verbal suffix -i)
NOM Sbj Nominaliser/Relativiser of intransitive subject (verbal suffix -i)
NQF Numeral quantifying suffix -he
NTR Neutral Form (verbal suffix: -a)
OLD Old/Traditional language; spoken by older people
PASS Passive (verbal suffix -noi)
PL Plural (verbal suffixes)
PLN Place name
PLO Plural Object (verbal suffix -naha)
PLT Polite/Honorific (enclitic =tQe)
POSS Possessive marker raj
PRB Probability (verbal suffix -kaj)
PRIR Purposive Irrealis (verbal suffix -nenu)
PRT Participle (Verbal suffix -i)
PRV Privative (verbal suffix -elanaala 'without')
PSB Possibility (verbal suffix -naka\ usually with negation)
PSN Person name

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
xviii Abbreviations

PTY Pity Marker (particle noane)

PURP Purposive Complementiser haanu
RAP Rapid Velocity (verbal suffix -uri)
REAS Reassurance (verbal enclitic =tau)
REC Reciprocal (particle it a)
RED Reduplicant (of verb root)
REM Remote past/future (verbal enclitic =/«)
REP Reportative (enclitic -he)
RHT Rhetorical question (clause enclitic =te)
STPL Stative plural (verbal suffix -ana)
SUB Subordinate (clause enclitic =ne)
SUG Suggestive (verbal suffix -riu)
TRN Tribal Name or Affiliation
VBL Verbaliser (noun suffix -oka)
VLI Valency increaser (postposition ke)
WIT Experience/witness evidential (enclitic =ra)
WRN Warning/Threat (verbal enclitic =naare)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
Map of Peru
[The shaded area marks the approximate location of Urarina territory.]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
1 Introduction

1.1 Linguistic profile

Urarina is an isolate spoken by less than 3,000 people in the area of the Rio
Chambira, Loreto Province, North-western Peru. The language exhibits a range
of unusual grammatical characteristics that are rarely or not at all found in other
languages.
One of the most remarkable properties of Urarina grammar is its pervasive
constituent order OVA/VS, which was classified as "non-existing" by
Greenberg (1963). At a later stage, a handful of Amazonian languages (mostly
from the Cariban family) were discovered which exhibit this constituent order
under certain conditions. However, Urarina probably represents the most
convincing example for this atypical syntactic structure, as is investigated in
§18.
Another feature unique to Urarina is the existence of a three-way distinction
for person marking on all verbs, as discussed in §11. Every verb can potentially
be person-marked according to three different paradigms. The choice of the
respective paradigm depends on a complex set of syntactic and pragmatic
conditions, which are investigated in detail. Among the relevant features for the
selection of the person inflection class are clause type and text genre, as is
further supported by statistical evidence.
Word classes, as described in §5 are a further point of particular interest.
Only few members of word classes other than nouns and verbs are underived or
represent homogeneous classes. Numerals, for instance, are split into native
ones, which are inflected as verbs, and loans from Quechua, which are
morphologically treated like nouns. In a similar way, there is no uniform class
of adjectives: again, most of them can be divided into special types of nouns
and verbs. In contrast, there is a closed class of clause introducers that share
certain syntactic features. These can be divided into optional and obligatory
types, which mark specific clause types, such as negative questions, prohibition,
or others.
The phonology exhibits a comprehensive system of alternations. Of
particular interest are certain morpho-phonological rules that involve both
segmental and prosodic factors: some person markers, for instance, only occur
on verb roots that have a specific number of syllables and end in particular
sequences of segments (cf. §3). The complex pitch-accent system of Urarina, as
analysed in §4 is characterised by alternations observed between different word

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
2 Introduction

classes in specific syntactic environments, can be characterised as most unusual.

Specifically, the tonal structure of Urarina suggests that there is a connection
between syntax and tone.
The polysynthetic verbal morphology of Urarina fits well into the
Amazonian context, as lengthy sequences of verbal morphemes to express a
variety of grammatical categories are common (cf. §12). In total, there are 24
suffixal and enclitic slots that can be filled by markers for core grammatical
concepts such as person, number, polarity, and aspect, but which also host
morphemes for velocity, assertion, probability, politeness, and evidentiality, to
mention a few. One remarkable aspect of this system is that several of these
morphemes exhibit order variation.
An intriguing as well as challenging factor for the documentation of Urarina
is the rapid pace at which the language is changing, as is described in §23.
Unusually high degrees of variation between the older and younger generations
are observed at nearly all levels of the grammar. There is a strong trend for
simplification and loss of features, such as the neutralisation of contrast
between former phonemes, the loss of an enclitic originally expressing negative
emotions, but also the introduction of new elements under the influence of
Spanish (cf. §12.3.7). I will refer to the language predominantly used by the
younger generation as "innovative" or "contemporary".

1.2 Geographical location

The main speaker area of Urarina is the Rio Chambira and its tributaries in the
Urarinas District, Loreto Province of Peru. As shown in the map in (1), the
language was originally spoken between the Urituyacu and Corrientes Rivers.
While it is reported that Urarina has been largely replaced by Spanish on the
Urituyacu River, it is still spoken on the Corrientes River, where it is subject to
increasing pressure by Spanish. As an estimate, the number of speakers is
between 2,000 and 3,000 (cf. 1.4).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Geographical location 3

(1) Urarina speaker area: overview with names of rivers and tributaries

The main speaker area nowadays is on the Tigrillo River (including Espejo and
Rayayacu) and the Chambira River and its tributaries. As can be seen in the
map in (2), the largest number o f communities (indicated by black squares) are
found in the lower area o f the Chambira, including the Patoyacu tributary. The
lower parts o f the Tigrillo and Chambira Rivers, however, also host most o f the
Mestizo and mixed communities (indicated by white squares). This grammar o f
Urarina is mainly based on fieldwork in Nueva Union, at the Espejo tributary,
which flows into the Tigrillo River.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
4 Introduction

(2) Chambira River basin and communities

Due to the remote location, access to the Urarina territory is rather difficult. A
single trip from Iquitos, the capital of Loreto District, to the mouth of the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Genetic affiliation 5

Chambira River takes up to two days by ship (lancha - a cargo ship which also
takes passengers). From there, it is another day by peque-peque (a small boat
powered by simple motor which is connected to a steel rod with a propeller) in
order to get to Nueva Union de Espejo. The community has a population of less
than 100 (which equals about 15 households) and is in relative isolation from
other places. As the village is surrounded by almost impenetrable swamps,
outsiders hardly find their way to the place and easily get lost, especially in the
rainy season. The community owns a boat, which is used sporadically to
transport crops to the Maranon River, from where they are shipped to Iquitos, in
order to sell them. Access to Espejo areas upriver is only possible by canoe.
The Urarina language is also known under a variety of other names. Most
importantly, the people traditionally refer to themselves as katga 'man', and to
their language as kat$a ere 'man's word'. However, this name is being replaced
by the term Urarina, whose origin is unknown, but there is a traditional
narrative that accounts for the introduction of this term. According to this story,
which has certain funny connotations, a Jesuit priest who first made contact
with the people (katQa) was presented with urari roots as a gift (about this
staple food, see §1.5.1) - and subsequently called them "Urari-na" (cf. Olawsky
2002:214ff). It is unclear whether this anecdotal account can be taken seriously.
Other names by which the Urarina have been known include Shimacu, and
Itucale, which are not in use any longer. Shimacu (or Simacu and other
variants) is understood as a depreciative term and according to Manus (2005
p.c.), could also be based on the Spanish term cimarron for 'salvage'. The name
Itucale (or Itucali and other variants) must have been introduced by error as it
describes the name of a subgroup or clan of Urarina called itokivari (cf. §1.5.3).

1.3 Genetic affiliation

The genetic affiliation of Urarina is a controversial issue, partly due to the fact
that veiy little is known about the language's history and about its neighbours.
Urarina has been classified as a member of at least four different languages
families by various authors (Panoan: Velasco 1960:v3, 208; Tupian: Figueroa
1986:253; Macro-Tucanoan: Shell & Wise 1971:14; Andean: Greenberg
1960:794 - all cited in Dean 1994(1):37). A comprehensive summary of the
various approaches towards a classification of Urarina is given in Cajas Rojas &
Gualdieri (1987:16ff). However, based on lexical and grammatical comparison
with the small amounts of data on languages supposedly related to Urarina, it is
difficult to find similarities of Urarina with any of the language groups
mentioned in the literature to date. While the term "unclassified" is one way to
deal with this difficulty, there is evidence to suggest that Urarina ought to be
described as an isolate - a solution also adopted by Cajas Rojas & Gualdieri

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
6 Introduction

(1987:19), who provide a study of Urarina phonology. One argument in favour

of assigning an isolate status to the language is the total lack of lexical overlap
with any of the languages spoken in the wider vicinity. This can be based on
comparative evidence from word lists provided by Loukotka (1968) and
Tessmann (1930). Specifically, the following languages are or were spoken in
the wider vicinity of the Urarina speaker area:
- Murato (a.k.a. Candoshi), an unclassified language perhaps distantly related to
Arawak
- Omurana (Zaparoan or isolate) - extinct
- Iquito (Zaparoan)
- Jebero (Cahuapanan)
- Cocama (Tupi)
- Yameo (Peba-Yaguan) - extinct

It should be noted that "wider vicinity" is not meant to imply actual proximity
to the Urarina territory in a strict sense. Given the unhospitable geophysical
conditions of the rainforest, a trip to any of the groups mentioned (except for
Omurana), would involve several weeks of travel by foot and canoe. As a
guideline, a canoe trip from the mouth of the Chambira up to Mangual may take
between two and three weeks. The map in (3) can only give a very approximate
location of the different languages and their distance from the Urarina territory
is not to scale.

(3) Languages surrounding the Urarina territory (approximate location - not to scale)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Genetic affiliation 7

The following comparison of a few key terms in some neighbouring languages

of Urarina is based on Loukotka (1968:118; 153-160).1 Remarkably, Urarina
words do not even exhibit a single similarity between any lexical items for any
language listed here.

(4) Lexical comparison of some key terms according to Loukotka (1968)

English Urarina Murato Iquito
'head' kutan modzik p-anak
'tooth' kata
'ear' irit£u
'eye' nutä k-achik pui-nami
'hand' bihi
'fire' isi sumaadzi inämi
'water' akau
'sun' enoto shäri niinami
'star' arasihje tsangächi narexa
'house' lureri pangozi ita
'woman' eene kisa muesaxi
'maize' katuri iwäto shakärok
'white' somahäj kantsirpi mosotin
'one' lejhii minumpta noki
'two' kuruatahäj tsimboro ko:mi
'three' nit^atahaj

English Cocama Jebero Yameo

'head' yakö moto wi-näto
'tooth' dzäi lätek wi-e
'ear' nämi wuioga wi-tiwe
'eye'
'hand' püwa
'fire' täta pön öle
'water' uni
'sun' köki natera
'star'
'house'
'woman' wäina

1
The transcription of Urarina follows the conventions used elsewhere in this grammar,
whereas the transcription of other languages is adopted from Loukotka.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
8 Introduction

English Cocama Jebero Yameo

'maize' abati chiter ogung
'white'
'one' wipi ala ? atsa pwiter
'two' mokoika katäta narämue
'three' motsapwöka käla pwiterorineo

Further four languages originally spoken in the wider vicinity are now extinct.
The table in (5) is based on wordlists by Tessmann (1930; his transcription is
adopted below). Note that on Pinche/Taushiro no data is available.
- Andoa (Zaparoan) - extinct
- Cahuarano (Zaparoan) - extinct
- Pinche/Taushiro (possibly Zaparoan or Candoshi) - extinct (one speaker left
in 2004)
- Pano(bo) (Panoan/unclassified) - extinct
- Chamicuro (Arawak) - extinct

(5) Lexical comparison of some key terms according to Tessmann (1930)

English Urarina Cahuarano Andoa Pano(bo) Chamicuro

'head' kuturi pa-nak anäke mäpu okäski
'tooth' katu puö-ka kixä seta awäxsi
'ear' iritpu pio-toko nomäko pawiki axcäi
'eye' nutä poi-nami nemiz buero awoxkui
'hand' bihi pio-ködika awuäsi muiskä ucixpa
'water' akaa mowäka ompäsko unixsa
'sun' enoto nianami poanänu wäri mosoxko
'house' lureri dahäpu taping axkoci
'woman' eene itemo maxi äibo molota
'white' somahäj musitena makusini hoso kamadläwa

Omurana (isolate) was the language in closest geographical proximity to

Urarina. Though it is now extinct, both Loukotka and Tessmann provide word
lists for Omurana. Still, a lexical comparison based on their data does not
exhibit any similarity with Urarina.

(6) Lexical comparison with Omurana based on Loukotka (1968:157) and Tessmann
(1930:252ff)

English Urarina (Loukotka) (Tessmann)

'head' kuturi na-nenyalok naneyalok
'eye' nutä an-atn (a)natun

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Genetic affiliation 9

English Urarina (Loukotka) (Tessmann)

'sun' enoto hena hena
'house' lureri ana ana
'woman' eene mparäwan mparäwan
'white' somahäj chalama calama
'fire isi ino injo
'maize' katuri aichia aicia
'one' lejhii nadzora naGoro
'two' kuruatahäj dzo ? ora Gooro

From the lexical items compared in the above tables, it is unmistakable that
none of these languages exhibits any lexical matches with Urarina. Besides, it is
unclear to what extent these groups may have been in contact with the Urarina.
There certainly has been some contact with the Cocama (Tupi), as Urarina
speakers even reported of some cases of intermarriage in the past. At present,
there is no contact between the two languages - and the Cocama language is
struggling despite the relatively large number of ethnic Cocama. Contacts with
some other groups are described as hostile, as these are characterised as
cannibals. There are several old narratives that describe the conflict between
those "indios" (bakaua) and the Urarina, who according to the stories defeat the
attackers in revenge. The Urarina names of these groups are ijioro - which
refers to the Shapra (Murato), fioano (which probably refers to the Pano(bo)),
and kasarena (unidentified - perhaps Cashinahua (Panoan)). Another narrative
mentions a man from the rameseto - referring to the Lamisto or Lama (Munichi
stock), now extinct. The Omurana, - once the group in the closest vicinity to the
Urarina, which even appeared to have a settlement on the Chambira and were
considered non-hostile, were called majßa.
The knowledge about grammatical structures of most of these languages is
very limited. Naturally, a polysynthetic morphology as a general typical feature
of Amazonian languages is also found in the nearby language families. Other
properties such as the presence of evidential markers, or pitch-accent occur all
too commonly in Amazonian languages as to postulate a relation between these
and Urarina. In contrast, for none of the languages listed above, similar
structures such as the properties highlighted in §1 have been reported. Based on
the lexical comparison and on the range of extremely uncommon features found
in Urarina, it is feasible to characterise Urarina as an isolate.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
10 Introduction

1.4 The current state of the language and its endangerment

The exact number of Urarina speakers remains to be determined, as information

about this varies from source to source. It is for certain that the number of
ethnic Urarina, which possibly ranges over 3,000, does not reflect the number of
its speakers, as language loss has been ongoing over several years. The actual
number of fluent Urarina speakers may be estimated at around 2,000. One threat
to the Urarina language and culture is the rising number of mestizo settlers.
These form their own communities or become part of a native community. In
mixed Urarina-Mestizo communities, mestizos typically control most
administrative functions and Spanish becomes the language of communication,
sometimes supported by school teachers who do not speak Urarina. Spanish is
also the language of communication with traders and other groups with
commercial interests who virtually invade the region in order to exploit the
natural resources. Traders also "employ" Urarina as cheap labour in exchange
for alcohol or other items. As a consequence, many Urarina face problems such
as alcoholism and subsequent financial dependency upon outsiders.
Different from what is stated in Olawsky (2002) (which is purely based on
knowledge from the Espejo village of Nueva Union) Spanish also is the
language of education, social, and economic interaction with non-Urarina
speaking population. In "mixed" communities, Spanish has become the
preferred language of communication. One alarming aspect of the language
change is the loss of a number of some grammatical features among speakers of
the younger generation (also cf. §23.2). The linguistic situation is paralleled by
the threat to most cultural practices as many traditional customs have already
been abandoned.
There has been no previous attempt to provide a comprehensive description
of either the Urarina language or the culture of its speakers. Among the few
studies on Urarina grammar is an unpublished thesis by Gualdieri & Cajas
Rojas (1987), which covers aspects of Urarina phonology and gives an
overview of previous attempts to classify Urarina. Another study is a paper by
Cajas Rojas (1989) on nasalisation. There also is an article published in Spanish
by Manus (1992) on subordinate clause and focus markers. Ron and Phyllis
Manus have been involved in Urarina Bible translation and literacy since 1960
and also have produced several unpublished manuscripts on some aspects of the
language. The most recent publications on Urarina are a text collection
(Olawsky 2002) and a few articles (Olawsky 2003, 2005). Apart from
Tessmann's (1930) chapter on the Urarina, cultural aspects have been mainly
studied by Dean (1994-2002), who has published a number of essays on

2 t
The Spanish term mestizo describes the descendants of European and native South
American parents. These are monolingual in Spanish.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Summary of social organisation 11

anthropological characteristics of the Urarina society. Tessmann (1930:268-

282), apart from a wordlist, also gives an overview of some cultural practices,
which, however, cannot all be confirmed for the present.
So far there has not been much material written in Urarina. The orthography
used in these is based on the spelling system used in the Bible translation, which
is in its final stage (to appear in 2005 or 2006). The orthography is based on
Spanish orthography, which makes it easy to read for someone who knows
Spanish. Of course, Urarina has a number of sounds that are hard to represent,
but an approximate pronunciation may be possible using the present
orthography, which is further described in § 1.7.1. There is an Urarina version of
the UN charter of human rights and the educational NGO Su Camino has also
published some literacy material in Urarina. A bilingual collection of traditional
narratives by Olawsky & Arahuata (2004) was distributed to Urarina teachers
for literacy purposes. As a result of a project recently started, there is a bilingual
education programme by the Red Educativa de la Etnia Urarina, which is fully
controlled by Urarina and has started to replace mestizo teachers with bilingual
teachers in most Urarina communities.

1.5 Summary of social organisation

According to oral accounts, the Urarina traditionally populated the Rio

Chambira and its tributaries, originally living in longhouses along the riverside.
The primary vegetation of the area is rainforest, with vast areas of densely
overgrown swampland. The Rio Chambira is the main waterway of the area,
with about ten tributary rivers flowing into it.

1.5.1 Means of subsistence

Most people still subsist on hunting, fishing, and food crops. Nearly all animals
that appear to be big enough to make a meal are eaten. Among the common
game are mammals such as tapirs, pacas, peccaries, monkeys, as well as
crocodiles and birds. Snakes and dolphins are not eaten, but jaguars may be
consumed.
Fishing is another important means of subsistence. There are four different
fishing methods; all types of fishing are carried out from a canoe:
a) Fishing with a net (which is left overnight or longer),
b) Fishing with a spear,
c) Using a fishing line and a hook (originally with the thorn of the garabata
liana),

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
12 Introduction

d) Killing (or stunning) fish by poisoning a fraction of the river - these are then
collected.

Plantains and cassava constitute the main vegetable diet. Cassava (Urarina
laano, also known as manioc, or yuca in local Spanish), is one of the most
important staple crops for the Urarina. Another widely used crop is a tuber
called urari (or variant erwari; Xanthosoma sagittifolium, Araceae family) - a
tuber typically grown next to the house and consumed grilled, fried, or as a soup
ingredient.
Some fruit, such as small banana, papaya, maize, and peanuts are also
known and grown, but less commonly. Most crops are also sold to traders. The
fields where these items are grown can be several hours' walk away from the
village. As most people have chickens, eggs also form a part of the diet.
Cassava beer, know as masato in local Spanish (Urarina barue) is a popular
alcoholic drink made from mashed cassava (and sometimes plantain), which is
left to ferment for up to four days. Various texts of the database show that social
drinking has been common since ancient times. Alcoholism is mainly a problem
in communities that depend on river traders who sell trago (high-volume liquor
made from sugar cane).

1.5.2 Material culture and daily life

The tradition of living in longhouses has been largely abandoned as most people
now live in individual houses. While these are typically built from natural
materials, modern building materials such as zinc roofs and nails are highly
valued. Traditionally, houses have a thatched roof and no walls. The floor,
which is about lm off the ground, is made from the trunk of the white nist palm
(Span, porta, Urarina ahtaraa). Poles are tied with lianas. The average lifetime
of a house is up to 5 years, but the roof, made from branches of the chevon palm
(elele) has to be renewed about every two to three years.
Among the items that are produced by the Urarina are blowguns and darts,
hammocks for infants, sleeping mats, clay pots, wooden spoons, baskets,
necklaces, and bags. Fibre from the chambira palm (risijie; or alternatively, the
aguaje palm, alaa - in swampy areas) is the most important material for all
kinds of weaving. While most handcrafted items are produced by women, men
engage in the fabrication of baskets, weapons, and carved items. This includes
paddles and dugout canoes, which are the main means of transport on the rivers,
often used in combination with walking. Items are usually carried with a
tumpline or rope across the forehead.
The division of labour is not clear-cut, but some tendencies can be observed.
Men are in charge of hunting and all different types of fishing, but women may
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Summary of social organisation 13

also fish with a net or with a fishing hook. Women may also accompany their
husbands on hunting expeditions that last for several days. Both sexes carry out
work on the field, such as harvesting, cutting weeds, or fetch firewood. Splitting
firewood is done by both genders, though more commonly by women. In
preparing a field for cultivation, men fell the bigger trees and clear the brush,
which are then left to dry and burnt. Women are responsible for cooking,
washing, and taking care of the children. Men occasionally smoke meat and
prepare their own meals during hunting trips. Cassava beer (Span, masato,
Urarina barue) is made by the women. A popular way of labour is a so-called
minga (a loan probably from Quechua), a kind of community work for which a
man invites the other people of a village to work on his field. The inviter then
provides masato for the entire working party.
There are strict regulations for the coexistence of men and women not
married to each other. For instance, all recordings of text with female narrators
had to occur under supervision of an adult male family member. Alternatively, a
woman could do the recording. Typically, men and women would eat
separately.
Emotional states are not normally shown in public. One aspect worth
mentioning is the avoidance of publicly showing anger. The verb for 'become
angry' (neheratia; cf (914)) is also used to translate the Spanish word for 'hate'
(odiar).

1.5.3 Ceremonial life and spirituality

Very little is known about the spiritual life and potential ceremonies, as the
Urarina do not easily share their religious traditions and most details of Urarina
spiritual life were not investigated in detail. Interestingly, there are no
ceremonies for marriage, nor have any rituals at death been observed (there are
no formal funerals). In Nueva Union, the dead are buried in a place about ten
minutes from the village, but the graves are not cared for. Marriages are usually
arranged by the parents. If no bride is available in the same village, the father
would take his son on a trip to another place.3 Brothers frequently marry women
from one family, where feasible. Polygamy is not very common, but legitimate.
However, the first wife has to agree to the husband's taking of an additional
wife. In some cases, a man marries several sisters. The marriage age is around
14 for women and at about 16 for men. There is no dowry, but the man simply
moves in to the wife's family and becomes part of it. This also involves
brideservice, as the husband's labour force now belongs to his new family.

3
In contrast to this, Manus (2005 p.c.) have observed that the girl's parents have to find
a suitable man to marry their daughter.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
14 Introduction

Thus, hunting expeditions are often conducted by two (or more, when including
children), involving the head of family and his son-in-law. A more detailed
account of kinship - and some politeness factors related to the kinship system -
is given in §13.1.
Possibly the only popular ceremony among the Urarina is the consumption
of ayahuasca, a hallucinogenic plant well known for its psychoactive properties.
It is a thick-rooted plant with long leaves and small flowers. The beverage is
based on mashed and boiled ayahuasca vines, blended with at least one other
plant. The Urarina attribute healing power and spiritual wisdom to the
consumption of ayahuasca. The exact rules for the ceremony are unknown to
the author, but participants are required to keep a specific diet for about a week
after participating. There are a number of items especially used for the
ayahuasca ceremony, such as ornaments for the leader of the ceremony.
Apparently, a spiritual leader who performs the ceremony (ayahuasquero), does
not generally have a higher status than other community members, apart from
his healing powers.
People also attribute healing powers to certain plants called biri (cyperus sp.;
piri-piri in local Spanish), which are used as medicinal plants. The Urarina treat
all kinds of diseases with these, which are especially planted and whose
application is strictly prescribed, often in combination with specific dietary
measures. There also are biri which are believed to have magical powers such
as to enable hunters to successfully find certain animals (for each animal there
is a separate biri), or for lovers in order to draw the beloved's attention to
himself or herself. Further functions also include fertilising or infertilising
properties.
Naming is a complex topic, but not much information about it is available at
this stage. New-born children receive Spanish names, but usually names are
given only several weeks after birth, apparently because the parents first wait to
see whether the child survives. Family names are handled in the same way as
the Spanish system; i.e. one name is the father's family name whereas the
second one is the mother's name. Most people also have unique nicknames that
are usually based on some special characteristic of the person or some
experience in the past. For example, one girl is called Lanahaj 'red one'
because she was "red" when she was born. Another person is called Airiu
because he moved to the Espejo from the Airico River. Yet another person is
called Buku 'bone' - apparently due to his skinniness. There also used to exist a
naming ritual in which children were given secret names by a shaman.
However, as this custom has been largely abandoned, no further information is
available on this matter, except that these names appear to have no particular
meaning, but are given as novel names by God. Similarly, not all the details are
known about some names that appear to be clan names. I know of seven names
in total, which are arabura, uhidi, lomajßari, itokoari, nafioiano, d^aitge, and

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Summary of social organisation 15

ahjauriara. According to one consultant, many people nowadays do not know

what "clan" they belong to.
Some aspects of the spiritual beliefs of the Urarina are encoded in traditional
narratives. In particular, many of these explain the creation of specific animals
or plants, or they account for certain aspects or circumstances of human life,
such as how women learnt to give birth, or why children welcome their fathers
when they come from a hunting trip, or how the howler monkey replaced the
spider monkey as the "chief of the monkeys". Three particularly important
characters that occur in most of these narratives should be mentioned:
1. kana huaaunera, "the one who created us" can be understood as
corresponding to a creator God. He is generally described as a caring
supernatural being who does good to his creatures. At least one narrative
also mentions the existence of kana huaaunera kalaui, "God's son", who is
the victim of the people's aggression in one story that results in a deluge.
2. kana huaaunera letono - the "envoy of God" is also equalled to Hirikuri by
some story tellers, though it is not entirely clear whether these refer to one
and the same person. He may be regarded as a kind of spirit appearing in
human shape and usually tries to trap people by creating all kinds of
difficulties and treacherous situations. For instance, he is the one who sends
a group of people into dangerous territory in order to retrieve the night
(which is to be created).
3. Lomai is another important character in many traditional tales. While her
exact relation to the other two characters is not clear, she seems to represent
the "Mother God" (cf. Dean 1994:38) - but again it is unclear whether
Lomai is identical to what occurs as kana neba "our mother" in other
narratives. Lomai's status clearly is high above humans, but she lives in
magical terrain with many spiritual beings (such as talking trees). Lomai
typically sets up strict rules and the people's failure to obey her orders then
results in punishment. In most stories, people end up being turned into some
animal or plant because they did not obey orders.

Over two dozen of these traditional narratives were recorded during the research
project on Urarina language and culture. The text selection in Appendix A
contains one short traditional narrative, which accounts for the creation of two
kinds of peccaries. Many narratives mix elements of traditional religion and
Christianity. For example, in one flood narrative, a man called Adam saves the
son of God from the other people. In the same account told by another speaker,
the character is called Noah. Due to obvious resemblance between characters
such as God's son (who is actually equalled with Jesus in one narrative, but not
in others), the boundaries between traditional and innovative spirituality are not
transparent. Nowadays, many Urarina are Christians (catholic and evangelical),

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
16 Introduction

and aspects of both religions have been merged. A more detailed investigation
of Urarina cosmology will be left to future studies. 4

1.6 Database and language resources

This grammar is part of a comprehensive research project on Urarina language

culture conducted over five years. It was initiated by the Research Centre for
Linguistic Typology (Melbourne) in 2000 and later funded by the Endangered
Languages Documentation Programme (London) from mid 2003. During five
separate field trips between one to six months length, a range of texts were
recorded and will be archived. In total, 126 texts were recorded, equalling about
17 hours of speech. While these include a wide range of different genres such as
conversations, picture descriptions, instructions, and sermons, most are
narratives, which can be further divided into several subtypes:
1) Traditional narratives: these typically account for the Urarina worldview in
terms of creation; others are stories that describe the courage of specific
people in the past, most commonly in defeating groups of enemies.
2) Hunting stories: in these, the storyteller describes a personal experience of
his own past.
3) Dramatic events experienced by others: the storyteller describes dangerous or
funny events that happened during his own lifetime - mostly reproducing the
second type.
4) History: typically, these narratives account for the history of a village,
describing how people first arrived and settled. This also includes
descriptions of how the people lived in ancient times, thus not necessarily
including the storyteller himself.

The text selection in Appendix A gives examples for some of these genres. Four
out of the five texts given there are narratives, which corresponds to the
proportion of narratives in the database, but is also based on the fact that other
texts were too lengthy as to be included here.
For all examples discussed throughout the grammar - if not stated otherwise
- it is tacitly implied that they are "natural" in one or another way, i.e. they are
either parts of speech from natural conversation between native speakers or
between the author and native speaker, or they are embedded in a larger text
independently produced by a native speaker. However, it is well-known that not
all of the questions linguists typically have about the structure of a language can
be answered easily from the occurrence of a given construction in text.
Especially, Urarina has such a variety of morphological phenomena that may

4
For a small collection of narratives, also see Olawsky (2002).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Database and language resources 17

occur only a few times in a specific combination in the entire text database. For
some aspects of the analysis, this may not be sufficient in order to make valid
generalisations about the possibilities and impossibilities of a construction, or
about their productivity. Only in such cases, I have looked for suitable examples
in the N e w Testament translation (abbreviated to N T hereafter), which, with its
abundance of about 218,000 words, is a useful source of examples. Some
further details about the N T translation are mentioned below. Where examples
are absent even in the NT, I relied on elicitation of appropriate examples. These
examples are usually not sentences that were simply translated from Spanish to
Urarina, but they result from questions that arise during the analysis of texts or
in conversations. Where elicited data was used (which is marked in the text) I
tried to match the judgement of several speakers. Overall, elicited examples
represent a very small minority. Some examples, such as short utterances,
phonological data, or word formation are based on observation and knowledge
of the language by the author.
As mentioned above, the overwhelming majority of examples are taken from
the database of natural texts, which are composed of different genres. Nearly all
examples are taken from the Espejo dialect (except §23.1 on dialect
differences). A few examples from other dialects in which no relevant
differences to the Espejo dialect apply are employed as well. 5 In total, four main
dialect areas can be identified, which are sketched in §23.1. Some sample
sentences (repeated in different chapters), which are practical for the
investigation of several different aspects of the grammar, sometimes highlight
their structure in different ways.
Where elicited examples or data from the N T are used, this is indicated. The
proportion of all examples adopted from the N T translation is less than 5%, and
used where specific constructions or combinations of morphemes or phrases
were not available from the text database. 6 While I do not regard examples from
translations as instances of natural text, the translation process implies that such
data is reliable and consistent with speech normally produced by Urarina
speakers. The process of the N T translation has undergone constant revision
over the last few decades and was largely conducted by native speakers of
Urarina. The translation is very free rather than literal and all examples are
reviewed by a range of speakers of Urarina. Each sentence, after being
translated from Spanish to Urarina, goes through a back translation into Spanish
by different speakers and amendments are made, where necessary. In the next

5
Where dialectal differences occur, this is indicated.
6
One half of these occur in the chapters on word classes and NP structure, where a
majority of examples consists of single words or phrases. In these chapters, they
supplement natural examples from the text database, typically in order to illustrate
syntactic or semantic variations.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
18 Introduction

step, further revisions to the grammar and style of each example follows. The
entire process involves a variety of native speakers. Thus, it is plausible to
assume that the N T examples are unlikely to be caiques from Spanish or that
they reflect transfer errors in any way. As far as can be predicted at the present
stage, the N T translation appears to be well understood by the Urarina
community.
Statistics regarding the sound inventory and word classes as presented in the
respective chapters are based on the lexical database - a wordlist of 3,650
entries at the time of completion of this grammar. This should represent a sound
basis for statements on frequency and distribution of the categories discussed.
Texts were recorded from 24 different speakers and it is impractical to list
the details for each individual here. A speaker profile for the recordings from
different dialects is provided in §23.1. The general grammatical description is
based on fieldwork in Nueva Union (Espejo River), and a large proportion of
examples are taken from stories told by the village's most elaborate storyteller,
Medardo Arahuata Manizari. He is the second-oldest man of the village (born
around 1952) and a monolingual speaker. Originally coming from the Tigrillo
River (San Lorenzo), he was one of the founders of the community in the
Espejo. Another important consultant, one of the main helpers in grammatical
analysis, is Julian Nuribe Vela, the village elder and representative (born in
1946), who is known in the entire Chambira basin as a sort of (unofficial)
Urarina ambassador. He is bilingual in Urarina and Spanish and did not learn
any traditional narratives from his ancestors, but has a comprehensive
vocabulary of Urarina and good understanding of grammatical structures. His
parents moved from the Urituyacu River to the Chambira River shortly after he
was born. From there, he moved on to various places, including the Tigrillo
River, where he stayed for several years. Then he followed his father-in-law
(Medardo Arahuata's father) to the Espejo and became one of the village
founders. In fact, most inhabitants of the Espejo community are in some way
related to each other and further important consultants include Medardo's
brother, Julian's sons, and a number of people related to either of these as in-
laws. A brief speaker profile for further consultants is given in §23.1.

1.7 Terminology and conventions

1.7.1 Transcription

The presentation of Urarina terms in this grammar is an intermediate solution

between a phonemic and a (broad) phonetic transcription, with a few
peculiarities. On the one hand, the transcription is phonemic in that most of the
predictable, non-essential alternations do not surface in the transcription. For
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Terminology and conventions 19

instance, postvocalic aspiration, nasal spreading, and most simplification rules,

as described in §2 (Phonology), are not included in the transcription. (7a) gives
a few examples for the representation of terms that include these features. On
the other hand, some phonological rules were considered in the transcription, as
listed in (7b). These include morpho-phonological alternations and some other
rules even though these may be entirely predictable. However, some of the rules
underlying these alternations are rather complex and as a general guideline, I
represent examples in a way that is close to their actual pronunciation, while
omitting additional non-essential features that may occur as optional variations.
As a result, Urarina speakers would be able to identify the words, if pronounced
as they are represented here.

(7) E x a m p l e s f o r representation of simplified transcription

a) Features not included in transcription:

Feature Representation in Phonemic Narrow Gloss
grammar structure transcription
Aspiration ate /ate/ [ahte] 'fish'
Affricatisation hauria /hauria/ [haad^a] 'earlier'
Monophthongisation haü /haa/ [häö], [hää] 'because'
Nasal spreading rua /rüa/ [rüä] 'side'
Glide in hiatus enua /enua/ [enuga] 'tree'
/aj/ [ej], [e] raj /raj/ [raj], [rej], [re] 'for'
Optional vowel satii /satii/ [satii], [saatii] 'all'
lengthening
Obligatory vowel kioara-H /kwara-«/ [kwaraaa] 'see'-IMP
lengthening

b) Features included in transcription:

Feature Representation Phonemic Narrow Gloss
in grammar structure transcription
Insertion of [k] hoara-knru-a /kwara-aru-a/ [kwarakurua] •see'-PL-3ps/A
(morpho-
phonology)
Palatalisation ari-ßa /ari-na/ [arijia] 'seek'-INF
/«/ [u], [0] eno-oru-a /enu-ara-a/ [enoorua] 'enter'-PL-
3ps/A
Vowel copying kihja /kiha/ [kihja] 'paddle'
Vowel raising ti-a /te-a/ [tia] 'give'-3ps/A

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
20 Introduction

With regard to rules that are subject to a high degree of variation, I have given
the examples in accordance with the original recording. This would sometimes
result in individual differences, such as the realisation of vowels that have
allophones in free variation, or the realisation of certain consonants in some
words, which may be subject to variation due to optional rules. These include
the variation between [u] and [o], [hw] and [fw], [r] and [d] or [t]. All variations
are also clearly identified in §3. Most variations depend on differences between
individual speakers, some on dialectal differences, but a few may also occur
speaker-internally. Note that all examples except for §23.1 on dialects are from
dialect zone A - Espejo (where some speakers may exhibit strong influence of
the Tigrillo dialect due to their heritage). Some examples are illustrated in (8).
These include phonological variations and lexical differences between speakers
or sometimes speaker-internally.

(8) Examples for representation of varying transcription

a) Phonology:
Variation Representations Phonemic Narrow Gloss
in g r a m m a r structure transcription
[u] ~ [0] omari, umari /umari/ [omari], [umari] 'basket'
[hw] ~ [fw] hwuera, fuera /fwuera/ [hwuera], [fuera] 'obstacle'
[d] ~ [t] dada-e, data-e /dada-e/ [dadae], [datae] 'touch'-3ps/E

b) Lexicon:
Representations Phonemic Narrow transcription Gloss
in g r a m m a r structure
itulere, tulere /itulere/ [itulere], [tulere] 'all kinds o f
akaaru, kaum /akauru/ [ahkauru], [kauru] 'they' (3pl)
naamanaa, (dialect-based) [naamanaa], [enamana] 'young man'
enamana
kami, kaami /kaa+nii/ [kanii], [kaanii] 'this'+'that'
= 'sort o f
ne, na /ne/, /na/ [ne], [na] (subordinate marker)
knane, kuanaj /kuanaj/ [kuane], [kuanaj] 'inside'

Out of practical (typographic) considerations, I have transcribed long vowels by

writing two vowel symbols, such as /aa/, [ee], etc. Generally, vowel length is
only spelled out where this is required in order to distinguish long vowels from
diphthongs or short vowels. Vowel length is not spelled out at morpheme
boundaries, where it is predictable. In addition, diphthongs are distinguished
from the vowel sequences, such as /a/+/i/ (which is represented as <ai>), as

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Terminology and conventions 21

opposed to as <aj> for the diphthong. Further details are given in (9) and will be
further analysed in the respective sections.

(9) Transcription of diphthongs and vowel sequences in roots

Underlying Phonetic Transcription Comments

in grammar
/aw/ [aw] au
/a«/ [a«] a«
/eu/ [eu] en
/ae/ [ae] ae
/ai/ [a]] a
i
/ej/ [ej] ej
/e.i/ [ee.i] ei
/a.i/ [aa.i] ai
/ e . of [ee.o] eo
/a.u/ [aa.u] aau except automatic lengthening
/a.«/ [aa.u] aau except automatic lengthening
/ a.e/ [aa.e] aae except automatic lengthening
/ e.u/ [eeu] eeu except automatic lengthening

Tone has been marked in chapters 2 to 4 only, as it may be relevant for the
discussion of some phonological structures. Tone is not marked in the rest of
the grammar, as it is not essential for the discussion of most topics and at least
partly predictable from the rules presented in §4. Also note that the transcription
of tone follows some specialised conventions due to typographical overlap of
nasalisation and tone marking. These are explained in §4. Another feature only
used for phonological analysis is the representation of /u/ as the underlying
phoneme for what can be realised as [u] or [o], alternatively (cf. §2.3).
When nasalisation occurs on a long vowel, it is marked on the first vowel
only, as the spread of nasalisation follows a predictable rule (cf. §2.8.10). For
instance, the expression for 'where is?' is transcribed as äaka, though
nasalisation applies to both VV segments, more narrowly transcribed as [ ä ä k a ] .
In diphthongs, nasalisation is marked on the first vowel, such as in haü
'because'.
A note is also in order regarding the orthography used in literacy materials
and in the N T translation. These are largely based on the Chambira dialect area
(zone C; cf. §23.1), since it has the largest number of speakers. 7 As a tendency,
the N T translation is narrower in that it includes some alternations in the

7
Note that this grammar is based on dialect zone A. However, the dialects are mutually
entirely intelligible as most differences are of a minor nature (cf. ch. §23.1).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
22 Introduction

transcription that are omitted in the transcription used here. This includes
optional vowel lengthening and most cases of nasal spreading. The orthography
is based on the Spanish spelling system and does not include any symbols apart
from this. This results in the representation of the central high vowel / « / as <u>.
The allophony between [u] and [o] is accounted for by writing any underlying
vowel /u/ (which can be pronounced as [u] or [o]) as <o> in the orthography.
Glides are represented by their full vowel counterparts, such as <i> for [j], <o>
for [w], etc., which is shown in detail in (10). Glides in hiatus are transcribed as
<h>. Nasalised vowels are followed by <n> in the orthography.

(10) Differences between Urarina orthography and IPA

IPA NT Example IPA Comments

[«] <u> <turuha> [ t « r « a ] 'arrive' High central vowel
[O], [U], [W] <o> <colane> [ k u l a n e ] 'without' [w] after <j>
<coi> [ k w i ] 'medicine'
[i], Ü] <i> <iniha> [inia] 'go up' [j] after <j>
<jiane> [hjane] 'urine'
[to], [d 3 ] <ch> <cacha> [katga] 'man' Voiced word-initially,
<cha> [d^a] 'what' otherwise [to]
[fw] <P> <ichafa> [ i t c a f w a ] 'rifle' but <jo>+vowel
pronounced as
[hw]+vowel
[h] Ψ <jaa> [haa] 'make' (Like Spanish)
[k] <c> <caa> [kaa] 'this' (Like Spanish)
[k] <qu> <quiha> [kia] 'eat' before <e>, <i> (like
Spanish)
? ?
[ ], [ u i l [g] <h> <eheeoca> [ e e e o k a ] 'shout' used as boundary symbol
<cuha> [ktaqa]'go' to indicate glottal stop,
<enuha> [ e n a g a ] 'tree' approximant [u|] after /«/,
or glide between two
vowels
Μ <sh> <shaaoca> [Jaaoka] 'kick'
Γι], [a], etc. <n> <arajiin> [arahn] 'many' Vowel nasalisation,
<asaan> [asää] 'inside' marked by <n> after the
vowel

Another rule regards the representation of diphthongs and long vowels, which is
similar to the transcription used in this grammar. Diphthongs are spelled as a
sequence of two vowels, such <ai> for [aj] and <ao> for [aw]. In contrast,
sequences of two non-identical vowels are spelled as a long vowel plus a short
vowel, eg., [aai] = <aai>.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Terminology and conventions 23

1.7.2 Organisation of examples

The representation of examples in this book may slightly vary with regard to the
context in which they occur. Most examples that require an interlinear
translation consist of three lines, which represent the Urarina transcription, an
interlinearised gloss, and a free translation in English, as exemplified in (11a).
In examples, where an instant understanding of complex syntactic structures is
essential, such as in the chapter on constituent order and some other chapters
that involve syntax, I have bracketed syntactic constituents and added a line to
specify these, as illustrated in ( l i b ) , where the arguments of the verb (V) are
clearly marked as A and O. In some cases, such as shown in (11c), marking of
arguments may be simplified in that only the clause structure is indicated, but
not every single constituent of each clause.

(11) Different representations of examples

a) "Short" example:
lana-hee-ka ka=raj laano
be.missing-DIM-3ps/A lsg=for cassava
Ί need (a little) cassava.'

b) Example with detailed syntactic information added:

unee bua basihjau-a alau=ne
[kinkajou bag] [steal-3ps/D] [spider.monkey=CND]
[Ο] [V] [A]
'When the spider monkey stole kinkajou's bag' 8

c) Example with simplified syntactic information added:

rdi hetau=te raj rauihirii be-uru-a=na hau, eseneta-e
[that HRS=FOC] [for really tell-PL-3ps/D=SUB because] [believe-3ps/E]
[A] [DEP] [V]
'He believed it because they really told him [that his wife was lost in the swamp].'

Where morphemes or words are in round brackets in the first line of an

example, this implies they are redundant or should be omitted according to
other speakers' judgement. Typically, this would regard the excessive use of
gap fillers (cf. §22.5). Where practical, commas are used in the Urarina
transcription of examples to indicate clause boundaries, such as to mark
different dependent clauses (cf. (1 lc)).

g
For an explanation of animal terms such as 'kinkajou' (a monkey-like mammal) see
Appendix B.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
24 Introduction

In some chapters, bracketing of the interlinearised translation (line 2 of

examples) is used to mark the respective structures under investigation. For
instance, in §7 (Possession), brackets are used to highlight possessive NPs, as
exemplified in (12a). Similarly, in §17 on serial verb constructions, brackets are
inserted to mark serial verbs. In a similar way, some syntactically complex
examples are marked in §6 on NP structure.

(12) Use of brackets in other chapters

a) Possessive NPs:
raj n=arai-tQuru raj hi-a ku-e
[POSS 3ps=family-PL] for tell-NTR go-3ps/E
'He went to tell his family.'

b) SVCs:
ahaenaa su-ahi-a=ne ku-u
weed [kill-CNT-NTR=FOC: l s g go-lsg/E]
Ί went to weed grass.'

There is a small number of frequently recurring words for which certain

simplifications are applied in their interlinearisation. The forms shown in (13)
are morphologically complex, but are either lexicalised or being lexicalised in
their current state. As they are particularly common words, they have been
translated as a single lexical item in the interlinearised version. The same
strategy is applied to a small number of idioms or other complex expressions
that recur throughout this grammar (e.g. 'our creator').

(13) Simplified forms in the interlinear translation

Morphological Exact glossing and Urarina form as Simplified glossing

structure structure used here
d%atoane-~t 'be how'-PRT d^atoanei 'how'
katoane-i ' b e like t h i s ' - P R T katoarufi Mike this'
rdtoane-i ' b e like t h a t ' - P R T nitoanel 'like that'
rihihe-ϊ ' b e like'-PRT rihihei •like'
rihe-1 (short form) rihel
aj-a=m rihihe-'i A U X - 3 p s / D = S U B 'like' ajam rihihei 'like'
kaa+mi, ka+mi 'this'+'that' kaami, kami 'sort o f ; G P F
kam+hvaawi-era lpl/in+'create'-AG kana+hoaawiera 'our creator'
amn-naa+kiuaaHn-era 'walk'-NOM+'create'- amHiiaa+kiuaaunera 'Creator of w a y s '
AG
kahe+kiuaawie-l 'from'+'create'-PRT kahe+kiuaauml ' f r o m there'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Terminology and conventions 25

Morphological Exact glossing and Urarina form as Simplified glossing

structure structure used here
amue+mue-ka RED+'walk'-3ps/A amuemueka 'wander'
balu+balu+bahiu-ka RED+RED+'wrap'- balnbalnbahiuka 'embrace'
3ps/A

Some other words have multiple meanings, as listed in (14). Since these
meanings, which include pairs such as 'give' and ' m a k e ' are semantically
related, it would be implausible to assume that they represent separate lexical
entries. However, the interpretation of the respective meaning depends entirely
on the context. Therefore, I found it useful to provide the appropriate glossing
in the respective examples. In other words, a word such as tia is glossed as
' g i v e ' in some examples and as ' m a k e ' in others.

(14) Words with multiple meanings

Urarina word Meanings

tia 1. 'give'
2. 'make' (also implying 'put', 'set')
hana 1. 'when'
2. 'instead'
3. 'inside'
ke 1. VLI (Valency increasing postposition)
2. INST (Instrumental postposition)
hauria 1. 'earlier'
2. 'first'
patona 1. 'certainly'
2. 'exactly', 'in fact'
raa 1. receive'
2. 'take'
ahaena 1. 'forest', 'jungle'
2. 'grass', 'weed'
kanii, kaanii 1. 'sort of
2. GPF (Gap filler; cf. §22.5)
=ne, =na 1. FOC: lsg
2. FOC: 1 pi
=ne 1. SUB (subordinate marker or complementer)
2. CND (conditional subordinate marker, also
with temporal function; see §20.1.1)

The different translations for the last two examples in (14) are based on
variations in their function. The form kaiiii (or variant kaanii) can function as a
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
26 Introduction

gap filler (e.g. in pause), or have multiple other functions, which are discussed
in §22.5. The variation between the enclitics for focus, which in principle can
occur with lsg or lpl reference, is due to variations investigated in §19.
Some words have an English translation that is impractical to display due to
its length. These are abbreviated in the interlinearised translation and include
some terms for animals or plants, such as illustrated with two different types of
peccaries in (15). Where an English term is unavailable, the local Spanish name
of plants and animals is provided. A list of the scientific names and local
Spanish names for all plant and animal terms is given in Appendix B. In cases
where the scientific name is unknown, only the name in local Spanish is given.
In a small number of examples, no exact identification was possible; these are
glossed as 'type of fish', 'type of tree', etc. A simplified translation was also
chosen for some words that involve a complex deictic definition, such as
locational demonstratives, which are discussed in detail in §5.6.2, but
abbreviated to 'there' or 'over there' in glossing them. In a similar way, the
glossing of the weak and strong prohibitive markers and their variants is
simplified to ' d o n ' t ' , but discussed in detail in § 15.2.

(15) Special glossing conventions

Urarina Exact translation Abbreviated translation

raana 'white-lipped peccary' 'w.l. peccary'
obana 'collared peccary' 'cl. peccary'
ßoaelu 'in ancient times' -> 'earlier'
näi (non-indicating, non- -> 'over there'
punctual distant locational
demonstrative)
kuae (non-indicating, non- 'over there'
punctual, non-distant
locational demonstrative)
kiua 'don't' (weak prohibitive) 'don't'
mhjauria, hjauiße 'don't' (strong prohibitive) 'don't'

The final line of each example is an intermediate solution between a free

translation and a literal representation of the sentence's grammatical structure.
In some cases, the reader may find the English glossing awkward - this is due
to the attempt to provide a translation that helps to follow the syntactic or
semantic structure. In some examples, I have accounted for a more literal
translation in square brackets. These are also used in order to supplement
information not given in the Urarina version of a sentence but implied by the
discourse, such as illustrated in (16a). In examples where certain information is
marked morphologically in Urarina, but does not surface otherwise, this is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Acknowledgements and personal notes 27

indicated in round brackets. In (16b), for instance, 'my son-in-law' is expressed

only through the suffix -am, which refers to a male in-law.

(16) Brackets in free translation

a) Square brackets:
itQaf(f)a-anu hana-te lejhii nii eene-etga mla-t(t)o-a=ne hau
shoot-lsg/D when=FOC one that woman-only fall-INTS-3ps/D=SUB because
'When I shot it [the monkey], as the female fell down ... [the male tried to escape].'

b) Round brackets:
kiuajtel kohwanoo ari-a ku-akaanu hau tabiit$a lejhii ru-hi-ana-e
again next.day seek-NTR go-lpl/ex because finally one find-DIM-ILT-3ps/E
'As we went to seek [turtles] again on the next day, he (my son-in-law) finally found
one.'

Where practical, in some examples, the context from which the example was
taken, is also added. This describes the situation in which a sentence was
uttered; alternatively, the free translation of the sentence preceding the example
is given, to allow a more thorough embedding of the example into a specific
context.
Where Urarina words are used in text, they are given in italic style. Verbs
are quoted in their citation form, if not indicated otherwise, as in itQaa 'do',
which morphologically consists of the root it$a plus the 3ps/A suffix -a. Note
that the addition of a slash and the capital letters /A, /E, and /D after person
markingYefers to the person inflection class as investigated in §11.

1.8 Acknowledgements and personal notes

When I first made contact with the Urarina, they had reason to be suspicious
about the inquisitive gringo who came to learn things about their language. All
too often the Urarina had been deceived and exploited by outsiders. I would like
to thank the people of Nueva Union de Espejo for their trust in receiving me and
my family in their community. Not only did they provide living space, but in a
growing relationship of trust over the years they also shared many of their
customs with me. Most importantly, they made great efforts in searching for
ways to let me understand the structure of their language. There are two people
whom I must mention first of all: Medardo Arahuata Manizari, a top storyteller
and musical artist, spent days if not weeks trying to remember the stories he had
heard in his childhood. This has layed the foundation for preserving an
impressive range of traditional narratives for future generations. Julian Nuribe

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
28 Introduction

Vela, my most senior and knowledgeable consultant, patiently helped me to

transcribe and analyse endless hours of text and other data. Moreover, he
answered the oddest questions an Urarina could imagine. I am also very grateful
to each of my other consultants, who include Vicente Arahuata Manizari,
Humberto Nuribe Arahuata, Manuel Nuribe Arahuata, plus numerous other
speakers who managed to overcome their fear of speaking into the microphone.
Without the efforts of Ron and Phyllis Manus it would have been extremely
difficult to establish a friendly and rewarding relationship between the Urarina
and myself. I also thank the Manus for their continuing friendship throughout
the project, as they provided invaluable logistical support and advice on cultural
and linguistic matters.
In a remote location such as the Chambira basin, without infrastructure and
communication to the outside world, it is hard to keep in touch with the rest of
the world. I would like to express my gratitude to Elisabeth Wyss from ILV
Peru for providing me with a shortwave transceiver for a number of fieldtrips.
The regular occurrence of her voice also was rather comforting during the
lonely months in the rainforest. Thanks also to Mary Ruth Wise and Margarethe
and Jorge Chavez, who have supported my project both morally and logistically
on various occasions. I would also like to thank Jerry Manus, Gerardo Sosa,
Christian & Susana Linkert, and Anita Vallez, who were among the people who
helped me in a variety of ways during my stays in Iquitos.
When I took up the project, I was aware that fieldwork would involve some
sacrifice of one or the other kind. Either I could travel alone - and leave behind
a distressed partner who has to cope with two little children all by herself, or,
alternatively, I could drag my family along to the field. During the five
fieldtrips on which this grammar is based, we tried both variants. I am proud of
my wife Michaela, who managed all kinds of unfamiliar situations extremely
well. This includes anything from splitting firewood and skinning animals to
avoiding stepping on tarantulas and snakes. Michaela also did a wonderful job
in getting the maps presented in this book into shape and in dealing with a few
further fiddly tasks in producing the manuscript. My son Manuel Mogli still
remembers a few words in Urarina and would like to go back to his jungle
friends some time. Canoe trips in the swamp and chasing wild bees are just the
activities five-year-old boys find so fascinating. My daughter Nadia, who
celebrated her second birthday in the village, has learnt a lot, too, such as to
respect the ever-present fire ants and a range of other creatures that do not
bother us elsewhere. In fact, we also thank the Almighty One for keeping all of
us in one piece. Nariha Cana Coaaunera coreniha.
Throughout the five years of research, the Research Centre for Linguistic
Typology (RCLT), La Trobe University (Melbourne) has been the base of my
studies. I am sincerely grateful to Bob Dixon and Sasha Aikhenvald for
initiating this project and for providing me with an ideal environment for
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Acknowledgements and personal notes 29

conducting my research. Sasha spent endless hours reading the numerous

versions of my script and many aspects would not have been fully explored
without her questions and comments.
The final two years of the project have been conducted under the generous
wings of the Hans Rausing Endangered Languages Project, London, which
provided a research grant to make this venture possible. Their support has not
only enabled me to explore the grammar of Urarina in greater depth, but it also
made other aspects of language documentation possible, which are still to be
published. The materials will also be accessible in the archives of the RCLT and
the Endangered Languages Archive (ELAR), London.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
2 Phonology

The sound system of Urarina contains 16 consonant phonemes plus a number of

other consonants that occur as allophones. One consonant, [hw], has lost its
phonemic status in the innovative language, while some older speakers of the
language still distinguish it from [fw]. There are five vowel phonemes and four
diphthongs, again with some alternations. Vowel length is lexically distinctive
in a few cases only, as there are only a handful of minimal pairs that differ from
each other by vowel length. Nasalisation is a distinctive feature throughout the
phonology. The basic syllable structure is (C)V(V), but exceptions from this
may be found in loans. Word length is an important category for phonological
and morphological alternations.
The analysis of Urarina phonology is divided into three chapters. §2
investigates the sound inventory and automatic phonological alternations. It also
discusses issues such as syllable structure, phonological word, and the role of
clitics. §3 deals with non-automatic regularities, i.e. morpho-phonological
alternations. Another fascinating property of Urarina phonology is its tone
system. While a description of its entire complexity cannot be covered here, the
basic tonal structure and some tonal alternations are discussed in §4.

2.1 Consonants

(17) Consonant phonemes

Labial Alveolar Palatal Velar Retroflex Glottal

Stop b t k
d
Labialised stop kw
Nasal m η η
Voiceless fricative s f h
Labialised fricative fw
Palatalised hj
fricative
Affricate tc
Lateral 1
Flap r

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 31

Urarina has only four stops and, in terms of phonemic distinction, the feature
[+/-Voiced] is distinctive only in one case (/d/ vs. /t/). However, some dialects
the distinction between /d/ and It/ tends to be neutralised, as is outlined in
§2.1.3. Voicedness contrast otherwise only exists as a non-phonemic alternation
for the affricates [t$ / CI3], which occur in complementary distribution (cf.
§2.1.9). There are nasal consonant phonemes in three different places of
articulation, plus a velar nasal whose occurrence is restricted to a predictable
environment (cf. §2.8.1). Labialisation of /k/ and /h/ and palatalisation o f / h / are
widespread features; while their occurrence is largely predictable by vowel
copying, the consonants /kw/ and /hj/ have phonemic status, as some examples
show (cf. §2.1.6; §2.1.12). [fw] and [hw] function as allophones in the
innovative language, but they were distinct sounds in the traditional language,
of which only a few remnants are documented (cf. §2.1.11). The rhotic /r/ has
several possible realisations and is transcribed as <r> throughout this grammar.
As investigated in §2.1.3, it also exhibits some variation with /d/.
The relative frequency for the occurrence of consonants is illustrated in (18).
The numbers indicate the frequency of types in the lexical database. The total
number of entries is 3,365, including ideophones, derived words, loans, affixes,
and onomatopoeias. The numbers for /kw/, /hj/, and [ji] represent their
phonemic occurrence, i.e. in non-predictable environments. That is, for /hj/,
only such instances were counted that do not occur in environment where /hj/ is
preceded by /i/ (which would trigger vowel copying; cf. §2.8.3).

(18) Statistical distribution of consonant phonemes

Consonant % Absolute no.

r 17.3 1,700
t 16.1 1,586
η 15.8 1,555
h 12.6 1,240
k 11.5 1,128
s 5.9 580
1 5.4 529
m 3.5 343
b 3.4 339
iQ 2.8 278
d 2.0 199
fw 1.3 131
kw 1.0 94
hj 0.8 75
η 0.3 31

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
32 Phonology

Consonant % Absolute no.

hw 0.3 27
f 0.1 9
(Total): 100.0 9,844

It is evident that out o f 16 phonemes, only five can be considered frequent (i.e.
10% or more), /r/ and ft/ are the most frequent consonants; nasal consonants
make up about a fifth o f all consonants and /h/ and IkJ are comparatively
frequent as well, whereas most other sounds are less common. In the following
sections, each consonant and its respective allophones are introduced. I have
grouped the discussion of the respective sounds according to allophonic
variations and similarity o f the occurring alternations.

2.1.1 /t/

The phoneme /t/ is not restricted in its distribution: it occurs as a syllable onset,
in word-initial and word-internal position with any vowel.

(19) Distribution o f / t /

a) In word-initial position:
/tarara-a/ [ta.ra.raa] 'pull'-3ps/A
/temule/ [te.mu.le] 'plant'
/te-a/ [ti.a] 'give'-3ps/A
/tururi/ [tu.ru.ri] "llanchama tree'
/turu-a/ [tu.ra.ä] ' arrive'-3 ps/A

b) In word-internal position:
/atane/ [a.ta.ne] 'earth'
/ate/ [a.te] 'fish'
/kuretea/ [ku.re.ti.ä] 'buy'-3ps/A
/nituane-a/ [ni.to.a.ni.ä] 'be like that'-3ps/A
/situ-a/ [si.ttt.ä] 'pass'-3ps/A

In the Chambira dialect [t] also occurs as an allophone o f /d/ in some words. In
the word for 'sky', for instance, the pronunciation may vary between [ d e t e ] and
[ d e d e ] (cf. §2.1.3 on /d/).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 33

2.1.2 Μ

In general, there are no restrictions in the distribution of the phoneme Irl. The
following examples illustrate its occurrence in different positions before
different vowels.

(20) Distribution of Μ

a) In word-initial position:
/ra-a/ [raä] 'receive'-3ps/A
/rela-a/ [re.laä] 'teach'-3ps/A
/rihe/ [ri.hje] ' p i f a y o tree'
/rüa/ [rü.ä] 'side'
/ruaka-a/ [ru.a.kaa] 'carry'-3ps/A

b) In word-internal position:
/ arai/ [a.raa.i] 'family'
/ere/ [e.re] 'word'
/bahari/ [ba.ha.ri] 'tinamou'(type of heron: Tinamus major)
/aruba/ [a.ru.bä] 'giant otter'(Pteronura brasiliensis)
/barae/ [ba.m.e] 'masato' (cassava beer)

What I represent as the phoneme Irl has a wide phonetic range between different
speakers and speaker-internally. The rhotic is normally realised as a retroflex
flap [f], but in some cases the tongue goes back not as far as retroflex position.
In fact, it may be close to an alveolar tap [r] (down to 20 msec occlusion time),
but typically has a longer occlusion time than that (40-90 msec). In some
variations the occlusion time can be longer (ca. 110 msec), which may result in
a pronunciation as a stop close to [d] (or [t] in some dialects). Most typically,
this realisation of Irl occurs before the front vowel III, which is a matter of
assimilation. In word-initial position, the rhotic tends to be pronounced as a
retroflex approximant [^J.
Cajas Rojas & Gualdieri (1987: 87) propose a systematic rule for the
variation of this consonant: "A retroflex affricate changes into a vibrant in
morpheme-internal position." In the examples they give (illustrated in (21a)),
the retroflex is pronounced as [r] (rather than [.(] within the root, but as [j] at
morpheme boundaries). In the Espejo dialect (cf. (21b)), the difference between
root-internal position and position at a morpheme boundary is not systematic:
the rhotic is pronounced as a flap or tap, with some variation, in any word-
internal position.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
34 Phonology

(21) Alternation between [r] and

a) Cajas Rojas & Gualdieri (1987: 87)

/lureri/ [lu.re.ri] 'house'
/rauru-i/ [{aa.ro.-i] 'be healthy'-2ps
/ ratiri-ri-t<?ä«=ni / 'leave'-IRR-lsg/A=ASS

b) Espejo dialect:
/lureri/ [lu.^e.^i], [lu.re.ri] 'house'
/ rauru-i/ [^au.^o.-i], [ .jau.ro.-i] 'be healthy'-2ps
/ratiri-ri-tc§«=ni/ [^a.ti.|;i.-|;i.-tQä«.=ni], 'leave'-IRR-1 sg/A=ASS
[^a.ti.ri.-ri.-t(^äö.=ni]

2.1.3 /d/

The fact that Irl can be realised with a long occlusion time sometimes makes it
difficult to distinguish from /d/. In fact, some of the younger speakers appear to
have neutralised the distinction and use /r/ and /d/ in free variation. Older
speakers of the Espejo dialect, however, clearly distinguish the two phonemes,
with one complication: speakers of the Chambira dialect (and possibly some
other dialects) realise /d/ as [t] in some words. For instance, the word /dede/
' s k y ' is pronounced [dete] in the Chambira dialect. This variation could be
regarded a factor that enhances the loss of distinction between /r/ and /d/ in the
innovative language. Otherwise, the phoneme /d/ is observed in most
phonological environments, as illustrated in (22).

(22) Occurrence of [d]

/dari/ [da.ri] 'frog'

/darane/ [da.ra.ne] 'mojara fish'
/darue/ [da.ru.e] 'pot'
/dada-a/ [da.daä] 'touch'-3ps/A (Chambira dial.: [dataa])
/hadaae/ [ha.daa.e] 'uakari' (Chambira dial.: [hataae])
/ bedajne-a/ [be.daj.ni.ä] 'visit'-3ps/A
/dede/ [de.de] 'sky' (Chambira dial.: [dete])
/jiadera-a/ [jia.de.raä] 'be sad'-3ps/A
/diaa-ka/ [di.aa.kä] 'throw around'-3ps/A
/koduunetu-a/ [ko.doo.ne.to.ä] 'be heaped up'-3ps/A
/duha-a/ [du.h u aa] 'fart'-3ps/A

The phoneme /d/ also occurs with loans such as doktoro 'doctor', dosmildos
' 2 0 0 2 ' , gobernadoro 'governor', or homomda ' c o m m u n i t y ' . The presence of
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 35

borrowings from Spanish, which distinguishes the two sounds under

investigation, would make a development towards the re-establishment of two
separate phonemes more likely; however, this does not exactly match the facts
at this stage, as a loss of the distinction is probable to occur in the future.

2.1.4 /b/

The bilabial stop occurs in word-initial and word-internal position. However, it

is not attested before the back vowel /u/ (or alternant [o]), except in loans from
Spanish.

(23) Distribution of/b/

a) Word-initial position in canonical words:

/baasu/ [baa.so] 'bad thing'
/bereri/ [be.re.ri] 'tarantula'
/bihi/ [bi.hi] 'hand'
/bak«/ [bä.kü] 'bone'

b) Word-internal position in canonical words:

/neba/ [nebä] 'mother'
/abe/ [a.be] 'piranha'
/nabe-a/ [na.bi.ä] 'fish with poison'
/kabuitu-a/ [ka.bö.i.to.a] 'splash (water)'-3ps/A

c) /bo/ in loans:
/bote/ [bo.te] 'boat' (loan from Spanish)
/bolsa/ [bol.sa] 'bag' (loan from Spanish)
/klabo/ [klä.bo] 'nail' (loan from Spanish)

2.1.5 /k/

The velar stop /k/ is found before /a/, /i/, /u/, and /«/. Before /e/, it is attested in
two examples only, and in loans.

(24) Distribution of/k/

a) Word-initial position in canonical words:

/kauatu-a/ [ka.wa.to.ä] 'be nice'-3ps/A
/ki-a/ [ki.ä] 'eat'-3ps/A
/kumasaj/ [ko.ma.saj] 'wife'
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
36 Phonology

/kuane/ [ka.a.ne] 'inside'

b) Word-internal position in canonical words:

/aka/ [a.kä] (3sg pronoun)
/niki/ [ni.ki] (Adversative marker)
/uku/ [u.kü] 'needle'
u
/akttiiha/ [a.ka«.h ä] 'entrance'

c) Rare occurrence of /ke/:

/fwakete/ [fwa.ke.te] 'package' (loan from Spanish)
/ke/ [ke] (Valency increase marker)
/keu/ [ke.o] 'gentry plant'

2.1.6 /kw/

The labialised velar can occur under two conditions:

a) As a result of vowel copying from preceding vowels (see §2.8.3), or
b) As a phoneme. In this case it occurs in root-initial position. Diachronically,
this probably results from labialisation due to vowel sequences: one could
assume that a word like hvenaa 'laugh' originally was hienaa. This
hypothesis is further supported by the fact that /kw/ does not occur before
/u/.
The words in (25) show examples for the occurrence of /kw/ in word-initial
position as well as in morphologically complex constructions such as
intransitivised verbs (where the root is preceded by the prefix m-).

(25) Occurrence of /kw/ in root-initial position

a) Word-initial position:
/kwaaane-a/ [kwaa.ö.ni.ä] 'create'-3ps/A
/kwatia/ [kwa.ti.ä] 'don't' (Prohibitive introducer)
/kwena-a/ [kwe.naä] 'laugh'-3ps/A
/kwitca-a/ [kwi.tcaä] 'heal'-3ps/A

b) Word-internal but root-initial position:

/ne-kwasaha-a/ [ne.kwa.sa.haa] ITR-'get hurt'-3ps/A
/ne-kwarate-a/ [ne.kwa.ra.ti.ä] ITR-'show'-3ps/A (= 'seem')

In a few examples, some of which are shown in (26), /kw/ is found root-
internally without being a result of labialisation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 37

(26) [kw] in root-internal position

/nekwäri/ [ne.kwH.ri] 'Rufescent Tiger-heron'

/ekwaj/ [e.kwej] 'down'
/d3akwalaitu-a/ [dja.kwa.laa.i.to.a] 'grab and pull'-3ps/A
/nakwaannel/ [na.kwaa.a.nee.T ] 'again'
/nekwehe/ [ne.kwe.he] 'shame'

One evident point of discussion is whether /kw/ is a combination o f /k/ and /u/
synchronically. This is in fact the case in some examples in which the initial
/kw/ is decomposed into two single parts after palatalisation (cf. 3.5.1). When
the word hverehe 'child' is combined with the 2sg possessive clitic i=, it is
realised as [i.tpu.e.re.he]. The transformation o f /kJ into [t$] is a result o f a
palatalisation rule and therefore the "stranded" labial glide is realised as a full
vowel. This is due to the fact that [t£w] is not a possible sequence in Urarina
phonology. Another case is the occurrence o f the associative marker h i - with
some nouns: the word it$ana for 'blood' is cited as [ k w i . t ^ a . n a ] , (underlyingly
/ k u - i t £ a n a / ) ; however, under certain conditions, it will occur without the
initial [kw] sequence, which s h o w s that it is a separate morpheme (also cf. §7
on possessives).
On the other hand, there are many words in which /kw/ is not a prefix, nor
subject to any alternation, as was illustrated in (25). The most convincing
evidence, however, is given by a minimal pair: the word for 'medicine' is
pronounced as [ k w i ] , whereas there is a word [ku.i], which means
'overripe/foul'. While it may well be the case that the /kw/ ~ [ku] alternation
indicates the diachronic origin o f /kw/, it does not support the claim that all
occurrences o f /kw/ are underlying combinations o f /k/ + /u/ synchronically.

2.1.7 /$/

The fricative /s/ does not exhibit any restrictions in distribution.

(27) Distribution o f / s /

a) In word-initial position:
/sau-a/ [saa.a] 'cut'-3ps/A
/ setu-a/ [se.tu.a] 'rot'-3ps/A
/sini-a/ [si.ni.ä] 'sleep'-3ps/A
/su-a/ [su.ä] 'kill'-3ps/A
/sari/ [su.ri] 'intestine'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
38 Phonology

b) In word-internal position:
/ arasane/ [a.ra.sa.ne] 'dew'
/eseneta-a/ [e.se.ne.taä] 'believe'-3ps/A
/anisihja/ [a.ni.si.hjä] 'main pole of the house'
/sunajjia/ [so.naj.jta] 'afternoon'
/es«/ [e.sa] 'chinbillo tree'

2.1.8 ///

The palatal voiceless fricative is quite infrequent and mainly occurs in loans.
Nevertheless, a few native examples are found.

(28) Distribution o f / / /

/Jaaeto-a/ [Jaa.e.to.ä] 'step on'-3ps/A

//abeeto/ [/a.bee.to] "guayaba tree'
//eremia/ [Je.re.mi.ä] 'pandisho tree'
/Juhwira/ [Ju.hwi.ra] 'blanket'(possibly a loan; origin unknown)

The occurrence of /J/ in word-internal position is even rarer and only attested
for a few words in the Espejo dialect. In the Airico dialect, Is/ tends to be
realised as [[] after l\l and after nasalised vowels. Though I cannot confirm this
for all cases, I have recorded a few examples in which this rule applies:

(29) [J] after /i/

/äasihe-i/ [äa./i.hee.T] 'a little' (but also [äa.tQi.hel])

/bisuito-a/ [bi.Ju.i.to.a] 'drop'
/ri/aeko-a/ [ri./ae.ko.a] 'be very small'-3ps/A
/le=fwaa/a/ [le.fwaa.Ja] 'one hundred' (loan from Quechua)
/i-su-a/ [i./u.a] 2sg=beat-3ps/A ('he beat you'; Airico dialect)

Counterexamples to this regularity are räasaa 'dance', äasaj 'wicked', usi

'fire' (but isi in some dialects, including the Chambira dialect), to?ii-si-ji
('sound'-CPL-NEG:3ps/A) 'it does not work', hasisi 'mixture', kuisia 'be
sorry', misi 'umbilical cord', and various others. Even though [J] is extremely
rare and mainly occurs in predictable contexts, its distinctive function in some
words cannot be denied. Therefore, it must be regarded a phoneme.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 39

2.1.9 /tQ/,[d7j

The voiceless alveopalatal affricate /tp/ mainly occurs word-internally as a

syllable onset. In word-initial position, it is usually realised with the voiced
affricate [ d j ] , which only occurs word-initially, before /a/ and /«/ (except in the
Quechua loan d^ufiuana 'chupana flute', where it is found before /u/). 9 There
are only a few examples where [tp] is attested in word-initial position, such as
in tg,äe 'also', tgoae 'on', tg,aaohwaa 'be ugly', and the "certainty" particle IQU.
It must be noted that no other examples than these four are attested and that
these may constitute exceptions: the word for 'on' is realised with an initial /i/
(as itQoae) in the Chambira dialect; the conjunction for 'also' is a member of a
peripheral word class; only the word for 'be ugly' seems to represent a
convincing example for /tp/ in initial position (it is unknown how this word is
realised in the Chambira dialect), but based on exception, I would still conclude
that the rule as exemplified in (30c) generally applies.

(30) Distribution of [tp] and [ d j ]

a) Word-internal occurrence of [tp]:

/kat$a/ [ka.tpa] 'man'
/ahitpa/ [a.hi.t^a] 'eldest son'
/ ajtQune/ [aj.t^u.ne] 'Espejo River'
/enejt^ä/ [e.nej.tgy] 'monkey'
b) Other occurrences of [t<s]:
/tcoae/ [tpo.a.e] 'on'
/tcäe/ [t<?äe] 'also'
/t^aauha-a/ [tcaa.o.hwää] 'be ugly'-3ps/A

c) Occurrence of [dg]:
/ djakari/ [d3a.ka.ri] 'crocodile'
/djala-a/ [dja.lu.a] 'distribute'-3ps/A
/d3«/ [d30] 'where'
/d3«h»leni-a/ [d 3 ii.h«.le.ni.a] 'stop'-3ps/A

In addition, there is the proclitic dz=, which occurs as an allomorph of the

object and possessive proclitic for 2n person singular /=. It is attached to vowel-
initial words and occurs with an epenthetic /e/ when attached to a consonant-
initial word, as in d^eteteriu below (also cf. §3.5.2)).

9
Note that there is a slight asymmetry between the voiced and voiceless allophones, as
the exact counterpart of [tp] would be the voiced alveopalaatal affricate [d?].
However, the voiced affricate tends to be realised as [d 3 ].
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
40 Phonology

(31) c?3= as variant of proclitic i=

/i=ukwala/ [d3u.kwa.la] 'your younger brother'

/i=ukwana/ [d3u.kwa.na] 'your field'
/i=uba/ [d3u.ba] 'your mother'
/i=ere/ [d3e.re] 'your word'
/i=tete-riu/ [d3e.te.te.ri.«] 'if I were you'

Also recall that /t^/ can be the result of palatalisation in words with the proclitic
i= and an initial /kw/, as was mentioned in §2.1.6.
There are some words that begin with the sound sequence /i/ + /t$/, such as
/ i t ^ a s u / 'companion'. The Γι/ in these examples is mute in rapid speech. It is
well imaginable that diachronically, the few examples that begin with the
voiceless affricate may have been preceded by l\l as well. Thus, the rule that
transforms initial /t$/ into [CI3] is almost without exception. As a consequence, I
do not consider [CI3] a separate phoneme. For the occurrence of /t$/ as an
allophone of /r/, see §3.2 on phonological alternations.

2.1.10 /h/

The glottal fricative /h/ can occur as a syllable onset with most vowels, in any
position. However, it is extremely rare before /«/, except in cases where this is
the result of vowel copying (e.g. / s u u . h a / - [sHu.h a a] 'heart').

(32) Distribution of/h/

a) In word-initial position:
/hanuri/ [hanori] 'back'
/here-a/ [he.ri.ä] 'want'-3ps/A
/hit^ana/ [hi.t^a.nä] 'blowgun'
/huhu-a/ [hu.hu.ä] 'increase'-3ps/A
/htftmo/ [huH.no] 'nice smell'

b) In word-internal position:
/ahaane/ [a.haa.ne] 'ashes'
/aheri/ [a.he.ri] 'stone'
/ahinia/ [a.hi.ni.ä] 'before'
/kajahuri/ [ka.ja.hu.ri] 'cumala tree'
/d3«hulene-a/ [d3n.hu.le.ni.a] 'stop'

A remarkable feature of /h/ is its occurrence between two syllables, since

Urarina syllables do otherwise not allow any consonants in the coda. This may
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 41

be characterised as an instance of postvocalic aspiration. The occurrence of this

phenomenon is optional and non-distinctive and can vary between speakers.
However, it is possible to identify certain features that favour the insertion of /h/
after vowels. Note that most speakers of the Espejo dialect do not have the
postvocalic aspiration before /l/, whereas it is observed in other dialects in this
environment.

(33) Postvocalic aspiration

After /a/ After lei After III After /u/ After /«/
Before [ah.te] [eh.to.e] [bih.to.a] [kuh.ti.a] [tnh.ttt.a]
It/ 'fish' 'owl monkey' 'trap' 'invite' 'mash'
Before [ah.ka] [eh.kti.a] [lih.kii.to.a] [muh.ku.a] [buh.ku]
Ik/ (pron. 3sg) 'nest' 'smash' 'burn' 'bone'
Before [d3ah.t£ö] [heh.t^a] [kih.tQa] [tuh.t^a] —

/til/ (dubitative) (HRS) 'boy' 'touch'

Before [ah.sae] [neh.sa.ri.a] [nuh.si.nye] [kuh.si.nya.ka] [ntih.se]
Isl 'inside' 'hunt' 'bait' 'abstain' 'town'
Before [kah.laa] [leh.le] — [muh.lu.a] [kuh.le]
III 'bite' 'tongue' 'cut off 'testicle'
Before [nah.kwaa.u.nel] [neh.kwe.he] — [uh.kwaa.e] —

/kw/ 'again' 'shame' 'deer'

Generally, postvocalic aspiration occurs after all short vowels (never after long
vowels), but in fewer contexts after /i/ and / « / (however, these are lower in
frequency than other vowels anyway). Thus, the quality of the preceding vowel
is not a criterion for or against aspiration. Postvocalic aspiration mainly occurs
before voiceless stops (including affricate ItQl and the labialised velar /kw/).
Aspiration also occurs before /s/ and /l/. The phenomenon does not occur before
Irl, /b/, and before nasal consonants. The absence of aspiration before /t$/, /l/,
and /kw/ after some vowels may again be due to the relatively low frequency of
these phonemes.
Manus (2001, p.c. about the Chambira dialect) attribute the aspiration to
both syllables involved and call it "pre-consonant onset"; their consultants
realised words with "something close to a geminate", such as in [ak.ka], [at.te].
This is in fact a dialectal variation found in the upper Chambira area (cf.
§23.1.8). Cajas Rojas & Gualdieri (1987: 57-64) treat this phenomenon as
aspiration and describe it as "a case of coarticulation conditioned by a preceding
High tone". Their account suggests that aspiration
"... occurs between a syllable with Η tone and a voiceless or lateral consonant.
Aspiration is present when the tone-bearing syllable is open and its nucleus is an
oral vowel; it never occurs with closed syllables or with nasalised vowels."

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
42 Phonology

This analysis cannot be confirmed as far as tone is regarded. Tone is normally

assigned to the last syllable of a word (in isolation; cf. §4). In trisyllabic or
longer words, the analysis of Cajas Rojas & Gualdieri does account for the
facts. For instance, in [ k a h . k u . n u ] 'daughter', the syllable which exhibits
postvocalic aspiration is not adjacent to the Η-toned, final syllable. Similarly, in
other examples, such as [ m u h . l u . ä ] 'cut o f f or [ k a h . s a . r e . n a ] (a tribal name),
there is an additional syllable between the aspiration and the Η-toned final
syllable. Thus, the only connection between tone and postvocalic aspiration is
that aspiration only occurs on non-H toned syllables.
As far as the involvement of a voiceless or lateral consonant is concerned,
their description is correct, as the table in (33) suggests. Cajas Rojas &
Gualdieri further mention correctly that each root typically has only one
aspiration, occurring on the first syllable in most of the cases. However, the
prefix/proclitic itself can be aspirated. There are no words in which two
aspirated syllables would follow each other. The examples in (34) show some
cases of prefixation.

(34) Postvocalic aspiration in prefix and root (data from Cajas Rojas & Gualdieri
1987)

[ka] + [kwöh.töri] -> ka=hvuturi [kah.kwu.tu.ri] lsg='head' ('my head')

[ne] + [ku.reh.ti.a] -> ne-kureti-a [neh.ku.reh.ti.a] ITR-buy-3ps/A ('he sold it')
In the Espejo dialect, there is only one aspiration per phonological word, i.e. the
pronunciation of the example in (34b) would be [ n e h . k u . r e . t i . a ] or
[ n e . k u . r e h . t i . a ] , alternatively. In summary, it is difficult to account for the
presence or absence of the aspiration, in particular since it is optional and non-
phonemic. However, the following contexts for the occurrence of aspiration can
be identified as conditional and apply to the phonological word. Postvocalic
aspiration occurs:
1. On non-final, short syllables.
2. On non-H toned syllables.
3. Before voiceless and lateral consonants.

2.1.11 /fw/, [hw]

There are two fricatives that phonetically only minimally differ from each other
and are in fact not distinguished by most younger speakers. While these were
distinguished at an earlier stage of the language, the two consonants /hw/ and
/fw/ have lost their phonemic distinction in most words. In the innovative
language, [hw] (a labialised glottal fricative) and [fw] (a labialised labiodental
fricative) are "nearly" in free variation. As a tendency, most words that begin

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 43

with one or the other consonant are now pronounced with /fw/; there are only
four examples of words that begin with /hw/. In addition, there are a number of
words for which the two main consultants could not agree on a "correct"
pronunciation even though both speakers are not part of the younger population.
Most younger speakers would also disregard the /hw/ of the examples in (35b)
and pronounce them with [fw].

(35) Distribution of /fw/ and /hw/

a) Agreed pronunciation with /fw/

/fwara-a/ [fwa.ru.a] 'cut'-3ps/A
/fwaa-a/ [fwau.ä] 'descend'-3ps/A
/ fwei/ [fwee.i] 'firewood'
/fwitQu-a/ [fwi.tQu.ä] 'swim'-3ps/A
/fwi-a/ [fwi.ä] 'fell'-3ps/A
/fwua-ka/ [fwö.a.kä] 'close'-3ps/A
/itpafwa/ [i.tca.fwa] 'rifle'
/kufwarihja/ [ku.fwa.ri.hja] 'birth hole'
/ne-fwa-a/ [ne.fwaä] ITR-'increase'-3ps/A

b) Agreed pronunciation with /hw/

/hwaaeku-a/ [hwaa.e.ko.a] 'be open'-3ps/A
/hwalaitu-a/ [hwa.läa.i.to.a] 'lift up'-3ps/A
/hwauitu-a/ [hwaa.i.to.a] 'scratch'-3ps/A
/hwulene-a/ [hwu.le.ni.a] 'move to the side'-3ps/A

c) Disputed pronunciation:
/fwuera/ [fwu.e.ra / hwH.e.ra] 'obstacle'
/fwaitu-a/ [fwä.i.to.ä / hwu.i.to.ä] 'cover'-3ps/A
/fwafwaka-a/ [fwa.fwa.kaä / hwa.hwa.kaä] 'lift up'-3ps/A

For older speakers (who speak the traditional language), the two sounds have
distinctive function in a few words, as one known minimal pair illustrates.

(36) Minimal pair /fw/ vs. /hw/ in the traditional language

/ufwa/ [o.fwä] 'father'
/uhwa/ [o.hwä] 'huicungu tree'

The scenario described here makes it difficult to state whether /hw/ and /fw/ are
actually different phonemes. It is evident that the distinction is in the process of
being lost and that sooner or later, /hw/ will cease to exist as a distinctive sound
of Urarina. As one can thus postulate that there is no distinction in the
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
44 Phonology

contemporary language, I do not regard [hw] and [fw] separate phonemes and
therefore transcribe the phoneme as /fw/. However, [hw] plays an important role
elsewhere in the phonology of Urarina: When /h/ follows /u/ in word-internal
position, it is labialised and thus is realised as [hw], as a result of vowel copying
from the left. In these cases, the transcription [hw] is retained (also cf. 2.8.3).

(37) [hw] as a result of vowel copying

/banetuhe-V [ba.ne.to.hwel] 'totally'

/nuha/ [no.hwä] 'shoulder'
/ rautuhe-a/ [rau.to.hwi.ä] 'be happy'-3ps/A
/seuha/ [see.o.hwä] 'big'

Another observation regards the use of /fw/ as the more prominent of the two
sounds in question: in loans, /fw/ is used to represent /p/, such as frequently
occurs in Spanish words:

(38) [hw] in loans

/afwaja/ [a.fwa.jä] 'papaya' (loan from Spanish)

/fwafwafwa/ [fwa.fwa.fwa] ' h u a p a p a ' (type of heron) (loan from Spanish)

2.1.12 /hj/

Similar to /hw/, /hj/ can be the result of vowel copying through /h/ (cf. §2.8.3).
But /hj/ also functions as a phoneme, mainly occurring in word-initial position.
Its distribution is restricted, as it occurs before the vowels /a/, /u/, and / « / only.

(39) Phonemic ence of [hj]

/hja-a / [hjaä] 'urinate'-3ps/A

/hjane-a/ [hja.ni.ä] 'leave/let'-3ps/A
/hjanarihi/ [hja.na.ri.hi] 'sugar cane'
/hjan«n«-a/ [hja.nu.na.a] 'pull'-3ps/A
/hjusi-a/ [hja.si.a] 'scratch'-3ps/A

In a few examples, /hj/ occurs in a different context, i.e. not in initial position
and not preceded by /if:

(40) /hj/ in word-internal position

/kahjaasi/ [ka.hjaa.si] 'cortadera plant'

/ahjaaune/ [a.hjaa.o.ne] 'santa maria plant'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 45

/kahjune/ [ka.hju.ne] 'clothing'

/kuhjuta-a/ [ku.hju.taa] 'bless/curse'-3ps/A
/huhja-a/ [huhjaä] 'make grow'-3ps/A

There are reasons to believe that /hj/ originally is a sequence of /h/ + /{/, at least
in some examples. Similar to the idea that [kw] could be a result of reduction
f r o m /ku/ (cf. §2.1.6), one could assume that the glottal fricative /hj/ is the result
of an underlying syllable /hi/, when another vowel follows. This is illustrated
with the examples in (41), which contain roots followed by a " m e r g i n g " suffix;
i.e. the final vowel of the root merges with the attached suffix. For example,
when the locational noun makuhi 'upriver' is followed by the location suffix - « ,
the final /i/ of the noun merges with the suffix and the word is resyllabified. A
similar situation applies to the two other examples, with some morpho-
phonological complications. The word for ' s h a r p e n ' contains the causative
suffix -a, which merges with the final syllable of the verb root.

(41) [hj] as a result of /hi/ + [V]

/makuhi-u/ [ma.ku.hja] 'upriver'-LOC

/ahi-a-a/ [a.hjaä] 'be.sharp'-CAUl-3ps/A
/huuhi-a/ [huu.hjä] 'grow'-3ps/A

2.1.13 ///

T h e lateral /l/ does not have any restrictions in distribution. It occurs in initial or
internal position of the word and can precede any vowel.

(42) Distribution of /I/

a) In word-initial position:
/lana/ [la.nä] 'husband'
/lenone/ [le.no.ne] 'food'
/likiitu-a/ [li.kii.to.ä] 'mash'-3ps/A
/lureri/ [lu.re.ri] 'house'
/lnae/ [ln.ae] 'nightmare'

b) In word-internal position:
/alaa/ [a.laä] 'agnaje tree'
/lele/ [le.le] 'tongue'
/ mamale-a/ [ma.ma.li.ä] 'make stick'-3ps/A
/mulu-a/ [mo.lo.ä] 'cut off -3ps/A
/alulari/ [a.lu.ltir.i] 'termite nest'
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
46 Phonology

2.1.14 /η/

For the alveolar nasal /n/, no restrictions in distribution apply, as it occurs

before any vowel in word-internal position as well as in word-initial position.

(43) Distribution of /n/

a) In word-initial position:
/nahari/ [na.ha.ri] 'liver'
/neda-a/ [ne.daä] 'stay'-3ps/A
/nisisi-a/ [ni.si.si.ä] 'wipe off-3ps/A
/nuri/ [no.ri] 'giant river turtle'
/nane/ [nu.ne] 'branch'

b) In word-internal position:
/ena/ [e.nä] now
/ahaane/ [a.haa.ne] 'ashes'
/ajrinia/ [aj.ri.ni.ä] 'outside'
/kunu/ [ko.no] 'light'
/enaa/ [e.nH.ä] 'tree'

2.1.15 /m/

Similar to /n/, there are no restrictions in the distribution of the bilabial nasal
/m/.

(44) Distribution of /m/

a) In word-initial position:
/amiane-a/ [a.mi.a.ni.ä] 'work'-3ps/A
/meseri/ [me.se.ri] 'cocona tree'
/ misi/ [mi.si] 'umbilical cord'
/mulu-a/ [mo.lo.ä] 'cut off-3ps/A
/maka-a/ [mu.ku.a] 'catch'-3ps/A

b) In word-internal position:
/ama-a / [a.maä] 'take along'-3ps/A
/temeeku-a/ [te.mee.ko.a] 'stand together'-3ps/A
/amiane-a/ [a.mi.a.ni.ä] 'work'-3ps/A
/kumu-a/ [ko.mo.ä] 'fall down'-3ps/A
/lemn-a/ [le.m«.a] 'sink'-3ps/A

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Consonants 47

2.1.16 /β/

In the Chambira dialect, /j1/ does not occur in word-initial position, but is
always preceded by Γι/ in all environments. This is based on a rule of
palatalisation that applies when /i/ is followed by /n/ (cf. §2.8.9). Consequently,
[ji] is not a phoneme in the Chambira dialect. However, in the Espejo dialect
described here, the palatal nasal is developing its status as a phoneme, since the
initial /i/ is usually omitted. What is iflaara in the Chambira dialect is realised
as paara by Espejo speakers (though variations may occur). In general, the
occurrence of /jV in initial position is rather rare: while this consonant occurs in
the lexical database 195 times, most of the occurrences refer to predictable
environments (i.e. after /i/); its occurrence as a phoneme, in initial position is
only 31 times (cf. (18)). An alternative analysis of /jt/ would be to assume that
an optional rule that deletes a word-initial /i/ before /n/ applies and thus results
in the realisation of words that begin with [ji].
The word-initial position is the only environment in which /ji/ is not
predictable. Otherwise, the palatalisation rule mentioned above applies. The
examples given in (45) show two possible pronunciations of the words. In the
Espejo dialect, the realisation without the /i/ is preferred.

(45) /jV in word-initial position

/jiadera-a/ [jia.de.raä], [i.jia.de.raä] 'be sad'-3ps/A
/jiäe/ [jiäe], [i.jiäe] 'already'
/jietunu/ [pe.to.no], [i.jie.to.no] 'each, every'
/jiuae=lu/ [jio.ae.ML [i.jio.ae.lö] 'earlier'

However, there are some words in which an initial /i/ is present before /n/. In
the two examples where /n/ is preceded by /i/ and followed by another /i/, the
palatalisation rule does not apply. When a different vowel follows, /n/ is
palatalised.

(46) [n] vs. [ji] after /i/

a) No palatalisation when /i/ follows:
/ini-a/ [i.ni.ä], *[i.jii.a], *[jii.a] 'go.up'-3ps/A
/iniu/ [i.ni.jo], *[i.jii.o], *[jii.o] 'meat'

b) Palatalisation when a different vowel follows:

/ijiaa/ [i.jiaä] 'fruit'
/ipunu/ [i.jio.no] 'ayahuasca'
/ijiuri/ [i.jio.ri] 'type of parrot'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
48 Phonology

The dialectal differences may indicate that the initial /ji/ in the Espejo dialect
originates from an initial sequence /i/ + /n/, which resulted in palatalisation; the
initial /i/ was then dropped. In fact, it could be expected that the words in (46b)
above) will lose their initial /i/.

2.1.17 The status of [ j ] and [w]

The palatal glide [j] is not a phoneme, but is realised in hiatus between two
vowels (cf. §2.8.1). It does not occur in word-initial position. However, it can
also be suspected that an underlying l\l is realised as [j] between two identical
vowels. All attested examples involve two /a/ vowels, which surround the glide
[j]. There are three different ways to analyse this, as is exemplified by haja
'horn'.
Analysis 1: /ha.ja/: [j] is a consonantal phoneme and functions as an onset for
the second syllable. This is a possible analysis, but one would wonder why
[j] does not occur in word-initial position then.
Analysis 2: /ha.i.a/: [j] is a realisation of l\l between two identical vowels, but
phonetically reduced to form an onset for the second syllable, due to
resyllabification.
Analysis 3: /haj.a/: [j] is underlyingly the offglide of a diphthong /aj/. This is the
most likely analysis since it accounts for [j] not occurring in word-initial
position and for the fact that it is not found between other identical vowels
(*/eje/, */ojo/, etc.).

(47) Occurrence of [j] between vowels

/haja/ [ha.jä] 'horn'; also: 'muchaca caspi tree'
/taj-a/ [ta.jä] 'pluck/collect'
/here-i/ [he.ri.ji] ' want' -N EG: 3 ps/A

Further examples illustrate the occurrence of [j] as part of the diphthong /aj/
without being followed by another /a/. In conclusion, [j] is not a phoneme.

(48) Occurrence of [j] in diphthongs

/uhajtc«/ [o.hwaj.tpa] 'clay bowl'
/ ajtu-a/ [aj.to.ä] 'say'-3ps/A
/bajte-a/ [baj.tiä] 'forget'-3 ps/A

Similar to its palatal counterpart, the labial glide /w/ has a transitional function
between /u/ and a following vowel (cf. §2.8.1). In other environments, the same
analysis as for the palatal glide can be applied to [w]. It does not occur in word-
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Phonotactics 49

initial position, except in the onomatopoeic verb zuawaahoa 'grunt'. In most

cases where it occurs between two identical vowels, the first syllable is a
diphthong /au/. Phonetically, [w] takes the place of an onset for the next
syllable; i.e. instead of [a.nau.e], the word for 'tamarin' is realised as
[a.na.we].

(49) Occurrence of [w] in other contexts

/larui/ [la.ro. wi] 'spear'

/anaue/ [a.na.we] 'tamarin' (type of monkey)
/arauata/ [a.ra.wa.tä] 'woolly monkey'
/kalaui/ [ka.la.wl] 'son'

2.2 Phonotactics

The previous section has shown that most consonant phonemes occur relatively
freely in word-initial and word-internal position, before any vowel. There are a
few deviations from this, as was discussed in detail. The following table gives a
summary of the phonotactic distribution for all consonant phonemes. As a
general observation, one can say that some consonants occur more freely in
word-initial position rather than word-internally. Regarding the occurrence
before certain vowels, it is observed that there are a few restrictions. With /«/,
the distribution is most restricted, but it should be noted that this is also the least
frequent vowel in terms of statistical frequency (cf. (52)). Another observation
is that there is a correlation between the frequency of consonants (as stated in
(18)) and the compatibility with vowels: the less frequent a consonant is, the
higher is the possibility that it may not be compatible with all vowels.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
50 Phonology

(50) Phonotactic distribution of consonants (sorted according to consonant frequency)

w- W- Before Before Before Before Before

initial internal /a/ /e/ Ν /u/ /a/
Irl As [JQ As yes yes yes yes yes
tri/ [r]
IM yes yes yes yes yes yes yes
/η/ yes yes yes yes yes yes yes
Ihl yes yes yes yes yes yes rare
IkJ yes yes yes rare yes yes yes
Isl yes yes yes yes yes yes yes
η/ yes yes yes yes yes yes yes
Iml yes yes yes yes yes yes yes
Ibl yes yes yes yes yes no yes
/tii/ As [d*] yes yes yes yes yes yes
Id yes yes yes yes yes yes yes
Ifwl yes yes yes yes yes no yes
Ikwl yes rare yes yes yes yes yes
/hj/ yes yes yes no no yes yes
W yes no yes yes no yes no
Ψ yes rare yes yes yes yes no

It should be mentioned that deviations from the regular phoneme inventory are
observed in ideophones (also cf. §5.14). Examples for these are combinations of
sounds that do not normally occur in a sequence, such as a stop followed by a
glottal stop (e.g. Ρεεε to imitate the sound of a rifle). However, since the
realisation of these sounds may vary substantially from speaker to speaker, they
are not included in the presentation of the sound system.
Another aspect is that speakers who are highly proficient in Spanish may
adopt consonant combinations are or even new consonants into their phoneme
inventory. Phonological structures occurring in Spanish loans include /p/, g/,
and a range of consonant clusters, such as stops followed by l\l or Irl. A further
feature sometimes copied from Spanish is the use of coda consonants in loans.
A more detailed discussion of this is found in §23.3.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Vowels 51

2.3 Vowels

Urarina has five vowel phonemes. One major point of discussion is the status of
/u/ (cf. (55)). Another issue is to what extent vowel length is distinctive or not
(cf. §2.4). None of the short vowels is subject to any restrictions regarding its
distribution, as is illustrated by the examples in (53). They occur as a syllable
nucleus in initial, internal, and final position of a word.

(51) Vowel phonemes

i
a u
e
a

The statistical distribution of vowels is partly unproportional, as /a/ is extremely

frequent, occurring in 34% of all lexical entries. The distribution of the other
vowels is more equilibrated, except for /«/, which is the least frequent vowel.

(52) Frequency of short vowels (based on a wordlist of 3,365 entries)

Vowel Absolute no. %

/a/ 3,610 34.0
/ u / , [o] 2,276 (767 + 1,509) 21.4
/e/ 2,025 19.1
/i/ 1,679 15.8
/«/ 1,032 9.7
(Total): 10,622 100.0

The words in (53) represent examples for the occurrence of each vowel in
different positions of the word.

(53) Distribution of vowel phonemes

/a/
/akanu/ [a.ka.no] 'snake'
/ate/ [a.te] 'fish'
/amu-a/ [a.mai] 'walk'-3ps/A
/kat$a/ [ka.tca] 'man'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
52 Phonology

Id
/elu/ [e.lo] 'rain'
/enanihja/ [e.na.ni.hjä] 'canoe'
/eresi/ [e.re.si] 'tomorrow'
/ahaane/ [a.haa.ne] 'ashes'

/i/
/iha/ Ei-hja] 'lighthouse tree'
/iniu/ [i.ni.jo] 'meat'
/it$a-a/ [i.teaa] 'do'-3ps/A
/itanitQa/ [i.ta.ni.tpa] 'mud'

/«/
/«-a/ [«.ä] 'come'-3ps/A
/ökari/ [ti.ka.ri] 'ant eater'
/aluktri/ [a.lu.lö.ri] 'termite nest'
/buk«/ [bu.ku] 'bone'

The preferred phonetic realisation o f the mid vowel /e/ is as a closed vowel. The
phoneme /»/ can be characterised as a high central, (slightly) rounded vowel,
with some (optional) variation. Depending on individual preferences, words, or
dialects, it may also be realised as [a] or [i].

/u/

The situation for this vowel is slightly more complicated than for the others, as
there is a high degree o f variation between [u] and [o] (with [o] typically being
realised as a closed vowel). The occurrence o f the two allophones is not entirely
predictable from the phonological environment, which indicates that these
vowels are in free variation. In fact, there exists no minimal pair that would
prove the distinctive status o f /u/ vs. /of. There also is no convincing evidence
as to which o f the two would be underlying, as both can occur in the same
contexts. However, the distribution o f the two can be described as connected to
certain preferences. For practical purposes, I will describe /u/ as the base
phoneme.
As there is variation among speakers and speaker-internally, it is not easy to
pinpoint the actual preferences for the realisation o f /o/. First o f all, the
preferences appear to be entirely arbitrary and I will not attempt to interpret
these as results o f phonetic regularities. Secondly, while there are some general
tendencies related to phonological environment, some words may be
pronounced differently from what would be expected, which reflects the reality
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Vowels 53

of free variation. The tables in (54a,b) give a fairly accurate overview of the
general preferences of occurrence for each allophone. For most tendencies,
some deviations are also stated. For instance, [u] is the preferred allophone after
/k/, with the exception of the stative/diminutive suffix -koa and, as a tendency,
in word-final position. Similarly, the affricate Ιίς,Ι normally triggers the use of
[u], but in the word for ' o n ' (and some other exceptions), [o] is preferred.

(54) Preferences for allophones [u], [o]

a) Preferred variant [u]:

After cons. Preferred [u| Exceptions
/h/ [hu]
/hj/ [hju]
Ac/ [tcu] tQoae ' o n '

b) Preferred variant [o]:

After cons. Preferred [o] Exceptions
/m/ [mo]
/μ/ [no]
/f/ [fo]
(Word-initial) [o] Before IkJ
/k/ [ko] Before Iii, /hj/, /ji/, /I/, /r/;
Prefix ku-·,
kukuri 'armadillo'
/s/ [so] Before Irl
/r/ [r] At word end;
Before III, Irl·,
erura 'owner'
Ν [to] Inchoative -tura\
itulere 'all kinds'
/η/ [no] Before Ik/, Μ
l\l [lo] Before Irl

Note that the occurrence of /u/ after other consonants is too rare as to indicate
any reliable preference. A further regularity is that a preference for
pronunciation as [u] may change to [o] before nasal consonants. For instance,
the word / k u m u a / 'fall down', tends to be realised as [komoa] due to the
presence of the nasal. Thus, the preference of [u] after /k/ is overridden. Further
exceptions regard loans of any kind, where pronunciation is adapted to the
source language. An example for this is [ k a r t u t ^ o ] from Spanish cartucho 'rifle
shell', which is pronounced with [o] despite the preference for [u] after /tp/.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
54 Phonology

N o n e of the preferences stated here are obligatory. Speakers do not perceive

an utterance as incorrect when a different pronunciation is intentionally
produced, using [o] instead of [u] or vice versa. Differences occur a m o n g
different speakers as well as in speaker-internal variation. Concluding f r o m the
non-distinctive status of /u/ vs. its rounded counterpart, the orthography
uniformly represents the vowel as <o> (while <u> is used to represent /«/).
In the following, I give examples for each allophone in word-initial position.
In (55a), words tend to be pronounced with [u], (but the realisation of the high
back vowel as [o] is possible). T h e same applies in opposite direction for the
e x a m p l e s in (55b). A s a tendency in word-initial position, [u] is preferred before
Irl, /k/, /kw/, and /s/, but [o] before [f], [b] and before nasals. For many words, it
cannot be determined what would be the " p r e f e r r e d " form. Similarly, the
pronunciation varies in vowel sequences.

(55) Occurrence of [u] and [o] in word-initial position

a) [u] as preferred pronunciation

/uku/ [u.ku] 'needle'
/ukwala/ [u.kwa.lä] 'younger son'
/urarina/ [u.ra.ri.jiä] 'Urarina'
/usa/ [u.sä] 'sin'

b) [o] as preferred pronunciation

/ufwa/ [o.fwä] 'father'
/ubaae-ka/ [o.baa.e.kä] 'get angry'-3ps/A
/umari/ [o.ma.ri] 'basket'

While I have characterised /u/ as the underlying vowel. I will follow the
"preferred pronunciation" in my transcription, i.e. each vowel will be
transcribed based on the actual realisation of each example given by the
speaker, while it will be assumed that the actual underlying representation is /u/.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Vowel length 55

2.4 Vowel length

In general, vowels of all qualities mentioned above occur as long vowels.

Regarding the realisation of /u/ and its variant [o], there is no general preference
for [uu] or [oo]. Overall, long vowels are less frequent than their short
counterparts and thus, their combination with consonants appears more
restricted. The reason for this may be due to their lower frequency; perhaps
some combinations not attested here are possible, but absent due to the lack of
occurrences in the database. Examples for each occurrence are illustrated by the
table below.

(56) Occurrence of long vowels in roots (sorted by consonant frequency)

aa ee ii uu uu
No aari eene ii — uune
onset Ίορα tree' 'woman' 'you' 'branch'
r raa reemae ruha fieruuka hiruulene
'receive' 'dog' 'grandfather' 'come off "thread'
t taa tnteema satu netooka hituuhe
'that' 'chop' 'all' 'hang up' 'fried'
η ηaa Hnee mi koJnuamo —

'say' 'kinkajou' 'that' 'the next day'

h ahaa heelej lejhn huuka huwio
'fever' 'same' 'one' 'flood' 'nice smell'
k kaa — likiitoa ekooka eköH
•this' 'smash' 'appear' 'above'
s saa nusei siiria ennasoom SHuha
'rat' 'tail' 'have' 'type of fish' 'heart'
1 alaa knleeka liiaka helöo muluntoa
'agitaje tree' 'wash' 'be tied' 'towards' 'be thick'
m maahet meeri -miι munraria nevwuka
'also' 'face' (JUSS suffix) (part of tree) 'hold on to'
b baaso beene biißa — —

'bad' 'female' 'old'

t<J itQai — — alajtQÖo mtQüutaa
(type of root) 'lancehead snake' 'go near'
d edaa — — hodoomtoa —

'outside' 'be heaped'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
56 Phonology

(cont.) aa ee ii uu
fw ifioaaja) afioeera fioiiaka nofiooo —

'100' (loan) 'stab' 'be hanging' 'rump'

kw hoaaeraa tofaoeekaa kiviitq,u nehoHuri
'be strong' 'lay down' 'shoulder blade' 'garabcitci
plant'
hj hjaare bihjei — — hjnnsia
'lizard' 'spindle' 'scratch out'

J 1 ßaara kiuipeeria — — —

'you.2pl' 'exploit'

J faetoa — — — —

'step on'

Similar to the frequency distribution of single vowels (cf. (52)), /aaJ is the most
common long vowel found in Urarina words, whereas / « « / again is
comparatively rare. Some gaps in the above-table may again reflect the fact that
combinations with lower frequency consonants are less likely to occur.

(57) Statistical distribution o f long v o w e l s

Vowel Absolute no. %

/aa/ 669 55.5
/ii/ 195 16.2
/ee/ 190 15.8
/ u u / , incl. [ o o ] 95 (21 +74) 7.9
/««/ 56 4.6
(Total): 1,205

The crucial question is to what extent vowel length is distinctive. On the one
hand, it is significant on the grammatical level, since the combination of roots
with suffixes results in vowel lengthening under certain conditions. One
example is the imperative -u. When it is suffixed to a verb whose root ends in
/«/, the form results in a long vowel (cf. §3.3).

(58) [ u u ] resulting f r o m a t t a c h m e n t o f imperative s u f f i x

/ama-a / walk -> /amu-u/ [a.mtui] 'walk'-JMP

/muku-a/ 'catch' -> /muku-u/ [mu.knu] 'catch'-IMP

However, there is no minimal pair that could be contrasted to this form. In

general, minimal pairs of any kind are hard to find in Urarina, due to the fact
that roots tend to be relatively long. In addition, even pairs that are segmentally

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Vowel length 57

identical can still d i f f e r from each other by their tonal structure, which gives
a l m o s t no r o o m f o r minimal pairs in a strict sense. All o f the f o l l o w i n g m i n i m a l
pairs are not clear-cut, as they involve non-lexical w o r d classes, loans, or
m o r p h o l o g i c a l l y c o m p l e x f o r m s in o n e or the other w a y . In addition, speakers
h a v e d i f f i c u l t i e s in recognising the d i f f e r e n c e b e t w e e n these pairs. T h i s s h o w s
that v o w e l length plays a very marginal role on t h e lexical level.

(59) Minimal pairs for vowel length

baanä (Irrealis introducer) vs. bana 'when'

baaso 'bad thing' vs. bäso 'vase' (loan, from vaso)
kaana 'father-in-law' vs. kana 'we' (lpl/in)
huataa (Negative introducer) vs. kiuata 'metal' (loan, from plata)

T h i s lack o f contrast is e m p h a s i s e d by the fact that - similar to the /u/ contrast

v o w e l length is variable in s o m e cases (the preferred version is underlined).

(60) Variable vowel length

kuraanaa, kuraanä 'chief

naarä, ßarä 'you' (2pl)
raaheniane, raheniane 'self
eene, ene 'woman'
beelaja, bei aja 'give as a gift'-3ps/A

A n e x a m p l e f o r v o w e l shortening is the realisation o f the d e m o n s t r a t i v e s b e f o r e

n o u n s : kaa and mi are m a y s u r f a c e as ka= and rii= respectively, (cf. §2.11 on
clitics). F r o m the relatively high degree of variation b e t w e e n long and short
v o w e l s in general, o n e w o u l d c o n c l u d e that length only has such a marginal
contrastive f u n c t i o n f o r Urarina p h o n o l o g y that it might be considered non-
p h o n e m i c . H o w e v e r , v o w e l length plays a significant role for a n u m b e r of
f a c t o r s outside the lexicon:
1. T o n e a s s i g n m e n t partly d e p e n d s on vowel length, as light and heavy syllables
are d e f i n e d in these terms, i.e. a syllable that contains a long v o w e l is
" h e a v y " . A s will b e c o m e clear in §4 on tone, syllable w e i g h t is relevant in
o r d e r to d e t e r m i n e the length of a word, on which in turn tone a s s i g n m e n t is
based. W h e n a disyllabic noun begins with a heavy syllable, it is treated like
a trisyllabic noun in terms of tone a s s i g n m e n t under specific c o n d i t i o n s (cf.
§4.6.1 on the possessive tone pattern).
2. Even t h o u g h very little information is available on stress, it is evident that
h e a v y syllables (syllables that contain a long v o w e l ) tend to be stressed. T h e
exact role stress plays is still to be determined.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
58 Phonology

3. Vowel length is significant for certain morpho-phonological processes. Plural

allomorphy is one example for this. The attachment of certain suffixes can
depend on the quantity of a root-final syllable, (as with the distribution of the
3ps suffix on verbs, cf. §3.1.3).
4. The combination of suffixes can result into long vowels and the resulting
forms can be distinctive from other forms. For instance, the 3ps/A form of
/eno-a/ 'enter' ends in a short /a/. When a causative suffix is added, the final
vowel becomes long: /eno-a-a/ ('enter'-CAUl-3ps/A) [e.no.aa].

Based on these aspects it the overall status of vowel length for Urarina
phonology becomes a bit clearer. On the one hand, there is a high degree of
variation between long and short vowels on the lexical level, with only minor
distinctive function. On the other, entire processes depend on syllable weight.
More research will be necessary in order to explore the exact conditions under
which vowel length is predictable or optional. In the meantime, as I have shown
that vowel length is relevant especially in complex morphological
environments, I will assume that it is phonemic.

2.5 Vowel sequences

Urarina allows most combinations of vowels in a sequence. In some cases, the

distinction of a diphthong from a vowel sequence is difficult: if the second
vowel of a sequence is /i/, /e/, /u/, or /«/, it may be neutralised due to shortening
in rapid speech (cf. §3.6.1). The main difference between vowel sequences and
diphthongs is that typically, the first of two vowels in a sequence is lengthened,
which contrasts from a combination of vowels that form a diphthong together. 10
While diphthongs form one rhythmic unit, vowel sequences differ from this in
that they form two separate syllables. An additional difference between vowel
sequences and diphthongs is the presence of hiatus-filling glides ([j], [w])
between vowels that represent separate syllables (cf. §2.8.1).

(61) Examples for attested vowel sequences

/ a.i/ [a.raa.i] 'family'

<
/a.e/ [a.laa.e.ri] caimitu tree'
/a.u / [aa.u.ne] 'cooked plantain
/a.«/ [kwaa.H.ni.ä] 'create'-3ps/A
/ e.i/ [fwee.i] 'firewood'

10
Lengthening mainly occurs with /a/ and /e/. However, it is unclear whether this is
systematic.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Vowel sequences 59

/e.u/ [see.o.hwä] 'big'

/e.u/ [ee.u] 'sieve'
/i.a/ [i.a.ni.a] 'burn'-3ps/A
/i.e/ [si.e] 'cotton'
/i.o/ [i.ni.o] 'meat'
/i.«/ [si.u.rä] 'bag'
/u.a/ [to.ku.a.ni.ä] 'compare'-3ps/A
/u.e/ [da.ru.e] 'pot'
/ u.i/ [ku.i.si.ä] 'be sorry'-3ps/A
/u.a/ [bu.ä] 'bag'
/u.e/ [ba.ru.e] 'masato' (cassava beer)
/u.i/ [ni.tpu.i] 'rib'

There are some combinations of vowels that cannot occur in a sequence of two
separate syllables. These are indicated as "—" in (62).

(62) Table of vowel sequences

a e i u »
a — a.e a.i a.u a.u
e — — e.i e.u e.u
i i.a i.e — i.u i.u
u u.a u.e u.i — —

u.a u.e u.i — —

The impossible combinations mainly involve /«/ and /u/. As a regularity, /«/
does not co-occur in any combination with /u/ in any order. Another illicit
sequence is [e.a], which is the result of a phonological rule: Id is realised as [i]
before /a/ (cf. (82)).
An important feature of vowel sequences is the occurrence of transitional
glides in hiatus between two subsequent vowels. These are [j], [w], and [uj],
which are inserted depending on the first vowel of a sequence. Their
distribution is described in §2.8.1.
There is only one example where a sequence of identical vowels occurs on
the lexical level: in the verb e?eoka 'shout', the sequence of two instances of Id
is interrupted by a glottal stop. However, the meaning of this verb as 'shout' is
clearly realised as an onomatopoeic word, imitating the nature of shouting.
However, a similar situation applies to the combination of verb roots that end in
/a/ when these are followed by the negative of the lpl form: in this case, the
root-final /a/ is followed by the negative suffix -a, which in turn is followed by
the initial /a/ of any lpl suffix, such as -akaa?m for lpl/ex. In combination, the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
60 Phonology

three instances of /a/ are again separated by a glottal stop, to result in

[ i t p a - ? a a k a a n u ] ' w e did not do it'. Since the occurrence of [?] is predictable in
all contexts, it is not considered as a phoneme (also cf. §2.8.8).

2.6 Diphthongs

As mentioned above, diphthongs contrast to vowel sequences [V.V] in that they

occupy only one syllable. They differ from long vowels in that their syllable
structure is to be represented as [V], as opposed to [VV] for long vowels.
Subsequently, there is no hiatus between the two components of a diphthong
that would be filled with a transitional glide. The second component involved in
a diphthong will be realised as an offglide. In Urarina, a variant of /«/ can
function as an offglide, in addition to the common [j] (for /if) and [w] (for /u/).
In addition, there is a rare diphthong /ae/. In the following, I list all occurring
diphthongs with examples for each.

(63) Diphthongs

e i u u
a ae aj au an
e —
ej — (en)
i — — — —

u — — — —

tt — — — —

(64) Examples for diphthongs

a) /aj/
/najne-a/ [naj.ni.ä] 'be able'-3ps/A
/ajrinia/ [aj.ri.ni.ä] 'outside'
/ ajtu-a/ [aj.to.ä] 'say'-3ps/A
/hajti/ [haj.ti] 'still'

b)/au/
/alau/ [a.läu] 'spider monkey'
/ itpau/ [i.taäu] 'life'
/auna-a/ [au.naä] 'hear, feel'-3ps/A
/mausa/ [mau.sä] 'blind person'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Diphthongs 61

c) / a « /
/akau/ [a.käu] 'water'
/kau-a/ [kaa.ä] 'go home'-3ps/A
/hauria/ [haö.ri.ä] 'earlier'
/auenu/ [aa.e.no] 'sacha culantro herb'

d) /a e/
/reemae/ [ree.mäe] 'dog'
/asae/ [a.säe] 'in'
/esijiae/ [e.si.jiäe] 'really'
/d3aereku-a/ [d3ae.re.k0.a] 'be angry'-3ps/A

e)/ej/
/ ejt$u/ [ej.tsu] 'cashapona tree'
/enejtp«/ [e.nej.t£«] 'monkey'
/atejjia-a/ [a.tej.jiaä] 'mortally wound'-3ps/A

f)/e«/
/karen/ /k«.ren/ 'after'

There are certain regularities that apply when full vowels are involved but do
not apply with the offglide of a diphthong. For instance, vowel copying is
triggered by /i/, /u/, or /«/ when these are followed by /h/, /k/, or Irl (for details
see §2.8.3). / r u r a n a / 'sweat' is realised as [ru.rwa.na]. However, vowel
copying does not apply through Ikl and Irl when any of the consonants
mentioned are preceded by a diphthong: /auri/ 'paucar bird' is pronounced
[au.ri], not *[au.rwi]. n
Some occurrences of diphthongs are non-phonemic. For instance, the
diphthong [e«] does not occur on the lexical level, except for one example, the
postposition kureu 'after'. However, 1 suspect that its underlying form is
/ k u r a a / i.e. [en] functions as a variant of /an/. 12 [en] also occurs in
combinations of /e/-final noun roots that are followed by the locative marker /-
ul.
The diphthong /ej/ is very rare, and there is some variation in its
pronunciation. Usually, an alternative pronunciation as [ee] is possible:

11
However, vowel copying is attested with diphthongs whose offglide spreads through
Ihl.
12
For instance, the form ku-re-ü ('go'-IRR-lsg/E) Ί would go' may be realised as
[kti.reö] or [ku.räa] in some dialects (cf. §23.1.6).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
62 Phonology

(65) Occurrence of [ej]

/enejtG«/ [e.nej.tQ«], [e.nee.tQtt] 'monkey'

/bahejrito-a/ [ba.hej.ri.to.a], [ba.hee.ri.to.a] 'be lazy'-3ps/A
/urerej/ [tt.re.rej], [ureree] 'jaguar'

In word-final position, [ej] functions as an allophone of /aj/, but is attested for a

few examples only. Alternatively, /aj/ is reduced to [e] or [i] in this context
(also cf. §2.8.7).

(66) [ej] resulting from final /aj/

/heelaj/ [hee.lej] 'same'
/komasaj/ [ko.ma.sej] 'wife'
/ekwaj/ [e.kwej] 'down'

[uj] is another diphthong that does not exist as a phoneme, but it occurs in
several examples. The fact that it is found in the same environment in all three
examples in (67) - between /k/ and /ji/ - indicates that the underlying structure
of these words is actually /kwiji.../, i.e. /kw/ followed by /i/ and /n/ (which is
palatalised in turn). As the phonetic difference between /kui/ and /kwi/ is
minimal (i.e. merely a matter of length), the forms may surface as [uj].

(67) Occurrence of [uj]

/kwijia/ [kuj.jia], [kwi.jia] 'in order to'
/kwijiadera-a/ [kuj.jia.de.raä], [kwi.jia.de.raa] 'be worried'-3ps/A
/kwijieti-a/ [kuj.jie.ti.ä], [kwi.jie.ti.ä] 'meet'-3ps/A

The representation of diphthongs as opposed to vowel sequences and long

vowels faces a problem as diphthongs show an ambivalent behaviour for tone
assignment: while they are normally treated as a single rhythmic unit and
subsequently as a single tone-bearing unit, there are specific conditions under
which a diphthong is split into two separate tone-bearing units. Normally, a
single rhythmic unit also corresponds to a single tone-bearing unit in prosodic
phonology, but this is not precisely the case here; there are specific conditions
under which a diphthong is split into two separate tone-bearing units (cf. §4).
Furthermore, some tone assignment rules are based on syllable weight; i.e. they
apply to heavy syllables ((C)VV) and these include syllable nuclei that
represent long vowels or diphthongs (cf. §4.6.1).
In segmental phonology, diphthongs are treated as a single rhythmic unit
being the nucleus of a light ((C)V) syllable. This becomes evident from a rule
that applies to verb roots that end in a vowel sequence or in a long vowel, but

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nasalised vowels 63

does not apply when the root ends in a diphthong (cf. §3.1.3). Thus, the
distinction between diphthongs and vowel sequences is reduced to the
segmental level; the distinction between diphthongs and long vowels is a
different one on the segmental and on the prosodic level.

2.7 Nasalised vowels

Nasalisation is a lexical feature of Urarina phonology, i.e. nasalisation on

vowels or diphthongs in roots or suffixes is not predictable. It can be assumed
that all vowels and diphthongs have nasalised counterparts, but on the lexical
level, not all of these are attested due to limitations of the database. The words
in (68) are examples for the following nasalised vowels and diphthongs. (Note
that nasalisation on long vowels and diphthongs is marked on one vowel only
for typographical reasons, but both V units are nasalised.)

(68) Nasalised vowels and diphthongs

/ä/ / aräala/ [a.räa.lä] 'tapir'
/e/ / meieer a / [me.lee.rä] 'servant'
Ν /aal/ [aa.T] 'jaguar'
/ü/ / rüa/ [rü.ä] 'side'
/«/ /ekö«/ [e.kü«] 'above'
/ä«/ /häö/ [häö] 'because'
/äe/ /tpäe/ [t£äe] 'also'

In addition to lexical nasalisation, all vowels and diphthongs can be nasalised as

a result of assimilation. Nasalisation spreads rightward through certain
segments (cf. §2.8.10).

2.8 Automatic phonological alternations

Phonological alternations in Urarina do not typically distinguish between types

of morphological units they apply to. They freely apply in root-internal position
and between roots and suffixes (or between different suffixes that follow each
other). This regularity also largely includes clitics. There is only one rule that
arguably does not apply across a clitic boundary: vowel copying is only attested
root-internally and across suffix boundaries, but not across clitic boundaries (cf.
§2.8.3). It must be noted that many of the alternations discussed in this section
involve the combination of roots with suffixes that begin with a vowel. This
creates a natural problem for the analysis of clitics, as none of the clitics

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
64 Phonology

occurring in Urarina begins with a vowel. Therefore, it can neither be proven

nor disproved whether these follow the same rules or not. However, given the
fact that clitics largely follow the same rules as suffixes in some other cases, I
will assume that a distinction between suffixes and clitics with regard to
phonological alternations is only marginal. It must also be mentioned that there
are no phonological rules that apply across word boundaries.
In addition to the automatic phonological rules discussed here, there is a
wide range of rules that apply to specific morphemes only. These are discussed
in chapter 3.

2.8.1 Insertion of transitional glides

Urarina has four transitional glides whose occurrence can be predicted in terms
of their phonological environment, [j] is realised in hiatus in vowel sequences
where the first vowel is /[/. (Note that the underlying vowel is Id in some cases,
but realised as [i] before /a/; cf. §2.8.4.)

(69) [j] in hiatus

a) In root-internal position:
/iana/ [i.ja.nä] 'older sibling'
/kulia/ [ku.li.ja] 'mass'

b) Between root and suffix:

/siiri-a/ [siiri.jä] 'have'-3ps/A
/ki-a/ [ki.ju] 'eat'-IMP

The labial glide [w] is realised in hiatus in vowel sequences where the first
vowel is /u/.

(70) [w] in hiatus

a) In root-internal position:
/rüa/ [rü.wä] 'side'
/etue/ [e.to.we] 'owl monkey'

b) Between root and suffix:

/u-a/ [u.wä] 'die'-3ps/A
/ajtu-a/ [aj.to.wa] 'say'-3ps/A

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Automatic phonological alternations 65

When /«/ is followed by another vowel (which is not /«/), the voiced velar
approximant [u\] functions as a glide between the two. The examples in (71)
show the application of this rule in root-internal position and between root and
suffix.

(71) [uj] is hiatus

a) In root-internal position:
/kaane/ [k«.u|a.ne] 'in'
/barue/ [ba.ra.iqe] 'masato'(cassava beer)
/ahjaui/ [a.hjatt.uji] 'obilla tree'

b) Between root and suffix:

/ra-a/ [rn.tqa] 'find'-3ps/A
/here-köra-i/ [he.re.ka.r«.u}l] 'want'-PL-NEG:3ps/A
/kau-a/ [kaa.tqa] 'return'-3 ps/A

When /«/ is followed by another vowel and it is preceded by a nasal onset or is

nasalised, the transitional consonant will be realised as a velar nasal [g]. This
applies to root-internal position, as well as between root and suffixes. Also note
the realisation of nasal spreading on vowels (cf. §2.8.10).

(72) Occurrence of [g]

/ amu-a/ [a.mu.gal 'walk'-3ps/A
/enua/ [e.nti.rja] 'tree'

Note that in the example enua above, nasalisation spreads all the way through
two vowels and /h/ from the initial consonant /n/ (thus surfacing [e.nü.gä'] in a
narrower transcription). This rule does not apply when a non-transparent
consonant (i.e. all except /h/) interrupts the nasal spread, as the example nukue
[ n u k t f i q e ] 'creek' illustrates: in this case, the transitional consonant is not
nasal: * [ n a k » g e ] (cf. §2.8.10)
[g] may also occur in loans before a velar stop. It should be noted that the
velar nasal also occurs in the respective source languages and has been adopted
into Urarina. The word 'minga' shown in (73) also has a velar nasal in Spanish
and according to Adelaar (2004, p.c.) "All the Quechua dialects assimilate the
nasal to velar position before a velar consonant." Since the occurrence of
[g] does not occur in native Urarina words, I do not consider CV+[g] a separate
syllable type.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
66 Phonology

(73) Occurrence of [g] in loans

/letQugka/ [le.t^ug.ka] Ί00'(loan from Quechua)
/miigka/ [miirj.ka] 'minga' (working session; loan from Spanish)

2.8.2 Assimilation of 7u/

In (62) it was shown that sequences of /«/ and fuf do not occur. This regularity
is due to the fact that assimilation takes place when /«/ is preceded by /u/. A
typical example for this is the suffixation of the 3pl suffix -uru to a verbal root
ending in /u/.

(74) /«/ [u], [o] with plural suffix

a) Assimilation to [u]:
/ku-uru-i/ [kuu.ra.i] 'drink'-PL-3ps.NEG/A
/ruku-«ru-a/ [ru.kuu.ra.a] 'pull out'-PL-3ps/A

b) Assimilation to [o]:
/masu-urn/ [ma.so0.ru] 'parakeet'-PL
/enu-ura-a/ [e.noo.ra.ä] 'enter'-PL-3ps/A

Corresponding examples of assimilation are found with other suffixes that

underlyingly begin with the vowel /«/, such as the form -ü for lsg/E and the
imperative -u.

(75) /«/ -> [u], [o] with lsg/E or imperative suffix

a) Assimilation to [u]:
/suru-ü/ [su.rüii] 'run'-lsg/E
/su-«/ [suu] 'kill'-IMP

b) Assimilation to [o]:
/ajtu-«/ [aj.töo] 'say'-lsg/E
/ajtu-«/ [aj.too] 'say'-IMP

While one might suspect that the assimilation of /u/ could be a morpho-
phonological process, as it is mainly observed with the plural suffix and with
the imperative form, it has to be acknowledged that Urarina does not have any
root-internal sequences of /u/ + /«/. It may thus be assumed that the absence of
such a sequence is due to the fact that assimilation has been a result of
diachronic variation as well. Since there is no material available to prove such

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Automatic phonological alternations 67

c h a n g e s , I will p r e s u m e that the rule that assimilates / « / to [u] or [o] is not a

special property of t h e plural and imperative suffixes, but a universal regularity
of Urarina phonology.

2.8.3 Vowel copying

T h e p r o c e s s of v o w e l c o p y i n g can be characterised as h o m o r g a n i c v o w e l
insertion a f t e r certain consonants, which include /h/, /k/, and Irl in Urarina. For
e x a m p l e , the c o n s o n a n t /h/ is a f f e c t e d by the presence of /i/, /u/, and / « / . A s a
result o f an optional rule, these v o w e l s spread rightward t h r o u g h the c o n s o n a n t
and are realised as palatalisation, labialisation, and "retroflexisation",
respectively. N o t e that all three s o u n d s that are a f f e c t e d by this rule are velar or
post-velar. V o w e l c o p y i n g operates word-internally, b e i n g in roots or roots with
derivational or inflectional affixes. For e x a m p l e , the labial glide o c c u r r i n g with
the /h/ in the word duhwa 'cedro masha tree' can be a s s u m e d to be a result of
/u/ s p r e a d i n g rightward. T h e r e are no e x a m p l e s in the language w h e r e a /h/-
initial syllable f o l l o w i n g a /u/ w o u l d not be labialised. In the s a m e w a y , /i/
carries t h r o u g h /h/ and c h a n g e s it into [hj], / « / operates in the s a m e w a y and
results in a m o d i f i e d variant of /h/, which I transcribe as [ h u ] here. Phonetically,
the /h/ is fronted and could probably be transcribed as [x] or [ χ ] (voiced
velar/uvular fricative). E x a m p l e s f o r all three alternations are given b e l o w . T h e y
illustrate both root-internal alternation and vowel c o p y i n g across s u f f i x
boundaries.

(76) Vowel copying through /h/

a) Copying of /u/ through /h/

/laauhiri/ [laa.u.hwi.ri] 'small (thing)'
/tasinuha-a/ [ta.si.jio.hwaa] 'be big'-3ps/A
/rautu-he-i/ [rau.to.hwee.T] 'be calm'-CNT-PRT

b) Copying of /i/ through /hI

/enaniha/ [e.na.ni.hjä] canoe
/kiha/ [ki.hjä] 'paddle'
/kutia-e/ [ku.ti.hje] 'call'-HORT

c) Copying of /u/ through /h/

/kuhana/ [ka.h u a.na] 'wind'
/nähe/ [nH.h u e] 'penis'
/kamu-ha-!/ [ka.m«.h u aa.T] 'bathe'-CNT-PRT

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
68 Phonology

In a similar manner, vowel copying may also occur through the velar stop IVJ.
However, in this case, the rule is restricted to the vowels /u/ and /«/. With /i/, no
vowel copying is observed. I have demonstrated in §2.1.6 that /kw/ functions as
a phoneme in word-initial position. If w e look at its occurrence in word-internal
position (including morphologically complex words), it becomes evident that
the labial portion of this sound is a result of vowel copying. (77) lists some
examples for this and also illustrates this process with the vowel /«/.

(77) Vowel copying through /k/

a) Vowel copying of /u/:

/kukajt^a/ [ko.kwaj.tea] 'inhabitant'
/ukana/ [u.kwa.nä] 'field'
/kukaerate-a/ [ku.kwae.ra.ti.ä] 'defend'-3ps/A

b) Vowel copying of /«/:

/«katu-a/ [tt.kua.to.a] 'be deep'-3ps/A
/ukari/ [u.k u a.ri] 'giant anteater'
/kwaauk-e/ [kwaa.H.k u e] 'think'-3ps/E

Word-internal [kw] is almost always preceded by /u/. However, this rule does
not apply to clitics. For example, a /kw/-initial word can be preceded by a
proclitic that does not contain /u/, such as in ka=hvedaj 'my visitor', where the
proclitic functions as a possessive marker for lsg. Another exception regards
the distributive suffix -ahua which in itself illustrates that /kw/ is not
necessarily the result of vowel copying.
Another consonant that can be affected by vowel copying is /r/, but the rule
tends to have a much more optional status in this case. In particular, its
application is limited to only a few speakers of the Espejo dialect, and it is less
common in other dialects (cf. §23.1.2). For instance, the word for 'Urarina' can
be pronounced with or without vowel spread: [urwarijia / urarijia]. A few
more examples for vowel copying through Irl are given in (78). Note that again,
/i/ is not subject to this rule. With /«/, the process is highly optional and rare.

(78) Vowel copying through Irl

a) Copying of /u/ through Irl·.

/kuraate-a/ [ku.raa.ti.ä], [ku.rwaa.ti.ä] 'feed'
/erari,/urari/ [u.ra.ri], [u.rwa.ri], [e.rwa.ri], 'urari root'
*[e.ra.ri)
/haatura-a/ [haa.tu.rä], [haa.tu.rwa] 'be ready'-3ps/A

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Automatic phonological alternations 69

b) Copying of Μ through Irl:

/kuraana/ [ku.raa.nä], [k«.r u aa.nä] 'chief; 'male'

The example urari is noteworthy as the actual pronunciation of the word in the
Espejo dialect is [ e r w a r i ] ; in this case the labialisation has remained although
the initial vowel was replaced by /e/ (the pronunciation urari is from the
Chambira dialect). Also note that there are counterexamples, showing that the
rule does not consistently apply in all words:

(79) Absence of vowel copying through Irl

/lureri/ [lu.re.ri] 'house'

/huura-ka/ [huu.ra.ka] 'open'-3ps/A (*[hu.rwa.ka])
/kurahe-a/ [ku.ra.hi.ä] 'desire' (but [ku.rwa.hi.a] = 'be happy')

There also is a palatalised form of the rhotic. However, this is not the result of
vowel copying, as one could expect, [rj] only occurs as a shortened version of
Iril before another vowel, as illustrated below.

(80) [rj] as a result of vowel copying

/ne-laaria-a/ [ne.laa.rjaa] ITR-'sit'-3ps/A

/nuriu/ [nu.rju] 'Marafion River'
/hauria/ [han.rja] 'earlier'

The vowel copying rule through Irl does not apply to clitics: when the emphatic
attitudinal marker =ra is attached to a host, it is never realised as [rwa] even
when the vowel lul precedes it. Similarly, vowel copying does not apply
through the reportative enclitic =he. This is in so far remarkable as enclitics
otherwise follow the same rules as suffixes.

(81) Absence of vowel copying through enclitics

a) With enclitic =ra:

/hetau=ra/ [he.tau.ra] (Hearsay evidential + EMF)
/ajtu-ö=ra/ [aj.töo.rä] 'say'-lsg/E-EMF

b) With enclitic =he:

/d3atoane-i=he/ [d3a.to.a.nee.i.he] 'be how'-PRT=REP
/su-a=he/ [suu.he] 'kill'-IMP=REP

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
70 Phonology

In summary, one can say that vowel copying occurs through consonants whose
place of articulation can be characterised as "back" as it involves velar,
retroflex, and glottal consonants.
2.8.4 Vowel raising

An important rule that accounts for the absence of the sequence /e.a/ in roots, as
mentioned in (62), is dissimilatory vowel raising: when /e/ is followed by /a/, it
is raised to [i]. The same rule also applies between roots and suffixes, or
between two suffixes: when a root-final Id is followed by a suffix that begins
with /a/, it is raised to /i/. When it is followed by a suffix that begins with any
other vowel, /e/ remains unchanged.

(82) Vowel raising /e/ [i]

a) Vowel raising applies before /a/:
/te-a/ [ti.ä] 'give'-3ps/A
/ here-a/ [he.ri.ä] 'want'-3ps/A

b) No vowel raising before other vowels:

/te-urti-a/ [tee.u.ru.a] 'give'-PL-3ps/A
/te-i/ ttee.T] 'give'-PRT

The fact that /e.a/ is not a possible vowel sequence of Urarina indicates that the
alternation described here is of a general nature.

2.8.5 Simplification of/aü/

The nasalised diphthong /a«/ can be simplified by deleting its component /«/ (in
stressed or unstressed syllable). Typically, the vowel /a/ is lengthened to [äa],
but shortening to [ä] may also occur. This rule is optional and speaker-
dependent, but it can occur in any position in a word.

(83) Simplification of /an/

/hat*/ [häu], [hää], [hä'] 'because'
/here-an«/ [he.ri.t£äö], [he.ri.tQäa], [he.ri.tpä'] 'want'-lsg/A
/aj-ri-a-ü=ni/ [aj.ri.äa.ni], [aj.ri.äa.ni], [aj.ri.ä.ni] 'do'-IRR-NEG-lsg/A=ASS

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Automatic phonological alternations 71

2.8.6 Affricatisation /ri/ ~ [rj] ~ [d?J

An optional rule observed in rapid speech only is the occurrence of the voiced
affricate [CI3] in word-internal position. As mentioned earlier (cf. §2.8.3), the
underlying syllable /ri/ can be reduced to [rj] in rapid speech if another vowel
follows. The resulting sequence [rj] + [V] is optionally realised as [ d 3 ] + [V], as
the following examples show. Note that all instances of this phenomenon shown
in (84) occur on the last syllable of the word, (which carries a Η tone).

(84) /ri/ ~ [rj] ~ [d3]

/hanria/ [hau.rja], [hau^a] 'earlier'
/ajriu/ [aj.rjii], [aj.dju] 'Airico River'
/kuriae/ [ku.rjäe], [ku.d3ae] 'far; distance'
/nemttri-«/ [ne.mu.rju], [ne.mu.d3«] 'lake'-LOC ('in the lake')

Note that the last example nemurju involves the locative suffix -«, which
merges with the final root vowel of any word it is attached to (cf. §3.4) and
therefore changes the final syllable into [rj«] as the result of a morpho-
phonological rule. The syllable is then further changed into its alternant ^ 3 0 ] .
The rule does not apply to other suffixes that begin with /«/. For instance, the
lsg/E suffix -ü does not merge with the last syllable and subsequently is
realised as a syllable sequence /i.«/.

2.8.7 /aj/~~[ej]

In word-final position, the diphthong /aj/ is pronounced as [ej]. As a further

simplification, [ej] can be monophthongised to [e]. As an exception, with the
word kanaanaj 'child', the diphthong tend to surface as [i].

(85) /aj/ [ej]

/kanaanaj/ [ka.naa.nej], [ka.naa.ni] 'child'
/raj/ [rej], [re] 'for/to'
/helaj/ [he.lej], [he.le] 'separate'
/komasaj/ [ko.ma.sej], [ko.ma.se] 'wife'
/itQa-i/ [i.tpej], [i.tpe] 'do'-NOMobj ('what he did')
/lanaha-i/ [la.na.hej], [la.na.he] 'be red'-NOMsbj ('red one')

The last two examples in (85) involve a suffix that merges with the vowel
preceding it and the final syllable therefore should be realised as [aj]. Through

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
72 Phonology

application of the described rule, words such as lanahaj are pronounced as

[lanahej] or [lanahe].

2.8.8 Insertion of glottal stop [?]

As mentioned in §2.5, the glottal stop has no distinctive function on the lexical
level. There is only one root (and derivations of it) in which ΠI occurs: the word
for 'shout/cry', e?eeoka (and the related form e?ee.ehoa). As this example may
be considered as onomatopoeic, it can be characterised as marginal anyway.
Interestingly, the glottal stop seems to function as a boundary segment between
two identical vowels of which the second one is long. This matches similar
constellations in words where the glottal stop occurs at suffix boundaries: ft!
occurs in vowel sequences in which a short vowel is followed by a long vowel
(or by another vowel sequence). In the negative form of some verbs, the glottal
stop is predictable by this rule: in [kwa.ra. ? aa.ü] Ί have not seen it', it
functions as a boundary marker between the root-final /a/ and the negative
suffix -a (followed by the lsg/E suffix -«). Similarly, it forms a boundary
between /a/ and -e-i (NEG-2ps, with the negative suffix being lengthened before
the 2ps suffix), as in [i.tpa. ? ee.i] 'you have not done it'. Apparently, the
insertion of /?/ is only necessary when a short vowel is followed by a long
vowel or by a vowel sequence; it does not apply in the opposite direction: in
[i.teaa.i] (/itpa-i/ - 'do'-NEG-3ps/A) 'he did not do it', a VV syllable is
followed by a vowel without involving a glottal stop (note that vowel
lengthening of the root-final vowel occurs before the 3ps negative form).
Since the occurrence of the glottal stop is predictable, it is not regarded a
phoneme. Regarding the nature of the examples described above, it can be
stated that the insertion rule for the glottal stop applies root-internally and
across suffix boundaries.

2.8.9 Palatalisation of Μ after ///

When the alveolar nasal /n/ (in its function as the onset of a following syllable)
is preceded by the front vowel /i/ or its realisation as an offglide [j], it is
palatalised to [ji].

(86) /η/ |ji]

a) Root-internal position:
/urarina/ [u.ra.ri.jia] 'Urarina'
/biina/ [bii.jiä] 'old person'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Automatic phonological alternations 73

/ ajna/ [aj.jiä] 'and, with'

b) Between root and suffix:

/ari-na/ [a.ri.jiä] 'seek'-INF
/ nesari-nakaar« / [ne.sa.ri.jia.käa.ä.rn] 'hunt'-'those who'
/siiri-naa/ [sii.ri.jiaa] 'have'-NOM

This rule also applies across clitic boundaries. As exemplified in (87), this
includes the final negative enclitic =ne, the subordinate marker =ne (with
conditional function here), the interrogative marker for polar questions =na, and
the attitudinal markers =naare and =naate.

(87) Palatalisation of /n/ to [ji] between suffix and clitic:

/be-i=ne/ [bee.i.jie] 'tell'-NEG.IMP=NEGF

/ajtu-i=ne/ [aj.to.i.jie] 'say'-2ps=CND ('if you say')
/kwitttk«-i=na/ [kwi.tH.ku.i.jia] 'know'^ps^INT ('do you know?')
/kutaj-ri-tQa=i=naare/ [ku.taj.ri.tpaa.i.jiaa.re] 'scold'-IRR-3ps=ASS=WRN
('he will punish him')

Palatalisation, such as all other rules discussed in this chapter, does not apply
across word boundaries, as in (88), where the initial /n/ of 'say' remains
unaffected by the preceding /i/. Interestingly, palatalisation occurs with the
copular auxiliary nei (which is a form of the copula), when it is attached to
numerals, nouns, and passives. However, the status of the auxiliary as a clitic-
like element is further discussed in §2.11.

(88) No palatalisation of /n/ to [ji] across word boundary

a)
kurete-i na-1 [ku.re.tee.i naa.T]; *[jiaal]
buy-2ps say-PRT 'saying you bought it'

b)
nii nukue [nii nn.ku.e]; *[jia.k«.e]
that river 'that river'

c)
ahina-i nii anofioa [a.hi.jiaa.i jiii a.no.fwä]; *[jtii]
sharp-NEG:3ps/A that knife 'That knife is not sharp.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
74 Phonology

2.8.10 Nasal spreading

Nasalisation generally spreads rightward through adjacent vowels and through

the consonant /h/. This rule applies on the word level and includes enclitics. It
can be initiated through nasal consonants or nasal vowels. Nasalisation affects
any non-consonantal segment if it is preceded by any nasal segment (including
nasal consonants). Typically, the syllables following a nasalised vowel will be
affected in that they are nasalised if their onset consonant is /h/. Nasalisation
will spread to adjacent vowels or diphthongs on its right until it is blocked by a
consonant (which is valid for all consonants except /h/).

(89) Nasal spreading

a) Root-internal spreading:
/rüa/ [rü.ä'j 'side'
/seeuha/ [see.ö.hwä'] 'big (thing)'
/ahäauri/ [a.hää.öri] 'turtle' (type)
/nuhe/ [nö.h*ge] 'excrements'

b) Spreading across suffix boundary:

/lana-hee-ka/ [la.nä.hee.kä] 'be missing'-DIM-3ps/A
/ amti-i/ [a.mö.T] 'walk'-2ps

c) Spreading across clitic boundary:

/ini-ri-tpa=i=he=ra/ [i.ni.ri.tcaa.lhe.ra] 'go.up'-IRR-3ps/A=ASS=REP=EMF
/am«-a=he/ [a.mü.ä.he] 'walk'-3ps/A=REP

However, nasalisation appears to be a feature that depends at least partially on

the speaker. Some speakers tend to spread nasalisation in both directions. This
makes distinctions particularly difficult as nasalisation is otherwise distinctive,
especially where suffixes are involved. Ambiguities will also occur where a
suffix -i, which may indicate NEG:3ps/A, or 2ps, is attached to a verbal root
that has a nasalised vowel, or a nasal consonant. For instance, the copula root
/ne/ spreads its nasality to an attached suffix -i intended for 2ps: 'you w e r e ' .
However, as the realisation of the form will be [nee.i], it can also be interpreted
as NEG:3ps/A 'he was not', or as the participle form 'being'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Syllable structure 75

2.9 Syllable structure

The prototypical syllable type found in Urarina is CV. However, syllables do

not require an onset. Besides, long vowels and diphthongs exist, as discussed
above, which gives a range of the following syllable types: V, VV, CV, CVV
(where VV stands for a long vowel). Coda consonants do not occur, and since
the insertion of [h] after short vowels is not phonemic, I do not list CVC here.
The following table gives an overview of co-occurring syllable types. In
particular, such sequences that would result in two adjacent long vowels are
excluded. The example for CVV.CVV is untypical as well, and no examples are
attested for words that contain VV.CVV sequences, involving long vowels only.
The absence of CV.VV may be based on the fact that the number of VV
syllables is generally lower than others.

(90) Combination of syllable types

jst j 2"d V VV CV CVV
V e.o.ri e. ? e.o.ka a.te a.laa
'termite nest' 'shout' 'fish' 'aguaje tree'
VV a.l — ee.ne —

'jaguar' 'woman'
CV ba.a — be.ru ku.raa
'bag' 'way' 'name'
CVV fwe.i — kwa.taa kö.raa.naa
'firewood' ('so that not') 'chief

Diphthongs, which I represent as having a single V position here, occur in the

following combinations:

(91) Syllable sequences involving diphthongs

V.CVjDipl,,!,] /ekwaj/ [e.kwej] 'down'
CV.CV[Diphlh] /tajbinaae/ [taj.bi.jiaa.e ] 'animal'
cv.v [ D i p h t h ] / niej/ [ni.jej] 'not at all'
vv.cv [ D i p h l h ] /äasaj/ [äa.sej] 'wicked'
V[Diphth].V /aua/ [a.wä] 'sister'
V[Diphth]-CV / ajna/ [aj.jiä] 'and'
cv [Diphth] .v /baja/ [ba.jä] 'after'

Syllable weight plays an important role for certain tone assignment rules (cf.
§4). An aspect that has not been explored in detail regards the circumstances
under which moraic structure is significant. In most contexts, a heavy syllable
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
76 Phonology

(represented as (C)VV) will be handled as one rhythmic unit. However, there

are some examples in which heavy syllables are split into two units for certain
tone assignment rules (cf. (178) on "tone type B" nouns, which assign a specific
tone pattern to a verb that follows). For instance, in constructions like [o.ba.na
ra.a.kää] 'he carried a peccary', the final syllable [kaa] carries a HL contour
tone, which can be analysed as consisting of a Η plus a L tone. However, such
structures only occur in trisyllabic words with final heavy syllables. The
example [ k w a . r ä a . i ] 'he did not see it' proves that heavy syllables in other than
final position are not affected for this specific tone assignment rule.
Interestingly, in another context, the weight of the first syllable of a word is
significant (also cf. (194)): in possessive constructions where a possessed noun,
(which follows the possessor), is in the plural form, its Η tone shifts from the
second to the initial syllable. Again, this regularity is restricted to this specific
construction. While it remains less than clear in what way moraic structure
should be taken into account, I adopt the notion of morae as significant for the
description of tonal structures in Urarina (cf. §4).
Another remark is in order on the syllable structure observed with some
loans. In particular, many loans from Spanish exhibit consonant clusters (CC) in
a syllable onset. While syllables involving a CC onset are absent from Urarina
phonology, the realisation of these depends much on a speaker's proficiency in
Spanish. Speakers who have none or little knowledge of Spanish, tend to
syllabify loans with intermediate vowels, such as (92). In contrast, speakers
who are proficient in Spanish, tend to pronounce the same loans with a
consonant cluster. Between the two, any transition between strict CV structure
and perfect Spanish pronunciation can be observed. Similarly, a final consonant
in Spanish will be realised with vowel epenthesis in Urarina. The kind of vowel
to be inserted (in both cases: CC cluster or final C) sometimes is resolved by
vowel harmony, as the inserted vowel tends to be identical to the vowel in an
adjacent syllable. Alternatively, the inserted vowel may be realised as a reduced
variant of /«/.

(92) CC clusters and coda consonants in loans

a) Initial CC cluster:
[kti.rä.bo] 'nail'; Span, clavo: [kla.bo]
[ta.rää.fwa] 'fishing net'; Span, trampa: [tram.pa]
[te'.fwo.lo] 'church building'; Span, templo: [tem.plo]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 The notion of "Phonological Word" 77

b) Coda consonant:
[do.ko.to.ro] 'doctor'; Span, doctor, [dok.tor]
[ba.ra.de] 'bucket'; Span, balde: [bal.de]
[be.se.ta] 'party'; Span .fiesta: [fjes.ta]

2.10 The notion of "Phonological Word"

Many languages distinguish the concept of "Phonological Word" from

"Grammatical Word". A phonological word is usually defined by the following
criteria:
- It is larger than the syllable
- It exhibits special word boundary or pause phenomena
- It carries stress or tone
- It exhibits specific prosodic features such as nasalisation or vowel harmony
- There are phonological rules that apply only within the phonological word
In comparison, grammatical words are often characterised in the following way:
according to Dixon & Aikhenvald (2002:19), "a grammatical word consists of a
number of grammatical elements which:
(a) always occur together rather than scattered through the clause (the criterion
of cohesiveness);
(b) occur in a fixed order;
(c) have a conventionalised coherence and meaning."

There is no doubt that Urarina recognises the notion of "Word" as opposed to

smaller units such as suffixes and clitics. However, it appears that the
phonological word and the grammatical word always coincide. As will be seen
in §2.11, the involvement of clitics hardly changes this attitude: clitics behave
largely in the same way as suffixes and form a word together with their host.
However, clitics exhibit a few properties that are different from affixes. Another
aspect related to the notion of "word" is word length, which is investigated in
the following section.

2.10.1 Word length

There is a tendency in Urarina to prefer words that are longer than a CV

syllable. However, this cannot be formulated as a minimality constraint for the
length of words: The following examples represent some of the shortest
possible words in Urarina. These include a few 3ps/E forms of verbs that have a

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
78 Phonology

C V root (plus a zero suffix for 3ps/E) and a few nouns. 1 3 The word for 'skirt' is
hu (referring to the cloth women wrap around their waist); 'rat' is sa; besides
that, there is the word u 'death'. The latter obviously is a derivation of the verb
ua ' d i e ' .

(93) CV words

/te/ Ite] 'give'-3ps/E

/ne/ ine] 'be'-3ps/E
/be/ [be] 'tell'-3ps/E
/hu/ [hü] 'skirt'
/sa/ [sä] 'rat'
/u/ [ύ] 'death'
/d3a/ td3ä] 'what'
/d3«/ td3«] 'where'
/k«/ [ktt] 'here/there' (past reference)
/ke/ [ke] (Valency increasing/instrumental postposition)

Otherwise, the majority of Urarina words are not only longer than CV, but
typically have more than one syllable. Most monosyllabic nouns and verbs (and
a f e w other monosyllabic words) contain a long vowel or a diphthong.

(94) CVV words, CV words containing a diphthong

/ha-a/ thaä] 'make'-3ps/A

/ra-a/ [raä] 'receive'-3ps/A
/sa-a/; /saa/ [saä] 'end'-3ps/A; 'rat'
/hau/ [häö] 'because'
/aj/ läj] 'do'-3ps/E (marked by a zero suffix)
/raj/ [räj] 'for/to'
/ka-«/ [käu] 'this'-LOC (= 'here')

Looking at nominal roots, it is evident that most nouns are trisyllabic (talking of
simplex nouns, i.e. strictly underived, canonical words that can occur in
isolation without further inflection). This result corresponds to a count
conducted by Cajas Rojas & Gualdieri (1987:43); their numbers are indicated in
the rightmost column of the table in (95). 14

13
"Zero suffix" more exactly is the result of merging between the 3ps/E suffix -e with a
root which ends in /e/ or /i/, as discussed in §3.4.
14
Cajas Rojas and Gualdieri (1987) state they based their count on 1,000 nouns, but do
not make further specifications on how they define this group. Given that they
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 The notion of "Phonological Word" 79

(95) Nouns: syllable count out of 700 (underived) simple nouns

Olawsky Percentage CR&G

1 syllable 10 1.4% 0.5%
2 syllables 219 31.3% 20%
3 syllables 336 48.0% 44%
4 syllables 125 17.9% 27%
5 syllables 10 1.4% 8%
6 syllables 0 0% 1%

Verb roots do not occur by themselves and are neither independent grammatical
nor phonological words. They are accompanied by the neutral suffix -a (or
allomorphs -ka or -tQa) in the citation form. Their typical length can be
deducted from the table displayed in (96). 15

(96) Verbs: word length of 527 (underived) simple verbs

Citation form including -a Entries %

1 syllable 24 4.6
2 syllables 220 41.7
3 syllables 225 42.7
4 syllables 52 9.9
5 syllables 6 1.1
Total 527 100.0

As becomes evident from the above table, the average length of an underived
verb in its citation form is two to three syllables. Verbs shorter than that are
comparatively rare; only 24 verb roots with CV structure are attested (which
occur as CVV syllables in citation form, e.g. ha-a [haa] 'make'). Verb roots
with four syllables are more frequent than that, but with less than 10%
constitute a minority. This clearly confirms the tendency already observed for
Urarina nouns: roots that have an average of three syllables are the preferred
type.

include nouns with six syllables (which are absent from my count), I suspect that
they included morphologically complex nouns in addition to simple roots.
15
I define underived verb roots as not containing a derivational suffix as discussed in
§10.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
80 Phonology

2.10.2 Other properties of words

As was stated in §2.1.10, a word can have only one postvocalic aspiration, at
least in the Espejo dialect. There is no known clitic that could be added to a
word and that would increase this potential. For instance, attaching the clause-
final enclitic =ra (realised as [tpa] after /i/) to the noun /kalaui/ 'son', can have
two realisations: a) [kah.la.wi.t^a], with aspiration after the first syllable, or b)
[ka.la.wih.t^a], with aspiration after the final syllable of the noun. However,
this restriction cannot be tied exclusively to the phonological word, since it
overlaps with the grammatical word.
Another word-related phenomenon observed in Urarina regards the
distribution of tones: every word (in isolation) tends to have exactly one Η tone.
When words are combined to syntactic constructions, different rules apply, but
this will be discussed in detail in the following section.

2.11 Clitics

Clitics are relevant for the discussion of the phonology of a language as they
typically exhibit phonological peculiarities. At the same time their status is also
important for the grammatical structure of a word. More accurately, clitics
represent a special kind of unit for the distinction between "word" and "affix".
It is widely agreed that clitics are phonologically deficient from a typological
point of view. Typically, this means that a clitic does not bear prosodic features
such as stress or tone. The situation in Urarina, however, is not quite as simple
as that, since the numerous clitics found in the language can be distinguished
from each other by a range of features. The grammatical properties and
functions of clitics, and their relative order when attached to a verb, are
discussed in §12. Here I will analyse the status of the clitics listed in (97) as
elements that share properties otherwise assigned to words and affixes.
However, as will become clear, it is not possible to divide clitics into clear-cut
classes, as the features they share with each other or with affixes and words
overlap.
One factor which clearly distinguishes clitics from affixes and words is the
fact that their selectivity is not restricted to a single word class, whereas this
applies to affixes. Many clitics can be attached to a variety of categories
including nouns, verbs, adverbs or other verb classes. Beside that, some clitics
may occur on variable positions within the clause, which is not the case for
affixes. In (97), a list of Urarina clitics and their function is given.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clitics 81

(97) Urarina clitics

le= 'one'
ka= lsg possessive or object proclitic
i= 2sg possessive or object proclitic
ni= 3sg possessive
ka=, ni=, ta= Short forms of demonstratives kaa, nii, taa
=t$e Politeness
=ni Witness evidential
=he Reportative
=lu Remoteness
=tau Reassurance
=na Interrogative
=ne Negative question; Prohibitive
=te Rhetorical question
=ra Emphasis
=ta Frustrative
=naare Warning
=naate Fear
=na, =ne, =te Focus
~ne Subordination

N o t e that le= is a special case, as it functions as the root occurring in the

numeral lejhii, which can be cliticised to numerals and s o m e other words. The
realisation o f /le/ as [lej] is due to a phonological variation (cf. §3.6.3). In
counting, the combination with the suffix -hi7 applies - which in turn is a well-
formed phonological and grammatical word, but composed o f a clitic and a
suffix. The root /le/ is obligatorily cliticised to all numerals o f the decimal
system that begin with 'one', i.e. 'ten', 'hundred', 'thousand', and 'million'. In
addition, /le/ can also be attached to words such as 'day', year', etc. (cf. §5.12).
In effect, it is difficult to state whether the combination o f le= to another word
represents an instance of compounding, or whether it should be described as a
case o f cliticisation. Since /le/ fulfils the criteria o f an proclitic phonologically,
it is glossed as le= here. Its possible status as an clitic are also supported by the
fact that it can be attached to more than one word class (i.e. nouns and
numerals). le= cannot occur by itself but always needs a host to attach to.
What all clitics have in common (and what distinguishes them from words)
is that they are always attached to a "host". A s such, they are bound morphemes
and cannot be used in isolation. The distinction between clitic and affix is
mainly based on the fact that clitics are more productive in that they can attach
to different kinds o f word classes, while affixes are usually bound to one type o f
category. In addition, a combination of word plus clitic does not occur as a

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
82 Phonology

lexicalised form (while the combination of word plus affix can be lexicalised).
O n e feature that enclitics and suffixes in Urarina have in common is that
enclitics undergo most phonological alternations in the same way as suffixes
(except for vowel copying; cf. §2.8.3). Not all enclitics are subject to
phonological alternations but this is due to the fact that their initial phonemes do
not alternate in general. In practice, there are three different alternations that can
be observed with enclitics:
1. All enclitics beginning with the nasal /n/ are subject to palatalisation of their
initial /n/ to |JL] after /i/ (cf. §2.8.9).
2. Nasal spreading applies to the reportative clitic =he in the same way as to
suffixes beginning with /h/ (cf. §2.8.10).
3. In the emphatic attitudinal marker =ra, the initial Irl undergoes palatalisation
to [tp], which complies with a morpho-phonological rule discussed in §3.2.
Clitics beginning with /t/, /tp/, or /l/ are not subject to any phonological
alternations. I therefore conclude that enclitics do not differ significantly from
suffixes with respect to their segmental properties and behaviour. However, one
difference between some suffixes and clitics should be mentioned: in §3.4, I list
a number of suffixes that merge with the root of a preceding verb. In contrast,
there are no enclitics that would merge with any segmental material that
precedes. Since the merging rule is of a morpho-phonological nature (rather
than a general phonological rule), it cannot be stated with certainty that this is a
criterion for the general distinction of suffixes and enclitics, but it may indicate
a tendency for their disparity. Regarding the order between affixes and clitics,
there is uniformity among enclitics, as they always follow suffixes. Proclitics
and prefixes do not co-occur as there is only one productive prefix in the entire
morphology.
T h e major distinction in Urarina is between proclitics and enclitics. There
are only a few proclitics, whereas enclitics form a much larger class. Other
features by which a classification of Urarina clitics is possible are listed below
and partly based on Aikhenvald (2002: 43). The relative order of clitics that co-
occur with each other and their grammatical function are discussed in §12 on
verbal morphology.
T h e most frequently occurring proclitics in Urarina are bound pronominal
forms. On the one hand, these can be attached to a noun: in this construction,
they have possessor function (cf. §7), which is illustrated in (98). The second
form given here is the variant with a full pronoun. Also note that a number of
morpho-phonological alternations apply to the combination of these proclitics
and the root (cf. §3.5; §7.1.3).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clitics 83

(98) Proclitics for lsg, 2sg, and 3sg with possessor function

a) lsg:
ka=hitgana / kanu hitQäna
lsg=blowgun
'my blowgun'

b) 2sg:
i=tQuerehe / ii kioerehe
2sg=child
'your child'

c) 3sg:
n=ahitga / raj ahitQa
3sg=older.brother
'his older brother'

On the other hand, when they are cliticised to a verb, they function as an object
pronoun as shown in (99). In this function, they can be followed by verbs or
postpositions. Note that the 3sg form does not occur as an object proclitic with
verbs, as 3ps objects do not surface (cf. §18.1).

(99) Proclitics for lsg, 2sg, and 3sg with pronominal object function
a) lsg with postposition:
ka=ke / kanu ke
lsg=VLI
'with/to/at me'

b) lsg with verb:

ka=kuruatahani-u j kanu kurnatahaniu
lsg=help-IMP
'Help me!'

c) 2sg with postposition:

i-tQej / ii räj
2sg=for
'for you'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
84 Phonology

d) 2sg with verb:

i=tQuara-ri-a-ü-ni=tä j ii kioarariäünita
2sg=see-IRR-NEG-1 sg/A-ASS=FRS
Ί will not watch you.'

e) 3sg with postposition

ni=tQ0ae / aka tgoae
3sg=on
'on top of it'

The status of the proclitics as clitics as being different from words is revealed
by their tonal structure: proclitics are fully integrated into the default tonal
structure of a noun (which surfaces with a final High tone) whereas the presence
of a full pronoun triggers certain tonal patterns that are inherent to nouns (for
details cf. §4.6.1).
In addition to the pronominal clitics, there are short forms for the three
demonstrative pronouns occurring in Urarina. When cliticised to a host, these
lose their vowel length: kaa is shortened to ka=, nii appears as m=, and taa may
occur as ta=. The shortened demonstratives have the same functions as their full
counterparts and occur as modifiers in a NP or as a noun. They can be shortened
in the NP or when they function as a noun in object position. In subject position
the short forms are not attested. Similar to the pronominal proclitics, the short
forms of the numeral and the shortened demonstratives are attached to the word
that follows and form a prosodic unit with it, i.e. they lose their tone and
become part of their host with respect to tone assignment.
Urarina enclitics form a larger group than proclitics. There are three main
properties that are relevant for the distinction of the different clitic types -
selectivity, position, and tonal behaviour.
Selectivity regards the type of host a clitic can be attached to. Some clitics
are cliticised to the final word of a clause, which can be of almost any word
class, depending on the structure of the clause. 16

16
"Any" word class includes classes with a high degree of lexical content, whereas
classes with a highly functional load, such as conjunctions, introducers,
interrogatives, interjections, and particles are excluded.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clitics 85

(100) Final negative enclitic - m attached to different word classes

a) With adverb:
hjauipe hja-i kau=ne
don't urinate-2ps here=NEGF
'Don't urinate here!'

b) With noun:
ta u-ri-tg,a=i ii komasaj=ße
NEGQ come-IRR-3ps/A=ASS 2sg wife=NEGF
'Won't your wife come?'

c) With verb:
nihjauria reemae hasasa-i=ße
don't dog untie-2ps=NEGF
'Don't untie the dog!'

Other enclitics are attached to verbs only, as is the case with mate and tau.
They can be combined with other clitics, as shown in (101), where =tau follows
the reportative =he. As a matter of fact, the frequency of these clitics is
relatively low (except that =tau forms the (lexicalised) hearsay marker hetau in
combination with =he; cf. §12.3.5); The marker =mate is observed in a handful
of examples only (and all of these utterances consist of a verb only). Therefore
it is difficult to exclude the possibility that they could appear on a word
different than a verb. The "warning" enclitic =iiaare is attested in combination
with nouns, but the example given in (101b) must be regarded as a verbless
exclamation, which rules out the attachment to a verb.17

(101) Enclitics typically attached to verbs

a) naate:
kunajti-a=naate
be.sick-3ps/A=FEAR
'Is he sick?'

b) naare:
akano=naare
snake=WRN
'There is a snake!' [implies 'be careful!']

17
The insertion of =naare in this example is optional. In most situations, people would
simply shout the word for 'snake' without any additional marking.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
86 Phonology

c) tau:
aj-a hetau kiuajtei ßoaelu
AUX-3ps/A (REP=REAS) again earlier
'Then, again, earlier there was [a story].'

Some enclitics can also occur on clause introducers, as shown in (102). More
exactly, the position of the clitic depends on the introducer: with the negative
question introducer ta and the prohibitive hua, the clitic must be attached to the
introducer (cf. (102a-c)), while with kiuatia and nihjauria, there is freedom of
choice (cf. (102d-e)). With other introducers than the ones listed below, clitics
only occur on the verb (cf. §5.10).

(102) Enclitics attached to introducers

a) ta + =na:
ta(=na) u-ri-fo=~i=pe / *ta u-ri-ki=l-pe=na
NEGQ=INT come-IRR-2ps=ASS=NEGF NEGQ come-IRR-2ps=ASS=NEGF=INT
'Won't you come?'

b) ta + =he:
ta-he ate muht-i=ße / *ta ate muku-i=ße=he
NEGQ=REP fish catch-2ps=NEGF NEGQ fish catch-2ps=NEGF=REP
'Have you not caught any fish? he asks.'

c) kiua + =he\
kzua=he su-a / *hva su-a=he
don't=REP kill-NTR don't kill-NTR=REP
"'Don't kill him!" he says.'

d) kiuatia + =he:
huatia=he itQa-i / kiuatia itg,a-i=hje
not=REP do-NEG:3ps/A not do-NEG:3ps/A=REP
"'He did not do it", he says.'

e) nihjauria + -he:
nihjauria=he su-i=pe / nihjauria su-i=ße=he
don't= RE Ρ kill-2ps=NEGF don't kill-2ps=NEGF=REP
'Don't kill him! he says.'

The "final" emphatic marker =ra is almost exclusively found in clause-final

position. However, the example in (103) illustrates its use after the introducer
ta. The meaning in this case implies a sort of reproach, claiming the addressee

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clitics 87

has actually seen it, but pretends not to. Also note that =ra cannot be attached to
other introducers.

(103) Enclitic =ra after negative question marker ta

ta-ra hoara-i=pe
NEGQ=EMF see-2ps=NEGF
'Didn't y o u s e e it?'

It should be mentioned that the enclitics =lu and =m occur in combination with
=tau as taum and taulu, which are lexicalised forms with adverbial function to
mean 'earlier' (cf. §12.3.5).
Another criterion for the distinction of clitic properties is the position within
the clause, as already mentioned above. A number of enclitics appear only in
clause-final position, while some can (or must) be attached to introducers if one
is present. The focus markers =m, =m, and =te (cf. §19), which may be
attached to any lexical word class, naturally occur after the element(s) they shift
into focus. This can be a phrase, a phrase plus an adverb, or a dependent clause.
The subordinate marker =ne always follows a dependent clause, which in turn
may end in any lexical word class.
The interrogative enclitic =na exhibits a peculiarity in so far as it does not
only occur in clause-final position and after introducers, but may also be
attached to the first constituent of a clause. In this position it implies focusing
emphatic function, similar to the focus markers mentioned above. The details of
this are further discussed in §21.

(104) Interrogative enclitic =na after first constituent

a)
d$a(=na) itQa-i ßonei
what(=INT) do-2ps like.this
'What (the heck) are you doing like this?'

b)
enejtQu=na kohiuajte-i
monkey=INT like-2ps
' D o y o u like m o n k e y ? ' [implying 'What do you like - m o n k e y ? ' ]

The third distinctive feature for Urarina enclitics regards their tonal structure.
From a cross-linguistic point of view, clitics tend to be prosodically deficient;
i.e. they typically do not have stress or tone. However, about half of the Urarina
enclitics do follow this tendency in that they carry a Η tone in addition to the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
88 Phonology

tone pattern on the word they are attached to. 18 Normally, a verb has only one
High tone, regardless its number of syllables and suffixes. The position of the
tone largely depends on the tonal structure of the word preceding it. Thus, if an
enclitic is attached to a verb and forms part of its prosodic structure, it would be
expected that it complies according to the tone pattern the verb would have
without the clitic. For instance, if the pattern assigns the High tone to the first
syllable, the clitic, which is in final position, will not bear a tone, as only one
tone per word can occur. This is indeed the case for a number of enclitics,
which are marked by "No" in (105), as they do not carry a tone. In contrast,
there are enclitics that always carry a Η tone, even though a tone is already
assigned to the rest of the verb. In this respect, they behave like independent
phonological words. The clitics =ne, =tau and =m carry a tone when directly
attached to a verb, but they do not when another clitic follows, (which in turn
carries a tone).

(105) Enclitics: summary of distinctive features

Selectivity Position Tone

=ne (NEGF) Any Clause-final No
=te Any Clause-final; in Always Η
comb, with =na
=ra Any Clause-final; after Always Η
some INTR
=ta Any Clause-final; after No
some INTR
=na, =ne, =te Any After first "chunk" Always Η
=ne (SUB) Any After [DEP] Always Η (except with
other clitic)
=naare Verbs, Nouns Clause-final Always final Η
=naate Verbs Clause-final Always final Η
=tau Verbs Postverbal Always Η (except with
other clitic)
=toe Verbs, INTR Postverbal; INTR No
=he Verbs, INTR Postverbal; INTR No
=lu Verbs; taulu Postverbal; INTR No
=ni Verbs; tauni Postverbal; INTR No
=na (INT) Verbs, INTR; Postverbal; INTR; Always Η (except with
initial phrase initial phrase other clitic)

ι ο

It is necessary to comprehend the general tonal structure of Urarina verbs, but this can
only be summarised here; for further discussion, see §4.2.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clitics 89

C o m p a r i n g the respective enclitics in (105) and the features assigned to each of

them, it becomes evident that it is not practical to represent them as natural
groups or classes of clitics. The distinctive features merely help to characterise
the similarities and differences between the clitics, but these do not neatly
coincide with each other. For instance, the fact that a clitic can be combined
with any word class does not imply its position within the clause, nor does it
correlate with its tonal structure.
T h e copula ma is not listed in (105), but it is particularly interesting even
though it does not represent a typical clitic. Yet, the copula exhibits clitic-like
behaviour in at least one way not listed in the table, but mentioned further
above: in the same way as suffixes and enclitics, it is subject to the phonological
rule of palatalisation: As illustrated in (106a,b), when it follows a word ending
in /i/, the initial /n/ of the copula it realised as [ji] (except when /i/ follows; cf.
§2.8.9). This is in contrast to other verbs in the same environment: full verbs do
not undergo palatalisation in this context (cf. (106c)).

(106) Palatalisation of /n/ with copula

a)
anuri pe-ene
parrot be-NEG:3ps/E
'It is not a parrot.'

b)
kasi=ße-Ί kuraanaa
'seven'='be'-PRT chief
'seven males'

c)
anuri na-a raj kuraa
parrot say-3ps/A POSS name
'Its name is "parrot".'

This rule takes almost idiosyncratic status when it comes to the behaviour of /n/
when the copula follows the participle suffix -1, which would be expected to
trigger palatalisation, as can be observed in (107): the first occurrence of the
copula (βέϊ) undergoes palatalisation because the passive form that precedes it
ends in /i/. However, the second instance of the copula (ne) is not palatalised
even though the participle form ends in /i/ as well. There are two possible
explanations for this behaviour: Firstly, palatalisation does not occur twice in a
row; i.e. the rule applies only once per Phonological Word. In this case it must
be postulated that the form ( i t g a n o i ß e l ) represents one Phonological Word,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
90 Phonology

which could be argued for as the copular part of it does not bear its own High
tone. The second copula (ne) could be considered an independent word, carrying
a tone, and therefore not being subject to the rule. However, this does not
explain the presence of High tones on the copula when it follows a noun.

(107) Only one instance of palatalisation of /n/ after /i/

a)
nitoaneT ÜQa-mi=ße-! ne
like, that do-PASS=be-PRT be:3ps/E
'It is done like that.'

b)
kiveta-noi=ße-l ne=lu
clear-PASS=be-PRT be:3ps/E=REM
'It was being cleared.'

Alternatively, one might assume that for some reason, the palatalisation rule is
not triggered after the participle form, whereas it does apply after the passive.
This view is supported by another example, where the rule does not apply
either. In fact, palatalisation of [ne] would change the meaning: with the same
sentence ending in soijie the jussive form would be understood, to mean 'that
man ought to kill the monkey'. Thus, another possible reason for not applying
the rule could be that ambiguities are to be avoided.

(108) No palatalisation of /n/ after the participle form

kaa katg,a=te enejtgu so-l ne

this man=FOC monkey kill-PRT be:3ps/E
'This man is killing a monkey.'

Which explanation is to be favoured should remain open. However, what can be

stated is that the copula is phonologically different from other verbs.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
3 Morpho-phonology

Urarina has a wide range of phonological alternations that do not apply in all
contexts, but which are specifically associated with certain morphemes. Due to
the complexity of the applicable rules and a high degree of idiosyncrasy the
discussion may appear discouraging at first sight. While 1 outline the
regularities in a basic manner, complications and deviations from the general
rules are also investigated, which demonstrates the array of conditions involved
in formulating morpho-phonological regularities. One intriguing aspect of some
of these rules is the relation to word length and, in part, the role syllable weight
plays for the assignment of certain affixes. A predominant role in this regard is
the allomorphy of plural and person marking: many forms involve alternations
that determine the presence or absence of the consonants [k] and [t$] (cf. §3.1).
Many morpho-phonological rules apply between roots and suffixes; these are
discussed in §3.1 to §3.4. Another set of rules regards the combination of pre-
root elements and roots, which is investigated in §3.5. Whether these are
specific to the respective morphemes or whether these are rules that generally
apply between roots and proclitics cannot be determined since only a small
number of preverbal elements (i.e. prefixes or proclitics) exist. §3.6 discusses a
number of optional simplifications that typically occur in rapid speech.

3.1 Alternations involving /k/ versus no onset on suffixes

3.1.1 Plural marker -uru on nouns

The plural suffix -uru has the allomorphs -kuru, and -t$uru, which are in
complementary distribution. However, the exact conditions for their occurrence
are not straightforward and some exceptions apply. A particular problem is the
distinction of the suffixes -uru and -kuru, as the conditions under which they
occur can only be predicted in terms of statistical tendency, rather than a 100%
proof solution. As a tendency, most nouns that take the plural with -kuru are
trisyllabic, but some nouns longer than that also take this suffix.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
92 Morpho-phonology

(109) Plural marker -kuru with trisyllabic and quadrisyllable nouns

a) With trisyllabic roots:

/enejtca-ara/ [e.nej.tpn.ktt.ru] 'monkey'-PL
/karaanaa-ara/ [ka.raa.naa.ka.ra] 'chief-PL
/lerano-ara/ [le.ra.no.kara] 'macana' fish-PL

b) With longer roots:

/tajbinaae-ara/ [taj.bi.jiaa.e.ka.ra] 'animal'-PL
/kirimata-tirtt/ [ki.ri.ma.ta.ka.ra] ' boquichico fish'-PL
/akaaraa-ara/ [a.ka.a.räa.ka.ra] 'pona palm'-PL

T h e list of trisyllabic nouns that take the plural suffix -kuru also includes loans
and proper nouns, which indicates a certain productivity of this rule:

(110) Plural marker -kuru with trisyllabic loans

/makina-ara/ [ma.ki.na.kara] 'machine'-PL (from Spanish)

/kokamo-ara/ [ko.kwa.mo.ka.ra] 'Cocama' (tribe name)-PL
/inoma-ara/ [i.no.ma.ka.ra] 'Inoma'-PL (family name)

Loans that have less or more than three syllables do not strictly follow this rule:
some are pluralised by -uru, others by -htru.

(Ill) Occurrence of -uru and -kuru with disyllabic and quadrisyllable loans

a) Disyllabic loans + -uru:

/bote-ara/ [bo.te.a.ra] 'boat'-PL
/balde-ara/ [bal.de.a.ra] 'bucket'-PL
/besta-ara/ [bes.tä.a.ra] 'fiesta, feast'-PL

b) Disyllabic loans + -kuru:

/kaa/u-ara/ [kaa./u.kara] 'cashew tree'-PL
/taasa-ara / [taa.sa.ka.ra] 'cup'-PL

c) Quadrisyllable loans + -uru:

/kalamina-ara/ [ka.la.mi.nä.a.ra] 'zinc roof-PL
/ misjonero-ara/ [mi.sjo.ne.roo.ra] 'missionary'-PL

d) Quadrisyllable and longer loans + -kuru:

/motosjera-ara/ [mo.to.sje.ra.ka.ra] 'chainsaw'-PL
/ heneradoro-ara/ [he.ne. ra. doro.ka. ra] 'generator'-PL

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Alternations involving /kJ versus no onset on suffixes 93

With regard to canonical nouns, there also are exceptions from the rule that
assigns -kuru to non-trisyllabic nouns. Note that the disyllabic loans that take
-kuru in (111b) have one feature in common: they both contain a heavy
syllable. The same phenomenon is observed with corresponding canonical
nouns in (112). Some of them are monosyllabic or disyllabic and end in a CVV
syllable, while others are disyllabic and begin with a CVV syllable. Thus, it can
be assumed that this property is significant, as also some tonal rules are related
to initial and final heavy syllables (cf. §4.2).

(112) Plural marker -kuru with disyllabic nouns that contain a long vowel

a) Long syllable before plural suffix:

/taa-ttr«/ [taa.ktt.ru] 'that'-PL (to mean 'those ones')
/kuraa-ttru/ [ka.raa.ka.m] 'name'-PL
/akaa-ttra/ [a.kaa.ktt.r«] 7oeplant'-PL

b) Long first syllable:

/laano-uru/ [laa.no.ktt.rtt] 'cassava'-PL
/ eene-yru/ [ee.ne.ktt.ru] 'woman'-PL
/raana-ura/ [raa.na.k«.rn] 'white-lipped peccary'-PL

This regularity is confined to CVV syllables that contain a long vowel;

diphthongs do not underlie this rule, as the following examples show:

(113) Plural m a r k e r - a m with disyllabic nouns that contain a diphthong

/akatt-ur«/ [a.ka«.«.rn] 'water'-PL

/alau-uru/ [a. lau. ύ. ΓΗ] 'spider monkey'-PL

Other cases that deviate from the 3-syllable condition are compounds or
derivational suffixes, such as the word for 'work' (which is a nominalisation
derived from the verb). Each of the examples in (114) to which -kuru is
attached has more than three syllables.

(114) Plural marker -kuru with compounds and long derived words

/amiane-naa-ura/ [a.mi.a.ne.naa.ktt.ru] 'work'-NOM-PL

/usi+biine-uru/ [ii.si.bii.jie.ktt.rtt] 'fire' + 'spark'-PL
(= 'sparks')
/atauari+kuraanaa-ttrtt/ [a.ta.wa.ri.kti.raa.naa.ktt.ru] 'chicken' +'chief -
PL (= 'roosters')

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
94 Morpho-phonology

Despite the tendency for a predictable plural inflection, there are

counterexamples that take the plural /-tar«/, especially such words that are
trisyllabic. Note that the example hohuajtQa 'inhabitant' may form its plural
with -kuru or with the variant -uru.

(115) Plural marker -uru with trisyllabic nouns

/kokwajtga-ara/ [ko.kwaj.tga.ka.ru], [ko.kwaj.tga.a.ra] 'inhabitant'-PL

/leotga-ara/ [le.o.tga.a.ra] 'other (person)'-PL
l
/kanaanaj-ara/ [ka.naa.naj.a.ra] child'-PL
/djaruba-ara/ [d3a.ru.ba.a-ra] 'Amazonian manatee'-PL

Otherwise, most nouns that form their plural with -uru have indeed a number of
syllables that are lower or higher than three.

(116) Plural marker -uru with nouns that have less or more than three syllables

a) Monosyllabic nouns:
/kwi-ara/ [kwi.a.ra] 'medicine'-PL
/sa-ara/ [sa.a.ra] 'rat'-PL

b) Disyllabic nouns:
/katga-ara/ [ka.tga.a.ra] 'man'-PL
/ate-ara/ [a.te.a.ra] 'fish'-PL
/tihja-ara/ [ti.hja.a.ra] 'foot'-PL

c) Quadrisyllable nouns:
/tariatga-ara/ [ta.ri.a.tpa.ara] 'yellow spotted river turtle'-PL
/enaniha-ara/ [e.na.ni.hja.ara] 'canoe'-PL
/hanulari-ara/ [ha.no.la.ri.a.ra] 'jaguar'-PL (any wild cat)

The occurrence of [tp] in the plural allomorph -tQuru is relatively simple to

explain through phonological context: IkJ is realised as [tp] after /i/; i.e. some
words whose syllable structure suggests that they take the plural marker -Joiru
occur with the plural suffix -tQuru, if they end in /i/. However, there is an
additional condition that involves word length: the rule only applies to
trisyllabic words, i.e. /i/ is the nucleus of the third syllable from the left.

(117) Occurrence of-tQuru

/lureri-ara/ [lu.re.ri.tga.ra] 'house'-PL

/kalaui-ara/ [ka.la.wi.tpa.ra] 'son'-PL
/anuri-ara/ [a.nu.ri.tpa.ra] 'parrot'-PL

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Alternations involving /kJ versus no onset on suffixes 95

/kukuri-urti/ [ku.ku.ri.tga.r«] 'armadillo'-PL

While this rule applies to the overwhelming majority of trisyllabic nouns that
end in /i/, there are some exceptions, such as the monosyllabic (CVV)
demonstrative ηίί 'that', which forms its plural in -tguru. (118) lists a few more
exceptions of nouns that are not trisyllabic but still take this plural suffix.

(118) Non-trisyllabic nouns with plural suffix -tQuru

/al-ar«/ [aa.lt^tt.rn] 'jaguar'

/hareheeri-«m/ [ha.re.hee.ri.tQtt.ru] 'thin person'
/kameranati-urtt/ [ka.me.ra.na.ti.tQtt.ra] (type of banana: guineo pindorito)-VL

3.1.2 Plural marker -uru in verbs

The plural suffix -uru and its allomorphs are also used as plural markers in the
3pi form of verbs, where they precede polarity and person suffixes, among
others (cf. §12.2.12). The conditions regarding their distribution are similarly
complex as the system for noun pluralisation. With monosyllabic roots, only
-uru is used. There is a preference for -uru to occur after disyllabic roots.

(119) Plural marker -uru with monosyllabic and disyllabic verb roots

a) With monosyllabic verb root:

/sa-ttru-a/ [sa.tt.rtt.ä] 'end'-PL-3ps/A
/ne-ura-a/ [ne.a.rn.a] 'be'-PL-3ps/A
/ku-uru-a/ [kutt.ra.a] 'go'-PL-3ps/A

b) With disyllabic verb root:

/räasa-uru-a/ [räa.sa.tt.rtt.ä] 'dance'-PL-3ps/A
/ahe-ura-a/ [a.he.tt.ra.ä] 'get.drunk'-PL-3ps/A
/tura-uru-a/ [ttt.r«tt.r«.ä] 'arrive'-PL-3ps/A

In contrast, some disyllabic roots take the plural suffix -kuru. The conditions
that underlie this alternation are not predictable.

(120) Plural marker -kuru with disyllabic verb roots

/itea-ttrti-a/ [i.tga.ka.rtt.a] 'do'-PL-3ps/A

/kwara-ttrii-a/ [kwa.ra.kii.rtt.ä] 'see'-PL-3ps/A
/here-«r«-ene/ [he.re.ku.re.ne] 'want'-PL-NEG:3ps/E

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
96 Morpho-phonology

The form -kuru is also used with most verbs that have a trisyllabic root. In
addition, it is attested with some derived forms. The examples in (121b) show
two examples that contain a reflexive prefix, one case that involves a causative
derivation, and an imperfective form. These examples are so heterogeneous that
it is impossible to make any generalisations about the use of -kuru.

(121) Plural marker -kuru with trisyllabic verb roots and longer forms

a) Trisyllabic roots:
/lenone-tfru-a/ [le.no. ne.kä.rti.ä] 'eat'-PL-Ops/A
/kufwia-uru-e/ [ku.fwi.a.ktt.re] 'felP-PL-3ps/E
/kurete-«rn-e=ln/ [ku.re.te.kü.re.M] 'buy'-PL-3ps/E=REM

b) Longer verb forms:

/ ne-tabatakaae-ttru-a/ [ne.ta.ba.ta.kaa.e.ktt.ru.ä] lTR-'reduce'-PL-3ps/A
/ kwitQa-erate-ttra-e / [kwi.tea.ra.te.ktt.re] 'heal'-CAU2-PL-3pl/E
/ne-rehete-Hrti-e=l«/ [ne.re.he.te.ku.re.ltt] 'be'-HAB 1 -PL-3ps/E=REM

In contrast, the examples listed in (122) again demonstrate that -uru is the
preferred suffix with roots that have more than two syllables, some including a
reflexive prefix. This makes it even more difficult to predict the occurrence of
the respective suffix on a certain form.

(122) Plural marker -uru with longer verb forms

/ kwitttku-ara-a / [kwi.tti.ktttt.rtt.a] 'know'-PL-3ps/A

/eseneta-uru-a/ [e.se.ne.ta.ii.rtt.a] 'believe'-PL-3ps/A
/kwara-he-tirtt-a/ [kwa.ra.he.tt.ra.a] 'see'-CNT-PL-3ps/A

The last example in (122) though is remarkable: the verbal root / k w a r a / ' s e e '
occurs with the plural allomorph -kuru when the plural directly follows the root,
as was observed in (120). When the continuous aspect suffix -he is attached to
the root, the plural that follows is realised by the suffix -uru. There are a
number of possible factors that may play a role in the assignment of the plural
suffix:
a) The structure of certain roots. Note that huaraa has a few other exceptional
properties regarding its inflection. It could be a lexical exception. However,
this does not directly account for the alternation in the assignment of the
plural allomorph.
b) The number of syllables involved, kioaraa is followed by -kuru as a
disyllabic root, but takes the suffix -uru when three syllables precede. The
problem is that this is exactly against the trend, as trisyllabic roots tend to be
pluralised by -kuru.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Alternations involving /kJ versus no onset on suffixes 97

c) It is possible that certain suffixes block the use of the suffix -htru. In fact,
the continuous suffix -ahe never occurs with -htru.

Thus, one may conclude that the distribution of -uru and -kuru with verbs is
predictable to a very low degree, i.e. for monosyllabic roots. This is the only
context for which there are no examples that involve the suffix -uru. Apart
from this, numerous factors such as length, morpheme structure, and lexical
exceptions seem to play a role in plural assignment, which makes it impossible
to formulate a rule. Regarding the question which allomorph may be the
underlying one, several views are possible. While one may argue that it is more
common to described a shorter allomorph as a reduction of a longer one, the
latter (i.e. -kuru) appears to be more restricted in its occurrence: most typically,
it is used in rather specific environments, e.g. with respect to word length and
after long vowels and vowel sequences, in order to avoid an accumulation of
three vowels, as supported by the analysis of several verbal suffixes in §3.1.3 to
§3.1.6. While these are rather specific contexts, the occurrence of -uru could be
viewed as the default case.
For the occurrence of the allomorph -tg,uru, the situation is different. As one
may predict based on its occurrence with nouns, it is attached only to verbs that
have a specific structure: most importantly, their stem must end with the vowel
/i/ (and [t$] is a result of palatalisation). Thus, -tQuru may be viewed as a
variant of -kuru in a clearly defined phonological environment. Note that I use
the word 'stem' intentionally, as some examples in (123) contain a root plus
other affixes in addition to plural and person marking. In the examples with the
completive marker -si, the stem (i.e. the extended root) ends in i\!. The root for
'hunt' contains the intransitivising prefix m-. Apparently, word length does not
play a role here, as the verbs have between five and six syllables.

(123) Occurrence of -t^uru in verbs

/siiri-«ru-a/ [sii.ri.tQH.rn.a] 'have'-PL-3ps/A
/kwaasi-um-a/ [kwaa.si.t£H.r«.a] 'be.afraid'-PL-3ps/A
/ huku-si-aru-ene / [hu.ku.si.t£H.re.ne] 'stab'-CPL-PL-NEG:3ps/E
/ra-si-urti-ene/ [ra.si.t£«.re.ne] 'receive'-CPL-PL-NEG:3ps/E
/ne-sari-Hra-a/ [ne.sa.ri.tQH.m.a] lTR-'hunt'-PL-3ps/A

Taking into account what was said about the occurrence of -tQuru with nouns,
one may generalise the distribution of the three plural allomorphs as follows:
1. -uru is the unmarked plural form for most nouns and verbs, but rare with
trisyllabic roots.
2. -kuru is used under the following conditions:
a) With most trisyllabic noun and verb roots

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
98 Morpho-phonology

b) With most disyllabic roots that contain a long vowel

c) With some roots that have less or more than three syllables.
3. -tQuru consistently occurs with trisyllabic nouns that end in /i/. It also occurs
with a few exceptional roots that have less or more than three syllables. With
verbs, its behaviour is in so far predictable, as it is restricted to verbal stems
that end in i\!.

3.1.3 Verbal alternation with f k j , [tg,] in the 3ps form

As we have seen in the previous section, plural allomorphy involves alternation

between three types of suffixes: without an onset (i.e. vowel-initial), with /k/,
and with /t$/. A similar situation is found with another inflectional category.
The person suffix for 3ps/A (which also functions as the neutral or citation
form) is vowel-initial, i.e. -a for most verbs. For a number of verbs though, it is
realised as -ka (and as -t$a for a few). In (124), only examples for the 3ps with
-ka are listed. Regardless of the length of these examples, there are two
conditions under which the use of the suffix -ka is required. The first option is
when a root ends in a long vowel:

(124) -ka with verb root ending in a long vowel

/huu-a/ [huu.kä] 'flood'-3ps/A

/tukuu-a/ [tu.kuu.ka] 'sew'-3ps/A
/lanaa-a/ [la.naa.ka] 'be red'-3ps/A
/ne-ruu-a/ [ne.ruu.ka] 'come off-3ps/A

The other context in which the suffix -ka is applied is with roots that end in a
sequence of two vowels, as illustrated in (125). It does not play a role whether
the vowels involved are long, short, or diphthongs, as becomes evident: in
kamua-ka 'bathe', two short vowels occur as a sequence, in ubaae-ka 'be
angry', the first vowel is long, in huee-ka 'close', the second vowel is long, and
in kuruae-ka 'be happy', the second vowel is a diphthong.

(125) Suffix -ka with verb root ending in vowel sequence

/kamua-a/ [ka.mö.a.kä] 'bathe'-3ps/A

/ubaae-a/ [u.baa.e.kä] 'get angry'-2pl
/huee-a/ [hu.ee. kä] 'grow'-3ps/A
/kuruae-a/ [ku.ru.ae.kä] 'be happy'-3ps/A
/kwaaa-a/ [kwaa.u.kä] 'think'-3ps/A

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Alternations involving /k/ versus no onset on suffixes 99

Note that the -ka suffixation rule does not apply when the root ends in a single
diphthong, rather than in a long vowel or a vowel sequence. Words such as
han-a 'throw' and beelaj-a 'give as a gift' prove this. When a root end in a
single vowel, the 3ps/A suffix also is -a.

(126) -a with verb root ending in diphthong

/itpau-a/ [i.tpa.wä] 'live'-3ps/A

/kutaj-a/ [ku.ta.ja] 'scold'-3ps/A
/basihjati-a/ [ba.si.hjau.ä] 'steal'-3ps/A

When the root ends in a long vowel /ii/, -ka is replaced by -t$a:

(127) -tga with verb root ending in /ii/

/hanii-a/ [ha.nii.tga] 'select'-3ps/A

/kaa+nii-a/ [ka.nii.tca] 'this' + 'that'-3ps/A ('he did like that')
/batei-a/ [ba.tee.i.tcä] 'break into small pieces'-3ps/A

3.1.4 Verbal alternation with [k], ftQj in Ips forms

In the lsg/A form, verb roots that end in a vowel sequence are realised with the
suffix -käu (instead of the regular -anu). This form is based on the element /ka/,
which was discussed above, and the suffix -w, (which otherwise occurs as the
suffix for lsg/E). These are combined to replace -anu, which again cannot be
used since it is vowel-initial and would thus violate the principle that disallows
three vowels in a row.

(128) lsg/A suffix -kau with verb roots that end in long vowel or vowel sequence

/ubaae-an«/ [u.baa.e.käö] 'be angry'-lsg/A

/huee-ana/ [hu.ee.käö] 'grow'-lsg/A
/kulee-anu/ [kö.lee.käö] 'wash'-lsg/A

There are a few more verbs which take -käu instead of -anu in order to mark
lsg/A. The difficulty with these verbs is that their classification in terms of
phonological structure is not as straightforward as for the verbs mentioned
above. The only feature that these verbs have in common is that their root ends
in the vowel /a/, but this is also the case for many other verbs that are not
subject to this rule.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 100 Morpho-phonology

(129) -käu with some exceptional verb roots

/itaa-ana/ [i.t^a.käu] 'do'-lsg/A

/kwara-anu/ [kwa.ra.kä«] 'see'-lsg/A
/ auna-antt/ [au.na.käö] 'hear'-lsg/A

Other verb roots ending in /a/ take the regular inflection for lsg/A. Thus, the
verbs listed in (130) can be regarded as exceptions. In addition, verb roots that
end in a vowel other than /a/ also follow the normal pattern.

(130) Regular lsg/A marker -anu

/baha-an«/ [ba.haa.na] 'ask'-lsg/A

/ ama-an«/ [a.mu.a.n«] 'walk'- lsg/A
/ere-anu/ [e.ri.a.n«] 'speak'-lsg/A
/ ari-ana/ [a.ri.a.nü] 'seek'-lsg/A

With some verb roots that end in /e/ or /i/, the lsg/A form is realised as -tgäu.
However, this seems to be lexically prespecified as a comparison of the
examples in (130) and (131) illustrates. While the examples /ari/ 'seek' and
/ere/ 'speak' in (130), receive the suffix -anu for lsg/A, the examples in (131)
differ from that: even though the underlying final root vowel of bajtia and heria
is /e/, as can be proven by other forms (e.g. the participle, which is bajtei, here1,
respectively), they form the lsg/A with -i(sä«. 19 Correspondingly, some verbs
with roots ending in /i/ take this suffix for lsg/A.

(131) -t$au after disyllabic verb roots ending in /i/

/bajte-anu/ [baj.ti.tQäö] 'forget'-lsg/A

/here-ana/ [he.ri.tQää] 'want'-lsg/A
/kurete-arttt/ [ko.re.ti.tpä«] 'buy'-lsg/A
/laati-an«/ [laa.ti.t£äö] 'need'-lsg/A
/kwaasi-ana/ [kwaa.si.tQäu] 'be scared'-lsg/A

Note that most monosyllabic verb roots that end with an underlying /e/ do not
follow this rule. The forms of the copula ni-anu and m-tQäu occur alternatively.
Some trisyllabic roots (and disyllabic roots that contain a long vowel) may
surface with /t^/.

19
Recall that the vowel /e/ always surfaces as /i/ before /a/, as was already observed as a
general phonotactic rule in 2.8.4. This results in the forms ending in [ianu] with the
suffix -anu.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Alternations involving /k/ versus no onset on suffixes 101

(132) Occurrence of -anu and -tQäu with monosyllabic and trisyllabic verb roots
a) With monosyllabic root:
/ne-anu/ [ni.a.nu] / [ni-tcäö] 'be'-lsg/A
/ te-anu/ [ti.a.na] 'give'-lsg/A
/be-ana/ [bi.a.na] 'tell'-lsg/A

b) With trisyllabic
/kuhjuri-ana/ [ku.hju.ri.t^äa] 'shout/cry'-lsg/A
/nesari-ana/ [ne.sa.ri.tpäa] 'hunt'-lsg/A
/kurete-ana/ [ku.re.ti.t$äa] 'pay'-lsg/A
/kanajte-ana/ [ka.naj.ti.tpaa] 'be sick'-lsg/A

Similar to the alternation for the lsg/A form, all lpl suffixes have allomorphs
beginning with /tp/ when they are directly attached to the root of any verb listed
in (132b). The regular suffixes are -akaanu (lpl/ex), -aka (lpl/du), and -akaatQe
(lpl/in). The alternations with [tp] are illustrated in (133).

(133) Occurrence of -tQaanu

/najne-akaana/ [naj.ni.tciaa.n«] 'be able'-lpl/ex

/nesari-akaana/ [ne.sa.ri.t^aa.na] 'hunt'-lsg/E
/laate-akaatce/ [laa.ti.tQaa.t^e] 'need'-lpl/in
/sini-akaatpe/ [si.ni.t^aa.tpe] 'sleep'-lpl/in

Another example for the systematic occurrence of [t£] is in the future tense.
After the tense/irrealis suffix - n , both lsg and 3ps personal suffixes are realised
with an initial [t$]. However, note that there is no assimilation of /k/ to [tp], as
could be expected in forms that involve the 2ps suffix -ki.

(134) [tga] in the future tense

/kö-ri-a=l/ [kti.ri.taa.T] 'go'-lRR-3ps/A=ASS
/ka-ri-äö/ [ku.ri.tQäö] 'go'-IRR- lsg/A
/k«-ri-ki=i/ [kn.ri.kn] 'go'-IRR-2ps=ASS

As one can conclude from the data presented in this section, a high degree of
idiosyncrasy is involved in the allomorphy of person suffixes that have three
types of alternants: vowel-initial forms (e.g. -anu), forms beginning with [k]
(e.g. -käu), and allomorphs beginning with [tp] (e.g. -tQäu). The following table
exemplifies this with the suffix for lsg/A. In fact, there are only a few contexts
where no variation occurs, i.e. most forms are not predictable. I have indicated
in brackets when a preference in terms of clear statistical tendencies can be

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
102 Morpho-phonology

identified. Especially with the final root vowels /e/ and I\I, no predictions can be
made.

(135) Summary of possible realisations of lsg/A

Final root vowel Monosyllabic root Disyllabic root Trisyllabic root

(or longer)
Any VV or V.V always -käu·, always -käu; always -käu·,
-tQäu after Γι/ -tQäu after /£/ -tQäu after /i/
/a/ -anu -anu -anu
-käu (rare) -käu (rare)
/e/ -anu -anu -anu
(changing to [i]) -tQäu -tQäu -tQäu
m -anu -anu -anu
-tQäu -tQäu
/u/ -anu -anu -anu
-käu (rare)
/«/ -anu -anu -anu

3.1.5 Verbal alternation with [k], [tg,] in the imperative form

A further instance of alternation between vowel-initial, [k]-initial, and [t$]-

initial suffixes is found in the plural form of the imperative, however, involving
additional factors. Corresponding to the distinction discussed above, word
length plays an important role for this allomorphy. The imperative for 2ps has
three different forms (cf. §15.1.1): singular, which is expressed by suffixation of
-u to the root; plural, which requires an additional suffix -toe to follow the
imperative marker, and distal imperative, which is expressed by the suffix -m,
preceding the imperative -u.

(136) Order of suffixes in the imperative form

IMP: itga-ύ! 'Do it!'

IMP-PL: itQa-u-tQe! 'Do (PL) it!'
DSTL-IMP: itQa-ni-u! 'Go do it!'
DSTL-IMP-PL: itQa-ni-u-tQe! 'Go (PL) do it!'

There are alternations that affect the shape of the imperative morpheme in some
of these forms, depending on the structure of the verb root and other factors.
The alternation between -u and its allomorphs -ku and -tQU highly depends on
the presence or absence of the plural marker and the distal suffix. For instance,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Alternations involving /kJ versus no onset on suffixes 103

the 2pl.DSTL form -ni-u-tQe is realised as -ni-tQU-tQe in a specific context, i.e.

the insertion of [t$] after the distance marker is triggered by the /i/ in /-ni/, but
realised only under additional conditions. Besides, there is a high degree of
variation between the two forms -ni-u-tQe and -m-tpu-tee, as both forms are
possible with most verbs (speaker-internal variation). In the table in (137), only
the preferred form is listed. Also note that in two contexts, no alternation
occurs: in the imperative singular (when no plural or distal suffix is present),
and in the distal imperative form (when no plural suffix follows). In these cases,
the imperative is always realised as -u. The forms in (137) only show examples
in which the plural suffix -t$e is involved.

(137) Imperative marker -u-tQe and allomorph -ku-tQe

Root Gloss IMP-PL DSTL-IMP.2ps-PL

itga- 'do' itQa-u-toe -ni-u-tQe
itQa-ku-tQe
enoala- 'cook' enoala-u-tQe -ni-u-toe
enoala-ku-tQe
sini- 'sleep' sini-u-tQe -ni-tQÜ-toe
sini-tQu-tQe
lenone- 'eat' lenone-u-toe -ni-u-tQe
fwi- 'fell' fiui-u-tQe -ni-tQu-toe
eru- 'meet' eru-u-toe -ni-tQu-toe
d^ana- 'cry' d^ana-u-tee -ni-tQÜ-tQe
beelaj- 'give as gift' beelaj-u-toe -ni-u-toe
memoa- 'fill' memoa-u-tge -ni-u-tQe
ahe- 'get drunk' ahe-u-tQe -ni-tQU-tQe
raatiri- 'leave' raatiri-u-toe -ni-u-tQi
kau- 'return' kaa-u-toe -ni-tQU-tQe
hau- 'throw' haau-tge -ni-tQu-toe
u- 'die' u-ü-tge -ni-tQU-tQe
ajto- 'say' ajto-o-tge -ni-tQu-toe
ajto-o-toe
kiuaau- 'think' kioaau-ku-tQe -ni-u-toe
kuruae- 'be happy' kuruae-ku-tQe -ni-u-tQe
? ?
e eeo- 'shout' e eeo-ku-tQe -ni-u-toe
kutia- 'call' kutia-ku-tge -ni-u-toe
fwitQu- 'swim' fioitQU-kü-t<^e -ni-u-tQe
kwara- 'see' kzvara-ku-tQe hoaa-ni-tQu-toe
bajte- 'forget' bajte-ku-tee -m-u-tQe

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
104 Morpho-phonology

Several remarks can be made about the factors that are responsible for the
distribution of the imperative allomorphs. It is evident from the transcription
that in addition to this, tonal phenomena are also involved. These will be
discussed in §4.3.
1) The regular imperative suffix is -«. It is attached to a number of verb roots of
different length. They can be monosyllabic, or can have two or three
syllables.
2) When the root-final vowel of a verb is a diphthong whose offglide is [w] or
[«], it is split up before the suffixation of the imperative morpheme. The first
component of the diphthong (which is /a/ in all attested examples) is
lengthened and the offglide merges with the imperative suffix -u. For
instance, the root / k a « / 'return' is combined with -u to result in
[kaa.a.(t$e)].
3) Roots that end in fu/ (also realised as [o], cf. (55)) assimilate the imperative
morpheme, which is then realised as /u/.
4) The use of -ku is more restricted: it does not occur with monosyllabic roots.
Instead, the verbs with which it is typically combined are familiar from the
other alternations that involve [k], as discussed in the previous sections. In
detail, these verbs are of the following structure:
a) Their root ends in a vowel sequence (e.g. huaau-ka, e?eeo-ka), or
b) their root ends with the vowel /e/ (e.g. bajte-a [bajtia]).
c) In addition, the two verbs huara-a and fiintQu-a take the -ku imperative in
combination with the plural marker. The verbs itQa-a and enoala-a can take
it as a possible variant. These could be treated as lexical exceptions (recall
that some exceptional features were already observed for itQaa and hoaraa).

The alternation between -m-u-tpe and -ni-tQU-tQe is difficult to analyse, since

the degree of variation has not been tested with a large number of speakers.
Besides, the "preferred" variant is obviously a very subjective judgement by the
respective speaker. However, apart from a few exceptions, the form -ni-u-tge is
the preferred one with those verbs that take the imperative -ku-tpe. Any further
attempt to account for this coincidence would only lead to speculations, but
hardly to substantial conclusions.

3.1.6 Verbal alternation with [k] in 2ps forms

The 2ps suffix -i has an allomorph -ki that occurs under specific conditions. In
particular, it is attested only in combination with the plural -tg,e to surface as
-kitQQ, as opposed to -itQe, which is the standard form. The conditions that
determine the alternation are quite complex and a summary with examples is
given in (138). While the form -itQe for 2pl is the regular one, the alternation
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 A It ernations involving /k/ versus no onset on suffixes 105

with /k/ as -kitQe is again related to the final root vowel of the verb. With a
number of verbs whose final root vowel is /a/, both variants are possible,
depending on speaker preference. However, the majority of verbs of this type
form their 2pl with -itQe only. A similar situation is found with verb roots that
end in /e/: again, some verbs can take both forms, but the majority occurs with
-itQe only. Interestingly, some verbs are not attested with -itQe but only with
-kitQe - this could be a coincidence due to lack of counterexamples (as I
assume), or it is the case that some verbs only occur with the alternant -kitQe.
With verbs that have / « / as their final root vowel, the situation is clearer, as all
but one verb take the form -itQe. With the remaining verbs a peculiarity can be
observed: verb roots that end with l\l always take the 2pl form -itQe. However,
when the 2pl follows the irrealis suffix -ri, it is consistently realised as -kitQe.
Interestingly, also the two verbs whose roots end in /ri/, siiria ' h a v e ' and heria
' w a n t ' , follow the same pattern, even when no irrealis suffix is present. There
are very few verbs whose root ends in /u/, and among these, the preference for
one or the other form is mixed. But when the 2pl form is preceded by the
habitual suffix -reheto, which itself ends in /u/ (realised as [o]), it surfaces as
-kitQe throughout.

(138) 2pl alternation -itQe vs. -kitQe

Final root Forms with -itpe Forms with -kitpe

vowel
/a/ Regular form: Variant/Some verbs only:
l
fouara-itge ( see'-2pl) kioara-kitQe ('see'-2pl)
itga-itge ('do'-2pl) itQa-kitQe ( ' d o ' - 2 p l )
esemta-itge ('believe'-2pl) eseneta-kitQe ( l believe'-2pl)
ßadera-itQe ('see'-2pl) ßadera-kitoe ('see'-2pl)
najßa-itQe ('finish'-2pl) mjßa-kitg,e ('finish'-2pl)
Id Regular form: -> Variant/Some verbs only:
ne-itQe ( ' b e ' - 2 p l ) w-kitQe ( l b e ' - 2 p l )
hjane-itge ('let'-2pl) hjam-kitQe ('let'-2pl)
najfle-re-itoe ( ' b e able'-IRR-2pl) riajße-kitQe ( ' b e able'-2pl)
ere-itQe ('speak'-2pl) here-kit^e ('want'-2pl)
te-it$e ( l give'-2pl) bajte-kitQe ( l forget'-2pl)
eseneta-ene-itQe ('believe'-NEG-2pl) kitrete-kitQe ( ' b u y ' - 2 p l )
ßadera-ene-itoe ('be sad'-NEG-2pl) kitruatahane-kitQe ('forget'-2pl)
/«/ -> Regular form: One example only:
hi-itge ('go'-2pl) nesaru-kitQe ('shove o f f - 2 p l )
turu-itQe ('arrive'-2pl)
hoitukn-itQe (lknow'-2pl)
basihjau-itQe ('steal'-2pl)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
106 Morpho-phonology

Final root Forms with -/fpe Forms with -kitge

vowel
HI Regular form: Some verbs; after IRR suffix /-ri/:
ari-itoe ('seek'-2pl) ar-ri-kitge=l ('seek'-IRR-2pl=ASS)
sim-itQe (Mive'-2pl) siiri-kitQe ('have'-2pl)
asi-it$e (*hide'-2pl) bajti-kitQe ('forget'-2pl)
im-itQe ('go.up'-2pl) rukit-ri-kitQe ('seek'-IRR-2pl)
/u/ -> Regular form (probably): Some verbs; always after HAB 1:
l
itQau-itoe ( live'-2pl) ajto-kiuitoe ('say'-2pl)
rautato-itqe ('be pleasant'-2pl) itgan-reheto-kvitoe ( l live'-HABl-2pl)
ne-reheto-kiuitQe ('be'-HABl-2pI)
kcoara-reheto-kimtge ('see'-HAB 1 -2pl)

Regarding the low degree of predictability for the alternating 2pl form it must
be concluded that it is highly lexically determined, with some phonological and
morphological factors playing a role.

3.1.7 Stem extension with /he/

There is another alternation for one group of verbs discussed in the previous
sections: when a verbal root ends in a long vowel or in a vowel sequence.
Before some suffixes, the syllable /he/ is inserted as a stem extension. For
instance, the participle form, which is normally -z, is realised as -hei in
combination with the respective verb roots. In some verbs, the second vowel of
the root-final sequence is deleted before suffixation of -hei. With the verb root
kurnae- in (139), the components /e/ of the diphthong /ae/ is elided.

(139) Stem extension /he/ with the participle form after vowel sequence
Root: / u b a a e / PRT: /ubaa-he-i/ [o.baa.hee.T] 'get.angry'-PRT
Root: / k u r u a e / PRT: /kurua-he-i/ [ku.ru.aa.hee.T] 'be happy'-PRT

This rule also applies to some verbs whose root ends in a diphthong: if the
diphthong ends in /j/ (such as in /aj/; note that there are no verbs whose root
ends in /ej/), the participle form is realised as -hei. With other diphthongs, the
rule does not apply and the regular participle suffix -z is used.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Palatalisation of /r/ after /i/ 107

(140) Stem extension /he/ after diphthongs

a) /he/ after diphthong /aj/:

Root: / b e e l a j / PRT: /beelaj-he-1/ [bee.laj.hjee.T] 'give as a gift'-PRT
Root: / k u t a j / PRT: /kutaj-he-1/ [ku.taj.hjee.T] 'scold'-PRT

b) -i after other diphthongs:

Root: / h a u / PRT: / h a u - 1 / [ha.wT] 'throw away'-PRT
Root: / k a « / PRT: / k a u - ! / [kau.T] 'return'-PRT

The application of the /he/-insertion rule before the participle form is not
surprising: as -7 is a vowel-initial suffix, one can conclude that the replacement
of a vowel-initial suffix by a consonant-initial suffix functions in order to avoid
a sequence of three vowels in a row. The fact that it also applies to diphthong-
final roots where [j] (underlyingly /i/) is the last element, is plausible, as the
insertion of /he/ avoids a clash between two identical vowels or between /e/ and
/i/.
Verbs with roots ending in a vowel sequence are subject to a few other
alternations regarding their inflection. With a verb such as kuruaeka ' b e h a p p y ' ,
the suffix for 3ps/E is -he instead of -e. Similarly, the suffix lsg/E is realised as
-heü (normally - « ) . Also the negative form involves the insertion of /h/, such as
in -hern instead of -ene with other verbs. Accordingly, the imperative of the
verbs under investigation is realised as -heu rather than -u, as shown in (141).

(141) Stem extension /he/ with other forms

Root: / k u r u a e / 3ps/E: / k u r u a - h e / [ku.ru.aa.he ] 'be happy'-3ps/E

Root: / n e k u l e e / IMP: / n e k u l e - h e u / [ne.kä.le.hee.a] 'wash'-IMP
Root: / u b a a e / PRT: / u b a a - h e - i / [o.baa.hee.T] 'get.angry'-lsg/E

It can be concluded that Ihl functions as a boundary-indicating element with

dissimilatory function. Less conceivable is why this should also affect roots that
end in a long vowel, since sequences of long vowel followed by a short vowel
regularly occur in the phonology.

3.2 Palatalisation of Irl after li/

Another morpho-phonological rule is the assimilation of Irl to [t£] when

preceded by /i/, which occurs with most suffixes and clitics (but not with all).
One context in which this regularity occurs is with the emphatic enclitic =ra, a
clause-final clitic mainly used in conversations (in the traditional language only;
cf. §2.11). When it is preceded by an ill, =ra is realised as =t$a.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
108 Morpho-phonology

(142) Irl [tc] in enclitic =ra

/asi=ra/ [a.si.tpa] 'hide'-(3ps/E)=EMF

/kwairi=ra/ [kwaa.i.ri.t^a] 'ayahuasca'=EMF

This type of assimilation is also observed with some suffixes that begin with Irl.
When they follow a root that ends in /i/, the initial consonant of the suffix is
assimilated to [t$], In particular, this applies to the habitual suffix -reheto, and
to the causative suffix. The latter is underlyingly -erate, but the Id is deleted
after vowel-final verb roots, thus leaving -rate. The causative suffix surfaces as
-tpate when the final root vowel is Ii I.

(143) Irl [t<?] with other suffixes

/aj-rehetu-a/ [aj.tce.he.to.ä] AUX-HABl-3ps/A

/tuni-rehetu-ttr«:-e=lö/ [to.ni.tce.he.te.kn.re.M] 'sound'-HAB 1 -PL-3ps/E=REM
/ari-erate/ [a.ri.t^a.te] 'seek'-CAU2:3ps/E

The same palatalisation rule applies when the 2ps proclitic i= is combined with
a word that begins with Irl. Since Urarina has only very few morphemes in
prefixal position (and none of these ends in /i/), only the 2sg clitic for
possession (with nouns), which also functions as the object proclitic (with verbs
and with postposition raj 'for'), has these properties.

(144) /r/ -> [tc] after proclitic i=

/i=raj/ [i.tsej] 'for you'

/i=rene/ [i.t£e.ne] 'your place'

However, the sequence HI + Irl is not generally banned from the phonology as is
also attested through numerous examples where this sequence occurs in root-
internal position:

(145) Occurrence of /ri/ in roots

/irit^u/ [i.ri.t£ii] 'ear'

/lureri/ [lu.re.ri] 'house'
/laauhiri/ [laa.u.hwi.ri] 'small'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Vowel lengthening 109

3.3 Vowel lengthening

Systematic vowel lengthening another alternation occurring between roots and

suffixes (also reported by Cajas Rojas & Gualdieri (1987:82f)). The vowels /a/
and Id are lengthened if a following suffix begins with a [+high] vowel, (i.e.
/u/, /a/, or /i/). The regularity regards both nouns and verbs. When the final root
vowel is identical to the initial vowel of certain suffixes, a long vowel emerges
(cf. (146a)). Vowel lengthening is also illustrated with suffixes beginning with
/i/ or /«/, respectively (cf. (146b,c)). It must also be mentioned that suffixes
which begin with /i/, such as the NEG:3ps/A morpheme -i, show a deviating
behaviour after verb roots ending in /e/: in this case, the root vowel surfaces as
Ν and the suffix, which is also -i, is separated from it by a glide, as can be
observed in the example tiji 'he did not give it' (cf. (146c)). There is only one
suffix which begins with /u/, as shown in (146d) - the velocity marker -uri.

(146) Vowel lengthening

a) Resulting in long vowel:
/ki-i/ [kii] 'eat'-2ps
/ra-a/ [räa] 'receive'-3ps/A
/ruku-uri-a/ [ru.kuu.ri.ä] 'pull out'-RAP-3ps/A

b) Before suffixes beginning with /«/:

/kitpa-ur«/ [ki.teäa.ti.ra] 'man/boy'-PL
/be-«r«-a/ [bee.ä.ru.ä] 'tell'-PL-3ps/A
/auna-tt/ [au.naa.ö] 'hear'-IMP
/ auna-ä/ [au.naa.ö'l 'hear'-lsg/E

c) Before suffixes beginning with /i/:

/be-i/ [bee.T] 'tell'-PRT
/ne-rela-i / [ne.re.laa.i] 'learn'-2ps
/auna-i/ [au.naa.l] 'hear'-NEG:3ps/A
/te-i/ [ti.ji] 'give'-NEG:3ps/A

d) Before suffixes beginning with /u/:

/te-uri-a/ [tee.u.ri.ä] 'give'-RAP-3ps/A
/taba-uri-a/ [ta.baa.u.ri.ä] 'grow'-RAP-3ps/A

Taking into account the peculiar behaviour of the negative suffix, it can be
concluded that vowel lengthening occurs after specific suffixes, which include

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
110 Morpho-phonology

the following: -u for imperative, -ü for 1 sg/E, -uru for plural (nouns and some
verb forms), -i for 3ps negative, -i for 2ps affirmative, and -i for the participle.
The lengthening rule does not apply when a [-high] vowel follows, such as
in the agentive suffix -era: in /itaa-era/ 'do'-AG ('the one who did it') the root-
final /a/ is not lengthened, but merges with the /e/ (cf. 3.4).

3.4 Merging between final root vowels and vowel-initial suffixes

In many instances, the attachment of a vowel-initial suffix to a root will not

affect the syllable structure of the root, as the initial vowel of the suffix forms a
separate syllable. However, there are some suffixes that consistently merge with
the final vowel of the root they are attached to.
As indicated earlier, suffixes that begin with the vowel /e/ show a peculiar
behaviour. As demonstrated in (147a), when the 3ps/E suffix -e is attached to a
root whose final vowel is /a/, one would expect the diphthong [ae] to occur.
However, the resulting vowel is [e], which may result from a simplification of
[ae].

(147) Combination of 3ps/E suffix -e with preceding /a/ resulting in [e]

l
/itqa-e/ [i.tae] do'-3ps/E
/eno-a-e/ [e.no.e] 'enter'-CAUl-3ps/E
/aua-e/ [au.a.ke] 'cry/shout'-3ps/E

Note that the /k/ in last example listed in (147) is predictable from the
allomorphy of 3ps forms that are attached to a verb ending in a vowel sequence
(cf. §3.1.3). Thus, the 3ps/A form for this root is -ka; [ke] is the simplified form
of [kae].
Interestingly, the merging to [e] occurs with some verbs only. With others,
the expected diphthong [ae] is observed. While the conditions for this are not
entirely clear, word length is likely to play a role. The examples in which a final
root vowel /a/ merges to [ae] with a following 3ps/E suffix -e are not disyllabic,
but have monosyllabic roots or roots that have more than two syllables.
However, the low number of examples may not suffice to prove this point.

(148) Combination of 3ps/E suffix -el with preceding /a/ resulting in [ae]
/na-e/ [näe] 'say'-3ps/E
/ha-e/ [häe] 'make'-3ps/E
/ne-rauta-e/ [ne.raa.tae] 'ITR-'heal'-3ps/E

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Merging between final root vowels and vowel-initial suffixes \\\

In verb roots that end with Γι/, the final vowel merges with the suffix -e and
results in [i], as illustrated in (149). In contrast to the above forms, this
alternation does not depend on the length of the root. Monosyllabic roots such
as /ki/ ' e a t ' are realised as [ki] in the 3ps/E form, corresponding to longer roots
that end with III as the examples in (149) illustrate.

(149) Combination of 3ps/E suffix -el with preceding root vowel /i/

/ki-e/ [ki] 'eat'-3ps/E

/siiri-e/ [sii.ri] 'have'-3ps/E
/toni-e/ [to.ni] 'sound'-3ps/E

O n e could characterise the process described above as an instance of

neutralisation/elision of initial /e/ in a suffix. Similarly, verbal roots that end
with /e/, are realised with a final [e] when combined with the suffix -e, i.e. root
vowel and suffix merge to a single vowel [e]. In this case, it cannot be
determined whether the root-final vowel or the suffix-initial vowel is elided.

(150) Merging of 3ps/E suffix -e with preceding root vowel Id.

/be-e/ [be] 'tell'-3ps/E

/here-e/ [he.re] 'want'-3ps/E
/kurete-e/ [ku.re.te] 'buy'-3ps/E

V e r b roots that end with the vowel /u/ exhibit two different patterns in
combination with -e. O n e type surfaces with an onglide [ w e ] , i.e. the final /u/
(or variant [o]) and the suffix -e are shortened, while the other - exceptional -
type simplifies this combination to [e]. T h e latter form appears to occur only
with verbs w h o s e roots end in /to/, as the examples show. N o t e that this does
not only apply to cases where the final root syllable /to/ is a suffix, but also to
the example /ajto/ ' s a y ' , which does not contain a derivational suffix.

(151) Combination of 3ps/E suffix -e with preceding root vowel /u/.

a) After /to/:
/kaua-to-e/ [ka.wa.te] 'be good'-INTS-3ps/E
/beai-to-e/ [be.u.i.te] 'cut'-INTS-3ps/E
/ ajto-e/ [aj.te] 'say'-3ps/E

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
112 Morpho-phonology

b) After other verbs ending in Inl

/muku-e/ [mu.kwe] 'burn'-3ps/E
/molo-e/ [mo.lwe] 'cut o f f - 3 p s / E
/su-e/ [swe] 'kill'-3ps/E

As a generalisation, the merging rule applies to all suffixes that begin with /e/,
including derivational suffixes and others. Specifically, these are the hortative
suffix -e, the negative suffix -ene, the agentive suffix -era, the privative suffix
-elanaala, and the causative -erate. The examples in (152) illustrate the
regularity with some of these forms. Note that the form beelajtga in (152a) is
the result of palatalisation: first, the /e/ of the agentive suffix -era is elided; the
consonant Irl is then palatalised to [tp], as it follows a diphthong ending in Ν
(cf. §3.2).

(152) Merging with other suffixes beginning with /e/

a) Agentive suffix -era:

/itsa-era/ [i.t^e.ra] 'do'-AG ('the one who did it')
/su-era/ [swe.rä] 'kill'-AG ('the one who killed it')
/beelaj-era/ [bee.laj.tca] 'give as present'-AG ('the one who gave it1)

b) Privative suffix -elanaala:

/it^a-elanaala/ [i.tee.la.naa.la] 'do'-PRV ('without doing it')
/su-elenaala/ [swe.la.naa.la] 'kill'-PRV ('without killing it')
/ sini-elanaala/ [si.ni.la.naa.la] 'sleep'-PRV ('without sleeping')

c) Causative suffix -erate:

/su-erate-a/ [swe.ra.ti.ä] 'kill'-CAU2-3ps/A
/te-erate-a/ [te.ra.ti.ä] 'give'-CAU2-3ps/A
/räasa-erate-e/ [räa.se.ra.te] 'dance'-CAU2-3ps/E

An exception from the merging process of root vowels with the 3ps/E suffix -e
is the monosyllabic verb root /«/ 'come'. The root vowel is not elided and thus,
the resulting form is not [uqe] as predicted, but [u.iqe], Note that merging
applies with phonologically similar verbs, given they have a syllable onset. The
3ps/E form of /ku/ ' g o ' is realised as [kiqe], with a single syllable.
As another result of morpheme combination, non-lexical diphthongs may
surface. Most typically, this involves combinations of a vowel-final verb root
with a vowel-initial suffix. One example for this is the combination of a verb
root whose final vowel is /a/ with the nominalising suffix -v. when these
morphemes are combined, the diphthong [aj] emerges (cf. (153a)); with the
locative suffix - u the resulting diphthong is [a«], as shown in (153b)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Merging between final root vowels and vowel-initial suffixes 113

(153) Non-lexical diphthongisation with suffixes -i and -u

a) With NOM suffix -i:
/(niitca) auna-i/ [au.näj] 'hear'-NOM 0 bj ('what (he) heard')
/(niitsa) te-i/ [tej] 'give'-NOM 0 b j ('what (he) gave')

b) With LOC suffix -u:

/temura-ö/ [te.mu.rau] 'edge'-LOC ('at the edge')
/ka-u/ [käu] 'this'-LOC ('here')
/ku-nemari-u/ [ku.ne.mu.rjü] ASC-'lake'-LOC ('in the lake')

However, not all suffixes form a diphthong with a preceding vowel. Some
suffixes coincide with a syllable boundary directly at their left, i.e. a vowel
sequence is created, as was described in §3.3. The participle suffix -7 is an
example for this. In the participle form for 'give', /te-l/ is pronounced [tee.i]
under normal conditions. However, in rapid speech, it can be shortened to [tej],
thus forming a positional or "casual" diphthong (cf. §3.6.1). This effect can
result in the creation of other diphthongs that do not have phonemic status.
Urarina has a few more suffixes that show peculiar behaviour in their
combination with roots. In addition, there is a diphthongisation rule that seems
to apply under even more restricted conditions. While the plural suffix -uru
forms a separate syllable and causes vowel lengthening on the final root vowel
when it is attached to a noun (cf. (146)), its initial vowel /«/ forms a diphthong
with the final root vowel if it occurs in the second noun of a possessive
construction. For instance, the plural of ela 'sleeping mat' is pronounced
[e.laa.urtt] in isolation, but realised as [ka.nti e.laa.ra] in 'my sleeping mats'.
In (154), a summary of combinations with different vowels is illustrated. It is
important to mention that merging does not apply to roots that end in a vowel
sequence or a diphthong. In that case, the suffix will form a separate syllable.

(154) Summary of merging processes

Final root Suffix Resulting vowel

vowel or diphthong
/a/ -i ( N O M ) [aj]
/e/ -i ( N O M ) [ej]
/i/ -i ( N O M ) [i]
/u/ -i ( N O M ) [wi]
/«/ -i ( N O M ) [uii]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 114 Morpho-phonology

Final root Suffix Resulting vowel

vowel or diphthong
/a/ -u (LOC) [aa]
/e/ -u (LOC) [ew]
Ν -u (LOC) [i«l
/u/ -a (LOC) [u]
/«/ -u (LOC) (unattested)

/a/ Any with initial Id lae], [e]

/e/ Any with initial Id [e]
/i/ Any with initial Id [i], [e]
/u/ Any with initial Id [we], [e]
/»/ Any with initial Id Ne]

3.5 Alternations with pre-root elements

For some regularities it is difficult to determine the domain within which they
apply. As mentioned, this problem occurs for alternations that apply before the
root, as Urarina has only a very small number of morphemes that occur in this
position. More exactly, there are only two types of preverbal morphemes: one is
the set of proclitics, used in object function with verbs and in possessive
function with nouns, and the other is the intransitiviser - the only productive
prefix found in Urarina.20 This section lists a number of rules that apply to
exactly these two kinds of morphemes. On the one hand, it would be possible to
say that this is a morpho-phonological rule, as it applies to these morphemes
only. On the other, the process is more general, as it applies not only to
proclitics, but also to the reflexive prefix.

3.5.1 Palatalisation of /k/ to [tg,]

The velar stop Ik/ is palatalised to form the affricate [t$] when it is preceded by
/i/. One example for this was mentioned before, with the example /i=/ +
/kwerehe/ realised as [itpuerehe] 'your son'. A further example is 'your wife',
where /i=komasaj/ surfaces as [itpuasej] (with the irregularity of /m/ being
elided, which is not systematic). One further example is / i = k u t « r i / [itputttri]
'your head'.

20
The associative prefixes ku- and b- are not productive; cf. §7.3.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Alternations with pre-root elements 115

Otherwise, assimilation to [ts] is not a general phonological rule as it does

not apply root internally. In the examples in (155), /if is indeed followed by /k/,
although certain restrictions apply: the adversative marker (cf. §22.5) is one of
only a f e w native words where the sequence /i/ + Ik/ occurs root-internally. All
other words listed below are morphologically complex (sometimes these may
be lexicalised derivations that cannot be split up; however, in these examples,
the suffixing components are known as productive derivational suffixes
elsewhere).

(155) Occurrence of /ki/

/niki/ [ni.ki] (Adversative marker)

/likii-to-a/ [li.kii.to.ä] 'smash'-INTS-3ps/A
/hinii-kil/ [hi.nii.kiT] 'together'
/ne-ri-ki=i/ [ne.ri.kff] 'you will be' ('be'-IRR-2ps=ASS)

Interestingly, Cajas Rojas & Gualdieri (1987) mention the palatalisation of

other consonants after /i/, which is not prominent in the Espejo dialect. In
particular, they report a change of /t/ to [tj], which is not observed with Espejo
speakers. According to Cajas Rojas & Gualdieri's data, i=teru ' y o u r axe' is
pronounced [i.tje.ru], which is likely to be based on the Chambira variety. In
the Espejo dialect, this pronunciation is not possible, but is [i.te.ru], with
optional postvocalic aspiration as [ih.te.ru].
Another kind of palatalisation, observed in the Airico dialect, is the
realisation of Is/ as [/] after i= (also cf. (29)). This pronunciation can also be
observed in the Espejo dialect, though with some speakers only. As a number of
inhabitants of the Espejo village have moved there from other places, this
dialectal feature seems to be taking hold, especially among younger speakers.

(156) /s/^[J]

/i=saabere/ [i./aa.be.re] 2sg='machete' ('your machete')

/i=su-ann/ [i./u.a.ntt] 2sg='kiH'-lsg('I kill you')

3.5.2 /i/^fdTj

A further regularity is observed when the 2sg proclitic i= is attached to a vowel-

initial root. In that case, a glide is inserted, which in turn is palatalised to [CI3].
However, these forms are characterised as a feature of the traditional language
by speakers of the Espejo dialect. This is mainly due to the fact that
procliticisation is disappearing in general in favour of analytic constructions for
pronominal objects and possessives.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
116 Morpho-phonology

(157) /i/ -> [d-ζ] with possessive proclitic

/i=enaniha/ [d3e.na.ni.hja] 2sg='canoe' ('your canoe')
/i=ere/ [dje.re] 2sg='word' ('your word')
/i=ukana/ [d3u.kwa.na] 2sg='field' ('your field')
/i=uba/ [d3U.ba] 2sg='mother' ('your mother')

3.5.3 Vowel elision with object proclitics and reflexive prefix

When a proclitic is combined with a vowel-initial root, the vowel of the clitic is
deleted. This applies to proclitics that occur before a verb (where they function
as object proclitics) as well as in prenominal position (in possessive function).
The lsg proclitic ka= loses its vowel /a/, while in the 3ps proclitic ra=, the
vowel /i/ is elided. The same rule is also observed with the intransitivising
prefix ne-, where the vowel Id is elided.

(158) Vowel elision with possessive proclitics and intransitivising prefix

a) Vowel elision with proclitics:

/ka=«haae/ [k«.h u aa.e] lsg='mouth' ('my mouth')
/ka=auna-i/ [kau.naa.i] lsg='hear'-2ps ('you hear me')
/ni=itpafwa/ [ni.tQa.fwa] 3sg='rifle' ('his rifle')

b) Vowel elision with intransitiviser ne-:

/ne-asi-a/ [na.si.ä] ITR-'hide'-3ps/A ('he is hiding')
/ne-itahe-a/ [ni.ta.hi.a] ITR-'get.lost'-3ps/A ('he got lost')
/ne-auna-he-a/ [nau.na.hi.a] ITR-'feel'-CNT-3ps/A ('he is feeling better')

3.6 Various rules of shortening or simplification

Similar to probably most other languages, Urarina reflects a number of

regularities that are the result of rapid speech and which are not observed on
words in isolation or in carefully pronounced examples. Most of these rules in
Urarina affect the final syllable of a word and can change the vowel quality,
length, or nasalisation. Many times it will result in a simplification of the
syllable, as features are removed. Most instances of simplification are not of a
general nature, but are specific to lexical items; therefore, they are discussed
here rather than in the chapter on general phonological rules.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Various rules of shortening or simplification 117

3.6.1 Simplification of /ex/

Many words, especially complex adverbs are composed of a root and the
participle suffix -1. When this morpheme is preceded by /e/, the underlyingly
resulting structure /el/ can be modified in different ways, as shown here with the
example mhoaaunel 'again' (lit. 'repeatedly').
a) De-nasalisation to [ej] as in [ n a . k w a a . u . n e j ]
b) Vowel coalescence to [ee] as in [ n a . k w ä a . H . n e e ]
c) Vowel shortening to [e] while nasalisation is retained as [ n a . k w a a . u . n e ]
d) Vowel shortening to [e] as in [ n a . k w a a . n . n e ]

Further examples for this are listed below. As an additional condition, vowel
shortening can only occur in unstressed syllables. Thus, the word kuajtel cannot
be simplified according to variant d).

(159) Simplification of /el/

/kwajte-1/ [kwaj-tef], [kwaj.tee], 'again'

[kwaj.te ], (*[kwaj.te])
/(nii) rihihe-1/ [(nii) ri.hi.he'J], [ri.hi.hee], 'being like (that)'
[ri.hi.he], [ri.hi.he ]
/hawia+ne-he-i/ [ha«.ri.a.ne.hej], [hau.ri.a.ne.hee], 'at first'
[hati.ri.a.ne.he], [ha«.ri.a.ne.he ]

3.6.2 Denasalisation of final -i

A further optional simplification rule is the loss of nasalisation on the participle

suffix -i when it is preceded by a vowel of the same quality.

(160) Denasalisation of final -/

/ini-i/ [inii], [inn ] 'go.up'-PRT

/ki-1/ [kil], [kff] 'eat'-PRT
/sini-i/ [sinii], [sinff] 'sleep'-PRT

N o t e that this rule does not apply to the participle form of the derivational suffix
-to, which is realised as -til. Thus, in the form for 'being nice', / k a u a - t n / , the
final syllable is not denasalised.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
118 Morpho-phonology

3.6.3 /e/~[ej]

The vowel Id can be pronounced as [ej] in certain contexts. Even though this
kind of diphthongisation mainly occurs in word-internal position, the exact
conditions for this rule are unclear, as it occurs in a variety of environments, i.e.
in accented or unaccented syllable, in final or non-final position, etc. One
context in which the rule is excluded is when Id is in word-initial position. Its
occurrence in word-final position is probably restricted to the example d^ale
listed below.

(161) Id ~ [ej]
a) Application of rule:
/d3a(e)+le/ [dga.le], [ d j a . l e j ] 'something'
/aheri/ [a.he.ri], [a.hej.ri] 'stone'
/netcara/ [ne.tpa.ra], [nej.tca.ra] 'maybe'

b) N o application:
/ahaane/ [ a h a a n e ] , *[ahaanej] 'ashes'
/atane/ [atane], *[atanej] 'earth'
/elo/ [elo], *[ejlo] 'rain'
/enoto/ [enoto], *[ejnoto] 'sun'

3.6.4 Omission or shortening of unaccented syllables

In fast speech, the initial syllable of some very common words or expressions
can be omitted optionally if it does not bear a High tone. The following are
some examples for this.

(162) Omission or shortening of unaccented syllables

/aka-Hr«/ [a.ka.tt.m], [katt.ra] 'they/them' (pronoun 3pl)

/hetau=te/ [he.tau.te], [tau.te] HRS+FOC
/ita/ [i.tä], [tä] (Reciprocal marker)
/itca-a/ [i.t$aä], [t^aä] 'make'-3ps/A
/ena+netohel/ [e.na.ne.to.hwe.l], [na.n(e).to.hwe.i] 'until now'

In a similar way, there is apocope, i.e. omission of a final unaccented syllable.

This is especially observed with the syllable /na/, which is part of the lpl/ex
suffix -akaanu and the lsg/A suffix -anu, before the subordinate marker =ne.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Various rules of shortening or simplification 119

(163) Omission of /na/

/k«-akaann=ne h a « / [k«.a.kaa.(na).ne.häö] 'go'-lpl/ex=SUB 'because'
('as we went')
/enu-akaanu=ne hana/ [e.no.a.kaa.(na).ne.hana] 'enter'-lpl/ex=SUB 'when'
('when we entered')
/it£afwa-ariH=ne h a n a / [i.taa.fwaa.(mi).ne.hana] 'shoot'-lsg/A=SUB 'when'
('when I shot')

An unaccented word-internal sound sequence /aha/ can be simplified to [aa] in

the word for 'help', kuruatahania is realised as [ku.rwa.taa.ni.a] in rapid
speech.
A sound sequence /nia/ can be shortened to [jia] in word-final position.
Most typically, this is observed with the very frequent words for 'be like that'
and 'be able', which appear to undergo a shift in pronunciation. In other words,
such as the postposition hirema 'for', this kind of simplification is possible, but
appears to be less frequent (cf. (164a)). Note that in most examples (except
kurema), the sequence /ni-a/ is composed of a verbal root whose last syllable is
/ne/ and the neutral/3ps suffix -a (by which the root-final /e/ is changed to [i]).
As demonstrated in (164b), the rule is not attested for examples in which /nia/
is part of a negative inflection such as in mjpenia {-erna is the suffix for
NEG:3ps/D).

(164) / n i a / - » [pa]

a) Application of rule:
/nitoane-a / [ni.to.a.ni.ä], [ni.to.a.jiä] 'be like that'-3ps/A
/najne-a/ [naj.ni.ä], [naj.jiä] 'be able'-3ps/A
/amiane-a/ [a.mi.a.ni.ä], ([a.mi.a.jiä]) 'work'-3ps/A
/kurenia/ [ku.re.ni.ä], ([ku.re.jia]) 'for'

b) No application:
/najne-ene-a/ [naj.jie.ni.a], *[naj.jie.jta] 'be.able'-NEG-3ps/A
/kwara-«rn-ene-a/ [kwa.ra.ka.re.ni.a], 'see'-PL-NEG-3ps/A
*[kwa.ra.ku.re.jia]

The phenomena discussed in this section are typical properties of rapid speech.
Further regularities that occur under this condition include the omission of
initial /•/ before /tp/ (cf. §2.1.9) and palatalisation of lx\l to [CI3] (cf. §2.8.6).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
4 Tone

Urarina can be characterised as a 'pitch accent language', following Yip's

(2002) definition below. As Urarina has an intriguingly complex system of
tonal alternations, only parts of it have been analysed at this stage. The data
presented in this chapter only highlights some of the major tonal features.
I shall continue to use the term 'accentual language' as a descriptive
convenience, but analytically the majority of such languages have lexical tones,
and what makes them special is that these tones have a small number of
contrasting tones, sparsely distributed, absent on some words, and frequently
lexically associated with specific TBUs (= tone-bearing units - noted inserted
by author). There is no absolute division between accent languages and tone
languages, just a continuum from 'accent' to 'tone' as the number and
denseness of tones increases, and they become freer to move around. ... [p.258]
Accentual languages typically have a lexical contrast between tone and no tone,
with each morpheme having a maximum of one tone or tonal complex whose
location must be lexically specified, and even morphologically complex words
often allowing only one tone to surface. Yip (2002: 257)
The terms "Η-toned" and "accented" are thus used interchangeably here.
Note that stress has not been analysed in detail. As a concise summary, it can be
stated that heavy ((C)VV) syllables inherently attract stress, which is manifested
by higher intensity/amplitude. "Accented" syllables can be characterised as
having a higher pitch than other syllables. A stressed syllable can be accented or
unaccented.
Typographical restrictions make it necessary to set up certain conventions
for the transcription of tone and nasalisation. (Particularly, the problem is
caused by the need to mark nasalisation and tone on the same vowel).

(165) Conventions for tonal marking

1) Single vowels not marked with a diacritic are toneless (i.e. Low-toned).
2) In Η-toned diphthongs, only the first element is marked with a diacritic (i.e.
v w to mean v w , e.g. in ka.naa.naj 'child', the final diphthong is H).
3) On long high-toned vowels [VV], only on the first V is marked with a
diacritic (i.e. v v to mean v v , e.g. in e.na.ma.naa 'young man', the final
syllable is all-H).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Various rules of shortening or simplification \ 21

4) When the second V unit of a long vowel is High, but the first V is Low, only
the High part will be marked: v v (-> to mean v v , e.g. in hua.raa ' s e e ' , the
final long vowel is LH)
5) Nasalisation is normally marked on the first element of a long vowel: v v
to mean v v ) . Exceptions: see 6)
6) Combinations of High-tone marking and nasalisation:
vv to mean: there is a LH contour on a nasalised long vowel).
x
v v (-> to mean: there is a HL contour on a nasalised long vowel).
v~ to mean: there is a H L contour on a nasalised single vowel),
vw to mean: there is a Η tone on a nasalised diphthong),
v w (-> to mean: there is a LH contour on a nasalised diphthong),
vw to mean: there is a HL contour on a nasalised diphthong).

The Urarina tonal system has a contrast between Η tone, which is lexically
specified, and toneless syllables, which are L(ow) by default. Urarina fits into
Y i p ' s characterisation of a pitch accent language as there is maximally one tone
per morpheme. More than that, this rule can be extended to words that contain
several morphemes and even larger units than that, as will become clear in this
chapter. The tone-bearing unit (TBU) appears to be the syllable at first sight, but
a f e w examples show that under certain circumstances, a syllable may carry
more than one tone. Therefore, it is more precise to define the mora as tone-
bearing unit. Despite the apparent simplicity of the tonal contrast ("tone versus
no tone"), there is a complex system that determines the position of the
respective tone on the units to which it is assigned.
Most words have a Η tone on their final syllable and thus, there are hardly
any minimal pairs that differ from each other only by tone. In fact, the database
contains only one such pair, as shown in (166). The word for ' t o p a tree'
exhibits a Η tone on the second part of the initial long vowel, and Falling (HL)
tone on the final syllable. There are only a f e w nouns of this type and these
clearly deviate from the "one tone per m o r p h e m e " principle just described.
However, most other nouns exhibit a final Η tone (cf. §4.1).

(166) Tonal minimal pair

i
aari arambaza bee' (Tonal class Type B)
ääri 'topa tree' (Tonal class Type D)

While the vast majority of nouns (and verbs) carry a final Η tone in isolation,
their tonal structure is lexically predetermined in so far as they are responsible
for tone assignment on adjacent words. This is, even though words normally
have a final Η tone, their underlying tonal structure may significantly differ
from that. Nouns can be divided into four tonal classes (called Types A, B, C, D

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
122 Tone

here) and, depending on the tonal class they belong to, Η tone is assigned to a
particular syllable of the word to the right (cf. §4.Iff). The example in (167)
illustrates this phenomenon with another rare minimal pair: the words for
i
achiote'> (referring to the tree or colour of Bixa orellana L.) and 'urine' have
identical segmental structure and identical prosodic structure on the surface, as
they both surface with a final Η tone in isolation. However, when they are
combined with a verb or an adjective, they cause different effects for the tonal
structure of these: with 'achiote', Η is assigned to the third/second syllable of a
word that follows the noun, whereas with 'urine', Η falls onto the final syllable
of a following word.

(167) Same surface structure but different underlying tonal structure

a) Nouns in isolation:
hjaane 'achiote (tree)' (Bixa orellana L.) (Type B)
hjaane 'urine' (Type C)

b) Nouns followed by a verb:

hjaane kiuara-kau

achiote see-lsg/A Ί saw the achiote.'

hjaane kiuara-käu
achiote see-lsg/A Ί saw the urine.'

c) Nouns followed by an adjective:

hjaane lanahaj 'red achiote'
hjaane lanahaj 'red urine'

T w o more examples for such underlying tonal differences are recorded with
akano 'snake' (Type B) vs. akano 'daughter-in-law' (Type C) and hajä 'muriaca
caspi tree' (Type A) vs. hajä 'horn' (Type B).
The examples in (167) prove the underlying difference between different
nouns even though they look alike on the surface. In other languages, similar
structures can be explained by floating tones. However, a thorough analysis of
the different tonal classes in this chapter will show that this approach does not
account for the phenomena observed in Urarina. Interestingly, the tone
assignment rules such as introduced above are interrelated with syntactic
structure: as a tendency, each constituent (rather than each word) receives only
one Η tone (cf. §4.2 to §4.5). Another important feature is the existence of
construction-type based tonal patterns that override any underlying tonal
structures, as is discussed in §4.6. It is also shown that a number of bound

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Tonal structure of nouns 123

morphemes have inherent tonal features that cause deviations from the
regularities observed elsewhere.
The position of the Η tone is predictable for most nouns and the majority of
verbs: it always falls on the final syllable. In contrast, the position of Η is not
predictable for stative and "long" verbs, (i.e. verbs which have four or more
syllables; cf. §4.7.1) and some adverbs (cf. §4.7.3). It should also be mentioned
that the assignment of a Η tone does not generally depend on syllable weight,
even though weight is relevant for certain tonal rules (cf. §4.2; §4.6.1).

4.1 Tonal structure of nouns

While most nouns have a final Η tone in isolation regardless their length or
syllable structure, some loans and a few canonical nouns may deviate from that.
Some of these are recent loans and therefore not stable, i.e. pronunciation can
differ from speaker to speaker (e.g. balde vs. bälde 'bucket': the second variant
is based on the Spanish stress pattern; this form would be used by speakers with
a higher proficiency in Spanish). Nouns of any length keep their final Η tone
when they follow an intransitive verb (in subject position, as in (168)).

(168) Intransitive verb + Noun subject

a) Disyllabic noun:
suru-ä katQä
run-3ps/A man
'The man runs.'

b) Trisyllabic noun:
kau-a ku.raa.naa
return-3ps/A chief
'The chief returns.'

c) Quadrisyllable noun:
u-ä si.m.he.rä
die-3ps/A spouse
'The spouse died.'

The nouns listed in (169) differ from the general structure as they do not have
their Η tone on the final syllable. Most of these carry a Η tone on the
penultimate syllable; they make up the majority of the nouns that differ from the
"Final H " pattern. A few other nouns exhibit a different pattern: These nouns
have a long vowel in their penultimate syllable, which carries a rising (LH)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 124 Tone

tone, followed by a falling (HL) tone on the final (short) syllable. As

mentioned, this is a comparatively rare tonal type of noun. It appears that all
examples with any of these two patterns belong to tonal Type D, most nouns of
which exhibit a final Η tone in isolation. In the light of the exceptions displayed
in (169), tone assignment for these nouns is not predictable.

(169) Some nouns that differ from the "Final H" pattern

a) Penultimate Η pattern:
aj.tQuM.ri 'giant armadillo'
a.ru.lä.ri 'fox'
ae.ra.na 'pucalupuna tree'

b) Contour pattern:
a.nää.he 'ubos tree'
a.lau.t.hja 'perο caspi tree'
a.loo.n 'panguana bird'

When nouns are followed by a transitive verb (i.e. when they are in object
position), they lose their final Η tone. In turn, the tone pattern of the verb is
changed. The impact nouns have on the tonal structure of the verb makes it
plausible to divide them into four tone classes:
- Type A: triggers an initial Η on a word that follows.
- Type B: triggers Η on the third syllable ("$3")of a word that follows.
- Type C: triggers a final-Η on a word that follows.
- Type D: triggers an All-L pattern on a word that follows.

In this context, where the noun functions as an object of a transitive verb, the
final Η tone of the noun is neutralised for tonal types A, B, and C. Type D
nouns differ from that, as they keep their Η tone; in this case, the verb remains
toneless. The details of tone assignment will be discussed in §4.2.
A verb without an overt object (being transitive or intransitive) has a final Η
91
tone. This structure is replaced by a pattern that is assigned through the object
noun of a transitive verb. A discussion of these patterns follows below.
One remark is in order regarding the frequency of the different noun types,
as illustrated in (170). 22 While Type A clearly is the most frequent one, there are

21
A transitive verb without an overt object refers to a 3ps object, which is typically
unmarked (cf. §18.1).
22
The total of 649 is based on 700 underived canonical nouns, including a few long-
established loans (also cf. table (1) in §5.1). However, for 51 nouns the tonal type
was not identified.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Tonal structure of nouns \ 25

fewer examples of the other types. The least frequent type is Type D, which can
be split into four separate subtypes, depending on further differences in tonal
structure.

(170) Statistical distribution of tone classes

Class Number of entries In %
A 266 41.0
Β 123 19.0
C 179 27.6
D 81 12.5
TOTAL 649 100.0

Since the membership to the tonal classes A to D is partly based on word length
and syllable structure (as will become clear later), an analysis of the word
structure is essential. The crucial distinction between Type A and Β nouns is
that quadrisyllable (or longer) nouns are typically members of Group A, but
never of Group Β (however, note that longer nouns may also fall into Group C
or D). Apart from this, it is not possible to distinguish Type A and Β nouns
based on word length. Both groups contain disyllabic nouns of different kinds:
CV.CV, CV.CVV, and CVV.CV (with C being optional in all cases). 23 The
tables in (171) to (173) list nouns in their form as preceding a verb, i.e. the noun
surfaces toneless for Types A, B, and C. Also note that there are no disyllabic
nouns that belong to Type D, apart from one exception (aari Ίορα tree'
mentioned in (166)).

(171) Disyllabic nouns: tone groups: typical members

Type A Type Β TypeC
CV.CV ka.ti ('black monkey') a.te('fish') bi.hi ('hand')
ka.t$a ('man') ti.hja ('foot') bu.ku ('bone')
CV.CVV a.kau ('water') a.lau ('spider monkey') be.ree ('child')
a.laa ('aguaje tree') a.maa ('hammock') i.jiaa ('fruit')
CVV.CV ee.ne ('woman') five.i ('firewood') ree.mae ('dog')
raa.na ('boar') laa.no ('yucca') bii.ßa ('old person')

In contrast to the nouns listed above, trisyllabic members are found for all tonal
types, as shown in (172).

23
I list this distinction here as it is relevant for some rules to be discussed in sections
4.2, 4.6.1, and 4.6.2. For nouns that have more than two syllables, the distinction is
not relevant.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
126 Tone

(172) Trisyllabic nouns: tone groups: typical members

Type A Type Β Type C Type D

hi.ht.ri le.w.ne ('food') d^a.ru.ba ('manatee') e.o.ri ('termite nest')
('armadillo') a.ka.w ('snake') ko.ma.saj ('wife') saa.be.re ('machete')
bi.to.a ('trap') ta.haa.e ('bird') ka.la.wi ('son') ajt.Qii.ne ('Espejo River')
fioau.a.na ('port') ka.ta.i ('hook') kioi.tg,a.na ('blood') e.na.ma.naa ('young man')
a.räa.la ('tapir')

A distinction between Type A and Β is impossible for about 80% of all nouns
(estimate). As mentioned above, words that contain more than three syllables
are not found in Group B. Type C nouns can be of any length.
One generalisation about Type D nouns is that they have at least three
syllables. 24 However, many quadrisyllable examples are no different from Type
A members; therefore, the number of syllables alone is not a satisfactory
criterion to predict the membership to a tonal class. One aspect that may
function as an indicator for Type D nouns is that proper nouns such as the ones
for 'Airico River' and 'Espejo River' are found in this class. Similarly, many
loans that have more than two syllables belong to Type D. In addition, many
compound nouns exhibit a tonal structure of Type D. In the present discussion, I
will restrict the analysis to simple nouns.

(173) Quadrisyllable nouns: tone groups: typical members

Type A: e.na.ni.hja 'canoe'

e.ru.a.ri 'mosquito net'
ha.no.la.ri 'jaguar'
lo.a.na.ri 'shed'
Type B: —

Type C: ku.fiua.ri.hja 'birth shed'

ku.ra.ru.hive 'soup'
si.ni.he.ra 'spouse'
hja.ra.ni.hi 'sugar cane'
Type D: a.nu.su.h(u)a 'plane'
ma.ku.sa.ja.rt 'pepper'
hi.ri.ri.jio 'bagre fish'
da.ra.ra.no 'lizard'

There is a good degree of arbitrariness as to predict which noun may belong to

what tonal class, but some observations may be noted, which could help in

24
The noun enamanaa 'young man' also has the variant naamanaa.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Verb Phrase: effects of nouns on the tonal structure of the verb 127

predicting the effect a noun has on the tonal structure of a verb, taking into
account statistical frequency.
1. If a noun has four (or more syllables), it is likely to follow Type A (i.e. it will
assign an initial-Η tone to a following verb).
2. If a noun has less than four syllables, it is likely to follow Type B. However,
this does not account for the high number of Type A nouns that have less
than four syllables.
3. Type C nouns may be of any length, but typically are trisyllabic or longer.
4. Type D nouns typically have 3, 4 or more syllables.

From this, one can conclude that the membership to tone class A or Β is hardly
predictable (except for "long" nouns). All four tonal types are also represented
in the set of free personal pronouns, as is shown in (174).

(174) Personal pronouns: tonal types

Type A lsg ka.nu

2sg ü
2pl fiaa.ra
Type Β lpl/in ka.na
TypeC 3sg a.ka
3pl a.ka.u.ru
Type D lpl/ex ka.na.kaa.nu

4.2 Verb Phrase: effects of nouns on the tonal structure of the verb

Each of the four noun types assigns a different tone pattern to a verb following
it. The respective tonal melody is superimposed to the final-Η structure of the
verb in isolation. In all four patterns, only one Η tone is realised within the
entire verb phrase (VP):
- Type A nouns consistently assign a Η tone to the initial syllable of the verb
(and the noun is toneless, i.e. All-L).
- Type Β nouns are realised All-L, but their effect on the verb turns out to be
more complicated; it will be discussed below.
- Type C nouns (realised All-L) occur with a final-Η tone on the verb. This
pattern is identical to the one found on the verb in isolation.
- Type D nouns consistently bear a Η tone, most typically on the last syllable.
However, no Η tone is assigned to the verb in this context.

These regularities are illustrated in (175), where the quadrisyllable verb

hvarakau Ί have seen (him)' is preceded by nouns of different tonal types.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
128 Tone

(175) Effect of tonal types on the verb hoarakau '('see'-lsg/A)

Type A: 'house' lureri hoa.ra.kau Ί have seen the house.'

Type B: 'cl. peccary' obana hoa.ra.kau Ί have seen the collared peccary.'
Type C: 'wife' komasaj hoa.ra.kau Ί have seen the wife.'
Type D: 'young man' enamanaa hoa.ra.kau Ί have seen the young man.'

The main difficulty is in describing the tonal pattern Type Β nouns trigger on
the verb. In general, they assign a Η tone to the third syllable of the verb, as
shown with obana in (175). As a complication, this is only valid if a verb has
"more than three" syllables. If a verb has exactly three syllables, the Η tone is
assigned to the penultimate (= second) syllable. However, with words that have
"something more" than three syllables in that their final syllable is heavy (VV
or CVV), the rule "H on third syllable" again applies: for, instance, even though
[ k w a . r ä a . i ] 'he has not seen it' has three syllables, its last syllable is light (V);
therefore, tone assignment follows the pattern for trisyllabic verbs (cf. (178a)).
The according structure is illustrated with the trisyllabic verb heria in (176a). In
contrast, the verb form [ru.a.kaa] (cf. (176b)) shows that the Η tone is assigned
to the third syllable of a trisyllabic verb if the last syllable is heavy.

(176) Tonal effects on verbs with three syllables

a) With trisyllabic verb heri-a ('want'-3ps/A) / last syllable is a short vowel:

Type A:'w.l. peccary' raana he.ri.a 'He wants the white-lipped peccary.'
Type B: 'cl. peccary' obana he.ri.a 'He wants the collared peccary.'
Type C:'dog' reemae he.ri.a 'He wants the dog.'
Type D: 'pepper' makusajari he.ri.a 'He wants the pepper.'

b) With trisyllabic verb ruakaa ('carry'-3ps/A) / last syllable is a long vowel:

Type A: 'w.l. peccary' raana ru.a.kaa 'He has carried the white-lipped
peccary.
Type B: 'cl. peccary' obana ru.a.kaa 'He has carried the collared peccary.'
Type C: 'dog' reemae ru.a.kaa 'He has carried the dog.'
Type D: 'pepper' makusajari ru.a.kaa 'He has carried the pepper.'

If a verb has two syllables only, the Η tone is assigned to the second (= last)
syllable (cf. (177b)). If the verb is monosyllabic, it still receives a Η tone, as
shown in (177a). Verbs that follow Type D nouns are exempt from this rule, as
predicted.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Verb Phrase: effects of nouns on the tonal structure of the verb 129

(177) Tonal effects on verbs with one or two syllables

a) With monosyllabic verb ki ('eat'-3ps/E):

Type A: 'w.l. peccary' raana Μ 'He has eaten white-lipped peccary.'
Type B: 'cl. peccary' obana ki 'He has eaten the collared peccary.'
Type C: 'dog' reemae Ja 'He has eaten the dog.'
Type D: 'pepper' makusajari ki 'He has eaten the pepper.'

b) With disyllabic verb ki-u ('eat'-IMP):

Type A:'w.l. peccary' raana Ja.u 'Eat the white-lipped peccary!'
Type B:'cl. peccary' obana ki.u 'Eat the collared peccary!'
Type C: 'dog' reemae ki.u 'Eat the dog!'
Type D: 'pepper' makusajari ki.u 'Eat the pepper!'

Further forms to exemplify the predicted patterns are given in the table in (178).
In each example, the tonal structure of the verbs follows the predicted rules for
the respective tone class of the noun that precedes, with a range of different
derivational and inflectional suffixes regardless the length of the verb. Even on
verbs with eight syllables, a Η tone is only observed on the first syllable after
Type A nouns. Similarly, the entire verb is L-toned after Type D nouns. Recall
that in these combinations, Type D nouns carry a Η tone on their final syllable.

(178) Further examples for tone patterns in the VP

Syll. Gloss Verb after Type A Verb after Type Β

katga 'man' akaw 'snake'
3 'see'-NEG:3ps/A kiva.ra.-i kiva.ra.-i
4 'see'-lpl/ex hoa.ra.-ka-a.m kioa.ra.ka-a.m
5 ' want' -PL-N EG-3 ps/A he.re.-kn.ru.-i he.re.ku.ru.i
6 'kill '-CAU1 -1 pl/ex sw-e.ra.ti.~a.kaa.nn siu-e.ra.ti.a.kaa.nn
7 'know'-C AU 1 -1 sg/A hvi.tH.k'-e.ra.ti.-a.nH fooi.tu.k'-e.ra.ti.-a.nu
8 'kill'-DSTR-PL-IRR-
sü.-a.hoa.-u.ru.-ri.-tQa-i su.-a.hoa.-u.ru-.ri-.tQa.=i
3ps/A=ASS

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
130 Tone

b) Types C and D
Syll. Gloss Verb after Type C Verb after Type D
reemae 'dog' enamanaa 'young man'
3 'see'-NEG:3ps/A hoa.ra.i hoa.ra.i
4 'see'-lpl/ex hva.ra.kaa.m kiua.ra.kaa.nu
5 'want'-PL-NEG-3ps/A he.re.kn.ru.1 he.re.ku.ru.i
6 'kill'-CAUl-lpl/ex sioe.ra.ti.a.kaa.nu sioe.ra.ti.a.kaa.nu
7 'know'-CAU 1 -1 sg/A kzoi.tn.k"-e.ra.ti.-a.nu
8 'kill'-DSTR-PL-IRR-
sit.a.kioa.u.ru.ri.tqa.T sit.a.kioa.u.ru.ri.t^a.l
3ps/A=ASS

4.3 Suffixes that cause deviations from the predicted tone patterns

It should be noted that some derivational or inflectional suffixes that are part of
the verb cause deviations from the patterns illustrated in the previous section.
However, the division into four tone classes for nouns is still practical and
suffices to state the general rules for tone assignment for at least two reasons:
1) There is a wide range of suffixes with different functions that can occur
within a verb without altering the predicted tonal pattern.
2) The rules apply regardless the length of a verb. Where this is not possible
(for instance when the rule "H tone on the third syllable" cannot apply
because the verb is shorter than that), additional rules with very specific
conditions apply. This includes verbs that have three syllables where their
final syllable is light (i.e. (C)V).

4.3.1 lsg suffix l-anuj

The rules sketched in §4.2 are challenged by some examples that involve the
lsg person suffix -anu. As shown in (179), the predicted tone patterns apply to
constructions that involve Type C and Type D nouns when followed by a verb:
after the Type C noun reemae 'dog', all verbs are realised with a final Η tone;
after enamanaa 'young man' (Type D), the verb is toneless. But irregularities
are observed on verbs that follow the two other tonal types: after a Type A noun
such as katQa 'man', the initial Η tone occurs on some verbs, but not on all. As
shown in (179), with 'send' and 'throw' the Η tone is on the first syllable of the
person suffix (-anu); with several other verbs (e.g. 'scold', 'exceed'), Η falls on
the second syllable rather than on the predicted first syllable. In contrast, the
expected tonal pattern is observed with the verb for 'know'. The pattern
assigned by Type Β nouns such as akano 'snake' also differs from the predicted

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Suffixes that cause deviations from the predicted tone patterns 131

pattern "H on third syllable" with some forms: while it follows this rule with the
verbs for 'send', and 'scold', it is on the second syllable with 'throw' and on the
fourth syllable with 'exceed' and 'know'. However, it should be noted that,
except with 'exceed', the Η tone falls on the first syllable of the suffix again,
just as with 'send' and 'throw' after Type A nouns.

(179) Deviating tone patterns with 1 sg/A suffix -arm

Gloss After T y p e A After T y p e Β After T y p e C After T y p e D

katQa 'man' akatio 'snake' reemae 'dog' enamanäa 'young
man'
'send'-lsg/A le.to.a.nu le.to.a.nu le.to.a.nu le.to.a.nu
'throw'-lsg/A haii.ä.nu haii.ä.nu hau.a.nu hau.a.nu
'scold'-lsg/A hi.hjü.taa.nu hi.hjii.tää.nu hi.hju.taa.nu ht.hju.taa.nu
'exceed'-lsg/A a.mu.ri.to.a.nu a.mu.ri.to.a.nu a.mu.ri.to.a.nu a.mu.ri.to.a.nu
'know'-lsg/A hui.tu.ku.a.nu kiui.tu.ku.a.nu kiui.tu.ku.a.nu kxoi.tu.ku.a.nu

From these data, it is apparent that the suffix -anu causes special tonal effects.
However, it is difficult to analyse these as unfailing rules due to the fact that
they are neither consistent with all verbs, nor with all tonal types.
The complexity of the tonal behaviour of the lsg/A suffix -anu is also
demonstrated by some verb forms that involve several suffixes, when the verb
occurs in isolation. In this case, a default final-Η pattern would be expected (cf.
§4.7.1), but this does not reflect the actual forms. As shown in (180), the Η tone
consistently falls on the first syllable of the suffix -anu.

(180) "Long" verb forms with lsg/A in isolation

le.to.a.-na.ha-a.hoä-ä.nu 'send'-PLO-DSTR-lsg/A Ί send them one by one.'

i.tQa.-ha.hoa.-na.hä-ä.nu 'do'-DSTR-PLO-lsg/A Ί made them one by one.'
i.tQaa.u.ri.-a.kiva.-na.hä-ä.nu 'do'-RAP-DSTR-PLO-lsg/A Ί quickly made them one by
one.'

4.3.2 Imperative form

Further deviations from the regular patterns can be observed with the imperative
form. It consists of the 2ps imperative suffix -u (or allomorphs), which can be
preceded by the distal suffix -ni and followed by the plural suffix -tge (cf.
§3.1.5). In (181), the attested tonal patterns with the plural form and the distal
form are displayed. These represent the forms when the verb occurs in isolation;

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
132 Tone

thus, tonal changes triggered by a preceding noun could not apply. However,
the existing patterns are rather complex even in this context.

(181) Tonal patterns for imperative forms

a) For imperative plural -u-toe
1 = v-«-t(je (LHL) Examples: si.ni.-u.-toe ('Sleep!' / PL)
e.re.-u.-tQe ('Speak!' / PL)
2 = v-«-tQe (LLH) Examples: a.ri.-u.-tQe ('Seek it!' / PL)
ra.ti.ri.-u.-toe ('Leave it!' / PL)
3 = v-«-t£e (L Η HL) Examples: ru.ku-ü.-toe" ('Pull it out!' / PL)
su-u.-toe ('Kill it!'/PL)
[*4 - v-tt-tpe] *(L Η L)

b) For distal imperative plural -ni-u-toe

2 = ni-tt-t^e (LLH) Examples: a.ri.-ni-u.-tQe ('go seek it!' / PL)
su.-ni-tQU.-tQe ('go kill it!' / PL)
4 = rii-ö-t<?e (LHL) Examples: si.ni.-ni-tQti.-tge ('go sleep!' / PL)
e.re.-ni-tQÜ.-toe ('go speak!' / PL)
5 = ni-a-t^e (HLL) Example: e.?e.o.-m-u.-toe ('go shout!' / PL)
[*1 = v-ö-tse] *(H Η L)
[*3 = v-«-tQe ] *(L Η HL)

In particular, it can be observed that a specific pattern in the imperative plural

tends to imply a special tonal pattern with the distal plural form (cf. (182)):
verbs that take pattern "2" in the plural form, also take this pattern when adding
the distal suffix. Verbs that follow pattern "1" without the distal form, take
pattern " 4 " with the distal form (except for one case: e?eoka 'shout' takes
pattern " 5 " - the only example attested for this). Verbs that follow pattern " 3 "
in the non-distal form, switch to pattern "2" when the distal suffix is present.

(182) Tone patterns of imperative forms for plural and distal plural

Root Gloss IMP.2ps-PL PL DSTL- DSTL

patt. IMP.2ps-PL patt.
lenone 'eat' le.no.ne.-u.-tQe 1 -ni-u-tQe 4
ere 'speak' e.re.-u.-tge 1 -ni-tQbi'tg,e 4
d^ana 'cry' d^a.na.-u.-toe 1 -ni-tQu-toe 4
taka 'encounter' ta.ka.-u.-tge 1 -ni-tQu-tge 4
kau 'return' kaa.-u.-tQe 1 -ni-tQu-tQe 4
hwitQU 'swim' him. t QU.-kit.-toe 1 -ni-u-toe 4
e?eo 'shout' e. ?e.o.-ku.-tge 1 -ni-u-toe 5

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Suffixes that cause deviations from the predicted tone patterns 133

Root Gloss IMP.2ps-PL PL DSTL- DSTL

patt. IMP.2ps-PL patt.
itga 'do' i.tQa.-u.-tge 2 -ni-u-tge 2
i.tga.-ku.-tQe
ari 'seek' a.ri.-u.-tge 2 -ni-u-t<^e 2
muku 'catch' mu.ku-u.-tQe 2 -ni-tQU-toe 2
fxoi 'fell' fiui.-u.-tge 2 -ni-tQU-toe 2
beelaj 'give as gift' bee.la.j-u.-tge 2 -ni-u-tge 2
raatiri 'leave' raa.ti.ri.-u.-tge 2 -ni-u-tge 2
hau 'throw' haa.-u.-tge 2 -ni-tgu-toe 2
molo 'truncate' mo.lo-o.-tQe 2 -ni-tQU-tge 2
kzuaau 'think' kuaa.u.-ku.-tge 2 -m-u-toe 2
'be happy' ku.ru.aa.e.-ku.-tQ 2 2
kuruae e -ni-u-tge
kutia 'call' ku.ti.a.-ku.-tQe 2 -ni-u-tge 2
bajte 'forget' baj.te.-ku.-tge 2 -ni-u-tge 2
kurete 'buy' ku.re.te.-ku.-tge 2 -ni-u-tge 2
su 'kill' su-u.-tQe 3 -ni-tQU-tge 2
sau 'cut' saa.-u.-tge^ 3 -ni-tQU-toe 2
u 'die' ύ-ύ.-ίςέ" 3 -ni-tgu-toe 2
ruku 'pull out' ruM-H.-tge 3 -ni-tQu-tge 2
ajto 'say' aj.to-o.-tGe 3 -ni-tQU-toe 2

Further peculiarities occur when a verb in the imperative plural/distal form is

preceded by an object noun. As could be expected, the tonal structure of the
noun has an impact on the tonal representation of the verb, but again, some
deviations from the general rules are observed. The examples in (183) give an
overview of the tonal structure with the (non-distal) imperative plural forms.

(183) Tone patterns of imperative plural forms preceded by noun

Gloss After Type A After Type Β After Type C After Type D

' throw'-
IMP-PL haa.-u.-tge haa.-ü.-tQe haa.-H.-tQe" haa.-u.-tge
' catch'-
IMP-PL mu.ku-it.-toe mu.ku-u.-t& mu.ku-u.-tQe mu.ku-u.-toe
'give'-
IMP-PL tee-u.-toe* tee.-u.-tge tee.-u.-tQe tee.-u.-tge
'see'-
IMP-PL kiua.ra.-ku.-t<ge hua.ra.-ku.-tge kiua.ra.-ku.-tQe hua.ra.-ku.-toe

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
134 Tone

It is difficult to recognise a consistent pattern from the above forms, apart from
the fact that verbs that follow a Type D noun behave as predicted, i.e. they are
toneless. Some verbs have a Η tone on their first syllable when following a
Type A noun, but others considerably differ from that. In a similar way, the
effects of Type Β nouns on verbs is hardly recognisable. When a verb is
preceded by a Type C noun, again, there is a range of different patterns on
different verbs.
When the distal suffix -ni is inserted into the verb form (cf. (184)), some of
the changes tend to become more systematic in that tone patterns appear to be
inherent to specific verbs: the verb for 'kill', for example, still follows the
regular pattern, with or without the distal suffix. Similarly, the verbs for ' c a t c h '
and ' b u y ' keep their Η tone on the same syllable when the distal form is
inserted. However, the verb for 'give' changes its pattern. When a verb with the
distal imperative plural is preceded by a Type Β noun, the Η tone is always on
the penultimate syllable (the suffix -u). The examples for verbs preceded by a
Type C noun vary with respect to their tone pattern. Only Type D nouns trigger
the predicted effect on the verb, as was also observed without the distal form.

(184) Tone patterns of distal imperative plural forms preceded by noun (verb with
DSTL-IMP-PL form)

Gloss After Type A After Type Β After Type C After Type D

'kill' sü.-ni.-tgu.-tge su.-ni.-tQu.-tQe su.-ni.-tQu.-toe su.-ni.-tQu.-tQe
'catch' muM-ni.-u.-toe mu.ku-ni.-u.'tQe mu.ku-n.i-ü.-toe mu.ku-n.i-u.-tQe
'buy' ku.re.te.-ni-.u.-tQe ku.re.te.-ni-M.-tQe ku.re.te.-ni-.u.-tQe ku.re.te.-ni.-u.-tge
'give' te.-ni.-u.-tQe te.-ni.-u.-toe te.-m.-u.-toe te.-ni.-u.-tQe

At this stage, the available data is too inconclusive as to allow reliable

generalisations that could describe the precise tonal pattern applied to
imperatives.

4.4 Noun Phrase: Noun + Adjective

If the tonal structure of a noun is responsible for tone assignment on a verb that
follows the noun, it could be expected that the same effects also occur with
other word classes that follow the respective noun. This is indeed the case with
adjectives. (Note that most words that function as adjectives in Urarina are
derived, or they are nouns in modifier position, cf. §5.4.) Words that function as
adjectives typically occur in postnominal position. When a noun precedes an
adjective, again, most nouns lose their final Η tone and are realised with an All-
L pattern (except for Type D nouns). Parallel to verbs, the tone pattern on the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Noun Phrase: Noun + Adjective 135

adjective changes according to one of four possible melodies. The tonal

melodies found with Types A, C, and D are straightforward as they follow the
patterns observed on verbs (cf. (185)). The effect of Group Β nouns, again, is
more complex, but generally follows the rules: as predicted, Η on the third
syllable is assigned to [laa.u.hwi.ri] 'small'. With seohwa 'big', Η is assigned
to the second syllable to result in [see.o.hwa], because it is trisyllabic and does
not end in a CVV syllable. Deviations from this pattern are found with the two
shorter (disyllabic) examples. The expected pattern would be a final-Η pattern.
However, instead of finding a single final Η tone on these words, they end in a
Falling tone, which could be the effect of some kind of spreading mechanism
not identified at this stage. After Type C and Type D nouns, the adjectives
behave as predicted and have a "Final-Η" or "All-L" pattern, respectively.

(185) Noun + Adjective: Adjectives that are subject to tonal changes

Adjective After Type A After Type Β After Type C After Type D

'small' laa.u.hwi.ri laa.u.hioi.ri laauhwiri laauhwiri
'big' see.o.hwa see.o.hwa seeohwä seeohwa
'bad' bäa.so bäa.so' baaso baaso
'new' na.tQa.si.pe na.tQa.si.pe
'good' ka.wa.tQa

A number of adjectives have a special inherent tonal structure that is not subject
to any changes triggered by a noun that precedes the adjective. These are listed
in (186). As a peculiarity, the words natQasiße and kauatga follow the Type D
pattern, but three other adjectives keep their tone after any type of noun.
natQasiße also follows the pattern triggered by Type C nouns, but has a stable
tone after Type A and Β nouns.

(186) Noun + Adjective: Adjectives with a fixed tone pattern

Adjective After Type A After Type Β After Type C After Type D

'new' na.tQä.si.ße na.tQä.si.ße
'good' ka.wa.tQa ka.ivä.tQa ka.wä.tQa
'good' ka.wa.ti ka.iuä.ti ka.iuä.ti ka.iuä.ti
'big' ta.bäj ta.bäj ta.bäj ta.bäj
'big' ta.si.ßo.hwaj ta.si.ßo.hwaj ta.si.ßo.hwaj ta.si.ßo.hwaj

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
136 Tone

4.5 Postpositional Phrase: Noun + Postposition

Postpositions behave in a similar way as adjectives. While they all have a final
Η tone in isolation, they are subject to tonal alternations when preceded by a
noun. In (187), only a few examples are given. Type C and Type D nouns are
unproblematic as they consistently assign their tonal pattern to all postpositions.
Type A nouns assign a Η tone to the first syllable of ajßa, kahe, a n d kolane, but
to the second syllable of kurenia and kuane (such as if the first syllable were
extrasyllabic). Type Β nouns behave as predicted, as they assign a Η tone to the
final syllable of disyllabic postpositions, but a Η to the penultimate syllable of
trisyllabic postpositions. The quadrisyllable kurenia receives a Η on the third
syllable.

(187) Noun + postposition

a) Regular pattern:
Postposition With Type A With Type Β With Type C With Type D
lureri 'house' ate 'fish' reemae 'dog' enamanaa
'young man'
ajfla 'with' lureri äjßa ate ajßä reemae ajßä enamanaa ajßa
kahe 'from' lureri kahe ate kahe reemae kahe enamanaa kahe
kulane 'without' lureri kolane ate kolane reemae kolane enamanaa kolane

)) Deviations after Type A nouns:

Postposition With Type A With Type Β With Type C With Type D
kurenia 'for' lureri kurenia ate kurenia reemae kurenia enamanaa kurenia
kuane 'inside' lureri kuäne ate kuäne reemae kuane enamanaa kuane

Deviations from the regular tonal patterns are also found with a few other
postpositions, but these will not be discussed here.

4.6 Overriding patterns

There are some environments in which the predicted four patterns for nouns do
not match the previously established patterns, at least not at first sight. Possible
explanations for this include extrasyllabicity, tonally prespecified suffixes, and
possibly some lexical exceptions, as described above. It is also not uncommon
for tone languages to replace ("override") the underlying tonal structure of
words with a consistent tonal melody under certain conditions. Typically, such
melodies are associated with particular grammatical constructions. The
possessive construction in Urarina is an example of this.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Overriding patterns ] 37

4.6.1 Possessive construction

Nouns that function as a possessor and precede another noun (the possessum),
trigger a certain tone pattern to occur on a following noun in a possessive
construction. This also applies to constructions where a pronoun has possessor
function (on possessive constructions, cf. §7.1). The possessor, (which has a
final-H tone in isolation), is realised with an All-L pattern, similar to nouns
occurring before a transitive verb. The tones on the noun that functions as a
possessum (in second position), changes according to a specific pattern. The
examples given in (188) involve the lsg pronoun kanu 'my' as a possessor
(recall that this pronoun is a Type A noun from a tonal perspective, cf. (174)).
Interestingly, the effect that kanu would normally have on a word that follows
(i.e. to assign a Η tone to the first syllable), is neutralised. Instead, the tonal
structure of the possessed noun largely depends on its length and syllable
structure: disyllabic nouns of Types A, B, and C receive a Η tone on their initial
syllable when they function as possessum. However, when a disyllabic noun
ends in a long vowel or in a diphthong, it is regarded "more than disyllabic" and
automatically is assigned a Η tone otherwise found with trisyllabic or longer
nouns. This is an intriguing feature, which matches the analysis suggested for
Type Β nouns; recall that a Η tone is assigned to the third syllable of trisyllabic
verbs only if the final syllable is heavy (cf. (178)).
Trisyllabic or longer nouns of Types A, B, and C follow a common pattern:
they receive a Η tone on the second syllable. It is plausible to characterise tone
assignment to the second, not penultimate syllable, since this also applies to
quadrisyllable nouns of all types.

(188) Possessive constructions: lsg pronoun kanu plus noun of Types A, B, C

Possessed noun: number of Possessum

syllables
Disyllabic Ν preceded by ka.tQa ('man') / Type A
possessor kanu laa.no ('cassava') / Type Β
ree.mae ('dog')/Type C
Disyllabic with final VV or a.laä ('aguaje tree') / Type A
diphthong a.lau. ('kinkajou') / Type Β
Trisyllabic lu.re.ri ('house') / Type A
ka.taa.l ('hook') / Type Β
Longer e.na.ni.hja ('canoe') / Type A
si.ni.he.ra ('spouse') / Type C

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
138 Tone

As shown in (189), Type D nouns deviate from these tone assignment rules:
trisyllabic Type D nouns get a Η tone on their first syllable, whereas nouns with
more than three syllables receive a Η tone on the second syllable.

(189) Possessive constructions: lsg pronoun kanu plus noun of Type D

Possessed Type D noun: Possessum

syllables
Trisyllabic Ν preceded by saa.be.re ('machete')
possessor kanu es.hve.la ('school')
Longer ma.kü.sa.ja.ri ('pepper')
a.näj.si.hje ('mucura plant')

The examples given in (188) and (189) all involve the l s g pronoun kanu, but
the word-length dependent patterns described above are not specific to
possessive pronouns. When full nouns function as possessor, the same tonal
pattern applies: the possessor noun is L-toned, whereas the possessed noun
receives an "initial-Η" or a "H on $ 2 " pattern, according to its length.

(190) Possessive constructions: Noun + Noun

Possessed noun: syllables Possessum

Disyllabic Ν preceded by ä.kau ('water')
possessor katga 'man' räa.na ('w.l. peccary')
ree.mae ('dog')
Trisyllabic lu.re.ri ('house')
ko.ma.saj ('wife')
ka.naa.naj ('child')

It can be concluded that the word length-based patterns are a feature of the
possessive construction in general. However, this only applies to possessive
constructions in juxtaposition, as described above. When the 3ps possessive
marker raj is involved (cf. §7.1.2), the possessor noun is L-toned, raj itself
bears a Η tone, and any possessum of Type A, B, or C receives a final Η tone
regardless its length.

(191) Possessive construction with raj

ka.t$a raj lu.re.rt 'the man's house'

ka.tqa raj a.kau 'the man's water'
ka.tQa raj ru.k'e.le 'the man's things'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Overriding patterns 139

With Type D nouns, the pattern is different, as the Η tone is found on the
penultimate syllable of the possessum.

(192) Possessive construction with raj and Type D nouns

kann raj saa.be.re 'my machete'
kanu raj ma.ku.sa.ja.ri 'my pepper'
ka.t$a räj ena.mä.naa 'the man's young fellow'
ka.tpa räj aj.t$ü.ne 'the man's Espejo River' (i.e. his homeland)

Urarina has an alternative way to mark possession for a certain group of words
(cf. §7.4). These forms represent a class of formerly inalienable nouns and had
an obligatory proclitic indicating possession. Typically, the nouns found with
such clitics represent kinship terms and body parts. 25 The interesting fact lies in
the tonal structure of nouns preceded by a proclitic: they consistently have a
final-Η tone, which makes them indistinguishable from underived nouns in
isolation. One may conclude from this that the proclitic is regarded a part of the
Phonological Word and therefore has no impact on the tonal structure.

(193) Proclitic forms vs. possessive construction

Noun (citation form) With POSS pronoun With proclitic

su.ri ('intestine') ka.nu+sü.ri ka.=su.ri
a.naa.e ('leg') ka.nu+a.naa.e k=a.naa.e
a.laa.ri.hjä ('chest') ka.nu+a.läa.ri.hja k=a.laa.ri.hja

4.6.2 Possessive construction with adjective

In combinations of a possessive construction with an adjective that follows the

possessum, another pattern applies, again depending on the length of the
possessed noun and on the tonal structure of the adjective. With adjectives that
do not have a stable tone, such as seohwa 'big' and laauhwiri 'small' (cf. §4.4),
the following tonal structure is realised: as in other possessive constructions, the
possessor is realised toneless and the possessum receives a Η tone according to
its length and syllable structure: Disyllabic nouns whose first syllable is light (V
or CV), receive a Η tone on their second syllable, disyllabic nouns whose first
syllable is heavy (VV or CVV), receive a Η tone on their first syllable. All
possessed nouns that have more than two syllables, get a Η tone on their second

25
These forms are being abandoned in the innovative language. Younger speakers
prefer the construction with a possessive pronoun + noun. However, note that vowel-
initial nouns have kept a 3ps proclitic /n/ in citation form (cf. §7.1.3).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
140 Tone

syllable, as shown in (194). The tonal type is not a relevant factor in this case,
as Type D nouns behave in the same way as other types. The adjective that
follows the noun receives an All-L pattern.

(194) Possessive constructions: Pronoun + Noun + Adjective

Possessor kanu +... Noun Adjective

Disyllabic noun a.läa Caguaje tree') see.o.hwa ('big')
ka.ti ('white monkey') laa.u.hwi.ri ('small')
Disyllabic noun with laa.no ('cassava') see.o.hwa ('big')
heavy first syllable ree.mae ('dog') laa.u.him.ri ('small')
Trisyllabic noun ka.täa.i ('hook') see.o.hioa ('big')
ka.naa.naj ("child') laa.u.hwi.ri ('small')
Longer noun ma.kü.sa.ja.ri ('pepper') see.o.hwa ('big')
si.ni.he.ra ('spouse') laa.u.hivi.ri ('small')

However, there is an additional alternation, depending on the tonal structure of

the adjective that follows the possessum: when the adjective has a stable tone,
i.e. if it does not take the tone pattern imposed by a preceding noun in noun-
adjective constructions, it will surface with its Η tone in addition to the
possessive tone. This is illustrated with tabaj, tasißohioaj (both meaning 'big' -
for discussion see §5.4.2) and kawati 'good' in (195), which both are adjectives
with unchangeable tones. Thus, in examples like this, the principle "One Η per
phrase" is violated.

(195) Possessive constructions with adjective

kanu kä.ti ta.bäj 'my big white monkey'

kanu ka.naa.naj ta.stßo.hwaj 'my big child'
kanu si.ni.he.ra kaxua.ti 'my beautiful wife'

One can conclude that the pattern "H on second syllable" is the regular melody
applied to all nouns in possessive constructions, including any suffixes or
elements belonging to the same syntactic phrase. The fact that in "simple"
possessive constructions, disyllabic nouns receive an initial-Η tone, could be
regarded as a mechanism that applies in case of syllable shortage (i.e. not
enough syllables for a pattern to be assigned), as was observed with verbs
before. An enigmatic aspect of this is that there is no obvious reason why a Η
tone could not be assigned to the second syllable of a disyllabic word:
apparently one "needs Three in order to count up to Two".

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Sentences: More than one tone 1 41

Regarding the "order" of tone assignment in this context, the rules applying
to noun-adjective constructions are neutralised in favour of the possessive
construction.

4.7 Sentences: More than one tone

In the previous sections it has been illustrated that on the basis of tonal classes
for nouns, syntactic phrases which involve a noun have a largely predictable
tonal structure. Most typically, there is one Η tone per constituent and its
position depends on the tonal class of the noun which precedes a verb,
adjective, or postposition. The verbs involved in verbal phrases analysed so far
were transitive. In §4.7.1, I will discuss different tonal types for verbs.
Comparable to the tonal structure of Type D nouns, there are some verbs that
appear to be lexically prespecified for tone, i.e. they deviate from the final-H
pattern. In §4.7.2, it will be shown that a clause with an overt noun subject
contains more than one tone. I then analyse the tonal peculiarities in a few
selected constructions such as clauses that involve adverbs (§4.7.3), focus
constructions (§4.7.4) and serial verbs (§4.7.5). There is a range of further
contexts that also exhibit special tone patterns, such as infinitive constructions,
subordinate clauses, participles, nominalisations, and sentences with introducer.
These involve a set of complex peculiarities, which will not be discussed here.
It should also be noted that, while certain principles can be identified, the
generalisations made here are tentative at the present stage.

4.7.1 Tonal patterns on verbs

Nearly all transitive verbs have a final Η tone when they are in isolation or
when they are followed by a subject noun. Note that a 3ps object of a clause is
usually not marked, i.e. a zero object is automatically understood as 3ps
pronominal referent. The examples in (196) show two transitive verbs with zero
objects; both are realised with a final Η tone, as is the subject. The tonal class of
the noun that functions as a subject does not play a role, since the tone
assignment rules apply from left to right, as observed in previous sections.

(196) Transitive verbs without overt object

a)
kwituku-ä katQä
know-3ps/A man
'The man knows him.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
142 Tone

b)
su-uru-ä itgauena-urü
kill-PL-3ps/A elder-PL
'The elders killed it.'

Some intransitive verbs and most stative verbs behave differently f r o m that. A
f e w transitive verbs that contain the intensifying derivational suffix -to, also
deviate f r o m the regular Final-Η pattern for verbs. There are three main tone
classes for verbs. As a general principle, most of the verbs that do not follow the
Final-H pattern, have more than three syllables and typically end in /ia/ or /oa/
(but note that these are not suffixes). The fact that only "long" verbs have a
tonal structure that shows up on the surface in some way corresponds to the
behaviour of Type D nouns as most of these are "long" also. Each group carries
a Η tone on a different syllable:
Type 1: Η on final syllable
Type 2: Η on penultimate syllable, but L on last
Type 3: H-H pattern on the last two syllables (infrequent)

(197) "Long" verbs (in isolation) (Final-L on Type 2 verbs is marked for convenience)

a) Type 1: Final H:
a.haa.to.ä 'be warm'
e.lo.a.to.ä 'be bitter'
a.mu.ri.to.ä 'exceed'

b) Type 2: Penultimate H:
kal.ta.to.a 'be difficult'
ka.wa.to.ä 'be good'
ka.sa.ti.ä 'scare'

c) Type 3: Penultimate and Final H:

he.no.e.to.ä 'circle' (trans.)
ku.noa.ti.a 'feed'
ha.la.i.to.ä 'make a hole'

These tonal patterns remain stable when the verb functions as the predicate of a
clause (such as 'it is green') and when a noun subject follows. In the following
example, all elements carry the same tones as in isolation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Sentences: More than one tone 143

(198) ahaaruto-a katgä

be.warm-3ps/A man
'The man is warm' [due to the heat]

In addition to the three most common types of tonal patterns on verbs, some
affixes can have specific effects on the tonal structure of verbs. One example for
this are the derivational suffix -hiohwa (and allomorphs, cf. §10.2.2), which
itself carries a Η tone on the syllable /hi/ (or variant [si]), as shown in (199a).
Another is the intransitivising prefix ne-, which assigns a Η tone to the
penultimate syllable of most verbs, such as illustrated in (199b). When a long
vowel represents the final syllable, it is realised as a falling (HL) tone.

(199) Tone patterns resulting from derivational affixes

a) Medium degree suffix -hiohioa:
so.ma.hi.o.hwaa 'be almost white'
be.rau.e.si.o.hwaa 'be green(ish)'

b) Intransitiviser ne-:
ne.hve.ra.ja 'care for oneself
ne.rau.tää 'get healthy'
ne.be.ta.kää 'relax'

An aspect that needs further investigation regards the tonal structure of Type D
nouns when following a verb from Types 2 or 3. In this case, the pattern on the
noun changes, as exemplified in (200), where the Η tone on the noun is on the
penultimate syllable followed by a falling (HL) tone on the long final syllable.

(200) Type D noun after "long" verb

ka.wa.to.a enamänaä
be.good-3ps/A young.man
'The young man is nice.'

4.7.2 VS and Ο VA constructions

As already indicated in the previous section, the tonal distribution "One Η per
constituent" is largely retained within a clause. This also applies to simple
sentences with the regular subject-final constituent order (VS or OVA; cf. §18),
where full nouns are present as arguments. In (201b), the object of the verb is a
Type A noun; thus, the verb gets its Η tone on the first syllable. Otherwise, the
Verb Phrase (VP) does not have other Η tones. Another Η tone is only found on

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
144 Tone

the subject 'man', which constitutes a separate syntactic phrase. As a general

principle, a subject NP always carries a Η tone on its own.

(201) Tonal structure in simple VS and OVA constructions

a) VS order:
u-ri-jia=T bartolome
[come-IRR-NEG:3ps/A=ASS] [PSN]
[VP] [NP]
'Bartolome will not come.'

b) OVA order:
enejtgu hoara-a katga
[monkey see-3ps/A] [man]
[VP] [NP]
'The man saw the monkey.'

A different tonal structure applies when object and subject are syntactically
more complex than the examples above. In (202), the object consists of a
possessive construction with an adjective ('my small bag') and the subject is
composed of a noun and an adjective. In this case, the object NP as well as the
verb receive a Η tone. The subject NP bears a Η tone by rule, as predicted
above.

(202) Tonal structure in complex OVA constructions

kanu stura laauhzuiri basihjau-ä katQa bäaso
[lsg bag small] [steaI-3ps/A] [man bad]
[NP]0 [V] [NP]a
'The bad man stole my little bag.'

An interesting alternation is observed with possessive constructions that

function as an object in a transitive clause: recall that disyllabic nouns that
function as possessum receive a Η tone on their first syllable, as in kanu bihi
'my hand' (cf. §4.6.1). When this type of NP is followed by a transitive verb,
the Η tone shifts to the second syllable, while the verb remains toneless, as
illustrated in (203a). However, the shift only applies to possessed nouns that
have a light (V or CV) first syllable, whereas the possessive tone pattern
remains stable on possessed nouns with a heavy (VV or CVV) first syllable (cf.
(203b). For all other nouns in this position, no tonal changes from the
possessive pattern apply.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Sentences: More than one tone 145

(203) Possessive constructions as object NP

a) Disyllabic nouns with light first syllable:

POSS + Noun With verb ('see'-3pl/A)
ka.nu bi.hi ('hand') ka.nu bi.ht hoa.ra.-ku.ru-. a
ka.nu ti.hja ('foot') ka.nu ti.hja kiua.ra.-ku.ru-. a

b) Disyllabic nouns with heavy first syllable:

ka.nu ree.mae ('dog') -> ka.nu ree.mae kiua.ra.-ku.ru-.a
ka.nu laa.no ('cassava') ka.nu laa.no kiua.ra.-ku.ru-.a

Examples such as discussed above demonstrate once more that the tonology of
Urarina is subject to an intriguingly complex system of rules. Further studies are
necessary in order to explore all the alternations also for contexts not
investigated here.

4.7.3 Clauses with adverbs

Most adverbs have a final Η tone, but there are a few that exhibit a different
tone pattern.

(204) Selection of adverbs

a) Final Η [e.nä] now

[e.re.si] 'tomorrow'
[ka.ra.ha.T] 'long (time)'
[he.raa.e ] 'slowly'
[rau.to.hwe.T] 'calm(ly)'

b) Other pattern [jio.ae.lu] 'at an earlier time'

[ta.bii.tea ] 'sometimes'

The position of adverbs is variable. The combination of adverbs with verbs in

simple constructions does not have any impact on the tonal structure of verbs or
adverbs.

(205) ena itQa-kuru-a 'They did it now' ('do'-PL-3ps/A)

itQakuruä ena 'They did it now'
itQakuruä eresi 'They do it tomorrow'
itgakuruä heräae 'They did it slowly'
itQakuruä ßoäelu 'They did it earlier'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
146 Tone

The intensifier [ha.ta.Y] 'very', can precede or follow a verb. Syntactically and
morphologically it behaves like an adverb. In both preverbal and postverbal
position, it does not have any impact on the tonal structure of the verb, and it is
not modified itself.

(206) Constructions with intensifier hataT

hatai' kaiuatoa 'It is very nice.'

kaioatoa hatai' 'It is very nice.'

4.7.4 Focus constructions

The basic constituent order of Urarina is OVA/VS, as the examples discussed so

far have illustrated. This order can be modified through the use of a focus
marker (=te, =ne, or =na), depending on the person of the subject (cf. §19). The
verb is obligatorily inflected with suffixes from the Ε-inflection type (discussed
in §11). While a wide range of different words or phrases can be shifted to the
front in this construction, I will discuss only noun fronting here. For this context
(207a), the following tonal patterns were observed:
- In examples with intransitive verbs, all nouns (here in clause-initial position,
and followed by the focus marker) are L-toned.
- The focus marker (an enclitic attached to the end of the phrase that is in
focus) receives a Η tone.
- The verb bears a final Η tone.

Note that in examples (207a,b,d), a subject is fronted, whereas the example in

(207c) exhibits an object followed by the focus marker. Thus, the tonal pattern
observed in these examples is not limited to subject fronting, but occurs with
any noun to which the focus marker is attached.

(207) Focus construction with intransitive verb

a) SV order with 3ps focus marker:

ka.tQa.=te si.ni
man=FOC sleep:3ps/E
'The man sleeps.'

b) AV order with 1 sg focus marker:

ka.nu.=ne kivi.tu.ku-u
lsg=FOC:lsg know-lsg/E
Ί know him.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Sentences: More than one tone 147

c) OV order with 1 sg focus marker:

a.ka.no.=ne hua.ra.-u
snake=FOC:lsg see-lsg/E
Ί have seen a snake.' [focus on 'snake']

d) SV order with 3ps focus marker, Type D noun:

e.na.ma.naa.-te si.ni
young.man=FOC sleep:3ps/E
'The young man sleeps.'

As seen in (207b), the situation is slightly different when a Type D noun is

involved. In this case, the Η tone is retained according to the tonal structure of
the noun. In this example, the Η tone occurs on the final syllable of the noun
and the focus marker remains toneless.
In a transitive clause, the tones on subject and focus marker behave in the
same way as in an intransitive clause. However, the object noun, which directly
precedes the verb, triggers the usual pattern on the verb, according to the tone
class of the noun.

(208) Focus construction with transitive verb and noun object (AOV order)

a) With Type A noun as object:

katg>a=te enejtQu sw-e
man=FOC monkey kill-3ps/E
'The man killed a monkey.'

b) With Type Β noun as object:

ka.tQa.=te o.ba.na hoa.ra-e
man=FOC peccary see-3ps/E
'The man saw a peccary.'

c) With Type C noun as object:

ka.nu.=ne ree.mae kiva.ra.-u
lsg=FOC:lsg dog see-lsg/E
Ί saw a dog.'

d) With Type D noun as object:

ka.nu.=ne e.na.ma.naa hva.ra.-ü
lsg=FOC:lsg young.man see-lsg/E
Ί saw a young man.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
148 Tone

4.7.5 Serial Verb Constructions

Urarina serial verb constructions typically consist of two verbs and their
combination is restricted. While verb 1 can be of almost any type and appears in
the "neutral" form (identical to the citation form), verb 2 usually is represented
by 'go' or 'come' (cf. §17.2). In approaches to describing the characteristics of
serial verb constructions, it is usually assumed that these represent a single
phonological unit (e.g. a phonological word). This is clearly not the case in
Urarina, if we assume that a phonological word has only one Η tone - at least
this has been demonstrated so far for the "regular" patterns. In Urarina serial
verbs, each verb carries a Η tone, as is illustrated in (209). In (209a), where verb
1 is intransitive, both verbs receive a Η tone on their final syllable. With
transitive verbs as verb 1, the verb pattern follows the regular principles for tone
assignment: with a Type A noun as object (as in (209b)), a Η tone goes to the
first syllable of verb 1. In contrast, Type Β nouns trigger a Η tone on the third
syllable of verb 1 (cf. (209c)). Verb 2 receives a final Η tone in all three cases.

(209) Tone assignment in serial verbs with 'go' and 'come'

a) With intransitive verb 1:

a.mu.-a ktt.-a
walk-NTR go-3ps/A
'He went hunting.'

b) With transitive verb 1 and Type A object noun:

raa.na sü.-a u.-ä
peccary kill-NTR come-3ps/A
'He came to kill peccaries.'

c) With transitive verb 1 and Type Β object noun:

o.ba.na hoi.tuM.-a u.-a
peccary know-NTR come-3ps/A
'He came to know peccaries.'

However, the examples in (209) show only very short forms of verb 2. When
the verbs 'go' or 'come' occur with one or several suffixes that make it at least
trisyllabic, the Η tone is found on the penultimate syllable. Note that suffixes
which trigger special tonal patterns may cause deviations from that (cf. §4.3).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Summary of tonal types 149

(210) Tone assignment in serial verbs with longer verb 2

a) With future form on verb 2:

a.mu.-ä ku.-ri.-tQä.=i
walk-NTR go-IRR-3ps/A=ASS
'He will go hunting.'

b) With distal imperative (polite) form on verb 2:

raa.na su.-a ku.-ni.-tQU.=toe
peccary kill-NTR go-DSTL-IMP=PLT
'Go kill peccaries!'

4.8 Summary of tonal types

The difference in tone assignment between the four different tone classes
proposed for nouns runs through this description of Urarina tone like a red
thread. Summarised in a simplified way, the respective tonal types of nouns
have the following effects on a word that follows.
Type A surfaces all-L and assigns an initial-Η tone to a word that follows.
Type Β surfaces all-L and assigns a Η tone to the third syllable of a word that
follows.
Type C surfaces all-L and assigns a final Η tone to a word that follows.
T y p e D contains a Η tone but does not assign a Η tone to a word that follows.

In isolation, nearly all nouns and most words of other classes have a final Η
tone in common. It is plausible to characterise this as the default pattern for
Urarina words: whatever underlying structure a word may have is neutralised
when the word occurs in isolation. However, one has to remember that some
words have an underlying tone pattern that differs from this: some adverbs
exhibit different patterns and a number of verbs have tonal structures that do not
match the default pattern. Therefore, it must be assumed that some words have a
lexically prespecified, stable tone.
Type D nouns appear the easiest to account for. In some way, the Η tone of
Type D nouns resists most (though not all) patterns that are superimposed to
any words that follow it. One could conclude that these nouns have lexically
specified stable tones; these change only in certain constructions. If we assume
that each syntactic constituent has only one Η tone and if this Η tone is already
assigned to the noun, any following word will be L-toned by rule. T w o further
aspects support the idea of Type D nouns being lexical exceptions: firstly, the
fact that the number of Type D nouns is smaller than that of other types, and

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
150 Tone

secondly, the fact that this group hosts most loans and proper nouns. A similar
approach can be applied to verbs and adjectives which have a "stable" tone.
Type C nouns very consistently follow their pattern. One possibility to
account for the tonal structure they trigger on a following word simply is to
assume that they are underlyingly toneless and therefore have no impact on any
word that follows them. In constructions that involve more than a word, one Η
tone must be assigned to a syntactic constituent. The result, which involves
lowering of the nouns plus a Η tone on the final syllable of the second element,
represents a realisation of the default rule: if no Η tone is assigned through
underlying structures, a default final-H tone will be assigned.
Type A nouns are the most frequent type in the lexical database. The fact
that a Η tone is consistently assigned to the first syllable of a word that follows
might indicate that these nouns have an underlyingly final Η tone. Since it is
suppressed by syntactic restrictions (e.g. "Assign the Η tone to the second
element of a constituent"), it shifts to the adjacent syllable to the right.
The most challenging group of nouns is represented by Type B. For these, it
is possible to establish a tone assignment rule that places a Η tone on the third
syllable of a word that follows the noun. However, it is difficult to characterise
this regularity as a plausible tonal rule. Certainly, it is not a spreading rule, as
the first two syllables following the noun are unaffected. Why would a Η tone
shift to the third syllable of a word, and, especially, how? To complicate things
even further, one can only "count" up to the third syllable (from left to right) if
more segmental material follows (i.e. a mora or a syllable, cf. (188)). And why
is syllable structure significant only in certain positions, such as in the above
example, at the end of a following word? Regarding Type Β nouns, these
questions remain unanswered at this stage.
It is evident that the tentative account given here is subject to speculations on
the underlying tonal structure of Urarina nouns. A deeper analysis that takes
into account all the regularities and the numerous deviations from these will be
indispensable. In further studies of Urarina tone, one should also take into
account the close relation between tone and syntax. In the introduction to this
chapter, I proposed that possibly each syntactic constituent, rather than each
word, has a single Η tone. In fact there is a tendency for many constructions to
follow this regularity, but some modifications to the claim will be necessary.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
5 Word classes

The division into word classes such as nouns, adjectives and verbs, is not
always straightforward in Urarina. Some words that are intuitively assumed to
be adverbs, conjunctions, or elements of other categories (from their semantic
content) happen to be inflected as verbs. Linguists typically distinguish between
lexical and functional classes. From a cross-linguistic perspective, the first type
usually includes nouns, verbs, adjectives, and adverbs in underived form, which
are discussed in §5.1. to §5.4. In most languages, these are open classes, i.e.
new words can be added to them, whereas the functional classes are typically
closed. In Urarina, the classes of adverbs and adjectives are underrepresented -
more exactly, they have only few underived members. The small group of
underived adverbs represents a closed class. However, the creation of adverbs is
possible through derivation with the participle form (cf. §5.3). In contrast,
ideophones appear to form an open word class along with nouns and verbs.
Verbs mainly function as the head of a (transitive or intransitive) clause, but
in derived form, they can have several noun-like functions through
nominalisation; when a verb is derived by the participle form, it also functions
as a verbal modifier. N o u n s function as the nominal head, as copula
complement, as an argument of a postposition, or as a noun modifier in certain
isolated cases (cf. §5.1.3). The basic relations between the major word classes
and functional slots are summarised in table (211). Note that adjectives are not
listed here, since they do not represent a homogeneous class in Urarina (cf.
§5-4).

(211) Word classes and functional slots

Function Verb Noun Adverb

Head of intr. clause yes no derived (PRT)
Head of trans, clause yes no derived (PRT)
Head of NP derived (NOM Sbj ) yes some
Modifier in NP derived (NOM Sbj ) some some
Copula complement derived ( N O M S b i ) yes no
Argument of postpos. derived ( N O M S b j ) yes yes
Modifier of verb derived (PRT) rare/derived (PRT) yes

There are some other sets of words that are often associated with separate word
classes in other languages, but which are subtypes of nouns in Urarina. These

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
152 Word classes

include some pronouns (cf. §5.5), demonstratives (cf. §5.6), and some numerals,
as will be shown in §5.12.

5.1 Nouns

The vast majority of Urarina nouns have either two or three syllables, but
quadrisyllable nouns are not uncommon, while nouns with five syllables are
exceptional. This also applies to monosyllabic nouns, which are a small
minority.

(212) Word length of canonical underived nominal roots

Number of syllables Number of entries In %

1 10 1.4
2 219 31.3
3 336 48.0
4 125 17.9
5 10 1.4
TOTAL 700 100.0

N o u n s can be marked for number, possession, and location (in some cases). The
unmarked form of a noun is the singular, which corresponds to the root and
which also functions as the citation form. Plural is marked by the suffix -uru,
which has the allomorphs -kuru and -tQuru. The alternation between these
f o r m s is partly predictable, as discussed in §3.1.1. In summary, trisyllabic nouns
tend to be pluralised with -kuru, but with -tg,uru if these end in /i/, while most
of the rest receive the suffix -uru. On the one hand, plural marking is optional
and can be omitted if the context implies plurality (cf. §8.7). On the other hand,
also nouns that are normally uncountable can receive a plural marker. In this
case, the meaning of 'different kinds o f is indicated, such as with ' w a t e r ' ,
akau-uru, (which means 'waters', i.e. rivers, creeks etc). Similar examples for
this are illustrated in (213).

(213) Plural of mass nouns

itQunaj-uru 'honey' (different kinds from different insects)

omane-uru/-kuru 'grass' (different types of grass by form and colour)
tebe-uru 'salt' (various bags of salt)
kuitQana-uru 'blood' (different types of blood, e.g. from people and
from animals)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nouns 153

A classification of nouns into different subclasses is possible through a number

of parameters summarised in (214). These parameters include the question
whether nouns of a certain class can co-occur with a demonstrative, an
adjectival modifier, or whether they function as a modifier themselves. Most
nouns can take a plural form, but not all. Others can be suffixed with a locative
marker. Prototypical nouns can be head of a possessive constructions, but nouns
from most other subclasses cannot, or only some members out of a specific
subclass can.

(214) Properties of noun subclasses

Ν subclass With DEM With adj. Function Take Take LOC Head of
modifier as modif. plural suffix POSS
Standard Ν yes yes no yes place/person yes
reference
Locational Ν yes yes some some yes some
Relational Ν as argument no no no no no
Proper Ν yes yes yes person yes no
names
Adjectival Ν yes (n/a) yes yes no yes
Pers. pronouns no no no 3pl only no no
Demonstratives (n/a) yes yes yes yes no

5.1.1 Proclitics and nouns

In the traditional language, some nouns are combined with possessive proclitics
for 1st and 2 nd person singular (cf. §5.5). Even though this kind of cliticisation is
being lost in the contemporary language, the possessive marking tends to be
preserved for kinship terms and body parts, which points towards a system in
which inalienable possession used to play a role. (Note that some nouns are
subject to phonological changes in their root when prefixed, but the data
available on the traditional language does not suffice in order to make
generalisations.) Possession for lsg is marked by the proclitic ka= and by i= (or
alternant d$=, cf. §3.5.2) for 2sg, as illustrated with kinship terms in (215).

(215) Possessive proclitics with kinship terms

ka=tanaa 'my mother's brother'

ka=rißaha 'my grandfather'
i=ßaka 'your father'
d$=ukioala 'your younger brother'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
154 Word classes

A more detailed analysis of nouns that can take a proclitic instead of a full
possessive pronoun is given in §7.4.

5.1.2 Nouns and the expression of location

Nouns referring to location do not represent a homogeneous group in Urarina.

However, there are subtypes of such nouns that share certain features. For
instance, some nouns that inherently refer to a location can be suffixed with the
locative marker -«. Urarina distinguishes between such "geographical" (or
inherently locational) nouns, as illustrated in (217) and relational nouns (cf.
(219)). The latter represent a heterogeneous group of words with members that
share some similarities but also exhibit differences, as is discussed in this
section. For example, most of these do not take the locative -«. In contrast to
locational nouns, most relational nouns obligatorily take an argument. A
summary of the various properties is given in table (216). The postposition he is
a valency increase marker that can also be used with instrumental function (cf.
§16.5). It is found with the relational nouns for 'left' and 'right'. With
geographical nouns, ke is not used in a locational sense, but may occur as a
valency increase marker to introduce an external argument (e.g. 'urinate "at"
something'; also see ex. (897) in §16.5).

(216) Summary of nouns related to location

Take an argument Take -u Occur with ke

"Geographical" nouns
(cf. (217), (218));
All place names;
dede 'sky'
nuse 'town'
nunaa 'forest'
not typically yes not typically
ahaenaa 'mountain'
atane 'land'
hjaruhioa 'urinal'
bahitguhwa 'toilet'
eshvela 'school'
(nukue) 'river'
Relational nouns A:
edaa (cf. (220), (221))
yes no no
akatia (cf. (223))
auama (cf. (224))

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nouns 155

Take an argument Take -Η Occur with ke

Relational nouns B:
itQüu (cf. (226)) no (inherent?) no
ekuu (cf. (226))
ahaa (cf. (225)) yes no yes
rautow (cf. (225))
ajrima (cf. (222)) optional no no
buhua (cf. (227)) optional yes no

The suffixation o f - « is exemplified with some words in (217). Note that this
suffix merges with the final vowel of the preceding syllable (cf. §3.4).

(217) Noun with locative marker -u

dede-u 'in heaven'

nese-u 'in town'
nauta-u 'in Nauta' (place name)
lima-u 'in Lima' (place name)
ikito-o 'in Iquitos' (place name)
bahitQuhwa-H 'in the toilet'

Note that the noun bahi-tQuhwa literally means 'place to defecate' and is
composed of 'defecate' plus a derivational suffix -ruha, which also occurs with
'urinate'. While this suffix is not productive otherwise, it appears to mark
locative nominalisation. In order to form a similar construction with other verbs,
the suffix -naha preceded by the nominaliser -naa is used, such as in
lenom-ma-naha ('eat'-NOM-PURP) 'place to eat'.
Nouns occurring with the locative suffix -« are typically found in
constructions that involve the copula or a motion verb. (Note that the noun
ahaenaa 'forest' (cf. (218b)) has an irregular form with -nu.) Without the
locative marker, the reading of the example 'He/she/it is in the jungle' (cf.
(218a)) changes to 'It is the jungle'. With motion verbs, suffixation of -u with a
limited number of nouns appears to be obligatory as well (cf. (218d)).

(218) Nouns + -u in sentences

a) With copula:
nuna-u=te tie
mountain-LOC=FOC be:3ps/E
'He is in the (high) jungle.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
156 Word classes

b) Irregular form -nu:

ahaenaa-nu=te rte
forest-LOC=FOC be:3ps/E
'He is in the forest.'

c) With place name and copula:

nauta-u-ne ne-ü
PLN-LOC=FOC be-lsg/E
Ί live in Nauta.'

d) With motion verb:

nauta-u=ne ku-u
PLN-LOC=FOC go-lsg/E
Ί went to Nauta.'

Urarina possibly has a few more Iocational nouns, which occur with -u, but
their number is not known. The locative marker cannot be attached to nouns
that refer to a geographical location where one "cannot exist" (e.g. with beru
'way'). Also nouns like 'house' are ruled out as they do not denote a
geographical location.
In addition to the set of nouns listed above, Urarina has a number of
relational nouns that never take the locative marker. These may be characterised
as describing part-whole relations.

(219) Relational nouns

edaa 'off, outside'
ajrinia 'outside'
ÜQÜU 'side'; 'near'
ahaa 'left (side)'
rautono 'right (side)'
auania 'edge'
ekäu 'top'; 'above'
akatia 'side'; next to'
bub.ua 'cover'

Some of these relational nouns (edaa, akatia, auania, ahaa) must take an
argument, which can be a full noun or the 3ps proclitic n=. They may differ
from each other with respect to their compatibility with the locative marker -u
and the co-occurrence with the postposition ke (listed in (216)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nouns 157

(220) edaa with proclitic

n=edaa=te ne
3ps=outside=FOC be:3ps/E
'It is at its outside.'

Further examples for each relational noun follow below. There are two
relational nouns that refer to 'outside', each of them used with different nouns.

edaa 'outside'; ' o f f

The relational noun edaa is used to describe the position of something outside
an object that does not have a roof or cover. This includes containers like
'glass', but is also used with beru ' w a y ' to mean ' o f f the track'. The word edaa
cannot be combined with the locative marker - u and it is not used with the
postposition he. Note that the word is realised as udaa in the Tigrillo dialect.

(221) Relational noun edaa

a)
nii baja-te beru edaa eno-1
that after=FOC way outside enter-PRT
'After that, entering [the area] outside the track ...'

b)
na1 nuse edaa ku-akaanu, nukue auania ku-t
over.there town outside go-lpl/ex river edge go-PRT
"We went out of the city, going by a river side.' [NT: Acts 16:13]

ajrima 'outside'

In contrast to edaa, the word ajrinia refers to 'outside' of an object that has a
roof or cover. It typically takes words such as ' h o u s e ' or ' s h e d ' as an argument,
but is also found with 'shade' (which makes sense if the tree that gives shade is
understood as a kind of cover). It is not combined with -u or he. Ajrima also
refers to the rectangular spot in front of or around the house that is kept free
from grass. While it can optionally co-occur with a noun, it does not take a
nominal argument.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
158 Word classes

(222) Relational noun ajrinia

a)
lureri ajrinia=te ne
house outside=FOC be:3sg/E
'It is outside the house.'

b)
hi ajrinia hetau kioajtel babaaku-ure
there outside HRS again be.sitting-3pl/E
'Again they were sitting outside.'

akatia 'in front o f

This noun always occurs with an argument. It is not combined with - u or with
he.

(223) Relational noun akatia

a)
lureri akatia
house next.to
'in front of the house'

b)
n^akatia launeto-a u-e hanulari
3ps=next.to be.sitting-3ps/A come-3ps/E jaguar
'The jaguar came to sit down in front of him.'

auama 'edge'

T h e w o r d auarua refers to ' e d g e ' or ' m a r g i n ' and needs an argument. It is not
compatible with -u or ke.

(224) Relational noun auama

a)
n=auania
3ps=edge
'at its edge'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nouns 159

b)
nemuri auania turu-1 ku nelaurja-1 lauek-e
lake edge arrive-PRT there sit.down-PRT be.sitting-3ps/E
'He arrived at the edge ( = shore) o f the lake, sat down and was sitting there.'

(a)haa 'left side'; rautono 'right side'

(a)haa 'left side' is another relational noun that takes an argument. Frequently,
the 3ps proclitic n= precedes this noun. In an alternative form o f ahaa, the
initial /a/ is elided. The corresponding word for 'right (side)' is rautono, which
also refers to 'peace'. An important difference between the two words is that
(a)haa always co-occurs with the word bihi 'hand'. This is not the case for
rautono, which is rarely combined with bihi (but one example is attested in the
NT). Both relational nouns can be followed by the word nitQahaj for 'side'.
They also may co-occur with the postposition ke, but are not compatible with - « .

(225) Relational nouns (a)haa, rautono

a) ahaa with proclitic:

n=ahaa bihi nitgahaj ke=te ne
3ps=left hand side INST=FOC be:3ps/E
'He is on the left (hand side).'

b) ahaa, rautono:
ii (a)haa bihi nitgahaj ke, ii rautono nitgahaj ke
2sg left hand side VLI 2sg right side INST
'one at your left (hand side) and one on your right hand side' [NT: Mark 1 0 : 3 7 ]

c)rautono:
kaa=te liboro rautono ke ne
this=FOC book right VLI be-3ps/E
'It is to the book's right.'

itQÜu 'near'; eküu 'top'; above'

The word itQÜu is related to the verb n-itQÜuta-a 'go near' (with the
intransitiviser ne-, cf. §16.2). As a relational noun it can be used in the sense o f
'in the vicinity o f (cf. (226a)). It does not take an argument and does not occur
with ke. However, the phonological structure of ÜQÜU may give rise to
speculations that it is inherently combined with the locative marker - « . The
same speculations regarding the compatibility with -u can be applied to the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
160 Word classes

word for 'top'. Eküu is also similar to itgüu in that it does not take an argument
and is not compatible with ke. Note that in (226b), eküu appears to have
modifying function, while it is a nominal head in (226c).

(226) Relational nouns itQitu and eküu

a) itgüu:
itgüu ni-a
near be:3ps/A
'He is near'

b) eküu modifying a noun with ke:

eküu enua ke sini-i
top tree VLI sleep-PRT
'sleeping on top, in a tree'
[Note: reading is not 'on top of a tree'; this would be expressed by a postposition]

c) eküu as head of NP:

ßäe ama-uru-a kuruataha-j katga-uru, eküu ne-ti-i
already take-PL-3ps/A two-NOM Sbj man-PL top be-INTS-PRT
'Two men already take [the poles], being on top [of the roof].'

buhua 'cover'
This word is used to refer to 'cover' in the sense of 'protection' and could also
be translated as 'in the cover o f or 'protected by'. In a locational manner it
could also be understood as referring to the location 'behind', since the
protecting object usually covers the hiding participant. It typically occurs with
the locative suffix -u (realised as -ju in (227b)), but is not observed with ke.
buhua may occur with an argument optionally. In (227a), it occurs as head of a
possessive construction ('their uncle's cover')."

26
The variation between -u and -ju for the locative suffix may be based on dialectal
differences for the noun. It seems that buhua is realised as buhuaj in the Chambira
dialect.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nouns 161

(227) Relational noun buhua

a) Head of possessive NP:
akauru itgene buhua-u hetau kwajtel babaturu-a ku-uru-a
3pl uncle cover-LOC HRS again sit.down-NTR go-PL-3ps/A
'They went hiding again in the cover of their uncle.'

b)
ku hetau buhua-ju ne-tü hüara-he
there HRS cover-LOC be-INTS.PRT see-CNT:3ps/E
'She was staying there watching from her cover.'

A summary of properties that distinguish different types of nouns related to

location was given in (216). It becomes evident that the various relational nouns
differ from each other with respect to three features: some obligatorily take an
argument, some may occur with the locative -«, and others can be combined
with the postposition ke, but not all share the same features.

5.1.3 Nouns as modifiers

One strategy for a noun to modify another noun is by nominalisation, i.e. a verb
is suffixed with a nominalising suffix in order to modify another noun. This is
used as a strategy for relativisation (due to the absence of relative constructions)
and is very productive. Urarina has three nominalising suffixes that can be
employed in this manner. One is the suffix -naa, which is otherwise used in
order to derive an abstract or agentive noun from a verb. A nominalised verb
with -naa can occur as a modifier of a noun in any syntactic function. The
position of the verb nominalised with -naa can be before or after the noun it
modifies. A detailed discussion of noun modifiers and their position within the
NP follows in chapter 6 on NP structure.
The examples in (228) illustrate the use of deverbal nouns resulting from
suffixation with -naa (underlined).

(228) Nominalised verbs with -naa as noun modifier

a) Postnominal modifier:
nii ranuna n-itahe-naa kuraa =te marja ne
that girl ITR-lose-NOM girl=FOC Maria be:3ps/E
'The name of the girl who got lost is Maria.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
162 Word classes

b) Prenominal modifier:
saatonom-a ne-naa lureri
be.last-3ps/E be-NOM house
'the last house'

The other nominalising suffix, -i, refers to the subject of an intransitive verb or
to the object of a transitive verb (cf. (229)). Again, the position of the
nominalised verb varies.

(229) Nominalised verbs with -i as noun modifier

a) Postnominal modifier:
ka=raj anofioa ahine-j ua-u
lsg=for knife be.sharp-NOMSbj bring-IMP
'Bring me a knife that is sharp!'

b) Prenominal modifier (subject function):

kau laulauelau-i kanaanaj
here go.by.canoe-NOMSbj child
'the child who was paddling here'

c) Prenominal modifier (object function):

kanu-atca hoaaune-j atane
lsg-only create-NOM0bj land
'the land that I created'

In example (229c), -i refers to a verb object 'what (I) create(d)'. Note that
another nominalising suffix, -era, refers to an A argument and the verb will take
an object as in mi lenorie te-era ('that' 'food' 'give'-AG) 'the one who has given
that food'. In (230), a verb nominalised with -era functions as a modifier. It can
occur before or after the nominal head.

(230) Nominalised verbs with -era as noun modifier

katQa rela-era eene / eene kat$a rela-era

man teach-AG woman / woman person teach-AG
'a woman who teaches people'

In addition to noun modification achieved through derivation, as discussed

above, there are some nouns that can modify another noun, but their use is
restricted and there are only a few (also see §6.5.4). In detail, only nouns that
specify membership in a specific group can function as modifiers. In particular,
this regards professions, or nouns whose semantic content coincides with

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nouns 163

categories typically assigned to adjectives (but note that these concepts are not
expressed by adjectives in Urarina, cf. §5.4). There are only few nouns that can
function as a modifier of another noun and these cannot be freely combined
with any noun, as the following examples show: the noun enamanaa 'young
(man)' can modify the noun kat$a for 'man/person', but its use with ' w o m a n '
or 'peccary' is not possible. Note that enamanaa typically occurs by itself. The
order of such modifiers as enamanaa is not fixed, as it can occur in prenominal
or postnominal position.

(231) "Adjectival concepts" as noun modifier

a)
enamanaa katga / katQa enamanaa
young.man man
'young man'

b)
*eene enamanaa *raana enamanaa
woman young.man w.l.peccary young.man

(232) Profession or mission as noun modifier

eene misjonera
woman missionary
'a woman who is a missionary'

It should be noted that these constructions are elicited and not very productive.
For instance, eene batiri 'a woman who is a priest' is not an acceptable
combination (though this may be based on semantic restrictions). Similarly the
modification of a noun through a noun that specifies tribal affiliation is
restricted: it is not possible with the tribal name for 'Cocama' (a neighbouring
group), while 'Urarina' is used as a modifier. Also the general term for 'Indio'
can function as a modifier.

(233) Tribal affiliation as noun modifier

a)
bakaua katga kiitga
Indio man boy
'son of an Indio'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
164 Word classes

b)
uraripa nuse
PLN town'
'Urarina village'

c)
*kanaanaj kohoamo
child TRN
[to mean: 'Cocama child']

A different type of noun that can modify another noun are nouns that specify
the category of a proper noun, such as place names. The order in these cases is
fixed for each modifier: with the modifier beetle, the proper noun occurs in first
position (cf. (234a)), but in the examples (234b,c), the proper noun is preceded
by the modifier.

(234) Person name as noun modifier

a)
lomaj b-eene
PSN ASCM-woman
'the woman Lomai'

b)
maria lana hose
PSN husband PSN
'Jose, husband of Maria' [NT: Matthew 1:16]

c)
rey herodesi
king PSN
'Herod the king' [NT: Matthew 2:1 ]

Another noun that can be used as a modifier is the word for ' c h i e f (which is
also used for 'male'). In (235), it precedes the proper noun. According to
information gathered in elicitation, the order could also be reversed, but the
example would be ambiguous then: huliä kuraanaa could be understood as
'chief Julian' or as a possessive construction 'Julian's c h i e f .

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Nouns 165

(235) Person name as postnominal noun modifier

kuraanaa huliä
chief PSN
'chief Julian'

In the Urarina N T translation, there are various examples for place names
combined with the respective noun denoting the kind of geographical
specification, such as 'river', 'town', or 'country'. The order in which these
occur is variable. Also note that the proper noun can occur without the
specifying noun. It is normally used in order to introduce a name when used for
the first time in a text. Note that the function of the associative modifying prefix
b- (also observed in b-eene in (234a)), which is discussed further in §7.3.2 is not
entirely clear.

(236) Place name as noun modifier, involving b-

a)
egipto b-atane
PLN ASCM-land
'(the land of) Egypt' [NT: Acts 7:40]

b)
kuniniku b-atane ku-fzoauana turu-Ί sini-akaanu
PLN ASCM-land ASC-harbour arrive-PRT sleep-lpl/ex
'We arrived at the port of Coninico and slept.'

The order of modifier and modified noun can be reversed with nuse (or variant
nuse in the Chambira dialect) 'town'. The word nukue for 'river', always occurs
before the proper noun.

(237) nuse and nukue occurring before the proper noun

a)
lejhii nese betafzuahee
one town PLN
'one (certain) town [called] Bethphage' [NT: Matthew 21:1]

b)
nukue taba-j eufrate auania
river be.big-NOM sb j PLN edge
'at the edge of the big river Euphrates' [NT: Revelation 9:14]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
166 Word classes

The particular properties of nouns and location (cf. §5.1.2) allows the
conclusion that there are separate subclasses of nouns. However, it is difficult to
classify the group of nouns that co-occurs with proper nouns as modifiers. From
the examples attested, one can conclude that these include nouns that refer to
location or types of person. Combinations with any further nouns are not
attested, but cannot be excluded.
A summary of their functions was given in (214). In addition, there are a few
subtypes of nouns that will be discussed in other sections of this chapter. These
include nouns with adjectival properties, personal pronouns, and demonstratives
(cf. §5.4 to §5.6).27 Note that interrogatives are not listed here as there is only
one underived interrogative (cf. §5.9). As stated in §7, nearly all noun types can
function as the head of a possessive NP (cf. (500)).

5.2 Verbs

Verbs can be subject to a large number of inflections. These include, person and
number, polarity, tense, aspect, mood and various kinds of modality/Aktionsart.
Valency-changing suffixes such as for passive and causative can be attached to
some verbs. Generally, verbs have three different forms of inflection for
person/number/polarity, depending on the clause type and/or the constituent
order (cf. §11). Another set of suffixes is used to cross-reference objects and
intransitive subjects under certain conditions. All grammatical categories are
realised as suffixes on the verb; the only productive prefix is the intransitiviser
ne- (cf. (240); §16.2).
Regarding transitivity, Urarina verbs fall into two main classes - transitive
vs. intransitive verbs, with several subclasses the details for which will be
discussed in §9. In summary, there are transitive, active intransitive, stative
intransitive, and reflexive or intransitivised verbs. However, these are not
clearly distinguished morphologically on the lexical level. The differences
become clear only through the use of derivational morphology. For instance, in
nominalisations, different suffixes are assigned to transitive and intransitive
verbs (cf. §6.6). The distinction of stative and active intransitive verbs is subtle
as stative intransitive verbs can take a special plural suffix, but are otherwise
hard to distinguish from active intransitive verbs (cf. §9.2.2). Ambitransitivity is
a marginal feature attested for one verb only, as is seen in §9.4. In §9.5, I will
also argue that Urarina has no ditransitive verbs. Any constituents that appear
external arguments at first sight are in fact adjuncts that can occur with any

27
Note that kinship terms are discussed in §13. While they may be regarded as another
subtype of noun, this distinction is only based on factors related to possession.
Otherwise, kinship terms exhibit exactly the same features as standard nouns.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Verbs 167

transitive or intransitive verb. The special properties of the copula are discussed
in §9.6.
In the following, I will briefly illustrate the function of each verb class.
Transitive verbs and active intransitive verbs are difficult to distinguish from
each other as they do not exhibit any differences on the morphological level. A
distinction is possible through their syntactic behaviour since transitive verbs
take an object as an argument whereas intransitive verbs do not. In (238a), the
transitive verb for 'hear' is shown with the lsg pronoun as an object. The
intransitive verb sinia 'sleep' does not take an object, as shown in (238b).
However, a typical feature of Urarina is that transitive verbs typically occur
without an object if it refers to a 3ps pronoun, as illustrated in (238c): in this
example, the 3pl object pronoun is implicitly known through the context, but is
not realised here (also cf. §18.1 on this subject).

(238) Transitive and active intransitive verbs

a) Transitive verb aunaa 'hear':
d^atoanei kann auna-i=ßa
how 1 sg hear-2ps=INT
'How do you hear me?'

b) Intransitive verb sinia 'sleep':

loanari asae siri raj beree-kuru ajßa
shed under sleep:3ps/E POSS child-PL with
'She slept under a shed with her children.'

c) No object with aunaa:

ku auna-elanaala sini-uru-a
there hear-PRV sleep-PL-3ps/A
'They slept there without hearing (them).'

Stative intransitive verbs, such as tabaa 'be big' exhibit only few differences to
active intransitive verbs that are not visible at first sight. While the plural suffix
for active intransitive verbs always is -uru (as with 'sleep' in (238c)), stative
verbs can take the plural suffix -ana (instead of -uru or in addition; cf. §9.2.2).
This is shown with the verb karetoa 'be long' below.

(239) Stative intransitive verb karetoa 'be long'

kareto-ana-a raj n-usitQue=ne hau

be.long-STPL-3ps/D POSS 3ps=fingernail=SUB because
'because his fingernails are long ...'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
168 Word classes

Reflexivity is expressed by the prefix ne-, which is attached to transitive verb

roots. The suffix has multiple functions and may also be described as a general
intransitiviser, as is discussed in §16.2. The examples in (240) show that ne-
(here realised as n-) transforms the transitive verb itQafiuaa 'shoot' into a
reflexive verb.

(240) Reflexive verb nitQafioaa 'shoot oneself

a) Transitive verb:
obana it$afwa-a kat$a
cl.peccary shoot-3ps/A man
'The man shot the collared peccary.'

b) With reflexive marker:

n-itQafioa-e lejhn kat$a
ITR-shoot-3ps/E one man
Ά certain man shot himself.'

Verbs that typically represent a class of ditransitive verbs in many languages

(e.g. 'give') are transitive verbs in Urarina and do not form a separate class.
This becomes evident through examples that lack an indirect object such as
shown in (241). Note that in this example, also the direct object is not present.
The issue is discussed in greater detail in §9.5.

(241) Transitive verb tia 'give'

ti-a hau hetau=te kuhwu-a ku-ure

give-3ps/D because HRS=FOC fish-NTR go-3pl/E
'As he gave [the fishing hooks] [to them], they went fishing.'

The copula has a variety of functions and may be characterised as an

ambitransitive verb with special properties. When relating to identity or
equation, it occurs with a copula complement and has the same syntactic
behaviour as transitive verbs; when relating to existence or temporary presence,
it behaves like an intransitive verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 169

5.3 Adverbs

While the suffixation of the participle form to a verb is a very productive way to
create adverbs, Urarina has only a limited number of underived adverbs.
Remarkably, almost all of them are time-related. A further feature shared by
these adverbs is that they preferably (not obligatorily) occur in preverbal
position. (However, note that otherwise the position of adverbs is relatively free
in the language, cf. §18.4.2.) Adverbs can be categorized into five types, being
related to time, place, manner, degree, and epistemics. Some isolated words that
do not fit into these subclasses may not be regarded as separate subtypes of
adverbs if they have only one member. Examples for such words are 'only' and
'also', which are not free adverbial forms, but bound morphemes. Since they
carry out functions typically assigned to adverbs, I discuss these under "Others"
(§5.3.6). The table in (242) summarises some of the functions of the adverbs
discussed in this chapter. All adverbs modify a verb, but not all can also modify
a noun. Another distinctive feature is whether they occur with a focus marker in
sentence-initial position, as some are not attested in this combination (but they
can occur in this position without a focus marker). The status of adverbs as
derived or underived is indicated by hyphens between the morpheme
boundaries, where applicable.

(242) Adverbs and their functions

Type Adverb Modifies Ν With FOC

Temporal ena 'now' yes yes
(cf. 5.3.1):
enene 'nowadays' yes yes
raka 'yesterday' yes yes
eresi 'tomorrow' yes yes
kohioamo 'the next day' yes yes
ßäe 'already' no no
hajti 'still' no yes
hauria 'first/before' yes yes
tabiitga 'finally' no yes
ßoae=lu 'earlier' yes yes
na-huaune-1 'again' yes yes
kiuajte-~i 'again' yes yes
kara-ha-ι 'long time' no yes
baja-hiri-i 'soon after' no yes
rihjeori-tü 'immediately' no yes

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
170 Word classes

Type Adverb Modifies Ν With FOC

Locational ka-u 'here' (my area) yes yes
(cf. 5.3.2):
nii-laj 'over there' (in/to that area) yes unattested
ta-elöo 'over there' (towards that yes unattested
direction over there')
Manner rauto-hxoe-1 'quietly' yes yes
(cf. 5.3.3):
m-toane-i 'like that' yes yes
suru-ti-i 'quickly' no yes
kauat^a-l 'well' no yes
hinii-kii 'together' no yes
hva-n-eori-1 'immediately' no yes
helaj 'separate' yes yes
hatal 'very' (Degree, cf. 5.3.4) yes yes
Epistemic heri-a-ne 'maybe/probably' no yes
(cf. 5.3.5):
bawtohwe-t 'totally' yes yes
raui-hi-ri-~i 'really/truthfully' no yes
esißae 'really'; 'truth' no yes
hetg.a 'maybe/certainly' no no
fiatona 'certainly' yes Yes

5.3.1 Adverbs related to time

Urarina has eight underived time-related adverbs. In (244) to (249) the use of
each of these is illustrated by examples.

(243) Time-related adverbs

ena 'now'
ena-ne/ene-ne 'nowadays'
raka 'yesterday'
eresi 'tomorrow'
kohwanoo 'the next day'
ßäe 'already'
hajti 'still'
hauria 'first/earlier'
tabiitQa 'finally'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 171

ena 'now'
enene/enane 'nowadays'

The adverb ena has a wider semantic range than the English word for 'now', as
it can also mean 'a moment ago' or 'in a moment', 'soon', referring to present,
past or future meaning. Apparently, the distinction of an exact point on the
timeline in the literal sense (of minutes and seconds) is not important in Urarina
language.

(244) Use of ena

a) Present reference 'today':

ahaaruto-a ena
be.hot-3ps/A now
'Today it is hot.'

b) Present reference to wider time frame:

poaelu ne-rehet-i trafioitge ena ku ne
earlier be-HABl-NOMSbj distillery now there be:3ps/E
'Now the distillery that used to exist earlier is there.' [Implying 'nowadays']

c) Future reference:
kaa ßaara atane ena-atQa=na lemu-ri-tQäu=ra
this 2pl land now-only=FOC:lsg flood-IRR-lsg/A=EMF
'Now I will flood your land.' [Implied: 'in a moment']

d) Past reference:
ena-te it$a-e
now=FOC do-3ps/E
'He did it right now.' [Implied: 'a moment ago']

In a similar way, the word enene (alternatively enane, depending on dialectal

differences) refers to a time period that can be characterised as 'nowadays', 'in
these days' (as compared to earlier. This word is obviously derived from ena,
but the function of /ne/ in enene is unclear, as ne has a large number of
homophones, mostly describing functional categories (e.g. subordinate marker,
focus marker, negative marker, etc.).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
172 Word classes

(245) Use of enene

a)
nete ßäe enene ne-ene
but already nowadays be-NEG:3ps/E
'But now it [this custom] does not exist any more.'

b)
enane ne-nakauru nijei nitoanei itQa-na najni-ji
nowadays be-those.who not.at.all like.that do-INF be.able-NEG:3ps/A
'The people who live nowadays cannot do anything like that.'

Urarina also has underived adverbs for the description of days in terms of
'yesterday', 'tomorrow', and 'the next day' (with respect to the time of
utterance, e.g. 'the day after tomorrow' if pronounced today).

(246) Underived adverbs for 'yesterday', 'tomorrow', 'the next day'

a) 'Yesterday':
raka=te u-e elo hatal tabai
yesterday=FOC come-3ps/E rain very much
'Yesterday it rained very much.'

b) 'Tomorrow':
eresi turu-i-tge-ne
tomorrow arrive-2ps-PL=CND
'when you arrive tomorrow'

c) 'The next day':

ßäe kohzuanoo hetau ate kuhzuu-a leto-a-naha-e
already the.next.day HRS fish catch-NTR send-PLO-3ps/E
'The next day, he already sent them fishing.'

Another adverb is the word for 'already', which is overwhelmingly frequent in

use. It is also used as a gap filler in very short sentences (cf. §22.5). Beside that,
an utterance such as u-a ('come'-3ps/A) for ' H e has come' would be untypical,
even though it is grammatically correct. A preferred way to express this would
be ßäe ua ('he has come already'). In most recorded texts, there are also many
instances where speakers simply insert ßäe at about any possible position in a
sentence in order to pause. In such cases, it may lead to multiple use of ßäe in
one sentence (which is not considered to be "good style", but which is what
most people do when telling a story). However, multiple occurrence of ßäe
within one and the same clause is not normally observed.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 173

T h e adverb hajti 'still' can be used with both affirmative and negative
clauses. In combination with negation it means 'not yet'.

(247) Use of hajti

a) 'Still':
hajti=te ne ißono ku-ur-era
stiIl=FOC be:3sg/E ayahuasca drink-PL-AG
'There are still those who drink ayahuasca.'

b) 'Not yet':
hajti huituku-a-ü u-ri-tQa=~i itQasu=ne
still know-NEG-lsg/E come-IRR-3ps/A=ASS brother=CND
Ί don't know yet if my brother will come.'

T h e adverb hauria can refer to 'earlier', 'before', or 'at first'. It also occurs in
combination with nehe-ι, w h i c h is the continuous participle form o f the copula.
In this form it can mean 'at first', 'in the beginning'.

(248) Use of hauria, haurianehel

a) 'At first':
hauria aasaern ajrinia ne-uri-1
first a.while outside be-RAP-PRT
'first waiting outside for a while ...'

b)'Earlier':
hauria ne-nakauru katga-uru
earlier be-those.who man-PL
'the people who lived earlier/before'

c) Derived form:
haurianehel eno-akaanu=ne hana
at.first enter- lpl/ex=SUB when
'when we entered [the Espejo River] for the first time'

Another underived adverb is the word tabiitQa 'finally'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
174 Word classes

(249) Use of tabiitQa

a)
rü-a hanutü hana, ßäe tabiitga hoara-a ku-ure
be-3ps/D morning when already finally see-NTR go-3pl/E
'When it was morning, they finally went to look.'

b)
nitQata+hanu-l kuane tabiitga ßäe ekoo-ka u-e
three+day-PRT inside finally already appear-NTR come-3ps/E
'After three days she [the lost girl] finally appeared.'

Some other time- or frequency-related words are verbs or they show signs of
verbal derivation/inflection. The adverb poae=lu 'earlier', 'in ancient times', is a
combination of a root that is also found as ßoae-si 'ancient person'. Though its
root /jioae/ is not a regular verb root, it is used with the remoteness marker =lu,
or alternatively with the recent past evidential =ni (for the function of these
enclitics, see §12.3).
With other roots (whose class membership cannot be specified as these may
be relics of the traditional language), the occurrence of the participle form -~i is
particularly frequent (cf. §20.3). Here, it functions as an adverbialiser. For
example, with the verb for 'repeat' kiuajti-a, it is used as a frozen adverb kiuajtei
'again', but literally, it may also be translated as 'repeating it'.

(250) Some morphologically complex time-related adverbs

jioae=lu 'before' ('earlier'=REM)

na-kioaaune-1 'again' (ITR-create-PRT)
huajte-t 'again' ('repeat'-PRT)
kara-ha-ϊ 'long (time') (Mong'-CNT-PRT)
baja-hiri-Ί 'soon after' ('after'-DIM-PRT)
rihjeori-tii 'immediately' ([unknown root]-INTS.PRT)

The words for 'again' may not be strictly time-related, but they refer to
frequency. The main difference between these two is in origin; in the present
state of the language, they are used interchangeably, kiuajtei and nahuaauneei
can occur separately or in combination, in any order. Semantically, this
combination usually refers to an action or event that has already been repeated
several times.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 175

(251) Use of hoajtei, nahoaaunei

a) hoajtei:
u-ι kiuajtel turu-uru-a
go-PRT again arrive-PL-3ps/A
'Going, they arrived again.'

b) nahoaaunei:
ne-uru-a baja ku nahoaaunei nala-a-ne hana
be-PL-3ps/D after there again drop-3ps/D=SUB when
'after they stayed there [for a long time], when he dropped a fruit again ...'

c) Combination of both: xxx

mhjauria hoajtei nahoaaunei ne-tabataka-he-i turu-i-tge=ne=ra
don't again again ITR-reduce-CNT-PRT arrive-2ps-
PL=NEGF=EMF
'Don't arrive being reduced in number again!'

d) Order variation:
nahoaaunei hoajtei muku-a hana ßäe hoajtei nahoaaunei ku-a
again again catch-3ps/D when already again again go-3ps/E
'When she started to make [Lit. 'catch' children] again, he went (hunting) again.'

Another morphologically complex adverb listed in (250) is the word karahäi,

referring to 'long (time)', which is formed from the verb karaha-a 'stay, last',
and the P R T suffix. Interestingly, similar forms are also used with regard to
' l o n g ' in the dimensional sense: kaare = 'length'; karatia ' b e long'.

(252) Use of karahal

karahal ku ne-ure
long-time there be-3pl/E
'They stayed there for a long time.'

Most other types of adverbs are derived, similar to the morphologically complex
time-related adverbs discussed here. Most commonly, these adverbs contain a
suffix -i, which is a very frequent form and otherwise occurs with verbs (cf.
§20.3). It has a wide range of functions and could also be described as a participle
form or a converb, making the function of the affected verb adverbial (cf.
§20.3.4.3). Some of these adverbs represent fossilized forms: though the probable
original structure of their composition can be recognized, their roots do not occur
in isolation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
176 Word classes

5.3.2 Adverbs related to place

The complex morphological structure of adverbs can also be illustrated with the
group of place-related adverbs, such as 'here' and 'there'. These are
combinations of a demonstrative base (a subtype of noun) with locative
suffixes: kaa 'this' (close to the speaker), mi 'that' (close to the listener), or taa
'that over there' (not close to any of the two; cf. §5.6).

(253) Complex locational adverbs

- ka-u ('here'), ku (in narratives)

- nii-lej ('in/to that area there')
- ta-elöo ('towards that direction over there')

Urarina has no underived place-related adverbs. Other concepts related to

location or direction are expressed by relational nouns such as 'outside' or
'opposite' (cf. §5.1.2).

5.3.3 Adverbs related to manner

Almost all adverbs of manner are morphologically complex. Again, most of

these contain the participle suffix, which is evident from the selection in (254).

(254) Some adverbs related to manner

rauto-hwe-i 'calm(ly)' ('be calm'-CNT-PRT)

28
ni-toane-1 'so/like that' ('that'-'(compare)'-PRT)
kauatca-1 'well' ('be good'-PRT)
suru-ti-1 'suddenly' ('run'-INTS.PRT)
hinii-ki-Ί 'together' ('count'-DER-PRT)

Urarina has two words for 'together', hi?vikn and hiniitQai, which are obviously
related. What appears to be the "root" ([hinii]), however, is not attested
elsewhere. The precise difference between the two forms is difficult to
determine, as they are used interchangeable with regard to meaning. However,
one difference is that hiniitQai can be prefixed with a proclitic (usually being
object or possessive in this position, depending on whether it co-occurs with a

28
The root *toania cannot occur without a demonstrative prefixed to it; however, it is
plausible to split the word into morphemes, as the root also occurs with the
demonstrative for 'this' in ka-toani-a 'be like this'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 177

noun or with a verb). In contrast, hirdikn does not occur with proclitics. Another
observation is that the distribution of the two words is also based on dialectal
differences: hiniikii is the form most widely used in the Espejo dialect, whereas
hiniitQäi appears to be predominant in other dialects, according to Espejo
consultants.

(255) Use of hiniiku and hiruitgal

a)
hiniiku=jia itQa-akaanu
together=FOC:lpl do-lpl/ex
'We did it together.'

b)
nii hana reemae hetau=te rüa sini hiniitQai
that when dog HRS=FOC side sleep-3ps/E together
'Then the dog slept together with him at his side.' [from Airico dialect]

Another interesting example for a complex adverb is hua-m-ori-t 'suddenly',

which is a sequence of a prohibitive introducer, followed by the copula, which
in turn has a velocity derivation and is inflected with P R T suffix. Literally, it
might be roughly translated as 'not being quickly'.

(256) Use of kioaneorit

eno-a ku-a hau hetau=te ßäe hatal hvaneorü

enter-NTR go-3ps/D because HRS=FOC already very suddenly
'When he entered [her place], it became light very suddenly.'

The adverb helaj 'separate' is probably underived and therefore an exception

from other manner adverbs. It can have both adverbial and adjectival function
for ' a l o n e ' . However, also occurring as a copula complement, it may occur in
the same environment as a noun.

(257) Use of helaj

a) 'Alone':
nii eerte ku helaj neda-e
that woman there separate stay-3ps/E
'That woman stayed there alone.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
178 Word classes

b) 'Separate':
helaj=te ne kanaka ana eruari
separate=FOC be:3ps/E lpl/ex mosquito.net
'Our mosquito nets are separate.' [i.e. for each of us]

5.3.4 Adverbs related to degree

The only word used to express degree is the intensifier hatai. Since it has some
peculiarities, it is difficult to decide to which word class it belongs. It can occur
with verbs (in preverbal or postverbal position) and with adverbs. It is also
attested for its co-occurrence with the quantifying noun itulere, as illustrated in
90
(258d); also cf. § 5 . 4 / ' However, the fact that it also ends in -7 supports the
assumption that it might be an ancient derivation from a verb (even though the
meaning of its root cannot be given).

(258) Use of intensifier hatai

a) Preverbal position:
hatai ahi-a=ne hau
very get.drunk-3ps/D=SUB because
'because she was very drunk'

b) Postverbal position:
ßadera-ure hatai katQa-uru
be.sad-3pl/E very man-PL
'The people were very sad.'

c) With adverb:
hatai ena
very now
'very soon'

d) With (quantifying) noun:

aka=te hatai tulere amuri-hiri-1 ne
3sg=F0C very all.kinds surpass-DlM-PRT be:3ps/E
'He surpasses very many different kinds.'

29
In the NT translation, it is also attested with the adjectival noun biifla 'old (person)'.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 179

5.3.5 Epistemic adverbs

Urarina has a f e w adverbs that function as epistemics. While three of the

adverbs listed below appear to be fossilised combinations of several
m o r p h e m e s , the words ßatorw 'certainly' and hetpa ' s u p p o s e d l y ' are probably
underived. N o t e that epistemic concepts are also expressed by the verbal suffix
-kaj, m a r k i n g probability (cf. §12.2.13)

(259) Epistemics

heri-a-ne 'probably' (Lit. 'want'-3ps/D=CND 'if he wants')

banetohive-1 'totally', 'too much' (unknown root + participle suffix)
raui-hiri-1 'really' ('right'-DIM-PRT)
hetQa 'supposedly', 'wrongly believing'
ßatona 'certainly'

T h e c o m p l e x adverb heriam expresses a positive probability of an event to

happen or have happened and can be used with present, past, or future meaning.
M u c h depending on the context, it may be translated in the range between
' m a y b e ' (referring to ' h o p e ' ) and ' p r o b a b l y ' (more regarding a suspicion).

(260) Use of heriane

a) 'Maybe'; referring to present:

heriane ni-a
maybe be.there-3ps/A
'Maybe it is there.'

b) 'Probably'; referring to future:

heriane=teuu-re=l
probably=FOC come-IRR:3ps=ASS
'He will probably come.'

c) 'Probably'; referring to past:

mi hau heriane enanihja itQa-fotru-i=lu
that because probably canoe make-PL-NEG:3ps/A=REM
'Therefore, they probably did not make canoes.'

heriam is also attested in combination with -he-1, the participle form of the
continuous aspect, which is normally attached to verbs. Its meaning does not
change substantially through this suffixation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
180 Word classes

(261) heriane in combination with -he-t

heriane-he-i tQU=ra kalau-itQa siiri
probably-CNT-PRT CRTN=EMF son-only have:3ps/E
'Probably it has had a baby.' [about a dead animal]

In contrast to heriane, the word hetQa (also pronounced hit$a in the Tigrillo
dialect) expresses an unfulfilled expectation. Using hetga, the speaker describes
a situation of which he knows that the action or expectation described in the
sentence is unfulfilled. For instance, in a sentence such as (262a) Ί said that the
ship would take us', the speaker now knows that this did not occur, whereas in
the original utterance he was convinced that the ship would take them. Similarly
in (262b), the speaker describes his own failure to do something which he
thought he had done. In (262c), the subject of the main verb ('come') is
different from the one in the dependent clause ('do'). In this case, the insertion
of the quotative verb naa is obligatory.

(262) Epistemic hetQa

a)
kana ama-ri-tQa=i hetQa na-anu
lpl/in take-IRR-3ps/A=ASS supposedly say-lsg/A
Ί said that it [the ship] was supposed to take us.' [Lit. "would supposedly take us"]

b)
itQa-käu hetQa
do-lsg/A supposedly
Ί supposedly did it.' [Implying: Ί thought I would do it but I did not.']

c)
itQa-a hetQa m-ι u-u
do-3ps/A supposedly say-PRT come-lsg/E
'(Wrongly) supposing that he had done it, I came.'

The adverb flatona indicates the positive probability of an event or action.

However, jiatona means 'certainly' in (263a), whereas in (263b) it has the
meaning 'exactly', 'in fact' - which does not express probability of any kind
but is an affirmation of what had been predicted before. The low number of
examples also does not provide enough evidence to state whether jiatona
specifically refers to the past, present, or future. The sentence displayed below
occurs in past context.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 181

(263) Use of ßatona

a) Meaning 'certainly':
ßatona =te aka ne
certainly=FOC 3sg be-3ps/E
'Certainly it was him.'

b) Meaning 'exactly':
ßatona nil ajto-a hau rihihei kauatü esißae teru tokuanei
exactly that say-3ps/D because like nicely really axe compared.to

hetau toku-hzue
Η RS be.lying-3ps/E
'Exactly as he [the guide] had said before, there was lying really nicely [something] like
an axe.'

Another word that is used in adverbial function esißae 'really' (also found in
(263b)). However, it can also function as a noun, as shown in (264b), where it
occurs as a nominal head.

(264) Use of esißae

a) Adverb:
esißae kauatga-ι kana+hoaaun-era saku-i
really good-PRT our.creator follow-NOMsbj
'someone who follows God really well'

b) Noun:
kana+hoaaun-era=te esißae rauta-e
1 pl/in+create-AG=FOC truth like-3ps/E
'God loves the truth.'

The meaning of 'really' is also expressed by rauihirii, which is the participle

form of the verb rauihiri-a 'to be right/straight'. This verb itself is
morphologically complex and composed of the noun raui 'right' followed by
the diminutive suffix -hiri and the participle suffix -7. It is also used in the sense
of'straight/clear'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
182 Word classes

(265) Use of rauihirii

a)
akauru raj rauihirii ku be
3pi to really there tell:3ps/E
'There she really told it to them.'

b)
rauihirii nii hamlari asi bihijejto-a hau ku-atga n=asi
really that jaguar nose strike-3ps/D because there-only 3ps=nose

kahe itgana tutuara+tutuara-Ί

from blood RED+make.flow-PRT
'As>he really stroke the jaguar's nose, making the blood flow right there out of his nose

In m a n y instances, rauihirii is used in combination with esiflae, with

emphasising function.

(266) Combination of rauihirii and esijiae

a)
hatal esijiae rauihiri-ι ku kana+kioaaun-era saku-i ne-tiakauru
very really be.right-PRT there our.creator follow-NOMSbj be-those.who
'those who really very much in truth are followers of God'

b)
atii kanakaana kuruatahane-u, esißae rauihirii ajto-a ii ere
please lpl/ex help-IMP really really say-3ps/D 2sg word

kuane hau rihihel

inside because like
'Please help us, as it is really and in truth said in your word ...' [From prayer]

5.3.6 Other morphemes with adverbial function

T h e expressions for ' a l s o ' , ' o n l y ' , and ' e v e n ' are independent words belonging
to an adverb class in many languages. Urarina has different strategies in order to
express these concepts. There are t w o f o r m s for ' a l s o ' : -netonaj and -nemaahel.
W h i l e I treat these as suffixes, it is difficult to distinguish them f r o m enclitics.
T h e only w a y in which they differ f r o m enclitics is their limited selectivity: both
can only occur on nouns (including demonstratives and pronouns), but are not

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs \ 83

attached to any other word class. They undergo the same phonological
alternations as other suffixes. For instance, when -netonaj is attached to a
vowel-final root, the initial nasal is assimilated. When it follows the root vowel
/«/, it is realised as [rj], based on a rule that operates root-internally and between
roots and suffixes. As a peculiarity occurring only with -netonaj, the /n/ is elided
after the vowel /a/ and the root final vowel is nasalised in turn (cf. (267c)).

(267) Assimilation of -netonaj and -nemaahel

a) Velarisation:
kanaka anu-rjetonaj eno-akaanu=ne kujßa-ra
lpl/ex-also enter-lpl/ex=SUB so.that=EMF
'that we also may enter'

b) Palatalisation:
kanaanaj-ßemaahei hmtuku-a
child-also know-3ps/A
'He also knows the child.'

c) Elision of /n/:
aka-etonaj aka-uru kaihje ini-a ku-e
3ps-also 3ps-PL behind go.up-NTR. go-3ps/E
'She also climbed up behind them.'

Distinguishing the two forms for 'also' from each other is a major challenge, as
their semantic and functional properties appear to be virtually identical. The
meaning of the two sentences contrasted in (268) was judged to be "the same"
by native speakers. However, note the unusual word order in these examples,
where the subject occurs in initial position without a focus marker (cf. §18.3).

(268) Two forms for 'also'

a) -netonaj:
akä-etonaj kivaa-na here
3sg-also see-INF want:3ps/E
'He also wants to see it.'

b) -nemaahel:
aka-nemaaheet huaa-na here
3sg-also see-INF want:3ps/E
'He also wants to see it.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
184 Word classes

A possible difference could lie in the history of the two suffixes. While this is
highly speculative, one assumption would be that -nemaahei is derived from
*ne-amaa-he-l (ITR-'take.along'-CNT-PRT), which might be translated as
'including'. The suffix -netonaj is even less transparent to analysis and cannot
be interpreted.
Interestingly, the two forms occur in combination in several examples of the
NT, which would further justify the assumption that -nemaahei can be glossed
as 'including'.

(269) Combination of -netonaj and -nemaahei

arahn katga-uru, eene-kuru-rjetonaj- pemaahei ... neheotehe-ure
many man-PL woman-PL-also-including be.added-3pl/E
'Many people and also including women ... were added.' [NT: Acts 5:14]

One can conclude that Urarina has no independent word for 'also'. Note that
there is the word tg,äe 'also', 'besides', which functions as a conjunction and as
an adverb (cf. §20.5.1).
In a similar way as the forms for 'also', 'only' is realized as a suffix -at$a
(or allomorphs), most typically attached to nouns (270a), but also found on
postpositions (cf. (270c)). It has an emphatic function and can also be translated
as 'self in some examples (but note that there is a separate reciprocal form).

(270) Use of-atga

a) With noun:
hjane-etQa he hetau=te ne-rua-kure
achiote-only INST HRS=FOC ITR-rub/smear-3pl/E
'They rubbed themselves with achiote only.'

b) With locational demonstrative:

ku-atga uahei
here-only until
'till here only (and not any further)'

c) With postposition:
lureri kuane-etQa =na=ra ne-betaka-na here-ü
house inside-only=FOC: 1 sg=EMF ITR-relax-INF want-1 sg/E
Ί want to relax only in the house.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adverbs 185

d) With adverbialised verb:

d^atoanel-tQa=ra sini batiri
how-only=EMF sleep:3ps/E priest
' H o w (only) does the priest sleep?'

The suffix -atQa can even occur on verbs where it is directly attached to the
verb root and followed by inflectional or other suffixes. In (271a), it precedes
the imperative. In this context, it has the connotation of 'simply', 'just'. In
(271b), it refers to 'only' in the sense of'nothing else'.

(271) -atQa suffixed to a verb

a)
lenone-etpa-u
eat-only-IMP
'Just eat!' [emphatic, considered impolite]

b)
kana nekivehe+d$anu-atQa-ki-tQe=ra
lpl/in shame+make.feel-only-2ps-PLT=EMF
' Y o u (will) just put shame upon us!'

An additional function of -atQa might be related to its emphatic use and its
meaning 'only', but is not translated as such in one particular construction: in
nominalisations with the derivational suffix -i. This morpheme can refer to the
subject of an intransitive clause or to the object of a transitive clause (cf.
§6.6.1). The attachment of -atQa is obligatory to the underlying subject of this
construction.

(272) Use of -atQa in nominalisations with -i

a)
nii-tga here-j itga-kure
that-only want-NOM 0 bj do-3pl/E
' T h e y did what he wanted.'

b)
nerauta-e nii lejhii katQa, akaw-atQa kunajta-j
get.well-3ps/E that one man snake-only hurt-NOMobj
'That one man got better, the one whom the snake hurt.'

The suffix -atQa has several allomorphs. Its realisation depends on the
preceding vowel.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
186 Word classes

a) After /a/, /«/, / u / and after the diphthong / a « / , it surfaces as [atpa].

b) After /e/, and /i/, the initial vowel of -atg,a is assimilated to the respective
vowel to result in [et$a], [it$a].
c) After vowel sequences and after the diphthong /aj/ (one example only), it is
realised as [tpa].

(273) Allomorphy of -at$a

Context Examples Gloss

After / a / katQa-atQa 'man'-only
jiaara-atQa 'you' (2pl) -only
After / « / enejtgu-atQa 'monkey'-only
kanu-atQa Τ (lsg)-only
After diphthong / a « / akau-atQa 'water'-only
kau-atQa 'here'-only
After / u / arusu-at^a 'rice'-only
letono-atQa 'envoy'-only
After / e / eene-etga 'woman'-only
lewne-etQa 'food'-only
After / i / lureri-itpa 'house'-only
loanari-itga 'shed'-only
After V sequence inio-tga 'meat'-only
kalaui-tQa 'son'-only
After diphthong / a j / kamanaj-tga 'child'-only

In this section, adverbs were categorised according to their semantic functions

and it became evident that temporal adverbs differ from other types
morphologically. In an attempt to list the syntactic properties of Urarina
adverbs, some of their syntactic functions were summarised in (242). Naturally,
all adverbs modify a verb, except for the suffixes for ' a l s o ' . (In order to modify
a verb with 'also' such as in 'he took it and also ate it', different strategies
would be applied, e.g. the use of discourse particles.) Some adverbs may also
m o d i f y a noun. For instance, ena katpa ( ' n o w ' + ' m a n ' ) would refer to 'this
t i m e ' s m a n ' , to imply 'the man I'm talking about n o w ' or 'the man w e are
seeing n o w ' . However, constructions of this type are elicited and do not
normally occur in context. Another distinctive feature of adverbs is whether
they can occur with a focus marker or not. As a matter of fact, only a f e w do not
have this option, as can be seen in the table. It is difficult to distinguish from
elicited examples whether an adverb modifies another adverb or whether at least
one of the adverbs refers to the verb. Apparently, the intensifier hatal and the
epistemic adverbs (except for hetQa) can modify another adverb, but examples
are not attested in context.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 187

5.4 Adjectives

Dixon (2004:3f) identifies four semantic classes that are typically expressed by
adjectives from a typological perspective: dimension, age, value, and colour. If
Urarina has an adjective class, its members are likely to be found in these
semantic groups. Further semantic classes include physical property, human
propensity, and speed. However, while these concepts are typically represented
by adjectives in other languages, they are mainly expressed by nouns or verbs in
Urarina (sometimes by special types of nouns or verbs). For instance, the
concepts for age are mainly expressed by nouns (cf. (286)), while all colour-
related terms are verbs (cf. (275)). Stative verbs can be nominalised to take
modifier function. Two dimension-related words come close to prototypical
adjectives with regard to their function, but since the morphological and
syntactic features of these do not match exactly, I conclude that there is no
underived class of adjectives in the language.
In defining the typical properties of adjectives, several tests are commonly
used in order to distinguish an adjective class from other word classes. The
main point of distinction then is whether they are "verb-like" or "noun-like".
The following criteria help to identify them for one of the two groups. In order
to compare a potential adjective with a noun, the following checkpoints (listed
here as based on cross-linguistic tendencies) should be involved:
1) Can the word function as a nominal head? - While this is obviously a typical
property of a noun, it is not one of adjectives.
2) Can it function as a noun modifier? - This is a typical characteristic of an
adjective, and in some languages also a feature of nouns. In Urarina, only a
very limited group of nouns can function as such.
3) Can it function as a copula complement? Even though this is a property also
associated with nouns, adjectives typically appear in this function.
4) Can the word take a noun plural? - This is another property that applies to
nouns. In number agreement (which Urarina has only in part and as an
optional feature within the NP) one may expect an adjective to be marked for
number. In any case this would help to distinguish between noun-like and
verb-like adjectives.
Properties that are typically assigned to verbs help to recognize whether a word
is verb-like or is an actual verb:
5) Can it take verbal suffixes?
6) Can the word be modified by an intensifier? 30

30
The intensifier hatal is typically used with verbs but, as seen in the previous section,
it can also modify a noun in some cases. Therefore, this test is not ideal for the
distinction between nouns and verbs.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
188 Word classes

A s u m m a r y of adjectival " c a n d i d a t e s " with respect to t h e s e properties is given

in ( 2 9 9 ) and (300). A n o t h e r typical test that is usually applied to s h o w w h e t h e r
a w o r d is an adjective, is the c o m p a r a t i v e . Since Urarina has no m o r p h o l o g i c a l
c o m p a r a t i v e , this test c a n n o t be used in order to d e t e r m i n e the status o f t h e
w o r d s in question (cf. §5.4.6).
U r a r i n a has a f e w underived w o r d s that can take f u n c t i o n s typically
a s s o c i a t e d with adjectives, but their use is restricted. M o s t w o r d s c a n n o t
f u n c t i o n as noun m o d i f i e r s and as copula c o m p l e m e n t s ; typically, they are
either n o u n - l i k e or verb-like, w h i c h leads to the conclusion that they are special
s u b t y p e s of n o u n s and verbs. T h e first t w o e x a m p l e s in ( 2 7 4 ) are regular stative
verbs; the third e x a m p l e (baaso ' b a d ' ) is a noun that can be used as a n o u n
modifier.

(274) Nouns and verbs referring to quality

nohwia 'be hard' (fully injectable as a verb)

ahaarutoa 'be warm' (fully injectable as a verb)
baaso 'bad person/thing' (injectable as a noun)

T h e r e are s o m e w o r d s , h o w e v e r that have properties not typically shared by

" r e g u l a r " n o u n s and verbs, w h i c h m a y lead to the a s s u m p t i o n that U r a r i n a d o e s
h a v e a small class o f " a d j e c t i v e s " . T h e other w a y w o u l d be to call t h e s e isolated
e x a m p l e s s u b g r o u p s of n o u n s and verbs, respectively, but, ultimately, this is a
mere matter of terminology.

5.4.1 Adjectival verbs

All c o l o u r t e r m s are represented by stative intransitive verbs (also called

a d j e c t i v a l v e r b s here). For their use in adjectival f u n c t i o n , they are derived by
t h e n o m i n a l i s i n g s u f f i x -i, to f o r m a noun m o d i f i e r . Alternatively, they can b e
inflected with the participle s u f f i x -1 (which can be applied to any v e r b ) to f o r m
a m o d i f y i n g clause (e.g. ' a horse, b e i n g r e d ' , cf. (302)). T h e list of colour t e r m s
in ( 2 7 5 ) exhibits that all colour terms end in [ h a j ] (or [ h w a j ] , which results
f r o m labialisation a f t e r /u/; cf. §2.8.3). T h e e n d i n g in -j o f these w o r d s is a
realisation of t h e n o m i n a l i s i n g s u f f i x -i (cf. §3.4).

(275) Colour terms

lanahaj 'red'
somahaj 'white'
hitQuhiuaj 'black'
berauesiohioaj 'green', also used for 'blue'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 189

kafwahaj 'yellow'

Some examples as illustrated with 'red' in (276) show that the respective roots
of these verbs regularly occur with other kinds of inflection. In the same way as
other stative verbs, they occur with different values for polarity, tense, and with
some derivations.

(276) Verbal use of colour term for 'red'

a) Nominalisation:
lana-ha-j
be.red-DER- NOM Sbj
'the red one'

b) Irrealis (future):
lana-ri-tQa=7
be.red-IRR-3ps/A=ASS
'It will be red.'

c) Negation:
lana-7a-i
be.red-NEG-NEG:3ps/A
'It is not red.'

While the final [j] of the words depicted in (275) is analysed as the
nominalising suffix -z, which merges with a preceding vowel (cf. §3.4), the part
[ha] is a less productive element of the morphology. This suffix occurs as a
derivational suffix with a few stative verbs only; most other stative verbs are not
compatible with it. Besides, the verbal roots referring to colour also occur
without /ha/. For instance, the verb for 'be red' is lanaa-ka. In addition to the
colour terms, the suffix -ha is also found with two other adjective-like words: it
occurs in tasijiohwaa 'be big', whose nominalised form tasißohzvaj also
functions as an adjective, and in seeohwa 'big', which is a noun-like adjective
(for semantic distinctions between the words for 'big' see (284)).
There is another suffix that can be attached to colour terms (but unattested
with other words): -hio (or -sio as in berausiohwaj 'green') indicates the
equivalent of the English derivational suffix '-ish' for these verbs (cf. §10.2.2).
The form lana-hio-hwa-j 'reddish', 'medium red' is also used to describe
colours similar to red for which no separate term exists, such as 'pink'.
Correspondingly, the form hitQua-hio-hiva-j is attested to describe 'purple'. The
word berauesiohzoaj 'green' is obviously derived in the same way, but the
meaning of its original root (*/bera«e/) is unknown. The combination of the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
190 Word classes

two suffixes to -hiohwa is also observed with some other stative verbs, as
illustrated in (277). For example, nohwaa 'be hard' occurs as nohweohwaa with
the meaning of medium degree 'be kind of hard', 'be hardish'. However, this
form cannot be used freely with other stative verbs.

(277) Use of -hiohiua (and related forms) with stative verbs

nohwaa nohweohwaa 'be hard'

tasißohwaa tasißohxoaaohwaa 'be big'
uretoa *urehiohwaa 'be brave'
anaitoa *anaihohwaa 'be heavy'
tabaa *tabahiohwaa 'be big'

A feature that distinguishes stative verbs from other verbs is the fact that they
can modify a noun when they are nominalised with -i. As opposed to the
occurrence of this suffix on intransitive verbs, where it refers to the S argument,
the suffixation to stative verbs refers to a quality and modifies a noun.

(278) Adjectival use of forms involving -i

a) With tabaa 'be big':

ate taba-j
fish be.big- NOM Sbj
'big fish'

b) With tasißohzoaa 'be big':

hanolari tasiflohioa-j
jaguar be.big- NOMSbj
'big jaguar'

Note that in all examples listed above, the modifier follows the noun. This
corresponds to examples in which the nominaliser -i occurs with an intransitive
verb (cf.(279)). However, note that in constructions with a transitive verb, the
nominalised verb occurs before the modified noun.

(279) Use of-i with intransitive and transitive verb

a) With intransitive verb:
äaka ka-kanaanaj laulauelau-i
where, is this=child go.by.canoe- N O M S b j
'Where is that child who was going by canoe?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 191

b) With transitive verb:

akano-atga kurw-i katga
snake-only swallow-NOMSbj man
'the man whom the snake swallowed'

In conclusion about colour terms, it can be said that these are consistently
represented by stative verbs. Derivation with -i can convert these into modifiers,
which is a strategy also found with other verbs.
Another word that can be described as an adjectival verb is äa/ihia 'be
small/little'. Its main use is in the nominalised form, which is used to mean 'a
little'. In this case, it takes the form / ä a f i h i - i / .

(280) Use of äafihia

a) Stative verb:
äafihi-a kaa lureri
be.small-3ps/A this house
'This house is small.'

b) Noun modifier:
äafihi-i nune baku-ri-1 u-a-e
be.small-NOMsbj branch break-RAP-PRT come-CAUl-3ps/E
'He broke off a little branch and brought it.'

c) Nominal head:
lejhu=te aasihi-i ama-e
one=FOC be.little-NOMsbj take-3ps/E
'One [man] takes along a little.' [The other one carries lot.]

A further candidate on the list of possible adjectives is the word for 'good'.
Similar to äafihii, this one is hard to classify because it takes a number of very
different functions. In some cases, morphological derivation is required, in
others, the word simply takes a different role without being altered. There are
two basic forms in which the root for 'good' occurs: kaua-tga and
kaua-toa. The root /kaua/, which these forms have in common never occurs in
isolation. Also /tpa/ is not a suffix found elsewhere in Urarina morphology
(only as a variant of -a after /i/; but note that the root /kaua/ does not end with
l\l). -toa is a frequent derivational suffix found with verb roots in order to
express intensity, or stativity (cf. §10.2.3). From this suffix, the forms
kaua-ti, a n d kaua-tii are derived, -tu is the participle form of -toa, and -ti is
the nominalised form of -toa (a morpho-phonologically irregular combination
of -to + -/). One difference between kauatQa and kanatoa is a semantic one:

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 192 Word classes

while kauatQa tends to refer to ' g o o d ' in terms of character (e.g. a good
worker), kauatoa implies a more visual quality such as ' n i c e ' , 'pretty',
'beautiful' or 'clean', depending on the context.
One function of ' g o o d ' is as a verb, kauatoa can be inflected for tense,
polarity, and person in the same way as any other stative verb.

(281) Use of kauatoa as a stative verb

a)
hatal kauato-a hau
very be.good-3ps/D because
'because she was very beautiful'

b)
kauato-ri-tQa-7 hanonaa
be.good-IRR-3ps=ASS day
'It will be a nice day.'

In its nominalised form with -z, the word functions as a noun modifier in
prenominal or postnominal position.

(282) Use of kauati

a) Prenominal modifier:
ni-a kauat-i tefzuolo-ne
be-3ps/D be.good-NOMsbj church.building=CND
'if there is a nice church building'

b) Postnominal modifier:
ni-a ka=raj teru kauat-i
be-3ps/A lsg=POSS axe be.good-NOM sbj
Ί have a good axe.'

With the suffix -tu, the word is used as an adverb.

(283) Use of kauatii

a)
kauati-i kusißa-u
be.good-PRT close-IMP
'Close it well!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 193

b)
aka=te kauati-i enua sau-e
3sg=F0C be.good-PRT wood cut-3ps/E
'It (the axe) cuts wood well.'

The peculiarities of the word kauatQa for ' g o o d ' are discussed in §5.4.2, since it
is not a stative verb.
Urarina has three different words for ' b i g ' , two of which are stative verbs:
tabaa and tasijiohwaa. These have the same properties as the other stative
verbs described above: with the nominaliser -i, they function as postnominal
modifiers; with the participle form, they can be adverbialised. The other word,
seohwa, is noun-like; the differences between the three words for ' b i g ' will be
discussed in the following section.

5.4.2 Adjectival nouns

While the three terms for ' b i g ' have different syntactic and morphological
properties, their semantic distinction is much less obvious. Generally there
appear to be almost no restrictions as to the use of each of them in modifier
function with nouns, as is clear from the table in (284). As a tendency, however,
one may say that there is a preference to use seohwa for artefacts and tabaj (the
nominalised form of the stative verb tabaa) for nature-related or abstract nouns,
while the use of tasißohwaj (from the stative verb tasißohwaa) is unrestricted.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
194 Word classes

(284) Words for 'big' in Urarina31

tabaj tasißohivaj seeohiua

Humans: 'man' 'man' 'man' (preferred)
'man' 'woman' 'woman' (= fat and tall)
Animals or parts: 'animal' 'animal' 'bull'
'jaguar' 'egg' 'chicken'
'peccary' 'peccary' 'peccary'
(less big than tabaj)
Nature/plants: 'world' — 'world'
'wind' — 'wind'
'rain' — 'rain'
'tree' — 'tree'
'lawn /grass field' — 'branch'
Artefacts: 'canoe' ('house') 'boat', 'canoe'
Abstract nouns: 'suffering' — (unattested)
'darkness' (unattested) (unattested)

Regarding the actual size of the modified object, consultation with different
speakers lead to inconclusive characterisations between the three words. While
with most speakers, tabaj appears to refer to a larger size than the other two
words, the size distinction of tasißohwaj and seohwa is not drawn consistently.
A more obvious distinction is of a syntactic nature. While tasijiohwaj can occur
as a noun modifier, it is less common in this function. The main and preferred
use of tasiflohwaa is as a verb. When it is used as a modifier, it evidently tends
to refer to a process 'having become big' with some nouns, e.g. with humans
and animals. A further restriction to tasißohwaa is that is does not occur with
abstract nouns. In contrast, the verb tabaa tends not to be used with some
human terms, such as 'woman', while its use as a modifier for 'woman' is
possible. In summary, the semantic distinction between the different words for
'big' remains challenging.
An important difference between seohwa and the other two terms is its
syntactic and morphological structure, seohwa is one of only two words that
matches the typical properties of an adjective, in terms of its syntactic
behaviour. Its ending in [hwaa] is clearly non-verbal, since no kind of
derivation or inflection is possible in this case, while it occurs with all stative
verbs that contain -h(w)aa as a suffix. Thus, the form *seohzoa-j or any kind of
inflection for tense or polarity is not possible. It is plausible to conclude that the

31
Nouns in brackets were reluctantly accepted as "possible", but unusual combinations
with the respective modifier. Dashes indicate that a combination is not possible.
"Unattested" combinations were not tested.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 195

two occurrences of [ h w a a ] in seohwa and tasißohzoaa, respectively, is a case

of homophony. In (285), the use of seohwa is illustrated.

(285) Use of seohwa

a)
bute seohwa itga-e
boat big make-3ps/E
'He made a big boat.'

b)
aka=te katpa seohwa ne
house=FOC man big be:3ps/E
'He is a big man.'

Note that seohwa can only occur in postnominal position, whereas some
deverbal adjectives are also found in prenominal position. Even though seohwa
does not occur as a copula complement, it is noun-like in that it can take a plural
form to mean 'the big ones'. From these features, one may conclude that
seohwa is either a noun with adjectival properties or an adjective with noun-like
qualities.
Further adjectival nouns include enamanaa 'young (man)' and biißa 'old
(man)'. Both are used as nominal heads and they can take a noun plural suffix:
the forms emmanaa-kuru, biißa-htru refer to 'young/old people', respectively.
The use of these words as copula complements is also a feature typically found
with nouns (but also with adjectives).

(286) Use of enamanaa, biißa as nominal head

a) Head of NP:
lejhu enamanaa kuriki-o-na
one young.man money-VBL-NOM
'a rich young man'

b) Head of NP:
turu-a biißa
arrive-3ps/A old.man
'The old man has arrived.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
196 Word classes

c) Copula complement:
enamanaa ni-a ka=katg,a
young.man be-3ps/A this=man
'This man is (a) young (person).'

d) Copula complement:
ßäe hatai biißa ni-a=ne sajhjel
already very old.man be-3ps/D=SUB although
'although they were very old (people) already' [NT: Hebrews 11:12]

Another feature of 'old' and 'young' is that they can co-occur with the
intensifler hatai. Both words can function as a modifier; however, enamanaa
can modify only animate nouns (not restricted to humans), while biißa is also
used with inanimate nouns.

(287) Use of enamanaa, biißa as noun modifiers

a) Human:
katga biißa(-kbtru)
man old(-PL)
'old man/(men)'

b) Animate:
raana enamanaa
w.l.peccary young
'young white-lipped peccary

c) Inanimate:
lureri biißa
house old
'old house'

The word for 'other', leot$a, typically occurs as a nominal head. It is also
attested in modifying position, but different to other modifiers. While most
adjectival modifiers occur in postnominal position, leotQa is only observed
before the head noun. Note that this word can refer to animate or inanimate
nouns.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 197

(288) U s e o f leotQa

a) As nominal head:
leotQa raj=ßa te-ü
other for=FOC:lsg give-lsg/E
Ί have passed it on to someone else.'

b) As modifier:
leotQa itgau-naa ke itgau-ru-a
other live-NOM VLI live-PL-3ps/A
'They live a different life.'

Turning to quality-related candidates for adjectives, again a range of structural

differences can be observed, baaso 'bad' is a noun, which is found as a nominal
head, meaning either 'bad thing' or 'bad person'. But similar to other adjectival
nouns, baaso is different from regular nouns in that it is also used as a
(postnominal) noun modifier. In this function it can modify animate and
inanimate nouns.

(289) Use of baaso

a) Head of NP:
ii=te baaso itga-i
2sg=FOC bad do-2ps
'You did bad things.'

b) Modifier:
katQa baaso
man bad
'bad man'

Related to baaso is the word aaso, which means 'foolish (person)' and refers to
animate nouns only. It functions as a nominal head and can take a noun plural,
as the examples in (290) show. It is not attested as a noun modifier, but this
could be due to its low frequency in use.

(290) Use of aaso

a)
ka=rihiite=ra erele+akane aaso
this=do.like:3ps/E=EMF mosquito+insect foolish
'The crazy mosquito did like that to me!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
198 Word classes

b)
aaso ajßa mausa rihihei ne-nakauru katpa-uru
foolish and blind like be-those.who man-PL
'Ye fools and blind!' [NT: Matthew 23:17]

Another word related to the two above is äasaj 'wicked, evil'. However, it does
not behave in the same way as baaso and aaso, as its syntactic use is different.
At first sight, it seems to be an adverb, as it modifies the (nominalised) verb
'live' in example (291), meaning 'those who live in an evil manner'. In other
examples, it is used as a noun modifier, but in this context, it only occurs with
the component mtQa (or alternatively netQU, its meaning and function is
unclear) to result in netQa-äasaj. This form can occur as a modifier in
prenominal and postnominal position. However, only very few examples are
attested, most of which occur in the Urarina New Testament. The variation
between the three words is not based on any productive derivational pattern. A
further occurrence of äasaj (without net$a) is as a nominal head, which
suggests that it is a noun.

(291) Use of (netpa-)äasaj

a) Adverb:
äasaj itQau-nakauru katQa-uru
evil live-those.who man-PL
'the wicked'

b) Head of NP:
d^ale äasaj ne-ϊ aj-i
something evil be-PRT AUX-NEG:3ps/A
'He is not anything evil, [he is just a priest]!'

c) Noun modifier:
katga netea äasaj / *katQa äasaj
man ? evil
'evil man'

Urarina has two words related to 'small': laauhwiri and aajihjei (which was
investigated in §5.4.1). Similar to seohzva 'big', laauhwiri is noun-like in that it
functions as a postnominal noun modifier without morphological derivation.
However, it is special in that it has two possible forms: under certain conditions,
it occurs as laauißaa. Since this is the only word that shows this kind of
variation, it is difficult to analyse, but, based on several examples, laauißaa
refers to plurality, or more precisely, to individuality or distributive function,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 199

implying ' o n e by o n e ' . This is especially illustrated by the examples (292d),

w h e r e the numeral for ' o n e ' is involved. Note, however, that in (292c), a plural
suffix is attached to laauißaa.

(292) Use of laauhwiri, laaupaa

a) laauhwiri
mi saabere laauhwiri ajßa ku-a hau hetau
that machete small with go-3ps/D because Η RS
'because he was going with that small machete'

b) laauißaa:
ate laauißaa hiririßo laauißaa
fish small type.of.fish small.PL
'small fish, small hiririnyo fish [were sticking out their heads]'

c) laauißaa:
ku kahe hetau ka=arulari laauißaa-kuru ne-ϊ hetau
there from HRS DEM=fox small-PL be-PRT HRS

kahe hetau nalute+naluti-i suru-a

from HRS RED-fall-PRT run-3ps/A
'Being [like] these little foxes, they kept on falling out of there [the tree hole] and ran
away.'

d) laauißaa:
lejhii kampana laauißaa, lejhü tihihemele rihihe-ι ne-naa,
one bell small.PL one pomegranate be.like-PRT be-NOM

leotQa kampana laauißaa

other bell small.PL
first one little bell, then a pomegranate, then another little bell ...' (Old Testament:
Exodus 39:26)

In s o m e elicited examples, speakers contrasted the two f o r m s of ' s m a l l ' for

singular and plural by adding the (optional) plural suffix -uru (surfacing as
allomorph -kuru in (293b)) to laauißaa when it occurred in a plural context. A
possible meaning of (293b) is to imply that various trees out of a group were
felled, i.e. when talking about an entity that is composed of many parts.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
200 Word classes

(293) Use of plural laauißaakuru

a) Singular:
enua laauhzuiri fiui-a
tree small fell-3ps/A
'He felled a small tree.'

b) Plural:
enua laauißaa-kuru fzui-uru-a
tree small.PL-PL fell-PL-3ps/A
'They felled small trees.'

The structure of the ending -ßaa also resembles the plurality suffix used with
stative verbs (cf. §9.2.2); however, the above examples demonstrate that this
form does not exactly mark plural. The lack of further forms that would occur
with -ßaa makes it difficult to make generalisations about its exact function. As
a tentative analysis, it might be characterised as being related to a distributive
plural.
The combination of aafihii and laauhzuiri is also common, as the following
examples show. In these cases, aafihii can be analysed as a nominal head
followed by the modifier laauhzuiri.

(294) Combination of aafihii and laauhzuiri

aafihi-i laauhzuiri ra-a kat^a-ne
be.small- NOM S bj small receive-3ps/D man=CND
'... when a man receives a little' [NT: Luke 16:10]

Another adjectival candidate that occurs as a noun modifier without derivational

morphology is the word (n)atQasiße 'new'. Most typically, it occurs as a noun
modifier. Both variants, with or without initial /n/ are used, with no apparent
distinction. Attempts to distinguish the two forms are inconclusive. Speakers
tried to characterise the distinction by different degrees of age, such as at$asiße
for 'quite new, but already used' and natg,asiße for 'brand new'. However, in a
later example, a one-year-old sewing machine was described with the term
natQasiße. Therefore, I assume that the unmarked form for 'new' is ?iat$asiße.
One aspect that distinguishes this adjective from the other two is that natg,asiße
can also occur as a modifier in prenominal position, as the example in (295b)
illustrates.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 201

(295) Use of natgasipe

a) Postnominal modifier:
tenfiuolo natgasipe itga-a iglesia =ne kujjia
church.building new make-3ps/D church=SUB so.that
'so that the church builds a new church building'

b) Prenominal modifier:
natgasipe itgau-naa
new live-NOM
'new life' [NT: various verses]

As illustrated in (296), natQasipe can also be used as a nominal head (e.g. in

object position) and as a copula complement:

(296) natQasijie as a nominal head

a) Head of NP (object):
natgasipe hvara-kau
new see-lsg/A
Ί have seen something new.'

b) Copula complement:
raj lureri=te natgasipe pe
POSS house=FOC new be:3ps/E
'His house is new.'

The above example clearly qualifies natQasipe as a noun, though with some
adjective-like features.

5.4.3 Multifunctional word "kauatQa "

The words discussed so far could be identified as either noun-like or verb-like.

This is not exactly the case with the word kauatQa for 'good'. While the
attachment of different suffixes clearly distinguishes the use of the different
forms of kauatoa 'be good' (cf. (281)), the distinction is less obvious with
kauatQa. One complication is that kauatQa behaves like a verb stem
morphologically even though this is a morphologically complex form already.
As the examples in (297) suggest, any markers of the usual grammatical
categories are suffixed to kauatQa.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
202 Word classes

(297) kauatQa with verbal function

a) Negation:
kauatga-i kuriae amu-a alau=ne
be.good-NEG:3ps/A far walk-3ps/D spider.monkey=CND
'It is not good that the monkey walks far.'

b) Future:
mi=te kauatQa-re=n satn raj
that=FOC be.good-IRR:3ps/E=ASS all for
'It will be good for all.'

c) Participle:
m-a=lu lejhii katQa kauatga-1
be-3ps/A=REM one man be.good-PRT
'There was (only) one good man.'

On the other hand, exactly this stem is used with other functions without being
modified. In the following examples, it functions as a noun, a (prenominal or
postnominal) modifier, and as an adverb, which makes its status rather
questionable.

(298) Other functions of kauatQa

a) Noun:
d$a kauatQa ni-ji=lu
what good be-NEG:3ps/A=REM
'There were no good things.'

b) Prenominal modifier:
ßaaohwa ahaana-kanei kau situ-e kaa kauatQa ate kosemane
how.much year-ago here pass-3ps/E this good tapir
'How many tears ago did this great tapir pass here?'

c) Adverb:
kauatQa hakuturu-a hau
good make.dry.out-3ps/D because
'as he made [the sea] dry out well ...'

Even if we assumed that kauatQa is a noun, one would have to explain its use as
a verb stem in the previous examples. Additional evidence against treating
kauatQa as a noun is found by examples where the participle marker -7 is
attached to kauatQa - which clearly makes it look like a verb. The only
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 203

conclusion that can be drawn from the peculiar behaviour of kauatpa is that this
word is exceptional with respect to its assignment to any of the word classes
and that it must be regarded as "multifunctional".

5.4.4 Summary of word class assignment for adjective-like words

In the discussion above I have illustrated the great diversity of Urarina

"adjectival candidates". As it turns out, these cannot be defined as a
homogeneous, independent word class, as they differ significantly from each
other by means of function, position, and morphology. In order to give an
overview of this, the tables in (299) and (300) categorise these words according
to various tests.
In (299), first, the crucial features that qualify a noun are contrasted with the
"adjectival nouns", as one may call them. The distinction is based on four
typical features that can be related to nouns, as was discussed at the beginning
of §5.4, plus feature no. 6 - the option of adding the intensifier hatai.

(299) Adjectival nouns32

1) Functions as nominal head/occurs headlessly
2) Functions as noun modifier (pre-/postnominal)
3) Functions as copula complement
4) Can take noun plural -uru
6) Can be modified by intensifier hatai

1 (head)2 (modif.) 3 (CC) 4N-PL 6 Intens.

Typical noun yes rare, only yes yes no
specific Ν
? leotga 'other' yes pre yes yes no
A enamanaa 'young' yes pre yes yes yes
A biißa 'old' yes post yes yes yes
A nat^asiße 'new' yes post, pre yes yes yes
V/H (netQa)äasaj 'evil' yes post, pre yes yes yes
V/H baaso 'bad' yes post yes yes yes
V/H aaso 'foolish' yes post yes yes yes
D seohwä 'big' no post rare yes yes
D laauhwiri 'small' no post rare no yes
A = Age // D = Dimension // V = value // Η = Human propensity

32
The criteria are numbered according to the numbering in the introduction to 5.4.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
204 Word classes

Table (299) shows that the syntactic and morphological structure of noun-like
"adjectival candidates" is highly heterogeneous. All three terms related to the
semantic concept of age fall into this category, but their syntactic behaviour
differs between the respective words, especially with regard to the position in
modifying function. Three words that refer to value or human propensity, such
as 'bad' and related concepts differ from each other in a similar way. However,
age-related as well as value-related terms can all be characterised as nouns, as
they follow the two main properties that are embodied by Urarina nouns, i.e.
they can function as nominal heads and as such also occur as copula
complements. They also form the same plural as nouns, though this is unclear
for two of the words. A property that distinguishes these words from
prototypical nouns is that adjective-like nouns can be accompanied by the
intensifier hatäi, while this is not normally observed with "regular" nouns.
The two terms relating to dimension listed in (299) are different from the
others in that they do not occur as nominal heads. Their use as copula
complement is possible, but only attested from elicited examples. Their function
as noun modifiers is restricted to postnominal position. Yet, they cannot be said
to form a - admittedly very small - class of adjectives because they differ from
each other with regard to plural formation. Thus, they must be described as
exceptional cases of adjectives in a language that does not have a class of
underived adjectives. However, what all of the words listed in (299) have in
common is that they are in fact different from prototypical nouns in that they
can function as noun modifiers, which is a property otherwise represented only
by a very limited group of nouns such as names and nominalisations (cf.
§5.1.3). This is a feature they share with verb-like adjectival candidates. Apart
form one exception, they only occur as noun modifiers after being derived by
the nominalising suffix -i, which of course is a feature inherent to any stative
verb. The words listed in table (300) shows a selection of typical verb-like
adjectives, which represent the semantic concepts of dimension, colour, and
value.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 205

(300) Adjectival verbs

2) Functions as noun modifier (pre-/postnominal)

5) Can take verbal suffixes
6) Can be modified by intensifier hatai

2) N-modifier 5) V-suffix 6) Intens.

Other stative verbs DER with NOM yes yes
D taba-a 'be big' DER with NOM yes yes
D aajihi-a 'be small' DER with NOM yes yes
D tasißohioaa 'be big' DER with NOM yes yes
C lanaa-ka 'be red' DER with NOM yes yes
(all colours)
V kauato-a 'be good' DER with NOM yes yes
V kauatQa '(be) good' yes (without DER) yes yes
D = Dimension H C - Colour // V = value

A s discussed above (298), kauatpa '(be) g o o d ' represents an exceptional case,

but it is probably acceptable to characterise it as a verb as well, since it takes
inflectional suffixes. In conclusion, it must be admitted that "verb-like
adjectives", or "adjectival verbs" do not represent a separate class but are a
subgroup of stative verbs that happen to coincide with the semantic concepts
typically represented by adjectives in many other languages.
A number of further words can function as noun modifiers, mainly through
derivation of a stative verb with the nominaliser -i. Examples for this are
maueki ( ' s h o r t ' ) from maueba ( ' b e short'), tQaauhwaj ( ' u g l y ' ) from
tQaauhwa-a ( ' b e ugly'), and elonaj ('bitter') from elona-a ( ' b e bitter'). These
occur as postnominal modifiers and can also function as nominal heads (e.g. 'a
short one'). Another form of derivation is the subtraction of verbal suffixes,
which only leaves the root and which in turn can function as a modifier. This is
exemplified by setu ('rotten') derived from the intransitive verb setu-a ( ' r o t ' )
and ii ( ' r i p e ' ) from ii-ka ('be ripe'; also ii-to-a). The modifier uhuari 'unripe' is
derived from nuhuarim-a ( ' b e unripe') and only occurs as a noun head as
nuhuari ('something unripe'). The word kiii for ' d e a d ' is obviously related to
the verb u-a ( ' d i e ' ; also cf. su-a 'kill'). It can function as a copula complement
but is not acceptable as a nominal head when preceded by a numeral. As a noun
modifier, it occurs after the head. Speaker judgements on the word seri ( ' o l d ' ;
origin or derivation unclear) differ with respect to its status as a noun head. In
elicitation, some find it acceptable as a head, others not, whereas it is always
accepted as a postnominal modifier.
In summary, noun-like "adjectives" resemble nouns in that all of them can
be copula complements and they can function as nominal heads (except for one,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
206 Word classes

which is derived). Another feature they share with nouns is that the noun plural
-uru can be attached to most of these. The main difference to other nouns lies in
their behaviour as noun modifiers. While a noun only rarely modifies another
noun, the words under discussion do so. The other important difference to
nouns is that these are not accompanied by the intensifier, while the adjective-
like nouns have this option.
All verb-like "adjectives" are derived; their behaviour is quite similar to that
of Stative verbs. Again, the main difference lies in their function as a modifier.

5.4.5 Sequences of noun modifiers

It is possible to combine several modifiers in a phrase, but this seems to be quite

rare in natural speech. All examples for such sequences are taken from the NT.
In addition, no particular pattern can be recognised that would postulate a
prescribed order of the different modifiers. The examples given below each
contain two adjectives, which are distributed around the noun in prenominal and
postnominal position, or both occur after the noun. It must also be noted that the
attested examples all involve deverbal adjectives, without any involvement of
adjectival nouns, which are not attested as sequences of modifiers. The first
example in (301) shows a sequence of stative verbs nominalised with /-i/. The
literal translation 'blackish red' is meant to describe a dark purple colour. The
second example illustrates the prenominal use of the nominalised form of 'be
white', followed by the noun, which in turn is followed by kanatga in
modifying function.

(301) Modifier sequence with deverbal modifiers through nominaliser -i

a)
nii kahjune lanaha-j hitQiia-hiohiua-j
that cloth be.red-NOMsbj be.black-DER-NOM S b j
'that purple cloth' [NT: Mark 15:20]

b)
somaha-j kahjune kauatga hikiaki-i
be.white cloth good shine-NOM Sb j
'a nice, white, shining cloth' [NT: Revelation 19:8]

More frequent than a sequence of nominalised verbs is the chaining of verbs that
carry a participle suffix, i.e. these function as adverbials. This is not surprising, as
sequences of several participle constructions (representing clauses with different
types of transitivity) are very common in Urarina (cf. §20.3). In a similar way,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 207

nominalised modifiers co-occur with participle forms, as example (302b) illustrates.

The order of nominalised verbs followed by a participle is not systematic, as other a
different order than the one observed below can occur with other examples.

(302) Modifier sequence with deverbal modifiers and participle -7

a)
eku m-a lejhii atajbijiaae tasipohwa-~i lanaha-1
there be-3ps/A one animal be.big-PRT be. red-Ρ RT
'There was one big, red dragon.' [NT: Revelation 17:3]33

b)
lejhii atajbipaae dragon lanaha-j taslpohwa-l
one animal dragon be.red-NOMSbj be.big-PRT
'a great red dragon,...' [NT: Revelation 12:3]

Since sequences of modifiers do not seem to be common in the language, one

would not expect specific strategies for their ordering to be employed in the
translation of given sentences.

5.4.6 Comparison

As mentioned in the beginning of §5.4, Urarina does not have a morphological

comparative. However, there are a number of alternative strategies that allow a
comparison of qualities. Comparisons that imply superiority or inferiority
scarcely occur in the text database; thus, the majority of the examples given for
these is elicited. Comparisons that refer to equality are much more frequent,
which might imply that this concept is more significant than other types of
comparison, in terms of cultural values. Any kind of comparison is made
through derived forms of verbs such as 'exceed', 'be less', or 'be like'. The
table in (303) given an overview of the various comparative strategies. This
includes the reference to the type of parameter, i.e. the quality or content with
respect to which a comparison is drawn. While it is common from a cross-
linguistic perspective that the parameter is expressed by words referring to
quality (e.g. adjectives), this is not the case in Urarina. In most cases, the
parameter of comparison is expressed by a clause.

33
The words eku and atajbipaae are Chambira dialect variants of the Espejo
forms ku and tajbipaee.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
208 Word classes

(303) Comparative strategies in summary

Type Function Parameter Index

Biclausal Superiority Trans it i ve/i ntran s i t i ve amurito-eri-7 (PRT)
'more' clause 'exceeding'
Biclausal Superiority Transitive/intransitive amuri-hi-1 (PRT)
'more' clause 'exceeding'
Biclausal Inferiority Intransitive clause fiuauta-eri-l (PRT)
'less' 'being less'
Biclausal Equality 'as' None rihihel 'being like'
Biclausal Equality 'as' None tokuanei 'compared to'
Monoclausal 'as' Fuses with index Derivation /-ni/ 'be like'
Monoclausal 'as' Adjectival noun Reciprocal marker ita
Monoclausal 'as' None rihianaa, tokuania
'be like'

For comparison with 'more' and 'less', a biclausal construction as shown in

(304) is applied. Literally the sentence can be translated 'Jorge is big, exceeding
Manuel'. The word amurihu used in this construction is the continuous
participle form of the root /amari/ (not occurring in underived form) from
which the transitive verb amuritoa 'exceed', 'surpass' is formed. The two
sentences show alternative word orders, with (304b) involving a focus marker.

(304) Comparison with amuritoa

a) Dependent clause following the main clause:

tasijiohzoa-a horhe manoel amurihii
be.big-3ps/A PSN PSN exceeding
'Jorge is bigger than Manuel.' [Lit. 'Jorge is big, exceeding Manuel']

b) Dependent clause in initial position:

manoel amurihii=te tasißohiva-e horhe
PSN exceeding=FOC be.big-3ps/E PSN
'Jorge is bigger than Manuel.'

With comparisons involving a verb that semantically implies inferiority, the

word fiuautaeru is used, which is derived from the transitive verb fzoauta-a 'be
less'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 209

(305) Comparison with fiuautaeria

aafiht-a manoel horhe fiuautaeni

be.small-3ps/A PSN PSN being.less
'Manuel is smaller than Jorge.' [Lit. 'Manuel is small, being less than Jorge.']

N o t e v e r y s e m a n t i c concept is expressed by single stative v e r b s such as in the

e x a m p l e s cited above. W h e n the standard of c o m p a r i s o n is a n o u n , a c o p u l a
clause is f o r m e d w h i c h then o c c u r s in the s a m e position and f u n c t i o n as
illustrated in (304) and (305). In the e x a m p l e s in (306), t h e p a r a m e t e r is
represented by an entire clause, which can involve a c o p u l a with c o p u l a
c o m p l e m e n t (306a), a transitive clause (306b), or a c o m p l e m e n t clause, such as
in ( 3 0 6 c ) . N o t e that the c o p u l a in ( 3 0 6 a ) o c c u r s in a derived f o r m , involving the
s u f f i x -to, w h i c h generally implies intensity or other kinds o f e m p h a s i s on the
v e r b (cf. §10.2.3). H o w e v e r , this is not an obligatory s u f f i x in this construction,
but m e r e l y e m p h a s i s e s the fact that M a n u e l really is old. T h e clause that refers
to the p a r a m e t e r is m a r k e d with square brackets.

(306) Comparison involving a clause

a) With copula clause:

biifla ne-to-a horhe manoel amurihn
[old.man be-INTS-3ps/A] PSN PSN exceeding
'Jorge is older than Manuel.'

b) With transitive clause:

akauru amurihirn hatai baaso itQa-1 itQau-i-toe=ne sajhjel
[3pl exceeding] very bad do-PRT live-2ps-PL=SUB although
'Even though you are doing much worse than they ...' [NT: Matthew 7:3] [Lit. 'though
you live doing very bad (things), exceeding them']

c) With complement clause:

itulere eene-kuru raj kauatQa it$a-a=ne amurihirn
[all.kinds woman-PL for good do-3ps/D=SUB] exceeding

i=tgej kauatga itga-e

2sg=for good do-3ps/E
'He did more good to you than to other women.' [NT: Luke 1:42] [Lit.: 'exceeding what
he did good to all other women, he did good to you']

N o t e that there are several variants o f derivation f r o m the verb f o r ' e x c e e d ' . T h e
d i f f e r e n c e b e t w e e n the different kinds o f derivation is uncertain since m o s t o f

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
210 Word classes

these are non-productive and require further study. One variant is amuritoeru as
illustrated in (307).

(307) Comparison with amuritoeru

hatal nii ißono amuritoeru itulere he kanii-hjel ne=lu

very that ayahuasca exceeding all.kinds VLI [sort.of-PRT be:3ps/E=REM]
'Exceeding that ayahuasca very much, it [the other drug] was like that (over all others).'
[Talking about the strength of ayahuasca]

The equivalent to a superlative construction is expressed with one of the two

forms mentioned above, with the nominal quantifier for 'all' (or 'all kinds') as
an object. Literally, the example in (308) means 'Manuel is small, being less
than all (others)'.

(308) "Superlative"

itulere fiuautaerü äafihi manoel

all.kinds being.less be.small:3ps/E PSN
'Manuel is smaller than anyone.'

Urarina employs several strategies to express comparison that involves equality.

One is of a morphological nature and involves the derivational suffix -ra, which
expresses "as ... as" and occurs with a limited number of verbs. It always forms
a transitive verb from a stative intransitive base verb. Examples for this are
given in (309).

(309) Comparative derivational suffix -ni

a)
tabaae-ka horhe tabaae-ni-a manoel
be.big:3ps/A PSN be.big-CMP-3ps/A PSN
'Manuel is as big as Jorge.'

b)
anaito-a aheri anai-ni-a raj maleta
be.heavy-3ps/A stone be.heavy-CMP-3ps/A POSS suitcase
'His suitcase is as heavy as stone.'

However, since the suffix is not especially productive, it is attested with a few
more stative verbs only (cf. (310a)). It is not compatible with verbs such as the
ones shown in (310b). This also excludes active intransitive verbs.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Adjectives 211

(310) Further verbs with/without derivational -ni

a) Possible:
ahaa aha-ni-a 'be as hot as'
kalarna kalano-ni-a 'be as dirty as'
ahiflaa ahißa-ni-a 'be as sharp as'
somahaa somara-ni-a 'be as white as'
raautoa raauto-ni-a 'be as tasty as'

b) Impossible:
äatQihia 'be small'
ateroa 'be tired'
tasißohwaa 'be big'
lanaaka 'be red'
(räasaa) 'dance' (active intransitive)

Another strategy to compare two nouns of equal status is a construction that

involves a reciprocal form. However, the example in (311), which is the only
one of this kind is peculiar: it translates literally as "Manuel and Jorge, they are
old (men) each other". The unusual feature about this sentence is that the
subject ('Manuel and Jorge') is divided and only 'Manuel' remains in final
(subject) position, preceded by the copula in the 3pl form.

(311) Comparison using the reciprocal form

horhe ajßa=te ita biißa ne-ure manoel

PSN with=FOC REC [old be-3pl/E] PSN
'Manuel is as old as Jorge.' / 'Manuel and Jorge are of the same age.'

A more productive way to compare two participants of equal status is with one
of several comparative verbs that express 'be like'. These include tohiama,
rihiitga (or plural rihianaa), rihitoa, and rihitohoaka. While there might be a
suspicion that these are a kind of copula (due to their semantics, which involves
identity to a certain degree), they are transitive verbs: all of these verbs take a
noun object, but cannot occur with an adjective such as nat$asijie ( ' n e w ' ; cf.
(299)), whereas the copula nia does. The difference in meaning between these
verbs is based on different kinds of perception, as is summarised in (312). The
verb tokoania is not strictly limited to any specific perceptual level and may be
characterised as a general comparative verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
212 Word classes

(312) Comparative verbs

a) tokoama - General comparison 'be like'

b) rihiitqa - Visual comparison 'look similar to' (e.g. person, animal)
c) rihitoa - Comparison of voice or behaviour 'sound/behave similar to' (e.g. bird)
d) rihitokoaka- Comparison of other senses, e.g. taste, smell, effect (e.g. venom)

Most typically, these verbs occur in the participle form, where they have
adverbial function (cf. (314)). T w o examples for their use as finite verbs are
given in (313).

(313) Different types of compari son

a) With tokuania:
raana tokuani-ji aräala
w.l. peccary compare-NEG:3ps/A tapir
'Tapir is not like white-lipped peccary.'

b) With rihitoa:
läe esißae baka rihito-a=m hau
just really cow do.like-3ps/D=SUB because
'because they were actually doing it like cows' [about reproductive act of tapirs]

In a similar way as amurihirii and the other verb forms that express for
superiority or inferiority, the comparative verbs listed above are used in the
participle form, sometimes preceded by a subordinate marker. Since the
combination with the participle marker does not occur when they are inflected
for person, one might argue that tokuanei and rihihei are lexicalised and
function as conjunctions (cf. §5.8). Literally, sentences such as (314a) can be
translated as 'Compared to (what) is now, there were no guns'.

(314) Participle forms tokuanei and rihihei

a) With tokuanei:
poaelu ena aj-a-ne tokuanei itgafiua ni-ji=lu
earlier now AUX-3ps/D=SUB compared.to rifle be-NEG:3ps/A=REM
'Earlier there were no guns as there are now.'

b) With rihihei:
aka heri-a=ne rihihei ne-uru-a katQa-uru=ne kujßa
3sg want-3ps/D=SUB like be-PL-3ps/D man-PL=SUB so.that
'... so that the people would be as he wanted.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Pronouns 213

Frequently, these forms occur with a demonstrative such as in kaa toknanei

('compared to this') or niirihihei ('like that'), which can also be cliticised to
ka=tohianei and ni=rihei (shortened). It may also be noted that there is a
transitive verb nitoania 'be like that', which is composed of the demonstrative
m= and the root /toane/ (not occurring by itself). Its most common occurrence
is in the participle form nitoanei to mean 'like that'.
In summary, Urarina has a number of different strategies that express
comparison. However, different from many other languages, the comparison
focuses on equality rather than indicating that someone or something is of
higher or lower rank than others. The absence of adjectives that would function
as the parameter of comparison results in the parameter being expressed as a
clause in most of the strategies.

5.5 Pronouns

Urarina marks person with bound pronouns on the verb. In addition to these,
there is a set of free pronouns which can be added to indicate emphasis. For
instance, the free pronoun is used as a focus in reply to a question ('Who killed
the monkey?' - Ί killed the monkey'). Apart from the common distinction
between 1st, 2nd, and 3rd person for both singular and plural, Urarina
differentiates between inclusive and exclusive in the 1st person plural and dual.
As Urarina does not have morphological case, personal pronouns do not differ
with respect to their function as subject or object of a verb. While the
pronominal system distinguishes between person and number, animacy does not
play a role. The pronouns for 3ps can refer to animate or inanimate nouns.

(315) Free pronouns

lsg kanu
2sg ii
3sg aka
lpl/in kana
1 pl/ex kanakaana
2P1 ßaara
3pl akaoru

Note that the dual form of the subject is marked on the verb, but there is no
separate free pronoun for it. Instead, the lpl inclusive pronoun is used in this
context. This is illustrated in (316), which is from a context where the two
protagonists appear as the only people in the world. While this implies a "dual"
interpretation of the lpl pronoun, kana (lpl inclusive) is used.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
214 Word classes

(316) Use of 1 pl/in pronoun kana in dual context

d$asn=te katoanei kanaanaj hanoa-he-i kana raj=ta

how.possibly=FOC like.this child give.birth-CNT-2ps 1 pl/in for=FRS
'How is it possible that you are giving birth to a child for us?'

The pronoun for lsg has a variant kiitQa that occurs in nominaliser
constructions with the suffix -i (cf. §6.6.1). In this construction, where the noun
derived from a transitive verb represents the relativised object of the clause, the
subject typically is marked with the suffix -at$a 'only', which results in
kanu-atga when it occurs with the lsg pronoun. Alternatively, the two
components can fuse and are realised as kiitQa, with no apparent semantic
difference. All other pronouns are regularly joined with -at$a in this context
without any alternation.

(317) 1 sg pronoun alternant kiitga

ii raj kiitQa te-j anofxua presta-u

2sg for lsg.EMPH give-NOM 0bj knife lend-IMP
'Lend me the knife that I gave to you!'

In addition to the free pronouns, Urarina has proclitics for the singular, which
are derived from the full forms (except for 3sg). These can be attached to nouns
and verbs. With verbs, they function as object proclitics; with nouns, they refer
to possessive pronouns (cf. §7.1.3). However, procliticisation mainly functions
as a feature of the traditional language; younger people prefer the use of free
pronouns rather than procliticisation. This applies to both personal and
possessive pronouns.

(318) Proclitics

lsg ka=
2sg i=
3sg n=, (ne=, ni=)

In particular, the use of the 3sg proclitic has been reduced even among older
speakers. While ka= and i= (or allomorph d$= for 2sg before vowel-initial
roots) are still common practice, the proclitic 11= is only observed with vowel-
initial roots. Before consonant-initial roots (319b), the use of the 3sg allomorphs
m = or m= is very rare: the two examples with these 3ps forms listed below
were rejected as ungrammatical even by an older speaker, even though these
forms occurred in a text by another speaker of the traditional variety of Urarina.
It is also obvious that the use of possessive proclitics is related to the matter of

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Pronouns 215

inalienable possession; this is discussed in §7.4. S o m e examples for the use of

proclitics in possessive function are illustrated below:

(319) Use of proclitics in possessive function

a) lsg and 2sg forms:

ka=hitQana 'my blowgun'
ka=ripaha 'my grandfather'
i=hitg,ana 'your blowgun'
i=tQuerehe-uru 'your children'

b) 3sg forms:
n=asi 'his nose'
n=uhoana 'his field'
?m=hitQana 'his blowgun'
?ne=batohwi 'their arrows'

For proclitics that occur with object function, only the f o r m s for l s g and 2sg are
attested, while the 3sg form is not observed. T h e reason for its absence as an
object proclitic could be that 3 rd person object pronouns do not surface in
general, but are implied through omission. While it is not ungrammatical to use
them as arguments of a transitive verb, they almost never occur in this context.
T h e m e a n i n g of ' h i m ' or ' t h e m ' is mutually understood f r o m the context of an
utterance. However, all three proclitics occur in combination with postpositions,
as shown in (320c-e).

(320) Object proclitics

a) lsg object proclitic with verb:

ka=kuruatahani-a kana+hoaaun-era=ne eseneta-anu
lsg=help-3ps/D our.creator=SUB believe=lsg/A
Ί believe that God helps me.'

b) 2sg object proclitic with verb:

i=tQuara-ri-a-u=ni=ta
2sg=see-IRR-NEG-1 sg/E=ASS=FRS
Ί will not watch you.'

c) lsg object proclitic with postposition ke:

ka-ke taka-i
lsg=INST encounter-PRT
'meeting with me'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
216 Word classes

d) 2sg object proclitic with postposition raj (assimilated to [t<?] here):

i=tcej
2sg=for
'for you'

e) 3ps object proclitic with postpositions tgoae, asae:

m=tQoae n=asae
3sg=on 3sg=in/under
'on top of it' 'under it'

The prefix found on intransitivised verbs, ne- (and variants), which may also
have reflexive function to some degree, possibly is a remnant of the 3sg object
proclitic, as the reflexive subject in this construction is identical to the object
argument. However, this prefix exhibits an allomorphy different from the one
found with proclitics (cf. §16.2).

5.6 Demonstratives

The Urarina deictic system distinguishes three degrees of spatial reference, each
of them involving the position of the participants:
kaa 'this' (close to the speaker)
mi 'that' (close to the listener)
taa 'that over there' (distant from both participants)

Regarding this type of distinction, one could also relate the three demonstratives
with first, second, and third person, respectively, kaa even has some degree of
phonological similarity, if compared with kanu ( T ) . The similarity between nii
and ii ( ' y o u ' ) is more subtle (different onset), taa and aka ('he/she/it') have
even less in common. However, the use of the demonstratives suggests that kaa
refers to something with regard to first person, nii to second person and so on.
Evidence for this is not only provided by the phonological similarities, but also
through the deictic function of the demonstratives. The correlation between first
person and demonstrative kaa is manifested by the fact that kaa relates to an
item in the space close to the speaker; similarly nii refers to something close to
the listener, implying a relation to second person. Correspondingly, taa refers to
something not close to either the speaker or the listener - thus, implying a third
person.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Demonstratives 217

5.6.1 Function of demonstratives

In combination with a noun, demonstratives occur in prenominal position such

as in kaalureri 'this house' or niisaabere 'that machete'. Apart from that,
demonstratives can also be used as nominal heads and as such, they can take a
plural form (which is identical to the noun plural -urn). However, they do not
take other modifiers in this function. The plural form does not imply any
distinction in animacy, as the examples in (321) illustrate. Furthermore, their
use in combination with conjunctions is very common, such as in mi hau
'therefore' (Lit. 'because of that'), niibaja ('afterwards, lit. 'after that'), or
niihana 'then' (Lit. 'in that (time)'). In this context, demonstratives are again
understood as a nominal head.

(321) Demonstrative as nominal head

a)
nii=te noane-naa ne
that=FOC lie-NOM be:3ps/E
'That is a lie.'

b)
nii-tQuru hjane-u=te kau-a ku-uru-ijie
that-PL leave-IMP=FOC return-NTR go-PL-JUSS
'Let them, so that they go home!'

In discourse, the spatial distinction between the three is not very strict and nii is
used much more frequently than kaa. The reason for this is that rtii is used under
many circumstances that are not exactly "close to the listener". In elicited
translation, when the context is not completely clear, nii is chosen as a kind of
default demonstrative. In narratives, nii has a discourse function, typically
referring to a noun that had already been introduced. In this case, the use of nii
does not imply that the described object is either close to the speaker or to the
listener. The example given in (322) illustrates this: in the first sentence, the
participants ( ' w o m a n ' and 'her daughter') are introduced. While sentence 2
contains 'her daughter's living place' as head of NP, sentence 3 does not
contain a noun that would refer to either participant. In sentence 4 then, 'her
daughter' is picked up again and this time, the NP is preceded by nii as an
anaphoric referral to a known participant (also cf. §22.4.2).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
218 Word classes

(322) Discourse function of nii

a) Sentence 1 :
lejhii eene=te su-e enua, raj kakunu rem bedajni-a ku hana
one woman=FOC kill-3ps/E tree POSS daughterplace visit-3ps/Dgo when
Ά certain woman was killed by a tree when she went to visit her daughter.'

b) Sentence 2:
rautohwe-i=te niki raj kakunu ne-naha turu-e
calm-PRT=FOC ADVRS POSS daughter be-LPP arrive
'She safely arrived at her daughter's living place.'

[c) Sentence 3: 'Arriving, [she] said: "I have come to see you.'"]

d) Sentence 4:
nii hau=te raj lureri asae ni-a=ne here
that because=FOC POSS house under be-3ps/D=SUB want:3ps/E

mi raj kakunu
that POSS daughter
'Therefore, ["that"] her daughter wanted her to stay in her house.'

It should be noted that the insertion of the demonstrative in the described

context is not entirely obligatory. Later in the same story, 'her daughter' is
mentioned again without the demonstrative.
Another use of mi is as the standard demonstrative in expressions like 'be
like that' (nitoania), 'like that' (niirihihei, cf. (314)), 'then' (nii hana which is
composed of nii plus a temporal conjunction), or 'therefore' (nii hau, literally
'because of that'). Demonstratives in modifier function typically occur in
phrase-initial position, but deviations from this are possible (cf. §6.1).

(323) Demonstrative in modifier function

a) With kaa\
kaa kana raj d^aura+atane
this lpl/in POSS flesh+land
'this our world'

b) With nii:
nii lejhii katQa
that one man
'that one man'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Demonstratives 219

In addition, the demonstratives also occur with proper nouns, such as names of
persons or locations, and with adverbs.

(324) Demonstrative with proper nouns and adverbs

a) With place name:

kaa ajtQune kuane
this PLN inside
'in this river, the Espejo'

b) With adverb:
kaa ene netohwel
this now until
'until ("this") now'

There also is a compound composed of kaa + nii, which means 'something like
this/that'. Apart from being employed as a hesitation marker , it can be used as a
head of N P and as a verb stem (cf. (440)). Note that the demonstrative pronouns
can be shortened to ka=, ni=, and ta= when they function as a modifier (cf.
§2.11).

5.6.2 Locational demonstratives

Each of the demonstratives can be combined with a number of elements that

form locative or directional expressions. The parameters involved in their
distinction involve indicating/pointing, which is further distinguished into items
that are in sight or not. Further parameters refer to punctuality, distance, and
location versus direction. A scheme that summarises these distinctions is found
at the end of this section.
The first distinctive parameter is indicating. There are three locational
demonstratives that exclude the possibility of supportive pointing with the hand,
finger, or head. These are ht (cf. (330)), nai, and kttaae (cf. (331)). All other
demonstratives allow physical indicating (even though this is not required if
clear from the context). These are further distinguished with respect to
visibility: one set of locational demonstratives refers to an object, person, or
area that is not in sight from the point of indicating, e.g. a field that lies in a
certain direction but cannot be seen from the location of the speaker. These
involve the combination of the demonstratives kaa or taa with the suffix -hia,
resulting in kahia 'over here' and tahia 'over there'. They can be used in a
locational or directional sense, as suggested by the examples in (325). For
instance, the example in (325d) involves tahia with the meaning of 'from over

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
220 Word classes

there'. In (325a), kahia is preceded by another occurrence o f kaa, which

indicates that kahia may have the function o f a noun 'this area' which is then
specified further by 'this'. A corresponding form for mi with -hia does not exist.
Instead, the form nii-tQae is used with the same function for second person.

(325) Use of -hia

a) Locational use of kahia:

hjauißera, huatia kaa kahia ka=atane=ke hjajni-a-akaanu
no not this over.here this-land=VLI urinate-NEG-lpl/ex
'No, we do not urinate here in this world.'

b) Locational use of tahia:

tahia aj-ni-u
over.there AUX-DSTL-IMP
'Do it over there!'

c) Directional use of kahia:

kahia u-a-u
over.here come-C AU 1 -IMP
'bring it here!'

d) Directional use of tahia:

ßäe hvajtel tahia turu-a u-e nii kati
already again over.there arrive-NTR come-3ps/E that black.monkey
'Again the black monkey arrived from over there.'

In combination with net (the participle form o f the copula), the form
ka-hia-ne-l Cthis'-LOC-'be'-PRT) is used to mean 'everywhere'. The
distinction for speaker reference is neutralised with this word.

(326) Use of kahianei

a)
kahianei=te m atanihi
everywhere=FOC be:3ps/E mosquito
'There are mosquitoes everywhere'

b)
kahianei=te ama-ϊ hoitQa-rate-kure
everywhere=FOC take-PRT heal-CAU2-3pl/E
'They took him everywhere in order to have him healed.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Demonstratives 221

Note that there is no corresponding form for *tahia-nei or *nihia-nei. However,

the word tahia-mhia-nel is used to mean 'in various places over there'.
For the locational demonstratives that involve a visible point of reference,
there are further distinctive features. The first of these parameters is punctuality,
i.e. it distinguishes between points of reference that are in a precisely definable
location, and other points, for which the exact location cannot be determined.
The set of punctual locational demonstratives consists of combinations of kaa,
mi, and taa with the locative marker -u. Recall that this suffix also occurs with
nouns referring to place (cf. §5.1.2). The demonstratives surface as ka-u, m-u,
and ta-u, respectively.

(327) Locative marker -u with demonstratives

a)
kau=na neda-na here-u
here=FOC: 1 sg stay-INF want-lsg/E
Ί want to stay here.'

b)
aj, niu te-u-toe
do there give-IMP-PL
'Yes, (just) put it there!'

The non-punctual demonstratives are further distinguished into locational and

directional ones. The locational suffix -laj (pronounced [lej]) can be attached to
any of the three demonstrative bases, but is attested in a few examples only. The
difference in meaning as compared to -u is that -laj implies a larger, less
specified area rather than an exact point of location. It could also be translated
with 'in the area o f .

(328) Use of -laj with demonstratives

a) With kaa\ ;
kaalaj raj kuraanaa ne-it
this.area for chief be-lsg/E
Ί am the chief for this area.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
222 Word classes

b) With mi:
niilaj kahe amu-u-tge ka=kioeraj-uru=ne hauria
that.area from walk-IMP-PL lsg=visitor=FOC: lsg first

ne-kabakabaaha-ni=t$a
ITR-bathe-DSTL.HORT=EMF
'Go over there, my visitors, since I first have to go take a bathe.'

The other non-punctual set is a combination of the demonstratives with the

direction marker helöo, which represents a directional 'towards', 'in direction
o f . While the word surfaces as helöo with nouns, it is attached to
demonstratives as -elöo to result in kaelöo, niilöo, and taelöon, respectively.

(329) Directional suffix -elöo

a)
ka-eloo raj bihi lera te-1
this-DIR POSS hand toy give-PRT
'attaching toys to (the area of) his hands'

b)
lejhü-tQa ta-elöo helu+helu-ki-Ί beru herahauito-a ktt-e
one-only that-DIR RED+jump-DER-PRT way cross-NTR go-3ps/E
One [peccary] has crossed the way running towards that direction over there.'

As mentioned above, there are three locational demonstratives that exclude

physical indicating. These do not occur as a set referring to the person marking
system, but are distinguished for punctuality in the first place. More exactly, ku,
which is only used in narratives, refers to a specific, definable location in the
past. The parameter of visibility is neutralised for h t and the actual distance to
speaker or listener is not relevant. The use of hi neither indicates that the
speaker was present on the spot at that time, nor implies that the action took
place in the location where the story was told. In some cases, ku is also used as
a discourse filler particle and may be used at various times in a sentence (cf.
§22.5).

(330) Use of it«

a)
ku ne-rehete pasto taba-j
there be-HABl:3ps/E field be.big-NOMSbj
'There used to be a big field here'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Demonstratives 223

b)
ku ajßa räasa-ure
there with dance-3pl/E
'They danced with her there'

The non-indicating, non-punctual locational demonstratives are further

distinguished for distance, näi refers to a distant place whose exact location
cannot be described. The non-distant equivalent kuaae refers to a closer place.
Note that it is realised as ukuaae in the Chambira dialect. Both words can be
used in a locational or directional sense. The locational character of näi is
illustrated in (331a), the directional sense in (331b). In (331c), the non-distant
character of kuaae (here occurring as ukuaae) is supported by the preceding
sentence, which involves a return from a distant to a nearby location.

(331) Use of näi and kuaae

a)
näi kanu atane ajriu kuane lejhü kurahe rajmöoro ne
there lsg land PLN inside one name-REP PSN be:3ps/E
'Over there in my place, the Airico, there was a man named Raimondo.'

b)
helaja näi ukiuana hana ku-e lejhü katQa
same.place there field inside go-3ps/E one man
Ά man went right there to his field (which was far away)'

c)
[Preceding: '... he returned from the Jordan ...']
ukuaae ni-a heena+tQurjka hanone-l=ße hana
there be-3ps/D four+ten-PRT day-PRT=SUB when
'when he was there for 40 days ...' [NT: Luke 4:2]

In summary, demonstratives behave like nouns in many different ways: they

can be used as nominal heads, they can take a plural form and they can be used
with locational and directional markers that are also used with nouns. However,
the use of these markers differ from nouns in that they can be joined with
demonstratives morphologically - a property that does not apply to full nouns.
One can conclude that demonstratives, similar to pronouns, are a subgroup of
nouns. The scheme in (332) summarises the functions of the various
demonstratives again.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
224 Word classes

(332) Locational demonstratives

Indicating Non-indicatinc

In siuht Not in sight Punctual Non-punctual

1 Lihui A-H
2 IlitQJC

3 Initio

Punctual Non-punctual Distant Non-

1 Lih distant
2 tiia mi knaac
3 tcHi

Locational Directional
1 k/ialnj 1 k-mloo
2 niilaj 2 niildo
3 taalaj 3 hiclöo

5.7 Postpositions

All adpositions of Urarina occur in postnominal position. The majority of

postpositions are locational. They differ from locational nouns in that they
cannot function as nominal heads. Postpositions take a noun argument, which
can also be represented by a demonstrative, a pronoun, or a proclitic.

(333) Postpositions

Locational: asae 'in', 'under'

asaa 'inside', 'covered by'
hana 'inside' (surrounded)
kuane, kuanaj 'inside' (container)
tgoae 'on'
nitQae 'near'
kahe 'from'
kaihje 'behind/after'
Locational + ahinia 'before'
temporal:
tial 'until' (non-continuing)
uahei 'until' (continuing)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 225

Temporal: netohwei 'until' (temporal)

baja 'after'
bana 'at the time when'
Other: kurema 'for' (beneficiary)
raj 'for/to' (beneficiary)
kularte 'without'
ajßa 'together with' (comitative)
sajhjei 'despite of
käu 'because of (reason/source)
tonoana 'while' (temporal)
najnia 'while' (contrastive)
kujßa 'so that' (purposive)

5.7.1 Locational postpositions

asae 'in/under'
The postposition asae is used with the meaning of 'in' or 'under', mainly with
objects that have a roof. These include the words for 'house', 'hut', 'shed', or
'church building'.

(334) Use of asae

a)
nit lureri asae ni-akwa-e barue, itulere, inio-kuru
that house under be-DSTR-3ps/E masato all.kinds meat-PL
'There were masato, all kinds of things, different types of meat in that house'

b)
mi banaao asae
that shelter under
'under that leaf shelter'

Related to asae is asäa, which is typically used with reference to 'water' and
other liquids, meaning 'covered by' or 'deep inside'. In any case it is implied
that the object referred to has contact with the object it is covered by. Note that
this form is also used in combination with other postpositions that refer to a
similar semantic concept. These include hana and htatie, which both can be
translated as 'in/inside' (cf. (336), (338)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
226 Word classes

(335) Use of asäa

a)
nala-a=ne hana jiäe asäa tom-a
drop-3ps/D=SUB when already deep.inside sound-3ps/A
'When he dropped it [the fruit], it sounded "inside" [the water = under water].'

b)
akau-ana asäa
water-inside deep, inside

'deep below in the water' [Talking about a canoe which sank]

c)
itanitQa hana asäa
mud inside deep.inside
'in the mud, covered by mud'

hana 'inside'
The postposition hana occurs with various kinds of nouns, which include
nature-related or spatial nouns like 'water', 'field', 'forest', but it is also found
with 'eye'. In opposition to asäa, hana does not imply that the object it refers to
would be covered. A distinction between location and direction is not relevant,
as the example 'he went to the field' (cf. (336a)) shows. In example (336d),
kaahana even implies 'in this (tree)', but this has to be understood from the
context as the sentence is from a story in which Adam spends an extended
period of time in a house-like tree. Otherwise, one would talk of being "on" a
tree (using fpoae; cf. (343)).

(336) Use of hana

a) With 'field':
n=uhuana hana=te ku-e
3sg=field inside=FOC go-3ps/E
'He went to ["inside"] the field.'

b) With 'puddle':
neru-a kutuka hana
play-3ps/A puddle inside
'He is playing in the puddle.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 227

c) With 'eye':
i=ßuta hana=ra
2sg=eye inside=EMF
'because o f you' [Idiomatic expression; lit. 'in your eyes']

d) With 'tree' (implied):

netQara wane ni-a kaa hana lenone=ta
maybe PTY be-3ps/A this inside food=FRS
'Maybe there is food [for us unfortunate creatures] in this [tree].'

Note that hana has a related form, which is attached to nouns as a suffix -ana.
The exact difference between the two is not entirely clear, as most speakers
characterise the two as identical. They can occur with the same nouns without
an apparent difference ('water' akau-ana ~ akau hana). Similarly, both forms
are found with 'forest' and 'bush', respectively. However, one speaker
described akau-ana as 'swimming in/on top of the water' and akau hana as
'standing in the water', for example when working with logs. If we follow this
hint, hana refers to a state in which an object is "deeper inside", but not
covered, while -ana mainly describes a state of being on the surface, but inside
a location, for instance, in the centre. Interestingly, two of the examples in (337)
are double-marked for location, as 'mud' and 'cÄevow' (a type of tree) each
carry the suffix -ana plus the locative marker -u. In this case, the meaning of
'covered by' is again implied. Another example with -ana is 'in the moonlight',
which makes any further interpretation of its semantic structure rather difficult.

(337) Use o f - a n a

a)
nuna hana=te ne
jungle inside=FOC be:3ps/E
'He lives in the jungle.' [= in the middle of the jungle]

b)
itanitg,a-ana-H
mud-inside-LOC
'inside the mud' [i.e. covered, which would not be the case if were absent]

c)
elelia-ana-u
chevon-inside-LOC
'in the chevon tree' [use in response to 'where?']

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
228 Word classes

d)
atene kunu-ana
moon shine-inside

'in the moonlight' [lit. 'in the shine of the moon']

kuatie 'inside'

With nouns referring to objects that can be container-like, the postposition

kuane is used. Most likely, this form, which is the common realisation, is a
variant of kuanaj, resulting from word-final simplification of /aj/ to [e] (cf.
§2.8.7). The underlying form only surfaces when it is used in the plural
kuanaj-uru. Examples for the use of this postposition are 'bucket', 'canoe',
'mouth', 'bag', but the word also occurs with 'way/path', 'river' and all other
waterways (which obviously "contain" water). A list of nouns with which htane
is attested is listed below. Analogous to other postpositions, there is no
distinction between locational and directional use. Basically, a three-way
distinction can be drawn. In an abstract way of thinking, all objects can function
as a container, even though it is more difficult to argue for nature-related terms.
Some objects are designed in order to contain a content, while others in addition
have a non-flexible margin that enables the object to contain substances or
objects. For instance, clothes may function as a "container", but they do not
have a fixed margin, such as a bucket has. In most cases, the contained object is
surrounded by the respective container, but not covered from above (except,
possibly, the 'mosquito net' - however, this is not seen as a "cover" as the
person within is not touched from above). In fact, buildings are also included in
the terms that are used with kuane. 'Hands' and 'body' are additional terms that
are attested in this function.
(3 3 8) Occurrence o f kuane

nuhuae 'mouth' (Content)

bua 'bag' (Content)
kahjune 'clothes' (Content)
eruari 'mosquito net' (ContentTBuilding")
karesero 'prison' (loan) (Building)
bäako 'bank' (loan) (Building)
lureri 'house' (Building)
hata 'hole for giving birth' (Margin)
nukua 'nest' (Margin)
bote 'boat' (loan) (Margin)
balde 'bucket' (loan) (Margin)
kururihja 'crib' (Margin)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 229

sartee 'fiypan' (loan) (Margin)

enanihja 'canoe' (Margin)
beru 'way' (Nature)
tiunu 'curve' (Nature)
kukuana 'sea' (Nature, Margin)
nemuri 'lake' (Nature, Margin)
nukue 'river' (Nature, Margin)
ajriu 'Airico River' (or any (Nature, Margin)
other river)
bihi 'hand' (Body)
haberuku 'body' (Body)

In addition to the functions illustrated above, htane is also used to refer to

abstract concepts. N o u n s that occur with it include results o f writing, such as
'story', 'law', or 'book', but also with the term for 'day' (as in 'in/after three
days').

(339) Occurrence of kuane with abstract concepts

erenaa 'story' Writing

kiritca 'book' Writing
raui 'law' Writing
hanone 'day' Abstract
suuhua 'heart' (not physically) Abstract

The f o l l o w i n g examples illustrate the use o f kuane with nouns relating to

different kinds o f semantic concepts.

(340) Use of kuane with different semantic concepts

a) Margin:
tQäe tajbißaae-kuru eno-a-u bute kuane
also animal-PL enter-CAUl-IMP boat inside
'Also make the animals enter into the boat!'

b) Nature/Margin:
kaa ajtQune kuane eno-akaanu=ne hana
this PLN inside enter-lpl/ex=SUB when
'when we first entered the Espejo River'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
230 Word classes

c) Building/Content:
hua eno-a eruari kuane
don't enter-NTR mosquito.net inside
'Don't enter the mosquito net!'

d) Content:
beree amaema-he hetau unee bua kuane
child wander-CNT:3ps/E HRS kinkajou bag inside
'The kinkajou carries his child in a bag.'

e) Building:
lureri kuane-etg,a=na=ra ne-betaka-m here-u
house inside-only=FOC:lsg=EMF ITR-relax-INF want-1 sg/E
Ί want to relax only inside the house.'

f) Body:
arasihje siiri bihi kuane
star have:3ps/E hand inside
'He has got the stars in his hand.' [NT: Revelation 1:16]

Note that kuane is also used with 'water', but in this case 'inside' refers to
someone who stays on the surface, swimming, rather than being covered by
water. The difference with 'water' occurring with hana is that hana does not
necessarily refer to an object on the surface, but more to the fact that it is partly
covered as in the example where people stand in the water. In some examples,
such as (341), kuane is used as an instrumental postposition:

(341) kuane with instrumental function

lejhn hit^ana-atQa kuane=te nesaritQäete-kure katga-uru

one blowgun-only inside=FOC hunt-3pl/E man-PL
'People hunted with a blowgun only.'

The table below summarises the differences between the various terms used for
'inside'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 231

(342) Summary of terms for 'inside'

Covered Contact Surrounded On surface Container

with cover
asae yes no no no no
asäa yes yes yes no no
hana no no yes partly no
-ana no no partly possible no
kuane no no possible possible yes

tgoae 'on (top)'

This postposition again is used with a range of nouns. With some, it can mean
'on the surface', such as with 'land', 'tongue', 'beach', or 'floor'. On the other
hand, it is also found with 'tree' (in locational or directional sense, depending
on the verb that may be involved). In addition, tpoae may be used in a figurative
sense 'on top of (something)' to mean 'besides'. The shape or size of the
respective object is not relevant.

(343) Use of tgoae

a) Direction:
kanu lele tgoae tohoeta-u-ra
lsg tongue on lay-IMP-EMF
'Put it onto my tongue!'

b) Passing over:
tgoae bauleneto-a nehere-kure enamhja kttane
on drive.over-NTR almost-3pl/E canoe inside
'They almost drove over him in (= by) the canoe.'

c) Figurative
tgäe=te tgu nii tgoae taka-i eshoela hoaauka-anu
also=FOC CRTN that on encounter-PRT school think-lps/A
'("Meeting") on top of that, I also think of [building] a school.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
232 Word classes

mtgae 'near'

The word mtQae is used with an argument to express 'near', 'in the vicinity o f ,
or 'in the area o f . It also functions as a locational demonstrative and can be
defined as referring to a non-punctual location or direction (cf. (332)). In all
examples displayed in (344), the unspecified character of the location is given
by the context. The directional function is illustrated in (344c), where the
context makes a translation as 'towards' more appropriate.

(344) Use of nitQae

a) Location:
akau nitg,ae
water near
'close to the water' [about the location of crocodiles]

b) Direction:
na-1 hetau akauru mtQae rialaeto-a
say-PRT HRS 3pl near throw-3ps/A
'Saying so, she launched [the shells] towards them.'

kahe 'from'
The postposition 'from' is multifunctional. It is used not only with nouns that
refer to a location, but also in an abstract sense with 'people' (e.g. 'learn from
somebody'), kahe can also be used in a temporal sense with adverbs of time;
however, in this case, the participle form o f ' c r e a t e ' , hoaaunei is added to mean
'beginning from'.

(345) Use of kahe

a) With location:
dede kahe=te hiua-e
sky from=FOC descend-3ps/E
'She came from heaven.'

b) With common nouns:

lureri kahe hiua-it
house from descend-lsg/E
Ί have descended from the house.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 233

c) With humans:
kahjwie kaja-hahoa-uru-a katqa-uru kahe
clothes seize-DSTR-PL-3ps/A man-PL from
'They took the clothes away from the people.'

d) With humans (abstract sense):

leotg,a-uru kahe nerela-t
other-PL from learn-PRT
'learning from the others'

e) With abstract meaning:

mi kahe=ne huituku-ü
that from^FOClsg know-lsg/E
'From that [experience] I know.' [i.e. I have Ieamt from this situation]

f) With numeral:
sauki kahe=te nitgataha-l setu-e ate
five from=FOC three-PRT rot-3ps/E fish
'Three out of the five fish have rotted.' [Elicited]

g) Indicating material:
hoata kahe itQa-noi
iron from make-PASS
'made of iron'

h) With location and kwaauner.

ikito kahe hoaaune-i
PLN from create-PRT
'Starting from Iquitos [we will move on].'

i) Temporal meaning:
eresi kahe kwaaune-i
tomorrow from create-PRT
'from tomorrow on' [Lit. 'creating from tomorrow']

kaihje 'behind'

This postposition is found with h u m a n s only. Its meaning ' b e h i n d ' is strictly
locational or directional (not temporal); it can also be used metaphorically in the
sense o f ' b e h i n d s o m e o n e ' s b a c k ' .

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
234 Word classes

(346) Use of kaihje

a)
akauru kaihje ini-a ku-a
3pl behind go.up-NTR go-3ps/A
'She climbed behind them.'

b)
kaihje muku-na heri-a-ne hana
behind catch-INF want-3ps/D=SUB when
'when he wanted to catch her [running] behind/after her'

c)
kaihje turu-1 fioanara basihjau-a
behind arrive-PRT banana steal-3ps/A
'He stole bananas behind [someone's back].'

5.7.2 Postpositions with locational and temporal function

ahinia 'before'

T h e postposition ahima ' b e f o r e ' can have temporal or locational reference. It

can also be realised as the variant ahntQa, which appears to have the same
function and meaning. In (347a), ahima is used with temporal function after a
pronoun, to imply ' b e f o r e our arrival'. T h e example in (347b) s h o w s that the
argument can also be realised as a proclitic.

(347) Use of ahinia

a) Temporal:
kanakaanu ahinia barue ahu-u=ra
lpl/ex before masato mash-IMP=EMF
'Prepare masato before we arrive!'

b) Locational:
hana=te n=ahima ne-ure arahit katQa-uru
instead=FOC 3ps=before be-3pl/E many man-PL
'And there were many people [standing] before him.' [NT: Matthew 21:8]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 235

c) Variant ahntga:
kanu ahutga ne enua
lsg before be:3ps/E tree
'The tree is [standing] before me.'

tial 'until'
Urarina has several terms for 'until', one of which is tial. This postposition is
generally used in an inherently directional sense. However, there also is the
postposition reetiai, where the meaning of /ree/ is unclear. This form has
temporal meaning, occurring after adverbs of time or after verbs.

(348) Occurrence of tial, reetiai

a) Locational use:
nukue kuane hoeta-akaanu=nuke tial
river inside clear-lpl/ex=PURP until
'[One could enter] up to the spot that we were supposed to clear.' [Lit. 'until where we
were to clear the river']

b) Temporal use:
hauria hitarii letQuyka-j katga rela-i=jie reetiai nihjauria
first all ten-NOMSbj man teach-2ps=SUB until don't

be-i=jie
tell-2ps=NEGF
'In the beginning do not tell [them] anything until you have taught all ten persons [how
to do it]!'

uahei 'until'
Another term for 'until' is uahei, which probably is a fossilised form of 'come'-
CNT-PRT, to mean 'coming'. The difference with tial lies in the fact that uahei
is continuing, while tial and reetiai define locational or temporal "boundaries".
For instance, in example (348a) 'up to the spot ...' refers to the endpoint of an
area to be cleared, whereas hi uahei (cf. (349a)) implies a temporary stop on a
trip that is to be continued.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
236 Word classes

(349) Use of uahei

a)
ku(-atQa) uahei
there(-only) until
'until there [and then going on]'

b)
ena nerereta-re-u=ni ßaa-hi-anu=ne haa ke uahei
now tell-IRR-1 sg/E=ASS know-CNT-lsg/A=SUB PURP VLI until
'Now I will tell stories as far as I know.' [but implying that he may know more]

5.7.3 Postpositions with temporal function

netohwei 'until'

A further expression for 'until' is netohwei, which is used with temporal

function only. Most typically, it occur with adverbs of time, such as ' n o w ' , or
' f o r e v e r ' . N o t e that the adverb ßoaelu, which normally m e a n s 'earlier', 'in
ancient t i m e s ' takes the meaning ' f o r e v e r ' in combination with the postposition
netohwei.

(350) Use of netohwei

a) Reference to present:
ena netohwei=te ne ißono ku-1 nii ajßa ne-nakauru
now until=FOC be:3ps/E ayahuasca drink-PRT that with be-those.who
'Until now there are those who live with that, drinking ayahuasca.'

b) Reference to future:
ajane rihihei ßoaelu netohwei kajritgi ne-ι ne-ri-h-tQe=l
like.that ever until poor.person be-PRT be-IRR-2ps-PL=ASS
'Being like that, you will live being poor people forever.'

O n e d i f f e r e n c e between netohwei and other expressions for 'until' that have

temporal function is that netohwei is also used with future reference (cf. (350b)),
while examples for this are attested for the other words. The following table
lists the main differences between the various terms. N o t e that the feature
" F u t u r e " is an additional, but not essential feature in order to distinguish the
four terms.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 237

(351) Summary of terms for 'until'

Locational Temporal Continuing Future

tiäi yes no no ?
reeitiäi no yes no ?
uahei yes no yes ?
netohwei no yes yes yes

baja 'after'
The word baja is used with a temporal function for 'after'. It frequently follows
the demonstrative nii, to mean 'after that', 'then'. Baja is not very widely
attested as postposition, but mainly occurs as a conjunction in combination with
the subordinate marker (cf. §5.8). In fact, there are no examples in the text
database where it occurs after full nouns. There are a number of examples
attested in the NT in which baja occurs after a pronoun, and two cases in which
it co-occurs with the nouns for 'death' and 'day', respectively.

(352) Occurrence of baja

a) After demonstrative:
nii baja=te nii amiam-naa saku-re=i leotQa
that after=FOC that work-NOM follow-IRR:3ps/E=ASS other
'After that, someone else will continue that work.'

b) After pronoun:
kann baja=te u-re=~i lejhii
lsg after=FOC come-IRR:3ps/E=ASS one
'After me, one will come ...' [NT: Matthew 3:11]

The form baja-hiri-i is a derivation that involves suffixes for diminutive and
participle and is used to mean 'a little later'.

(353) Use of bajahirii

mi baja-hiri-i=te päe ku-e
that after-DIM-PRT=FOC already go-3ps/E
Ά short time after that, he already went away.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
238 Word classes

bana 'when'
In the same way as baja, the word bana can function as a postposition or as a
conjunction (cf. §5.8). It has a temporal meaning and is translated as 'at the time
o f . It is used predominantly with the demonstrative.

(354) Use of bana

a) With demonstrative:
nii bana hetau=te
that when HRS=FOC
'in those days (it is said)'

b) With weekdays:
tomijo bana-te nebetaka-e
Sunday when=FOC relax-3ps/E
O n Sunday(s), he relaxes.'

c) With nominalisation:
jwaelu kana ne-huaau-naa bana
earlier lpl/in ITR-create-NOM when
'earlier, at the time of our creation'

d) With full noun:

itQauena-uru iteau bana
ancient-PL life when
'at the time of the life of the ancient people'

The word also occurs as a component of the complex interrogative d^a+bana

'when' and is composed of the word for 'what' (which can also function as an
indefinite pronoun; cf. §21.1.11), and 'when' (or 'time'). Further occurrences
are observed in neebana 'for the first time' and in the adversative introducer
nabana. However, the morphological composition of these expressions is not
transparent.

5.7.4 Postpositions with other functions

kurenia 'for'

The postposition kurenia is probably related to the root /kure/ that occurs with
the meaning for 'change, exchange', and recently also for 'price'. Related words
include hiretia 'buy, change for', (which can be segmented into /kure/ and /tia/

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 239

' g i v e ' ) and kurerama 'replace', (which is morphologically opaque).

Correspondingly, kurenia functions in a semantic range between three slightly
different concepts:
a) Benefit/recipient 'for' or maleficiary 'against'
b) Exchange ' f o r ' or 'in order to get'
c) Source 'due to', 'thanks to'

Note that the recipient of an item or action can be interpreted as beneficiary or

maleficiary, depending on the context. For example (355b), 'for you' is part of a
curse, as something bad is being transmitted to the recipient.

(355) Use of kurenia

a) Beneficiary:
kana+kwaaun-era kurenia=ne itQau-akaanu
lpl/in+create-AG for=FOC:lpI live-lpl/ex
'We live for God.'

b) Maleficiary:
mtoanei ne i=toej kurenia=ra
like.that be:3ps/E 2sg-for for=EMF
'It shall be like that for you [forever].' [Meant as a curse]

c) Exchange:
kalaui kohvaerati-a aräala kurenia-nemaahel
son defend-3ps/A tapir for-also
'He [the jaguar] defended his baby and also for [i.e. in order to get] the tapir [the jaguar
attacked].'

d) Payment:
kurete-ri-tQäH=rü kaa kurenia
pay-IRR-1 sg/A=ASS this for
Ί will pay for this.'

e) Source
naria ebanhelio, ramo kurenia leotQa-uru misjonero kurenia
thanks Gospel PSN for other-PL missionary for
'Thanks [= due] to the Gospel, thanks to Ramon, thanks to the other missionaries [we
live in peace].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
240 Word classes

raj 'for', 'to'

The frequently used postposition raj is identical to the possessive marker (cf.
§7.1.2). This overlap is not coincidental, as raj follows a recipient. If compared
to the structure of a possessive construction such as kanu raj lureri (lsg + POSS
+ 'house') 'my house', a literary translation could also be "house for me". The
origin of raj goes back to the nominalised form of raa 'receive': /ra-i/ would
then refer to 'what (he) received'. However, as a postposition, raj consistently
occurs with recipient function translated as 'for' or 'to'. It follows animate and
inanimate nouns; in (356a), it occurs after a locational demonstrative.

(356) Occurrence of postposition raj

a) 'For'; with locational demonstrative:

kaalaj raj kuraanaa ne-ü
this.area for chief be-lsg/E
'I'm the chief for this area.'

b) 'For'; with noun:

ena=te amiane kana+kiuaaun-era raj
now=FOC work:3ps/E our.creator for
'Now he works for God.'

c) 'To'; with noun:

ka=raaheniam=na lajna-u=lu gobernador raj
lsg=self=FOC:lsg write-lsg/E=REM governor for
Ί myself wrote to the governor'

Note that the example in (356b) resembles the one in (355a) 'we live for God'
and mainly differs by the use of postposition. In fact, the difference between
kurenia and raj in this particular context is marginal; kurenia appears to
emphasise the beneficial character. However, the recipient function of kurema is
only one of three different options, while for raj it is the only function. Since
kurenia is more specialised in its meaning, it can be assumed that it expresses
the benefit (or misfit) in a stronger way than raj does.
One additional function of raj is in expressions that refer to obligation (cf.
§18.3). In general, this involves the construction ne+te+ne, which is a
combination of subordinate marker plus focus marker, plus copula, literally 'it
is that'. As a variant of this construct, the focus marker =te can be preceded or
replaced by raj, which results in a literal translation as 'it is for that'. Even
though the three components of this form can still be identified, it appears to be
lexicalised, or it is on the way to being lexicalised, as the variation between
netene and neraj(te)ne may suggest.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 241

(357) Occurrence of raj in 'must' constructions

a)
satit ku-i-toe nerajße
all drink-2ps-PL must
'You all must drink it.'

b)
atauari su-anu nerajtene
chicken kill-lsg/D must
Ί must kill a chicken.'

hilane 'without'

Urarina has two ways to express 'without'. Kulane is used as a postposition

with nouns only, while there also is a verbal suffix -elanaala (cf. §12.4; §20.4).
In a very general sense, hilane negates the N P which is in its scope. This
includes animate and inanimate nouns and may be understood as a negation of
possessive, instrumental or comitative function. Further attested examples in
addition to those listed in (358) include 'clothes', 'sin', ' f o o d ' , 'work', 'people',
and 'fishing hook'. In (358b), hilane follows d^a, which here occurs in
quantifying function; negation in this example is expressed by hilane (and
further emphasised by nijej 'not at all').

(358) Use of Mane

a)
itQafiua kulane=te ne-ure=lu katQa-uru
rifle without=FOC be-3pI/E=REM man-PL
'The people lived without rifles.'

b)
nijej d^a kulane
not.at.all what without
'Without anything at all [I had to defend myself].'

ajfla 'together with'

This postposition has a strictly comitative function as its instrumental use is
excluded. (Note that the instrumental function is expressed by ke, which also
functions as a valency increase marker, cf. §16.5.) Ajßa refers to 'in company
o f and occurs with animate and inanimate arguments, including demonstratives

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
242 Word classes

that may refer to an abstract concept such as 'this (situation)'. At first sight,
ajßa may seem to be a phrase conjunction, as in the example 'Juan and
Manuel' (cf. (359a); however, the context implies that this is to be read as
'Manuel with Juan'. In addition, the verb in this example appears in the 3sg
form, which indicates that it is Manuel who goes, but in the company of Juan.
Beyond that, there are very few examples, which would appear to conjoin two
nouns or other syntactic phrases with each other. Thus, it can be concluded that
ajßa is a postposition. The example in (359b), ajßa occurs without an overt
argument, because it refers to a 3ps object pronoun, which is typically omitted.

(359) Use of ajßa

a) Human referent:
hwaa ajßa manoel=te nukue kuane ku-re=i
PSN with PSN=FOC creek inside go-IRR:3ps/E=ASS
'Juan will go to the river with Manuel.'

b) Human referent without overt noun:

hjane-na here-lanaala nii rihihe-ϊ ajßa räasa-uru-a hana
let-INF want-PRV that be.like-PRT with dance-PL-3ps/D when
'when they danced with her like that, without wanting to let her (go)'

c) Inanimate referent:
kunaj ajßa=te hatal nekajritga-e
sickness with=FOC very suffer-3ps/E
'He has suffered a lot with illness.'

d) Inanimate referent:
saabere ajßa neda-a=ne hau
machete with remain-3ps/D=SUB because
'because [the rifle] had remained with the machete'

e) Abstract referent:
ni-a mi ajßa ne-nakauru katga-uru
be-3ps/A that with be-those.who man-PL
'There are people who live with that (ayahuasca drinking).'

Ajßa is also used in expressions such as nii hau ajßa te ('that' + 'because' +
'with'=FOC), literally meaning 'with (the situation being) because of that'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Postpositions 243

sajhjei 'despite o f
Another postposition that is also used as a subordinate conjunction is sajhjei
'although'. It is only attested with the demonstrative nii 'that' to mean
'nevertheless', but does not occur with any other nouns. It can also occur as the
variant asajhjei, with no apparent condition.

(360) Use of sajhjei

a)
mi sajhjei be-naa here-ne
that despite tell-INF want-NEG:3ps/E
'Nevertheless [even though they urged her], she did not want to tell.'

b)
nii asajhjei nijej ere-naa towre-lanaala ke-et$a balubalubaluu-ka
that despite not.at.all speak-NOM obey-PRV VLI-only embrace-3ps/A
'Nevertheless, without obeying, he embraced only her.'

käu 'because o f
The postposition käu refers to the reason of source of an action or event and can
be translated as 'because o f , 'out or' or 'due to'.

(361) Postposition käu

a)
atane huu-ka akau=ne hana nesoone käu d^ana-i
land flood-3ps/D water=SUB when fear because.of ciy-PRT
'As the water flooded the earth, [they were] crying for fear and said ...'

b)
hatal eluna-i hoasi-jia käu hetau aj
very be.terrorised-PRT be.scared-NOM because.of HRS AUX:3ps/E

nii ranuna
that girl
'Being so terrorised, the girl did it out of fear.'

There are two more terms that can be used with the demonstrative, but their
main function is as clause conjunctions. These are two words for 'while',
tonoana and najrtia. With nii, they are translated as 'in the meantime', 'during
that'. With other nouns, their use is not attested. Their structure will be further

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
244 Word classes

discussed in §5.8. In the same way, the word kujfla 'so that', which mainly
functions as a purposive subordinate conjunction, can follow a demonstrative to
express 'for that purpose'. However, its use in this function is very rare.
The table in (362) gives an overview of postpositions and their occurrence
with nouns and as conjunctions. Two types of postpositions are predominant:
those that occur with nouns and demonstratives, but do not function as
conjunctions (Type A), and those that can take this function in addition to their
function as postpositions (Type B). Three postpositions (Type C) only occur
with demonstratives and as conjunctions, but not with other nouns. Three other
postpositions are unattested with demonstratives; the word netohzuei is peculiar
as it only occurs after temporal adverbs. A further feature which allows to
distinguish the various postpositions is their reference to location or direction.

(362) Functions of postpositions

Gloss With With other As Referring

DEM nouns conjunction to location
or direction
Type A:
asae 'in/under' yes yes no both
kuane 'inside' yes yes no both
tg-oae 'on' yes yes no both
kaihje 'behind / yes yes no both
after'
kurema 'for' yes yes no none
raj 'for/to' yes yes no none
hilane 'without' yes yes no none
ajpa 'together yes yes no none
with'
käu 'because of yes yes (no) none

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Conjunctions 245

Type B:
ham 'inside' yes yes yes both
kahe •from' yes yes yes DIR
ahima 'before' yes yes yes LOC
bana 'at the time yes yes yes none
when'
baja 'after' yes yes yes none
Type C:
tomana 'while' yes no yes none
najnia 'while' yes no yes none
sajhjel 'although' yes no yes none
kujpa 'so that' yes no yes none
hvataa 'so that not' yes no yes none
hijßa 'so that' yes no yes none
Others:
nitQae 'near' unattested yes no both
tia~i 'until' unattested yes no DIR
uahei 'until' unattested ? no DIR
netohwei 'until' no no no none

5.8 Conjunctions

While Urarina has a comparatively large number of clause conjunctions, there

are no phrase conjunctions. The only word that would be a candidate for "and"
(ajjia) is a postposition rather than a conjunction. Subsequently, two nouns or
noun phrases are associated with each other by simply listing them as a
sequence. The enumerative verb naaohwaa, which could be roughly translated
as 'these are' or 'do so' (when referring to an enumeration of clauses), is
employed (optionally) to indicate enumeration. This verb is inflected in the
same way as other verbs for various categories, which include person, number,
irrealis (future), imperfective aspect, and negation and can also be nominalised
and take the participle form. It marks enumeration and, according to Manus
(p.c. 2003), appears to specifically emphasise that the items to which it refers
constitute separate entities, as opposed to items that are "mixed". The verb can
refer to two or more nouns that precede it, but is never used with one noun only.
Its use also excludes the meaning of 'and so on', i.e. naaohwaa implies that the
list of items mentioned is complete.
The examples listed in (363) occur without naaohwaa. (363a) refers to a
meal, which implies that the separate status of the involved items is not
relevant; therefore, they occur in a sequence without the enumerative verb. The
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
246 Word classes

example in (363b) is elicited and was given as a possible translation for 'you
and P. The use of naaohwaa in (363c), where it occurs in the participle form,
illustrates the listing of a number of food items. However, naaohwaa can also
function as to "list" a chain of dependent clauses. In example (363d) naaohwaa
occurs in the 3pl/E form and refers to an enumeration of events, here expressed
through two participle clauses (also cf. §5.16).

(363) Sequence of nouns

a)
fwanara take
banana egg
'plantain and egg' [a meal]

b)
kanu ii
lsg 2sg
'you and me'

c)
ku ne imo, itulere, fioanara, laano, naaohwa-i
there be:3ps/E meat all.kinds banana cassava ENUM-PRT
'There was meat - all kinds of things, bananas and cassava.'

d)
raj mnaae+buku ra-l, bajbihi ra-ϊ, riaaohwa-ure
POSS leg+bone receive-PRT upper.poles receive-PRT ENUM-3pl/E
'Taking its vertical poles and taking the upper poles, they do it.' [about house-building]

As seen in the previous section, Urarina has a number of words that can take
postpositional or conjunctional function. In fact, all subordinate conjunctions,
(which are listed in (364)), except for kttreu 'after' and hanamane 'if are
attested as postpositions after the demonstrative nii. Their main use is as
subordinate clause conjunctions, i.e. they normally follow a verb. Following
this, one could alternatively analyse conjunctions as actual postpositions that
can take a dependent clause as an argument. Except for tQäe and riete, which
function as clause coordinators, all Urarina conjunctions introduce a subordinate
clause. These follow the subordinate enclitic =ne, which is attached to the last
element of the clause. No elements can occur between =ne and a conjunction;
however, the subordinate marker may be omitted in a subordinate construction.
=ne also has a variety of functions as accompanied by a conjunction, it can
introduce a temporal or conditional clause (cf. §20.1.1). Also note that the verb

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Conjunctions 247

of a subordinate clause is marked by the "D-form", i.e. the person inflection

type that applies to verb in a dependent clause. Apart from the presence of =ne,
this can be conveniently used as a criterion in order to distinguish subordination
from coordination, as the D-form only occurs with subordinate conjunctions.

(364) Conjunctions
a) Subordinate conjunctions (with subordinate marker =ne):
Temporal: hana 'when'
bana 'at the time when'
ahirna 'before'
baja 'after'
kureu 'after'
tonoana 'while'
najrda 'while', 'whereas'
kahe 'from' (Temporal and locational)
Others: hau 'as/because' (Reason)
kujßa 'in order to' (Purpose)
huataa '(in order) that not' (Purpose)
sajhjel 'although' (Concessive function)
hanardane 'if (Condition)
rihihei 'like' (cf. §5.4.6) (Comparison)
tokuanei 'compared to', 'like' (cf. §5.4.6) (Comparison)

b) Clause coordinators:
tgäe 'also'
ne+te 'but'

5.8. J Temporal conjunctions

hana 'when'
The conjunction hana is mainly employed to mark a subordinate clause related
to time. However, there is some interesting relation to discourse structure, as
will be examined in §22.4.5. In particular, this regards the relation between
hana and häu (cf. §5.8.2; §22.4.5).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
248 Word classes

(365) Use of hana

a)
amu-a=ne hana=te su-e hanolari
walk-3ps/D=SUB when=FOC kiII-3ps/E jaguar
'When he went (hunting), a jaguar killed him.'

b)
haurianehe-1 kaa ajtgune kuane eno-akaanu=m hana
for.the.first.time this PLN inside enter-lpl/ex=SUB when

ri-ei=lu ne-naanaha
be-NEG:3ps/A=REM be-LPP
'When we entered this (river) Espejo for the first time, there was no place to live.'

To account for the semantic and functional difference between the conjunctions
hau and harm, which are the two most frequently used ones, is a challenge. A
clue towards a solution may lie in the semantic difference in their use with the
demonstrative nii. In this context, the meaning is invariably 'therefore' for
nii hau, but 'then' for nii hana. Any similarities in function or semantic content
thus occurs in relation to discourse; this will be analysed in §22.4.5 on
discourse.

bana 'when'
There is a certain degree of overlap in the meaning of bana and hana. The
conjunction bana is generally much less frequent than hana. In correspondence
to its occurrence in postpositional function, it refers to a time or era, rather than
a particular event. In contrast, hana indicates the time of a particular event.
While one might suspect from this a relation to habitual aspect, this is not the
case. The presence of bana is not restricted to a certain verb type, nor associated
to a particular type of aspect marking, as the examples in (366) illustrate.

(366) Use of bana as a conjunction

a)
ne-ni-a iglesia=ne bana
be-NEG-3ps/D church=SUB when
'(at the time) when there was no church'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Conjunctions 249

b)
hanulari su-anu=lu=ne bana
jaguar kill-lsg/D=REM=SUB when
'(at the time) when I killed jaguars'

ahinia 'before'

As a conjunction, ahinia has temporal function (whereas it can also have

locational reference when used as postposition; cf. §5.7.2).

(367) Use of conjunction ahinia

a)
dosmildos ke turu-si-a=ne ahinia hestjona-anu nerajtene
2002 VLI arrive-CPL-3ps/D=SUB before delegate-lsg/D must
Ί must delegate it before the year 2002 arrives.'

b)
eno-aka=ne ahinia nukue kuane hveta-aka
enter-lpl/du=SUB before river inside clear-lpl/du
'Before we enter the river, we clear it [of logs].'

baja 'after'
Similar to its use as a postposition, its occurrence of baja with a subordinate
clause has a temporal function and focuses on the sequence or order of events.

(368) Use of baja as a conjunction

a)
setu-ahva-a enua=ne baja kau ßäe ne-naha
rot-DSTR-3ps/D tree=SUB after here already be-LPP

ti-a u-akaanu=ln
give-NTR come-1 pl/ex=REM
'After the trees rotted one by one [in the other place], we already came here to make a
place to live.'

b)
raneto-a lomaj beene=ne baja hetau
weave.cloth PSN female=SUB after HRS
'After Lomai had been weaving, [she said ...]'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
250 Word classes

kureu 'after'

The conjunction htreu is not attested as a postposition. However, it is extremely

rare in the database and is not attested at all in the NT. The difference between
baja and kureu only consists of a nuance as both could be translated as ' a f t e r ' .
However, the use of kureu implies an additional connotation, which focuses on
the fact that the " n e w " event (i.e. the second event of a sequence) occurs "in
e x c h a n g e " of the former activity, while the first activity may still be going on.
For instance, if example (369a) contained baja instead of kureu, the temporal
order of events would be emphasised. With kureu, instead, the fact that the
speaker could not be there when his grandson died is stressed. The use of this
conjunction makes the relation between "going" and the event that happened in
the background more dramatic, as if to describe the contrast between the two
events. Example (369b), which is based on context-related elicitation,
emphasises the fact that the work, in a way, "replaces" the previous action of
eating.

(369) Conjunction kureu

a)
ikito ku-anu=ne kureu=te u-e naao
PLN go-lsg/D=SUB after=FOC die-3ps/E PSN
'After I had gone to Iquitos, Naao died.'

b)
lenom-a=ne kureu amiani-a ku-e
eat-3ps/D=SUB after work-NTR go-3ps/E
'After he had eaten, he went to work.'

tonoana, najma 'while'

Urarina has three different strategies to express the concept of ' w h i l e ' , and the
differences between these are marginal. While one morpheme for while is a
suffix, there are two words that function as a subordinate conjunction. As
illustrated in (370), the conjunction tonoana is used to refer to actions or events
that occur simultaneously.

(370) Use of tonoana

a)
amiani-akaanu=ne tonoana nebetaka-ure leotQa-uru
work-lpl/ex=SUB while relax-3pl/E other-PL
'While we are working, the others are relaxing.'
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Conjunctions 251

b)
tQoaetQani-a ku-a kitQa=ne tonoana=te kaihje ku-e eene
go.ahead-NTR go-3ps/D man=SUB while=FOC behind go-3ps/E woman
'While the man went ahead, the woman went behind.'

The other conjunction for 'while' is najrna. Unlike tonoana, najrna has a
contrastive function, by which it sets the action described in the subordinate
clause in contrast to the action of the main clause. For instance, the same
sentence as the first example in (370) is possible with najnia instead of tonoana.
The difference in meaning is explained through contrast, which, for instance,
can be manifested by the fact that with najma, the two groups of people do not
work together in the same place. A possible translation for the first sentence in
(371) would thus be with 'whereas'.

(371) Use of najma

a)
amiani-akaanu=ne najma nebetaka-ure leotQa-uru
work-lpl/ex=SUB while relax-3pl/E other-PL
'Whereas/while we are working, the others are relaxing.'

b)
kau m-akaanu=ne najma, estadosunidos ne ramoo
here be-lpl/ex=SUB while PLN be:3ps/E PSN
'Whereas/while we are here, Ramön is in the United States.'

Note that 'while' may also be expressed by a verbal suffix -lai (also cf. §20.4).
The main difference to the two conjunctions discussed here is that -lai only
occurs with same subject function.

kahe 'from'
The word kahe is relatively rare as a conjunction, whereas it is more frequently
used as a postposition. Similar to its use after nouns, it also has locational
function when following a clause. However, a temporal component is always
given through the context, as illustrated in the examples in (372). For instance,
the fact that 'they have returned from fishing' (which is meant as an indication
of the direction), automatically implies that the process of fishing occurred
earlier in the timeline than the return.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
252 Word classes

(372) Use of kahe as a conjunction

a)
pae kau-a ku-uru-a ate kuhivu-uru-a=ne kahe
already return-NTR go-PL-3ps/D fish fish.with.hook-PL-3ps/D=SUB from
'They already go home from fishing.' [Lit. 'from catching fish']

b)
ajto-a nil ranuna=ne kahe kmtuku-i
say-3ps/D that girl=SUB from know-NEG:3ps/A
'They don't know [it] from what the girl says.'

5.8.2 Other conjunctions

häu 'because'

The subordinate conjunction häu is mainly related to cause. However, similar to

the English 'as' or 'since', this meaning can be expanded to a partly temporal
function. Further discussion on this follows in chapter §20.1.2 on subordinate
clauses. As with all other subordinate conjunctions, the use of the subordinate
marker =ne is common, but not obligatory.

(373) Use of hau

a)
tasißohwa-a hau=te kana kasate
be.big-3ps/D because=FOC lpl/in frighten:3ps/E
'Because it is big, it frightens us.' [About jaguar]

b)
ku kwara-1, hatai kauato-a häu, here-Ί, muku-t ajpa
there see-PRT very be.nice-3ps/D because want-PRT catch-PRT with

räasa-ure
dance-3pl/E
'They saw her, and because she was very beautiful, they desired her and danced with
her.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Conjunctions 253

kujßa 'in order to'

This conjunction has purposive function best translated as 'so that' or 'in order
to'. While the syntactic difference to other forms related to purpose
(haanu,-naha\ cf. §20.2.4.1), is evident in that only kujßa introduces a
subordinate clause, it is not clear-cut what could be the semantic differences.
Note that the translation 'in order to' is not meant to imply a same subject
construction (as in English), but that the subjects of dependent and main clause
can be different.

(374) Use of kujßa

a)
mi häu=te tururi molo-ore ruku-uru-a=ne kujßa
that because=FOC llanchama.Xree cut-3pl/E pull.out-PL-3ps/D=SUBso.that
'Therefore, they cut the llanchama tree in order to pull her out.'

b)
satonohei=te ikito-o ama-ure hoitQa-kuru-a doktoro-kuru=ne kujßa
in.the.end=FOC PLN-LOC take-3pl/E heal-PL-3ps/D doctor-PL=SUB so.that
'In the end, they brought him to Iquitos so that the doctors (would) cure him.

huataa 'so that not', 'lest'

hoataa (and its variant kauatia\ occurring only in the traditional language) is
the negative equivalent of kujßa. Accordingly, the subordinate clause can have
a different subject than the main clause. Generally, the fact that the subordinate
marker already implies negation leaves the dependent verb in the affirmative
mood. When the subordinate clause is in initial position, it can be emphasised
by the introducer nabana, as shown in (375c).

(375) Use of hoataa

a)
itga-u eno-a akau=ne huataa
do-IMP enter-3ps/D water=SUB so.that.not
'Do it so that the water (would) not enter!'

b)
anofwa dada-a kanu beree=ne hoataa ajto-o
knife touch-3ps/D lsg child=SUB so.that.not say-lsg/E
Ί said it so that my child would not touch the knife.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
254 Word classes

c)
nabana hoena-uru-a=ne hvataa enuhe tauhja-1,
that.not laugh-PL-3ps/D=SUB so.that.not pepper collect-PRT

makusajari tauhja-i, murara-l ku-ure

red.pepper collect-PRT put.in.mouth-PRT go-3pI/E
'They went, gathering peppers, putting big peppers into their mouths, so that they would
not laugh.'

sajhjei 'although'

sajhjei is a subordinate conjunction with oppositional (concessive) function. It

has a variant asajhjel, which is in free variation with sajhjei.

(376) Use of sajhjei

a)
enua ari-anu=ne sajhjei ne-ni-a=na hau, nii baja
tree seek-lsg/D=SUB although be-NEG-3ps/D=SUB because that after
'even though I looked for a stick, (as) there was none, [but] then ...'

b)
na-a=ne asajhjel hetau ku aua-na here-lanaala
say-3ps/D=SUB although HRS there hear-INF want-PRV

ku balubalubaluu-ka hau
there embrace-3ps/D because
'though she spoke like this, (as) he did not listen and kept on embracing her ...'

A construction related in function with regard to opposition involves two

elements: ria natu. However, while this expression can also be translated as
' t h o u g h ' , neither of the two elements is a conjunction; the form is composed of
a form of the copula and atn, which could be interpreted as the participle form
of the verb for 'insist'. An additional meaning of this construction refers to
' e v e n ' , ' u p t o ' , but in this context it requires a different syntactic structure:
while it is possible to separate ma and natu from each other in the ' t h o u g h '
construction, the two components are inseparable when they are used with the
meaning of ' e v e n ' . In addition, this function does not allow the inflection of ma,
which makes it questionable that the two functions would be identical. For
instance, in (377a), the copula ma occurs in the lpl/ex form and in example
(377b), the two components are separated by the subject mikatga 'that m a n ' .

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Conjunctions 255

Most other examples (such as (377c)) exhibit manatn as an inseparable and

invariable unit. Therefore, I transcribe manatn as one unit in this context.

(377) Use of ma natu

a) Meaning 'although': inflection for person:
βaara rihihei katQa ne-ϊ ni-akaanu natn=na, enejtgu ne-1
2pl like man be-PRT be-lpl/ex though=FOC.pl monkey be-PRT

ne-huara-kaanu ßaara raj=na hau

ITR-see-lpl/ex 2pl for=SUB because
'Even though we are people like you, (because) we appear (as) monkeys to you.'

b) Meaning 'although': non-contiguous occurrence:

kana letoa-i ni-a amu-naa+kwaaun-era natu=ta
lpl/in send-PRT be-3ps/D walk-NOM+create-AG though=FRS
'[How is that possible] ... even though the Creator of Ways has send us'

c) Meaning 'even':
nehelau ninitQU kohvajtQa-kuru nianatn=te nitoanei uhoana ha-ure
other.side PLN inhabitant-PL even=FOC like.that field make-3pl/E
'Even the people from the other side of the Rio Chambira made their fields like that.'

hanamam ' if
The conjunction hanamane is used to emphasise irrealis in a subordinate clause,
explicitly referring to an unfulfilled condition (also cf. §20.1.1.3). Irrealis is not
marked as a suffix on the verb of the dependent clause; the hypothetical status
of the clause is solely indicated by the conjunction. However, irrealis can be
marked in the main clause (which is not obligatory). In (378a), the conditional
function of the dependent clause is additionally emphasised by the introducer
baana (cf. §5.10.6); one of the clauses of this sentence is marked for irrealis. In
(378b), the main clause has irrealis inflection.

(378) Use of hanamane

a)
baana itQuuni-a=ne hananiane raj kalaui-tQuru muku-akat(x
if be.near-3ps/D=SUB if POSS son-PL catch-lpl/in
'If its creatures had been near, we would have caught it [about a peccary].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
256 Word classes

b)
d^a siiri-tgau hananiane nix itga-re
what have-lsg/D if that do-IRR:3ps/E
'He would do that if I had what?'

As mentioned at the beginning of this section, all conjunctions discussed in the

above sections co-occur with the subordinate marker =ne, which can also be
omitted. =ne can also occur without being accompanied by a conjunction and
even as such it has a range of meanings. For instance, it can be used with
temporal function or as a conditional marker. These issues will be discussed in
detail in §20.1.1 on subordinate clauses without conjunctions.
Two further conjunctions, rihihei and toknanei, were already introduced in
§5.4.6. These occur with a comparative function and are derived from verbs,
whereas all other conjunctions are underived.

5.8.3 Coordinating conjunctions

tQäe 'also'
There is only one monomorphemic word that can have coordinating function
between two clauses, t$äe 'also'. Its function is restricted to the coordination of
clauses or sentences. An alternative translation of tgäe is 'besides'. It generally
occurs in initial position of the second clause, but deviation from this may
occur. A few peculiarities regarding its syntactic behaviour are discussed in
chapter 20.5 on coordination. One might also argue that t$äe is an adverb that
operates on the clause level. However, there also are reasons that favour its
treatment as a multifunctional word which includes its function as a
conjunction: firstly, its typical occurrence between clauses or sentences, which
is untypical for all other adverbs (also cf. §18.4.2) and, secondly, its function to
relate information given in separate clauses or sentences, rather than modify a
verb.

(379) Occurrence of tgäe

a)
bute haa-u kuane eno-a akau=ne hoataa-ra
boat make-IMP inside enter-3ps/D water=SUB so.that.not=EMF

tQäe i-=hiniitQäi lauri+lauri ne-nakauru eno-a-u

also 2sg=together RED+group be-those.who enter-CAU 1-IMP
'Make a boat so that the water will not enter! Besides, make all kinds of living things
enter!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Conjunctions 257

b)
nil hau hetau=te, t$äe ßäe ne-molo-a raj misi=ße
that because HRS=FOC also already ITR-cut-3ps/D POSS umbilical.cord=CND

amu-a ku-a na-iße

walk-NTR go-3ps/A say-JUSS
'Therefore, also, when his [the baby's] umbilical cord already falls off, she should say
that he (her husband) goes hunting.'

It is common for tgäe to be combined with an (optional) focus marker and an

emphatic marker, such as in tgäe (te) tQU. While it may be difficult to express
the difference in English, this construction implies an emphatic use o f ' b e s i d e s ' ,
or 'on top of it' (cf. §5.15 on particles).

nete 'but'
nete is a clause coordinating conjunction for 'but', focussing on
counterexpectation. It conjoins two clauses or sentences.

(380) Use of nete

a)
m=a=lu enua kauatQa nete ßäe enene ne-ene
be-3ps/A=REM tree good but already nowadays be-NEG:3ps/E
'There was good wood, but now there isn't any more.'

b)
teru siiri-a-ü nete kann iteasu=te siiri
axe have-NEG-lsg/E but lsg brother=FOC have:3ps/E
Ί don't have an axe, but my brother has [one].'

Interestingly, nete can be interpreted as a composition of the subordinate marker

=ne plus the focus marker =te. This is illustrated in example (381), where nete
(cross-referencing a 2ps or 3ps subject) is in free variation with nene, which
contains the lsg cross-referencing focus marker =ne. However, it should be
noted that the variant nene is less common and tends to be replaced by nete for
good. The same applies to the form nena, which involves the 1 pi focus marker
=na.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
258 Word classes

(381) Variation between nete and nene

a) Variant nene:
aum-kau, ne+ne hoara-ene-u
hear-1 sg/A SUB+FOC:lsg see-NEG-1 sg/E
Ί heard it, but I did not see it.'

b) Variant nete:
auna-käu, ne+te hoara-ene-ü
hear-1 sg/A SUB+FOC see-NEG-1 sg/E
Ί heard it, but I did not see it.'

There is another reason which suggests that nete is on its way to being
lexicalised (which is why it is glossed as monomorphemic 'but' elsewhere).
Firstly, a verb that is preceded by a focus marker (such as =te) always occurs in
the Ε-form (cf. §11.2). However, this is not obligatory when nete precedes the
verb. The Ε-form can be chosen as well as the Α-form, which shows that the
grammatical peculiarities of the focus marker as a component of nete have been
neutralised.

(382) Conjunction nete with Α-form or E-form

nete hoara-ene-ü / huara-?a-ü

but see-NEG-1 sg/E see-NEG-1 sg/A
'But I did not see it.'

It can be concluded that the conjunction nete is in transition to being lexicalised.

While it cannot be described as genuinely monomorphemic, yet its use as
described in (381) suggests that it may soon reach this status.

5.9 Interrogatives

Urarina has only one basic interrogative of which all others are derived. d$a is
the word for 'what/who', which is combined with other morphemes in order to
express other interrogative concepts. The component d$a surfaces in all these
interrogatives except in the word for 'where'. However, it is plausible to assume
that d$u is a fossilised form of d$a plus the locative marker -u. The loss of /a/
would thus be due to lexicalisation. The list in (383) contains the most
important combinations of d$a with other elements. A complete analysis is
given in §21.1.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clause introducers 259

(383) Interrogatives

'what/who'
d^a-nu 'why' (with PURP)
d$a-toane-i 'how' (with verbal DER)
d^a-bana 'when' (with postposition 'when')
d$a-elöo 'what(ever) way' (with DIR)
dyu 'where' (with LOC)

As a general rule, the position of all interrogatives is clause initial. An important

aspect is that there is no distinction with regard to animacy, as will be further
discussed in §21.1.1.
The terms for 'how much' ßatahaa and ßaaohwaa do not belong to this
class. The behaviour of these suggests that they are verbs, in a similar way as
the native numerals discussed in §5.12. A detailed discussion of their function is
given in §21.1.10, while (384) illustrates their use in two examples.

(384) Interrogative verb ßatahaa and ßaaohioaa 'how much'

a) ßatahaa with person suffix:

ßataha-a kure - sauta ni-a
be.how.much-3ps/A exchange five be-3ps/A
'Exactly how much is it/does it cost?'- 'Five.' [Example from Manus 2002, p.c.]

b) ßaaohwaa with participle suffix:

ßaaohwa-Ί arusu here-i
be.how.much-PRT rice want-2ps
'How much rice do you want?'

5.10 Clause introducers

Urarina has a small class of words that function as clause-initial elements to

indicate clause type and other categories, such as negation or mood. Their
presence is in addition to verbal morphology. The fact that their syntactic
position is specified and that most of them require a particular type of inflection
on the verb or elsewhere in the clause suggests that these are a separate word
class, even though this may be typologically unusual. The properties these
words have in common involve their occurrence in clause-initial position and
the co-occurrence with a specific inflection of clause type.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
260 Word classes

(385) Clause introducers

rdhjauria, hjauijie "Strong" prohibitives (require NEGF clause enclitic)

ku>a "Weak" prohibitive (requires NTR inflection on verb)
kwatia Negative (optional; emphasis, with negative inflection)
huane Hortative (with HORT inflection)
nab ana Apprehensive (optional; always co-occurring with hoataa 'so
that not')
baana Irrealis / Conditional (optional; with conditional subordinate clause)
ta Negative question (with final negative marker)
d^atera Suggestive clause (with special inflection)

In the following sections, the use of each of the introducers is illustrated.

5.10.1 "Strong"prohibitives "nihjauria", "hjauiße" 'don't'

The clause introducers nihjauria and hjauijie (or variant βaauipe) 'don't'
function as "strong" prohibitive introducers (cf. §15.2.1). Their presence is
associated with the clause final negative enclitic =ne. The verb is also marked
for 2ps (-2) when any of the two are used as a direct command. Subsequently, in
the plural, the verb carries suffixes for 2ps and plural, followed by the final
enclitic. The final emphatic enclitic =ra can follow in addition and is attached
to =ne. Also note that nihjauria and hjauiße occur in sentence-initial position
and that a subordinate clause may occur between the introducer and the main
verb, such as in (386b) and (387b) where a participle clause occurs in this
position, and in (387a) where the intervening subordinate clause has a
temporal/conditional function.

(386) Introducer nihjauria

a) Singular:
nihjauria kaa temule ki-i-ße
don't this plant eat-2ps=NEGF
'Don't eat this plant!'

b) Plural:
nihjauria huajtei nakiuaaunel ne-tabataka-he-1 turu-i-tQe=ne=ra
don't again again ITR-reduce-CNT-PRT arrive-2ps-PL=NEGF=EMF
'Don't (PL) return again being reduced in number!'

The example in (387c) even exhibits a rare double use of the final negative
marker, occurring after the verb and after the adverb that follows the verb.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clause introducers 261

Possibly, hauria is meant as juxtaposition to the clause and is marked with the
enclitic in order to indicate that it is still in the scope of the negation (otherwise
one would expect the marker to occur only once - after the adverb).

(387) Use of mhjauria with 2sg form

a)
nihjauria kati eru-i-ße, saate-i=ße
don't black.monkey meet-2ps=CND kill-2ps=NEGF
'When you meet a black monkey, do not kill him!'

b)
mhjauria kanu kururu-1 ka=kiuara-i-ße-ra
don't lsg hide-PRT lsg=see-2ps=NEGF-EMF
'Don't watch me [by] hiding yourself.'

c)
nihjauria be-i-ße hauria=ne
don't tell-2ps=NEGF first=NEGF
'Don't tell anyone at first [but only later]!'

Prohibitive introducers are also used with 3ps subjects. These take the regular
suffix for 3ps/A in this context. The final negative enclitic =?ie is attached to the
end of the clause.

(388) Use o f nihjauria with 3ps form

a)
nihjauria tQäe kati saate-nana-a=nel
don't also black.monkey kill-ILT-3ps/A=NEGF
'His compadre must not kill black monkeys!'

b)
nihjauria nii rihihei eno-ana-a=ne
don't that like enter-ILT-3ps/A=NEGF
'His compadre must not enter like that!'

Regarding the origin of mhjauria, it may be suspected that it is the fossilised

irrealis form of a verb (though the exact meaning of its root is unclear). In the
NT translation, various forms of a verb root /nihjau/ meaning 'die; be thrown
away' are attested, as in mhjau-re=l (IRR:3ps/E=ASS; found in Matthew 5:21),
mhjau-i-tQe-ne (2ps-PL-NEG:3ps/E; Matthew 5:30), and rnhjau-i=jiehoataa
(2ps=SUB 'so that not'; Matthew 18:8). Thus, a literal translation of mhjauria
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
262 Word classes

would be '(he) would die', which might allow a semantic connection to a

prohibitive in the sense that "something would happen if you do it". This would
also explain the fact that nihjauria as a "strong" prohibitor. The origin of
ßaaoiße is not known.

5.10.2 "Weak"prohibitive "kwa " 'don't'

In addition to the prohibitive introducers discussed above, hua functions as a

"weak" prohibitive (investigated in detail in §15.2.2). The use of hua requires
neutral marking on the verb with all persons. 34

(389) Use of hua with 2sg form

a)
lava kuruata-sa-1 ti-a
don't two-times-PRT give-NTR
'Don't tell it twice!'

b)
hua hjä=ra uba-hei ajto-a=ra
don't just=EMF be.angry-PRT say-NTR=EMF
'Don't speak in anger!'

Plural is marked on the verb, as well as the polite form (which is identical to the
plural for 2ps and can only be distinguished by the context). In both cases, /tpe/
(which is a suffix for plural marking but an enclitic for politeness marking) is
suffixed to the verb (also cf. § 12.2.16.1).

(390) Use of hua with 2pl form

a) Plural:
hua ubaae-ka-tge
don't be.angry-NTR-PL
'Don't (PL) get angry!'

34
This statement is hypothetical as no examples for 3ps are attested. However, the fact
that 2ps is not explicitly marked indicates that person marking is neutralised in this
form.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clause introducers 263

b) Plural:
kwa ka=su-a-t$e=ra
don't 1 sg=kill-PL=EMF
'Don't (PL) kill me!'

c) Polite:
kwa bi-a=t$e
don't tell-NTR=PLT
'Don't tell him!' [Polite; man talking to woman]

5.10.3 Negative "kwatia" 'not'

Evidently related to Jaua through morphology is the introducer huatia (whose

literal meaning is unclear). It is used optionally to emphasise a negative clause
(in which the verb is inflected regularly according to the paradigm for
negation). This results in the unusual situation of a prohibitive being the
unmarked form and the "simple" negative occurring as the marked form on the
verb, huatia can be further emphasised by the insertion of nijej 'not at all'. In
the same way (but with a pause) it can be preceded by an expression for ' n o ' ,
such as mtohuara, or hjauijiera.

(391) Use of hoatia

a) hoatia by itself:
ka=hitQihi ke ni-a=ne hoatia kauatg,a-ri-ji=ta
this=side VLI be-3ps/D=CND not good-IRR-NEG:3ps/A=FRS
'If it is in this manner, it would not be good!'

b) With nijej:
mjei hoatia hoeru-i=lu ajtQune kuane
not.at.all not be.clear-NEG:3ps/A=REM PLN inside
'The Espejo (river) was not cleared at all.'

c) With hjauijiera:
hjauijiera, hoatia ne-hi-ßi kaa ßaara atane
no not be-DIM-NEG:3ps/A this 2pl land
'No! Your world is not big.' [Lit. 'your world is not anything']

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
264 Word classes

d) With nitokioara:
nitohoara, hoatia i=tQoara-ri-a-u=ni-ta
no not 2sg=see-IRR-NEG-1 sg/E=ASS=FRS
'No! I won't watch you.'

Note that there is another word related to the negative, which is the conjunction
kiuataa 'so that not', mentioned earlier (cf. §5.8.2). A further introducer that
resembles kwa at least phonologically, is the hortative introducer huane, though
it is not related to negation.

5.10.4 Hortative "kwane " 'let's'

The hortative introducer is frequently (but optionally) followed by a focus

marker (which is =ne/=na for lps and =te for all other persons, cf. §19.2.1).
Kzuane can also be accompanied by the attitudinal enclitic =ra in the traditional
language (cf. (392b)). The verb is inflected with the hortative suffix -e for lsg.
While the English translation of huane as 'let's' may imply that the speaker
is talking to others, this is not the connotation in Urarina, as the hortative in the
lsg is a kind of self-encouragement. In (392b), for instance, Adam, who in this
particular narrative is the only man in the world, is talking to himself.

(392) kcuane with hortative form -e for lsg

a)
huane ra-e
let's receive-HORT
'Let me get it!' [Said by character who is about to grab an axe]

b)
kioane=na kau kuhwa-~i huara-e=ra
let's=FOC:lsg here wait-PRT see-HORT=EMF
'Here I will wait and see' [Said by character who is hiding in order to watch what will
happen]

Examples that exhibit the form -ni represent the result of merging between the
distal suffix -ni and the hortative -e (cf. §15.1.2).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Clause introducers 265

(393) kioane with distal form -ni for lsg

a)
hoane hoaa-ni-0
let's see-DSTL:HORT
'Let me go to see!'

b)
huane hauria ahajna-ana ku-ni-0-tga
let's first forest-inside go-DSTL:HORT-EMF
'Let me go into the bush first!'

In the plural, a listener is involved, and there is a further distinction between

lpl/ex -akaanu, dual -aka, and inclusive plural -akatge (or variants -akaatge,
-tQaatoe). Again, the distal suffix -ni may occur on the verb to express a
movement away from the present location ('let's go fetch it'; also cf. §15.1.2).

(394) kioane with lpl forms

a) With dual form -aka:
hoane=te i=tgej ate ti-aka
let's=FOC 2sg-for fish give-lpl/du
'Let's give you a fish'

b) With lpl/in form -akatg,e:

hvane=te raa-ni-tgaatQe
let's=FOC receive-DSTL-lpl/in
'Let's go fetch it!'

5.10.5 Apprehensive "nabana" 'so that not'

Nabana is an adversative introducer used in initial position of a subordinate

clause. It always co-occurs with the subordinate conjunction huataa. In the
same way as some other introducers, its use is optional.

(395) Use of nabana

a)
nabana hvena-kuru-a=ne kioataa enuhe tauhja-1 ... ku-ure
that.not laugh-PL-3ps/D=SUBso.that.not pepper gather-PRT go-3pl/E
'They went,... putting peppers in their mouths so that they would not laugh.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
266 Word classes

b)
nabana aheri ke ne-kunajtaa-i=ße huataa
that.not stone INST ITR-hurt-2ps=SUB so.that.not
'so that you don't get hurt by a stone' [NT: Luke 4:11]

5.10.6 Conditional "baana" 'if'

Another optional introducer is baana, which occurs with conditional clauses and
reinforces the condition ' i f . It is comparatively rare and only attested in very
few examples, baana co-occurs with the subordinate enclitic =ne when this is
used as a conditional marker, or with the word hananiane ' i f (cf. §20.1.1.3).

(396) Use of baana

a)
baana ii here-i=ße hananiane kanu kuruatahane-i
if 2sg want-2ps=CND if lsg help-2ps
'If you want, you can help me.'

b)
baana ßaara here-ki-tpe=ne hananiane kanaka ana reroa-ri-i-toe
if 2pl want-2ps-PL=CND if lpl/ex replace-IRR-2ps-PL
'If you (PL) wanted, you (PL) would replace us.'

5.10.7 Negative question introducer "ta"

Another introducer is the obligatory marker for negative questions, ta. It is

always accompanied by the clause-final clitic =ne. However, further enclitics
such as the frustrative =ta or the emphatic marker =ra can be attached to it (in
the traditional language). Note that the inflection on the verb follows the
affirmative paradigm, as negation is only expressed by ta (cf. §14.5; §21.3).

(397) Negative question introducer ta

a)
ta u-ri-ki- ße
NEGQ come-IRR-2ps=NEGF
'Won't you come?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Quantifiers 267

b)
ta ate ra-uru-a katQa-uru=ne=ta
NEGQ fish receive-PL-3ps/A man-PL=NEGF=FRS
'Have the people not caught any fish?'

5.10.8 Suggestive "d$atera"

Suggestive clauses are introduced by d^atera, which can be roughly translated

as ' h o w i f , 'what i f , or 'what about . . . ' (cf. §15.5). The presence of this
introducer coincides with a special person marker for l s g and with the jussive
form for 3ps subjects (other forms are unknown).

(398) Suggestive clause with d^atera

a)
d^atera läatpa tur-a-riu
how.about ship arrive-CAUl-SUG: lsg
'How [nice would it be] if I made the ship arrive!'

b)
d$atera hitarn kau tete-riu
how.about all here make-SUG:lsg
'What if I put all [those things] here?' [Talking about items that some one else had
promised.]

5.11 Quantifiers

Most (if not all) Urarina quantifiers are derived or belong to other word classes:
while the majority can be analysed as fossilised participle forms of verbs, others
resemble nouns with regard to their syntactic and morphological properties. It is
not always possible to identify the original base of each quantifier, but their
morphological complexity is always evident from at least one known
component. In (399), a list of quantifiers is given, followed by a detailed
description for each. The list also includes some indication of the known parts.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
268 Word classes

(399) Quantifiers

itulere 'all kinds (of)'

lauri+lauri 'various (groups of)' (RED+'group'/'kind')
taba-uru 'some' ('be big'-PL)
taba-ϊ 'much' ('be big'-PRT)
kuruatahe-i 'little' ('two'-PRT)
lejhii- ße-hi-1 'not even one' ('one'-'be'-CNT-PRT)
arai, ara-hii 'many', 'various' (cf. arai 'family/tribe'; arajsat 'several times')
hita-t, hita-ri-i 'all' ( l finish'-PRT/-RAP-PRT)
sati-i 'all' ('end'-INTS-PRT)
d^n-ni 'nobody' ('where' + unknown morpheme)

5.11.1 Quantifier "itulere" 'various (kinds of) '

T h e w o r d itulere (or abbreviated form tulere) is found as a prenominal modifier,

but it also functions as a noun. In this context, it is used as a nominal head and
can take a noun plural (cf. (400a,b)). Its semantic content also refers to
' e v e r y t h i n g ' , 'all kinds of things'.

(400) Use of itulere

a) As head of NP:
bihi-tga ke=te itulere itQa-kure=lu katQa-uru
hand-only INST=FOC all.kinds do-3pl/E=REM man-PL
'The people did everything with their hands.'

b) As head of NP with plural:

atane huu-ka akau=ne hana itulere-uru ne-rutu-hahoa-e
land flood-3ps/D water=SUB when all.kinds=PL ITR-transform-DSTR-3ps/E
'When the water flooded the earth, everything was transformed.'

c) As modifier:
m-a itulere rukuele kioata kahe it^a-noi
be-3ps/A all.kinds things metal from make-PASS
'There were all kinds of things made of iron'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Quantifiers 269

5.11.2 Quantifier "laurilauri" 'various'

This quantifier is a reduplication of the word for ' g r o u p ' or ' k i n d ' , lauri, which
is also used on its own, sometimes together with the short form of the numeral
f o r ' o n e ' (as in example (401c)). It can function both as a nominal head and as a
modifier, occurring in prenominal or postnominal position. W h e n it functions as
head of N P , its reduplication refers to plurality ('all groups'). However, note
that reduplication is not in general a productive strategy for pluralisation (cf.
§10.3).

(401) Use of laurilauri

a) Head of NP:
lauri+lauri kahe kuraanaa beene naaohwa-i-tQa
RED+group from chief female ENUM-PRT-only
'From the various (species) [choose] a male and a female!'

b) Head of NP:
lauri+lauri kahe u-nakauru
RED+group from come-those.who
'those who come from all the [other] communities'

c) Prenominal modifier:
nitoanei hetau=te edara ne-1 amu-ure=lu le=lauri
like.that HRS=FOC water.people be-PRT walk-3pl/E=REM one=group

katQa arahtt
man many
Ά large group of people was transformed into water people like that.'

d) Postnominal modifier:
lenone lauri+lauri
food RED+group
'various types of food' [i.e. of all kinds, from different places]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
270 Word classes

5.11.3 Quantifier "tabauru" 'some'

A form derived from tabaa 'be big' is used as a quantifier for 'some (people)',
used with humans only. This term is formed by suffixation of the noun plural
-uru to the (otherwise verbal) root. Tabauru is used as a nominal head with
quantifying function. In addition, it also appears to modify a noun, as examples
from the NT show. However, an alternative analysis of the modifier
construction in (402b) would be to describe tabauru as a nominal head of a
possessive construction 'some of the women'. This would make sense in so far
as tabauru never occurs as a modifier in prenominal position, whereas most
other quantifiers are variable in their position.

(402) Use of tabauru

a) Head of NP:
tabauru=te niki hoituku-ure nete ßäe it$a-kur-ene
some=FOC ADVRS know-3pl/E but already do-PL-NEG:3p/E
'Some people know; but they do not make any more [the mixture].'

b) Modifier:
eene-kuru tabauru
woman-PL some
'some (of the) women' [NT: Mark 15:40]

5.11.4 Quantifier "tabai" 'much', 'enough'

The participle form of the verb tabaa 'be big' is used to refer to 'much, a lot' in
an uncountable sense. Its position can either be preverbal or postverbal. Note
that it is used with adverbial function and does not modify a noun.

(403) Use of tabai

a)
tabai hja-a häu hetau=te
much urinate-3ps/D because HRS=FOC
'as she urinated very much'

b)
tabäi=te ne inio aräala su-uru-a=ne
much=FOC be:3ps/E meat tapir kill-PL-3ps/D=CND
'There was meat in abundance when they killed the tapirs.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Quantifiers 271

5.11.5 Quantifier "kuruatahei" 'littlefew'

Apart from the verb äatQihia 'be little' (cf. (280)), the expression kuruatahe-1 is
used to refer to small quantities. Literally, it is the participle form of the verb for
' t w o ' , thus 'being two'.

(404) Use of kuruatahei

nuse kahe=te ne-lu kuruatahei kauatQa

town from=FOC be:3ps/E=REM little good
'There were few good things from town.'

5.11.6 Quantifier "lejhijiehei" 'not even one'

T h e expression lejKi-ße-he-i (or variant lejhlßehü) 'not even o n e ' is composed

of the numeral for 'one' plus the continuous participle form of the copula. Since
lejtiiflehel is always associated with negation on the verb, is can be translated
literally as "not being one". It is not used as a noun modifier, but has adverbial
function. However, in example (405b) (and in several examples from the NT), it
may also be interpreted as a nominal head.

(405) Use of lejhlßehet

a)
nii häu=te lejhlßehü kauti-a=ne here-kur-ene
that because=FOC not.even.one remain-3ps/D=SUB want-PL-NEG:3ps/E

ißono ho-era ke
ayahuasca drink-AG VL1
'He wants that not even one remains who does not drink ayahuasca.'

b)
kahjune hja-naa ißa-uru-i, kauatga+najßehel, lejhlßehü
cloth wear-NOM know-PL-NEG:3ps/A pretty.nothing not.even.one
'They did not know anything about wearing clothes, not even one of them.'

5.11.7 Quantifier "arahii" 'various', 'several', 'many'

T h e quantifier arahii (and its variant aral) occurs as a quantifier, but, due to its
ending in the participle suffix, can also function as an adverb. Regarding its
phonological shape, one could speculate that arahii is related to the noun arai

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
272 Word classes

'family/tribe/clan'. When arahii occurs as a quantifier, it can be in either

prenominal or postnominal position. There is no distinction between countable
and uncountable items, since arahii can occur with both. The quantifier also
occurs in combination with the suffix for "times": araj-sai means 'various
times'. It can be translated as 'many', 'several', or 'various', as speakers vary in
their description of what amount or quantity is involved.

(406) Occurrence of arahii

a) Quantifier: prenominal position:

u-akwa-a arahii katQa-uru
die-DSTR-3ps/A several man-PL
'Several people died.'

b) Quantifier: postnominal position:

ku ne-rehete-kure katga-uru arahii
there be-HABl-3pl/E man-PL several
'Many people used to live there.'

When arahii functions as an adverb, it typically occurs in initial position. It can

be followed by a focus marker and a noun as in (407a) or by the verb, which in
turn is followed by the subject (cf. (407b).

(407) Quantifier arahn with adverbial function

a) Followed by noun:
arahn=te fioanara ti-ahua-ure
several=FOC banana give-DSTR-3pI/E
'They gave lots of bananas [to me].'

b) Followed by verb:
arahu=te ne hanolari ßäe amu-aka=ne
many=FOC be:3ps/E jaguar already walk-lpl/du=CND
'There are many jaguars [around], when we walk.'

5.11.8 Quantifiers "hitaru", "satu" 'all'

Urarina has two quantifiers for 'all', hitaru and satu, which both can function
as nominal heads and as noun modifiers. As head of an NP, satu refers to
humans, while the use of hitaru in this context is restricted to non-human
referents. This distinction appears to be neutralised when these quantifiers
function as noun modifiers (cf. (410)).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Quantifiers 273

(408) hitarii and satn as head of NP

a) hitarii with non-human reference:

hitarü ka=raj kajßa-ku-tQe=ra
all lsg=for give.back-IMP-PL=EMF
'Return (PL) everything to me!'

b) satn with human reference:

ißono ku-a lejhii itcauena=ne=te satii ku-uru-a nerajne
ayahuasca drink-3ps/D one elder=CND=FOC all drink-PL-3ps/D must
'When one elder drinks ayahuasca, all have to drink it.'

The example in (409) shows that satn can be combined with the privative suffix
for 'without', which is normally attached to verbs. This supports the view that
both quantifiers for 'all' are fossilised forms of verbs.

(409) satii plus privative suffix

mtoanei n=ukuana ha-ure ßäe tabauru sati-elanaala

like.that 3ps=field make-3pI/E already some all-PRV
'Some make their fields like that, but not all.'

The word hitarn is derived from the transitive verb hitaa 'finish'. It has the
alternative form hita-1, which is the participle form of hitaa, while hitarii
consists of hitaa plus the participle form of the suffix -ri for velocity. The origin
of satn is can be analysed as derived form of intransitive verb saa ' e n d ' , to
which the participle form of the intensifier -toa is attached. This is also
confirmed by the fact that, similar to hitarii/hitai, the quantifier satii has a
variant, sa-1, which again indicates a morphologically complex structure, at
least diachronically.
The semantic difference between hitarii and satii is difficult to determine as
both are used in similar contexts. Both words occur in prenominal position
when used as a modifier. However, there is a slight difference with regard to
their preference: obviously related to the literal meaning of hitarii or hitai,
which is derived from the transitive verb for 'finish', these quantifiers are used
to express that something is taken away, is consumed or finished. In contrast,
satii does not imply this meaning.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
274 Word classes

(410) hitarü and satii in modifier function

a) hitarii with human reference:

hauria hitarü letQurjka-j katga rela-i=ße reetial
first all ten-NOMsbj man teach-2ps=SUB until
'until you have taught all ten people'

b) hitarii with non-human reference:

hitarii nii temule-kuru kivaae
all that plant-PL power
'the power of all those plants.'

c) satii with human reference:

satii enamanaa-kuru=te kana kuruatahane-re=l
all young.man-PL=FOC 1 pl/in he!p-IRR:3ps/E=ASS
'All the young people will help us.'

d) satii with non-human reference:

satii temule kauatpa
all plant good
'all good plants' [NT: Matthew 7:17]

The word d^uni listed in (399) will be discussed in §6.4 and §21.1.11. The
following overview summarises the different functions of the quantifiers
discussed in this section. One distinction is that some quantifiers may function
as head of an NP, but others not. A further distinctive property regards their
function and position as modifiers. In addition, most quantifiers have other
specific features that does not apply to others. More details about the position of
each quantifier are found in §6.4.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Numerals 275

(411) Summary of quantifier features

Quantifier Nominal head Modifier Other features

tabauru 'some' yes no humans only
kuruatahei 'little' unattested prenominal adverbial
tabal 'much' no: adverb no uncountables
lejhu jiehi 'not even one' yes no with NEG
itulere 'all kinds (of)' yes prenominal can take PL
satii 'all' yes prenominal cf. 'end' (intr.)
laurilauri 'various' yes prenominal cf. lauri 'group'
preferred
aral, arahn 'many', no prenominal also adverbial
'various' preferred
hitai, hitarii 'all' yes prenominal cf. 'finish'
preferred (trans.)
d^um 'nobody' yes prenominal (see §21.1.11)

5.12 Numerals

Urarina numerals represent a mixed class that can be divided into native and
borrowed forms. The native numbers include lexemes from 'one' to 'five',
while the other numerals are loans from Quechua. This includes words for up to
'thousand', while 'million' is a loan from Spanish. The two types of numerals
further differ from each other in that the native numerals are verbs, whereas the
loans are to be analysed as nouns.

(412) Urarina numerals

Urarina Quechua
1 lejhia (huk)
2 kuruata(ha)a (ishkay)
3 nitpatahaa (kimsa)
4 heena (chusku)
5 saukia (pichqa)
6 sauta soqta
7 kasi, käasi qanchis
8 hoosa, fiooosa pusaq
9 isköo isqon
10 (le=)tg^ka chunka
100 (le=)fiuaafa pachak
1000 (le=)iuarar]ka luaranqa

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
276 Word classes

The word for 'million', arguably a very recent borrowing, is (le=)midyi '(one)
million', is based on Spanish millon.
The term for 'one' can be described as a combination of the root /le/ with the
participle form of the continuous suffix (surfacing as -hü here). The problems
regarding the status of le= being a form "between" clitic and root were
discussed in §2.11. The word for 'two' is related to 'help', kuruatahania.

(413) Occurrence of short form le= 'one'

a) With numeral:
le-fioaaja
one=hundred
'one hundred'

b) With noun:
le=hitQu
one=day
'one day'

Interestingly, native and borrowed numerals reflect syntactic and morphological

differences: while the native numbers 1-5 are a subtype of verbs, the borrowed
forms 6-9 are nouns. This corresponds to the status of numerals in the source
language Quechua. The morphological difference between the two types
becomes evident through their use in context. Native numerals are inflected like
verbs, which is illustrated by the 3sg marking suffix in (414a). Literally, this
example means 'Three are his wives'.

(414) Verbal function of numerals

a)
nitQataha-a raj komasaj
three-3ps/A POSS wife
'He has three wives.'

b)
kuruataha-a kanu nuta
two-3ps/A lsg eye
Ί have two eyes.'

When used as modifiers, the nominalising suffix -i is attached to native

numerals. Alternatively, they can be inflected with the participle suffix -7.
Typically, numeral modifiers occur in prenominal position.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Numerals 2ΊΊ

(415) Modifier function of numerals: with nominalising suffix

a)
ni-a=lu lejhi-i katga kauatQa-i
be-3ps/A=REM one-NOM Sbj man good-PRT
'There was one good man'

b)
lejhii eene=te hana nitQataha-j fiuanara siitQa-i laueke
one woman=FOC instead three-NOM Sb j banana hold-PRT be.sitting:3ps/E
'One woman instead is sitting, holding three bananas.'

c)
le=tQurjka-j katga-uru=te kaa ita kuruatahane-ϊ amiane
one=ten-NOM Sb j man=PL=FOC this REC help-PRT work:3ps/E
'Ten people are working [here], helping each other.'

Note that the borrowed words for ' t e n ' , ' h u n d r e d ' , and 'thousand' are treated
like verbs with respect to their occurrence as modifiers, as they take the
nominalising suffix (cf. (415c)). However, they are not attested as full verbs as
the native numerals are. Thus, they are not used in the same way as the
examples in (414).
T h e borrowed numerals 6-9 exhibit a morphological and syntactic structure
that is different from the verbal numerals, since they behave like nouns. In order
to modify another noun the participle form of copular auxiliary nei must be
cliticised to the numeral. This auxiliary can also be used in the continuous form,
as example (416c) illustrates.

(416) Borrowed numerals as modifiers

a)
käsi=ße~i kuraanaa käsi=ße~t beene
seven=AUX-PRT chief seven=AUX-PRT female
'seven males and seven females'

b)
isköo=ne~i katQa
nine=AUX-PRT man
'nine people'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
278 Word classes

c)
iskoo=ne-he~i atauari kurete
nine=AUX-CNT-PRT chicken buy:3ps/E
'He bought nine chickens.'

In combinations of single numerals to complex ones, the decimal system is

employed. In adding numbers, the highest number is in initial position, followed
by a smaller one ("10+5"). In multiplications, the multiplier precedes the
multiplied.

(417) Combined numerals

'20' kurua[ta]+tQuijka "2 χ 10"

'80' hoosa+tguqka "8 χ 10"
'200' kuruata+fiuaa/a "2 χ 100"
ΊΓ le=tQuqka+lejhia "1 χ 10+ 1"
'16' le=tQuqka+sauta "1 χ 10 + 6"
'10,000' le=tQuqka+waraqka "1 χ 10 χ 1000"
'100,000' le=fivaa fa+iuararjka "1 χ 100 χ 1000"
'2 Γ kurua+tQuqka+lejhia "2 χ 10+ 1"
Ί56' le=fivaafa+sauki+tQuyka+sauta "(1 χ 100) + (5 χ 10)+ 6"
'294' kuruata+fiuaafa+iskoo+tGuqka+heena "(2 χ 100) + (9 χ 10) + 4"

The inflection in noun modifier constructions depends on the final part of the
compound: if the last component is of the native or verbal type, the compound
will follow the verbal morphological pattern; if the final element is between six
and nine, the whole word will be treated like a noun.

(418) Combined numerals as noun modifiers

a) Verbal type: final numeral is 'two'

le=tQU[jka+kuruataha-j katQa
1 = 10+2-NOMsbj man
'twelve people'

b) Loans: final numeral is 'six'

kuruata+fioaafa+käsi+tQuyka+sauta=w-~i nii bote kuane ni-akaanu
2+100+7+ 10+6=AUX-PRT that boat inside be-lpl/ex
'We were 276 [people] in the ship.' [NT: Acts 27:37]

N o u n s that refer to a temporal unit such as ' d a y ' , ' w e e k ' , ' m o n t h ' , or ' y e a r ' can
be combined with the root of a verbal numeral. In this case, all inflectional

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Numerals 279

suffixes (which would normally occur on the verbal numeral) are attached to the
end of the word, as exemplified in (419a), where the numeral root for ' o n e ' is
cliticised to the noun for ' m o n t h ' : in this case the participle marker is suffixed
to the final part ( ' m o n t h ' ) . O n e could say that through the combination of the
noun with a numeral verb a compound is formed which itself b e c o m e s a verb.
This is also shown by the example in (419c), where the verb ' b e four years of
a g e ' is created. 3 5

(419) Compound formation with verbal numerals

a)
le=ate-i
one=month-PRT
'one month'

b)
ku neda-i nitQata+haanaka-i heräe ku-ahe-ι ßäe rauta-a
there stay-PRT three+year-PRT slowly go-CNT-PRT already like-3ps/A
'Staying there for three years, he slowly started to like it.'

c)
heena+haanuka-a kanu kalaui
four+year-3ps/A lsg son
'My son is four years old.'

W h e n temporal nouns are combined with borrowed numerals, a similar situation

applies: the auxiliary follows the noun rather than being attached to the
numeral.

(420) Compound formation with borrowed numerals

a)
kasi+haanaka=ne-he-~i lana ajjia itQau-e=lu
seven+year=AUX-CNT-PRT husband with live-3ps/E=REM
'She had lived with a husband for seven years.' [NT: Luke 2:36]

35
The word for 'year', ahaana, exhibits a number of peculiarities: in combination with
a numeral, the initial /a/ is elided; there also are the variant haanaka and haanuka. An
alternative approach would be to assume that (a)haana is combined with the
verbalising suffix -(o)ka to form a verb, but this construction is not attested with other
nouns. It is unclear why the example with 'month' should not behave accordingly.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
280 Word classes

b)
sauta+haanuka-m-a
six+year=AUX-3ps/A
'He is six years old.'

The number of times of an action or event is expressed by the suffix -sa, which
is added to the numeral root. Again, this can be followed by the nominaliser -i
or the participle marker -7. The numerals for 'six' and 'seven' (and possibly
others that are not attested here) exhibit a peculiar behaviour as one would
expect the auxiliary form nel to occur after borrowed numerals. However, it is
attested in this form as well as with the nominaliser -i normally found with
verbs only. This could indicate that the distinction between native and borrowed
numerals is on its way to being neutralised.

(421) Counting the number of times

a) With N O M suffix:
kuruata-sa-j
two-times-NOMsbj
'twice'

b) With PRT suffix:

hü a kuruataha-sa-1 ajto-a=ra
don't two-times-PRT say-NTR=EMF
'Don't say it twice!'

c) Variants with 'six':

sauta-sa-j sauta-sa=ne-l
six-times-NOM S b j six-times=AUX-PRT
'six times' 'six times'

d) Variants with 'seven':

kasi-sa-j kasi-sa=ne-~i
seven-times-NOM s b j seven-times-AUX-PRT
'seven times' 'seven times'

Fractions in the arithmetic sense are not a traditional part of Urarina society.
The word katäu 'centre', 'middle' is also used for ' h a l f . In order to measure an
amount or quantity, recently introduced containers such as cups, buckets, etc.
are used. In this case, the suffix -he is attached to the respective noun in order to
express 'a ... o f . As well, the noun is prefixed with the short form for 'one' (no
examples for other quantities are attested). Note that with native words that

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Numerals 281

refer to a measuring unit, such as ' b u n d l e ' or ' h e r d ' , the suffix -he is not used.
In general, -he does not occur with any function related to quantity elsewhere in
the grammar.

(422) Measuring units

a) 'Cup':
le=tasa-he asejte
one=cup-NQF oil
'a cup of oil'

b) 'Plate':
le-plato-hiue ate
one=plate-NQF fish
'a plate of fish'

c) 'Bundle':
le=itQu ra-ι enua
one=bundle receive-PRT wood
'taking a bundle of wood'

d) 'Group':
le=lauri raana
one=group w.l.peccary
'a herd of white-lipped peccaries'

For other kinds of measuring, such as dimension or time, loans from Spanish
(metoro, minuto, oral) are used in combination with the Urarina words for
'length', ' w i d t h ' etc.

(423) Measuring time and size

a)
le=tQuyka metoro ni-a raj kaare
one=ten meter be-3ps/A POSS length
'Its length was 100m.'

b)
harautono=te kuruata+tQugka metoro ne
width=FOC two+ten meter be:3ps/E
'Its width is 20m.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
282 Word classes

c)
hajti kuruata+tQuyka minuto ne-laanala turu hana
still two+ten minute be-PRV arrive when
'20 minutes are missing till he comes.'

It should also be mentioned that the word for 'one' may be used as a kind of
delimitative for 'being the only (thing)', as the following examples illustrate.
Typically, it co-occurs with the suffix -at$a 'only', which is attached to the
delimited noun.

(424) lejhn used in the delimitative sense

a)
lejhii hitQana-atQa kuane=te nesaritQaete-kure katga-uru
one blowgun-only inside=FOC hunt-3pl/E man-PL
'People hunted with a blowgun only.'

b)
lejhti buruari-itga=te kurete-kure=lu nuse kahe
one poison-only=FOC buy-3pl/E=REM town from
'They only bought the poison in town.'

Urarina has no ordinal numbers. Also note that Jebero, which is one of the
languages in the area, has numerals from 1-5, while six and the following are
borrowed from Quechua as well (cf. Tessmann 1930). Tessmann also says that
Urarina has a word for 'six' - [dautani] - which is not attested in my database,
nor is it known to any of my consultants. Instead, it is likely to refer to
sauta nei, which is the borrowed numeral for 'six' followed by the participle
form of the auxiliary. The geographically proximate language Omurana (cf.
§1.3) has numerals from one to ten, which are neither related to Urarina
numerals, nor to Quechua.
There is no word for 'number', as there was no need for it before the
introduction of schools. The teachers nowadays use the Spanish word 'numero'.
The verb haniitQa 'select' is also used for 'count'.

5.13 Interjections

The following list represents a number of interjections that commonly occur in

narratives. It is understood that this list is not exhaustive.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Ideophones 283

(425) Interjections

ah (surprise)
ah-he (fatigue)
ahse (disgust)
auh-heha (surprise+desire)
eehe 'yes'
ha, he, hn (mark begin o f speech)
zva (suddenly happening; disapproval)

Some of these interjections reflect phonological irregularities, such as an initial

labial glide in wa, or the otherwise illicit form hit. Also note that postvocalic /h/
is obligatory in these words, where indicated.

5.14 Ideophones

Ideophones constitute an open class. They include onomatopoeia for various

kinds of sounds and noises. As a matter of fact, these elements are subject to
variation among speakers. In narratives, I also encountered the sound of an
identical object in different passages of the texts being realised in a slightly
different way by the same speaker.

(426) Ideophones: selection

haahaha (laughter)
hau-hau-hau (dog barking)
hii-hii-hii-hii (threatening cry)
hjau (happening easily and quickly)
hwa (falling d o w n )
kuru-kuru-kuru-kuru (bird croaking)
ßaaau (desire and lust)
raae (slowly falling d o w n )
7
raj (falling down to break)
rouu-rajhi (tree falling d o w n )
sehte (sth. falling into water)
tohwoj (sth. falling into water)
tQunenenenee (sizzling)
tiitQarararara (sliding d o w n )
tVi (falling d o w n after sliding)
toho-tohö-tohö-tohö (applying magic)
unee-unee-unee-unee (child crying)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
284 Word classes

The class of ideophones must be regarded an open class since new noises may
be added with each new situation or noise to be described. There also is
considerable variation among speakers with respect to the imitation of noises,
which can result into the production of sounds not otherwise represented in the
phonological inventory (some examples are given in §22.1). Also the number of
times of reduplication is subject to variation.

5.15 Particles

The number of particles that cannot be assigned to any of the word classes
discussed above is relatively small. Nevertheless, their function and meaning
are difficult to define. It seems that all of them imply emphasis of some kind.
These particles are observed in conversations or in direct speech; in narratives
they may be part of quotations of the stories' protagonists. While not much
information about the "emphatic" particles tQU, hjä, and läe is available, their
phonological structure and the fact that they can be attached to words of
different classes might suggest that they are enclitics. However, since speakers
tend to identify all particles as separate words, they clearly differ from clitics,
which were discussed in §2.11. The fact that all particles can be followed by the
enclitic final emphatic =ra does not provide evidence with regard to their status,
as =ra clause enclitic that could attach to any word class in clause-final
position. Thus, there is not enough evidence as to add any particles to the list of
clitics.

5.15.1 Certainty emphasiser "tQU "

The particle tgu typically occurs in combination with other elements, such as
the clause coordinator t$ae, or after a focus marker, such as in aka=te t$u 'he
(certainly)' (cf. (428a)). It puts emphasis on the element it follows, which
makes it likely to be a focus marker, possibly to be translated as 'certainly'.
However, its function is different from the set of focus markers =ne, =na, and
=te, which also function as cross-reference markers (cf. §19). In (427a), tQU
follows the clause coordinator tpäe, which itself is marked with a clitic for
focus; the particle emphasises the fact that magnetic ground will be there in
addition to the other things already mentioned in the context. Similarly, the
occurrence of the particle in (427b), this time without =te, "spices" the utterance
with additional emphasis.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Particles 285

(427) Emphasiser IQU following tgae

a)
tQäe=te tQU ne-re=~i atane maamale-j
also=FOC CRTN be-IRR:3ps/E=ASS land stick-NOM Sbj
'Besides, there will be magnetic ground.'

b)
tQäe tQU hoa eno-a eruari kuane raj kofiua
also CRTN don't enter-NTR mosquito.net inside POSS compadre
'Also, his compadre ought not enter into the mosquito net!'

(428) Emphasiser tgu following words of other classes

a) After pronoun:
aka=te tQU ni-a tenjente+gobernador=ne kujßa projekto
3sg=FOC CRTN be-3ps/D governor=SUB so.that project

itQa-e=lu
do-3ps/E=REM
'[It was] also he [who] developed a project that there be a governor.'

b) After adverb:
esijiae=te tQU aj
really=FOC CRTN AUX:3ps/E
'He really did it.'

c) After verb:
nahari h-i IQU nifo, tako
liver eat-2ps CRTN ADVRS PSN
'Tako, have you [really] eaten liver [which you should not have done]?'

Further examples illustrate that tg,u itself can be followed by an emphatic clitic,
=ra.

(429) Emphasiser tQU followed by =ra

a)
ßäe tQU=ra ne-tabatakaae-kuru-a kanaanaj-uru=ra
already CRTN=EMF ITR-reduce-PL-3ps/A child-PL=EMF
'Already my sons are reduced [in number].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
286 Word classes

b)
he, nii-tQa here-j tQu=ra ka=kiueraj
hey that-only want-NOMobj CRTN=EMF lsg=visitor
'Well, this is what my visitor wanted.'

5.15.2 Emphasiser "hjä" 'just'

This particle has mainly an encouraging character, possibly translatable as

'simply/just' in English; comparable to doch in German geh doch (nur) ("simply
go, since it will be good"). All examples in (430) co-occur with a form of
command, but are "softened" by the presence of hjä.

(430) Emphasiser hjä with commands

a) Self-encouragement:
d^aha-1 hjä ajane
come.on-PRT just yes
'Come on, yes' [he said and stole a piece of shell]

b) With distal imperative:

te-ni-u hjä
give-DSTL-IMP just
'Come on, just go give it to him' [soft command]

c)
turu-ni~u=he=te hjä ii kzvara-kuru-mn=tQa
arrive-DSTL-IMP=REP=FOC just 2sg see-PL-JUSS=EMF
'Just go and arrive there so that they might see you.' [Meant as encouragement towards
priest who is afraid that he would be killed by the natives]

In the following two examples hjä occurs with declarative forms, but could still
be translated as 'just', having a possibly "softening" function.

(431) Emphasiser hjä with declarative forms

a)
ii kauatg,a=nuke=na hjä nekäune itQa-ü=ra
2sg good=PURP=FOC: 1 sg just more do-lsg/E=EMF
36
Ί did it more for your sake [than mine].'

36
The exact function and meaning of nekäune are unclear.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Particles 287

b)
ki-i=ße kujßa=na hjä te-ü=ra
eat-2ps=SUB so.that=FOC: lsg just give-lsg/E=EMF
Ί gave it to you to eat.' [Seeing that he doesn't eat it]

Very c o m m o n is the combination of the two particles tQU and hjä.

(432) Combination of tgu and hjä

a)
ajto-a tQU hjä
say-3ps/A CRTN just
'Yes, he did say so.'

b)
eehe, aj-a tgu hjä kotaeto-a
yes AUX-3ps/A CRTN just be.tiring-3ps/A
'Yes, indeed it is very tiring.' [Why don't you replace us]

c)
aj-i=ta, ßäe t$u hjä=ra hitaru
AUX-NEG:3ps/A=FRS already CRTN just=EMF all

kajpa-kaanu=ra kurura
give.back-lpl/ex=EMF Milord
'It is not so, we [do] have returned everything, Milord.' [Contradicting an accusation]

5.15.3 Contrastive emphasiser "läe"

This particle marks emphasis referring to a contrast to a previous situation.

Usually, this refers to a temporal change, as most occurrences of läe are after
ßäe ' n o w ' or other time-related adverbs. In all examples listed in (433), the
meaning of läe could be translated as ' n o w y e s ' / ' b u t now (as opposed to
earlier)', indicating a change of situation. It is evident that läe follows words of
different classes and types: while most occurrences are in fact with adverbs, the
particle is attested after pronouns, nominalised verbs, focus markers, and
postpositions.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
288 Word classes

(433) Occurrence of läe with temporal reference

a) After ßäe:
ßäe läe ku tutgara-l ßäe ku situ-a akauru
already CHNG there become.straight-PRT already there pass-3ps/A 3pl

ahe
drunkenness
'Already becoming sober there, their drunkenness already passed.' [whereas they did
not feel anything before]

b) After ßäe=ra:
ßäe=ra läe bati-a
already=EMF CHNG dry-3ps/A
'It has dried already.' [after it was flooded for a long time]

c) After adverb ena:

ena läe eno-na najni-tg,aanu
now CHNG enter-INF be.ablelpl/ex
'Now we can enter.' [which was not possible before]

d) After postposition:
he, ka=raj läe hitarü nii kari kajßa-ku-t$e=rä
hey 1 sg=for CHNG all that shell give.back-IMP-PL=EMF
'Now give me all the shells back!' [after you kept refusing to do so]

The examples in (434) illustrate that läe is also used in order to mark a change
of situation that is not time-related. In both examples, it appears to have focus
function, which is consistent with the contrastive function suggested with
temporals above. The sentence in (434a) implies that 'someone else will do it or
has done it'. Example (434b) refers to a monkey that lives on one particular
tree, as opposed to another species of monkey, which moves from tree to tree -
which was the topic of the sentence preceding it. One can conclude from this
that läe is a discourse particle for a change of situation.

(434) Non-temporal use of läe

a)
kanu läe itQa-ri-a-ü=m
lsg CHNG do-IRR-NEG-lsg/E=ASS
Ί will certainly not do it.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Particles 289

b)
aka=te läe helaja-tga ne-lauria-Ί it$au-e=ra
3sg=FOC CHNG same.place-only ITR-sit-PRT live-3ps/E=EMF
'He lives sitting in a single place.'

T h e d i f f e r e n c e b e t w e e n the three " e m p h a t i c " particles m e n t i o n e d in this section

can be f u r t h e r illustrated by a set o f elicited e x a m p l e s that d i f f e r f r o m each other
only by the c h o i c e of the particle. A s s h o w n in (435), these roughly c o n f i r m that
tg,u f o c u s e s on the w o r d it is attached to ('certainly . . . ' ) , hjä has " s o f t e n i n g "
character, and läe m a r k s a c h a n g e o f situation, with the last t w o r e f e r r i n g to the
entire clause. It should also be m e n t i o n e d that t g u d o e s not o c c u r with
imperatives, w h e r e a s the other t w o particles do.

(435) Distinction of particles

a) tQu:
aka=te tgu itga-e
3sg=FOC CRTN do-3ps/E
'Certainly he did it.'

b) hjä:
aka=te hjä itga-e
3sg=FOC just do-3ps/E
'He simply did it.' [e.g. A in an argument between A and B, where Β claims that C did
not do it]

c) läe:
aka=te läe itQa-e
3sg=FOC CHNG do-3ps/E
'Yes, he did it indeed.' [whereas before he was still trying]

W h e n tQU and hjä are c o m b i n e d in this context, the m e a n i n g slightly c h a n g e s

and implies surprise and counterexpectation.

(436) Combination of particles

aka=te tpu hjä itQa-e

3sg=FOC CHNG just do-3ps/E
'But he did it indeed.' [even though you did not think so, while I was convinced he
would do it]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
290 Word classes

5.15.4 Particle "tQuisi"

Not much information is available on a forth particle, tQuisi, which occurs as

tgaisi in the traditional language. When contrasted in elicitation to the three
examples in (435), it was claimed to imply counterexpectation, or to confirm
that something one was hoping for actually occurred.

(437) Particle tQuisi

aka=te tguisi itga-e

3sg=FOC CEXP do-3ps/E
'But yes he did it' [even though I was not sure he could make it]

5.15.5 Pity marker "noane "

The particle noane (also functioning as the root of the verb 'lie' - which is
usually mentioned by Urarina speakers when asked about the particle), marks
self-pity, compassion and can be translated as 'unfortunately'. The emphasis,
however, is not to show compassion for another person, but to express one's
own pitiful situation. For instance, when this word is used in a statement such as
'unfortunately I don't have any more rice', it primarily refers to the speaker's
own miserable situation rather than the intention to comfort the person who
asked for rice. The database contains only a few examples, some of which are
given in (438). In (438c), where noam is used with the imperative, it has a self-
humiliating connotation.

(438) Pity marker noane

a)
netitgara, noane kuhoarue-hee-ka katga=ne
maybe PTY remain-DIM-3ps/D man=CND
'Poor me, maybe if someone has remained [of all those who have died]...'

b)
nalu-a-atga-anu noane
fall-only-lsg/A PTY
'Poor me, I fell down.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Word classes and inflection 291

c)
atii noane ka=raj arusu te-uri-u
please PTY lsg=for rice give-RAP-IMP
'Please give a little rice to me poor guy!'

There are a f e w further particles, which are discussed elsewhere. These include
the reciprocal marker ita (cf. §16.3) and the hearsay marker hetau (cf. §22.5).

5.16 Word classes and inflection

In the previous sections, I described the various word classes and their specific
properties. It has become evident that verbs are inherently different from nouns
in a number of features: they take arguments, they are inflected for person,
number, polarity, and a number of further grammatical categories. They also
have a citation form that ends with the neutral form (which is identical to the
citation form and the form for 3ps/A), i.e. every verb in citation form ends with
an /a/ (or allomorphs). Nouns, in contrast, do not take arguments, and their
citation form is the singular, which can end in any vowel of the sound system. A
link between nouns and verbs can typically be achieved by derivation, which,
for instance, derives a noun from a verbal root, or vice versa. This process
usually requires morphological elements, typically realised as suffixes in
Urarina.
O n e example for derivation is the verbalisation from a noun root through
-oka, as described in §7.3.1. Even though it may seem awkward to create a verb
for 'sister' or ' a x e ' , this is a regular type of derivation in the language. A more
complicated situation is described below. In some examples (although these are
only a few), words that are doubtlessly non-abstract nouns receive a suffix that
is otherwise not used as a category-changing morpheme, but a suffix used with
verbs, the participle suffix -i or its equivalents for the continuous form. In
example (439a), the noun for ' b a c k ' (body part) is inflected; in (439b), the form
is suffixed to kliri, which is a type of tree (also referring to its fruit). The use of
the participle with this noun is understood as 'apply (the j u i c e of) this fruit to
someone'.

(439) Use of the participle form with nouns

a)
niki ajßa u-a=ne sajhjei ke hanori-l amu-a
but with come-3ps/D=SUB although VLI back-PRT walk-3ps/A
'But even though he had come with him, he walked (away), turning his back on him.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
292 Word classes

b)
ßäe nii kuri ke kuri-hja-1 ku mita-e
already DEM jagua.tree VLI 7'agwa-CNT-PRT there take.out-3ps/E
'Applying that jagua fruit [to the baby], he took [the baby] out.'

While this does not seem to be a productive pattern, it occurs in isolated cases.
There are several ways to account for this:
1. If we assume that the participle suffix (and its continuous counterpart) is
actually a derivational morpheme rather than inflection, the problem is
resolved. The dilemma is that it can be attached to any verb without further
conditions - which would be atypical for derivation. Another problem is that
a participle form is not typically word-class changing, as it is functionally
associated with verbs. A further problem is that the distinction between
inflection and derivation in Urarina is of a rather scalar kind and it is not
useful to strictly separate the two terms (also cf. § 10).
2. Perhaps a more natural way to account for the "verbalisation" of nouns is to
regard the conversion of a noun into a verb as zero derivation, such as
frequently happens in other languages. Similar to the example in (439)
above, English also derives a verb from the noun for 'back'. The resulting
verb can even be inflected for past as in 'back-ed'.

Since both solutions are suitable to describe the process observed here, I would
not like to determine which one is superior. However, given that the
phenomenon is only observed with some isolated cases, the second solution
may appear more likely.
Partly related to the issue of zero derivation is the following example:
kaa+rtii (also simplified to ka+rni) is a compound formed from two different
demonstratives: 'this' and 'that' (cf. (440)), which together mean '(a) sort o f ,
'something like this', 'something like that' and can function as a nominal head
or as a modifier. It is also used as a "filler" (as in pause), but most interestingly,
it functions as a verb, in which case the suffixes for person and number are
attached.

(440) Use of kaanii

a) Modifier:
lae hvajtei kanii banajt^u ne-1 hetau d^uhuturu-a ku-e
CHNG again sort.of type.of.heron be-PRT Η RS sit.down-NTR go-3ps/E
'[Then] again, becoming a sort of banajtQu heron, he went to sit down [on a branch].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Word classes and inflection 293

b) Head of NP:
ku ari-uru-a hau hetau ku ni-a kanii ni-a
there seek-PL-3ps/D because HRS there be-3ps/A sort.of be-3ps/A

fwamra+kata kahe ne-naa aaune+baka

banana+ripe from be-NOM plantain+soup
'As they looked there, there it was, there was something like that, plantain drink made
from ripe bananas.'

c) Verb:
ku najßal, lenone-ϊ, kaanii-tQuru-a hau,... ßäe nuku-hiue
there nicely eat-PRT sort.of-PL-3ps/D because already get.dark-CNT:3ps/E
'As they were eating nicely, doing this sort of thing,... it already became dark.'

Another interesting issue of Urarina word classes is reflected by a n u m b e r of

verbs that represent semantic concepts not typically associated to verbs in other
languages. T h e following examples contain verbs for '(be) like t h a t ' , '(be) in
v a i n ' , and the enumerative verb already mentioned in §5.8.

(441) Unusual concepts represented by verbs

a) Enumerative verb naaohiuaa:

kuriki ra-1 ke kalamina kurete-l, Idabo kurete-l,
money receive-PRT INST zinc.roof buy-PRT nail buy-PRT

semento kurete-l naaohwa-katge netene

cement buy-PRT ENUM-lpl/in must
'We must get money and buy zinc roof, nails, and cement with it.'

b) Verb 'be like that':

nitoani-a akauru it$au
be.like.that-3ps/A 3pl life
'Their life was like that.'

c) Verb 'in vain':

ate ari-a ku-a nukuj-a=ne hana nijej eru-i
fish seek-NTR go-NTR in.vain-3ps/D=SUBwhen not.at.all find-NEG:3ps/A
'When he went fishing in vain, he did not find anything.'

While s o m e suffixes that typically occur with verbs are also found on word
classes other than that, there are noun suffixes that occur on word classes other
than nouns. M o r e specifically, the noun plural -uru (or allomorphs) is suffixed

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
294 Word classes

to members of other word classes. The examples listed in (442) represent

relatively rare cases, but they illustrate that noun plural forms can occur on
postpositions and verbs. This occurrence is unexpected, as Urarina does not
show a high degree of morphological agreement otherwise (cf. §8.7. Therefore,
it remains unexplained why the examples kahe-uru and kuanaj-uru below have
plural markers, as 99% of other postpositions are unmarked for number, even in
contexts that imply plurality. The examples listed below are taken from
narratives; further examples for the plural suffix attached to postpositions
occurred in elicitation. In (442b), the plural marker is used anaphorically as the
noun it refers to is omitted. However, this does not automatically require the
attachment of the plural suffix to the postposition, as plural marking is highly
optional. Other conditions, such as discourse-related factors have not been
investigated due to the low number of examples.

(442) Plural marker -uru with postposition

a) With kahe:
ka=iritgu hatgu-uru kuane kahe-uru itanitQa karuru-1
lsg=ear hole-PL inside from-PL mud clean-PRT
'... cleaning the mud from the inside of my ear holes ...'

b) With kuanaj:
kuanaj-uru nianatn hja-ur-e
inside-PL even urinate-PL-3ps/E
'They even urinated in them [the clothes].'

The following examples illustrate that plural marking on postpositions is not

obligatory in comparable contexts. However, it is unclear whether -uru could
be attached to the postpositions occurring in (443).

(443) Postposition with plural marker

a) With kahe:
ßäe kahjune kaja-hakioa-uru-a katga-uru kahe
already clothes seize-DSTR-PL-3ps/A man-PL from
'They took the clothes away from the people'

b) With kuane·.
kulümi itulere nukue-kuru kuane ... akauru raj ate=ra
let.there.be all.kinds river-PL inside 3pl for fish=EMF
'Let there be fish for them ... in all kinds of rivers.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Word classes and inflection 295

Similarly, the form of the verb illustrated in (444) lacks an explanation. In this
case, (which is an isolated example) the participle form of 'say' is followed by
the plural in order to indicate that the content of the sentence was
communicated to several people.

(444) PL marker -uru after participle form

ka-sitga-a kukuri na-l-tguru nii ku raj huhjara-uru reruee-ka

lsg=scare-3ps/A armadillo say-PRT-PL that there POSS parent-PL disturb-3ps/A
'Saying [to them], "the armadillo is scaring me", he bothered his parents.'

Less surprisingly, the plural -uru is observed with the quantifier itulere (which
was shown to be a noun, cf. (400), with demonstratives (which can function as a
nominal head, cf. §5.11.1), and with numerals (in nominalised form). In (445c),
plural is marked on the nominalisation of the numeral verb for 'two'. Note that
in this case, the plural suffix precedes the nominaliser -i, as a regular feature of
inflection (cf. §6.6.1).

(445) Plural marker with other word classes

a) Demonstrative:
i=tQuerehe-uru i-tguasej nii-tguru ajßa ini-u-ra
2sg=child-PL 2sg-wife that-PL with go.up-IMP-EMF
' G o and climb with your w i f e and your children, with those.'

b) Quantifier:
atane huu-ka akau=ne hana, itulere-uru nerutii-hahoa-e
land f l o o d - 3 p s / D water=SUB when all.kinds-PL change-DSTR-3ps/E
'When the water flooded the earth, all kinds o f things changed.'

c) Numeral:
ku huajtei nii kuruataha-ur-i
there again that two-PL-NOM S bj
'(There) again, those t w o [men] [said] ...' [Lit. 'those w h o were t w o ' ]

The findings of this section can be summarised as follows:

- Morphemes usually associated with verbs can also be attached to some noun
roots and to the lexical combination of demonstratives, which (probably)
undergo zero derivation.
- The inventory of verbs is further extended through the existence of some
semantic concepts that are rarely expressed by verbs in other languages.
- The plural marker -uru, which is identical for verbs and nouns, can also be
attached to postpositions, demonstratives and other word classes.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
296 Word classes

One can conclude that the morphology of Urarina is flexible to a certain degree
in that it is not strictly bound to specific word classes. Furthermore, it becomes
evident that there hardly are any morphological structures other than the one
also found on verbs. From this it can be concluded that Urarina is a highly verb-
prominent language.

5.17 Summary

The analysis of different word classes in Urarina has shown that this system is
unusual in various respects. Urarina exhibits a strong predominance of verbs
and nouns, as most other word classes are either verb-like or noun-like, if not
subclasses of nouns and verbs. Adjectives are one example of this, as they do
not represent a separate underived class, but occur on a scale between nouns
and verbs, with a variety of features that are normally assigned to verbs and
nouns. Adverbs are special in that only a small number are underived (temporal
adverbs), whereas most other words of this class are derived from verbs.
Pronouns and demonstratives can be analysed as subtypes of nouns;
postpositions are noun-like in that they can take a plural marker, and quantifiers
do not represent a homogeneous word class as most members are derived verbs.
Furthermore, numerals have been analysed as special types of nouns and verbs,
respectively. What remains are small classes of highly functional categories
such as conjunctions, interjections, and introducers (which may represent a
typologically rare word class). The verb-prone character of Urarina is further
confirmed by the fact that the morphology is vastly based on suffixes that also
occur on verbs. The following table illustrates the statistical distribution of
different word classes based on 3,365 lexical entries.

(446) Statistical distribution of word classes

Word class Frequency In %

Verbs 1,854 55.1
Nouns 1,006 29.9
Adverbs 127 3.8
Bound morphemes 99 2.9
Idioms 70 2.1
Names 66 2.0
Ideophones 35 1.0
Postpositions 21 0.6
Conjunctions 18 0.5
Numerals 14 0.4
Particles 14 0.4
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
 Summary 297

Word class Frequency In %

Quantifiers 12 0.4
Introducers 10 0.3
Interrogatives 8 0.2
Pronouns 8 0.2
Demonstratives 3 0.1
Total: 3,365 100.0

The predominance of verbs and nouns in the lexicon is overwhelming, whereas

other word classes have much fewer members. It should be noted that the above
statistics include loans (less than 4%), which mainly occur as nouns and in
place and person names, but also as verbs and in idioms. The bound morphemes
include affixes and clitics. It would be possible to summarise the figures even
further, e.g. by combining pronouns, demonstratives, and interrogatives as
pronouns, but I have chosen to represent the figures in smaller subclasses as this
allows more precise statements about class membership.
The high concentration on nouns and verbs is also manifested by the fact that
the complex morphological system of Urarina, which contains a large number
of suffixes, is restricted to nominal and verbal markers. 3 7 If one goes even one
step further, the plural marker found on nouns can be analysed as identical to
the 3ps plural marker of verbs. Thus, there is no productive inflection
specifically assigned to word classes other than verbs (like adverbs and
adjectives). For a language with a complex system of inflection this seems to be
typologically unusual. While many other languages with rich morphologies
assign category-specific affixes to each word class, Urarina inflection is almost
exclusively verbal.

37
However, recall that some clitics operate on the clause level and can be attached to
any word level irrespective of its class assignment (cf. §2.11).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:32 PM
6 Noun Phrase structure

The order of elements within the Urarina Noun Phrase (NP) is variable to a
certain degree. Speaking in a simplified manner, all kinds of modifiers can
occur in prenominal position, with some peculiarities, which are discussed in
this chapter. In principle, a division into two groups can be made: modifiers that
can also follow the noun and those that cannot, or, expressed in a different way:
modifiers with a fixed position within the N P and modifiers that exhibit a
certain degree of variation regarding their position. The "restricted" group of
words, which obligatorily occur in prenominal position, consists of
demonstratives, possessives, and numerals. 38 Adjectives, nominalised verbs,
quantifiers, and special modifying nouns can occur before or after the noun, the
details of which will be investigated in §6.1 to §6.5. In §6.6,1 will also discuss
the position and structure of relative constructions, which are realised as
nominalisations and are therefore part of the NP.
While the high degree of order variation makes it almost impossible to state
the "unmarked" order of modifiers within the NP, it can be said that most
typically, all modifiers except adjectives precede the head noun, with
demonstratives and quantifiers preferably occurring before other modifiers. The
variable order of constituents within the NP is only partly predictable, such as
with some adjectival modifiers (cf. §6.5); otherwise, it is unclear whether order
variation has any semantic or pragmatic effects. It must be noted that
syntactically complex NPs (involving two or more modifiers) do not very
commonly occur in the database. Since the N T translation contains a much
wider range of examples for complex NPs, many of the cases illustrated in this
chapter are adopted from there. However, the lack of natural examples results
into a number of questions that cannot be answered, i.e. it is impossible to
predict all possible and impossible combinations for each and all nominal
modifier. Another issue is the ambiguity of some examples with regard to the
scope of some modifiers. As a general tendency, the rightmost modifier has
scope over the head of N P and any further modifiers that precede it have scope
over that constituent (right-branched bracketing). However, in some cases,
which all involve quantifiers, the scope of the quantifier can differ from that.
With some lengthy examples from the NT, the scope may be ambiguous. The
following sections discuss the details of the common occurrence of different

38
Numerals may be regarded a subtype of quantifiers, but since they exhibit a different
syntactic behaviour than other quantifiers, they are discussed separately. Also see the
structural differences between quantifiers and numerals as discussed in §5.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Demonstratives 299

modifiers, exemplified by the most predominant structures. A summary with a

tentative overview o f possible positions is given in ( 4 9 8 ) . A s a tendency, the
following order o f modifiers, which excludes quantifiers, is c o m m o n : D E M +
P O S S R + N U M + N + A.

6.1 Demonstratives

Demonstratives usually take the first position in the Ν Ρ and are not normally
preceded by any other modifier, with the exception o f some quantifiers (cf.
§ 6 . 4 ) . T h e y also occur before possessives (i.e. before the possessor, which in
turn precedes the possessum). T h e following list shows those combinations
from the database in which demonstratives are attested. T h i s summary is
illustrated by the examples in ( 4 4 8 ) .

(447) Attested combinations involving demonstratives

[DEM + N] (cf. (448a))

[DEM + [POSSR + N]] (cf. (448b))
[DEM + [MOD + N]] (cf. (448c))
[DEM + [NUM + N]] (cf. (448d))
[QNF + [DEM + N]] (cf. (449a))
[DEM + [QNF + N]] (cf. (449b))

(448) Position of demonstratives

a) Before head noun [DEM + N]:

kaa d$aura+atane huu-hwe=lu lomaj beene
[this flesh+land] flood-3ps/E=REM PSN female
'Lomai flooded this world.'

b) Before possessive constructions [DEM + [POSSR + N]]:

kaa kana raj d^aura+atane
[this [lpl/in POSS flesh+land]]
'this world of ours'

c) Before modifying noun and head noun [DEM + [MOD + N]]:

nii lomaj beene
[that [PSN female]]
'that woman Lomai'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
300 Noun Phrase structure

d) Before numeral and head noun [DEM + [NUM + N]]:

kau-e nii lejhn katQa
return-3ps/E [that [one man]]
'That one man went home.'

With some quantifiers, the occurrence of a demonstrative is in second position,

i.e. following the quantifier. However, the reverse order (with the demonstrative
in initial position) is possible with some quantifiers (also cf. §6.4).

(449) Demonstrative and quantifiers

a) After quantifier (with most quantifiers) [QNF + [DEM + N]]:

hitarii kaa katQa-uru amia-rate-1
[all [this man-PL]] work-CAU2-PRT
'making all these people work'

b) Before quantifier (few cases) [DEM + [QNF + N]]:

kaa arahn katQa-uru raj
[this [many man-PL]] for
'for all these people' [NT: John 6:5]

6.2 Numerals

Numerals have a comparatively fixed position within the N P in that they follow
demonstratives and precede nouns. However, the order between possessives and
numerals may vary, as will be shown by the examples in (452).

(450) Attested combinations involving numerals

[NUM + N] (cf. (451a))

[NUM + [N + A]] (cf. (451b))
[[NUM + N] + QNF] (cf. (451 c))
[DEM + [NUM + N]] (cf. (448d))
[NUM + [POSSR + N]] (cf. (452a))
[POSSR + [NUM + N]] (cf. (452b))

Numerals always precede the noun, as is illustrated in (451). Note that a

reversion of the order (e.g. katQa letQurjkaj for (451a)) would change the
function into a possessive construction to mean 'ten of the m e n ' . In example
(451c), the numeral is expressed as lelauri ( ' o n e group'); the bracketing of this
example is different as the head is followed by a quantifier (cf. §6.4).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possessives 301

(451) Position of numerals

a) [NUM + N]:
le=tQurjka-j katQa
[one=ten-NOMSbj man]
'ten men'

b) [NUM + [N + A]]:
lejhii katQa baaso
[one [man bad]]
'one bad man'

c) [[NUM + N] + QNF]:
le=lauri katga arahii
[[one=group man] several]
O n e group of many people [became water people].' [I.e. a large group of people]

When a possessive construction forms part of the NP, the preferred order is with
the numeral in initial position, followed by the possessive (cf. (452a)). There is
only one counterexample where the numeral occurs between possessum and
possessor, which is displayed in (452b).

(452) Numerals and possessives

a) [NUM + [POSSR + N]]:

letQutjka-j kanu beree-kuru
[ten-NOMsbj [Isg child-PL]]
'my ten children'

b) [POSSR + [NUM + N]] (elicited):

noe nitgataha-j kalaui-tQuru
[PSN [three-NOM Sbj son-PL]]
'Noah's three sons'

Since the example in (452b) is elicited and represents a rare case, one can
conclude that the preferred position of numerals is before possessives, and
obligatorily before the head noun, but after demonstratives.

6.3 Possessives

In the previous section, it was already mentioned that the order between
numerals and the possessor within an N P is variable. The position of possessors

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
302 Noun Phrase structure

is otherwise fixed in that they never follow the head noun and never precede
demonstratives. Certain quantifiers, (exemplified by satü 'all' in (454d)) can
occur before a possessive.

(453) Attested combinations involving possessives

[POSSR + N] (cf. (454a))

[DEM + [POSSR + N]] (cf. (448b), (454b))
[NUM + [POSSR + N]] (cf. (452a), (454c))
[POSSR + [NUM + N]] (cf. (452b))
[[QNF + POSSR] + N] (cf. (454d))

(454) Position of possessives

a) [POSSR+ N]:
kanu ßaka
[lsg father]
'my father'

b) [DEM + [POSSR + [POSSR + N]]]:

nii hurju-kuru kuraanaa lureri asae
[that [Jew-PL [chief house]]] under
'in the house of that chief of the Jews' [NT: Matthew 9:23]

c) [NUM + [POSSR + N]]:

lejhü urarißa-uru kuraanaa
[one [Urarina-PL chief]]
'one chief of the Urarina'

d) [[QNF + POSSR] +N]:

aka=te satü akauru kuraanaa ne-ι ße
3sg=FOC [[all 3pl] chief] be-PRT be-3ps/E
'He is the chief of all of them.'

Possessives also occur in recursive manner, as the following example illustrates.

Note that while such chains are common in the NT, there are no recorded
examples for these in natural discourse.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Quantifiers 303

(455) Possessive chains

romaw-kuru raj soodaro-oru kuraanaa letono rauta-a hesoso=ne

[Roman-PL POSS [soldier-PL [chief [envoy]]]] heal-3ps/D PSN=CND
'when Jesus healed the envoy of the chief of the soldiers of the Romans' [NT: Matthew
8:5]

6.4 Quantifiers

T h e position o f quantifiers within the N P is not h o m o g e n e o u s , as it d i f f e r s

b e t w e e n d i f f e r e n t quantifiers. T h e r e f o r e , the list in ( 4 5 6 ) m u s t b e used with
caution, as it does not apply to each quantifier in the s a m e w a y .

(456) Attested combinations involving quantifiers

[QNF + N] (cf. 458), (459))

[QNF + [DEM + N]] (cf. 449a), (460a))
[[QNF + POSSR] + N] (cf. 454d))
[QNF + [POSSR + N]] (cf. 460d))
[QNF + [NUM + N]] (cf. 460b))
[QNF + [N + A]] (cf. 460c))
[N + QNF] (cf- 462a,d))
[DEM + N] + QNF]] (cf. 462c))
[[POSSR + N] + QNF] (cf. 462b))
[DEM + [QNF + N]] (cf- 449b), (461a))
[DEM + [QNF + [MOD + N]]] (cf- 461c,d))
[DEM + [QNF + [REL + N]]] (cf. 461b))

T h e table in (457) gives t h e possible and preferred positions f o r each quantifier.

T w o q u a n t i f i e r s o c c u r in p r e n o m i n a l position only, w h e r e a s another is m a i n l y
f o u n d in postnominal position. M o s t other quantifiers that occur in m o d i f i e r
f u n c t i o n a r e p r e d o m i n a n t l y observed in prenominal position, with a (rare)
option o f o c c u r r i n g a f t e r the noun. T h e situation with tabauru ' s o m e ' will be
illustrated in (464).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
304 Noun Phrase structure

(457) Position of quantifiers with modifier function

Quantifier Prenominal Postnominal

laurilauri 'various (groups of)' yes rare
arahü 'many', 'various' yes rare
hitarü 'all' yes rare
satü 'all' yes (unattested)
itulere 'all kinds o f yes no
d^uni 'nobody' yes no
tabauru 'some' rare (yes)

F o r the q u a n t i f i e r s laurilauri, arahü, hitaru, and satü, the m o s t f r e q u e n t

position is b e f o r e the noun, itulere and d^uni (or variant d^unej) are f o u n d o n l y
in this position. T h e f o l l o w i n g e x a m p l e s illustrate a variety of c o m b i n a t i o n s
with t h e s e w o r d s in p r e n o m i n a l position.

(458) Obligatory prenominal position of itulere and d$uni [QNF + N]

a) itulere:
nitgae hetau=te ne=lu itulere bakaua
over.there HRS=FOC be:3ps/E=REM all.kinds Indio
'Over there were all kinds of [dangerous] indios.'

b) d^uni:
rüjej d^unej katQa beraj-ßaa najß-ene
not.at.all nobody man care.for-INF be.able-NEG:3ps/E
'Absolutely nobody can take care of him.'

(459) Quantifiers that prefer prenominal position [QNF + N]

a) arahü:
arahü alau ru-anu-m hau
many spider.monkey find-lsg/D=SUB because
'as I had found [a group of] many spider monkeys ... [I shot them].'

b) hitanT.
hitarü akauru raj d^alu-a hau
all 3pl for distribute-3ps/D because
'as he distributed [the clothes] to all of them'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Quantifiers 305

c) satii:
satü kat$a-uru=te ku-re=i
all man-PL=FOC go-IRR:3ps/E=ASS
'All the people will go [to the election].'

d) laurilauri:
tgäe i-hiniitQal lauri+lauri ne-nakauru eno-a-u
also 2sg=together RED+group be-those.who enter-CAUl-IMP
'Also let enter with you all species of those that exist.'

When other modifiers are present in addition to one of the quantifiers

mentioned above, the most typical occurrence of the quantifiers hitaru and satü
is in phrase-initial position, where they precede all other modifiers (cf. (460a-
c)). The quantifier arahii does not occur before demonstratives, but it is attested
before a possessor, as shown in (460d). laurilauri is unattested in these
positions.

(460) Quantifier in phrase-initial position

a) Quantifier hitarii before demonstrative [QNF + [DEM + N]]:

hitaru nii kari kajßa-ku-toe=ra
[all [that shell]] give.back-IMP-PL=EMF
'Give all those shells back!'

b) Quantifier hitarü before numeral [[QNF + [NUM + N]:

hauria hitarn letgurjka-j katQa rela-i=ße reetial
first [[all ten-NOMSbJ] man] teach-2ps=SUB until
'until you have taught all ten people'

c) Quantifier satii before noun followed by adjective [QNF + [N + A]]:

satii temule kauatpa
[all [plant good]]
'all good plants'

d) Quantifier arahii before possessive [QNF + [POSSR + N]]:

kaa nese w-uru-a arahü kanu katQa-uru=ne hau
this town be-PL-3ps/D [many [lsg man-PL=SUB]] because
'because there are many of my people in this town'

It should be noted that the scope of quantifiers is not 100% predictable, as is

seen from examples (460d) as opposed to (454d). While (460d) follows the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
306 Noun Phrase structure

general tendency of right-branched bracketing, the other example exhibits the

scope of a quantifier over the possessor, rather than referring to the head noun.
While the preferred position of prenominal quantifiers is in initial position of
the NP, there are exceptions from this as well. The following examples show
that quantifiers can be preceded by demonstratives.

(461) Quantifier in non-phrase initial position

a) hitarii: [DEM + [QNF + N]]

nii hitarii akauru ne-uru-a=ne kimtuku-a=ne baja
[that [all 3pl]] be-PL-3ps/D=SUB know-3ps/D=SUB after
'after he knew that they were all there' [NT: Acts 21:19]

b) arahii: [DEM + [QNF + [REL + N]]]

nii arahii ku-nakauru katga-uru
[that [many [go-those.who man-PL]]]
'those many people who had gone' [NT: Mark 5:25]

c) satii: [DEM + [QNF + [MOD + N]]]

nii satii hesoso eseneta-era-uru
[that [all [PSN believe-AG-PL]]]
'all those who believed in Jesus' [NT: Luke 9:23]

d) itulere: [DEM + [QNF + [MOD + N]]]

nii itulere kiuituku-ur-era katQa-uru
[that [all.kinds [know-PL-AG man-PL]]]
'those various people who know it' [NT: Matthew 2:2]

In contrast to the examples in (461), there also are instances where the
quantifiers laurilauri, arahu, and hitarü are found after the noun. Note that this
is not attested for satii.

(462) Quantifiers in postnominal position

a) arahii: [N + QNF]:
hi ne-rehete-kure katQa-uru arahii
there be-HABl-3pl/E [man-PL several]
'Many people used to live there.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Quantifiers 307

b) hitarii: [[POSSR + N] + QNF]:

kana+hoaaun-era kalaui-tguru hitarii tauhja-a=ne kujßa
[[Our.creator son-PL] all] gather-3ps/D=SUB so.that
'... in order to collect all the children of God ...' [NT: John 11:52]

c) hitarii: [DEM + N] + QNF]]

mi hau hetau=te mi nune-uru hitarii itulere nukue ne-l-tguru
that because HRS=FOC[that [branch-PL all]] all.kinds river be-PRT-PL

ni-akioa-a ku-e=lu
be-DSTR-NTR go-3ps/E=REM
'Therefore, all those branches turned into all the rivers that there are.'

d) laurilauri: [N+ QNF]:

lenone lauri+lauri
[food RED+group]
'various types of food' (elicited)

e) satii: [[DEM + N]+ QNF] (from Pucuna dialect):

kaa komomda satii lenoni-tQaanu
[[this community] all] eat-lpl/ex
'The entire village ("we") all ate.'

An interesting variation in the position of quantifiers is their occurrence before

a verb, such as illustrated in (463), an example from the Pionero dialect. In this
case, the function of arahii has to be interpreted as adverbial, even though the
translation may suggest that the quantifier is part of the NP. Similar examples
are also attested in the N T translation.

(463) Quantifier with adverbial function (from Pionero dialect)

arahii wawako-a kakuri hau

many bark-3ps/D trompetero. bird because
'because many trompetero birds were making noise' [lit. "being many, they made
noise"]

In contrast to most of the quantifiers discussed above, tabauru 'some' normally

occurs in postnominal position. A possible interpretation for its occurrence after
the head is based on the fact that tabauru also functions as a noun. Under these
circumstances, it could be regarded a part of a possessive construction, as the
below examples illustrate: instead of functioning as a postnominal quantifier, it
is plausible to interpret tabauru as the head of NP, preceded by a possessor

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
308 Noun Phrase structure

(also cf. §5.11.3). However, while this position is the more common one,
tabauru can also occur before the head, where it functions as a modifier (cf.
(464b)).

(464) Position of tabauru

a) After noun (head of NP):
eene-kuru tabauru
woman-PL some
'some (of the) women'

b) Prenominal (elicited):
tabauru urarißa-uru ßäe ere-uru-i n=ere he
some Urarina-PL already speak-PL-NEG:3ps/A 3ps=word VLI
'Some Urarina (people) already do not speak their language.'

In summary, it can be said that the position of quantifiers in the NP is relatively

free, as compared to other components. While initial position is the preferred
one for most quantifiers, these do not only exhibit variations from that, but there
also are differences between different quantifiers.

6.5 Nominal and noun-like modifiers within the NP

Regarding the position of adjectives within the NP, it is important to remember

that Urarina has no class of underived adjectives, but a variety of words that can
take function usually assigned to adjectives. The main distinction is between
noun-like adjectival modifiers and deverbal adjectives. Therefore, the position
of adjectival modifiers is not homogeneous (cf. §5.4). Some occur in
postnominal position only, which is the unmarked variant, while others can
occur in prenominal or postnominal position. As a reminder, the respective
words are listed again in (465). Recall that most words (including deverbal
adjectives composed of a verb root and a nominalising suffix) can also function
as head of an NP. There are only two words that cannot function as nominal
head.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Nominal and noun-like modifiers within the NP 309

(465) Position of adjectival modifiers in the NP

Adjective Status Position

laauhwiri 'small' Modifier only postnominal
seohzua 'big' Modifier only postnominal
leotQa 'other' Can be Ν head prenominal
(netQa)äasaj 'evil' Can be Ν head pre-/postnominal
(n)atQasijie 'new' Can be Ν head pre-/postnominal
enamanaa 'young (man)' Can be Ν head pre-/postnominal
kauat^a 'good' Can be Ν head pre- /postnominal
baaso 'bad' Can be Ν head postnominal
biißa 'old' Can be Ν head postnominal
äafihii 'little' Deverbal adj. prenominal
kauati 'good' Deverbal adj. pre-/postnominal
tabaj 'big' Deverbal adj. pre-/postnominal
tasißohiuaj 'big' Deverbal adj. postnominal
lanahaj 'red' (all colours) Deverbal adj. postnominal

T h e above table shows that there are three m a j o r groups of adjectival modifiers
- those that those that function as noun modifiers only (cf. §6.5.1), those that
can also function as head of N P (cf. §6.5.2), and deverbal adjectives, as
discussed in §6.5.3. Their position within the N P is highly heterogeneous.
Subsequently, the combinations of adjectives with other elements within the N P
can be expected to vary with respect to the adjective that is involved, which is
in fact the case. For other differences between the different types of adjectives,
see §5.4.4.

6.5.1 Adjectives in the NP

T h e w o r d s laauhwiri and seohwa are the only ones that do not function as a
nominal head, but exclusively occur with m o d i f y i n g function. Their position in
the N P is restricted to occurrence directly after the noun.

(466) Position of laauhzuiri and seohiva: Ν + A

a) laauhiviri:
kanaanaj laauhzuiri
child small
'little child'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
310 Noun Phrase structure

b)seohioa:
bute seeohwa itQa-e
boat big make-3ps/E
'He made a big boat.'

The addition of a prenominal modifier does not have any impact on the position of
these adjectives. The examples in (467a,b), illustrate the occurrence of these words
in combination with numerals and demonstratives. In the same way, they occur in
this position when the head noun is preceded by a possessor (cf. (467c)).

(467) Occurrence of laauhwiri and seohioa with other modifiers

a) With numerals [NUM + [N + A]]:

lejhii saabere seohioa te-ure
[one [sword big]] give-3pl/E
'They gave [him] a big sword.' [NT: Revelation 6:4]

b) With demonstratives [DEM + [N + A]]:

nii kampana laauißaa te-u
[DEM [bell small]] give-IMP
'give (it) little golden bells' [Old Testament: Exodus 28:34; also verses 28, 33; 39, 25]

c) With possessor [POSS + [N + A]]:

kanu kamanaj laauhwiri
[lsg [child small]]
'my little child'

In summary, laauhzoiri and seohioa are most restricted in their function and
position, as they always modify a noun (but do not function as a head) and
consistently occur in postnominal position.

6.5.2 Adjectival nouns that can function as nominal head

The group of adjectival nouns that can also function as a nominal head is much
more diverse than the group mentioned above. Adjectival nouns differ from
other nouns in that they typically occur as modifiers of a nominal head. They
are also different from other modifying nouns, which mainly comprise proper
and "status" nouns (cf. §6.5.4). While the majority of adjectival nouns resemble
adjectives in that they occur as modifiers directly after the noun, there are some
whose position can vary, netgäasaj 'wicked', nat$a$i]ie 'new', and enamanaa

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Nominal and noun-like modifiers within the NP 311

' y o u n g ' m a y also p r e c e d e the head noun. T h e e x a m p l e s in ( 4 6 8 ) first s h o w the

t w o n o u n - l i k e m o d i f i e r s that obligatorily f o l l o w the head noun.

(468) Adjectival nouns in postnominal position

a) biißa:
lureri biißa
house old
'old house'

b) baaso:
hanonaa baaso bana
day bad when
'on a bad day' [e.g. a rainy day]

T h e variable position o f netQäasaj, natQasiße, and enamanaa is illustrated in ( 4 6 9 a -

f) below.

(469) Modifiers netQäasaj, natQasiße, and enamanaa with variable position

a) Postnominal position: natQasiße

tefwolo natQasiße icha-a iglesia=ne kujßa
church.building new make-3ps/D church=SUB so.that
1
'so that the church builds a new church building

b) Prenominal position: natQasiße

ßäe natQasiße hanonaa ke turu-a-ne bana
already new day VLI arrive-3ps/D=SUB when
'at the time when a new day arrives'

c) Postnominal position: netQäasaj

katQa netQäasaj
man evil
'evil m a n '

d) Prenominal position: netQäasaj (here as dialectal variant netQuäasaj)

netQuäasaj tajbißaae
evil animal
'the beast' [NT: Revelation 12:13]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
312 Noun Phrase structure

e) Postnominal position: enamanaa

katga naamanaa
man young.man
'young man'

f) Prenominal position: enamanaa

enamanaa katga
young.man man
'young man'

Semantic or pragmatic differences for the variation in order of the above

examples are not known and the order variation appears to occur quite freely.
However, with the word kauatQa 'good', which also may occur before or after
the head of NP, speakers tend to recognise tiny nuances for the interpretation of
each order. While the example in (470a) is the more regular order for a modifier
and is interpreted with some contrast ("good food" as opposed to "bad food"),
the example in (470b) has an almost generic reading such as 'he is a good guy'
(without implying contrast), possibly comparable to the distinction between
mujer linda and linda mujer in Spanish.

(470) Variable position of kauatQa

a) kauatQa after head of NP:

lenone kauatQa te-i-tQe
food good give-2ps-PL
'You give him good food.'

b) kauatQa before head of NP:

aka=te kauatQa katQa ne
Ssh^FOC good man be-3ps/E
'He is a good man.'

In contrast to other adjectival nouns, the word leot$a 'other' occurs only in
prenominal position. Reversing the order such as illustrated in (471c) changes
the meaning to a possessive reading, literally meaning 'the ones different from
the Urarina'. Also note that leotg,a always occurs adjacent to the head noun.

39
In Romance languages, an adjective following the noun refers to a property in
differentiation to another, possibly contrary property ('nice, not bad'), whereas the
adjective in prenominal position implies that the quality expressed by it is naturally
associated with the noun ("generic" property).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Nominal and noun-like modifiers within the NP 313

(471) Position o f leotga

a) Before noun:
leotQa rukuele=te laati-tgaatge
other things=FOC need-lpl/in
'We need other things.'

b) With quantifier:
arahti leot$a-uru katQa-uru
many other-PL man-PL
'many other people'

c) Possessive:
urarifla-uru leotQa-uru
Urarina-PL other-PL
'other people than Urarina'

6.5.3 Deverbal adjectives

Adjectival modifiers that are derived from verbs by the nominalising suffix -i
do not behave in a uniform way. Words of this type preferably occur in
postnominal position, as displayed in (472). However, some cases are also
attested before the head noun: Examples with kauati 'good' and tasipohwaj
'big' are given in (473). Different from other deverbal adjectives, aafihii
'small' is only observed in prenominal position (cf. (474)).

(472) Deverbal adjectives in postnominal position

a) hitQuhivaj:
hjanarihi hitQuhzua-j
sugar.cane be.black-NOM S bj
'black sugar cane' [Lit. 'sugar cane that is black']

b) tasißohivaj:
lureri tasißohiua-j hoara-kau
house be.big-NOM s b j see-Ips/A
Ί have seen a big house.' [Lit. 'house that is big']

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
314 Noun Phrase structure

c) tabaj:
ku ne-rehete fiuasto taba-j
there be-HABl :3ps/E field be.big-NOM Sbj
'There used to be a big field.' [Lit. 'field that is big']

The postnominal position of the modifiers presented above is also attested with
other verbs that are nominalised with -i. However, this does not correspond to
their prototypical position. As shown in §6.6.1, other intransitive verbs with the
nominaliser -i preferably occur before the head. This allows the conclusion that
deverbal adjectives have a slightly different syntactic behaviour than other
nominalisations with -i.
There are some attested examples for deverbal adjectives that occur in
variable position. In the same way as the adjectival noun kauatQa (cf. (470)),
the deverbal adjectives kauati, tasißohzoaj, and tabaj entail a slightly different
meaning when appearing before the head noun. The interpretation implied by
this order refers to a more generic, almost compound-like nature of the NP.

(473) Deverbal adjectives in prenominal position

a) kauati:
ni-a kauat-i teefivolo=ne
be-3s be.good-NOMsbj church.building=CND
'if there is a nice church building' [implying: not "nice" as opposed to "ugly", but one
of the nice sort, such as the people in the other village have]

b) tasißohwaj:
tasißohwa-j läatga
be.big-NOM sb j ship
'big ship' [implying "not just any boat as opposed to others, but the big type that goes
on the Maraflon River"]

The word äafihii for 'small' behaves in a different way, as it is exclusively

found before the head. Since its occurrence is most frequent in non-modifier
functions (e.g. as a nominal head), there are very few examples of its presence
as a noun modifier.

(474) Position of äafihii

äafihi-i mustasa ukue

be.small-NOMsbj mustard seed
'a small mustard seed' [NT: Luke 13:18-19]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Nominal and noun-like modifiers within the NP 315

At this stage, it must be assumed that the pre-head position of äafihii is the only
possible one, as no other examples are attested. Also note that an adjective can
be preceded by the intensifier hatal In example (475), this combination results
into the structure Noun + Intensifier + Adjective.

(475) Adjective modified by intensifier

m-a lejhii aheri hatal kauati kulu-1
be-3ps/A one stone very good colour-PRT
'He has a stone with a very nice colour.'

6.5.4 Other modifying nouns

As mentioned in §5.1.3, there are some nouns that can modify another noun.
Differences between these modifying nouns are observed with respect to their
position within the NP. Some nouns occur before the head, whereas others
follow the head. Modifying nouns can be proper nouns or can be related to
status or profession. As "status" is a difficult to define term, the cover term
"social position" is used in (476), which also includes "profession". 40 Another
important factor to be mentioned is that constructions that involve nouns, which
modify another noun are not frequent. As many examples are taken from the
NT, some of these could be caiques, which would also explain the variability in
position.

(476) Position of modifying nouns in the NP

Modifying noun Position
Social position e.g. htraanaa 'chief Postnominal
Tribal affiliation e.g. uraripa 'Urarina' Prenominal
Place names e.g. nauta Prenominal
Person names e.g. lomaj Pre-/postnominal

An example for "status" is the one in (477a), where 'woman' is further

specified as 'envoy of God'. However, this example was elicited and may not
necessarily represent a typical case. While it is possible to form such
constructions, these do not occur in the data corpus. The example in (477b)
faces another challenge as it is not entirely clear which noun is the head: are we
talking about priests, or about chiefs? Both interpretations are possible. In

40
The term "status" may be more justified as a characterisation of the modifying noun
in rtii raj bitoasipe aräala 'his booty, the tapir', where 'tapir' (araala) is specified by
'his booty'. However, such examples are extremely rare (cf. ex. (1) in §20.3.5).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
316 Noun Phrase structure

addition, both examples given in (477) are ambiguous and could be interpreted
as possessive constructions such as ' G o d ' s envoys' w o m a n ' and 'the chiefs of
all priests', respectively.

(477) Position of nouns referring to status

a)
kana+hoaaun-era letono eene
Our.creator envoy woman
'a woman who is an envoy of God' (elicited)

b)
satti saseredote-kuru htraanaa-kuru
all priest-PL chief-PL
'all chief priests' [NT: Matthew 2:4]

With nouns referring to a profession, a similar problem arises. Again, the

following example is elicited and it is unclear whether this can be regarded a
natural example for a modifier construction. A possible alternative way to
express the content of this phrase would be by nominalisation such as 'a w o m a n
w h o is a missionary' (cf. §6.6).

(478) Position of nouns referring to profession

eene misjonera
woman missionary
'a woman who is a missionary' (elicited)

A s mentioned, the combination of person or place names with full nouns is not
very c o m m o n in Urarina, which could mean that the examples from the N T
discussed below are caiques. However, they will be mentioned for the sake of
completeness, as this will illustrate which combinations are possible (even
though they might not be typical). The example (479a) is from a narrative and
therefore "natural". Also note that the second noun is prefixed with b-, which
occurs in this kind of construction, but only if the head noun begins with a
vowel (cf. §7.3.2). The other two examples are more difficult to interpret: again,
it is unclear which of the nouns is the head of the phrase, since person names
can occur with head function without an accompanying noun. Since Urarina
speakers find it absurd to see a difference between something that can be
translated as "King Herod" or alternatively "Herod the king", there is little
evidence for either analysis.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Nominal and noun-like modifiers within the NP 3 17

(479) Position of nouns referring to person names

a) With 'woman': name + head noun

lomaj b-eene
PSN ASCM-woman
'the woman Lomai'

b) With possessive and 'husband': head noun + name

maria lana hose
PSN husband PSN
'Jose, husband of Maria' [NT: Matthew 1:16]

c) With 'king': head noun + name

rey erodesi
king PSN
'Herod the king' [NT: Matthew 2:1]

N o u n s that refer to tribal affiliation are not much easier to analyse with regard
to their function as noun modifiers. The reason for this is that they typically
occur as nominal heads themselves. For instance, rather than the (elicited)
example in (480a), the variant without katQa ' m a n ' would be more natural.
Here, ' m a n ' is understood as the head noun modified by 'Indio' (indicating
what type of man is referred to). 41 The second example 'Urarina settlement'
could also be interpreted as a possessive construction: "town of Urarinas".

(480) Position of nouns referring to tribal affiliation

a) With 'Indio': MOD + head noun

bakaua katQa kiitQa
Indio man boy
'son of an Indio' (elicited)

b) With 'Urarina': name + head noun

urarißa nuse
TRN town
'Urarina settlement' (elicited)

The N T offers a variety of further examples that involve place names. However,
it appears that the position of these is not strictly fixed. In some examples, the

41
The word bakaua is used as a general term for all native indian groups other than the
Urarina, usually with a negative connotation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
318 Noun Phrase structure

place name precedes the head noun, but it occurs in postnominal position in
other phrases.

(481) Prenominal position of place names

a) With atane 'land':

hodaa b-atane ke ne-naa belee nese
PLN ASCM-land VLI be-NOM PLN town
'(the town of) Bethlehem, which is in Judea' [NT: Matthew 2:6]

b) With itpatane 'mountain':

sinai b-itpatane-u
PLN ASCM-mountain-LOC
'at Mount Sinai' [NT: Galatians 4:30]

c) With nese 'town':

taba-i nauta nese
be.big-NEG:3ps/A PLN town
'[The town of] Nauta is not big.'

(482) Postnominal position of place names

a) With nukue 'river':

nukue taba-j eufrate auania
river be.big-NOMSbj PLN edge
'at the edge of the big river Euphrates' [NT: Revelation 9:14]

b) With nese 'town':

lejhii nese betafiuahee
one town PLN
'one (certain) town (called) Bethphage' [NT: Matthew 21:1]

c) With komonida 'community' (loan from Spanish):

komonida taual ne-nakauru rela-a
community PLN be-those.who teach-3ps/A
'He teaches those who live in the community of Tagual.'

There also are some attested examples in which a complex NP modifies

another. In (483), two such cases are given. In the first sentence, a possessive
construction 'my people' occurs in initial position and is modified by another
complex NP 'the inhabitants o f Israel', which consists o f a place name and a
head noun. However, it is difficult to determine which of the two NPs
represents the head of the whole (or whether both are heads). Similarly, the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Nominal and noun-like modifiers within the NP 319

second example represents a combination of NPs for which it is unclear which

one is the head.

(483) Combination of complex NPs

a)
kann katQa-uru urarißa-uru
1 sg man-PL Urarina-PL
'my people the Urarina'

b)
kana hoaaun-era israel kiuaaun-era
lpl/in create-AG PLN create-AG
'God, the creator of Israel' [Lit. our creator, the creator of Israel') [NT: Matthew 15:31]

In summary, it is difficult to describe the behaviour of modifying nouns since

they are hardly attested in natural text. The table in (476) gives a rough idea of
their position, but it should be kept in mind that nouns which modify other
nouns are rather uncommon in Urarina.

6.5.5 More than one adjectival modifier

Sequences of modifiers are not really common; all attested examples are from
the NT; comparable structures are confirmed for elicited examples not listed
here. Also note that the bracketing for these examples is tentative. The
examples in (484) involve numerals and adjectives and illustrate that more than
one adjectival modifier can occur within one NP. In two of the sentences, the
adjectival noun netQäasaj 'wicked' occurs before the word baaso 'bad (thing)',
which is can function as a modifier, but is a nominal head here. In the other
example, netQäasaj precedes the nominal head, which is further modified by
postnominal tabaj.

(484) Numeral with adjectival modifier

a) [NUM + [[A + A] + REL]]:
kuruataha-l netQäasaj baaso itga-rauru ama-ure
[two-PRT [[wicked bad] do-those.who]] give-3pl/E
'They took two people who had done wicked bad things.' [NT: Luke 23:32]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
320 Noun Phrase structure

b) [DEM + [NUM + [[A + A] + REL]]]:

ajßa hiniikn nii kuruataha-j netQäasaj baaso it$a-rauru te-ure
with together [that [two-NOM Sb j [[wicked bad] do-those.who]]] give-3pl/E
'They put those two who had done wicked bad things together with him.' ' [NT: Luke
23:33]

c) [NUM + [A + [N + REL]]]:
lejhii netQäasaj tajbipaae taba-j
[one [wicked [animal be.big-NOM Sbj ]]]
'one big wicked animal' [NT: Revelation 11:7]

T h e f o l l o w i n g t w o e x a m p l e s again contain several adjectival m o d i f i e r s each. In

( 4 8 5 a ) , t w o n o m i n a l i s e d v e r b s f o l l o w the noun in a s e q u e n c e ; the description o f
' p u r p l e ' is realised as "blackish red". In (485b), there are three m o d i f i e r s : o n e
d e v e r b a l a d j e c t i v e o c c u r s in initial position, while a n o t h e r a p p e a r s at the end o f
t h e N P . In addition, the noun is directly f o l l o w e d by the m u l t i f u n c t i o n a l
a d j e c t i v e kauatQa 'good'.

(485) Several adjectival modifiers

a) DEM + Ν + REL + REL (bracketing unclear):

nii kahjune lanaha-j hitQua-hiohwa-j
that cloth be.red-NOMsbj be.black-DER-NOM Sbj
'that purple cloth' [NT: Mark 15:20]

b) REL + Ν + A + REL (bracketing unclear):

somaha-j kahjune kauatga hikiaki-i
be.white cloth good shine-NOM Sbj
'a nice, white, shining cloth' [NT: Revelation 19:8]

T h e split of several m o d i f i e r s into p r e n o m i n a l and p o s t n o m i n a l position is also

c o n f i r m e d by elicited e x a m p l e s . N o t e that the o c c u r r e n c e o f m o d i f i e r s as
a f t e r t h o u g h t s is not attested in Urarina.

6.6 Relativisation

T o call a section of a c h a p t e r on N P structure " R e l a t i v e c l a u s e s " m a y not s e e m

a p p r o p r i a t e at first sight as they do not constitute i n d e p e n d e n t c l a u s e s in
U r a r i n a . G r a m m a t i c a l categories such as tense, aspects etc. cannot be e x p r e s s e d
in a relative construction. In fact, there is no particular construction or
m o r p h o l o g y f o r relative clauses, but various kinds o f n o m i n a l i s a t i o n are

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Relativisation 3 21

employed for clause relativisation. The resulting noun can be used as an

argument in any syntactic slot of a clause, i.e. any argument can be shared
between main and relative clause. One type of nominalisation (with -i), was
briefly mentioned in §6.5.3. The following is a complete list of types of
nominalisation, which can be used in order to modify a noun.
1. Nominalisation with suffix -i for intransitive subjects and objects (cf. §6.6.1)
2. Nominalisation with -era for agents (cf. §6.6.2)
3. Nominalisation with -naa for intransitive verbs (cf. §6.6.3)
4. Constructions with -nakauru 'those who' for intransitive subjects (cf. §6.6.4)
In the following sections, I will discuss the different types of nominalisation
in their function as heads of the NP. The position of relative clause equivalents
within the N P will be investigated in §6.6.5.

6.6.1 Nominalisation with "-/"

The suffix -/, (which merges with a preceding vowel; cf. §3.4), occurs in
relativising function with transitive and intransitive verbs. With intransitive
verbs, the resulting noun refers to a subject, as is illustrated in (486).

(486) Nominaliser -i with subject function

a)
ku nii ajrinia lauhu-i rausi
there [that outside sit-NOM Sbj ] bite:3ps/E
'It bit the one that was sitting outside.'

b)
lomaj=te hanunaa siiri-i-tQa ne=ra
PSN=FOC [night have-NOM Sbj -only] be-3ps/E=EMF
'Lomai is the one who has got the night.'

With transitive verbs, the nominaliser -i is used in object position. In this

construction, the A argument is marked with the suffix -atg,a 'only', while the
nominalised verb marked with -i represents the Ο argument. The sentence in
(487b) is an example for a plural co-occurring with the nominaliser. In this case,
the suffix -uru (or allomorphs) precedes -i.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
322 Noun Phrase structure

(487) Nominaliser -i with object function

a)
kanu-atQa baha-j ti-ji
[lsg-only ask-NOM 0 bj] give-NEG:3ps/A
' H e did not give me what I asked for.'

b)
kiitQa kiuaauna-ur-i itahe-ri-tQäu=ni
[lsg:only create-PL-NOM o b j ] destroy-IRR-lsg/A=ASS
Ί will destroy those which I have created.'

6.6.2 Nominalisation with "-era"

A lexicalised example that contains the agentive suffix -era is the word for
' G o d ' , which is composed of the lpl/in pronoun kana followed by the agentive
form of the verb for 'create': kana kiu a ami-era literally means 'the one who
created us'. The suffix -era is productively used to refer to an A argument, as
some more examples illustrate. Note that in the examples in (488a), the plural
-uru precedes the suffix -era. A nominal ised verb involving -era can function
as a nominal head or precede the head of an NP.

(488) Agentive nominaliser -era with head function

a)
hajti=te ne ißono ku-ur-era
still=FOC be:3ps/E ayahuasca drink-PL-AG
'There are still those who drink ayahuasca.'

b)
lenone te-ra ari-a niki amuemu-he
food give-AG seek-NTR A D V R S wander-CNT:3ps/E
' H e walked around (in vain) to seek the one who had provided the food.'

There is a tendency for some verbs to occur with reverse order between plural
and the suffix -era with no apparent difference in meaning (cf. (489)). Instead of
-ur-era, the order of the two suffixes is reversed to -era-uru. There are only a
few attested examples that exhibit this variation, but it may be assumed that it is
possible with other verbs as well. Also note that certain morpho-phonological
alternations take place with the examples in (489), even though with some
deviations: in knhjutaa 'bless' (also 'curse'), the root-final /a/ is elided before
the attachment of -era, while this does not apply to letoaa 'send', where the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Relativisation 323

root-final /a/ is retained. In aria 'seek' the suffix-initial Id is elided, which

would result into the intermediate form -ra\ however, the Irl is assimilated to
[tp] after the root-final vowel /i/. In the variants where -era follows the plural
suffix -uru, the initial /«/ of the plural suffix is elided, which results into the
surface form [urera] for /-ura-era/ (also cf. §3.4).

(489) Order variation between agentive and plural suffix

kuhjutaa 'bless' kuhjut-era-uru / kuhjuta-ur-era 'those who blessed

him'
letoaa 'send' -> letoa-era-uru / letoa-ur-era 'those who sent it'
aria 'seek' -> ari-tgauru / ari-ur-era 'those who sought it'

There are no records about the systematic shift from one form to the other. Also
recall that the form with the preceding plural suffix is the typical position
occurring in the nominalisation with -i (cf. (487b)). This position of the plural
suffix corresponds to the non-final position of plural markers on verbs (cf.
§12.2.12). Since nouns do not take any other inflection than plural, the "final"
position of the plural on nouns is inconclusive. However, the fact that plural can
also be expressed after the agentive -era could have evolved under influence
from Spanish.

6.6.3 Nominalisation with "-naa"

The nominalising suffix -naa is.used with two different functions. Firstly, it
occurs as a nominaliser that forms noun describing an action or the result of an
action from a transitive or intransitive verb, such as illustrated in (490). The
derived noun can occur in any argument slot.

(490) Suffix -naa as action nominaliser

a) From intransitive verbs:

amiane-naa ('work'-NOM) 'work' (noun)
noane-naa ('lie'-NOM) 'lie' (noun)
baune-naa ('apply magic'-NOM) 'magic'
ajtöo-hivä ('say'-NOM) 'saying' / 'what he said'

b) From transitive verbs:

ate+su-naa (Tish'+'kill'-NOM) 'fishing'
kanaanaj+hanoa-naa ('child'+'give birth'-NOM) 'child-bearing'
kahjune+hja-naa ('garment'+'wear'-NOM) 'wearing of clothes'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
324 Noun Phrase structure

Note that the nominalising suffix -naa is irregular with the verb root /ajto/
'say': as shown in (490a), the form is realised as ajtöohwä. This form is
exceptional and does not occur with any other verb.
In some cases, the reading as an action noun does not apply, but the
nominalised verb functions as a relative clause (in (491a) as head noun; in
modifying function in (491b)). It is unclear whether the forms are polysemous
for action/result or agentive reading. All attested examples for this form involve
intransitive verbs or the copula nia. However, it is unclear which intransitive
verbs receive abstract meaning and which ones are used in a relative clause.

(491) Nominaliser -naa forming a relative clause

a) As head noun:
na-e hetau nii kana+huaaun-era letono ne-1 ne-naa
say-3ps/E HRS that our.creator envoy be-PRT be-NOM
'... said the one who was the envoy of God'

b) As noun modifier:
lejhii eene kunajte-naa=te u-e
one woman be.sick-NOM^FOC die-3ps/E
Ά certain woman who was sick has died.'

As a matter of fact, the use of -naa in relativising function is restricted to

intransitive verbs and mainly attested with the copula, such as in (491a). This
could be a coincidence or a systematic feature, but an answer to this can only be
given with more data available.

6.6.4 Constructions with "-nakauru"

Another suffix with relativising function is -nakauru. Its translation as 'those

who' indicates that it refers to plural and has an internal morphological
structure. A possible analysis of its structure is to assume that -nakauru is
composed of the nominaliser -naa plus an irregular form of the plural as -kauru.
Interestingly, this form only occurs with intransitive verbs. This also includes
verbs with the reflexive/intransitivising prefix ne-, such as 'sell', which is
composed of the intransitiviser and 'buy', (which in turn is a lexicalised
combination o f ' c h a n g e ' and 'give'; cf. (492a)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Relativisation 325

(492) Occurrence of plural nominaliser -nakauru

a) With 'sell' (with intransitiviser ne-):

heriane=te uu-re=~i rukuele ke ne-kurete-nakauru
maybe=FOC come-IRR:3ps/E=ASS things VLI ITR-buy-those.who
'Maybe the traders will come.' [Lit. 'those who sell things']

b) With copula:
tQäe i=hiniitQal lauri+lauri ne-nakauru enoa-u
also 2sg=together RED+group be-those.who make.enter-IMP
'Also make those who are of different types enter!'

The fact that -nakauru only occurs with intransitive verbs suggests that -naa in
its function as a relativiser also is restricted to intransitive verbs. However, it
remains unclear why the intransitive examples in (490a) receive abstract
meaning.

6.6.5 Relative clauses within the Ν Ρ

Regarding the structure of the NP, the position of nominalising/relativising

elements is subject to some variation. The most typical position of the
relativised constituent is before the head noun, which itself is in phrase-final
position. Any other modifiers of the head noun occur before the relative clause.
This structure can be observed with all four types of nominalisation, as the
examples in (493) show.

(493) Relative clauses as modifiers before the head noun

(Relativised constituents in brackets)
a) With -i and subject reference:
ku ne-rehete-kur-i kat$a-uru
there be-HAB 1 -NOM s b j] man-PL
'the people who used to live there'

b) With -i and object reference:

ii raj kiitQa te-j anofwa presta-u
[2sg for lsg.EMPH give-NOM 0 b j ] knife lend-IMP
'Lend me the knife that I gave to you!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
326 Noun Phrase structure

c) With -naa:
nauta kahe u-naa kat$a=te turu-e ikito-o
PLN from come-NOM man=FOC arrive-3ps/E PLN-LOC
'The man who came from Nauta has arrived in Iquitos.'

d) With -era:
katga ki-ur-era bakaua-kuru
man eat-PL-AG Indio-PL
'the indios who ate people'

e) With -nakauru:
saata marta kahe turu-nakauru katga-uru=te asejte baha-e
PLN from arrive-those.who man-PL=FOC oil ask-3ps/E
'The people who came from Santa Marta asked for oil.'

While the position o f the relativised clause as illustrated above is typical,

examples with a different structure are attested as well. The N P s in ( 4 9 4 )
involve a relative clause that f o l l o w s the head noun. N o t e that examples for as a
relativiser -i with object reference is not attested in this position.

(494) Relativised clause following the head noun

a) With -i and subject reference:

äaka ka=kanaanaj laulauelau-i
'where.is this=child go.by.canoe-NOM Sbj
'Where is that child who was going by canoe?'

b) With -naa:
mi ranuna nitahe-naa kuraa=te marja ne
that girl get.lost-NOM name=FOC PSN be:3ps/E
'The name of the girl who got lost is Maria.'

c) With -era:
lejhn katQa barue ku-era ne-ι ne-naa
one man masato drink-AG be-PRT be-NOM
'one man who drinks ayahuasca'

d) With -nakauru:
letguijka-j kat$a-uru lomaj atane he amu-nakauru
ten-NOMsbj man-PL PSN land VLI walk-those.who
'the ten men who went to the land of Lomai'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Involvement of other word classes as modifiers 327

The following table summarises the order of the different types of relative
clauses with respect to the head of an Ν P. It becomes evident that the position
of the relative clause before the head is more typical, while the occurrence of
relative clauses in post-head position is comparatively rare or unattested for.
Again, it is unknown whether the variation in position of a relative clause has
any impact on the interpretation.

(495) Position of relative clauses within the NP

Before head Ν After head Ν
NOM sbj with -z; object reference yes no
NOM0bj with -z; subject reference yes rare
AG with -era yes rare
NOM with -naa yes rare
NOM plural with -nakauru yes rare

6.7 Involvement of other word classes as modifiers

Adverbs can occur as modifiers within the NP, or, more precisely: they modify
a verb, which is then nominalised. In this context, adverbs precede the
relativised clause; demonstratives typically occur before the adverb. Note that
this is in contrast to their occurrence in clauses, where the position of most
adverbs is relatively free (cf. §18.4.2).

(496) Adverb ßoaelu in NP with relativised clause

a)
ßoaelu ne-nakauru katQa-uru
earlier be-those.who man-PL
'those who lived earlier'

b)
nii ßoaelu urarißa-uru
that earlier Urarina-PL
'those ancient Urarinas'

The following examples show the presence of several adverbs in a relative

clause embedded in an NP. However, the order between a demonstrative and a
relative clause is fixed, as the demonstrative always occurs before the relative
clause, including any adverbs that modify the nominalised verb. In (497a,b), the
demonstrative precedes the locational adverb kau and the temporal adverb ena,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
328 Noun Phrase structure

respectively. In (497c), there are two adverbs: kau occurs in initial position,
while the demonstrative kaa modifies the temporal adverb ena, which follows,
to mean "this now".

(497) Demonstratives and adverbs in the NP

a) [DEM + [ADV + REL]]:

kaa kau ne-nakauru
[this [here be-those.who]]
'these who are here'

b) [DEM + [ADV + [REL+ Ν]]]

kaa ena ne-naa hanonaa
[this [now [be-NOM day]]]
'the day, which is now'

c) [ADV + [DEM + [ADV + [REL + N]]]]:

kau kaa ena ne-nakauru tabauru=te u-uru-ne-re=l
[here [this [now [be-those.who some=FOC]]]] die-PL-NEG-IRR:3ps/E=ASS
'Some of those who are here now will not die.' [NT: Matthew 16:28]

6.8 Summary

For several reasons it is not possible to represent the order of Urarina noun
modifiers in a hierarchical manner:
1. Some classes of modifiers are not homogeneous in that they contain some
members, which occur in one position, while others occur in a different
position.
2. Some classes of modifiers contain members that are variable and occur in
various positions.

The attested combinations of different kinds of elements within the N P are

summarised in (498), where for each modifier type, it is indicated whether it
occurs in prenominal or postnominal position and which other modifiers it can
precede or follow, respectively. Note that the relative order between modifying
nouns and relative clauses is unclear due to lack of examples.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Summary 329

(498) Summary of noun modifier combinations

(N = head; P O S S R = possessor; REL = nominalised verb; M O D = m o d i f y i n g

noun)

Position Ν DEM NUM POSSR A QNF MOD REL

Ν Precedes n/a no no no some some some some
Follows n/a yes yes yes some some some yes
DEM Precedes yes n/a yes yes yes yes yes yes
Follows no n/a no no no yes no no
NUM Precedes yes no n/a no yes yes yes yes
Follows no yes n/a yes no yes no no
POSS Precedes yes no yes n/a yes yes yes yes
R
Follows no yes no n/a no yes no no
QNF Precedes some yes yes yes yes n/a yes yes
Follows some yes yes yes no n/a no no
A Precedes some no no no n/a no yes yes
Follows some yes yes yes n/a yes no yes
MOD Precedes some no no no no no n/a ?
Follows some yes yes yes yes yes n/a ?
REL Precedes yes no no no yes no ? n/a
Follows some yes yes yes yes yes 9 n/a

It is observed that the variability with regard to the position to each other
mainly concerns word types with a high lexical load, i.e. adjectives, modifying
nouns, and nominalised verbs. These types, which happen to relate to noun-like
categories, typically occur adjacent to the head noun, but have a comparatively
free order amongst each other. In contrast, demonstratives, which have a more
functional use, have a relatively fixed position in the NP and need not occur in
adjacency to the noun. Numerals, which also are less noun-like than other
modifiers, are fixed in their position in that they precede the head noun;
however, their position with regard to other modifiers is comparatively free.
Quantifiers, (which are not a separate underived word class, cf. §5.11), exhibit
the most diverse order pattern as they do not constitute a homogeneous group.
While a hierarchical scheme for the order of noun modifiers is not a
favourable approach to their representation, certain tendencies can be
summarised as follows:
1. Demonstratives, numerals, and possessor obligatorily occur in prenominal
position. Demonstratives typically occur on the left edge of an NP.
2. Of all other modifier types, at least some members can occur in prenominal
position.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
330 Noun Phrase structure

3. Demonstratives tend to occur at the left edge of an NP; their position is

relatively fixed. Only quantifiers can precede demonstratives.
4. While numerals occur before the head noun, they follow demonstratives and
possessives. Thus, for those types of modifiers that obligatorily precede the
noun, there is a preferred order: DEM + POSSR + NUM + N.
5. Adjectives, nominalisations (with relative clause function), and other noun-
like modifiers typically occur close to the head of the NP. At least some
members of these three modifier types can occur in prenominal or
postnominal position, but most adjectives occur after the noun, while most
relative clauses occur before the noun.
6. Adjectives precede other noun-like modifiers and relative clauses. However,
a preferred order between the three noun-like modifiers is difficult to draw,
as these groups are heterogeneous.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
7 Possession

This chapter discusses different ways of marking possession in Urarina. In the

NP, this is realised through possessive pronouns or proclitics, with an optional
possessive marker between possessor and the head (cf. §7.1). The different
strategies to express possession in the NP slightly differ from each other with
respect to function or occurrence with certain nouns only, which is summarised
in (513). On the clause level, a verb for 'have' is used, but as an alternative, the
copula can be employed to indicate possession, as is investigated in §7.2. In
addition, there are several affixes that are related to possessive meaning (cf.
§7.3). Furthermore, the issue of alienable vs. inalienable possession is discussed
in §7.4. While inalienability was a contrastive feature in the traditional
language, it is now hardly possible to draw a clear line between inalienably and
alienably possessed nouns. There are several different features that suggest a
remaining certain degree of inalienability for body parts and kinship terms, but
not for other nouns. One such feature is the ability of some nouns to take a
proclitic instead of a full possessive pronoun. However, these nouns do not
exactly coincide with those that exhibit another feature, which is the option of
occurring with a possessive marker. These and other features highlight the fact
that the system of possessive marking in Urarina is in a transitional stage
between synthetic and analytic use of possessives, with a trend to free pronouns,
possibly under the influence of Spanish.

7.1 Possession in the NP

The roles of possessors and possessums in a possessive constructions are

relatively free in that there are no restrictions with regard to the types of nouns
that can function as a possessor or possessum (also cf. §5.1). A possessor can be
human, which is expressed through a full noun, a proper noun, or a pronoun.
But also nouns on a lower scale of the animacy hierarchy can function as
possessors. The following examples show different kinds of nouns found in
possessor function, which have been categorised with respect to animacy. Note
that examples for abstract nouns as possessors are not attested.

(499) Types of possessors (possessor is underlined)

Pronoun: kanu lureri 'my house'
Human: d"z,=uhoa\a komasaj 'your brother's wife'
Proper noun: Oriasi komasaj 'Uria's wife'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
332 Possession

Spirit: kana+hoaaunera letono 'the envoy of God'

Animal: alau bihi 'spider monkey's hand'
Body part: bihi kuturi 'hand's head' (= 'finger')
Plant: hi nori kari 'the bark of a hinyori tree'
Nature-related: nuna kohoajtQakuru 'the inhabitants of the jungle'
Inanimate: lureri alo 'the roof of the house'

Similarly, different types of nouns function as a possessum, including humans,

kin terms, body parts, or inanimate/abstract nouns. However, pronouns do not
occur as possessed nouns; examples with ' G o d ' , proper names for persons, or
nouns referring to spirits are unattested in this function.

(500) Types of possessum (possessum is underlined)

Human: söodaaro-oru kuraanaa ('soldier'-PL + ' c h i e f ) 'the chief of

the soldiers'
Animal: raj reemae (POSS + 'dog') 'his dog'
Body part: kann bihi ( l s g + 'hand') 'my hands'
Plant: raj enua (POSS + 'tree') 'his tree'
Nature-related: raj mm (POSS + 'branch') 'his branch'
Abstract noun: akauru igau (3pl + 'life') 'their life'
Inanimate: akauru raj linterno fwoko (3pl + POSS + 'flashlight' + 'bulb')
'the lamp of their flashlight'

Spatial terms such as locational nouns, which can be preceded by a possessive

clitic, also occur in possessive constructions, as in n=akatia 'next to him' (lit.
'(at) his n e x t ' ) 4 2 The issue of inalienable nouns will be discussed in §7.4. At
this stage it should only be mentioned that proclitics do not obligatorily occur
with nouns in possessive functions.
A possessor always precedes the possessum. The possessum is the head of
the construction, which is determined by the semantic structure of the NP.
Possession may be marked in different ways:
a) Simple juxtaposition without a marker (cf. §7.1.1)
b) Insertion of possessive marker raj (cf. §7.1.2)
c) Marking with proclitics (cf. §7.1.3)
d) Multiple marking by combination of the above strategies (cf. §7.1.4)

42
However, an overt object need not be present, but can be zero. In this case, a 3ps
object will be understood (also cf. §5.1.2).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possession in the ΝΡ 333

7.1.1 Possessive construction without marker

The examples in (501) illustrate zero-marked juxtapositions with possessive

function, which is the most common strategy used to represent possessive
constructions. In the following examples, all possessive NPs are marked with
square brackets.

(501) Possessive juxtaposition

a)
ajtQune kuane ama-ure neba rene
PLN inside take-3pl/E [mother place]
'They took her to the Rio Espejo, to her mother's place.'

b)
alau bihi beeuite nii unee
[spider.monkey hand] cut.off:3ps/E that kinkajou
'The kinkajou cut off spider monkey's finger.'

7.1.2 Possessive marking with "raj "

The word raj was already discussed regarding its functions as a postposition
'for'/'to' (cf. §5.7.4); beside this, it also functions as a beneficiary marker with
indirect objects (cf. §9.5). Another function of raj is as a possessive marker.
Basically, it occurs in two different contexts:
a) As an optional marker between possessor and possessum.
b) As a substitute for the 3ps pronoun in possessor function.
In its function as a 3ps possessive marker, raj intrinsically refers to the
semantic content o f ' h i s ' , 'her', or 'its', which can be derived from its origin as
the nominalised form of the verb raa 'receive': when the root /ra/ is combined
with the nominaliser -i with object reference, it literally means 'what he
received' (cf. §6.6.1). In this function, it typically occurs as a possessive marker
for 3ps with any possessum. Interestingly, it replaces the 3ps pronouns aka (3sg)
and akauru (3pl), which only very rarely occur in possessor function. While a
form such as aka lureri 'his house' has emphatic status and is extremely rare,
the representation as raj lureri is the unmarked form.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
334 Possession

(502) raj marking 3ps possessor

a)
ruaka-1 ama-ure raj loanari asae
carry-PRT take-3pl/E [POSS shed] under
'They carried him and took him under his shed.'

b)
ne-fioaufwajte raj nune
ITR-break.off-3ps/E [POSS branch]
'His branch broke off [the one on which he was sitting]'.

Another way to interpret the occurrence of raj as a possessive marker is to

assume that a form such as raj nune (cf. (502b)) is a short version of
aka raj nune, with the 3ps pronoun aka omitted. This approach would be
plausible in so far as third person pronouns are also systematically omitted
when they function as objects of a transitive verb.
Examples involving raj are attested with all types of nouns, including those
that represent body parts or kinship terms. However, with most of these nouns,
(which may have been inalienably possessed at an earlier stage; cf. §7.4) a 3ps
possessor is unmarked for possession, which is consistent with the general
concept of zero marking for 3ps in Urarina. This also applies to other nouns
when the possessor is known from the context.

(503) Possessive meaning without overt possessor

a) With body part:

kwajtei lele tokiuete
again [tongue] lay:3sg/E
'Again he put it onto his [the envoy's] tongue.' [i.e. not his own]

b) With kinship term:

rautohioe-ϊ beree amae-ama-he
be.calm-PRT [child] RED-take.along-CNT:3ps/E
'He carries his children comfortably.'

c) With other noun:

kuriae bahi-tQuhwa kahe amu-1
distance [defecate-where.to] from walk-PRT
'walking far from his toilet' [about the monkey]

A s briefly mentioned above, pronouns very commonly occur as possessors, as

the examples in (504) illustrate. Further examples are also attested with all other
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possession in the ΝΡ 335

personal pronouns. Their form is identical to the free personal pronouns already
discussed in §5.5.

(504) Pronouns in possessor function

a) lsg:
d^anu=na kann bua basihjau-i-ta
why=INT [lsg bag] steal-2sg=FRS
'Why have you stolen my bag?'

b) 3sg:
ißa-ahe-ne aka komasaj
know-CNT-NEG:3sg/E [3sg wife]
'His wife doesn't know.'

c) lpl/in:
kana bihi afioera-1 hau-aka=ne=te taka-e
[lpl/in hand] follow-PRT throw-lpl/in=CND=FOC encounter-3ps/E
'If we throw it as if it would follow our hand, it hits.'

The above examples illustrate that pronouns occur in simple juxtaposition as

possessors in an NP. However, raj can function as an (optional) possessive
marker between possessor and possessum, which applies to pronouns as well as
to full nouns. Literally, a construction such as katQa raj hireri ('man' + POSS +
'house') may be translated as "the man his house". If we follow the literal (but
lexicalised) meaning of raj as 'what he received', the translation would result in
'the house that the man received'. Therefore, it is not entirely clear whether raj
should be understood as marking the head noun or the possessor. When no noun
surfaces as a possessor, raj could be interpreted as marking the head, as in
raj lureri ('his house'). However, in katQa raj lureri, it appears to refer to the
possessor ('man'), given the literal translation as 'for the man' or 'what the man
received'. Given that raj functions as a postposition in other contexts, one might
adversely assume that it marks the possessor. I will leave the decision whether
Urarina is head or dependent marking with respect to possessives open at this
stage. The examples in (505) illustrate the use of raj as a possessive marker
between possessor and head noun.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
336 Possession

(505) Presence of raj between pronominal possessor and possessum

a)
β aara raj eruari temura sururuhu-u-tga
[[2pl POSS mosquito.net] edge] fix-IMP-EMF
'Fix the edges of your mosquito net!'

b)
akauru raj batia, akauru ruria ßäe nnri=ße-J amu-e
[3pl POSS pot] [3pl tub] already turtle=be-PRT walk-3ps/E
'Their pots and containers already turned into turtles'

Recall that the presence of raj is entirely optional, which is also supported by
e x a m p l e s such as (505b), where two possessive constructions occur in
enumeration, of which only o n e is marked with raj. In fact, the actual difference
between constructions with raj vs. those without it is not straightforward. While
speakers declare both forms as being identical in meaning, there is one
distinction, which will be discussed in §7.4 on inalienable possession. In any
case, the construction without raj is the unmarked, more c o m m o n form, while
its presence between possessor and possessum seems to indicate emphasis of
s o m e kind.
T h e possessive marker raj also occurs after possessors that are full nouns or
demonstratives, and with the interrogative pronoun for ' w h a t ' to mean ' w h o s e ' ,
but it is much less c o m m o n l y used in this context, as compared to its occurrence
after pronouns. In fact, the sentence in (506a) could also be interpreted as 'take
those intestines of his'. The example (506b), which involves the proper noun
' M a n u e l ' , is elicited and the reading without raj is the preferred one.

(506) Optional presence of raj between other types of noun possessors and
possessum

a)
nii (raj) suri ra-u
[that POSS intestines] receive-IMP
'Take the intestines of that one!'

b)
manoel (raj) saabere ne
[PSN POSS machete] be:3ps/E
'It is Manuel's machete.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possession in the ΝΡ 337

7.1.3 Possess ive marking w ith proclitics

When pronouns occur in possessor function, there is the option of using

proclitics, (which also function as object proclitics on verbs). Instead of
kanu lureri 'my house', the form ka=lureri can be used. The difference between
the two constructions is of a diachronical nature, as proclitics are a feature of
the traditional language, whereas free pronouns are preferred in the
contemporary language. However, where proclitics still occur, this is mainly
with nouns that used to be inalienably possessed (mostly kinship terms and
body parts; cf. §7.4).
As mentioned in §3.5, Urarina has the proclitics ka= for lsg (simplified to
[k] before vowel), i= for 2sg (realised as d$= before vowel), and the 3ps
proclitic m= (occurring as n= before vowel).

(507) Possessive marking with proclitics for lsg, 2sg

a) lsg:
ka=hitQana 'my blowgun'
k-enanihja 'my canoe'

b)2sg:
i=hitQana 'your blowgun'
d$=uba 'your mother'

Marking for 3ps possessive is optional and will be zero in many cases (cf.
(503)). While the cliticisation of the allomorph n= to vowel-initial nouns is
rather common, 3ps clitics with consonant-initial nouns are unusual. However, a
few cases for cliticisation of ni= are attested, one not involving a noun, but the
postposition tQoae 'on (top)' (cf. (508c)).

(508) Possessive marking with proclitic for 3ps

a) Allomorph n=:
n=itpau itQuterati-a=ne kujßa
3ps=life save-3ps/D=SUB so.that
'in order to save his life'

b) Allomorph ni=\
raaheniane ni=hitQana te-1, kuane nesari-tQuru-a nerajpe
self 3ps=blowgun give-PRT inside hunt-PL-3ps/D must
'They had to make their own blowguns and hunt with them.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
338 Possession

c) Allomorph m= with postposition:

m=tgoae launeti-Ί hera ne-1 raj nuta hjusi
3ps=on be.sitting-PRT side be-PRT POSS eye scratch:3ps/E
'It [the jaguar] was sitting on top of him and scratched out both his eyes.'

The use of the 3ps proclitic is not restricted to the singular, as becomes evident
from the following examples, where it occurs in plural context. Note that the
absence of a plural marker on the cliticised noun may not be a systematic
feature, as plural marking is highly optional in general. Possession with lpl and
2pl forms cannot be expressed by proclitics, but will be realised by a full
pronoun followed by the head noun (with the possibility of inserting raj).

(509) Possessive marking with 3ps proclitic in plural context

a)
enene n=uhoana ha-uru-a katga-uru=ne hau
nowadays [3ps-field] make-PL-3ps/D man-PL-SUB because
'as the people already make their fields nowadays'

b)
aheri+ku-teru ke n=enanihja te-1
stone+ASC-axe INST [3ps=canoe] give-PRT
'making their canoes with stone axes'

The lsg proclitic ka= can be conjoined with the optional possessive marker raj
to karaj. The function of this form corresponds to kanu raj, with no apparent
difference in meaning. Again, the existence of two alternative forms probably is
the result of diachronic changes, since the use of proclitics is a feature of the
traditional language.

(510) Use of ka=raj

a)
ßäe tQu=ra noane saa-uru-a ka=raj kanaanaj-uru=ra
already CRNT=EMF PTY end-PL-3ps/A [lsg=POSS chi!d-PL=EMF]
'My children are gone already.'

b)
kaa tokoanel ka=raj kamanaj hano-a-e
this compared.to [lsg=POSS child] give.birth-CAUl-3ps/E
'He gave birth to my child like this.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possession in the ΝΡ 339

Interestingly, the lsg proclitic can also be combined with other postpositions. In
one case it is combined with the valency increase marker he. In another, ka= is
observed with the word for 'self to form ka=raaheniane Ί m y s e l f . However,
combinations with the possessive marker raj are not attested for the two other
proclitics.

7.1.4 Multiple possessive marking

Multiple marking of possessive forms is not unusual in Urarina. Basically, there

are two kinds of multiple marking. In the first case, the possessive marker raj is
followed by a possessum that is cliticised with an additional possessive
proclitic, as illustrated below.

(511) Double marking with raj + n=

a)
raj n=itQafioa t$oae kodoonete
[POSS 3ps=rifle] on heap.up:3sg/E
'It [the tree bark] heaped up on his gun.'

b)
raj n=arai-tQuru raj bi-a ktt-e
[POSS 3ps=family-PL] for tell-NTR go-3ps/E
'He went to tell his family'

The reason for this sort of double possessive marking could be explained by the
fact that the nouns involved in the above examples originally were inalienable
nouns and thus obligatorily carried the 3ps proclitic, which is not obligatory at
this stage. Since the proclitic is no longer perceived as a possessive marking, the
insertion of raj takes this function. However, there is an isolated example
shown in (512), which illustrates that double marking is also possible with 2sg
forms. Though the marking is of a different structure, the argument of
lexicalisation cannot hold, since combinations of nouns with first and second
person clitics are clearly perceived as bimorphemic. One can therefore conclude
that multiple possessive marking is an optional feature of the grammar. As far
as its function is concerned, it can only be speculated that some kind of
emphatic function is implied. The (rare) example below contains a free 2sg
possessor pronoun followed by a possessum that is cliticised with the 2sg
proclitic.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
340 Possession

(512) Double marking with 2sg

ii dz,=arai β e-t ne-uru-a=ne rihihei kauru raj

[2ps 2ps=family] be-PRT be-PL-3ps/D=SUB like 3pl for

ere-m-u=ra
talk-DSTL-IMP-EMF
'Since they are like your family, you go talk to them!'

It should also be noted that possession is recursive, i.e. chains of possessors

such as in 'the chief of the soldiers of the Jews' occur in the language. An
example for this was given in §6.3 (cf. (455)), but also cf. (505a)).

7.1.5 Summary

In summary, Urarina N P s involving possessives occur in a variety of ways. In

all cases, the position of the possessor is fixed, as it precedes the possessum.
Possessive marking is optional and can be zero, or involve a marker that
otherwise means ' f o r ' . Alternatively, proclitics can be involved in possessive
marking, with or without the marker raj. The main issue is whether the presence
of these different strategies indicates any functional differences. The high
degree of optionality and the fact that even double marking of possession is
possible might indicate that the differences are due to diachronical changes of
the system. The use of proclitics in general is in a stage of transition towards the
exclusive use of unbound pronouns. This could explain the " m i x " between
f o r m s involving proclitics as opposed to juxtaposition, probably without any
contrastive implications.
An account for the presence vs. the absence of raj turns out to be more
difficult, as the exact function of "emphasis" cannot be determined at this stage.
It is not clear whether the kind of emphasis marked by raj reflects contrast or
focus of some kind, neither do aspects of discourse structure offer any obvious
solution. The origin of this marker is quite evidently related to "receiving"
something and this explains the fact that raj does not occur in possessive
constructions with nouns that used to be inalienably possessed (i.e. they are not
"received" but possibly inherent; cf. §7.4).
T h e table in (513) gives an overview of the different kinds of possessive
marking in the NP. While simple juxtaposition is the default case where two
nouns are involved, the form with raj is typical when the possessor does not
surface as a noun. The variant where raj intervenes between two nouns is
usually observed with formerly inalienable nouns and puts emphasis on the
recipient status of the possession. The same interpretation applies to the forms

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possession with verbs 341

involving karaj, but this may be understood as a transitional form between the
traditional and the innovative language, as aspects of both variants (i.e. clitics
and unbound markers) have been combined. Nouns that are only marked with a
proclitic are typical for the traditional language and usually occur with formerly
inalienable nouns. Double marking is a comparatively rare phenomenon and
merely supports the hypothesis that this system in transition is subject to certain
instability.

(513) Summary of possessive forms in the NP

Marking Example Function/Restrictions Occurrence

N+N kanu lureri default, with all nouns common
'my house'
raj + Ν raj lureri 3ps (non-emphatic), with common
'his house' all nouns
Ν + raj + Ν kanu raj lureri emphatic, mainly with optional
'my house' formerly alienable nouns
ka=raj + Ν ka=raj lureri emphatic, recipient optional, transition
'my house' from traditional to
innovative
Proclitic + Ν ka=lureri with formerly inalienable traditional language
'my house' nouns
raj + raj n=itQafioa (relic) rare
proclitic + Ν 'his rifle'

7.2 Possession with verbs

Urarina has two verb-based strategies that refer to possession. One is with
siiria, a transitive verb for 'have'; the other involves the copula in combination
with a free or bound pronoun followed by the possessive marker raj.

7.2.1 Copula strategy

Possessive marking that involves the copula can be in the affirmative or

negative form (also cf. §9.6.5). Literally, this construction can be translated as
'... is (not) (there) for me' ('is mine') for the lsg form. The pronoun can be
realised as a full pronoun or as a proclitic in this case (cf. (514a,b)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
342 Possession

(514) Copula strategy

a) Affirmative, lsg full pronoun:

atauari=te kamt raj ne
chicken=FOC lsg POSS be:3sg/E
'The chicken was mine.'

b) Affirmative, lsg proclitic:

m-a ka=raj teru
be-3ps/A lsg=POSS axe
Ί have an axe.'

c) Negative, lsg proclitic:

rti-ji ka=raj kuriki
be-NEG:3ps/A lsg=POSS money
Ί don't have money.'

d) Negative, lpl/ex pronoun:

ni-ji kanakaanu raj lenone
be-NEG:3ps/A lpl/ex POSS food
'We don't have food.'

For other forms than lsg, cliticisation is not possible, but the combination of
copula, full pronoun, and raj applies, as illustrated with lpl/ex in example
(514d).

7.2.2 Possession with'have'

The other strategy for indicating possession with verbs involves the verb siiria.
Unlike possessive verbs in many other languages, siiria is a regular transitive
verb without any apparent paradigmatical restrictions. It can take a causative
form and is regularly inflected for most (if not all) grammatical categories; there
even is a derived form of the verb that involves the intransitiviser ne- (cf.
§16.2): ne-siiriatia means 'get married', which might be interpreted as a kind of
intransitive form of "having" (a partner).
The possessive construction with siiria involves a "closer" kind of
possession than the copula construction. Basically, it refers to ownership or
long-term possession, as opposed to the copula strategy, which seems to imply
short-term possession. The expression for 'be pregnant', for instance, can be
represented as kanaanaj siiria 'have a child'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possessive derivation 343

(515) Use of siiria

a)
uhuae+kuturi kahe haja siiri-t$uru-a hau
deer+head from horn have-PL-3ps/D because
4
as they had a horn (made from) a deer head ...' [i.e. the village owned one]

b)
nihjauria nekajritQa-i=ße kuriki siiri-i=ße
don't suffer-2ps=NEGF money have-2ps=CND
'Do not suffer when you have money!'

T h e d i f f e r e n c e b e t w e e n siiria and the c o p u l a strategy is f u r t h e r illustrated by t h e

f o l l o w i n g e x a m p l e . In e x a m p l e (516a), s o m e o n e is characterised as the o w n e r
o f a m o n k e y , implying that he k e e p s it as a pet. In contrast, the second s e n t e n c e
implies that the m o n k e y will serve as a meal.

(516) Contrasting possession with verbs

a) With siiria:
enejtQH siiri-tpäu
monkey have-1 sg/A
Ί have a monkey [permanently, as a pet]'.

b) With copula:
ni-a ka=raj enejtQu
be-3ps/A lsg=POSS monkey
Ί have a monkey [temporarily, as a meal].'

7.3 Possessive derivation

7.3.1 Verb suffix "-oka"

T h e verbaliser -oka is a productive derivational s u f f i x w h i c h derives a v e r b f r o m

a n o u n root. 4 3 M o s t typically, the resulting v e r b implies that the subject o f the
n e w l y created v e r b refers to a possessor. For e x a m p l e , if a v e r b is derived f r o m
' s i s t e r ' , it will m e a n ' t o h a v e a sister' (and not ' t o be a sister', which w o u l d be
a n o t h e r logical possibility; cf. (517a)). A s the s u f f i x is relatively infrequent,
m o s t o f the f o l l o w i n g e x a m p l e s are elicited. T h e f o r m s illustrated in ( 5 1 7 )

43
-oka has the allomorph -koa. Also note that the final /a/ may be subject to elision when
followed by another vowel.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
344 Possession

include terms that refer to kinship and body parts, which are most typical in
their occurrence with the verbaliser. Kinship terms also include the words for
' w i f e ' , 'husband', and 'child'.

(517) Derivation of kinship terms and body parts with verbaliser -oka

a) Kinship terms:
baj-oka
sister-VBL:3ps/A
'He has a sister.'

b) Body parts:
bihi-oka
hand-VBL:3ps/A
'It has hands.'

Similar to possession expressed by siiria, the derivation with -oka refers to long-
term possession. The use of this suffix is not restricted to the two types of nouns
mentioned above, but also occurs with other nouns, as the examples in (518)
illustrate. Further nouns to which -oka can be attached include enamhja 'canoe',
safioato 'shoe' (loan from Spanish zapato), saabere 'machete' (loan from
Spanish sable), and hiriki 'money' (loan from Quechua qullqi), which shows
that the suffix is not restricted to native words. What all these nouns have in
common is that they represent important possessions to the owner. Even though
they cannot be characterised as inalienable, they are related to long-term
possession. The functional difference between the suffix -oka and the verb siiria
cannot be determined at this stage.

(518) Derivation of other nouns with verbaliser -oka

a)
lureri-oka
house-VBL:3ps/A
'He has a house.'

b)
teru-k-ure
axe-VBL-3pl/E
'They have axes.'

Elicited derivations involving nouns that are not typically subject to long-term
possession, such as katai 'fishing hook' or barue 'masato' (cassava beer) are
considered "funny", but possible. As a second thought, these could be
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possessive derivation 345

interpreted as referring to someone who trades these items, but speakers

suggested that this would be even better expressed by forming a clause with
siiria, i.e. katai siiria ' h e has hooks'.
It should also be noted that -oka has the allomorph -koa. Typically, -koa
replaces -oka when the respective noun ends in a vowel sequence, a diphthong,
or in a long vowel. Thus, the process involved could be described as a case of
metathesis. Otherwise, metathesis is not a regular process of Urarina phonology.

(519) Occurrence of -koa as a verbaliser after vowel sequences

beree beree-koa (*beree-oka) 'He has a child.'

usei usei-koa (*usei-oka) 'It has a tail.'
kuraa kuraa-koa (*kuraa-oka) 'He has a name.
reemae -> reemae-koa (*reemae-oka) 'He has a dog.'

7.3.2 Prefix b-

There are a few words that occur with the prefix b-. One aspect that makes it
difficult to define the function of this element is that it is not productive and
only occurs with a few examples, mainly nouns. The reason for discussing it in
this chapter is that b- appears to be related to some kind of modifying
associative function. The other problem is that its occurrence is not limited to
nouns only, but is also found on one verb. Urarina speakers characterise the
difference in meaning between words that carry b- and those that do not as
being related to the differences between traditional and innovative language.
However, a complication is that its presence or absence is not consistently
linked to the traditional vs. the innovative language, as will become clear below.
The examples in (520) show the occurrence of b- as attached to a noun that
follows a proper noun; its function is to characterise the status or kind of the
preceding proper noun. 44

(520) Occurrence of b-

a) With atane 'land'

galilea batane '(the land of) Galilee' [NT: Matthew 2:22]
hodea batane '(the land of) Juda' [NT: Matthew 3:1]
asia batane 'Asia' [NT: Acts 2:9]

44
Due to the fact that the use of b- is not productive, the list of examples given in this
chapter can be considered nearly exhaustive.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
346 Possession

b) With eene 'woman'

lomaj beene 'the woman Lomai'
maria beene 'the woman Maria'
käsi-jiel beene '(being) seven females'

H o w e v e r , with other examples, which do not involve proper nouns, the facts are
less clear. For instance, arasihje 'star' is attested without any d i f f e r e n c e in
m e a n i n g as barasihje in the same text. T h e distinction w a s characterised by the
(elderly) speaker as a variation, since people mix words of the traditional and
innovative language: whereas the forms with b- are described as part of the
traditional language, the innovative language tends not to attach this prefix to
nouns.
T h e prefix b- is also observed with at least one verb (no more are attested in
m y data), the verb (b-)asihjaua 'steal'. Again, the difference is linked to
diachronic factors, but, surprisingly, while basihjaua is characterised as the
standard term in the innovative language, asihjaua is considered part of the
traditional variation.

(521) Occurrence of b- with the verb for 'steal'

a) Without b-:
kwatia asihjau-ßa-ü
not steal-NEG-lsg/E
Ί have not stolen it.'

b) With b-\
unee bua basihjau-a alau=ne
kinkajou bag steal-3ps/D spider.monkey=CND
'When spider monkey stole kinkajou's bag' [Story title]

c) ' T h i e f :
basihja-uru-re=l asihjauto-oru
steal-PL-IRR:3ps/E=ASS thief-PL
'The thieves will steal it.'

N o t e that in the example (521c), the word for ' t h i e f , asihjauto does not carry
the prefix (and there is no variant in which it would). Thus, the developmental
stages by which language change has taken place remains enigmatic. A related
question m a y be in what w a y the nouns baaso and aaso for ' b a d ' and ' f o o l i s h '
(already discussed in §5.4.2), are related. In this case, there is a clear semantic
distinction; however, the relationship between the two words can hardly be

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possessive derivation 347

denied, baaso seems to be a more general term, while aaso appears to be

restricted to humans.
In summary, there is no satisfying answer to the question in what way the
element b- changes the meaning of words. Since it is not a productive
morpheme, its relevance for the grammar is only marginal. An interesting
aspect, however, is that its presence with one word class is characterised as a
feature of a previous developmental stage of the language, whereas this is
described as a more recent feature with the other.

7.3.3 Associative prefix "ku-"

Another morpheme that appears in prenominal position is the associative

marker ku- (or variant AD-), even though it is not quite clear whether this is an
adequate term to describe it.45 Associative constructions typically are related to
possessives - which does not exactly match the function of ku-. While it occurs
between two nouns, it does not imply a possessive relation between the two. It
seems that in most cases, it actually functions as a conjoining element of two
nouns that form a compound. However, there are only a small number of words
that occur with ku-, and they are not common members of any semantic or
morphological class whatsoever. A look at the occurrence of ku- reveals that all
words in which ku- is found are compounds that involve two nouns.

(522) Occurrence of ku- in compounds

a) aheri+ku-teru 'stone'+ASC-'axe' 'stone axe'
b) obana+ko-biri 'cl. peccary'+ASC-'magic' 'peccary magic' [i.e. magic that
helps to hunt cl. peccaries]
c) laano+k-uhvana 'cassava'+ASC-'field' 'cassava field'
(Variant:
laano+kokana)

In all three examples cited in (522), the use of ku- is productive, as the first
noun can be exchanged for another noun with a similar semantic structure. For
instance, the same constructions could also be composed for 'wood axe', 'tapir
magic', or 'banana field', respectively.
A feature that all these examples share is that none of the examples in (522)
expresses a possessor-possessum relationship; on the contrary, the second
component of each compound appears to specify the first noun. For instance, it
is a "magic for peccary (hunting)", a "field for cassava (planting)", thus

45
The prefix is simplified to [k] before /u/ and to [kw] before other vowels.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
348 Possession

specifying the function of the first noun; with aheri+ku-teru, an "axe (made) of
stone", the first noun refers to the material. For the time being, I will refer to
compounds involving hi- as "associative type" compounds (ASC). Examples
such as the ones in (522) are in contrast to other compounds such as listed in
(523), which do not involve the prefix hi-.

(523) Compounds without ku-

a) tihja+suri 'leg'+'belly' 'calf

b) aaune+baka 'plantain'+'soup' 'plantain soup'
c) akau+hjane 'water'+'urine' 'urine water'[referring to urine that became
water]

Except for the example in (523c), all of the compounds in (523) can be analysed
as possessive constructions, which would also apply to further examples. 46 This
also demonstrates that hi- is not a marker generally used in compounds. It
should also be mentioned that all of the words that occur in second position of
(522) occur without hi- in other contexts. Interestingly, there are a couple of
words that contain hi- in their citation form and in some other environments,
but they lose it under specific conditions. The two words that are attested for
this particular behaviour with regard to hi- both refer to body liquids, as
illustrated in (524).

(524) Occurrence of ku- in citation form

a) 'Fat':
kio-erene, erene 'fat' (citation form)
raana ho-erene 'pig fat'
raj ho-erene 'his fat'
oliba ho-erene 'olive oil' [Old Testament: Exodus 29:40]
kaa erene 'this fat'
erene somahaj 'white fat'

46
There are different possible analyses for the example 'water'+'urine' in (lc). Due to
the particular nature of the resulting liquid, any of the two components could be
interpreted as the head. If 'urine' is the head, the interpretation 'urine for (turning
into) water' would match the suggested analysis. If 'water' were the head, this
compound would be very unusual since there are no other known compounds in
which the first component functions as the head.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Possessive derivation 349

b) 'Blood':
hoitgana 'blood' (citation form)
raana hu-itQana 'pig blood'
kanu hu-itgana 'my blood'
nii itgana 'that blood'
itQana somahaj 'white blood'

Based on the above forms, the following observations can be made:

1. Even though the element hi- must be regarded a prefix, it can occur in the
citation form o f the words for 'blood' and 'fat' (itQana is unattested as a
possible citation form for 'blood').
2. In possessive constructions and in compounds that imply a possessive
relationship, hi- is retained.
3. In combination with demonstratives and when the noun is modified by a
(deverbal) adjective, hi- is absent.

It is evident that here, the function of hi- differs from its function observed with
the examples in (522), in particular with regard to possessive constructions:
while the forms in (522) strictly exclude hi- from constructions that convey
possessive meaning, the prefix clearly occurs in this context with 'blood' and
'fat'. The only reasonable way to explain this difference would be to assume
that the combination o f hi- with these two words is being lexicalised for some
reason.
However, what both examples from (524) have in common with other nouns
is the fact that they do not carry the prefix in combination with demonstratives
and with adjectival modifiers. The differences and similarities are summarised
in the table below.

(525) Presence vs. absence of ku-

Citation form ASC type POSS type POSS With With

compounds compounds constructions DEM modifiers
tern 'axe' yes no no no no
nemuri 'lake' yes no no no no
biri 'magic' yes no no no no
ahoana 'field' yes no no no no
hoerene, erene 'fat' no yes yes no no
hoitg,ana 'blood' no yes yes no no

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
350 Possession

7.4 Inalienable possession

The term "inalienable possession" has been mentioned in various passages of

this chapter, without going into detail as to what extent this feature is relevant
for the distinction of nouns. Recall that inalienably possessed nouns were not
mentioned as a subclass of nouns in §5.1. The reason for this is that one can
suspect that Urarina had inalienably possessed nouns at an earlier stage, but this
feature has been largely lost over time. However, there still are some relics that
relate to it and this section is meant to explore the fragments that exhibit traces
of inalienable possession.
Based on cross-linguistic comparison and on logical considerations, inalienable
possession is most common with body parts and kinship terms, as these are not
normally possessed in the sense of legal ownership (cf. Chappell & McGregor
1995). In fact, this is the area in which Urarina still shows obvious traces of
inalienability. Some traces such as the ones discussed in the previous section ο
not represent clear-cut boundaries between alienable and inalienable nouns, as
another distinctive feature is long-term versus short-term possession, which
exhibits a certain degree of overlap with alienability. If we assume that
alienability was a distinctive feature at a former stage of language development,
the compatibility of proclitics with certain nouns should play a role, too, as
procliticisation is a feature being lost in the innovative language. The table in
(526) provides a (non-exhaustive) list of various body parts, together with the
respective proclitics for all three persons, where applicable.

(526) Body parts and cliticisation

Citation Gloss lsg 2sg 3sg

form
alaarihja 'chest' k=alaarihja d 3=alaarihja n=alaarihja
(n)anaae 'leg' k=anaae d^=anaae n-anaae
(n)usitQue 'fingernail' ka=jsitQue d^=usitQue n=usitQoe
nulunu 'neck' ka=lunu d^-ulunu raj nulunu
suri 'intestines' ka=suri i=furi raj suri
hanuri 'back' ka=häaori i=hjaaori raj hanuri

As exemplified by the terms for body parts in (526), there is a high degree of
irregularity between the various nouns. First of all, it must be stated that all
nouns can also occur with the full pronoun instead of the proclitic (e.g.
kanu suri 'my intestines'). Some nouns that begin with a vowel, such as
alaarihja 'chest', take shortened allomorphs of the proclitics to indicate
possession while their citation form is without the 3ps proclitic n=. Others, such
as (n)anaae 'leg' differ from that in that they have two possible citation forms

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 inalienable possession 351

(with or without n=). Yet other vowel-initial nouns always occur with /n/ in
their citation form while they occur without it when preceded by a possessive
proclitic for lsg/2sg (e.g. nuluntt 'neck'). Most nouns that begin with a
consonant (e.g. suri 'intestines') take (unreduced) proclitics for lsg and 2sg but
the possessive marker raj for 3sg (cf. §7.1.2). However, there are a number of
such nouns that are subject to further phonological changes when they occur
with a clitic; hanuri ( ' b a c k ' ) is only one example for this. Further examples for
such changes will be displayed in (529).
In the same way as body parts, kinship terms represent a semantic class of
nouns that can occur with a possessive proclitic, especially among older
speakers. Examples for this are given in (527).

(527) Kinship terms and proclitics

With proclitic (traditional) Citation form (current) Gloss

ka-tana tana 'my (maternal) uncle'
ka-lana lana 'my husband'
k=uhuala ukiuala 'my younger sister'
k=akano akano 'my daughter-in-law'

While body parts and kinship terms represent natural groups of nouns that could
be expected to be inalienable, other nouns may occur with possessive proclitics
as well. These include items that are significant possessions relevant for daily
life, such as the words for ' b l o w g u n ' , ' f i e l d ' , and others, as illustrated in (528).
Other nouns of this type refer to types of fruit, fields, firewood, and even pets.
In contrast, wild animals that are not held as pets, such as 'jaguar' cannot be
combined with a proclitic. One speaker also rejected the cliticisation of ka= to
' p e c c a r y ' , until he remembered that someone in the village had actually raised
one; in this case, the occurrence with the proclitic is acceptable. The example
' m y grasshopper' is based on the supposition that this specific one will be used
as fishing bait. In any case, long-term vs. short-term possession is not a relevant
factor here, rather whether something can be possessed or not.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
352 Possession

(528) Other nouns with proclitics

Citation form With proclitic Gloss

a) reemae ka=reemae 'my dog'
b) hitQana ka=hitQana 'my blowgun'
c) itQafiua ka=jtQafioa 'my rifle'
d) laano ka=laano 'my cassava'
e) fwei ka=fwei 'my firewood'
f) arusu k=arusu 'my rice' [loan from Spanish arroz]
g) ela k=ela 'sleeping mat'
h) sisiitQU ka=sisiitQu 'my grasshopper'

Note that even loans such as 'rice' can occur with a proclitic (cf. (528f)). In this
case, one could assume that the loan from Spanish arroz is quite old and thus
was subject to procliticisation before the loss of proclitics began. Alternatively
(or in addition to that), the loss of proclitics might be a relatively recent
development. In any case this emphasises the transitional state of many
examples.
As mentioned above, some nouns have different roots in their citation form
as compared to their occurrence with proclitics. The words in (529) represent
examples for this alternation. Note that these are not regular phonological
alternations of current Urarina but examples for systematic diachronic changes.

(529) Citation forms and procliticisation

a) Body parts:
With proclitic Citation form Gloss Change
(traditional) (current)
ka=buu buku 'my bone(s)' Insertion of /k/
ka=naari nahari 'my liver' Insertion of /h/
ka=tia tihja 'my foot' Insertion of /h/
k-amihjaneo amianeo 'my work' Insertion of /h/
ka=hx>ii bihi 'my hand' /kw/ /b/; insertion of /h/
ka=tg,aitQU tahitQU 'my ankle' /tς,Ι /t/; insertion of /h/
ka=lueri lureri 'my house' Insertion of Irl

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Inalienable possession 353

b) Kinship terms and others:

With proclitic Citation form Gloss Change
(traditional) (current)
ka=banuhi banui 'my son-in-law' Insertion of /h/
ka=nemae nebaene 'my maternal aunt' /m/ /b/; Insertion of /ne/
ka=kalabi kalaui 'my son' Ibl /u/ (or [w])
ka=kaaunu kakunu 'my daughter' Insertion of ikJ
ka=ktuaj baj 'my sister' /kw/ -> lb/
ka=fioaara fiuanara 'my bananas' Nasalisation -> /n/

A recurring feature in these examples is the absence of /h/ in the older

(procliticised) words, whereas /h/ is present in the current citation forms, (which
are now preceded by a free pronoun in possessive constructions and do not
undergo alternation at the present stage). However, further kinds of changes are
found with some other nouns, such as the change from /kw/ to /b/. As a general
principle, words in the innovative language become longer through the insertion
of additional material. Unfortunately, the list of available examples for this is
too short in order to draw valid conclusions about the original forms. However,
the high degree of variation between the various body part terms with respect to
cliticisation as observed in (526) and the changes shown in (529) confirm that
cliticisation is a process in transition, suggesting a development towards more
synthetic structures.
On the basis of the data presented above it is still difficult to distinguish
between nouns that used to be inalienably possessed and those that were not,
since a large number of nouns may occur with a possessive proclitic, given the
appropriate context. There is another feature that is evidence for the transitional
state of the possessive system in the innovative language: the optional
possessive marker raj in combination with a preceding pronoun, as in kanu raj
' m y ' , only occurs with formerly alienable nouns, but never with formerly
inalienably possessed nouns (also cf. §7.1.2). As illustrated in (530),
kann raj teru (or alternatively kanu teru) is a possible expression for 'my axe',
while ' m y foot' can only be realised as ka?iu tihja or ka=tihja, but not as
*kanu raj tihja.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
354 Possession

(530) Use of raj as criterion for distinction of formerly [+/-alienable] nouns

Without raj With raj Gloss Semantic

class
kamt kuturi *kanu raj kuturi 'my head' Body part
kanu tihja *kanu raj tihja 'my foot' Body part
kam suuhua *kanu raj suuhua 'my heart' Body part
kanu buku *kanu raj buku 'my bones' Body part
kanu ukwala *kanu raj uhoala 'my younger brother' Kinship
kanu beree *kanu raj beree 'my child' Kinship
kanu kalaui *kanu raj kalaui 'my son' Kinship
kanu komasaj *kanu raj komasaj 'my wife' Kinship
kanu amiane-o *kanu raj amiane-o 'my work' Abstract
kanu amiane-naa kanu raj amiane-naa 'my work' Abstract
kanu lureri kanu raj lureri 'my house' Property
kanu ukwana kanu raj uhoana 'my field' Property
kanu enardhja kanu raj enamhja 'my canoe' Property
kanu teru kanu raj teru 'my axe' Property
kanu reemae kanu raj reemae 'my dog' Property
kanu atane kanu raj atane 'my land' Property

From the above list, several conclusions can be drawn:

1. Formerly inalienably possessed nouns match the semantic classes for body
parts and kinship terms, as expected. Apart from blood relationships, also the
association to terms such as ' w i f e ' , ' h u s b a n d ' , and 'child' is regarded as
kinship. 4 7 (This also includes in-law relationships, which are not attested
here; however they exhibit the same behaviour with respect to proclitics as
other kinship terms).
2. Certain abstract nouns were regarded as inalienable. The word amiane-o,
which contains the (rare) nominaliser -o is inalienable, whereas the other
word for work, amiane-naa (containing the nominaliser -naa) is not. Due to
the absence of further examples it remains difficult to determine how a
distinction can be drawn.
3. Other nouns that would suggest a high degree of inalienability, such as
essential properties, tools, or pets, do not match the same criteria as body
parts and kinship terms. It can therefore be concluded that the relics of
inalienability refer exclusively to humans or human attributes and possibly a
f e w abstract nouns.

47
There are different words for 'child'. With kanaanaj 'child' (a generic, not strictly
relational term) raj can be used.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
8 Number marking

Singular is the formally unmarked form for number marking. Plural can be
marked on most nouns, including mass nouns or abstract nouns. In these cases,
the plural refers to 'different kinds o f (cf. §5.1 and examples in (534)).
Therefore, it is not practical to distinguish between countable and uncountable
nouns. Principles of number marking correlate with semantic subclasses of
nouns, in that most nouns related to location do not take plural. Plural is marked
on verbs, but can also occur on some postpositions (cf. §8.2). A plural marker
may cross-reference a constituent, most typically the subject of a transitive or
intransitive clause, or the object of a transitive clause under certain conditions to
be analysed in §8.3.
Note that the use of the term "plural" here is a simplification. Strictly
speaking, the term "non-singular" would be more precise for some areas of the
grammar. Specifically, this regards PSC verbs (verbs of posture, shape, and
colour; cf. §9.2.3), where a distinction between different types of number
marking is more complex - here, the term "non-singular" was adopted in order
to distinguish plural (which is more than four or five here) from dual, paucal,
and other forms, which are "more than one", but less than plural. For all other
verbs, this distinction is irrelevant, i.e. plural is in fact "any more than one". As
terms that would result from the "non-singular" terminology (such as "non-
singular-distributive" and non-singular-object") would be too inconvenient as to
use them throughout this entire grammar, I chose to employ the term "plural"
(glossed " P L " ) out of convenience, for all contexts other than with PSC verbs.
Urarina has two predominant plural suffixes. One is -t$e, which occurs with
verbs only; the other is -uru. The latter functions as a noun plural marker, but it
is found in a number of other contexts as well. For instance, it also functions as
a pluraliser in verbal marking for 3ps. Other suffixes referring to number are the
plural object marker -naha for transitive verbs, the distributive suffix -ahua,
and the plural marker -ana for stative verbs.

(531) Overview of plural markers

-toe 2ps subject (all verbs)

-uru 3ps subject (transitive and active intransitive verbs)
-ana 3ps subject (stative intransitive verbs)
-naha 3ps object (transitive verbs)
-ahoa 3ps subject distributive (intransitive verbs)
3ps object distributive (transitive verbs)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
356 Number marking

This chapter will analyse the use of these suffixes in detail. One thing that all
plural markers have in common is the tendency to be optional. While plural
marking on nouns or as an instance of morphological agreement on verbs may
be the more frequent option, it is not uncommon to use the singular form if
plurality is implied by the context. In relation to this, the correlation between
number and humanness will be examined (cf. §8.7).
Further peculiarities of number marking occur with position, colour, shape,
and colour (PSC) verbs. These distinguish different forms for singular, dual,
paucal, greater plural, and some for distributive plural, as investigated in
§10.2.1. As this way of expressing number has derivational character, it is
discussed in the chapter on word formation.

8.1 Verb plural for 2ps

The verbal suffix -t$e occurs as a plural marker for contexts involving 2ps,
which involves 2pl subject marking on verbs in the declarative mood and for
imperative marking. It also functions as a politeness marking enclitic in
conversation between members of the opposite sex and in specific in-law
relationships (cf. §13.2).

(532) Plural suffix -t$e

a) With imperative:
lomaj kalaui raa-ni-tQU-tQe=ra
Lomai son receive-DSTL-IMP-PL=EMF
'Go and get (PL) the son of Lomai!' [Talking to a group of people]

b) With 2pl declarative (here in the subordinate clause):

bute haa-u=te kuane nit$ute-i-tQe=ne kujßa=ra
boat make-IMP=FOC inside escape-2ps-PL=SUB so.that=EMF
'Build a boat so that you (PL) escape in it!'

In conversations between certain in-laws and between men and women, the
h o m o p h o n o u s enclitic =t$e is used as a polite form, independently of person.
For instance; in (533a), =t$e is attached to a 3sg form 'the trap snapped', since
the speaker is addressing his son-in-law. In (533b), it is cliticised to the neutral
form of the verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Plural marker for 3pi 357

(533) Suffix =tge as a politeness marker

a) Talking to father-in-law:
ßäe raj nakana raj huru-a=tge bitoa, na-a=ne sajhjel
already POSS son-in-law for snap-3sg/A-PLT trap say-3sg/D=SUB although
'Though he said to his son-in-law "the trap has already snapped" ... [he did not hear
him].'

b) Talking to woman:
hoa bi-a=toe
don't tell-NTR-PLT
'Don't tell him!' [Conversation between man and women]

The politeness marker can also be attached to introducers, as is investigated in

§13.2. The fact that =tg,e can be combined with different word classes supports
the view that it is homophonous with the plural suffix for 2ps, but functions as a
separate morpheme.

8.2 Plural marker for 3pl

The plural marker -uru has a wider distribution than the suffix -t$e, since it can
occur on nouns, postpositions, and verbs, where it functions as a cross-
referencing plural marker for 3ps subjects. In (534), the wide range of the
occurrence of -uru (or allomorphs -kuru, -t$uru) is illustrated with different
subclasses of nouns. Except for relational nouns, place names, and some other
locational nouns, there are no semantic restrictions on the type of noun that can
be pluralised (cf. §5.1). When the plural suffix is attached to mass nouns such as
'water', 'fish', or 'banana', it refers to various kinds of referent. The plural
marker -uru also occurs with demonstratives and in the free pronoun for 3pl
aka-uru, (which is derived from aka for 3sg; cf. §5.5).

(534) Plural suffix -uru with different subclasses of nouns

a) With mass nouns:

kaa akau-uru
this water-PL
'these waters' [i.e. rivers, ponds, lakes]

b) With abstract nouns:

leotQa-uru amiane-naa-kuru
other-PL work-NOM-PL
'other works' [i.e. work projects]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
358 Number marking

c) With demonstratives:
mi-t$uru=te eskiuela raj ne
that-PL=FOC school for be:3ps/E
'Those (things) are for the school.'

d) With person names:

makusi-tQuru
PSN-PL
'the Macusis'

e) With adjectival nouns:

enene ne-nakauru enamanaa-kuru katga-uru
nowadays be-those.who young.man-PL man-PL
'the young people who live nowadays'

Note that plural marking on person names such as in (534d) is elicited and not
common otherwise. In this form, it refers to several families with the same
name. The plural marking on adjectives such as in (534e) is highly optional
(and elicited as well). In fact, the variant without the plural suffix is preferred.
The plural marker -uru is also observed with most postpositions, although
its use in this context is rather uncommon. It refers to the plural of the
postpositional argument, e.g. 'from the towns', 'into the prisons', etc. However,
these examples (both form the NT; (535c,d) are elicited) are described as "rare"
by Urarina speakers They are described as implying some kind of distributive
plural, such as 'out of each town', 'into various prisons', etc, as indicated in the
glosses. Example (535d), is not referring to several tables, but rather to different
areas on the same table. One can conclude from this that the (rare and optional)
presence of a plural suffix on postpositions emphasises the individuality of the
referents.

(535) Plural suffix -uru with postpositions

a) kahe:
nese-uru kahe-uru ku-ahva-i saku-ure atane tQoae katQa-uru
town-PL from-PL go-DSTR-PRT follow-3pl/E land on man-PL
'The peopled followed him out of the towns to the hills.' [NT: Matthew 14:13]

b) kuane:
karesero kuam-uru te-rate-naha-reheto-o=lu
prison inside-PL give-CAU2-PLO-HABl-lsg/E=REM
Ί used to throw them into prisons.' [NT: Acts 22:4]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 PIural markerfor 3pl 359

c) kurerna:
lenone kurenia-kuru fzoanara te-ure
food for-PL banana give-3pl/E
'They gave bananas for various food items.'

d)tgoae:
mesa tQoae-kuru=te ne kuruataha-j radiofioonia(-kuru)
table on-PL=FOC be:3ps/E two-NOM radio(-PL)
'The two radios are on the table.' [i.e. on different parts of the table]

Similar examples for the plural suffix occurring on postpositions are also
attested for other postpositions, such as kaihje 'after', raj 'for', and ajjia 'with',
but not for the postpositions hana 'inside', asae 'under', and tiai 'until'. Since
the plural in combination with postpositions is rather infrequent in general, this
should not be taken as a reliable way to further classify postpositions.
A more regular use of -uru is as a 3ps plural marker on verbs. This applies
to all declarative (affirmative and negative) forms for all person inflection types
(A/E/D-forms), and to 3ps imperative forms (cf. §15.1.3). Interestingly, the
order of suffixation is different from the marking for 2pl. While the pluraliser
-t$e follows the suffix for 2ps, -uru precedes the person markers (cf. §12.2).

(536) Plural suffix -uru as 3pl marker

a)
hoara-he-uru-a=lu katga-uru
see-CNT-PL-3 ps/ A=REM man-PL
'The people were watching him.'

b)
enanihja kuane amu-uru-i~lu
canoe inside walk-PL-NEG:3ps/A=REM
'They did not go by canoe.'

Note that -uru has the allomorphs -kuru and -tQuru, as was discussed in §3.1.1.
Further alternations include the elision of the final /«/ before any suffix that
begins with /e/, such as in / m « k ö - « r a - e / ('catch'-PL-3ps/E) 'they caught her',
which realised as [ m u k a u r e ] . The initial /«/ of the plural suffix is assimilated
to [u] or [o], when the respective vowel precedes it: /molo-ura-a/ ('cut'-PL-
3ps/A) 'they cut it' surfaces as [moloorua].
The suffix -uru is also found in nominalisations. In a verb that is derived
with the agentive -era, the plural typically precedes the derivational suffix, as in

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
360 Number marking

(537a). However, as was discussed in §6.6.2), there is a variation in the ordering

of suffixes, as the reverse order is also possible, such as in (537b).

(537) Variation in suffix order

a) Order PL + AG:
katai kulane ate su-ur-era
hook without fish kill-PL-AG
'those who kill fish without a hook'

b) Order AG + PL:
eseneta-era-uru
believe-AG-PL
'those who believe it'

With verbs that contain the habitual aspect marker -reheto (cf. §12.2.6), an
optional double occurrence of -uru is attested with a number of examples,
which is shown in (538). (Note that the 3pl/E form -kure is actually composed
of -uru and the person suffix -e, but written jointly for convenience.)
Alternatively, plural can be marked only once, in either position (before or after
the habitual suffix): ne-rehete-kure=lu or ne-ure-rehete=lu (cf. §12.2.12)

(538) Double plural marking with habitual suffix

ht akauru ajßa karahai ne-ure-rehete-kure=lu

there 3pl with long.time be-PL-HABl-3pl/E=REM
'They used to stay with them for a long time.'

8.3 Plural objects

Urarina has a suffix that marks plurality for the object of a transitive verb.
However, its use is restricted to certain conditions. A verb is obligatorily
marked with the suffix -naha under the following circumstances:
1. When the object is human and refers to 3pl, and
2. When the object does not surface as a noun or pronoun in the clause.
As shown in (539b), the plural suffixes for subject and object can occur in the
same verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Plural objects 361

(539) Use of plural object marker -naha

a)
hoajtel hanii-ßaha-1 kiuajtel letoa-naha-a
again select-PLO-PRT again send-PLO-3ps/A
'Again he selected them and sent them again,'

b)
kaanii-tpa hau, su-ahe-naha-uru-a hau
sort.of-3ps/D because kilI-CNT-PLO-PL-3ps/D because
'As they did that, because they were killing them, ... [one of the young people said ...].'

If any of the t w o conditions does not apply, (i.e. if an object surfaces, or if it

refers to a non-human object), -mha is not employed, as illustrated in (540).

(540) Absence of plural object marker -naha

a) Non-human referent:
kwara-kur-eni-a=ne tonoana
see-PL-NEG:3ps/E=SUB while
'while they did not see them [the jaguars]' [i.e. the men did not see the two jaguars
mentioned before]

b) Presence of object noun:

ktt akauru rela-e
there 3pl teach-3ps/E
'There she taught them.'

N o t e that the anaphoric presence of an object noun need not be present in the
s a m e clause. When an object surfaces in a participle clause before the main
verb, the plural object suffix is not used. This is illustrated in (541), w h e r e
' p e o p l e ' functions as the object of the dependent clause 'inviting t h e m ' . In this
case, -naha is not marked with the main verb ' m a k e w o r k ' .

(541) No PLO marker after object presence in participle clause

katga-uru kutihja-1, amia-rate

man-PL invite-PRT work-CAU2:3ps/E
'He invited the people and made them work.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
362 Number marking

8.4 Plural "-ana" on stative verbs

A special plural suffix, also discussed in §9.2.2, is the form -ana, which only
occurs on stative intransitive verbs.

(542) Plural suffix -ana with stative verbs

a)
kotaeto-am-atQa-a tg,u=ra
be.tired-STPL-only-3ps/A CRTN=EMF
'They must be (really) tired.' (about the poles of the house)

b)
kareto-ana-a raj n=usitQue-ne hau
be.wide-STPL-3ps/D POSS 3ps=fingemail-SUB because
'because his fingernails are wide ...' [Example by Manus, p.c.]

Alternatively, plural can also be marked with -uru on a stative verb, which
would result in the variant kareto-oru-a 'they are wide' for (542b). However,
double marking is also possible, as the example in (543) illustrates. It is
unknown whether there is any semantic or pragmatic difference between the
three different ways of realisation. While the examples in (542) are from natural
contexts, speakers characterise the form in (543) (from elicitation) as an
additional appropriate variant to express plurality with a stative verb.

(543) Double plural marking with stative verbs

taba-ana-uru-a
be.big-STPL-PL-3ps/A
'They are big.'

8.5 Distributive plural

T h e suffix -afo.ua (or allomorphs -foua and -hah.ua, as discussed in §12.2.7),

refers to plurality in a particular sense, as it stresses the individuality of a
number of subjects or objects, which could be translated as ' o n e by one'. 4 8 It
can be attached to transitive and intransitive verbs, i.e. if the S or Ο argument
has plural reference. The distributive meaning is implied in all the examples.
Similar to the plural object marker -naha, the distributive plural shares the same
conditions for its occurrence: it refers to human participants, and, as a further

48
Note that PSC verbs can take a different distributive plural form (cf. §10.2.1)
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Distributivepiural 363

condition, the object must be absent. However, different from -naha, the suffix
-ahva occurs with transitive and intransitive verbs. The subject plural marker
-uru may optionally occur, but plural reference is understood automatically.
Examples for the occurrence of -ahua with the plural object marker -naha or
with the 2ps plural suffix -t$e are attested in elicitation.

(544) Distributive suffix -ahva with intransitive verbs

a) With subject plural:
ßäe satii ahi-ahva-uru-a=ne hau
already all be.drunk-DSTR-PL=SUB because
'as all of them (each) got drunk'

b) Without subject plural:

titiiritQere-lanaala ni-akiua-e
bend-without be-DSTR-3ps/E
'Without bending [the needles] they were there.' [I.e. each of them were there, trying to
fish]

When the distributive occurs with a transitive verb, the distributive meaning
refers to the object. For instance, example (545a) implies that the speaker
caused different people or groups of people to work, at different times or places.
Thus, the distributive plural form correlates with S/O syntactic pattern, which is
different from the general S/A pivot on the syntactic level (note, however, that
this is unmarked; cf. §22.4).

(545) Distributive suffix -ahoa with transitive verbs

a)
amia-rate-reheto-hua-u=lu
work-C AU2-H AB 1 -DSTR-1 sg/E=REM
Ί used to make each of them work.'

b)
uku ti-ahoa-l akauru letoa-e
needle give-DSTR-PRT 3pl send-3sg/E
'Giving each of them needles, he sent them.'

As example (545b) shows, the use of -ahva is excluded if an object is present

(in this case, the 3pl pronoun akauru 'them').

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
364 Number marking

8.6 Dual and lpl forms

Urarina has three forms for lpl, based on a distinction between inclusive versus
exclusive, and dual, but this three-way distinction is realised in verbal marking
only. There are only two free pronouns for lpl, which are kana (including the
listener) and kanakaanu (excluding the listener), regardless of the number of
speakers involved. The verbal suffix -akatQe for lpl inclusive includes the
speaker and the listener, under the condition that the listener or the speaker
group consists of more than two. The suffix -akaanu excludes the listener and
refers to the group of speakers only. The dual suffix -aka includes the listener,
but the total of persons involved is not more than two. Even though the three
suffixes are grammaticalised, they can be split up into the following
components, morphologically (with some morpho-phonological alternations; cf.
§12.2.16):
--aka lpl/du: Base
- -aka-anu 1 pl/ex = Base + 1 sg/A
- -aka-tQe 1 pl/in = Base + PL

The dual form, which is the base form from a morphological point of view, also
represents the smallest number of participants, but it includes the listener. The
lpl inclusive form can be described as a morphological pluralisation of the dual,
indicating an increased number of listeners. The lpl exclusive form can be
analysed as a combination of the base form and the suffix for lsg/A, thus
emphasising the role of the speaker. In (546), examples for each form are given.

(546) Use of suffixes for lpl

a) lpl/du:
huane=te kaa lejhn kari ama-aka=te hvara-kuru-mu
let's=FOC this one shell take-lpl/du=FOC see-PL-JUSS
'Let's take one of the shells so that they [the others] shall see it.' [One man to another,
as a plan between two]

b) 1 pl/ex:
uhuana ha-naanaha kiuituku-a-akaanu=lu
field make-LPP know-NEG-lpl/ex=REM
'We did not know where to make [our] fields.' [Narrator telling the author about how
they came into the Espejo area]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Number agreement 365

c) lpl/in:
d$aha-u=te lenone-tQaatQe na-1 ku raj beree-kuru
come.on-IMP=FOC eat-lpl/in say-PRT there POSS child-PL

ajjia ku lemne-htre
with there eat-3pl/E
'Saying "come on, let's eat", he ate there with his [three] sons.'

As a development in the contemporary language, the dual form with -aka is

increasingly used to refer to the lpl inclusive form, i.e. there is a trend to
neutralise the distinction between dual and lpl inclusive. Even though the
speakers involved in an event may be more than two, the dual is frequently used
by speakers of the younger generation when talking in a group. The dual form is
also used in instructions as a discourse strategy, as is investigated in §22.6.

8.7 Number agreement

In general, Urarina has number agreement (also referred to as concordance)

between verb and subject (S/A). When the subject is a noun, there can be
concordance with respect to number, i.e. with singular nouns, 3sg will be
marked on the verb. With plural nouns, 3pl is also realised on the verb.

(547) Number agreement between subject and verb

a)
itqa-kuru-a leotQa-uru
do-PL-3ps/A other-PL
'The others did it.'

b)
itulere rukuele ke=te ne-kurete-ure akauru
all.kinds things VLI=FOC ITR-buy-3pl/E 3pl
'They sell all kinds of things.'

However, personal pronouns and other subject referents are optional and
omitted most of the time (unless emphasis is intended or when the context
requires an implicit referent). Thus, in a typical Urarina clause, the question of
number concordance does not apply. However, even when a subject noun is
present, number concordance between subject and verb is not obligatory. One
possible, but not very common type of omission is to leave noun and verb
unmarked although the context is clearly plural. In (548a), for instance, plural is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
366 Number marking

not marked on either verb or noun, even though the context provides this
information. (The sentence is from a story where two men escape from two
jaguars).

(548) Singular marking in plural context

a)
nete akauru ßasi-lanaala amu-e nix hanulari
but 3pl harm-without waIk-3ps/E that jaguar
'But without doing them any harm, those jaguar(s) went (away).' [About two jaguars
mentioned earlier]

b)
akauru ke itulere it^a-rate
3pl VLI all.kinds do-CAU2:3ps/E
'They sent them to do all kinds of things.' [About the patrons, mentioned in the plural
earlier in the text].

The most frequent type of lack of number agreement occurs between a noun in
the plural form and a verb that remains unmarked for plural. However, since the
absence of number agreement occurs with different kinds of nouns, the criteria
for omission of the plural form would be difficult to define, if there are any.

(549) Absence of plural agreement: Plural on noun, but singular on verb

a)
ißoaelu ne-nakauru katQa-uru hetau =te eru-rehete
earlier be-those.who man-PL HRS=FOC find-HABl:3ps/E
'It is said that the ancient people found them.'

b)
eruari-uru ahaiße ne-ι ku-e
mosquito.net-PL paiche.fish be-PRT go-3ps/E
'The mosquito nets turned into paiche fish.'

In contrast, lack of number agreement is not observed in the opposite direction,

i.e. there are no examples of singular subject nouns associated with 3pl verbs. In
addition, there are nouns whose unmarked value is of a collective nature. These
are not strictly uncountable, as they can receive a plural marker, usually meant
to refer to different types. Examples for this are the words for 'fish', 'banana',
and other crops. Such collective nouns typically occur in the singular, even
when plurality in some sense is implied.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Plural agreement in the ΝΡ 367

(550) Nouns with collective reference

a)
ena netohwei=te ate muku-ure katai ke
now until=FOC fish catch-3pl/E hook INST
'Until now they catch fish(es) with a hook.'

b)
βoaelu kahe hvaaune-l fivanara ki-akaanu
earlier from create-PRT banana eat-lpl/ex
'Since ever we have been eating bananas.'

8.8 Plural agreement in the NP

Similar to the absence of plural marking between verb and subject, there are
limitations to plural agreement between the head of an NP and its modifiers.
This section investigates the morphological shape of noun modifiers in plural
context and it will become clear that animacy (more exactly: humanness) plays
a partial role for the presence or absence of number concordance. Further details
on the structure of the N P are discussed in §6.

8.8.1 Numerals and quantifiers

One context which demonstrates that plural marking on nouns is omittable, is

with numerals and quantifiers. While plural can be marked on the noun when it
is preceded by a numeral, this is not obligatory. In fact it is more common not to
mark a noun for plural when it is accompanied by a numeral. However, there is
an important distinction between nouns with regard to animacy: in the majority
of examples of the data corpus and in most examples from the NT, nouns
referring to humans are more likely to be marked for plural in combination with
numerals than not. An example for this is given in (551a). In contrast, if a noun
does not refer to a human, it is unlikely to be marked for plural if a numeral is
present (cf. (551b)).

(551) Plural marking with numerals

a) Nouns with human referent marked for plural:
mtQataha-ϊ raj beree-kuru-atQa ajßa ini-a ku-e
three-PRT POSS child-PL-only with go.up-NTR go-3ps/E
'He climbed up with his three sons.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
368 Number marking

b) Nouns with non-human referent not marked for plural:

heena-he alau bihi+kuturi
four-CNT:3ps/E spider.monkey hand+head
'The spider monkey has four fingers.'

N o t e that other options are possible, too: nouns with human reference can occur
without a plural marker, and, correspondingly, nouns that do not have human
reference, can bear a plural. However, note that (552b) is elicited and
understood as emphasising the fact that nitgatahaj lureri refers to something
like "all three houses". Thus, it appears that plural marking in this context, if
enforced, implies individuation and definiteness.

(552) Other patterns for plural agreement or its absence

a) Nouns with human reference not marked for plural:

raana ari-a ku-e kuriae kuruataha-j katQa ajßa
w.l.peccary seek-NTR go-3ps/E far two-NOM man with
'He went far to look for white-lipped peccaries together with two men'

b) Nouns with non-human reference not marked for plural:

mt$ataha-j lureri-tguru asae ku-akaanu
three-NOM Sb j house-PL under go-lpl/ex
'We went to [all] three houses.'

With quantifiers, the same regularities apply: nouns with human reference are
typically marked with the plural, while other nouns are not. Exceptions apply,
as illustrated in (553c,d).

(553) Plural marking with quantifiers

a) Nouns with human referent marked for plural:

satn enamanaa-kuru=te kana kuruatahane-re=l
all young.man-PL=FOC lpl/in heIp-IRR:3ps/E=ASS
'All the young people will help us.'

b) Nouns with non-human referent not marked for plural:

setu-akzua-a arahn enua
rot-DSTR-3ps/A many tree
'Many trees rotted.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Plural agreement in the NP 369

c) Nouns with human referent not marked for plural:

hitarii nil leotQa rei
all that other king
'all those other kings' [NT: Luke 14:32]

d) Nouns with non-human referent marked for plural:

hitarii kanu temule-kuru
all lsg plant-PL
'all my plants' [NT: Luke 12:18]

In conclusion, one must say that "everything is possible", but not everything is
common. Animacy is at least partially relevant, as nouns with human reference
are more likely to attract plural marking than nouns that are lower on the
hierarchy scale for animacy. As example (553) suggests, the major distinction is
between humans and others, which leaves animals in the group of nouns
preferably unmarked for plural. However, due to lack of examples, it cannot be
stated whether there is a gradual tendency for the presence or absence of plural
marking between animals and inanimate referents.

8.8.2 Other noun modifiers

Modifiers other than numerals and quantifiers do not have an inherent plural
reference, but they can take the plural suffix -uru. These include
demonstratives, adjectives, modifying nouns, and nominalised verbs in
relativising function (cf. §6.5). Different from numerals and quantifiers,
animacy or humanness does not have any impact on the presence or absence of
plural agreement with these modifiers.
Demonstratives do not exhibit any number agreement within the NP. Their
plural form, such as in mi-tQuru 'those' only occurs in contexts where they are
nominal heads themselves. There also are adjectival nouns that can be suffixed
with the plural marker (cf. §5.4.2). Some of these only occur in the plural when
they function as nominal heads themselves. One word, which commonly occurs
in the plural when functioning as a noun modifier, is leotQa 'other'. It is attested
with the plural suffix -uru irrespective of the animacy or humanness of the head
noun.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
370 Number marking

(554) Plural agreement with leotQa

a) With [+human] noun head:

leotQa-uru katQa-uru ajpa
other-PL man-PL with
'with the other people'

b) With [-human] noun head:

kauatga~l=te ku-re=l leotQa-uru amiane-naa-kuru
good-PRT=FOC go-IRR:3ps/E=ASS other-PL work-NOM-PL
'The other works will go well.'

However, there are examples involving leotQauru that do not exhibit plural
agreement with the head. The NP in (554b) could also be expressed in a similar
way without plural marking on the head noun, as leotQa-uru amianenaa, which
clearly shows that number marking is an optional feature. The examples in
(555) demonstrates further variants: in (555a), the head noun is not marked for
plural; in (555b), leotQa lacks a plural marker in combination with a [+human]
head noun. (555c) shows a plural form on the [-human] head noun, but not on
leotQa.

(555) Absence of plural agreement with leotQa

a) Plural marking on leotQa with [+human] head noun:

leotga-uru misjonero kurenia
other-PL missionary for
'for (the sake of) other missionaries [we have accomplished all this].'

b) N o plural on leotQa with [+human] head noun:

nii leotQa katQa-uru
that other man-PL
'those other people'

c) N o plural on leotQa with [-human] head noun:

akauru himitQäi ku-ure leotQa bote-uru
3pl together go-3pl/E other boat-PL
'The other boats went together with them.' [NT: Mark 4:36]

The word leotQa exemplifies what is characteristic for other adjectival nouns as
well. For instance, the words baaso 'bad' and kauatQa 'good' are adjectival
nouns that occasionally occur in the plural form. As nominal heads, they refer to
'bad/good thing', respectively, and their plural form correspondingly describes
'bad/good things'. As noun modifiers, they can occur with human or non-
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Plural agreement in the NP 3 71

human nouns. While it is reasonable to assume that plural marking with baaso
and kauat^a is optional, there are no attested examples, in which they would
occur in the singular when modifying a pluralised head noun.

(556) Number marking of baaso and kauatQa

a) baaso:
satii enua baaso-oru
all tree bad-PL
'all bad trees'

b) kauatQa:
mi katQa kauatga-uru
that man good-PL
'those good men' [NT: Matthew 23:35]

The word enamanaa, which also functions as a noun modifier, only occurs with
[+human] nouns, because of its reference to 'young man'. In all attested
examples with plural context, it exhibits number agreement with the head noun.

(557) Plural agreement with enamanaa

enene ne-nakauru enamanaa-kuru katga-uru kioituku-ur-ene
nowadays be-those.who young.man-PL man-PL know-PL-NEG:3ps/E
'The young people who live now do not know it.'

8.8.3 Plural agreement with relative constructions in modifier function

Further types of modifiers are represented by relative constructions, as

discussed in §6.6. When the inherently pluralised relativising suffix -nakauru is
involved, agreement with the head of the NP always occurs. This results from
the fact that the suffix refers to humans - at least no conflicting examples are
attested, which again might indicate that a distinction with respect to humanness
plays a role for the presence of plural agreement. In the examples in (558),
plural is marked on all words within the respective NPs.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
372 Number marking

(558) Number agreement with suffix -nakauru

a)
enene itaau-nakauru enamanaa-kuru katQa-uru
nowadays live-those.who young.man-PL man-PL
'the young people who live now'

b)
aheri+ko-teru ke teru-kure=lu jioaelu ne-nakauru katQa-uru
stone-ASC-axe VLI axe-VBL:3pl/E=REM earlier be-those.who man-PL
'The people who lived earlier had stone axes'

Correspondingly, there also is plural agreement when other types of

nominalisation are involved, such as the agentive suffix -era and the S/O
nominaliser -i (cf. §6.6.1). The fourth nominalising suffix, -naa, which typically
refers to abstract (or agent) nouns, only occasionally occurs with the plural
form, such as in amiane-naa-kuru ('work'-NOM-PL) for 'works'.

(559) Number agreement with suffixes -era and -i

a) With -era:
katQa rela-ur-era profiueta-kuru rihihei
man teach-PL-AG prophet-PL like
'like the prophets who taught the people' [NT: 2. Peter 2:1]

b) With -i:
kau m-rehete-kur-i katQa-uru
here be-HAB 1 -PL-NOM s man-PL
'the people who used to live here'

The relation between humanness and plural marking is again manifested by the
above examples. The following table gives an overview for the attested
examples of plural agreement with the different modifier types.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Plural marking on other categories 373

(560) Plural agreement within the Ν Ρ

Plural aj;reement
Modifier with |+human] Ν with [-human] Ν
Numerals frequent rare
Quantifiers frequent rare
leotQa 'other' optional optional
baaso 'bad' optional optional
kauatga 'good' optional optional
enamanaa 'young (man)' yes no
REL with -nakauru yes no
REL with -era yes no
REL with -i yes no
REL with -naa optional optional

8.9 Plural marking on other categories

A s m e n t i o n e d in §5.16 on w o r d classes, the plural s u f f i x -uru can also be

attached to postpositions (cf. (535)). In addition, -uru is also observed with
passivised verbs. O n e possible analysis is to a s s u m e that the f o r m that
c o r r e s p o n d s to passive is a special kind of nominalisation, as suggested in
§16.1: a derivational s u f f i x -no, possibly with n o m i n a l i s i n g f u n c t i o n , is attached
to t h e root, f o l l o w e d by t h e n o m i n a l i s e r -i. A s stated in §6.6.1, -i refers to S with
intransitive v e r b s and to Ο o f a transitive clause. A s the passive is analysed as a
t y p e o f n o m i n a l i s a t i o n , the o c c u r r e n c e of the plural m a r k e r is not particularly
surprising.

(561) Plural -uru with participle form

a)
itulere rukuele, hoata kahe itQa-mi-kuru
all.kinds things metal from do-PASS-PL
'all kinds of things, made from iron'

b)
ka=raj ii-tQa te-mhim-uru katQa-uru
lsg=for 2sg-only give-PASS-PL man-PL
'the people who were given to me by you / the people you gave me' [NT: John 17:8]

In s u m m a r y , plural marking, especially with respect to third person, is a highly

optional c a t e g o r y in Urarina. A n i m a c y plays a role in that n o u n s with h u m a n

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
374 Number marking

reference are more likely to be marked for plural than inanimate nouns. This is
particularly true for 3ps subject suffixes -uru and -ana, but does not apply to
the 3ps object suffixes -naha and -ahoa, which are obligatory in the respective
contexts. This corresponds to the occurrence of the 2ps plural marker -toe,
which also is obligatory in plural context.
The fact that number marking is further specified with PSC verbs is
intriguing, as a distinction between singular, dual, paucal, and plural is not
attested for other areas of Urarina grammar. The occurrence of a distributive
plural form also found with PSC verbs is paralleled by the distributive suffix
-ahoa for other verbs (though these are morphologically unrelated with each
other). The dual is distinguished for lpl with all verbs, but not in the set of free
pronouns. In conclusion, it can be stated that the number marking system
exhibits a high degree of asymmetry.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
9 Verb classes

Urarina verbs can be formally divided into several classes with respect to their
transitivity. Principally, they fall into two major classes, transitive and
intransitive verbs, with a number of subclasses. In this regard, stative
intransitive verbs are distinguished from other verbs by a special plural suffix
that is only used with this subclass (cf. 9.2.2). Intransitive verbs that refer to
position, shape, or colour (called "PSC" verbs here) form a further subclass of
intransitives, which is manifested through regularities in the derivational system
(cf. §9.2.3). Reflexivity may be expressed by verbs that are prefixed with the
morpheme Tie-. These verbs might be considered an additional type of
intransitive verb, but it must be noted that the function of the prefix is not
mainly to mark reflexivity. Instead, it has general intransitivising functions that
may derive a reflexive or intransitive verb from most transitive verb roots (cf.
§9.3; §16.2). Within the class of transitive verbs, there is a subgroup of "affect"
verbs, which differ from others with respect to their derivational behaviour. In
particular, they can be intransitivised by some suffixes also occurring with PSC
verbs (cf. §9.2.3). Another interesting feature of the verbal system is that
Urarina has no ditransitive verbs, as is shown in §9.5. Verbs such as 'give',
which would be ditransitive in many other languages, are regular transitive
verbs that cannot be formally distinguished from others. Urarina also has a
copula, which carries out functions that partly apply to transitive and
intransitive verbs (cf. §9.6). Ambitransitivity is not a predominant feature in
Urarina, as is briefly shown in §9.4. Furthermore, there are some verbs that
refer to a question, plus a few more verbs with special properties. These do not
represent a homogeneous class; they are listed in §9.9. Interrogative verbs are
further investigated in §21.6.
The table in (562) gives a summary of distinct morphological features of
each verb class (and subclasses), listing the morphemes that are inherent to each
kind, but are not shared by other verb classes.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
376 Verb classes

(562) Summary of verb class properties

Verb class Special properties

Transitive Agentive -era·, passive suffix -noi (§9.1)
"Affect" transitive Size suffixes (§9.1)
Active intransitive Causative -a; Nominaliser -nakauru (§9.2.1)
Stative intransitive Plural -ana (§9.2.2)
PSC Size suffixes (§9.2.3)
"Reflexive" Intransitiviser ne- (§9.3)

9.1 Transitive verbs

Transitive verbs involve an obligatory object. However, when an object is a 3 rd

person pronoun, it is unmarked in Urarina, which makes it difficult to see
whether a verb is transitive or intransitive when it is not used with a full noun in
object function. For example, Urarina has three verbs for 'eat', the transitive
verbs kia (for solid food, such as meat) and lia (for "soft" food, such as fruit),
plus the intransitive verb lenonia (for any type of food). In context, kia and lia
can be used without an overt nominal object, but the meaning will still be that
the agent 'has eaten it', not that he/she has 'had a meal' (as would be implied by
the use of lenonia). Similarly, in the following example, the object does not
surface as it is obvious that the snake bit 'him'.

(563) Transitive verb without overt object

akano ne~i neruto-ι kala-a

[snake be-PRT be.transformed-PRT] [bite-3ps/A]
[DEP] [V]
'Turning into a snake, she bit [him].' [about a woman]

Apart from semantic factors, differences between transitive and intransitive

verbs mainly occur on the derivational level. The agentive nominaliser -era (cf.
§6.6.2) and the passive suffix -mi (cf. §16.1) can only be attached to transitive
verb stems, but not to intransitive ones. In addition to that, the plural object
suffix -naha is found on transitive verbs only (cf. §12.2.7).
There is a subclass of transitive verbs which is composed of verbs that refer
to cutting, breaking, and other activities that affect the integrity of an object.
These are discussed in §9.2.3, as their properties are closely related to those of
PSC verbs.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Intransitive verbs 311

9.2 Intransitive verbs

Intransitive verbs do not take an argument, apart from the subject. There are a
couple of suffixes that cannot be attached to transitive verb stems, but only to
intransitive verbs. These are the causativising suffix -a (cf. §16.4.2) and the
plural nominaliser -nakauru (cf. §6.6.4).
Urarina has three subclasses of intransitive verbs, which exhibit certain
morphological differences on the derivational level. They can be divided into
active intransitive verbs, stative verbs, and verbs referring to position, shape,
and colour. The latter also includes some verbs describing the "status" of a
subject (e.g. 'be cut'); these are derived from transitive verbs that describe
"affective" action such as cutting, breaking, or splitting (also cf. §9.2.3).

9.2.1 Active intransitive verbs

Typical active intransitive verbs are sinia 'sleep', ahia 'get drunk', and ua 'die'.
The semantic roles of these verbs do not affect their grammatical behaviour in
any way, as the subject can be an agent, patient, or theme. They can increase
their valency by the postposition ke, but note that this postposition is also used
with other verb types. This issue will be further discussed in §16.5. on valency
change, but two examples are given below in order to illustrate the function of
ke (whose use as an instrumental postposition is also discussed in §5.7.4). In
both examples of (564), ke is used because an additional argument [E] is used
with an active intransitive verb. In the first example, the intransitive verb 'live'
is subject to this. In the second example, the verb for 'climb', which is
intransitive can also be translated as 'go up(ward)'.

(564) Increasing the valency of intransitive verbs by ke

a)
nehelau itQau-naa ke itQau-uru-a=na hau
[other.(side) live-NOM VLI] [live-PL-3ps/D=SUB] [because]
[Ε] [V] [CNJ]
'because they live a different life'

b)
enua ari-l ke ini-uru-a=ne hana
[tree seek-PRT] [VLI] [go.up-PL-3ps/D=SUB] [when]
[DEP] [Ε] [V] [CNJ]
'looking for a tree, when they climbed it' (lit. 'when they went up "at" it')

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
378 Verb classes

9.2.2 Stative intransitive verbs

On a semantic basis, it would be difficult to establish a difference between

active vs. stative intransitive verbs. While all stative verbs express physical
states, this may also appear to be the case for some active intransitive verbs. For
instance, ateroa 'be tired' and alalakaraa ' b e dangerous' are active intransitive
verbs. However, there is one morphological feature by which stative verbs
differ from other intransitive verbs: While active intransitive verbs are inflected
for 3pl with the plural suffix -uru, as in atero-oru-a ('be tired'-PL-3ps/A) 'they
are tired', stative verbs are usually pluralised with -ana (or allomorphs). For
instance, 'they are big' is realised as taba-ana-a ('be big'-STPL-3ps).
Alternatively, but less common, the form with -uru as in taba-uru-a is
49
possible, with no apparent difference in meaning. Verb types other than stative
verbs generally form their plural with -uru.

(565) Stative verbs and non-singular marker -ana

Citation/3sg 3pl/A Gloss Semantic type

form
ahaaruto-a ahaaruto-nana-a 'be hot'-(STPL)-3ps Physical
property
anaito-a anaito-ana-a 'be heavy'-(STPL)-3ps Physical
property
memuto-a memuto-ana-a 'be full'-(STPL)-3ps Physical
property
eloato-a eloato-ana-a 'be bitter'-(STPL)-3ps Physical
property
aafihi-a äafihi-ana-a 'be small'-(STPL)-3ps Dimension
taba-a taba-ana-a 'be big'-(STPL)-3ps Dimension
haraauto-a haraauto-ana-a 'be wide'-(STPL)-3ps Dimension
karato-a karato-ana-a 'be long'-(STPL)-3ps Dimension
berauesiohwa-a berauesio-liana-a 'be green'-(STPL)-3ps Colour
(irreg.)
kaitato-a käitato-ana-a 'be difficult'-(STPL)-3ps Difficulty
kauato-a kauato-nana-a 'be nice'-(STPL)-3ps Value
ureeto-a ureelo-ana-a 'be wild/brave'-(STPL)-3ps Human
propensity

49
The difference probably is of a dialectal nature. The Espejo dialect prefers the variant
with -ana. As a result of dialect mixing, the combination of both plural suffixes may
also be heard.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Intransitive verbs 379

The total number of Stative intransitive verbs might be around three dozen.
Thus, the verbs listed in (565) do not represent an exhaustive list, but partly
reflect the semantic distribution of Urarina stative verbs. The majority of verbs
in this subclass refer to physical property or dimension, but they also extend
into other semantic areas such as colour, difficulty, value, or human propensity
as can be seen from the rightmost column of (565).
It also is evident that many of these verbs contain the suffix -toa. The
conclusion, however, is not that -toa functions as a general marker for stative
intransitive verbs. This derivational suffix marks intensity or emphasis in
general and may also occur with transitive verb roots (cf. §10.2.3).

9.2.3 Verbs of posture, shape, and colour

There is a large group of intransitive verbs (possibly about a hundred) that

significantly differ from others with respect to their derivational behaviour:
principally, they are distinguished for size (big vs. small), which is expressed by
the suffix -e for 'small' or -a for 'big' (singular forms). These suffixes are
always followed by -hoa or -ka (3ps/A forms) to express a posture, shape, or
colour. (Note that the details are discussed in §10.2.1). For instance, 'be sitting'
is referred to as lau-e-koa for small subjects (e.g. a child), and lau-a-ka for big
subjects. Altogether, there are 24 derivational forms, i.e. 12 for each size. These
are further divided into different derivational types (each distinguished for size).
Type A derivations result in intransitive stative verbs, such as heku-e-koa 'it is
lying on its side'. Type Β derivations have an inchoative function, referring to
motion, such as heht-lema 'it is lying down on its side' (big item; but
hehi-turaa for small items). Type C has causative function (e.g. heku-taa 'he
laid something small on its side') and Type D refers to a function that may be
called "transitive state". The resulting intransitive verb describes the state
resulting from a caused action and may also be characterised as having a
passivised character. An example for this is heku-haa 'something big is lying on
its side / it has been laid on its side by somebody'. In this case, the process of
lying down is not observed by the speaker.
In addition to this distinction, there are suffixes that express further concepts
such as number (singular, dual, paucal, greater plural, collective plural). Within
the four types of derivation listed under (566a-d), there is a range of
"derivational" number distinctions. While Type A distinguishes five different
types of number, the other derivational types recognise less number distinctions.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
380 Verb classes

(566) Productive derivational suffixes for position verbs

a) Derivation A (Resulting in intransitive state; e.g. 'be standing'): singular,

dual, paucal, greater plural, distributive plural.
b) Derivation Β (Inchoative; resulting in intransitive action; e.g. 'stand up'):
singular, paucal, greater plural.
c) Derivation C (Resulting in transitive action; e.g. 'make stand'): singular,
plural.
d) Derivation D (Resulting in transitive state, e.g. 'stand as a result of a cause'):
singular, plural.

An important distinction to stative intransitive verbs is that position verbs do

not generally take the stative plural marker -ana.50 Conversely, stative verbs are
not compatible with any of the derivational suffixes found with position verbs.
The verbs described up to this point strictly involve meanings that refer to
posture, such as be lying, standing, sitting, or hanging. Each of these postures
can be further divided into different types of sitting, standing etc. For 'be lying',
for instance, there are lexical distinctions for the kind of lying. Depending on
whether the item is lying with its face up or down, whether it is lying on its side,
or whether it is lying straight or in a bent position, etc., there are different
lexemes. In addition, the shape of the respective item is also relevant: long
items or round items imply a different way of lying than flat items. Thus, there
are at least eight different verbs for 'be lying' and similar numbers for other
postures. Those also include verbs for 'be heaped up', 'be tied together', or 'be
sticking out', just to mention a few. The list in (567) gives an overview of some
verbs describing different types of posture, and referring to different kinds of
objects.

50
However, it should be mentioned that the stative plural suffix -ana may optionally
occur with the greater plural form of Type Β derivations (with no apparent difference
in meaning), e.g. as in lau-tura-ana-a '(they) are sitting down'. Also note that the
plural for small items of Type D derivations contains the form -ana, as in
lau-eka-ana-a 'they have been seated'.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Intransitive verbs 3 81

(567) Posture verbs

a) 'Be lying'
Small items Big items Position Shape
tuku-ee-koa tuku-a-ka be lying round and longish items
[tukweekoa] without a "mouth'Vopening;
e.g. pen, bin with lid on,
banana; soccer ball, egg
laa-e-koa laa-ka be lying on its flat and long objects that
back (face up) have a "mouth'Vopening;
e.g. ship, person (lying on
back); banana leaf
me-e-koa mi-a-koa be lying on its flat and long objects that
face (face down) have a "mouth'Vopening;
e.g. ship, person (lying on
back); banana leaf
heku-e-koa heku-a-ka be lying on its narrow and flat (longish)
side things; e.g. canoe with a
load on one side only; also:
dog; person; gumboot
tutoa-e-koa tutga-a-ka be lying straight long items; e.g. snake
se-e-koa si-a-koa be lying long items; e.g. rope, snake,
tortuously, piece of cloth
slightly bent
bau-e-koa bau-a-koa be lying partly on long or flat items; e.g. pole,
top of sth., stand pen, notebook
out, also leaned
on top of sth.
daraa-e-koa dara-a-koa be lying together, round items without a
next to each other mouth; e.g. rifle shell; cf.
tukueekoa

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
382 Verb classes

b) 'Be standing'
Small items Big items Position Shape
dzuhu-e-koa d$uhu-a-ka be standing high/tall items with one or
two long legs; e.g. pole,
slipper standing; chicken,
soccer goal, pifayo tree
te-e-koa tii-a-ka be standing items with four legs; e.g.
dog, house, fish in water if it
looks as if having feet,
crocodile with raised body,
millipede, monkey
teru-e-koa teru-a-koa be standing on its wide and thin items; e.g.
edge slipper; notebook on its
edge

c) 'Be sitting'
Small items Big items Position Shape
lau-e-koa lau-a-ka be sitting short items (not tall); e.g.
bird, monkey, dog,
gumboot, renaco tree (with
many roots)
baba-e-k)a baba-a-koa be sitting together, e.g. people
next to each other

d) 'Be hanging'
Small items Big items Position Shape
hjunu-e-koa hjunu-a-roa be hanging, fixed long and tight, e.g. cable,
(small but at two points rope for drying clothes
long)
lefoa-e-koa lekoa-roa be hanging (not long items; e.g. rope
tightly), fixed at (banana-shaped)
two points
d^ato-e-koa d^ato-a-roa be hanging down soft items that bend/fold
easily; e.g. rope over a
branch, towel
fioii-tQoa fiuii-a-ka be hanging and wide items; e.g. hammock,
spread out mosquito net
tihii-tgoa tihii-a-ka be hanging at one unspecified shape, e.g. bag,
end of a rope, be bottle, pot
hanging down

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Intransitive verbs 383

Small items Big items Position Shape

kodo-e-koa kodo-a-koa be heaped up e.g. grass
lunu-e-koa lunu-a-koa stick out, be a little e.g. crocodile with head out
outside of the water
risii-tQoa risii-a-ka be squatted persons only; not monkeys
tutu-e-koa tutu-a-roa be stuck in sth. e.g. pole; also about
mosquito sitting
tihja-e-koa tihja-a-ka walk slowly, go e.g. ship, animal
(motion verb)

There are further verbs that can take suffixes according to the size of the
subject, which do not exactly refer to posture. In particular, these are verbs that
describe shapes and colours. The productivity of the derivational suffixes with
these verbs is slightly lesser than with position verbs, which are almost 100%
compatible with the suffixes. In combination with shape verbs, some gaps occur
in the paradigm (cf. §10.2.1); the three verbs describing the basic colours black,
white, and red are fully compatible with the derivational suffixes.

(568) Some shape and colour verbs

Small items Big items Gloss

habe-e-koa habi-a-ka 'be round or roundish'; e.g. stingray
hala-e-koa hala-a-ka 'be open', or 'have a hole'
hjara-e-koa hjara-a-koa 'be smaller than the other'; "Big": 'be a little
smaller than the biggest'
maka-e-koa maka-a-koa 'be flat or thick, table-like'; about items that
can be piled up; e.g. book
turu-e-koa turu-a-roa 'be narrow'; e.g. path; cloth
soma-e-koa soma-a-koa 'be white'
hitQu-e-koa hitQu-a-koa 'be black'
lana-e-koa lana-a-ka 'be red'

There also are some verbs referring to shape that are not compatible with size
suffixes at all, as they are members of the stative or active intransitive verb
classes. These include examples such as muluutoa 'be thick', karatoa 'be
long', or haräatoa 'be wide', which are stative verbs that take the plural suffix
-ana. The verb mahiitga 'be pointed' is an active intransitive verb, regarding its
morphological properties. The same applies to the colour verb berauesiohwaa
'be green', though it is unclear from what root it may be derived. The verb
seems to be of newer origin and can also refer to 'blue'; traditionally, the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
384 Verb classes

Urarina distinguish only black, white, and red, which is in correspondence with
cross-linguistic tendencies.
A further verb type that can be combined with the size suffixes discussed
above are what could be called "affect" verbs. These include verbs of cutting,
tearing, splitting, smashing, stabbing, or other processes that reduce or mutate
the shape, integrity, or size of an object - a group of about 30 verbs. However,
there are certain limitations to the productivity of size suffixes within this
group:
1. The definition of "affect" verbs is restricted in that it does not include verbs
such as 'hit', 'kill', 'push', etc., but only those referring to one of the
activities listed above.
2. Out of the verbs that match this definition, most, but not all are compatible
with size suffixes.
3. Those that are compatible with these suffixes do not take the entire range of
suffixes from the paradigm, with differing degrees of compatibility. This is
exemplified with the verb saua 'cut' in ex. (607) (§10.2.1).

The suffixes of derivational Type A transform "affect" verbs into intransitive

verbs and assign a stative or resultative function to these. However, the
resulting verbs are only distinguished for size in a few cases; otherwise, size
distinction does not apply. Instead, the use of the derivational suffix for "small
+ posture" (-e-koa) typically implies the degree of being cut, etc., as is further
investigated in §10.2.1. The interpretation seems to be determined lexically and
may differ from verb to verb. For instance, the transitive verb moloa 'cut (off)'
has the stative form molo-e-koa 'be cut', but at the same time it implies that
something is 'cut a little so that it is almost cut o f f . While this form does not
automatically imply that the respective subject is small, the form derived with
-a-koa (for big items with position verbs) explicitly refers to a big item that is
cut. For most other verbs of this type, there is no derivation with -a-koa. The
verb sau-e-hoa (from sau-a 'cut with one cut') means 'be cut a little with one
cut, almost in h a l f .
It has been shown that verbs of position, shape, and colour form a natural
subclass of intransitive verbs that is distinct from other subclasses in terms of
derivational behaviour. Affect verbs can be classified as a subgroup of transitive
verbs on similar grounds.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Reflex ivity 385

9.3 Reflexivity

Urarina has no intrinsically reflexive verbs, but uses alternative strategies to

express reflexivity. Most importantly, there is a form that functions as a general
intransitiviser. The prefix ne- is productively attached to transitive verbs in
order to create an intransitive verb. For instance, the transitive verb for 'lose' is
transformed into an intransitive verb 'get lost' by prefixation of ne-. The same
prefix is also used to mark reflexivity. A detailed discussion follows in §16.2.1
on valency changing mechanisms.

9.4 Ambitransitivity

It is difficult to detect the existence of ambitransitive verbs in Urarina, as 3 rd

person objects are unmarked, i.e. if a transitive verb occurs without an overt
object, it automatically implies an (unmarked) 3 rd person object like 'it', 'him',
or 'them', which is only understandable through the context. One clue in order
to detect whether a verb is transitive or intransitive is the constituent order when
overt arguments such as subject or object are present. For instance, the verb
muhia 'burn' may occur with one or two overt arguments, as depicted in (569).
For both examples, the translation is unambiguous, but this is partly due to
pragmatic/semantic factors: it is obvious that the tree in (569b) does not play an
active role in the burning process. Therefore, with this ambitransitive verb,
there is an S/O relation in terms of argument structure.

(569) Verb tnukua 'burn'

a) Transitive use:
enua muku-anu
tree burn-lsg/A
Ί burned the tree.'

b) Intransitive use:
muku-a enua
burn-3ps/A tree
'The tree burned.'

Apart from the verb mukiia, no other ambitransitive verbs are known.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
386 Verb classes

9.5 Ditransitivity

Urarina has no ditransitive verbs. When a verb requires more than the t w o
thematic roles of agent and undergoer or theme, its valency is increased by ke.
However, there are no verbs that would require a participant marked by ke as its
basic argument structure. Verbs that typically take a goal or recipient in other
languages are transitive in Urarina. A recipient phrase can occur as a
beneficiary postpositional phrase (PP) with almost any verb. There are some
verbs that frequently take a PP with the postposition raj 'to/for', which also
functions as a possessive marker (cf. §7.1.2 on possession; also see §5.7.4). T h e
verbs that most commonly occur with this postposition include verbs of saying
and verbs of giving.

(570) Verbs most commonly occurring with raj

naa 'say' (intransitive; quotative verb)

ajtoa 'say' (intransitive)
mritoa 'say so' (intransitive)
eria 'speak' (intransitive)
bia 'tell' (transitive)
tia 'give' (transitive)
d^alua 'distribute' (transitive)

T h e following examples illustrate in what way raj is used as a beneficiary with

the verbs in question. Note that the indirect object may not be overt if it is 3sg.
In this case, only raj will surface to mean 'for/to him/her' (e.g. (572a)). The
additional participant may occur in preverbal or postverbal position.
T h e verb naa can be characterised as a quotational verb. It always follows a
quotation. Thus, it may not be strictly intransitive as one could argue that it
takes the quotation as a complement. However, the quotation is not marked with
a complementiser or any other marker (whereas complement verbs occur with
the subordinate markers =na or =ne\ cf. §20.2).

(571) Quotative verb naa used with raj

a)
nihjauria kati saate-nana-a=ne na-e hetau
[don't black.monkey kill-ILT-3ps/A=NEGF] [say-3ps/E HRS]
[QUOT] [V]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Ditransitivity 387

mi eene raj
[that woman for]
[PP]

'"Don't kill black monkeys!" he said to that woman.'

b)
na-e hetau noe raj kana+kivaaun-era
[say-3ps/E HRS] [Noah for] [lpl/in+create-AG]
[V] [PP] [S]
'... said God to Noah'
There are three further verbs of speaking, all of which are intransitive. The
following examples illustrate each of these in combination with a raj-phrase,
which refers to the listener o f an utterance.

(572) Intransitive verbs of speaking used with raj

a) niritoa 'say so'

mrito-a raj kana+kvaaun-era=ne hau hetau =te
[say.so-3ps/D] [for] [lpl/in+create-AG=SUB] [because HRS=FOC]
[VI] [PP1] [SI] [CNJ]

itg,a-e=lu
[do-3ps/E=REM]
[V2]
'As God said so to him, he did it.'

b) ajtoa 'tell, say, speak'

kati raj=pe ajto-ö=ni
[black.monkey for=FOC] [tell-lsg/E=WIT]
[PP] [V]
Ί just told [it to] the black monkey.'

c) eria 'speak, talk'

kaum raj ere-ni-u=ra
[3pl for] [speak-DSTL-IMP=EMF]
[PP] [V]
'Go and talk to them!'

Another verb of speaking is bia, which can be translated as 'tell' on the one
hand, but also as 'advice' or 'announce' on the other. In contrast to the verbs
illustrated in (572), bia is transitive.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
388 Verb classes

(573) bia 'tell/announce' used with raj

a)
te-ι turu-i hetau raj be-ure
[give-PRT] [arrive-PRT HRS] [for] [tell-3pl/E]
[DEP1] [DEP2] [PP] [V]
'Handing over [Lomai's son], arriving, they told it to [the Creator of Ways].'

b)
nii raj katai be-ure
[that for] [hook] [tell-3pl/E]
[PP] [Ο] [V]
'They showed [lit. "announced"] the hooks to that one.'

The fact that the ra/-phrases shown with the above verbs function as an adjunct
is easily shown through the examples in (574), where the verbs bia and ajtoa
occur without any mentioning of the goal or undergoer, i.e. without raj. The
same structure also applies to the other verbs of saying.

(574) Verbs of speaking used without raj-phrase

a) bia:
dede kahe ni-a komasaj bi-a=na hau
[sky from be-3ps/A wife] [teIl-3ps/D=SUB] [because]
[QUOT] [V] [CNJ]
'because he told (them) that he had a woman from heaven'

b) ajtoa:
enua ra-a ht-i=ße kujßa=na ajto-ö=ni
[wood take-NTR go-2ps=SUB so.that=FOC] [say-lsg/E=WIT]
[DEP] [V]
Ί told (you) to go and fetch wood.'

From a universal perspective, the most typical concept of a ditransitive verb is

related to the verb for 'give', as its semantic structure intuitively implies the
presence of a recipient. The following examples show that, similar to the verbs
of speaking, Urarina verbs of giving can be accompanied by a PP referring to a
recipient, which is usually expressed through raj. However, it represents an
optional phrase, as will become clear below.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Ditransitivity 389

(575) Verbs of giving used with raj

a) d$alua 'distribute':
akauru raj lenone d^alu-e
[3pl for] [food] [distribute-3ps/E]
[PP] [O] [V]
'He distributed food to them.'

b) tia 'give':
akauru raj eruari te loinaj
[3pl for] [mosquito.net] [give:3ps/E] [PSN]
[PP] [Ο] [V] [A]
'Lomai gave them a mosquito net.'

In the same way as verbs of speaking, 'give' and 'distribute' may also occur
without explicit mention of a recipient/indirect object, which clearly shows that
the involvement of the raj-phrase is not part of their argument structure, as is
shown with tia in (576).

(576) tia 'give' used without raj-phrase

a)
h-i -jie kujßa=ne te-ü
[eat-2sg=SUB so.that=FOC:lsg] [give-lsg/E]
[DEP] [V]
Ί gave [it] [to you] so that you would eat it.'

b)
hjauijie te-i-toe=ne
[don't] [give-2ps-PL=NEGF]
[INTR] [V]
'Don't give [it] [to him]!'

An alternative translation for tia could also be 'deliver' or 'provide', as these

verbs represent its argument structure as a transitive verb more adequately in
English. The verb tia also has an additional meaning 'make', where it refers to
constructing, producing, or setting up.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
390 Verb classes

(577) tia with the meaning o f 'make'

a)
m=hitQana ti-a nerajße
[3ps=blowgun] [make-3ps/D must]
[O] [V]
'He had to make his blowgun'

b)
aheri+kuteru ke n=ertanihja te-1
[stone+axe INST] [3ps=canoe] [make-PRT]
[PP] [O] [V]
'making their canoes with stone axes ...'

There also is another verb for 'give'; which is clearly transitive. The verb sia
could substitute tia in most examples (except for 'make'), but in practice is
mainly used in the imperative form when asking for something. Different from
tia, the recipient functions as the Ο argument, whereas the item to be given
must be marked with the valency increase marker ke, such as illustrated below:

(578) Verb sia 'give'

a)
atii katai ke kam Sl-u

[please] [hook VLI] [lsg] [give-IMP]

[EXPR] [PP] [O] [V]
'Please give me fishing hooks!'

b)
ßäe ii si-ri-tQäu-ni
[already] [2sg] [give-IRR-lsg/A=ASS]
[ADV] [Ο] [V]
Ί will already give you.'

The fact that the verbs o f saying and giving are either transitive or intransitive
verbs, eventually leads to the conclusion that Urarina has no ditransitive verbs.
The verbs under investigation are transitive verbs that can take a beneficiary as
an adjunct. Thus, the insertion o f a phrase with raj can be characterised as an
adjunct. Note that there are no formal criteria for the distinction between
arguments and adjuncts. This becomes even more evident through examples
with verbs whose argument structure does not imply an indirect object, such as
'work', 'be difficult', or the copula 'be' (cf. (579)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Copula 391

(579) Intransitive verbs with adjunct raj

a)
amiane-reheto-a=lu fioatoro-kuru raj
[work-HAB 1 -3ps/A=REM] [patron-PL for]
[V] [PP]
'He used to work for the patrons'

b)
tabauru raj kalta-e
[some for] [be.difficult-3ps/E]
[PP] [V]
'For some it is difficult'

c)
akauru raj aaune+baka, erene+kutuuhue, nii-tQuru ni-a
[3pl for] [plantain+soup fat+fried.food that-PL] [be-3ps/A]
[PP] [O] [V]
'There were for them plantain drink, fried food, all that.'

A l m o s t all verbs can t a k e an a d j u n c t with raj in order to express a b e n e f i c i a r y

f u n c t i o n . A p a r t f r o m the intransitive e x a m p l e s above, these also include Stative
intransitive verbs (e.g. ' b e g o o d ' ) , transitive v e r b s (e.g. ' k i l l ' ) , and
intransitivised verbs (e.g. ' s e l l ' ) . Restrictions only o c c u r in terms o f s e m a n t i c
incompatibility. E x a m p l e s f o r this are the intransitive verb d^anaa ' c i y ' , ua
' d i e ' , and sinia ' s l e e p ' . With these verbs, the postposition hirema ' f o r the
b e n e f i t o f can be used instead (cf. §5.7.4).

9.6 Copula

E v e n t h o u g h the c o p u l a d o e s not represent a separate " c l a s s " , it exhibits s o m e

f e a t u r e s that d i f f e r f r o m other verbs. Urarina has one copula, nia, w h i c h is a
m o r p h o l o g i c a l l y regular verb and is subject to inflection for all verbal
c a t e g o r i e s except the causative and c o m p l e t i v e aspect, w h i c h are not attested. 5 1
In a w a y , the copula implies a certain sense of ambitransitivity: it is similar to

51
It is impossible to determine whether the copula is "transitive" or "intransitive".
While morphologically, it lacks the distinctive properties of a transitive verb (such as
the compatibility with passive and agentive morphemes), on the syntactic level, the
copula resembles a transitive verb (cf. §18.1), as the copula complement occurs in
object, i.e. clause initial position and the copula subject follows the verb, in the same
way as the subject of a transitive verb.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
392 Verb classes

intransitive verbs when it occurs with a copula subject (CS) only, but resembles
a transitive verb when a copula complement (CC) is present (cf. §9.6.1). In
relation to existence, the copula takes only a subject (cf. §9.6.4), while with
other functions, a copula complement is involved. The copula complement
relates to the following functions:
- Identity or equation (cf. (§9.6.1)
- Attribution (cf. §9.6.2)
- Locative or NP followed by a non-locational postposition (cf. §9.6.3)
- Existence or temporary presence (cf. §9.6.4)
Additional features of the copula include its functions in possessive
constructions (cf. §9.6.5), and as an auxiliary (cf. §9.6.6).

9.6.1 Identity or equation

One function of the copula complement relates the subject of the predicate to a
specific identity or states an equation between the two (but the difference is not
formally marked). As with other verbs, the subject does not obligatorily surface
as an NP, but can be marked by a bound pronoun (i.e. a person suffix) on the
copula. The constituent order corresponds to that of a transitive clause, with the
copula subject in postverbal position and the copula complement in preverbal (=
object) position. The subject can be shifted to the front when a focus marker is
used.

(580) Identity/equation with copula complement

a) Without overt subject NP:

kauatQa-i kuraanaa ni-a=ne=ta
[good-NEG:3ps/A] [chief] [be-3ps/D=CND=FRS]
[Main] [CC] [V]
'It is not good that he is chief.'

b) With overt subject NP:

enamanaa ni-a ka=kat$a
[young.man] [be-3ps/A] [this=man]
[CC] [V] [CS]
'This man is young.' (lit. 'a young man')

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Copula 393

c) With subject NP shifted to the front:

aka=te raana ne
[3ps=FOC] [w.l.peccary] [be:3ps/E]
[CS] [CC] [V]
'It is a white-lipped peccary.'

T h e copula complement can also be a N P involving a measuring unit, such as

'meter'.

(581) Measuring unit as copula complement

letQurjka-j metoro ni-a raj kaare

[ten-NOMsbj meter] [be-3ps/A] [POSS length]
[CC] [V] [CS]
'Its length is 10m.'

N a m e s only occur as copula complements if the copula subject is the noun for
' n a m e ' . Otherwise, there are alternative constructions to express naming. O n e is
with the quotative verb naa for ' s a y ' , as illustrated in (582b). The e x a m p l e in
(582c) shows the use of the c o m p o u n d verb kuraa+tia ( ' n a m e ' + ' g i v e ' ) for
naming.

(582) Naming

a) With copula:
kanu kuraa=te hiväa
[lsg name=FOC] [PSN] [be:3ps/E]
[CS] [CC] [V]
'My name is Juan.'

b) With verb naa\

hwäa naa kanu kuraa
[PSN] [say-3ps/A] [lsg name]
[QUOT] [V] [S]
'My name is ("said") Juan.'

c) With compound verb kura+tia:

arauata ne-l=te kanu kura+te-uru-a katga-uru
[PSN be-PRT=FOC] [lsg] [name-give-PL-3ps/A] [man-PL]
[DEP] [O] [V] [A]
'The people call me Arahuata.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
394 Verb classes

9.6.2 Attribution

In §5.4 it was stated that Urarina has no homogeneous class of adjectives, but a
number of adjective-like words, with a range of properties that may differ from
word to word. One of the features mentioned was the possibility for a word to
occur as a copula complement (which does not apply to all adjective-like
words). However, some are indeed found as copula complements and again,
their position can be clause-initial (object position) or following a fronted
copula subject (cf. (583b)). Note that examples with the adjectival noun plus
focus marker are not attested, which could be due to the fact these mainly
function as modifiers.

(583) Adjectival noun as copula complement

a) Regular constituent order:

laauhwiri ni-a kaa ate
[small] [be-3ps/A] [this fish]
[CC] [V] [CS]
'This fish is s m a l l . '

b) With focus:
itQasu itea-naa=te baaso
[brother do-NOM=FOC] [bad] [be:3ps/E]
[CS] [CC] [V]
'What my brother does is bad.'

9.6.3 Locative

The copula complement can also function as a locative, with or without

postposition. It is realised as a noun that refers to a location or as a locational
demonstrative (cf. §5.6.2). Since there is no specific marking for this function,
the difference between identity/equation and locative function only becomes
clear through context. For instance, the reading of (584a) as 'The snail is an
entrance' is ruled out by the context, or by general knowledge of logical
relations.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Copula 395

(584) Copula complement as locative

a) With noun:
akuuhua ni-a ka=eruri
[entrance] [be-3ps/A] [this=snail]
[CC] [V] [CS]
'This snail was in the entrance.'

b) With locational demonstrative:

tahia=te ne
[over.there=FOC] [be:3ps/E]
[CC] [V]
'He is over there.'

c) With noun and postposition:

rf3«nifaja=ie lureri asae ne
[somebody=FOC] [house under] [be:3ps/E]
[CS] [CC] [V]
'Someone is in the house.' [i.e. under the roof]

In other cases, the copula complement is a postpositional phrase with other than
locational reference. For instance, it can also involve a non-locational
postposition such as ajjia ' w i t h ' or raj ' f o r ' .

(585) Copula complement with non-locational postposition

a) With ajjia:
kalaui ajßa ni-a nil kati hau
[son with] [be-3ps/D] [that black.monkey] [because]
[CC] [V] [CS] [CNJ]
'as that black monkey was with child'

b) With raj:
mi-tQuru=te eshoela raj jie
[that-PL=FOC] [school for] [be:3ps/E]
[CS] [CC] [V]
'Those things are for the school.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
396 Verb classes

9.6.4 Existence or temporary presence

W h e n the copula is used with the copula subject only, its function m a y relate to
existence (being permanent or temporary). T h e copula subject occurs in
postverbal position as this is the regular position for a subject (cf. §18). In
e x a m p l e (586b) the copula subject is syntactically complex ( ' s o m e t h i n g f o r
hunting without s u f f e r i n g ' ) .

(586) Copula subject referring to existence

a) Postverbal copula subject:

ne-ri-tga=l atane katg,a mamale-ra
[be-IRR-3ps/A=ASS] [ground man make.stick-AG]
[V] [CS]
'There will be ground that makes people stick.'

b) Complex postverbal copula subject:

ßäe ni-a kajtehe-lanaala nesari-tQuru-a=ne haanu=na hau
[already] [be-3ps/D] [suffer-PRV hunt-PL-3ps/A=SUB PURP=SUB][because]
[ADV] [V] [CS] [CNJ]
'Because there already is [something] in order to hunt without suffering.'

Similar to the existential function, the copula in combination with a copula

subject can refer to temporary presence in a place and is translated as 'live' or
' s t a y ' . Structurally, there is no difference to existence involved, as additional
constituents will occur in the same way as described in the respective sections
above (i.e. copula subject only or locational copula complement). Thus, the
distinction again depends on the context. In (587a), the copula c o m p l e m e n t is a
place n a m e ; in (587b), the location is expressed by a postposition.

(587) Copula referring to temporary presence

a)
pukalpa ni-anu-ne bana
[Pucallpa] [be-lsg/D=SUB] [when]
[CC] [V] [CNJ]
'At the time when I lived in Pucallpa'

b)
raj lureri asae ni-a=ne here nii raj kakunu
[POSS house under] [be-3ps/D=SUB] [want:3ps/E that POSS daughter]
[CC] [V] [Main]
'Her daughter wanted her to stay in her house.' (= Locative copula complement)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Copula 397

9.6.5 Possession

The copula subject in combination with the possessive marker raj (lit. ' f o r ' / ' t o ' )
marks possession (cf. §7.2.1). Typically, this involves a proclitic attached to raj,
such as in fc?=ra; (lit. 'for me'), which follows the copula in existential function.
Examples such as in (588), literally mean '(it) is there for me', which makes it
plausible to compare this construction to the one mentioned in (585b).
However, the constituent order is different, as the copula must precede the
possessive. As shown in (588b), the corresponding negative form implies non-
existence or non-presence (literally 'my axe is not there'). 52

(588) Possession with copula

a) Affirmative:
ni-a ka=raj pekepeke
[be-3ps/A] [lsg=POSS] [boat.engine]
[V] [CC] [CS]
Ί have a boat engine.'

b) Negative:
m-ji kanu teru
[be-NEG:3ps] [lsg axe]
[V] [CS]
Ί do not have an axe.'

9.6.6 Auxiliary

Urarina has a construction in which the main verb of a clause is suffixed with
the participle form and followed by a copula. The copula in this construction
hosts all inflectional suffixes normally found on the main verb. For instance,
instead of marking plural, negation, and person on the verb root /k«ra/ 'name'
in (589a), this verb receives a participle suffix and the inflection (except for the
distributive suffix) is marked on the copula. Literally, the translation would
result in 'they were not being with names'. Alternatively, the copula could be
omitted and the verb would surface as kura-hahua-ur-em ('name'-DSTR-PL-
NEG:3ps/E). The difference between the two variants is unclear, as speakers
describe meaning of both constructions as "exactly the same". The variation
occurs in the traditional as well as in the innovative language and is observed in
different dialects. However, there is variation between speakers: some regard

52
The presence or absence of raj depends on factors discussed in §7 on possession.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
398 Verb classes

the combination of participle and copula as "colloquial". At this stage, the exact
difference between the two alternant forms cannot be confirmed. One might
preliminarily characterise the function of this construction as a matter of style.

(589) Participle form followed by inflected copula

a)
ipoaelu aj-a=ne rihihel kuraa ke kura-hakioa-l
[earlier AUX-3ps/D=SUB like] [name VLI] [name-DSTR-PRT
[DEP] [PP] [V]

ne-ur-ene
be-PL-NEG: 3 ps/E]
'They were not with names like in ancient times.'

b)
lesahei, ßäe kanaanaj siiri-ϊ ni-a =ne
[once] [already] [child] [have-PRT be-3ps/D=CND]
[ADV] [ADV] [Ο] [V]
'once, when she was already having a baby' (= when she was pregnant)

The construction occurs with transitive as well as with active intransitive verbs,
but there are no attested examples for stative or PSC verbs being followed by
the copula. However, the copula itself can be the main verb of the clause, as
shown in (590). This results in two copulas occurring in a sequence, the first
being suffixed with the participle form and the second carrying the inflectional
suffixes.

(590) Two copulas in a sequence

a) Affirmative:
nitoanei ni-a=ne
like.that be-3ps/D=CND
'when it came to that / when it was like that' [i.e. that the woman was about to give
birth]

b) Negative:
nijei kana+huaaunera beree ne-i ni-ji
[not.at.all] [our.creator child] [be-PRT be-NEG:3ps/A]
[ADV] [CC] [V]
'He is absolutely not a child of God.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Auxiliary "aja" 399

Note that a similar construction is translated as a relative clause, with rienaa,

which is a nominalisation of the copula. The following example (591) contains
a participle form, followed by nenaa (lit. 'so that we could be those who are
obeying').

(591) Participle + nominalised copula

d$=ere tonora-ι ne-naa najm-tg,aanu=ne kujßa

[2sg=word] [obey-PRT be-NOM] [be.able-lpl/ex=SUB] [so.that]
[O] [CS1] [V2] [CNJ]
'so that we could be [people] who obey your word'

In a few examples, the participle form of the copula before another verb has the
meaning of 'become' or 'turn into'. This is the case in certain examples with the
motion verbs kua 'go' and amua 'walk' (also cf. §18.3). However, this function
can be characterised as an idiomatic complex predicate since it is only attested
in this specific combination. In general, there is no formal distinction between
'be' and 'become'.

(592) Idiomatic meaning of copula as 'become'

a) With amua:
tariatga rn-i amu-e
[turtle] [be-PRT walk-3ps/E]
[CC] [V]
'They turned into turtles.'

b) With kua:
kaum raj eruari ahaijie ne-1 ku-e
[3pl POSS mosquito.net] [paiche.fish] [be-PRT go-3ps/E]
[CS] [CC] [V]
'Their mosquito nets turned into paiche fish.'

9.7 Auxiliary "aja"

Apart from the copula, Urarina has an auxiliary aja that functions as an
intransitive "dummy" verb and refers to any action previously mentioned or
known through the context. There may be an explicit mentioning of the verb it
is referring to, as in (593), where aja refers to the transitive verb 'destroy'. This
example also shows that the verb aja is referring to may have any transitivity
value, while aja itself is intransitive.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
400 Verb classes

(593) Auxiliary referring to a previously mentioned verb

ka=d,z)aura+atane itahe-ri-t$au=ni=tQa
[this=flesh+world] [destroy-IRR-1 sg/A=ASS=EMF]
[01] [VI]

aj-a=ne hau bute ha-u

[AUX-3ps/D=SUB because] [boat] [make-IMP]
[DEP2] [Ol] [VI]
Ί will destroy this world. Because it is so, build a boat!'

Alternatively, the auxiliary can refer to an action that is obvious from the
context but not explicitly mentioned by a verb in the clause preceding it. In
(594), the semantic content o f 'doing so' refers to 'urinating', which had been
mentioned several sentences before in the course o f the story (as indicated by
the free translation preceding the sentence containing aja).

(594) Auxiliary referring to action known through the context

[Preceding: 'So she squatted and urinated, and she urinated very much. Therefore she
kept on flooding the earth. The water already came flooding the hall; there they were
playing the flute. The water kept rising, and even though they were dancing on and on,
she was doing it (aja). When the water was rising up to their ankles, they kept on
dancing and playing the flute and drum, and when the water rose up to their calves, they
danced and still did not want to release h e r . . . ' ]
ku hajti aj-ahe-i saku-e
there still AUX-CNT-PRT follow-3ps/E
'She still kept on doing so there.'

N o t e that, in the same way as other intransitive verbs, aja can occur with any
inflectional suffixes (cf. §12). In addition, the auxiliary is also used as a short
form for affirmation or negation, ajanu ( A U X - l s g / A ) and aj-i ( A U X -
N E G : 3 p s / A ) are common forms to express 'yes' and 'no', respectively (also cf.
§14.4.4). A n example for its use as a negator is given below. In this case, aji is
understood as referring to the situation as a whole: 'it is not so'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Verb I ess clauses 401

(595) Auxiliary used to mean 'no'

[Preceding: "'Why are they doing this to me [watching me while I sleep]?" As the priest
said so, his assistant said [to them], "Maybe you sometimes want to kill a priest." As he
said so, they [replied] ...']
aj-i=ta kwatia su-akaanu=ne kujßa aj-a-akaanu=ta
AUX-NEG:3ps/A=FRS not kill-lpl/ex=SUB so.that AUX-NEG-lpl/ex=FRS
'No. We did not do so [i.e. watch him] in order to kill him.'

9.8 Verbless clauses

Under certain, very specific conditions, the copula is obligatorily omitted. Most
typically, verbless clauses occur as questions that involve a demonstrative, such
as illustrated in (596). Both examples contain the demonstrative kaa 'this'
(surfacing as a clitic in (596b)) and an interrogative pronoun. Example (596c)
illustrates the impossibility of using a copula in this context. The sentence is
ungrammatical, because it does not contain any deictic device such as the
demonstrative. The correct way to express 'What is this?' would be d$a htal

(596) Verbless clause with interrogative pronoun

a) Full demonstrative:
d$a teru kaa
[what axe] [this]
[CC] [CS]
'What (also 'whose') axe is this?'

b) Cliticised demonstrative:
d$a raj ka=teru
[what for] [this=axe]
[CC] [CS]
'Whose axe is this?' (lit. 'for whom')

c) No demonstrative: ungrammatical
*d$a ni-a / *d$a ne
[what] [be-3ps/A] what be-3ps/E
[CC] [V]
(To mean: 'What is it?')

As in (597), a verbless clause is also employed in a polar question that contains

a demonstrative.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
402 Verb classes

(597) Verbless clause with polar question

raj nuhue=na mi
[POSS excrement=INT] [that]
[CC] [CS]
'Is that its excrement?' [Pointing at the intestines of a peccary that is being cut into
pieces]

There also are examples for verbless clauses that do not contain a
demonstrative. However, these have deictic reference only provided by the
context, as the speaker is pointing at the object he is referring to. An important
difference to the ungrammatical example in (596c) is that verbless constructions
with this kind of deictic reference only occur with polar questions, but not with
content questions. Also note that the interrogative clitic =na is obligatory in this
context.

(598) Verbless clause with visual deictic reference

a)
ii enejtgu=na
[2sg monkey=INT]
[CC]
'Is (that) your monkey?' (Pointing at the monkey on the girl's lap)

b)
ii raj=pa
[2sg for=INT]
[CC]
'Is it yours?' (Pointing at the notebook on the table)

Another way to indicate deixis is with the word mihoa 'look here/there' or
'this/that is ...', comparable to the French voici/voilä. It is used with declarative
function only.

(599) Verbless clause with nuhoa

a)
nukwa ii lenone
look.here 2sg food
'Here is your food.' (Giving it to the listener)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 "Special" verbs 403

b)
rtukwa raj nuhue
look.here POSS excrement
'This is its excrement.' [Holding the peccary's intestines in his hand]

A verbless clause can also occur as a reply to a verbless question of the type as
illustrated above. For instance, a question such as 'whose is this?' may have the
answer kamt raj (lsg + POSS) to mean '[it's] mine.' These and further cases of
clauses that lack a verb can be characterised as results of ellipsis that occur as a
reply to a question, mainly realised as one-word utterances. For example, the
reply to the question 'who did it?' could be kanu T ; or the question 'what did
you cook?' may be answered as riiru 'howler monkey'. A more common way
to answer these questions, however, would be to include a full verb, such as Ί
did it' or Ί cooked howler monkey'.

9.9 "Special" verbs

Urarina has a number of verbs that do not form a homogeneous class, but which
exhibit unusual features that are not compatible with the properties as defined
for the verb classes discussed in this chapter. The following list gives an
overview of these verbs, with reference to the chapters in which they are
discussed.

(600) Special verbs

naaohwaa Enumerative verb, cf. §5.16

nukuja 'be in vain', cf. §17.2
ßaaohwaa Interrogative verb 'how much', cf. §21.6
ßatahaa Interrogative verb 'how much', cf. §21.6
pasia Interrogative verb 'do how', cf. §21.6
d$atoama Interrogative verb 'be.how', cf. §21.6
tonaa Interrogative verb 'sound how', cf. §21.6
äaka Interrogative verb 'be where', cf. §21.6

The numeral verbs for 'one' through 'five' may also be added to this list.
However, note that not all numerals behave in the same way (cf. §5.12). Thus,
they are not regarded a separate verb class.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
10 Word formation

Urarina morphology does not only exhibit a large number of inflectional forms
(cf. §12), but also has a complex system of derivation. On the one hand, this
involves word class-changing derivation, such as the various types of
nominalisation and the (less productive) verbalisation (see §10.1). On the other
hand, there is a wide range of verbal derivations, which do not involve a change
of word class. While it is impossible to describe the system in its entire
complexity (there might be as many as three dozens of verbal derivational
suffixes), the most productive processes will be investigated in §10.2.
Reduplication (see §10.3) is another instance of word formation that applies to
verbs. Urarina has no less than five different types of verbal reduplication (plus
subtypes), each of which have a slightly different grammatical function. There
also are a number of different compounding types, most of which result in
complex nouns. These are investigated in §10.4.

10.1 Word-class changing derivation

All morphology in Urarina applies to either nouns or verbs, with verbs being
predominant in attracting affixes, whereas only few affixes occur on nouns. As
a result, any word-class changing derivation either changes a noun into a verb
or vice versa.

10.1.1 Nom inal is at ion

In §6.6, the different kinds of relativisation were already discussed in detail. A

verb or clause can be relativised through nominalisation, which is divided into
four different types, as summarised here:
1. Nominalisation with suffix -i for intransitive subjects and objects; attached to
intransitive or transitive verbs.
2. Nominalisation with -era for agents; attached to transitive verbs only.
3. Nominalisation with -naa for agents or abstract nouns; mainly attached to
intransitive verbs.
4. Constructions with -nakauru 'those who' for intransitive subjects in the
plural.
5. Nominalisation with -naha, with locational purposive function 'a place to ...'
(cf. §20.2.4.1)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Word-class changing derivation 405

There is one further type of nominalisation, which is however not productive

and only found with a few words. It involves the nominaliser -o (or allomorph
-no) and appears to create an abstract noun from a verbal root. The known
examples are given in (601). These include a few nouns referring to dimension,
such as 'depth' and 'width', which result in abstract nominalisation. The nouns
leto-no 'envoy' and amiane-o 'work', differ from these, as they are instances of
object nominalisation.

(601) Nominal isation with -o

haraauto-a 'be wide' haraauto-no 'width'

ukuum-a 'be deep' ukuun-o 'depth'
amiam-a 'work' -» amiane-o 'work' (e.g. as in kanu amianeo 'my work')
letoa-a 'send' leto-no 'envoy'

Another, type of derivation with a limited degree of productivity is found with a

number of nouns that end in /e/, corresponding to verbs ending in /ia/ in the
citation form. It should be noted that all verb roots listed in (602) end in /e/; the
realisation as [i] before /a/ is the result of a general phonological raising rule (cf.
§3.4). Thus, one could argue that the process responsible for the derivation is
zero derivation, i.e. a noun is derived from a verb root without any segmental
material being added to the verb root. From a semantic perspective, it also
seems plausible that the nouns are derived from a verbal base, rather than vice
versa, though for some entries, this might be unclear (e.g. 'fat' is more likely to
be the base for 'become fat'). Another observation is that all base verbs for this
derivation type are intransitive.

(602) Derivations ending in -e

Verb Noun
ahi-a 'get drunk'-3ps/A ahe 'drunkenness'
eri-a 'speak'-3ps/A ere 'word'
hjam-a 'urinate'-3ps/A hjane 'urine'
lenoni-a 'eat'-3ps/A lenone 'food'
ßaueni-a 'be silent'-3ps/A ßauene 'silence'
rani-a 'weave'-3ps/A rane 'cloth'
hoereni-a 'become fat'-3ps/A hoerene 'fat'

There are a number of further examples for zero derivation that do not involve
the final root vowel /e/. In the same way as the examples in (602), the nouns in
(603) can be assumed to be derived forms of the respective base verbs, all of
which are intransitive again.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
406 Wordformation

(603) Further zero derivations

Verb Noun
toni-a 'sound'-3ps/A torn 'sound'
nesari-a 'hunt'-3ps/A nesari 'booty'
raui-tQa 'be straight'-3 ps/A raui 'right'
sim-a 'sleep'-3ps/A sini 'sleep'
hareheeri-a 'be thin'-3ps/A hareheeri 'thin person'
durana-a 'sweat'-3ps/A durana 'sweat'
hoitQana-a 'bleed'-3ps/A kwitQana 'blood'
kunu-a 'shine'-3ps/A kunu 'light'
itgau-a 'live'-3ps/A itgau 'life'

The semantic structure of the nouns resulting from this kind of derivation is
rather heterogeneous. While quite a number of nouns refer to abstract concepts
such as 'sound', 'sleep', or 'drunkenness', others refer to physical persons or
materials. For some nouns it may be implausible to assume that they are derived
from verbs, such as 'light' would derive from 'shine, or 'blood' from 'bleed'.
The derivation could also apply in the opposite direction. The decision has to be
postponed to future studies.

10.1.2 Verbal isation

Derivation with -oka is the only productive process that derives a verb from a
noun root. Since this verbalising suffix has already been discussed in §7.3.1, I
will only summarise its function here: Derivations with -oka (or allomorph -koa,
after vowel sequence) most typically refer to long-term possession, such as in
lana-oka 'she has a husband'. The suffix only applies to certain nouns, which
include terms for kinship and body parts, but also other important possessions.
Another case of derivation that has already been previously mentioned is the
use of verbal inflection forms such as the participle form or personal suffixes
with some nouns (cf. §5.16). The examples for this include words such as
hanuri-i 'turning one's back' (from hanuri 'back') and kuri-1 'applying jagua
fruit' (from Jairi'jagua fruit'), but this type of derivation is not productive.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other types of verbal derivation 407

10.2 Other types of verbal derivation

10.2.1 Size and number derivations

Verbal derivations that form another verb from an existing one are very
common and follow a complex set of rules. In fact, the division into different
subclasses of intransitive verbs is based on these distinctions. There is a set of
derivational suffixes that exclusively applies to verbs of position, shape, and
colour (PSC verbs), and to certain "affect" verbs (as defined in §9.2.3).
Sufflxation of these morphemes to a verbal root may determine the transitivity
of the emerging verb. For instance, one group of suffixes creates intransitive
stative verbs (Type A); another has an inchoative function (Type B). Type C
derivation generates a transitive verb, and Type D reflects a "caused state", as
exemplified in (604) (also cf. §9.2.3).

(604) Derivation types

Type A: Stative position, e.g. 'be seated'
Type B: Inchoative; motion into a position, e.g. 'sit down'
Type C: Causation of Type A, e.g. 'make somebody sit'
Type D: Caused/inchoative state, e.g. 'caused to be seated as a result of somebody
causing it'/ or 'out of a motion'

A peculiarity of all four types of derivation is that they involve different

suffixes with respect to size and number. Size is distinguished between 'small'
and 'big', the choice of which depends on the speaker's judgement and on the
context. The distinction of number varies according to the type of derivation,
independent from normal plural marking with the suffix -uru. The tables in
(605a-d) give an overview of the different suffixes, illustrated with the verb root
'be standing' (on one or two feet).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
408 Word formation

(605) Productive derivational suffixes for position verbs

a) Derivation A (Resulting in intransitive state):

Number SMALL BIG Example (Small / Big)
Singular (1) -ekoa -akoa; -aka d^uhu-ekoa / d^uhu-aka
Dual (2) -r(V)hatoa -neria d^uhu-ruhatoa / d^uhu-neria
Paucal(3-5) -r(V)haa+RED -nere+RED d^uhu-ruha+d^uhu-ruhaa /
d^uhu-nere+d^uhu-neria
Greater plural na- (VV)-koa na- (VV)-koa m-d^uhu-aakoa
(4+)
Distributive -roa -roa d$uhu-aroa
plural

b) Derivation Β (Resulting in intransitive action / motion):

Number SMALL BIG Example (Small / Big)
Singular -turaa -lenia; -netoa d^uhu-turaa /d^uhu-lenia
Paucal(2-4) -tura+RED -lene+RED d^uhu-tura+d^uhu-tura-uru-a /
d^uhu-lene+d^uhu-lene-uru-a
Greater na- (VV)-turaa na-_JVV)- na-d^uhuu-turaa
plural (4+) turaa

c) Derivation C (Resulting in transitive action / causative):

Number SMALL BIG Example (Small / Big)
Singular -taa -l(V)itoa d^uhu-taa /d^uhu-luitoa
Non-singular -(V)ne+RED- -(V)sia d^uhuu-ne+d^uhuu-nia /
(2+) (nia) d^uhuu-sia

d) Derivation D ( Resulting in transitive state / causativised):

Number SMALL BIG Example (Small / Big)
Singular -ekaa -haa d$uhuu-ekaa /d^uhu-raa
-raa
Non-singular -ekaanaa -(VV)haria d^uhu-ekaanaa jd^uhu-haria
(2+)

With T y p e A derivations, up to five different forms are distinguished for each

size. The existence of a special derivational form referring to exactly two
participants is exceptional in Urarina morphology, as the dual is not a persistent
feature. It otherwise only occurs in the distinction of first plural suffixes on
verbs, where it marks a dual inclusive ' w e ' (cf. §12.2.16). Type A also
recognises a special paucal form that refers to 'a f e w ' , which may comprise
three, four, or up to five participants as a subject. The interpretation of ' f e w ' is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other types of verbal derivation 409

subject to the context: Out of a group of a hundred, four may be considered as

"few", whereas four out of ten could be regarded as more than that. Eventually,
there also is speaker variation for the interpretation of "few".
A further form describes a greater plural, referring to ' m a n y ' (i.e. four or
more), which must be part of one "group", or items gathered at the same place
as a subject. 53 Another form explicitly refers to a distributive plural, which
implies that 'various groups' or collectives are present at different places.
Examples for the respective contexts are given in (606). For both kinds of plural
(greater plural and distributive), the size distinction into big/small is neutralised.
It is also possible to refer to more than five without the collective connotation.
In this case, the regular plural suffix -uru is attached to any of the derived
forms optionally. For instance, the singular form can be marked with the plural
suffix (e.g. hekueko-oru-a 'they (small) are lying on their sides') to imply 'more
than one', but not specifying the number any further.

(606) Interpretation of different number types

a) Dual:
mee-rihjatoa
'be lying face down', about flat long items; here: two crocodiles are lying.

b) Paucal:
tuku-erihja+tuku-erihjaa
'be lying', about round lengthy items; here: four rifle shells.

c) Greater plural:
na-baba-akoa
'be sitting together', about persons as couples; here: various couples being seated in one
place.

d) Distributive plural:
te-roa
'be standing', about items on four or more legs (incl. houses); here: various settlements
in different parts of the river, one not visible from the next one.

The number distinctions with Type Β derivations are fewer and slightly differ
from Type A. Here, the paucal form refers to "two to maximally four", due to
the absence of a specific dual form. Again, there is an overlap between the

53
It is obvious that the presence of four or five subject items could be represented either
by the paucal or by the greater plural form. The choice depends on whether the
speaker regards the number as "few" or "many" in the light of the context.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
410 Word formation

paucal and the greater plural, as the latter may be applied with a presence of
four or more subjects. A distributive plural form does not exist for this
derivation type. Derivation types C and D are different again, as they only
distinguish between singular and plural, including what would be paucal or
greater plural with Types A and B.
The derivations listed here are nearly 100% productive for all position verbs,
with a few exceptions where a verb may not be compatible with all the forms.
In addition to these verbs, the same suffixes may also be combined with certain
other verbs. These include some verbs that refer to shape, such as mauekoa 'be
short', habeehoa 'be round', or tuniehoa 'be narrow' just to mention a few.
These behave like posture verbs as they are entirely compatible with the
derivational suffixes mentioned above. With Type Β derivation for inchoative
function, this implies that something "becomes" round, short, or whatever shape
is relevant. The same principle also applies to the verbs that refer to the colours
'black', 'white', and 'red'.
There is an additional group of verbs that can undergo some of the
derivational processes described here. These are derived from specific transitive
verbs that involve cutting, breaking or other kinds of action that affect the
integrity of an item (cf. §9.2.3). With these verbs, the suffixation of size and
number suffixes is much less productive than with the other verb types; i.e.
suffixation of the respective ffix is not possible. The table in (607) shows the
most productive example of compatibility between an "affect" verb and the
derivational suffixes under investigation.

(607) "Affect" verb saua 'cut' with derivational suffixes

Type A SMALL BIG Interpretation
Singular (1) sau-ekoa sau-akoa 'be cut a little' / 'be cut
normally'
Dual (2) sau-rahatoa sau-neria 'be cut at two parts'
Paucal(3-5) sau-rahaa+RED saunere+sauneria 'be cut at several (more)
parts'
Greater plural na-sattaa-hoa na-sauaa-koa 'be cut all over, in various
(4+) parts, everywhere'
Distributive *sau-roa *sau-roa —

plural
Type Β
Singular *sau-turaa *sau-lenia —

Paucal(2-4) *sau-tura+REO *sau-letie+RED —

Greater plural *na-saau-turaa *na-saau-turaa —

(4+)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other types of verbal derivation 411

TypeC SMALL BIG Interpretation

Singular sau-taa *saau-luitoa 'is cutting' (transitive)
Non-singular saaune+saauma *saau-sia 'is cutting into several
(2+) (small) pieces' (transitive)
TypeD
Singular sau-ekaa saauha+saauhaa 'is cutting' (transitive)
Non-singular sau-ekaanaa saauha+saauharia 'is cutting various'
(2+) (transitive)

As is evident from (607), there are a number of gaps in the paradigm. In

particular, the size distinction is neutralised for many forms. With derivation
Type A, the distinction is retained, but an additional connotation is involved:
talking about something small automatically implies that it is cut "a little",
whereas the use of the form for big items implies a regular or normal manner of
cutting. The distributive plural form is not compatible with saua. Another
difference to position verbs is that none of the forms for derivation Type Β
(inchoative) applies to "affect" verbs. The function of derivation types C and D
significantly differs from what was observed about position verbs. Firstly, the
size distinction for Type C suffixes is neutralised. Secondly, with saua, the
suffixation of Type C suffixes takes an aspectual meaning 'is cutting', while the
"causative" function disappears, which is plausible as the verb root is transitive
already. The plural form saaune+saauma implies that the object is small (but
there is no corresponding form for big objects). Type D derivations imply that a
small vs. big subject is acting, again with an aspectual connotation.
Alternatively, it can be understood that the object that is being cut is small vs.
big. This ambiguity corresponds to the one also observed with Type A
derivations, with the difference that Type D derivations result in a transitive
verb.
There are significant differences between the various "affect" verbs, which
makes it difficult to describe the degree of productivity for this verb type. While
saua is compatible with many of the suffixes, fewer derivations are possible for
most other verbs of this type. Further investigation is to follow.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
412 Wordform at ion

10.2.2 Derivations of degree

There are three suffixes that describe the degree to which an activity or process
is performed. Except for a few cases, these all refer to states related to posture,
shape, or colour; this applies to active intransitive and transitive verbs as well,
which may be converted into PSC verbs. The productivity of each suffix differs
with respect to the verb types it is attached to. In the tables (608) to (613), I
have indicated the degree of productivity by referring to the estimated quantity
of verbs found with each derivation type.

Diminutive "-eri"
The suffix -eri (or allomorph -heri) can be attached to any verb type, with
different degree of productivity. The transitivity value of each verb is not
affected by suffixation -eri. With PSC verbs, the suffix occurs after the
respective size suffix, as in lau-eho-eria 'be sitting a little'. While this suffix
implies a "small" subject when occurring by itself, size tends to be neutralised
in combination with -eri. Though the corresponding form for big items as in
lau-ako-eria is theoretically possible, speakers prefer to use the form for small
subjects (lau-eho-eria) to refer to big items as well. For most verbs, a size
distinction is either uncommon or impossible, as indicated in (608). The
semantic function of -eri refers to 'a little', 'not entirely', or 'partly'.

(608) Derivational suffix -eri

Verb type Example Meaning Productivity

Position lau-eko-eria 'be sitting a little' (small item); size All verbs
can be distinguished but uncommon
Shape habe-eko-eria 'be a little round'; size can be All verbs
distinguished but uncommon
Colour soma-eko-eria 'be a little white'; size can be All verbs
distinguished but uncommon
"Affect" sau-eko-eria 'be cut a little bit' Most verbs
transitive
Transitive kala-eria 'bite a little' Few verbs
Active obaae-heria 'be a little angry' Some verbs
intransitive
Stative setee-heria 'be a little wet' Some verbs
Reflexive nerutu-hweria 'change a little' (intr.) Some verbs

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other types of verbal derivation 413

As indicated in the far right column of (608), the suffix -eri is less productive
with verbs other than PSC and affect verbs. While most affect verbs can take
this suffix, there are only a few other transitive verbs that can occur with -eri.
With active and stative intransitive and with verbs intransitivised by rte-, the
suffix is slightly more productive.

Diminutive "-heriiri"
Another diminutive suffix, -heriiri (or allomorph -eriiri) expresses 'very little',
referring to a lesser degree than -eri. This morpheme describes states that occur
only ' a tiny little bit'. It is less productive than -eri, as it is not observed with
position or shape verbs, and it is not attached to any transitive verb. However, it
can occur with colour verbs and with active and stative intransitive verbs, as
well as with reflexive verbs.

(609) Derivational suffix -heriiri

Verb type Example Meaning Productivity

Position — None
Shape — None
Colour soma-heriiria 'be a tiny little bit white' All verbs
"Affect" — None
transitive
Transitive — None
Active sini-heriiria 'be a tiny little bit sleepy' Some verbs
intransitive
Stative setee-heriiria 'be a tiny little bit wet' Most verbs
Reflexive nerutu-heriiria 'be changed a tiny little bit' Some verbs

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
414 Word formation

Medium degree marker "-ohwa" and variants

A further distinction of degree is expressed through the suffix -ohwa, which has
the phonologically-conditioned allomorph -kohwa and the variants -koa and
-hiohioaa, whose distribution is not predictable, but which refer to the same
semantic concept. The meaning of these suffixes involves the notion of
' m e d i u m ' , 'not entirely but almost', 'more or less', or 'quite', which expresses
that a state or action has not been entirely accomplished to its normal degree,
but to a higher degree than with the other t w o suffixes (-eri and -heriiri). -ohwa
and its variants are very productive as they can occur with verbs of any class,
with minor differences in productivity. The transitivity value of a verb remains
unchanged. Some verbs can take two variants of this suffix without any
apparent distinction in meaning. Other verbs take only one specific form;
however, the distribution of the different variants cannot be tied to specific
subclasses, as is evident from the examples below. Size distinctions again vary
according to the type of base verb.

(610) Derivational suffixes -ohiua, -koa, and -hiohiua

Verb type Example Meaning Productivity

Position me-eko-ohwaa 'be almost lying' (small All verbs
item); size is distinguished
Shape maka-ako-hwaa; 'be almost flat' (big item); All verbs
makaa-hiohwaa size is distinguished
Colour soma-ah-hwaa; 'be almost white' (big item); Most verbs
somaa-hiohioaa size can be distinguished but
uncommon
"Affect" sau-ako-hwaa 'be almost cut' (big item); Some verbs
transitive size can be distinguished but
uncommon
Transitive auna-ohwaa 'understand more or less' Some verbs
eseneta-koaka 'believe more or less'
Active sim-ohwaa 'sleep almost normally' Some verbs
intransitive
Stative taba-ohwaa 'be quite big'; Some verbs
helaa-hiohwaa 'be quite cold'
Reflexive nebetaka-ohwaa 'relax more or less' Some verbs

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other types of verbal derivation 415

10.2.3 Derivations of manner

Urarina has two derivational suffixes that are attached to verbs in order to
express velocity or duration. Both convey the idea that something is performed
quickly or with one single motion. One of the suffixes goes a step further as it
refers to an action that happens 'passing by', or 'for a short moment only'.
There also is a general intensifying suffix that has a wide range of functions.

Suffix "-alaito " and variants

The suffix -alaito and its variants -aito, -luito, and -ito express the concept of
something that happens 'once', 'within a moment', 'with one movement', or
'quickly'. For instance, the verb molo-itoa implies that something is cut 'with a
single cut'. An interesting peculiarity of this suffix is its transitivising function:
verbs of any class become transitive when -laito is attached. Transitive affect
verbs are not increased in their valency, i.e. they are not causativised. As some
other forms discussed in this chapter, -laito has a number of variants whose
distribution is not easy to predict. However, it is obvious that the variant -ito is
mainly found with position verbs and affect verbs. Further variants include the
forms -aito and -luito.

(611) Derivational suffix -(a)laito, variants -aito, -luito, -ito

Verb type Example Meaning Productivity

Position lau-itoa 'make sit quickly/at once'; All verbs
about big items; (cf. small
lautaa)·, lauluitoa is
uncommon
Shape maka-laitoa 'make flat'; about big items; cf. All verbs
makaitoa\ cf. makataa (small)
Colour soma-laitoa 'make white'; about big items; All verbs
cf. somaitoa
"Affect" sau-itoa 'cut with one movement' All verbs
transitive (while holding it); sauluitoa is
uncommon
Transitive — None
Active — None
intransitive
Stative kalano-aitoa 'make dirty at once' Uncommon
Reflexive — None

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
416 Word form at ion

Suffix "-ahaito"
The suffix -ahaito (or allomorph -haito) refers to short duration and describes
concepts such as '(while) passing by', 'quickly', 'easily with one movement',
'once', 'for a short moment', frequently translated as de paso in Spanish.
Remarkably, it transitivises stative intransitive, but not active intransitive verbs.
Furthermore, it is not compatible with PSC verbs. When suggested in
elicitation, speakers refer to the form with -laito as having a similar function.
Thus, the minute conceptual difference between the two suffixes, which lies in
duration vs. velocity, is neutralised for these verbs. This makes sense in so far
as an expression such as 'make stand (while) passing by' would be hard to
interpret.

(612) Derivational suffix -(a)haito

Verb type Example Meaning Productivity

Position — None
Shape — None
Colour — None
"Affect" sau-ahaitoa 'cut while passing by' All verbs
transitive (without holding it; cf.
sauitoa)
Transitive hoara-häitoa 'look while passing by' Most verbs
Active eri-ahäitoa 'talk while passing by' Most verbs
intransitive
Stative taba-ahaltoa 'become big quickly' (e.g. Few verbs
river)
Reflexive nejwa-haltoa 'increase quickly, for a Most verbs
moment only'

Suffix "-to"
The suffix -to, as illustrated in (613), can be characterised as a general
intensifier. It is found with all verb classes except position verbs. Its function
includes a range of meanings, such as 'strongly', 'deeply', 'very', 'quickly',
'actually', or 'various times', much depending on the context and the verb -to is
used with. With shape and colour verbs though, a special function is involved,
as the suffix then refers to 'for a short moment', i.e. it has exactly the function
as otherwise assigned to -ahaito (cf. (612)). A clause such as 'be red for a short
moment' will usually imply 'something red occurred for a short moment, (e.g.
while passing by)'. Another observation about shape and colour verbs is that
they only occur in the form for "big" items. It is also possible to create an

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other types of verbal derivation 417

intensified form of shape and colour verbs, but in this case, other suffixes have
to precede the intensifier -to. Since these morphemes are not productive
otherwise, they will not be discussed here.
Stative verbs represent a yet different case as many of these inherently
contain the suffix -to, which with some examples may be identifiable as the
intensifier. However, this does usually not imply an intensifying function of this
suffix. In fact, nearly all stative verbs that contain -to do not have a form
without it. Also note that suffixation of -to never changes the transitivity of a
verb.

(613) Intensifying suffix -to

Verb type Example Meaning Productivity

Position — None
Shape habi-a-toa; 'occur for a short moment All verbs
being round'; about big item;
habe-reru-toa 'be very round'
Colour lana-a-toa; 'occur for a short moment All verbs
being red'; about big item;
soma-ra-toa 'be very white'
"Affect" huku-toa; 'stab intensely/various Most verbs
transitive teru-toa times';
'cut actually'
Transitive tarara-toa; 'pull intensely'; Most verbs
su-toa 'kill many/various
times/successfully'
Active sini-toa; 'sleep deeply'; Most verbs
intransitive turu-toa; 'arrive quickly';
ßadera-toa 'be very sad'
Stative eloa-toa; 'be bitter' Many verbs
kaua-toa 'be good'
Reflexive nebetaka-toa; 'relax well/a lot'; Most verbs
nefwa-toa 'increase much'

There is another pair of derivational suffixes that may be mentioned here. The
suffixes -ruruka and -ruruhja combine functions of distribution and motion, for
big vs. small items, respectively. For instance, hoitgam-rurukaa can be
translated as 'be moving bleeding at various parts' (based on the root for
'bleed'), about big items; whereas muku-ruruhjaa (based on ' b u r n ' ) means ' b e
moving burnt at various parts', but referring to small items. Examples are
observed with transitive, intransitive, affect, and stative verbs, but it is unclear
whether these suffixes are compatible with other verb classes.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
418 Wordformation

10.3 Reduplication

Urarina has five different kinds of verbal reduplication, two of which have
additional subtypes. Their functions exhibit a high degree of overlap between
aspectual interpretation ('repeatedly'), spatial distributive ('in various areas'),
and motion ('move/run'). The table in (614) gives an overview of the attested
types. All types are instances of total reduplication; differences between the
various types are indicated by additional morphemes such as suffixes.

(614) Types of reduplication (R = Root; underlined):

No. Type Example Function

la RED+R-[VV]ka, teme+teme-eka Position/state and motion; 'move in a
RED-a+R-akoa certain state or position'; big items
(cf. (616))
lb RED-e+R-ekoa lau-e+lau-ekoa Position/state and motion; 'move in a
(cf. (617)) certain state or position'; small items
2a RED+R- lau+laa-oka Position/state and quick motion; 'run
[VV]oka in a certain state or position'; big
(cf- (618» items
2b RED+R-eka lau+lau-eka Position/state and quick motion; 'run
(cf. (619» in a certain state or position'; small
items
3 RED-e+R-eka sau-e+sau-eka Aspectual: repeatedly, as a habit
(cf. (620))
4 RED+R-a sau+sau-a Aspectual meaning (repeatedly);
(cf. (621)) PSC verbs: Distributive 'in various
areas'; big items
5 RED+R-eria teme+teme-eria Distributive 'in various areas'; small
(cf. (622)) items

It should be noted that most types are very productive with PSC and affect
verbs, but do not or hardly occur with other verb classes (except Type 3). Thus,
it may not be surprising that some reduplication types occur as "pairs", each
distinguished for big and small items, as is observed with suffixal derivation. In
fact, each type of reduplication is enriched by the addition of other segments
that are partly also found in suffixal derivation, such as -akoa and -ekoa with
Type la and lb reduplications, respectively. The productivity of each
reduplication type is summarised in (615), which illustrates the major role PSC
verbs play for this type of word-formation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Reduplication 419

(615) Summary of productivity with reduplication types for different verb classes

Type la Type lb Type 2a Type 2b Type 3 Type 4 Type 5

Position Most All All All None All All
Shape All All All All None All All
Colour All All All All None All All
"Affect" Some Some Few Few All Most Most
transitive
Transitive None None None None Most Most None
Active None Few None None Most Few None
intransitive
Stative None None None None Few Few None
Reflexive None None None None Most Affect V None

Type la: RED+R-fWJka

The first two reduplication types ( l a and lb) only apply to verbs that can also
take size-distinguishing suffixes otherwise (i.e. PSC and affect verbs). Type l a
explicitly refers to big subjects; affect verbs are intransitivised through the
derivation. There are two possible interpretations for this form: the first one is
of a distributive nature, expressing 'in various parts/areas'. For instance, a cloth
can be white in various parts/areas (but coloured differently in other
parts/areas). The second reading refers to motion, such as 'be moving in lying
position'. Though this reading is only given for position and affect verbs, it is
plausible that it also applies to shape and colour verbs (though not attested with
specific examples). Conversely, the distributive interpretation might also apply
to position verbs, as it does with the other classes.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
420 Wordformation

(616) Reduplication RED+R-[VV]ka (big items), RED-a+R-akoa

Verb type Example Meaning Productivity

Position baua+bauakoa 'be moving in lying position' Most verbs
Shape maka+makaaka 'be flat in various parts/areas' All verbs
Colour soma+somaaka 'be white in various parts/areas' All verbs
"Affect" soaj+soaika 'be moving in broken state/shape'; Some verbs
transitive or: 'be broken in various parts/areas'
Transitive — None
Active — None
intransitive
Stative — None
Reflexive — None

Type lb: RED-e+R-ekoa

Exactly the same function as for Type l a applies to Type l b reduplication, with
the difference that it refers to small subjects. An additional difference is that
there are a few active intransitive verbs that can be subject to this reduplication
type. In this case, the interpretation can be aspectual, such as in 'talk several
t i m e s ' , (but also implying 'in different places'; cf. (617)). With these verbs, the
size distinction is neutralised.

(617) Reduplication RED-e+R-ekoa (small items)

Verb type Example Meaning Productivity

Position laue+laueha 'be moving in a sitting position' All verbs
Shape makaae+makaaekoa 'be flat in various parts/areas' All verbs
Colour somaae+somaaekoa 'be white in various parts/areas' All verbs
"Affect" soaj+soaitga 'be moving in broken Some verbs
transitive state/shape'
Transitive — None
Active ere+ereeka 'talk several times' Few verbs
intransitive
Stative — None
Reflexive — None

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Reduplication 421

Type 2a: RED+R-[W]oka

This form slightly modifies the meaning of the reduplication type discussed
above as it implies not only the motion component but specifically refers to
"fast motion", such as 'be running'. For instance, lana+lanaaoka (cf. (618))
refers to something white that runs while passing by, or is visible for a very
short moment only. Again, there is a division into big and small items: Type 2a
reduplication refers to big subjects. The productivity of this form corresponds to
the one found with Type la reduplication as it only applies to PSC and affect
verbs. The latter undergo a change in transitivity as they become intransitive,
describing a state such as 'something that has been cut is running'. With other
transitive verbs, Type 2a reduplication is possible in elicitation, but uncommon.

(618) Reduplication RED+R-[VV]oka (all about big items)

Verb type Example Meaning Productivity

Position lau+laaoka 'be moving quickly in a sitting All verbs
position'
Shape maka+makaaoka 'be moving quickly as something All verbs
flat' (e.g. turtle)
Colour lana+lanaaoka 'be moving quickly as something All verbs
white'
"Affect" sau+saaoka 'be moving quickly as something Few verbs
transitive cut'
Transitive (kala+kalaaoka) = kala-erarurukaa Uncommon
Active — None
intransitive
Stative — None
Reflexive — None

Type 2b: RED+R-eka

Parallel to the size distinction between Type la and lb reduplications, Type 2b

represents the "small" counterpart to Type 2a. Its productivity corresponds to
this type as well.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
422 Wordformation

(619) Reduplication RED+R-eka (all about small items)

Verb type Example Meaning Productivity

Position lau+laueka 'be moving quickly in a sitting All verbs
position'
Shape habe+habeeka 'be moving quickly as something All verbs
round'
Colour soma+somaaeka 'be moving quickly as something All verbs
white'
"Affect" sau+saueka 'be moving quickly as something Few verbs
transitive cut'
Transitive — None
Active — None
intransitive
Stative — None
Reflexive — None

Type 3: RED-e+R-eka
Another reduplication type adds the vowel /e/ to the first component (the
reduplicant) and the suffix -eka to the second. It significantly differs from the
types listed above in various aspects: firstly, no size distinctions apply to the
verbs that are subject to reduplication. Secondly, the productivity is a mirror
image to the other types, as PSC verbs are not compatible with this form,
whereas transitive and intransitive verbs are. However, there are only a few
stative verbs that are found with this type of reduplication. Remarkably, their
meaning in combination with this reduplication type corresponds exactly with
the "motion + state" function of Type la ('be moving in a certain way'),
whereas this is not the case with other verb classes and Type 3. The function
with these verbs is of an aspectual nature to imply 'regularly', 'repeatedly',
'again and again', or 'as a habit'. The transitivity value of a verb is not affected.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Reduplication 423

(620) Reduplication RED-e+R-eka

Verb type Example Meaning Productivity

Position — None
Shape — None
Colour — None
"Affect" saue+saueka 'cut regularly, again and All verbs
transitive again'
Transitive kalaae+kalaaeka 'bite regularly, various times' Most verbs
Active turue+turu-eka 'arrive several times, or at Most verbs
intransitive various places'
Stative kauatQaae+kauatQaaeka 'be moving nicely' Few verbs
Reflexive nerutue+nerutueka 'be changing again and again' Most verbs
(e.g. TV screen)

Type 4: RED+R-a
Another reduplication type related to aspectual function is formed by simply
reduplicating the root and adding the respective person suffix to this stem (or,
the neutral suffix in citation form). Its function involves the quick repetition of
an action in short intervals (except with stative intransitive verbs where it refers
to various "parts" of a state). 'Repeatedly' here implies that something is done
once and then again after a short break. There can also be a connotation of a
progressive form 'now', which depends on the context. For example, sau+saua
may imply 'he is cutting it quickly several times in a sequence', as indicated in
(621).
With PSC verbs, the meaning slightly differs from this, as it shifts to a more
spatial function, such as 'in various parts'. The example mee+mia 'make lie
down' explicitly implies a distributive function, meaning that the object(s) are
put to various places, such as leaves that are laid on the ground in order to dry
them. There is a variation in interpretation, which may also imply 'various
times'. It should also be noted that all PSC verbs become transitive when used
with this reduplication type. However, this does not apply to the other
intransitive verb classes. Another difference between PSC and other verbs is
that the distributive meaning only applies to the former, whereas an aspectual
meaning applies to the latter. Note that with reflexive verbs, only those derived
from affect verbs may undergo Type 4 reduplication. Otherwise, this form
occurs only with a few active and stative intransitive verbs.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
424 Wordformation

(621) Reduplication RED+R-a

Verb type Example Meaning Productivity

Position mee+mia 'make lie down in various parts'; e.g. All verbs
drying leaves; about big item
Shape habe+habia 'make round in various parts'; also All verbs
'several times'; about big item
Colour lana+lana 'paint red in various places'; about big All verbs
item
"Affect" sau+saua 'cut quickly several times in a Most verbs
transitive sequence'; about big item
Transitive hoara+hoaraa 'look several times' (e.g. turn around Most verbs
and look again)
Active räasa+räasaa 'dance again and again' (taking Few verbs
intransitive breaks)
Stative kaltate+kaitatoa 'various parts of it are difficult' Few verbs
Reflexive nemolo-nemoloa 'it breaks in various parts/areas' All refl.
affect verbs

Type 5: RED+R-eri
Reduplication with the addition of -eri is restricted to PSC and affect verbs also.
The diminutive function as it occurs in derivations with -eri does not apply to
Type 5 reduplication. Instead, this form has a distributive function, meaning 'in
various parts'. This corresponds to the meaning observed with Type 4
reduplications. The difference is again based on size distinction: Type 5
explicitly refers to small items, whereas this is not the case with Type 4. With
Type 4, PSC and affect verbs imply the involvement of a 'big' subject, (but note
that this meaning does not necessarily apply to the other verb classes with this
form).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Compounding 425

(622) Reduplication RED+R-eri (all about small items)

Verb type Example Meaning Productivity

Position me+meeria 'be lying in various parts/areas' All verbs
Shape maka+makaaeria 'be with or have various flat parts' All verbs
Colour soma+somaaeria 'have various white parts' All verbs
"Affect" sau+saueria 'be cut in various areas' Most verbs
transitive
Transitive — None
Active — None
intransitive
Stative — None
Reflexive — None

10.4 Compounding

Nine more or less productive types of compounding can be identified. All types
involve nouns in some way, and each type has its own semantic structure. In
most cases when a verb is involved the head of the compound will still be a
noun, sometimes through nominalisation of the verb. However, the occurrence
of verbs in compounds is comparatively rare. Only in one type of compounding
is the verb the head of the construction (cf. (632)). Except in compounds that
involve the associative marker hi- (cf. (625)), the interpretation of the internal
structure is hardly predictable, that is, it is mainly a matter of lexicalisation
and/or intuitive interpretation. In most compound types, the second element
(typically a noun) functions as the head. Deviations from this occur with such
cases where the compound can be interpreted as a noun-modifier construction
similar to the ones observed with adjectives. The table in (623) gives an
overview of the common types of compounding.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
426 Word formation

(623) Types of compounding

No. Type Head Structure

1 N +N N2 Possessive NP (Ν 1 = possessor)
2 Ν + ku-N N2 Associative NP (Ν 1 = source or goal)
3 N+N N2 N1 = modifier
4 N+N exocentric / N2 Possessive NP
5 V+N exocentric / Ν Ν modified by V
7 N+N N1 N1 modified by N2; N2 = nominalised verb
6 N+V Ν Ν modified by V (but V can be interpreted as
nominalised)
8 V+N Ν Ν modified by V
9 N + V; V Ν is the object of V
N+N N2 N1 is the object of N2 (nominalised verb)

The semantic and tonal structure of Type 1 compounds corresponds to that of

any possessive NP, with a High tone on the second syllable of N 2 (with certain
alternations; cf. §4.6.1). However, this tonal behaviour is not exclusive to this
kind of compound, as also most other types may occur in this shape, but without
the interpretation as a possessive NP.

(624) Type 1 compounds

kumi+temura 'light' + 'source' 'generator'; lit. 'source of

light'
tihja+siiri 'lower leg' + 'belly' 'calf
bihi+kuturi 'hand' + 'head' 'finger'
bihi+kuturi+nebä 'hand' + 'head' + 'mother' 'thumb'
enoe+nunera+i μόηο 'liana' + 'forest' + 'ayahuasca' 'wild ayahuasca'
isi+kuräanaa 'fire' + 'chief 'major piece of firewood
from the fire'

A formal distinction to other compounds is found with Type 2 compounds,

which involve the associative prefix ku~. Recall that this type implies reference
to a source or goal, such as "axe made of stone" (source) or "stable for
chickens" (goal; cf. §7.3.3).

(625) Type 2 compounds

aheri+ku-teru 'stone' + ASC-'axe' 'stone axe'

atauari+ku-lueri 'chicken' + ASC-'house' 'chicken stable'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Compounding All

Compound Type 3 does not have a possessive interpretation, though formally it

cannot be distinguished from Type 1. The tonal structure of most examples
corresponds to that of Type 1, as well as its morphological shape, (which simply
is a concatenation of two nouns). However, semantically, Type 3 compounds
are different, since they describe a noun where N2 is modified by N l . For
instance, laano+d^akari (cf. (626)) is not "cassava's crocodile", but a crocodile
whose appearance is associated with the shape and colour of a cassava plant,
(which refers to the belly of the crocodile). Referring to the attributes of N2 in a
similar way, kukuri+eriauru is a type of wasp whose nest has the shape of an
armadillo. One possible translation for these examples would be with a
comparative 'like' such as in 'cassava-like crocodile', 'armadillo-like wasp',
and 'water-like urine'.

(626) Type 3 compounds

laano+d^akäri 'cassava' + 'crocodile' 'white crocodile'
alajtQÖo eruari 'lancehead snake' + (type (type of root called jergon sacha)
of root)
kukuri + enäuru 'armadillo' + wasp' (type of wasp whose nest has the
shape of an armadillo)
akau + hjäne 'water' + 'urine' 'urine water' (from a story where
urine turned into water to flood the
earth)

Types 4 and 5 are exocentric compounds, as neither of their components may be

characterised as the head, at least in semantic terms. (Note that
morphologically, N2 is the head as it is subject to plural inflection.) Some
examples for Type 4 have a name compared to attributes of certain objects, such
as the banana species whose literal name is "tooth of huapapa bird" (referring
to its beak; cf. (627)). Technically, all of these compounds correspond to
possessive constructions, but they have a metaphorical interpretation.

(627) Type 4 compounds

baka+isitoo 'cow' + 'breast' (type of plant called teta de
vaca)
atari+nutäae 'paufd bird' + 'eye' (type of nut)
fwafivafwa+katu 'huapapa bird' + 'tooth' (type of banana called sapucho)
hu+asahaj 'skirt' + 'under' (N) (cloth used as underwear by
women)
ahäaori+kwit Qana 'turtle' + 'blood' 'birth-mark' ('mole')

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
428 Word formation

Type 5, which is rather rare, has the same properties as Type 4, with the
difference being that the first part of the compound is a verbal root.

(628) Type 5 compounds

lura+kuttiri 'be bald' + 'head' 'bald person'

maue + bihi 'be short' + 'hand' 'person missing a finger'

Compounds of Type 6 are usually quite transparent as they involve a noun

followed by a modifier (cf. §6.5). For instance, 'red banana' is ambiguous as it
could be interpreted as an ad hoc compound to describe any sort of red banana
(even if it were painted). Alternatively, it refers to the specific type of species as
implied here. The example amiri+siri 'bat' in (629) may be an exception from
this, as it is not transparent.

(629) Type 6 compounds

fwanara + larnhaj 'banana' + 'red' 'red banana' (of guineo type)

akauijio + hitQiihzoaj 'boa' + 'black' 'black boa' (species)
anuri + sirt 'parrot' + 'old' 'bat'

Type 7 is very rare as only two examples are attested. It involves a noun
followed by a verb root that functions as a relativising modifier. This is unusual,
as such a construction would normally involve a nominalising suffix, such as
described in §6.6. Instead, there is no morphological marking on the verb
whatsoever, which implies that its occurrence could be interpreted as a kind of
zero derivation. Similar to the examples in (603) further above, one could
postulate that the verb root in the examples in (630) represents a derived noun.
Note that plural marking is applied to the compound as a whole, i.e. after the
second component. The example nuhte+suru 'tributary river' can be
characterised as being highly fossilised, as not all speakers recognise the
component /suru/ as the root of the verb 'run'.

(630) Type 7 compounds

enanihja + ihjaulu 'canoe' + 'flying' 'airplane'; lit. 'flying canoe'

nukue+suru 'river' + 'run' 'tributary river'; lit. 'running river'

There also are a few examples of compounds that exhibit the structure verb +
noun, with a similar function as Type 7. For instance, homo+enua is a 'tree that
has fallen', which points at a relativisation of the noun. Slightly differently,
simhera is not simply a 'companion who sleeps', but more specifically refers to
a 'companion to sleep with'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Compounding 429

(631) Type 8 compounds

homo + enua 'fall' + 'tree' 'fallen tree'

ere + enua 'talk' + 'tree' 'talking tree'
sini + hera 'sleep' + 'companion' 'wife'

Type 9 is the only kind of compound where the verb is the head, such as the
term for 'advice', which literally means 'give heart'. One could possibly argue
that these constructions are in fact phraseologisms or instances of idiomatic
expressions. However, in this case, their tonal structure should be predictable
f r o m the noun, as the noun determines on which syllable of the verb a High tone
will occur (cf. §4.1). While the prediction is correct for suna+bia and
hikoala+tia in (632a), it does not hold for suuhua+tia, where the noun should
therefore assign a Η tone to the final syllable of the verb, according to its tonal
type. While this may imply that it is a lexicalised compound, the low number of
examples suggests that it would be premature to draw final conclusions.
In a related construction (cf. (632b)), the verb of this compound type is
nominalised with the agentive suffix -era (cf. §6.6.2), which results into a N + N
compound with Ν1 being the object of the nominalised verb N2.

(632) Type 9 compounds

a)
suuhu+tia 'heart' + 'give' 'consult', 'advice'
suna+bia 'afternoon' + 'announce' 'make a sound to announce the
afternoon' (as done by cricket)
k-ukwala+tia 'younger sibling' + 'give' 'give birth to another child'

b)
mami+hioa-era 'beard' + 'throw away'-AG 'razor' (lit. the one that
throws away the beard')
kunu+muku-era 'light' + 'burn-AG' 'lighter' ('the one that
burns the light')
katQa+kuhiounaka 'man' + 'image' + 'pull out'-AG 'camera' (the one that
+ ruku-era takes man's image')

There a few examples for compounds that do not match any of the types listed
in this section. These do not appear to follow productive patterns but some shall
be listed here.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
430 Word formation

(633) Other compounds

enoto mitu nitQahaj N+V+N: 'sun' + 'come out' + 'side' 'East'

enoto su. nit ς, aha j N+V+N: 'sun' + 'get dark' + 'side' 'West'
maua-ri-seihia 'be short' + (RAP) + 'tail' 'be with a short tail'
muku+muku-ne RED + 'burn'-(Suffix) 'burnt field'
suna+mtQae 'afternoon' + 'towards' 'afternoon'

It should also be mentioned that there is a verb root that is combined with the
interrogative pronoun d^a or a demonstrative to form the verbs as shown in
(634).

(634) Verb formation with demonstratives

d^a+toaria 'be how'

ka+toartia 'be like this'
ni+toama 'be like that'

In this chapter, I have described the most prominent aspects of Urarina word
formation, which reflects the complexity typically found in Amazonian
languages. Particularly interesting is the derivation with size suffixes and their
further division into different number types, as this is not found elsewhere in the
grammar. The variety of postures expressed by the wide range of verbal roots is
another prominent feature of Urarina morphology, which is further
complemented by a range of reduplication types with different degrees of
productivity. There also is a number of compounding types, which most
typically result in a noun as the output of the morphological process.
It is understood that the information given in this chapter merely scratches
the surface of word formation in Urarina and leaves ample opportunities for
future studies.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
11 Person inflection classes

Each Urarina verb can occur with up to three different verb forms in the
grammatical categories of person and polarity marking (and, partly, irrealis). I
distinguish these forms in the following way:
- D-form: occurring in finite dependent clauses only
- Α-form: 3sg is -a
- Ε-form: 3sg is -e
While one may feel tempted to describe these types as "conjugation", this term
could be easily misinterpreted. The phenomenon that is known from other
languages, such as the Romance languages, is fundamentally different from
Urarina in that in a language such as Italian, the class of verbs is divided into
subclasses that each take a specific set of inflectional markers. In Urarina, the
distinction is not inherent to the verb, but depends on the grammatical context
in which a verb occurs. That is, every verb can be inflected for any of the three
types if it occurs in the respective context. Similar phenomena are known from
a few other languages only. For example, some Arawak languages such as Bare,
Warekena, and Baniwa mark a focussed subject by a special cross-referencing
prefix in addition to other inflectional categories (cf. Aikhenvald 1995: 152). In
the Chadic languages Hdi, Lele and Mina, there is a morphological distinction
between "pragmatically independent clauses" and "pragmatically dependent
clauses (cf. Frajzyngier 2004). Different paradigms for person marking are also
reported for Menya (Trans New Guinea Phylum; cf. Whitehead 1987: 41) and a
few other Highlands languages of that area. Menya has two sets of subject
cross-reference markers the choice of which depends on the division into final
and medial verbs. However, a threefold distinction into different paradigms, as
found in Urarina may be considered unique.
In Urarina, the forms occurring in independent clause are further
distinguished according to a variety of conditions, as will be investigated in this
chapter. To make a choice with regard to the terminology, I will refer to the
different forms as "(Person) inflection classes" or "(Person) inflection types",
alternatively. This seems appropriate as the most prominent differences between
the forms appear on the person suffixes.
The distribution of the different forms for each person inflection class in
Urarina is subject to a complex set of conditions from different grammatical
areas. There are grammatical contexts that obligatorily require the choice of a
certain form and others, which influence the choice of form, but which
represent only tendencies. In (635), the obligatory conditions are listed in
summary. It should be mentioned that the citation form is also identical with the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
432 Person inflection classes

neutral suffix -a, but it cannot be determined whether this suffix is simply
homophonous or whether polysemy exists between either of these forms. The
occurrence of the Α-form in greetings is marginal, since only two greetings are
known (cf. § 11.3). The most crucial areas of distinction are clause type and the
presence of a focus marker. The occurrence of the D-form is restricted to one
single context only, which is in (finite) dependent clause. The Ε-form is
required in any sentence that contains a focus marker. Further distinctions
between Ε-form and Α-form are observed in different types of interrogatives.

(635) Summary of obligatory contexts that determine the choice of person inflection
type
Context A-form E-form D-form
Citation form obligatory impossible impossible
In greetings obligatory impossible impossible
In polar questions obligatory impossible impossible
With introducer buatia obligatory impossible impossible
In content questions impossible obligatory impossible
FOC marker precedes verb impossible obligatory impossible
In dependent clause impossible impossible obligatory

It is evident that the contexts listed in (635) cover only a small part of the
grammar, as they refer to very specific, mainly syntactic structures. The
restriction of D-forms, which exclusively, but obligatorily occur in dependent
clauses, narrows down the choice of forms in other environments to the two
remaining types, Α-form and Ε-form. Their distinction, however, is subject to
non-obligatory contexts, which involve polarity and syntactic complexity, as
well as style. Details will be investigated in §11.2 and §11.3. The table in (636)
gives an overview of the factors that play a role when obligatory conditions (as
listed in (635)) do not apply.

(636) Summary of optional contexts that influence the choice of person inflection
type

Context A-form E-form

In short utterances typical uncommon
(independent clauses)
With negated verb typical with 3ps possible
In narrative style possible typical
After a dependent clause possible typical
Before a dependent clause possible typical

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Compounding 433

A difficulty in identifying which context eventually leads to the choice of form

is that some optional conditions may overlap. For example, the co-occurrence
with short utterances may be found in narrative texts as well as in other genres,
and negation may be a feature occurring in any sort of text, in any clause type.
The morphological distinctions between the three person inflection types are
discussed in greater detail in §12.2.16 and will not be anticipated here. For
convenience, the paradigms are listed again in (637). In summary, the following
aspects can be stated:
1. The main differences between the person inflection classes regard the person
markers for l s g and 3ps (singular and plural), whereas there are no
distinctions between person markers for all 2ps and lpl forms.
2. Α-form and D-form coincide in the affirmative (realis; cf. (637a)).
3. The negation marker coincides for all Ε-forms and D-forms, but is different
for Α - f o r m s (cf. (637b,d)).
4. The choice of the irrealis marker is subject to complex conditions related to
the person inflection classes (cf. (637c)).

(637) Person inflection paradigms

AFF, A-form AFF, E-form AFF, D-form

lsg -anu -Ü -anu
2sg -i -i -i
3sg -a -e -a
lpl/ex -akaanu -akaanu -akaanu
lpl/du -aka -aka -aka
lpl/in -aka-tge -aka-tge -aka-tQe
2pl -i-tge -i-tge -i-toe
3pl -uru-a -ur-e -uru-a

b) Paradigm for negative marking (realis)

NEG, A-form NEG, E-form NEG, D-form
lsg -a-ü -ene-ü -eni-anu
2sg -e-i -ene-i -ene-i
3sg -e-i /-0-i /-ji -ene-0 -eni-a
lpl/du -a-aka -eni-aka -eni-aka
lpl/in -a-aka-toe -eni-aka-tge -eni-aka-tQe
lpl/ex -a-akaanu -eni-aka anu -eni-akaanu
2pl -e-i-tge -ene-i-tQe -ene-i-toe
3pl -uru-i-0 -ur-ene-0 -ur-eni-a

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
434 Person inflection classes

c) Paradigm for irrealis (affirmative)

AFF, A-form AFF, E-form AFF, D-form
lsg -ri-tQäu -re-ü -ri-anu
2sg -ri-ki -ri-ki -re-i
3sg -ri-tQa -re -ri-a
lpl/du -ri-tQaa -ri-tQaa -ri-aka / -ri-tQaa
lpl/in -ri-tQaa-toe -ri-akatoe -ri-akatoe
lpl/ex -ri-tQaanu -ri-akaanu -ri-akaanu
2pl -ri-M-toe -ri-ki-toe
3pl -uru-ri-tQa -urn-re -uru-ri-a

d) Paradigm for irrealis (negative)

NEG, A-form NEG, E-form NEG, D-form
lsg -ri-pa-ü -ene-re-u -ene-ri-anu
2sg -ri-e-i -ene-re-i -ene-re-i
-ri-a-i
3sg -ri-ßa-a -ene-re -ene-ri-a
lpl/du -ri-a-aka -ene-ri-aka -ene-ri-aka
lpl/in -ri-a-aka-tge -ene-ri-aka-tQe -ene-ri-aka-tQe
1 ρ l/ex -ri-a-akaanu -ene-ri-aka anu -ene-ri-aka anu
2pl -ri-e-i-toe -ene-re-i-tQe -ene-re-i-tQe
3pl -uru-ri-ßa-a -uru-ne-re -uru-ne-ri-a

11.1 D-form

The distribution of the D-form is the one easiest to describe, as it is only found
in one environment. It occurs on the finite verb of a dependent clause, usually a
subordinate clause. Other dependent clauses, such as the participle clause or the
complement clause cannot be inflected with the D-form as these are marked
with non-finite markers, while inflectional features are marked on the main verb
only (cf. §20). The same structure occurs in different subject complement
clauses (cf. §20.2). A subordinate clause is generally marked by the enclitic
=ne, which has a variety of functions and may or may not occur with
conjunctions. However, I have chosen the term D-form (for " D e p e n d e n t "
clause) since it is common practice to omit this marker (while still using the D-
form of the verb). In addition, speakers of the younger generation have
developed a complementation strategy, which is formally identical with
subordination (cf. §20.1.1). In (638), some examples for the use of the D-form
are given. These involve different types of dependent clauses. Examples (638a-
c) show the "regular" (i.e. traditional) contexts in which the D-form occurs. In

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Ε-form 435

these examples, the subordinate marker =ne is present with different functions.
In (63 8d), however, it is absent. This example exhibits the innovative use of the
D-form in a complement clause with hau (for details, cf. §20.1.2.2).

(638) Use of the D-form

a) With subordinate marker and conjunction:
mtoanei hetau=te kat$a lemu-e=lu lomaj,
like.that HRS=FOC man sink-3ps/E=REM PSN

edara ne-t ku-uru-a=ne kujßa

water.people be-PRT go-PL-3ps/D=SUB so.that
'Lomai sank the people like that, so that they would become water people.'

b) With subordinate marker only (temporal function):

unee bua basihjau-a alau=ne
kinkajou bag steal-3ps/D spider.monkey=CND
'When the spider monkey stole the kinkajou's bag' (Title of a narrative)

c) With subordinate marker and complement verb:

nii hau=te raj lureri asae m-a=ne here
that because=FOC POSS house under be-3ps/D=SUB want

nii raj kakunu

that POSS daughter
'Therefore, her daughter wanted her [mother] to stay in her [daughter's] house.'

d) Without subordinate marker, in innovative complement clause:

kaa ajto-hoäu hau auna-i=ßa
this say-lsg/D because hear-2ps=INT
'Have you heard what I said?' (Lit. "because I said this")

It can be concluded that the occurrence of the D-form is straightforward and

predictable. Therefore, the subsequent discussion will focus on the other two
person inflection types.

11.2 E-form

There is one context that strictly requires the choice of the Ε-form: this is the
case when a sentence contains any constituent in focus position. For instance,
the presence of a focus marker in a sentence always results in the marking of the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
436 Person inflection classes

inflected verb with suffixes from the Ε-form. The exact functions of focus
marking are discussed in §19.1. There are three enclitics that mark focus, whose
distribution depends on person and number marking. The enclitics =ne and =rw
mark l s g and 1 pi, respectively, but are subject to free variation (i.e. =ne can
mark lpl and =na can mark lsg, depending on speaker preference and dialectal
variation). The marker =te is used for all other contexts and thus is the most
frequent one. The units that are marked for focus do not strictly coincide with
constituents, but as a tendency, they do so in most examples. Most typically, the
focused items are a N P (in any function), an adverb, or a conjunction phrase (cf.
§19.3). These may include a particle, such as the hearsay evidential hetau in
(639c,d).

(639) Ε-form occurring with focus marker

a) FOC after subject:
aka=te saatono-a ku-re=l dosmildos ke
3sg=FOC end-NTR go-IRR:3ps/E=ASS 2002 VLI
'It [the assignment] will end in 2002.'

b) FOC after object:

lejhu eene=te sw-e enua, raj kakunu rem bedajni-a kuu
one woman=FOC kill-3ps/E tree POSS daughter place visit-NTR go

hana
when
Ά certain tree killed a woman when she went to visit her daughter.'

c) FOC after adverb:

ßoaelu hetau=te enejtQu-htru kuraanaa ne-ι ße=lu alau
earlier HRS=FOC monkey-PL chief be-PRT be:3ps/E=REMspider.monkey
'In ancient days, the spider monkey was chief of the monkeys.'

d) FOC after conjunction phrase:

nii hau hetau=te mi hanulari su-1,
that because HRS=FOC that jaguar kill-PRT

ku sau-hi-ure nii katga-uru

there cut-CNT-3pl/E that man-PL
'Therefore, the people killed the jaguar and cut him into pieces.'

It is not atypical for an entire subordinate clause to be focused, as is shown in

(640). As predicted, the verb of the main clause occurs with a person marker for

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Ε-form 437

the Ε-form, while the verb of the dependent clause is marked with the D-form.

(640) Ε-form after focused subordinate clause

hanotü linternia-a ku-uru-a=na häu=te fotaae
at.night hunt.with.flashlight-NTR go-PL-3ps/D=SUB because=FOC there

hano-ore
spend.night-3pl/E
'As they went to hunt at night, they spend the night there.'

A related context, in which the Ε-form is used obligatorily, is in content

questions. This environment is strongly related to focus, as the use of an
interrogative pronoun in this question type intrinsically highlights one particular
constituent of a clause. In fact, the interrogative is marked with the focus
marker in most cases, as is illustrated in (641). Examples are shown for
different interrogatives, such as d^a 'what/who' (the interpretation depending
on the context, cf. §21.1.1), d$anu ' w h y ' , and d^atoanei ' h o w ' (on the
morphologically complex structure of these interrogatives, cf. §5.9).

(641) Ε-form in content questions with focus marker

a) With 'who':
d$a=te najfle-re kivara-a ku-a=ne=ta
who=FOC be.able-IRR:3ps/E see-NTR go-3ps/D=SUB=FRS
'Who would be able [to find out] if he went?'

b) With 'what':
d^a=te itga-e, nitoanel kanaanaj-uru rela-~i=ta
what=FOC do-3ps/E like.that child-PL teach-PRT=FRS
'What does he do, teaching the children like that?'

c) With 'why':
d^anuMe kaaunei k=aj-te-kure=ta
why=FOC like.this lsg=AUX-INTS-3pl/E=FRS
'Why are they doing [that] like this to me?'

d)
d^atoanei=te kuhwu-ure katQa-uru=ta
how=FOC fish-3pl/E man-PL=FRS
'How did the people fish?'

Apart from the compulsory occurrence of the Ε-form in combination with a

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
438 Person inflection classes

focus marker, it is also obligatory for the Ε-form to occur with an interrogative
even when no focus marker is present. In (642), content questions involving
'who', 'how', and 'where' are used without a focus marker, but the verb in each
example is still inflected with the E-form.

(642) Ε-form in content questions without focus marker

a) With 'who':
d$a kaa ajte
who this say:3ps/E
'Who said this?'

b) With 'how':
d^atoarm-tQa=ra sini batiri=ta
how-only=EMF sleep:3ps/E priest
'How (on earth) does the priest sleep?'

c) With 'where':
d$u ku-re-u
where go-IRR-lsg/E
'Where would I go?'

All other contexts in which the Ε-form is observed are non-obligatory. There
are, however, environments in which this type is strongly preferred as opposed
to the Α-form. One such environment regards sentences in which the regular
constituent order OVA/VS is modified to AOV/SV by a frontshift of the
subject, but without the use of a focus marker. Recall that this type of
construction with content questions clearly requires the use of the Ε-form, as
demonstrated in (642). With affirmative sentences, there is a strong preference
for the Ε-form, as shown in (643). However, the fronting of a constituent
without a focus marker is very rare, and bound to very specific conditions (cf.
§18.3).

(643) Ε-form with subject frontshift, without focus marker

lejhü katQa jiäe kanaanaj siiri

one man already child have:3ps/E
'One man was about to have a child.'

Exceptions from this are mainly found with the negative form, such as shown in
(644). In this example, the subject is in focus position. The use of a focus
marker is excluded by rule as it does not co-occur with a negative clause (cf.
§19.2.4).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Ε-form 439

(644) Subject frontshift without focus marker in negative clause

enene ne-nakauru enamanaa-htru kat$a-uru kwituku-ur-ene

[nowadays be-those. who young.man-PL man-PL] [know-PL-NEG:3ps/E]
'The young people from nowadays do not know it.' [about traditional naming]

A completely different context in which the Ε-form is not obligatory, but

preferred, is related to style/genre. In particular, the Ε-form is predominant in
narratives, which is also confirmed by statistical evidence presented in § 11.4. A
few examples are given below. Note that the distribution of the focus marker is
overwhelmingly widespread in any text type, which makes the occurrence of E-
forms outside this context less common - except in narratives. This genre
exhibits numerous cases in which the Ε-form is chosen without any implication
of focus function, as is shown in (645). All examples given here are taken from
a tale that reports how the monkey taught the woman to give birth. Overall, the
Ε-form can be described as predominant in most narratives (with some variation
based on speaker-specific preferences; cf. (658) vs. (659)).

(645) Ε-form in narratives

a)
ßäe nehano-a=ne haa ke turu-e
already give.birth-3ps/D=SUB PURP VLI arrive-3ps/E
'The time to give birth already arrived.'

b)
mi bajahirii ßäe turu-a u-e
that shortly.after already arrive-NTR come-3ps/E
'Shortly after that, he [the monkey] already arrived.'

c)
mi hau, ßäe mi rihel atii suuh-e
that because already that like nevertheless hide-3ps/E
'Therefore, he hid himself like that nevertheless.'

Another non-obligatory context in which the use of the Ε-form is typical (and
usually more frequent than the Α-form in the same context) is in combination
with complex syntactic structures. Characteristically, what I describe as
"syntactic complexity" here involves a dependent clause that precedes or
follows the main verb. In many cases, this is a subordinate clause (cf. §20.1),
but it may also be represented by a complement clause or by a participle clause.
Some examples are given in (646). I generalise this context as being a "complex
syntactic structure" as opposed to simple sentences that involve a single clause

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
440 Person inflection classes

only. Regarding the example from a narrative in (646a), it cannot be stated for
certain whether the choice of the Ε-form is due to style or a result of complex
syntactic structures. Therefore, a similar example is also given from a non-
narrative text in (646b).

(646) Ε-form with complex syntactic structures: subordinate clause

a) Subordinate clause precedes verb (from flood narrative):

mi huara-a kana+huaaun-era=ne hau, raj kalaui letoa-e
that see-3ps/D our.creator=SUB because POSS son send-3ps/E
'As God saw this, he sent his son.'

b) Subordinate clause precedes verb (from conversation over radio):

ni-a eri-a=ne haanu ke=ne mi ajto-i hau ora
be-3ps/D speak-NTR=SUB PURP VLI=CNDthat say-2ps because hour

ke ii baha-re-it-ni
VLI 2sg ask-IRR-1 sg/E=ASS
'If there is anything to say, I will call you at the time you said.'

The examples in (647) illustrate the use of the Ε-form in a sentence that
contains a participle form. Since these are from a conversation, i.e. a non-
narrative text type, it can be excluded that genre is the relevant factor for the
choice of person inflection type.

(647) Ε-form with complex syntactic structures: participle

a) Participle clause precedes verb (from conversation over radio):

ßaara kurenia tQu hjä kana+huaaun-era baha-he-1 kau m-ü
2pl for CRTN just our.creator ask-CNT-PRT here be-1sg/E
Ί am praying here for you (PL).'

b) Subordinate clause and participle precede (from conversation over radio):

ii orden te-i hanardane, bäako kahe ruku-t
2sg order give-2ps if bank from pull.out-PRT

leotQa bäako kuane tee-re

other bank inside put-IRR:3ps/E
'If you gave your order, he would withdraw [the money] from the bank and put it into
another bank.'

It can be concluded that examples containing an Ε-form are occasionally found

in conversations as well, even though the use of the Α-form is prevalent in non-

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Α-form 441

narrative text types. The importance and ranking of the various conditions is
further discussed in § 11.4.

11.3 A-form

Similar to the Ε-form, the Α-form is obligatory in a few contexts, and optional
(but typical) in others. As a general tendency, the distribution of the Α-form is
less restricted than the Ε-form. While its presence is excluded from any
sentence that contains a focus marker (or which involves structures that imply
focus, such as content questions), the Α-form is found in most other
environments. Its use is less common in narratives, but still possible, especially
under certain conditions that will be investigated in this section.
Two marginal contexts in which only the Α-form occurs are the citation
form of a verb and greetings. Any verb, when cited in isolation (for instance, as
a reply to the question 'what does this verb mean in Urarina?'), occurs with a
final -a (or allomorph), which represents the 3sg form of the Α-form. However,
this form does also correspond to the neutral form, which is used as a non-finite
marker on verbs in serial verb constructions (cf. § 17) and in prohibitives with
Java (cf. §15.2.2). Since there is no way of determining whether the citation
form exhibits polysemy with one or the other, or whether this is in fact
homophony, it cannot be stated with certainty that the suffix in citation forms
really is an instance of the Α-form. However, since speakers intuitively interpret
a verb form such as letoaa (cf. (648)) as 'he sent him', 1 will assume that the
citation form indeed represents an instance of the A-form.

(648) Suffix -a

a) letoa-a 'send'-citation form '(to) send'

b) letoa-a 'send'-3ps/A 'He sent him.'
c) hua letoa-a 'don't'+'send'-NTR 'Don't send him!'

The Urarina do not traditionally greet each other. However, nowadays two
expressions are used, probably as a result of the influence of Spanish.
Specifically, these are used as counterparts of buenos dias ('good morning',
'good day') and buenas tardes ('good afternoon'). The expressions used for
these greetings do not literally correspond with this meaning, as they are
intransitive verbs whose meaning refers to the daylight conditions: hanoa
literally means 'clear up', 'become light', and sua refers to 'become dark (at
dusk)'. These verbs are used with the 3sg/A form and combined with the final
emphasis marker ra.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
442 Person inflection classes

(649) Suffix -a for 3ps/A in greetings

a) hano-a=ra 'clear.up'-3ps/A=EMF (lit. 'it has cleared up') 'Good morning';
'Good day'
b) su-a=ra 'become.dusk'-3ps/A=EMF (lit. 'it has become dark') 'Good
afternoon'; 'Good evening'

There also is a word that is used for 'thank you', but its origin is less clear than
for the two greetings mentioned above. The expression naria, (which is fully
lexicalised, whereas the other two greetings may still be understood in the sense
of light conditions) could have its origin as a composition of the 3ps object
proclitic n= and the verb for 'seek', occurring in the 3ps/A form. If this were the
case, the literal meaning of naria would be 'he has sought it' and would involve
an instance of the Α-form, similar to the other greetings. However, this remains
mere speculation, since no diachronic evidence is available.
Another condition that obligatorily requires the use of the Α-form is in polar
questions, being affirmative or negative. This is illustrated by the examples in
(650) for affirmative polar questions.

(650) Α-form in polar questions

a)
he, tunakuri-a jiaara raj kioairi
hey function-3ps/A 2pl POSS ayahuasca
'Hey, does your ayahuasca work?' (asked by Lomai after offering ayahuasca to her
visitors)

b)
raj hoit^ana ra-uru-a=ne huaasi-tQuru-a=na
POSS blood receive-PL-3ps/D=SUB be.afraid-PL-3ps/A=INT
'Are they afraid to take its blood?' (referring to a bowl of peccary blood)

The Α-form also is obligatory in negative polar questions, as is illustrated in

(651). In addition to inflection by the Α-form, negative questions are
accompanied by the introducer ta and the clause-final enclitic =ne (cf. §14.5,
§21.3).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Α-form 443

(651) Α-form in negative polar questions

a)
ta nahari heri-a ka=d^uhu-he-j=jie
NEGQ liver want-3ps/A this=stand-CNT-NOM=NEGF
'Does this one who is standing here not want any liver?'

b)
ta u-ri-tQa=i ii komasaj=jie
NEGQ come-IRR-3ps/A=ASS 2sg wife=NEGF
'Won't your wife come?'

Another context that requires the use of the Α-form is in negative clauses, when
they co-occur with the negative introducer hvatia 'don't' (cf. §14.4.1). In this
case, the Ε-form cannot be used.

(652) Α-form in negative clause with faoatia

hwatia enanihja kuane amu-uru-i=lu
not canoe inside walk-PL-NEG:3ps/A=REM
'They did not go in canoes.'

All other contexts in which the Α-form occurs are non-obligatory, but represent
tendencies. There are two contexts where this person inflection type is
extremely frequent: in "short" utterances and when the main verb is marked for
3ps negative. A "short" utterance can be defined as one that consists of an
inflected verb only or an inflected verb with maximally one NP or adverbial. In
these cases the Α-form is overwhelmingly typical; only a few examples for the
Ε-form are attested in this context. As the examples in (653) illustrate, this does
not only involve utterances that may be understood as answers to questions (cf.
(653a)), but also short observative statements as the ones in (653b), that
occurred in a picture description. Even in narratives, which otherwise prefer the
Ε-form, the final sentence that indicates the end of the story typically is a short
statement in which the Α-form is used (cf. (653c)).

(653) Α-form in short utterances

a) (Conversation over radio):
jiae auna-käu / huatia auna-7a-ü
already hear-lsg/A not hear-NEG-lsg/A
Ί understand (already).' / Ί did not understand.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
444 Person inflection classes

b) (Picture description):
iani-a usi
burn-3ps/A fire
Ά fire is burning.'

c) (End of narrative):
ßäe sa-a
already end-3ps/A
'It is finished.'

Another typical context in which the Α-form is preferred is in verbs that are
marked for 3ps negative. The suffix -/, which occurs in this case, is much more
common than the corresponding Ε-form -eve. Interestingly, this applies to all
genres, including narratives, where a preference for the Ε-form would be
expected. The examples in (654) give evidence for this with sentences from
different text types. Both examples are taken from a narrative about the life of
the ancient people. While the use of the Ε-form is prevalent elsewhere in the
text, the majority of verbs that are inflected for 3ps negative in this text occur in
the Α-form. N o n e of the sentences can be characterised as "short", and both
examples contain a subordinate clause, which was mentioned as a typical
trigger for the Ε-form. Also note that the negative introducer fauatia is not
involved here.

(654) Α-form with 3ps negative in narratives

a)
nii hau heriane enanihja itQa-kuru-i=lu, atane
that because probably canoe make-PL-NEG:3ps/A=REM land

tQoae amu-uru-a=na hau

on walk-PL-3ps/D-SUB because
'Therefore, they probably did not make canoes, because they walked on land.'

b)
nii baja enane ne-nakauru nijej nitoane-j
that after nowadays be-those.who not.at.all be.like.that-NOM

itQa-na najm-ji, kioituku-eni-a=ne hau

do-INF be.able.-NEG-3ps/A know-NEG-3ps/D=SUB because
'After that, the people who live nowadays cannot do something like that, because they
don't know it.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 445

The presence of the Α-form in this context demonstrates that some factors
which determine the choice of person inflection type are ranked higher than
others. Apparently, the criterion "3ps negative" is more significant than the
criterion "narrative style". It must be emphasised that this applies only to non-
obligatory contexts. In environments that obligatorily require a particular form,
such as a sentence involving a focus marker, these tendencies do not apply: the
Α-form cannot replace the Ε-form in this case.

11.4 Discussion

In the previous sections I have shown that there are certain environments that
obligatorily require one of the three possible person inflection types. This is a
convenient aspect for linguistic description, as the occurrence of forms is
predictable to a reliable extent. However, the other factors that influence the
choice of form represent much less reliable criteria, as they are mere tendencies
or preferences. In this section, these preferences will be challenged by counter
examples for each condition, which illustrates the fact that they are non-
obligatory. These may be quite atypical in some cases and "possible" in others.
Aiming at a solution, I will then approach the probability of choice for certain
forms with statistical data.
The example in (655a) shows the possible occurrence of the Α-form in
narratives. In the example, the main verb follows a participle clause. The use of
the Α-form is atypical, as complex syntactic structures, especially in narratives,
usually coincide with the Ε-form. This illustrates the potential for variation
between the two forms in non-obligatory contexts. In (655b), the occurrence of
the Α-form in the context of complex syntactic structures is further exemplified,
here with the finite verb occurring before a dependent clause. However, other
aspects, such as the fact that (655b) is from a conversation, might be relevant
for the choice of form in this example. This confirms that the choice of the E-
form in complex syntactic environments, such as illustrated in (647), is a mere
preference that does not exclude variation.

(655) Occurrence of the Α-form in atypical contexts

a) In narratives, after participle clauses:

na-Ί hetau mi bahe-ϊ, nii itanitQa kahe hetau ruku-a
say-PRT HRS that dig-PRT that mud from HRS pull.out-3ps/A
'Saying so, he dug and pulled him out of that mud.' (from flood narrative)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
446 Person inflection classes

b) In conversation, with complex syntactic structures (before subordinate clause):

kxuatia raj n~itg.afioa ruku-erati-jia-u, ku-naa
not POSS 3ps=rifle repair-CAU2-NEG-lsg/A go-INF

najpe-ni-anu mekamho hau

be.able-NEG-lsg/D mechanic because
Ί did not have his gun repaired because I could not go to the mechanic.'

c) In conversation, with complex syntactic structures (before subordinate clause):

alaa ama-a ke ne-kureti-a ikito-o=ne kujjia
aguqje take-3ps/A VLI ITR-buy-3ps/D PLN-LOOSUB so.that
'They have taken aguaje [fruit] with them in order to sell it in Iquitos.'

Similarly, the Ε-form can be found in contexts that do not match the preferences
as indicated in (636). The sentences in (656) illustrate uncommon occurrences
of the Ε-form. (656a) shows the use of this form in short utterances, as part of a
narrative. In (656b), the Ε-form occurs in a conversation, which is in opposition
to the style used in narratives, and in (656c), a 3ps negative form is observed
with the Ε-form. This particular example illustrates that the preference "A-form
with 3ps negated verbs" is violable.

(656) Occurrence of the Ε-form in atypical contexts

a) In short utterances:
le=sahi ke hoara-ü
one-time VLI see-lsg/E
'Once I experienced it.' [from narrative: hunting story]

b) In non-narrative genres:
eresi nii ora ke tQu hjä ajto-re-u=ni
tomorrow that hour VLI CRTN just say-IRR-lsg/E=ASS
Ί will tell tomorrow just at that time.' [from conversation]

c) With 3ps negative:

na-a häu hetau, atii ku hjam-na here-kur-ene
say-3ps/D because HRS nevertheless there let-INF want-PL-NEG:3ps/E

hjam-na here-lanaala ku ajpa raasa-ure

let-INF want-PRV there with dance-3pl/E
'As she said so, they still did not want to let her [go].' [from flood narrative]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 447

The examples shown above demonstrate that the preferences stated in (636) are
subject to exceptions. However, in order to further support the actual weight of
those preferences, I represent these preferences in terms of statistical
distribution. 54 Verb forms from ten selected texts of different speakers and
genres were counted in order to determine the relevance of the suggested
criteria that influence the choice of person inflection form. The overview in
(657) provides information on the following aspects:
1. 283 (=58%) of all inflected forms occurred in predictable contexts. Most
typically, this was through the presence of a focus marker, in dependent
clause or in questions and imperatives. The rest of 205 verbs were subject to
choice by preference, according to the optional conditions discussed above. 55
This also includes instances of 2ps or lpl occurrence, as these forms are not
distinguished in the different person inflection classes.
2. In narratives told by three different speakers, there is a preference for E-
forms (33 instances) as opposed to A-forms (22 examples). Details regarding
the nature of the contexts in which the Α-forms occurred will be investigated
below.
3. Inflected verbs used in conversation exhibited an overwhelming prevalence
for Α-forms, which is in agreement with the predictions made above.
4. Picture descriptions do not belong to the narrative genre, even though the
nature of speech in this case involves 'telling', but not from the speaker's
own experience (or through hearsay). This is manifested by the distribution
of person inflection types in Descriptions 1-6, told by two different speakers
(details discussed below). In these texts, only one instance of an Ε-form was
recorded, as opposed to 52 occurrences of Α-forms. Description 7 was
subject to other influences, as will be discussed below.

54
Note that these figures are based on a small selection of representative texts only.
However, it is plausible to assume that 3 total of 488 inflected verbs constitute a
sound basis in order to illustrate the general tendencies.
55
The addition of forms does not always match the total ("62+4"), as there were four
cases in which an Α-form was incorrectly used in a context that required the E-form,
i.e. one speaker in a conversation used the Α-form despite the presence of a focus
marker in the same sentence. See (1) for details.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
448 Person inflection classes

(657) Summary of inflected forms in selected texts

Text / Genre, Speaker Predictable/Total Rest A-form E-form

T i l / Narrative, HN 8/9 1 1 0
T12/Narrative, HN 7/8 1 1 0
T13 / History/Narrative, JN 26/46 20 12 8
T16 / Narrative, MA 10/43 33 8 25

Τ 34 / Conversation 108/179 71 62+4 9

T35A+B /Description 1-5, HN 52/78 26 25 1

T36A+B / Description 1-5, JN 9/34 25 25 0
T35C / Description 6, HN 8/9 1 1 0
T36C / Description 6, JN 11/12 1 1 0

T35D / Description 7 (story), 31/49 18 15 3

HN
T36D / Description 7 (story), 13/21 8 0 8
JN

Total: 283/488 205 155 54

In the following, each selected text is investigated in terms of distribution

between A- and Ε-forms. For this, three parameters were observed: utterance
length, syntactic complexity of the sentence, and negation of 3ps forms.
"Predictable" forms (as explained above) were not analysed any further.
- "Short" is defined as consisting of a verb accompanied by one constituent
only, a verb with an adverbial, or a verb by itself.
- The criterion "Complex" refers to syntactic complexity and implies that the
sentence contains a dependent clause that precedes or follows the inflected
verb. I also include participle constructions and complement clauses as part
of this criterion.
- Since some criteria may occur in combined form (e.g. "short" and
"negative"), the sum of occurrences in "Rest" forms does not match the total
of "Rest".

Text 13, surprisingly contained more instances of the Α-form as opposed to the
Ε-form, which would have been expected to occur more frequently, since this
text represents the narrative genre. However, all occurrences of the Α-form are
found in preferred contexts for the Α-form, most frequently with 3ps negatives.
The majority of those found with negatives also co-occurred with complex

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 449

syntactic structures, which explains the high figures here. However, the general
style of the narrative is exhibited by the predominant use of the Ε-form in other
contexts that are not subject to any preference criteria.

(658) Distribution of forms in T13 - Narrative (Speaker: JN)

Predictable/Total: 26/46
Rest: 20
Person inflection type chosen on Rest:
All Short NEG:3ps Complex Other
A-forms: 12 0 9 9 0
E-forms: 8 0 0 8 0

Texts 11 and 12 (cf. (657)), which are narratives told by another speaker, are
not conclusive, as in each text, all verb forms but one occurred in predictable
contexts, i.e. with a focus marker or in dependent clause. In both cases, the story
was finished by a "short" statement containing the Α-form to indicate the end.
Text 16, again told by another speaker, contained less predictable contexts
and is a rich source of contexts for which a preference for one of the two person
inflection types had to be made. The majority of verb forms occurred in the E-
form, most of which have complex syntactic structures. It must be noted that the
presence of subordinate clauses and other dependent clause types are a typical
feature for narratives in general; thus, the predominance of Ε-forms is not
surprising. Α-forms, however, did not occur in this specific context. Most of the
Α-forms attested in this text occurred in short utterances - typically representing
citations of speaker interaction, which are part of the story.

(659) Distribution of forms in T16 - Narrative (Speaker: MA)

Pred ictab le/Total: 10/43

Rest: 33
Person inflection type chosen on Rest:
All Short NEG:3ps Complex Other
A-forms: 8 7 0 1 0
E-forms: 25 1 1 24 0

A different type of genre is represented by a conversation between two speakers

over short-wave radio (who were untrained in this and simply conducted the
conversation in the same way as any other conversation, except for a number of
repetitions). One peculiarity was that one of the speakers, towards the end of the
conversation, used the Α-form in combination with a focus marker. However,
this was described as incorrect in the subsequent analysis of the text. A large

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
450 Person inflection classes

number of verb forms occurred in short statements, such as Ί will do it' and Ί
did not understand you'. In addition to the fact that the Α-form is typical in
conversations, it also occurred frequently with complex syntactic environments.
This aspect indicates that non-narrative style is a more relevant feature than the
preference for the Ε-form with complex structures. However, this was the
context in which most occurrences of this form were attested. One may
conclude from this, if Ε-forms occur in non-narrative texts, they preferably do
so within a complex syntactic environment.

(660) Distribution of forms in T34 (conversation)

Predictable/Total: 108/179
Α-form despite
4
FOC marker
Rest: 71
Person inf ection type chosen on Rest:
All Short NEG:3ps Complex Other
A-forms: 62 42 16 15 0
E-forms: 9 0 0 6 3

Another set of texts is represented by picture descriptions. The data from two
different speakers was selected for statistical analysis here. The texts
T35A+B/T36A+B each involve five little texts based on the description of five
images that showed situations of daily life for the Urarina, which included
hunting, cooking, and house building. The verb count exhibits an overwhelming
predominance of the Α-form, with only one Ε-form occurring in 51 inflected
verbs (in non-predictable contexts) altogether.

(661) Distribution of forms in T35A+B, T36A+B (picture descriptions)

a) T35A+B (Speaker YIN)
Predictable/Total: 52/78
Rest: 26
Person inflection type chosen on Rest:
All Short NEG:3ps Complex Other
A-forms: 25 10 0 3 12
E-forms: 1 1 0 0 0

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 451

b) T36A+B (Speaker JN)

Predictable/Total: 9/34
Rest: 25
Person inflection type chosen on Rest:
All Short NEG:3ps Complex Other
A-forms: 25 0 0 0 25
E-forms: 0 0 0 0 0

As shown in (661a), both speakers have a strong preference for short utterances
that described the various actions depicted in the images. In many cases, the
utterance involved a verb preceded by an object or followed by a subject, i.e.
with only one NP being present. In other examples, the sentences contained
more than one NP, but were still significantly shorter than typical sentences
occurring in narratives (cf. (661b)). However, having defined the criterion
"short" as "Verb plus one constituent or adverbial", these examples are
classified as "Other". For one speaker (JN), this applied to all given examples,
as illustrated in (662b,c).

(662) Typical examples from Picture descriptions

a) Utterances defined as "Short" (here V-S):

sauki-a isi+lahe
five-3ps/A fire+piece
'There are fives pieces of firewood.' (from T35A+B, speaker HN)

b) Utterances defined as "Other" (here O-V-A):

obana su-a katg,a
cl.peccary kill-3ps/A man
'The man has killed a collared peccary.' (from T36A+B, speaker JN)

c) Utterances defined as "Other" (here Adv-O-V):

ßäe obana ru-a
already cl.peccary find-3ps/A
'He has already found a collared peccary.' (from T35A+B, speaker HN)

Texts 35C and 36C (picture description 6 by two speakers) differed from the
other images in that the task was to describe geometrical shapes. In this context,
both speakers used a focus marker in almost each example, which resulted in
the use of the Ε-form and thus is not conclusive for the analysis of preferred
forms in non-obligatory contexts. However, texts T35D and T36D (picture
description 7) exhibited interesting differences between the two speakers, as the
task was to describe about a dozen images that were part of a biblical parable

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
452 Person inflection classes

(involving a Samaritan who comes to the rescue of a man who was robbed).
While one speaker (HN) proceeded in a similar way as with the previous
images, the other (JN), switched to narrative style, as he was very familiar with
the story. This also became evident by the fact that he mentioned aspects in his
description that were not depicted in the images (e.g. names). Different from the
previous picture description, his speech involved more complex syntactic
constructions, which corresponds to the complexity of the given images. This
probably contributed to the ample use of the Ε-form, while no occurrences of
the Α-form were recorded. The first speaker (cf. (663a)) also used more
complex structures, but tried to stick to the task of description, which resulted in
the use of the Α-form even in most syntactically complex environments.

(663) Distribution of forms in Picture description 7

a) T35D (Speaker HN}

Predictable/Total: 31/49
Rest: 18
Person inflection type chosen on Rest:
All Short NEG:3ps Complex Other
A-forms: 15 3 0 7 5
E-forms: 3 0 0 2 1

b) T36D (Speaker JN)

Predictable/Total: 13/21
Rest: 8
Person inflection type chosen on Rest:
All Short NEG:3ps Complex Other
A-forms: 0 0 0 0 0
E-forms: 8 0 0 7 1

In summary, the following observations can be made about the distribution of

A/E-form in non-obligatory contexts:
[1] The Ε-form is the preferred person inflection type used in narratives. One
typical feature of narratives is also the elaborate use of complex syntactic
structures, which in turn coincides with the use of the E-form.
[2] With verbs that are inflected for 3ps negative, the Α-form is preferred. This
preference is ranked higher than the ones mentioned in [1], as it is also
widespread in narratives and with complex syntactic structures.
[3] Short statements typically attract the Α-form. This is also observed in
narratives, where short utterances hardly occur. If they do, they normally
are citations of story characters interacting with each other. The use of the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Conclusion 453

Α-form here could be interpreted as an instance of "normal" style

embedded into a narrative.
[4] In daily life conversations between speakers, the Α-form is strongly
preferred (if not excluded by the presence of focus markers). This
preference also applies to sentences that involve dependent clauses or
similar complex syntactic structures. If an Ε-form is used at all in
conversations, it is likely to occur in this context.
[5] In characterisations of simple events, such as picture descriptions, the A-
form is preferred, including complex syntactic environments. This is
especially confirmed by Description 7 (cf. (663)), where one speaker
makes extensive use of this person inflection type, even in complex
structures, whereas another speaker switches to narrative style, which
results in the use of the Ε-form throughout.

From these observations, it can be stated that the Ε-form is the functionally
more marked person inflection type, as it only occurs in one specific genre,
which overlaps with the other typical environment where it is used, in complex
syntactic structures. However, its preference is easily overridden by other
factors that imply a preference for the Α-form, such as 3ps negation, utterance
length, and embedding of non-narrative style. The Α-form is functionally less
marked and can therefore be characterised as the default form.

11.5 Conclusion

Each Urarina verb can occur in three different forms that may be called person
inflection classes or types. The morphological differences between these classes
are manifested by separate sets of person markers for each class, plus disparities
in the marking of polarity and mood. The functional distinction between the
three types is divided into obligatory conditions that pre-determine the choice of
a certain form and preferences or tendencies that apply when no obligatory
context is specified.
For each person inflection type, there is at least one obligatory context that
requires the use of a certain form, as is depicted in (664). In dependent clauses,
the finite verb is inflected with suffixes of the D-form. When a sentence
contains a focus marker, the main verb occurs in the Ε-form; this also applies to
content questions, with or without focus marker, since the presence of an
interrogative pronoun implies focus. In contrast, verbs in polar questions are
marked with the Α-form. The options that remain could be called "Declarative".
In one special case, i.e. when the introducer kwatia determines the negative
status of the clause, the Α-form is obligatory. All other contexts are subject to
optional preferences.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
454 Person inflection classes

Given this distribution of obligatory criteria, one could claim that the three
different types actually represent different clause types: similar to the
declarative, as opposed to imperatives and questions, one may characterise
dependent clauses as a clause type that require special inflection, as is common
in other languages (e.g. if compared to the subjunctive mood in Spanish or
French). As an adjustment to the previous characterisation of sentences that
contain a focus marker, one could now describe this kind of construction as a
separate clause type. Representing a further type of clause, the Ε-form is used
on the main verb whenever a focus marker is present at the beginning of the
sentence. Other clauses, including dependent clauses or quotations, may form
part of the sentence.

(664) Obligatory criteria for the choice of person inflection class

Utterance

Dependent Independent
clause clause
I
D
FOC Non-FOC (no overt focus marker)

Question Declarative

Content Polar huatia Others

The main criterion for the choice of form in "Other" declarative clauses seems
to be genre. The Ε-form, i.e. the same form that marks focus constructions, is
preferred in narratives only, which at the same time represent a natural
environment for syntactically complex sentences - another criterion that favours
the occurrence of the Ε-form. The Ε-form may also occur in other genres, but
this is comparatively rare and typically related to a co-occurring complex
syntactic structure. The Α-form, in contrast, functions as a default form for
other text types. In addition, it also occurs in narratives in specific contexts.
These involve short utterances (with up to one constituent being present, apart
from the verb) and verbs that are inflected for 3ps negative. The reason for this
last preference is obscure. On the one hand, it is true that the Α-form -i is
shorter than the corresponding Ε-form -ene, which further contributes to the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Conclusion 455

tendency of having "short" utterances. On the other, the polysynthetic structure

of Urarina makes it quite evident that this is not a language that makes every
effort to shorten phonological or syntactic structures. In contrast to what one
might expect from a form that frequently overrides other criteria such as
"narrative style" or "syntactic complexity", there is no overgeneralisation of the
preference for the Α-form in this context with focus constructions, i.e. the A-
form of 3ps negatives is never used in constructions involving a focus marker.
The figure in (665) again illustrates the various preferences in summary.
Bold capital letters stand for the preferred choice, whereas the forms in brackets
are possible, but less likely to be used in the respective context.

(665) Optional preferences for the choice of person inflection class

Declarative

Default Narrative

A, (E)
Short Long NEG:3ps

Α, (Ε) E, (A) A, (E)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
12 Verbal morphology

Urarina is a predominantly suffixing language with 17 suffixal and 7 enclitic

positions, also called position classes or slots here. 56 However, person is the
only feature that is obligatorily marked on every finite verb. Other categories
only surface when they have to be expressed. For instance, polarity is only
marked for the negative; number is only marked in the plural form, etc. In
contrast, the potential for the morphological realisation of grammatical
categories is considerable and, given the circumstances, a large number of
features can be marked in one verb. In (666), the maximal structure of a verb is
projected, including all suffixal and enclitic slots.

(666) Urarina verb structure

a) Suffix slots
Slot no. Function
1 Causative-1 (intr. verbs)
2 Causative-2
3 Possibility, with negation
4 Aspect: continuous (order may vary)
5 Impersonal passive; with negation
6 Aspect: habitual (order may vary)
7 Distributive (order may vary);
Plural object (order may vary)
8 Velocity (order may vary)
9 Diminutive/Counterexpectation (order may vary)
10 Aspect: completive
11 In-law talk
12 Plural-1 (order may vary)
13 Probability
14 Irrealis
15 Negation
16 Person
17 Plural-2

56
One may also say that there are only five enclitic slots, since the last two contain
clause clitics. However, as their most typical occurrence is on a verb, they will be
discussed in this chapter.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Conclusion 457

b) Enclitic slots
18 Politeness
19 Assertive (Future)
20 Evidentiality: Witness
21 Evidentiality: Reportative (order may vary);
Remoteness (past) (order may vary)
22 Reassurance
23 Interrogative; Negative question
24 Attitude/Emotion: Emphasis, Frustrative, Warning, Fear;
Rhetorical question

Natural speech does not provide examples in which all 24 suffix/clitic slots
listed in (666) would occur in a row; however, apart from some exceptions,
their combination would be possible at least theoretically. On average, only
between two and four positions are filled. In (667a), an example for a typical
verb structure is given. The second example (667b), where the root is followed
by three suffixes and three enclitics, can be regarded longer than standard
length.

(667) Examples of typical verb structure

a)
itQa-si-uru-kaj-a
do-CPL-PL-PRB-3ps/A
'They have probably finished doing it.'

b)
su-naha-ana-i=tQe=he=ta
kill-PLO-ILT-NEG:3ps/A=PLT=REP=FRS
'It is said that he (my father-in-law) unfortunately did not kill them.' [Man talking to
woman]

It should also be remarked that it is difficult, if not impossible to formally

distinguish inflection from derivation in Urarina. The criteria which might play
a role for this distinction involve productivity and lexicalisation, or whether a
change of word class applies, whether a function has more grammatical or
lexical character, i.e. whether the addition of a form changes the meaning of a
verb. While Payne (1990:130-135) argues that the distinction between inflection
and derivation may be viewed as a continuum, she describes a number of
distinctive features between derivation and inflection, as summarised here:

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
458 Verbal morphology

Derivational morphology Inflectional morphology

1 - is not correlated with - correlates with something
something elsewhere in the elsewhere in the syntactic
syntactic structure structure
2 - is typically non-productive - is productive
3 - has non-predictable meaning - has predictable meaning
4 - does not participate in - participates in paradigms of
paradigms of oppositions oppositions
5 - results in new lexical items - does not result in new lexical
items
6 - is word class changing - does not change class
7 - occurs towards the root - occurs towards the edges of
words
8 - results in (substantial) change
in meaning

For Urarina verbal suffixes, most of these criteria cannot be clearly applied. For
instance, the velocity suffix -uri (slot 8) clearly qualifies as derivational
according to criteria 1 and 4, but also matches the criteria for inflection 2 and 6.
Conversely, velocity could be a grammatical category of Urarina. In fact, the
suffix occurs with all types of verbs and speakers do not intuitively recognise it
as a "different" verb as opposed to the form without the velocity suffix. For the
other criteria, any attempt to unmistakably separate derivation and inflection is
relative, as is the case for many polysynthetic languages. The causative suffix -a
could be regarded as "derivational" in that it results in new lexical items (e.g.
'come' + CAU results into 'bring'). However, its function being clearly of a
grammatical nature suggests that it could be described as "inflectional".
Thus, as a precise distinction between derivation and inflection is not always
practicable in Urarina, the division as suggested in this grammar ought not to be
taken as dogmatic. All morphemes investigated in this chapter have the
following properties in common:
1. They are not word class changing.
2. They are very productive, i.e. they can be combined with any verb or with
almost all members of a certain verb class. (For instance, it is obvious that a
plural object marker only occurs with transitive verbs.)
3. They are easily compatible with suffixes of other grammatical categories; i.e.
it is possible to assign a morpheme to a specific slot. In particular, all
morphemes discussed here can be combined with a person marking suffix.

Note that I have not included here size suffixes and other morphemes branded
"derivational" as discussed in §10, as they are not easily compatible with

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Preßxal positions 459

suffixes of other slots. Furthermore, their compatibility is highly restricted, as

they only occur with one specific verb class.

12.1 Prefixal positions

Urarina has only one position that occurs before the verb root. This slot can be
occupied by one of the two following forms, which are mutually exclusive:
a) The general intransitivising prefix ne-, which must be considered derivational
(cf. §9.3), or
b) An object proclitic for lsg or 2sg (3ps is unmarked; other forms are marked
through free pronouns where applicable - cf. §5.5).

Since neither of the two morphemes above is part of the inflectional system, it
can be stated that Urarina has a purely suffixing inflectional morphology. Note
that there are a few rare prefixes in the language, but they do not mark verbal
inflection. These include the associative marker hi- on nouns, (which appears to
be highly lexicalised, cf. §7.3.3), and, similarly, the even less productive prefix
b-, also found with a few nouns only (cf. §7.3.2).

12.2 Suffixal positions

12.2.1 Suffixal slot 1: Causative-1

The position closest to the root can be filled by the causative suffix -a, which
occurs with intransitive verbs only. It occupies a separate slot as is evident from
the fact that it can be combined with the other causative marker (cf. §12.2.2).
The details of the exact function and allomorphy for both causative suffixes are
discussed in §16.4.

(668) Causative with suffix -a

a)
kam kotnasaj u-a-anu
lsg wife come-CAUl-lsg/A
Ί have brought my wife.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
460 Verbal morphology

b)
tgäe kanaanaj-uru eno-a-e=lu
also child-PL enter-CAUl-3ps/E=REM
'He also made the children enter.'

12.2.2 Suffixal slot 2: Causative-2

The suffix -erate occurs with intransitive as well as with transitive verbs. The
examples in (669) show the use of -erate as attached to the root, which is its
most common position. It can be followed by suffixes of all other slots. Below,
its occurrence is illustrated before suffixes such as for habitual aspect and plural
object. 57

(669) Causative with suffix -erate

a)
akauru ke itulere itQa-rate
3pl VLI all.kinds do-CAU2:3ps/E
'He made them do all kinds of things.'

b)
mi hau=te kujßadera-Ί ari-t$ate
that because=FOC be.worried-PRT seek-CAU2:3ps
'Therefore, he was worried and had her sought.'

The examples in (670) illustrate that both causative suffixes can occur in
combination (and therefore occupy separate slots). The first example represents
a causativisation of the intransitive verb for 'enter', whose valency is increased
through the addition of -a. Then, in turn, it is suffixed with -erate, which results
in the meaning 'cause to make enter'. In (670b), the -a causative is attached to
the root / k u / 'drink', resulting in the irregular form ku-hvaa 'give to drink',
which in turn is followed by -erate. The semantic content of this form is a
double causative 'cause to give to drink'.

57
There is one example in the NT where the plural object marker precedes the
causative: te-naha-erate ('give'-PLO-CAU2:3ps/E) 'he had them delivered' (NT:
Acts 1:2). This is due to the variability in position of the PLO suffix.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 461

(670) Occurrence of both causative forms as a sequence

a)
fwedoro eno-a-erate karesero kuane herodesi
PSN enter-CAU 1 -CAU2:3ps/E prison inside PSN
'Herod had Peter thrown into prison.' [NT: Acts 12:4]

b)
bahaae+bahaae-rattru ere+ere-nakauru ku-kwaa-erate-ma
RED+ask-those.who RED-speak-those.who drink-CAUl-CAU2-INF

najpe-re=l kana+kioaaun-era
be.able-IRR:3ps/E=ASS our.creator
'God will be able to have [someone] give drink to those who cry to him day and night.'
[NT: Luke 18:7]

12.2.3 Suffixal slot 3: Impossibility

The suffix -naka is comparatively rare in occurrence, which is partly based on

dialect differences and partly on grammatical restrictions. There only is a
handful of examples in the database. Principally -naka can be described as an
intransitivising suffix to imply the degree of possibility of difficulty. As it only
occurs with negative polarity in the Espejo dialect, its function could be referred
to as "impossibility". An example for this is the form su-naka-i, composed of
the root 'kill' and -naka "-able" plus the 3ps/A negative suffix, to mean "it is
not kill-able" or 'it is hard to kill' (e.g. about a dangerous animal). According to
some Espejo speakers, the form can be used in the affirmative ("it is killable")
in the Chambira dialect, whereas Espejo speakers prefer a periphrastic form
with najnia 'be able' in this context. In the negative, it can be inflected for any
person and it can be combined with tense, aspect, or other grammatical
categories.

(671) Occurrence of impossibility suffix -naka

a)
su-naka-reheto-oru-i
kill-PSB-HAB 1 -PL-NEG:3ps/A
'They used to be hard to kill.' [Lit. "not easily killable"]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
462 Verbal morphology

b)
sw-erate-mka-i
kill-CAU2-PSB-NEG:3ps/A

'It is difficult to find someone who would kill it.' [Lit. "not easily make-kill-able"]

12.2.4 Suffixal slot 4: Aspect-1: Continuous

Perfective is the unmarked aspect in Urarina. This regards formal marking as
well as its function: a verb unmarked for aspect usually refers to the perfective.
There also is a suffix used to mark completive aspect (cf. §12.2.10).
Imperfective aspect is split into various forms: the continuous/durative form
-ahe (and allomorphs), the past habitual -reheto, and the habitual form
-nahaaoka. The latter two are discussed in §12.2.6.
The morpheme -ahe refers to a continuous or durative action and applies to
any tense. For instance, it can be used with past reference (as indicated by the
remoteness enclitic =lu; cf. §12.3.4) or with the future tense (indicated by the
irrealis form, cf. §12.2.14). It is also very common in combinations with the
participle marker -z, which results in the form -aheΐ (or allomorphs). Most
commonly, the continuous marker occurs adjacent to the root and is followed by
a person suffix, or by the participle marker, as mentioned above. The examples
in (672) show some cases in which other suffixes co-occur with -ahe. For
instance, it can be followed by suffixes referring to negation or tense and there
are no apparent restrictions.
(672) Occurrence of continuous suffix -ahe
a) With negation:
nete enene päe mrianaj itQa-he-ur-ene
but nowadays already all.that do-CNT-PL-NEG:3ps/E
'But nowadays they don't do all that any more.'

b) With future tense:

ii auna-he-ri-tQäu=ni, na-e hetau mi kati
2sg hear-CNT-IRR-1 sg/A=ASS say-3ps/E HRS that black.monkey
Ί will be listening to you, said the black monkey.'

The suffix can also occur with the velocity suffix -uri, as illustrated in (673).
The combination of this marker with the continuous form in (673) may be
surprising at first sight, since durativity and rapidity do not appear to be
semantically compatible. The form refers to a situation where the protagonist
was still listening and suddenly died.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 463

(673) Continuous aspect and velocity marker

hi auna-he-uri-a=ne baja tukuhzua-t taka-l, ßäe hi

there hear-CNT-RAP-3ps/D=SUB after be.lying-PRT meet-PRT already there

u-e
die-3ps/E
'Hearing [these words], he immediately fell down and died.' [Lit. 'after he quickly
heard it'; NT: Acts 5:5]

As indicated above, -ahe has a variety of allomorphs. While -ahe may be

regarded as the underlying form, there are some variations: as a general
phonological rule, the suffix is realised as -ahi before /a/ (cf. §2.8.4). Also, any
insertion of the glides [j] and [w] is the regular result of vowel copying from the
left (cf. 2.8.3). The suffix is realised as -he after verb roots that end in /a/ (such
as hvara 'see' and itga 'do') and after long vowels and vowel sequences.
With a few verbs, the continuous suffix is realised as -eka. These include the
verbs haa 'make' (CNT haa-eka), raa 'receive' (CNT raa-eka), and amaa 'take
along' (CNT amaa-eka) - which all end in /a/. However, as other verbs with a
similar phonological structure take the allomorph -ahe, the three verbs must be
regarded exceptions.
As another peculiarity, the continuous participle form of some verbs is not
-ahe-ι, as expected, but -aha-ι. This form consistently occurs after verbs whose
root ends in /i/. The surfacing palatal glide in the form [hjai] is the result of
vowel copying (cf. §2.8.3). Evidently, the form [hjei], which would be expected
is not possible with the verbs listed in (674), but this sequence can occur
elsewhere, such as in rihihjel 'being like', or sajhjel 'although'.

(674) Allomorph -hja-1 for continuous participle

tom-hja-Ί ' sound'-CNT-PRT

kuti-hja-i 'call'-CNT-PRT
kufioi-hja-l 'fell'-CNT-PRT
tohwi-hja-ϊ 'circle'-CNT-PRT

The continuous form in combination with the participle is also rather frequent in
adverbials. Interestingly, the durative meaning of the suffix tends to be
lexicalised in this context. For instance, the word rauto-hwe-ι ('be calm'-CNT-
PRT) 'calm(ly)' always contains a continuous form even though no reference to
aspect may apply. In fact, it is not uncommon to find the same verb with and
without -ahe within the same sentence, with the same meaning. This is
exemplified in (675), where the priest's assistant provides the people with
fishing equipment: in this example, 'give' is marked for continuous aspect once,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
464 Verbal morphology

while the other occurrence remains unmarked. The "optionality" of the suffix
depends on the context as there is no fixed rule that would state that it has to
appear once per clause, for instance.

(675) "Optional" use of continuous suffix

kauru letoa-a hau hetau=te, ku-ure niki, nii nuta kuane

3pl send-3ps/D because HRS.FOC go-3pl/E ADVRS that eye inside

kufiui te-he-ι, kauru raj sitQU-nemaahei te-1

line give-CNT-PRT 3pl for rope-also give-PRT
'As he sent them, they went, and he put lines through the "eyes" [of their needles] and
gave them ropes as well.'

O n e might wonder why the two suffixes -ahe (continuous) and -reheto (habitual;
cf. §12.2.6) are occupying separate morphological slots, despite the fact that
they both mark aspect. The reason for this is that there are a f e w examples, in
which -ahe co-occurs with -reheto, directly preceding it, as illustrated here:

(676) Continuous suffix -ahe preceding habitual -reheto

a)
obana su-ahe-reheto-a
cl.peccary kill-CNT-HABl-3ps/A
'He used to be killing collared peccaries.'

b)
satii ririanaj hetau=te raj suuha-atQa kuane huara-he-rehete=lu
all 'all.that' HRS=FOCPOSS heart-only inside see-CNT-HABl:3ps/E=REM

raj neba
POSS mother
'His mother used to keep everything [he said] in her heart.' [NT: Luke 2:51 ]

The continuous suffix is also observed in combination with suffixes of most

other slots, but is not attested with some. These include the distributive (slot 7),
which is considered "bad style". In this case, a combination of the two
grammatical features would be expressed by a participle form plus the verb
sakua (lit. ' f o l l o w ' ) , as in su-ahe-i=te saku-e ( ' k i H ' - C N T - P R T = F O C ' f o l l o w ' -
3ps/E) 'he continues to kill'. Note that the continuous suffix is also
incompatible with the velocity marker -uri and the completive -si.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 465

12.2.5 Suffixal slot 5: Impersonal passive

The impersonal suffix marker -era can be attached to any transitive verb stem,
including those that represent causativised intransitive verbs such u-a 'bring'
("make come") and verbs with the causative suffix -erate. However, the
occurrence of the impersonal passive is infrequent and the information available
on its use is limited. In the database it is only attested in a few examples such as
illustrated in (677). In all examples, it occurs with the negative form, to imply
'nobody'.

(677) Negative impersonal passive form -era-i

mi baja kwituku-era-i d^arihi tga'tti kurerani-a
that after know-IMPS-NEG:3ps/A which ever replace-NTR

eno-re=i teniente gobernadoro ne-1

enter-IRR:3ps/E=ASS village.chief be-PRT
'After that, it is unknown who ever will enter to replace him as the village chief.' [Or:
'Nobody knows who ...']

In fact, forms other than the 3ps negative (-erai) only occurred in elicitation, as
well as combinations with aspect and tense markers. In elicitation, affirmative
forms such as sw-era-a ('kill'-IMPS-3ps/A) 'it was killed' were also confirmed
as correct but these examples are not attested in actual examples from texts.

12.2.6 Suffixal slot 6: Aspect-2: Habitual

The suffixes -reheto and -nahaaoka refer to habitual action. The distinction
between the two is with regard to ongoing vs. completed actions or events. The
suffix -reheto is almost always related to past tense, where it describes a habit
that is not current any more. Thus, one may speak of a "completed habit" here.
The only complication is that this form also can be used with the future tense
(based on elicitation only), where the "completed" meaning becomes less
transparent, implying a meaning such as 'he will be doing it'. When used with
past reference, other elements such as the remoteness marker =lu or a temporal
conjunction (e.g. bana 'at the time when') frequently (but optionally) co-occur
with this form. Alternatively, a reference to the past can be made by an adverb,
or it is evident from the context.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
466 Verbal morphology

(678) Habitual suffix -reheto

a) Past reference with =lu:

ahe-reheto-hoäu=lu
get.drunk-HAB 1 -1 sg/D=REM
Ί used to get drunk.'

b) Past reference with conjunction

amiane-rehete-m-anu=ne bam
work-HAB 1 -NEG-1 sg/D=SUB when
'when I did not use to work'

c) Past reference with adverb:

mi rihei, hitQu nehesißahei hetau=te ne-rehete ßoaelu
that like day pure(ly) HRS=FOC be-HABl:3ps/E earlier
'Earlier, it used to be pure day.' [i.e. there was no night]

The suffix -reheto is subject to certain phonological and morpho-phonological

alternations in combination with other suffixes. As a result of a general
phonological rule, the initial Irl is changed to [t$] if it is preceded by IM or by a
diphthong that ends in [j].

(679) Allomorph -tpeheto

aj-tQeheto-kwaatQe=ne rihihei
AUX-HAB 1 -1 pl/in=SUB like
'like we used to do before'

The final vowel of the imperfective suffix is /u/ underlyingly (but preferably
pronounced as [o] after /t/). However, this vowel is elided when certain suffixes
follow. In particular, this is the case with the 3ps/E suffix -e: as observed in
(678c), the form for HABl:3ps/E is realised as -rehete (but underlyingly would
be -reheto-e). The vowel lol in -reheto is also elided before the negative suffix
-ene (or variants), and it is realised as [e] before the plural suffix -kuru (or
variants), and, as illustrated below. Note that this alternation is not a general
morphological rule, but it also applies to two more forms that end in /to/: ajto-a
-> ajt-e ('say'-3ps/E) and the intensifying suffix -toa which is realised as -te in
the 3ps/E form. Other verbs such as bito-a 'pass' do not follow this rule: the
3ps/E form of bito-a is bito-e.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 467

(680) Variant -reheto

a) Before -kuru:
ku akauru ajßa karahal ne-ure-rehete-kure=lu
there 3pl with long-time be-PL-HABl-3pl/E=REM
'They used to stay with them for a long time.'

b) Before negative form:

rni aj-toehet-em-a=ne bam
that AUX-HAB1 -NEG-3ps/D when
'at the time when he did not use to do that'

The alternation by which the final /o/ of the habitual suffix is absorbed when -e
or -ene follow is not a general rule extendable to all suffixes that begin with /e/
and follow the habitual: in fact, the negative suffix -e in te-reheto-e-i-tg,e=lu
('give'-HABl-NEG-2ps-PL=REM) 'you did not use to give it' does not have
any impact on the final vowel of the habitual.
Again, the occurrence of -reheto is comparatively infrequent. Therefore, the
fact that it is rarely attested in combination with suffixes from other
morphological slots could be coincidental. For instance, there are no examples
in which -reheto would be followed by any markers for aktionsart or by a
participle suffix.
The habitual suffix -mhaaoka cannot be combined with -reheto. It refers to
regularly repeated actions or events that have not been completed or ceased.
Thus, a past reference can be given, but it is implied that the action or event is
still current and likely to be repeated again. Different from -reheto, the suffix
-mhaaoka is not compatible with the future tense.

(681) Uncompleted habitual -nahaaoka

a)
kafiue ha-nahaaoka hwäa
coffee make-HAB2:3ps/A PSN
'Juan always makes his coffee.'

b)
raana su-nahaaoka
w.l.peccary kill-HAB2:3ps/A
'He always kills white-lipped peccaries' [About every four to five months]

The suffix -nahaaoka also involves some morpho-phonological irregularities,

which are illustrated by the paradigm in (682). Specifically, the vowel [o] is
deleted in several forms to result in [nahaak...]. Another irregularity is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
468 Verbal morphology

manifested by the merging of the consonant /k/ with some of the personal and
negative suffixes. For negated lpl/E forms, there are variants.

(682) Irregular person forms with suffix -nahaaoka

AFF, A/D-form NEG, A/D-form AFF, E-form NEG, E-form

lsg -nahaakäu -nahaauäu -mhaakoo -nahaakeneü
2sg -nahaauki -nahaauei -nahaauki -nahaauei
3sg -nahaauka -nahaaui -nahaake -nahaakene
lpl/in -nahaaukaatoe -nahaauaakaatQe -nahaaukaatoe -nahaakemakaatoe
-nahaauaakaatQe
lpl/ex -nahaaukaanu -nahaauaakaanu -nahaaukaanu -nahaakemakaanu
-nahaauaakaanu
lpl/du -nahaaukaa -nahaauaakaa -nahaaukaa -nahaakeniakaa
-nahaauaakaa
2pl -nahaaukitge -nahaaueitge -nahaauhtoe -nahaaueitQe
3pl -nahaaukurua -nahaaukurui -nahaaukure -nahaaukurene

12.2.7 Suffixal "zone " 7: Distributive, Plural object

The two suffixes that occupy this position represent a "zone" rather than a set of
mutually exclusive set of suffixes: the distributive suffix -ahva (and variants)
and the plural object marker -naha (with variant -βaha after /i/). The function
and distribution of both suffixes is discussed in §8 on number marking and will
only be summarised here. The plural object marker occurs under the following
circumstances:
1. When the object is human and refers to 3pl, and
2. when the object does not surface as a noun or pronoun.
3. With transitive verbs only.

(683) Plural object marker -naha

hoajtel kuruata+t^urjka hanii-ßaha-i, kiuajtel letoa-naha-a

again two+ten select-PLO-PRT again send-PLO-3ps/A
'Again he selected 20 [of them] and again he sent them.'

The conditions for the occurrence of the distributive marker -ahva are similar
and can be described as follows (also cf. §8.5). It is used
1. with reference to human participants only, and
2. when the object does not surface as a noun or pronoun (with transitive verbs).
3. Different from -naha, -ahoa occurs with transitive and intransitive verbs.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 469

(684) Distributive marker -ahua

m-a hau hetau=te ku hetau rietoo-hioahva-ure

say-3ps/D because HRS=FOC there HRS go.to.sleep-DSTR-3pl/E
'As she said so, they each went to sleep.'

A peculiarity is that the presence of one of these suffixes does not automatically
exclude the other. Both suffixes can occur adjacent to each other - and in
variable order. This makes it difficult to assign each suffix to a separate slot,
since the order cannot be determined. Examples for this are given in (685). In
(685a), the distributive marker precedes the plural object suffix, and in (685b),
the reverse order applies.

(685) Combination of plural object marker and distributive suffix

a) DSTR > PLO:

su-akwa-naha-a
kill-DSTR-PLO-3ps/A
'He has killed them one by one.'

b) PLO > DSTR:

letoa-naha-akwa-anu
send-PLO-DSTR-1 sg/A
Ί have sent them one by one.'

In addition to the variable order of -ahva and -naha, these can also change
position with regard to other suffixal slots. As illustrated in (686), -naha may
precede the habitual suffix (slot 6).

(686) Position variation with habitual suffix

karesero kbtane-uru, te-rate-naha-reheto-ö=lu, satu kitQa-uru,

prison inside-PL give-CAU2-PLO-HABl-lsg/E=REM all man-PL

eene-kuru
woman-PL
Ί used to send to prison all, men and women.' [NT: Acts 22:4]

In another type of variation, the suffixes of zone 7 can follow suffixes of slot 8
or 9 (cf. (687)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
470 Verbal morphology

(687) Distributive following slots 8, 9

a)
itQa-hee-hahoa-a
do-DIM-DSTR-3ps/A
'He [a weak person] made one by one.'

b)
su-uri-hje-hahoa-a
kill-RAP-CNT-DSTR-3ps/A
'He is rapidly killing one by one.'

As seen in the example in (687), -ahva has an allomorph -hakwa, which occurs
after verbal roots that end with the vowel /a/. Note that the first example in
(688) exhibits the opposite order compared to (685b). Below, the suffix is
realised as -hahoa, since it directly follows the root, whereas in (685), it
surfaces as -akwa, following the plural object suffix.

(688) Alternation -akwa vs. -hahua

letoa-hafava-naha-anu
send-DSTR-PLO-1 sg/A
Ί sent them one by one.'

As is evident from the examples in (687), the insertion of /h/ to form -hakzua is
not restricted to the occurrence of the distributive suffix after verbal roots. For
instance, in (687a), /h/ is inserted after the diminutive suffix -hee and in (687b)
it occurs after the continuous suffix -ahe. Recall that under normal conditions,
any sequence of /e/ followed by /a/ (such as the final vowel of -ahe or -hee)
would be realised as /i.a/ (cf. §2.8.4). The two suffixes seen here appear to be
resistant against this phonological change, and in order to avoid the illicit
sequence fe.af, a glottal fricative /h/ is inserted. However, recall that this is not a
general phonological rule, but restricted to the morpheme -ahva. A similar
situation applies to verbs whose roots end in /a/, as mentioned above: /h/ is
inserted here in a dissimilatory function, in order to make the morpheme
boundary more transparent. The situation is further complicated as the number
of syllables of a verbal root also seems to play a role: after disyllabic roots such
as ama-a 'take along', ari-a 'seek', or huhi-a 'stab', the distributive is realised
as -ahva. After most trisyllabic roots, such as asihjau-a 'steal', itQua-ka 'tie',
and kureti-a 'buy', the allomorph -hahua is chosen. However, there are a
number of exceptions to this regularity, as also some disyllabic verbal roots take
the variant -hahva. These include hoara-a 'see' and bajti-a 'forget';
conversely, a few trisyllabic roots occur with -ahva, such as koituku-a 'know'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 471

and kunajti-a 'be sick'. As the exact conditions for the distribution of the two
allomorphs remain obscure, one has to assume that there are quite a number of
exceptions from the tendencies described above. Also note that for some verbs
(e.g. kajfla-a 'return'), both variants are acceptable.
It should also be mentioned that there is an ordering irregularity between the
causative -erate and the PLO suffix -naha. There are several examples where
PLO precedes the causative even though the former normally occupies slot 2.
The form sw-erate-naha-a ('kiir-CAU2-PLO-3ps/A) 'he has them killed' is
equally acceptable as the variant su-naha-erati-a ('kill'-PLO-CAU2-3ps/A).

12.2.8 Suffixal slot 8: Velocity

The velocity suffix -uri (and allomorphs), which occurs in this position refers to
rapid speed, but may also have other functions, such as diminutive or politeness
(cf. §13.4.1). The interpretation of its semantics depends on the context. In
(689), its use clearly suggests a reference to speed.

(689) Suffix -uri referring to velocity

a)
aresipe muku-uri-1, enamhja kuane hau-to-a=ne ahariri
mojara.fitsh catch-RAP-PRT canoe inside throw-INTS-3ps/D gamitana.fish

ne-1 nerutu-e
be-PRT tum.into-3ps/E
'He quickly caught a mojara fish and when he threw it into the canoe, it turned into a
gamitana fish.'

b)
ku kuru-ri-1, naboaaunei hoajte-u=ra, na-a hau
there swallow-RAP-PRT again repeat-IMP=EMF say-3ps/D because
'He quickly swallowed it, and as he said "do it again", [he caught another fish].'

Only when -uri co-occurs with an imperative form, is it associated to politeness

(also cf. §13.4.2). In (690), the function "velocity" is interpreted as diminutive
("a little"), which in turn takes polite function with the imperative.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
472 Verbal morphology

(690) Politeness function of -uri

a)
basihu kuak-uri-u
a. while wait-RAP-IMP
'Please wait a little while!'

b)
ka=kulate-ri-u-toe=ra
1 sg=extend.visit-RAP-IMP-PL=EMF
'Please [still] stay with me!'

A peculiarity is that the position of -uri is variable with respect to its co-
occurrence with suffixes of slot 7 (distributive, plural object). In some
examples, -uri follows these suffixes, as illustrated below:

(691) RAP suffix following slot 5 suffixes (elicited)

a)
itQa-hahoa-uri-a
do-DSTR-RAP-3ps/A
'He quickly made one by one.'

b)
itQa-hahoa-naha-uri-anu
do-DSTR-PLO-RAP-1 sg/A
Ί quickly made various one by one.'

As -uri does not occur between -ahoa and -naha, it is evident that it occupies a
separate slot. However, its position after these suffixes is not fixed, as in other
examples, it is found before -ahua and -naha. At this stage, there are no clues as
to whether this involves any semantic or pragmatic difference and will be
accepted as an arbitrary variation.

(692) RAP suffix preceding slot 5 suffixes (elicited)

a)
itQaa-uri-ahoa-a
do-RAP-DSTR-3ps/A
'He quickly made one by one.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 473

b)
itQa-uri-ahoa-naha-anu
do-RAP-DSTR-PLO-1 sg/A
Ί quickly made them one by one.'

The velocity suffix has the allomorphs -ri, -uri, and -kuri, whose distribution is
determined by a complex set of factors, some of which have not been explored.
When a root ends in /«/, the velocity suffix is realised as -uri, as turu-uri-a
('arrive'-RAP-3ps/A), saru-uri-a ('push'-RAP-3ps/A), and temu-uri-a
('plant'-RAP-3ps/A) illustrate. When the root-final vowel of the verb is /e/ or
/i/, the velocity suffix is realised as -hiri in most cases, such as in here-kuri-a
('want'-RAP-3ps/A), kusi-kari-a ('bite'-RAP-3ps/A), or kurete-hiri-a ( ' b u y ' -
RAP-3ps/A). However, this rule does not account for the forms sim-uri-a
('sleep'-RAP-3ps/A), te-nri-a ('give'-RAP-3ps/A), and ere-uri-a ('speak'-RAP-
3ps/A), even though the same conditions in terms of the root-final vowel are
given.
Word length does not seem to be the relevant factor, since most verb roots
involving -hiri or -uri are disyllabic. In some cases, the velocity suffix surfaces
as -ri, with some variation. For instance, 'he made it quickly' can surface as
itQa-ri-a or itQa-kuri-a. Apart from this variation, the form -ri occurs with most
verbs whose root ends in a vowel sequence. Examples for this are lurua-ri-a
('lay'-RAP-3ps/A), aua-ri-u-tge ('hear'-RAP-IMP-PL), or naurutaae-ri-u=ra
('displace'-RAP-IMP=EMF), but there are counterexamples (e.g. ubaae-kuri-a
('be angry'-RAP-3ps/A).
The alternations of the velocity suffix are summarised in (693). Including all
exceptions, the distribution of allomorphs is so complex that it is hard to make
out a systematic rule at this point. The only alternation that is 100% predictable
is the realisation of the morpheme as -uri after /«/, and some regularities as
mentioned above.

(693) Velocity suffix: Attested allomorphs in different phonological contexts

Context -ri -uri -kuri -uri

(root ending)
After VV yes no yes no
After /a/ yes yes yes no
After /e/ yes yes yes no
After l\l yes yes yes no
After /u/ (optional) yes yes no
After / « / (optional) no no yes

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
474 Verbal morphology

12.2.9 Suffixal slot 9: Diminutive/Counterexpectation

The form -hee can refer to 'something or someone small or w e a k ' , as illustrated
in (694a). The diminutive function usually refers to the subject of the clause,
but, depending on the context, may also be understood as assigning diminutive
function to the core a meaning of the verb (i.e. 'a little'). In other examples, the
use of -hee here implies surprise about the fact that the person w h o performed
the action is able to do it despite his/her weakness. The reference to
counterexpectation is also expressed in (694b), which implies that the happy
person was expected to be sad. Alternatively (depending on the context), (694b)
can mean that the person is "a little happy", i.e. the form can have direct
diminutive function. While one could hypothesise that diminutive and
counterexpectation are homonyms, there is a relation between the t w o
functions: The "surprise" component of -hee is indeed based on the fact that the
agent has characteristics which imply a diminutive interpretation. For instance,
the examples in (694) have the connotation that the agent is indeed small or
weak (or sad, as for (694b)).

(694) Suffix -hee referring to counterexpectation

a)
itQa-hee-ka
do-DIM-3ps/A
'He (surprisingly) did it.' [Talking of a child or old person, thinking that he would be
too weak to do it.]

b)
kuruae-hee-ka
be.happy-DIM-3ps/A
'He is (unexpectedly) happy.'

A diminutive meaning is also implied in (695), where the diminutive refers to

"only one (person), if any at all", as part of a narrative, where a man is the only
survivor after a flood.

(695) Diminutive function of -hee

netitQara, wane kukwarue-hee-ka kat$a=ne

maybe PTY remain-DIM-3ps/D man=CND
'Poor me, maybe if someone has remained [of all those who have died] ...'

With the verb lama 'be missing', the diminutive as in lana-hee-ka 'a little is
missing' (to imply Ί n e e d ' ) is used with a politeness connotation, suggesting

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 475

that the listener is expected to help. Example (696a) illustrates the polite
function of -hee with lanaa, implying that a little firewood would be enough in
order to account for the need.

(696) Nuances of diminutive function with lanaa

fwei lana-hee-ka ka=raj

firewood be.missing-DIM-3ps/A lsg=for
Ί need a little firewood.' [Lit.: I am lacking a little firewood.']

Despite the polite use of -hee with lanaa, the purely diminutive function may
also be implied with this verb; the variation is only evident from the context. In
(697a), -hee is used as the woman in this narrative almost perfectly complied
with what she was told and missed out only one last task.

(697) Diminutive function with lanaa

itQa-1 lejhti lana-hee-ka=ne hana be

do-PRT one be.missing-DIM-3ps/D=SUB when tell:3ps/E
'Doing it, she left out one [only] and told it.'

The variation in meaning of the suffix -hee can be summarised by the following
functions:
- Counterexpectation
- " W e a k " agent
- "A little" (object reference)
- Politeness

To summarise these different types of usage in one term is difficult, which is

why I suggest that the suffix is in fact a diminutive with nuances in function.
Again, the context of utterance is the major criterion for the distinction of these
functions.
In all attested examples, -hee follows the velocity marker -uri. There is no
example for the occurrence of -hee before the velocity marker -uri, which
confirms the fixed order between these suffixes. However, the examples in
(698) show that the diminutive suffix precedes the suffixes assigned to slot 7.
This order can be accounted for through the variability in position of these
suffixes for plural object and distributive.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
476 Verbal morphology

(698) Diminutive -hee following RAP suffix -uri, but preceding slot 7 suffixes
(elicited)

a)
su-uri-hjee-mha-a
kill-RAP-DIM-PLO-3ps/A
'He (a weak/small person) killed them quickly.'

b)
su-uri-hjee-hahua-a
kill-RAP-DIM-DSTR-3ps/A
'He (a weak/small person) quickly killed one by one.'

T h e suffix -hee is subject to general phonological rules such as nasal spreading

(resulting in -hee after nasal consonants, cf. §2.8.10) and vowel copying (e.g.
resulting in -hjee after /i/, cf. §2.8.3). Note that the 3ps/E form is realised as -ka
when following -hee since it ends in a long vowel. The suffix also has the
allomorph -[V]he, which causes lengthening of the vowel that precedes it.
Examples of this are listed in (699).

(699) Vowel lengthening before velocity suffix

Root 3ps/A With RAP suffix

/ari/ 'seek' ari-a ariihia
/molo/ 'cut' molo-a moloohia
/te/ 'give' ti-a teehia
/nalu/ 'fall' nalu-a naluuhia
/ama/ 'take along' ama-a amaahia

A s a variation, some verbs can take -hee or /-[V]he/ alternatively. The rules for
the assignment of the respective allomorph are again inconclusive. The table in
(700) summarises the possible combinations.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions All

(700) Allomorphy of velocity suffix

Context -hee -[V]he

(root ending)
After VV yes yes
After /a1 yes yes
After /e/ yes yes
After /i/ no yes
After /u/ yes yes
After mJ no yes

12.2.10 Suffixal slot 10: Completive aspect

The suffix -si refers to completive aspect and often has an additional
connotation of "achievement". In the examples in (701), -si refers to an action
or event that occurred "before" or "previously", in relation to the time of
reference.

(701) Completive aspect marker -si

a)
najßa-si-anu=w hana=te turu-e
finish-CPL-1 sg/D=SUB when=FOC arrive-3ps/E
'He arrived when I had already finished it.' [Lit. 'When I had finished it before, he
arrived.']

b)
dosmildos ke turu-si-a=ne ahinia, hestjona-anu rajtene
2002 INST arrive-CPL-3ps/D=SUB before direct-lps/D must
Ί must direct it before the year 2002 arrives.'

The sentence in (701b) refers to an event that will have taken place in the future,
i.e. will be past only then. However, future tense is not marked. Instead, the
word ahinia 'before' (otherwise used as a postposition) functions as a
conjunction after the subordinate marker =ne. However, the completive is not
attested in co-occurrence with the irrealis to mark future tense. The
interpretation as "achievement" is particularly strong in combination with
negative forms. The examples in (702) show the use of -si with negative
suffixes. Note that the completive is not compatible with any aspect markers
such as continuous or habitual.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
478 Verbal morphology

(702) Completive marker -si

a)
βadera-ure temu-si-m-a mi aheri+ko-teru aari ke=ne
be.sad-3pl/E cut-CPL-NEG-3ps/D that stone+ASC-axe topa.tree VLI=SUB

häu
because
'They were sad because their stone axe could not cut the topa tree.'

b)
muku-si-ße kanakaana katai=tga
catch-CPL-NEG:3ps/E lpl/ex hook=EMF
'Our fishing hook did not catch [the fish].'

12.2.11 Suffixal slot 11: In-law talk

The Urarina system for politeness involves one form that is used in order to
refer to certain in-laws and a few other persons when referring to them (but not
when talking directly to them; cf. §13.3, where the exact conditions of use are
investigated). For instance, a male person talking about his son-in-law or father-
in-law uses this form, which is expressed as a verbal suffix -ana (and allomorph
-nana). The origin of this suffix is easy to trace, as 'father-in-law' is kaana and
'son-in-law' means nakana, which both have phonological similarity to the
suffix. The examples in (703) illustrate the position of -ana following suffixes
that occur earlier in the linear order.

(703) In-law talk suffix -ana

a) After root:
mhjauria tgae kati saate-nana-a=ne
don't also black.monkey kill.by.blowgun-ILT-Sps/A^NEGF
'He (the godfather) also ought not to kill black monkeys!'

b) After completive suffix:

mi hana=te kaana hanatgäe nijej eru-si-ßana-ene
that when=FOC son-in-law besides not.at.all fmd-CPL-ILT-NEG:3ps/E
'[But] then, besides, my son-in-law did not find [any turtles] at all.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 479

c) After diminutive suffix:

itg>a-hee-nana-a=tge
do-DIM-ILT-3ps/A=PLT
'He (my little in-law) has done it.' [Talking to a woman]

As evident from the examples in (703) there is a phonological alternation. Most

consistently, when -ana follows a long vowel, it is realised as -nana (cf. (703c)).
However, the same allomorph is also used in (703a), where the verb root ends
in /e/, and in (703b), following III (which in addition causes palatalisation).
Again, the exact conditions for this alternation are inconclusive. Since the
number of examples that involve the ILT suffix is very low in the database, the
exact conditions for the allomorphy remain uncertain.

12.2.12 Suffixal slot 12: Plural-1

As discussed in §8 on number marking, Urarina has two different plural

markers used with verbs. The plural suffix -uru (and allomorphs -kuru,
-tQuru), which also functions as a noun plural, is used with 3ps only and occurs
before the person markers (slot 16). The examples in (704) illustrate the
position of -uru in relation to the distributive, irrealis, and negative suffixes.
The alternations regarding the realisation of the plural marker as -kuru and
-tQuru have already been discussed in §3.1.2. Another peculiarity is the 3pl/E
form -ure, in which the plural is merged with the person suffix -e by deleting
the final /«/ of the plural suffix (cf. (704a)). This rule applies in all contexts in
which the suffix that follows -uru begins with the vowel /e/, as is also
illustrated by the form herehtrene 'they do not want' in (704c), where the plural
is followed by the negative suffix -ene.

(704) Position of plural suffix -uru

a) After DSTR suffix (slot 7):

fwila ti-ahoa-ure
battery give-DSTR-3pl/E
'They distributed batteries.'

b) Before IRR suffix (slot 14):

leotga amiane-naa kwara-kuru-re=~i
other work-NOM see-PL-IRR:3ps/E=ASS
'They will look for other works.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
480 Verbal morphology

c) Before NEG suffix (slot 15):

hjane-na here-kur-ene
let-INF want-PL-NEG :3 ps/E
'They did not want to let her go.'

The position of -uru is variable to a certain degree, as it can occur before the
habitual suffix -reheto (slot 6), as a variation (cf. §12.2.6). Thus, both forms for
'they used to live', as illustrated in (705), are possible (with no apparent
difference in meaning). These examples are confirmed by a number of examples
in the N T translation (not listed here), where both suffix orders occur in free
variation. Remarkable is the (unexplained) repetition of the plural suffix in
(705b), where a form of -uru occurs before and after the habitual suffix.

(705) Variable position and double occurrence of -uru

a)
itg,au-ru-rehete=lu / itQau-rehete-kure=lu
live-PL-HAB 1:3ps/E=REM Iive-HAB 1 -3pl/E=REM
'They used to live'

b)
ku akauru ajßa karahal ne-ure-rehete-kure=lu
there 3pl with long.time be-PL-HABl-3pl/E=REM
'They used to stay with them for a long time.'

12.2.13 Suffixal slot 13: Epistemic modality: Probability

The suffix -kaj expresses probability with the connotation of "convincing

evidence" or "unspecified logical conclusion", which makes it plausible to
regard this suffix as a marker for epistemic modality. It can be translated in
various ways, such as 'it must be the case that', 'it seems that', 'probably', or
'certainly', always implying some degree of suspicion, -kaj is the last suffix
occurring before person and polarity marking and can only be followed by
affirmative or negative suffixes. It is incompatible with the irrealis suffix -ri (cf.
§12.2.14), as they exclude each other on the pragmatic level, i.e. if something is
likely to be in a certain manner, the irrealis reading is not applicable.
In most examples, the probability form co-occurs with the adverb heriane,
which can be translated as 'maybe/probably', and, in combination with -kaj
implies 'certainly'. However, the use of heriane is not compulsory. Also note
that the occurrence of the emphatic enclitic =ra in the examples in (706) is a

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 481

coincidence as all sentences displayed here are taken from texts that involve the
traditional language.

(706) Probability suffix -kaj

a) After PL:
heriane edaae-kuru-kaj-a=ra
probably be.shy-PL-PRB-3ps/A=EMF
'They must be shy.' [About people who seem to have disappeared]

b) After RAP:
hirikuri=te ... heriane ra-uri-kaj-a=ra
PSN=FOC probably receive-RAP-PRB-3ps/A=EMF
'Hirikuri quickly must have taken it.' [Suspecting a spirit of theft]

c) With copula:
d$=asi+hatGu kuane=te=ra heriane ne-kaj-a=ra
2sg-nose+hole inside=FOC=EMF probably be-PRB-3ps/A=EMF
'Certainly it must be in your nose.' [Again suspecting spirit of stealing]

A morphological peculiarity is that the person suffix for 3ps, which follows -kaj
is always -a, including the Ε-form, which would normally result in -e, i.e. the
expected 3ps/E form -kaj-0 or -kaj-e does not exist (cf. (706b)). With other
person suffixes, no restrictions apply. However, when -kaj is followed by the
2ps suffix -i, the form is realised as -ka-i [-kaai], as illustrated in example (707).

(707) Probability suffix -kaj with 2ps

ii=te it Qa-ka-i
2sg=FOC do-PRB-2ps
'You must have done it.'

12.2.14 Suffixal slot 14: Mood/Future tense

Urarina does not have any formal marking explicitly used for tense. For
instance, the remoteness marker =lu is usually used with past reference, but it
can also refer to remote future when used with some words (as in
jioae=lu netohwei 'forever' (cf. §5.7.3). There also is no suffix that would mark
future tense on its own. Instead, future tense is a construct of the irrealis marker
-ri (and allomorph -re) and the assertive enclitic =ni, which is added in slot 18
(cf. §12.3.2). Without this clitic, the form is understood as irrealis. While -ri can

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
482 Verbal morphology

be translated as 'would', the addition of the assertive expresses the certainty of

an action or event to happen in the future.

(708) Irrealis suffix -ri

a)
hjauißera, ka=kutaj-ri-a ofiva=ra
no lsg=scold-IRR-3ps/A father=EMF
'No! My father would scold me [if I urinated here].'

b)
ruru-te najpe-re kuraanaa m-a=ne=ra
howler. monkey=FOC be.able-IRR:3ps/E chief be-3ps/D=CND=EMF
'The howler monkey would be able, if he were the chief.'

Note that with the irrealis reading, mood is only marked on the main verb, but
not on the verb of a subordinate clause, as illustrated in (708b). In fact, the
irrealis marker is the only verbal suffix that does not normally occur in
dependent clauses, except in elicitation (cf. §20.1.1.3). The sentences in (709)
show the interpretation as future tense when -ri occurs in combination with the
assertive enclitic.

(709) Irrealis suffix -ri plus assertive marker for future reference

a)
kaa d^aura+atane itahe-ri-tQäu=ni=tQa
this flesh+land destroy-IRR-1 sg/A=ASS=EMF
Ί will destroy this earth.'

b)
kwitu1oi-erate-re-ü=m
know-CAU2-IRR-1 sg/E=ASS
Ί will let him know.'

The realisation of -ri is subject to the general phonological alternations as

described in §2.8. After /i/, Irl is changed to [t$], resulting in -t$e, as in sim-tge
( c sleep'-IRR:3ps/E) 'he would sleep'; when the suffix is followed by /a/, it
surfaces as -ri, which is illustrated by example (708a). As a morpho-
phonological rule, when -ri is combined with the 3sg/E form -e, the two suffixes
merge to -re (cf. (708c)). The form of the assertive enclitic =m is subject to
additional alternations, which will be described in §12.3.3.
Further alternations related to the irrealis marker are illustrated by the
paradigm in (710). These changes can be summarised in the following way:
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 483

T h e irrealis suffix (affirmative) is realised as -ri:

- With all lpl forms
- Before all personal suffixes of the A-conjugation
- With 2ps forms of E-conjugation
- Before 1 sg/D form

Elsewhere, the irrealis suffix surfaces as -re, which makes it possible to assume
-re as the underlying form. Alternatively, (as implied here) one could assume
that -ri is the underlying form and describe the occurrence of -re as occurring
only in the forms for lsg/E, 3pl/E, and in 2ps/D. A s -ri is the allomorph with a
wider distribution, I will refer to this form as the underlying morpheme.

(710) Paradigm for irrealis (affirmative)

AFF, A-form AFF, E-form AFF, D-form

lsg -ri-tQäu -re-ü -ri-anu
2sg -ri-ki -ri-ki -re-i
3sg -ri-tQa -re -ri-a
lpl/du -ri-tQaa -ri-tQaa -ri-aka / -ri-tQaa
lpl/in -ri-tQaa-tge -ri-akatoe -ri-akatoe
lpl/ex -ri-tpaanu -ri-akaanu -ri-akaanu
2pl -ri-ki-tge -ri-ki-tQe -ri-ki-toe
3pl -uru-ri-tQa -uru-re -uru-ri-a

In addition to the changes of the suffix -ri, the shape of the person markers that
follow the irrealis is subject to alternation:
- When the irrealis is realised as -ri, a following /kJ should change to [t$], but
it does not in this context, instead, 2ps forms surface as -ki.
- The lsg/A form is irregular.
- In the lpl forms, the A conjugation is distinct from the other conjugations,
while otherwise A- and D-conjugations coincide (i.e. in the declarative
mood).
- In the 3ps/E form, /e/ merges with the irrealis suffix.

Further alternations apply when the irrealis is combined with a negative suffix.
In the Α-conjugation, -ri is followed by the negative marker, which varies
between -e and -a, depending on person (also cf. the following §12.2.15). In the
lsg/A, and 3ps/A form, the negative -a is realised as -βa, for which there is
good reason: when the affirmative form follows the irrealis, it is realised with
[tp], e.g. ne-ri-tQa=i ( ' b e ' - I R R - 3 p s / A = A S S ' h e will be'). The palatal nasal that
precedes the negative in this particular context is inserted for dissimilatory
purpose. This happens in a similar way with the negative irrealis form for

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
484 Verbal morphology

3ps/A, which is -ri-ßa-i, as opposed to the affirmative form -ri-tQa. The 2sg/A
is variable with regard to the negative marker, as it can surface as -a or -e. In all
other forms, negation is marked by -ene (or -ne with 3pl/E,D).
The realisation of irrealis and negative in the other conjugations follow
similar morpho-phonological alternations as the ones observed in the
affirmative irrealis. However, the order between the two suffixes is reversed,
i.e. negation precedes the irrealis rather than following it.

(711) Paradigm for irrealis (negative)

NEG, A-form NEG, E-form NEG, D-form
lsg -ri-jia-u -ene-re-ü -ene-ri-anu
2sg -ri-e-i -ene-re-i -ene-re-i
-ri-a-i
3sg -ri-ßa-a -ene-re-0 -ene-ri-a
lpl/du -ri-a-aka -ene-ri-aka -ene-ri-aka
lpl/in -ri-a-aka-tQe -ene-ri-aka-tQe -ene-ri-aka-tge
lpl/ex -ri-a-akaanu -ene-ri-akaanu -ene-ri-akaanu
2pl -ri-e-i-toe -ene-re-i-tge -ene-re-i-tge
3pl -uru-ri-ßa-a -uru-ne-re-0 -uru-ne-ri-a

12.2.15 Suffixal slot 15: Negation

Urarina has several negative suffixes, which are used with regard to person,
conjugation class, and other factors, as depicted in the paradigm for person and
polarity marking in (712). Since the different negative markers all fulfil the
same function, I consider them as projections of one morpheme, even though
they have quite different forms. While it is not possible to identify one
particular suffix as "underlying", the different forms can be understood as
allomorphs of each other.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 485

(712) Paradigm for negative marking (realis)

NEG, A-form NEG, E-form NEG, D-form

lsg -a-it -ene-ü -em-anu
2sg -e-i -erte-i -ene-i
3sg -e-i /-0-i /-ji -ene-0 -eni-a
lpl/du -a-aka -eni-aka -eni-aka
1 pi/in -a-aka-tge -eni-aka-toe -eni-aka-toe
lpl/ex -a-akaanu -eni-akaanu -eni-akaanu
2pl -e-i-tge -ene-i-toe -ene-i-toe
3pl -uru-i-0 -ur-ene-0 -ur-eni-a

The regularities for the distribution of the respective negative markers can be
summarised in the following way:
- With all Ε-forms and D-forms of the verb, the negative suffix is -ene (or -era
before /a/). In contrast, Α-forms are subject to a further division.
- With all lps/A forms, the negative suffix is -a.
- With all 2ps/A forms, the negative suffix is -e.
- With all 3ps/A forms, the negative suffix is -i.

The last peculiarity listed above (regarding -i) occurs in the negative marking
for 3ps forms: in this context, the negative suffix -i and the person suffix are
fused, i.e. the form is realised as -i. For 3ps/A, the expected person suffix -a is
elided, while in the Ε-form, the 3ps/E suffix -e is elided. In the D-form, such
elision does not take place: -a regularly occurs as a person marker. However,
there is a variant for NEG:3ps/A, which consists of the negative marker -e
followed by -i. I suspect that this might be the underlying form, which can be
simplified to -ji in the following way: the suffix sequence -e-i is realised with
hiatus after a vowel-final verb stem (i.e. [jei]), from which the form [ji] can be
interpreted as a simplification. A matter of concern might be whether -i in the
3pl/A form -uru-i functions as a person marker or a negative suffix. If we
assume that -i functions as a person marker for negative 3sg/A, we may
conclude that the corresponding 3pl form -uru-i is a simplification of the (non-
existent) form -uru-e-i.
In addition to these paradigmatic regularities, a few morpho-phonological
rules apply. In particular, a glottal stop is inserted in certain contexts. For
example, the negative suffix -a for lps/A is realised as -?a after some verbal
roots that end with /a/. In a similar way, it is -?e for the negation of 2sg (cf.
(713c)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
486 Verbal morphology

(713) Negative allomorphs with glottal stop after roots ending in /a/

a) lsg:
eseneta-7a-ü
believe-NEG-1 sg/E
Ί don't believe it.'

b) 1 pi:
ne-betaka-7 a-aka
ITR-relax-NEG-1 pl/du
'We did not relax.'

c) 2sg:
kauatQa-1 itQa-?e-i
good-PRT do-NEG-2ps
'You did not do it well.'

The variation with the insertion of a glottal stop also occurs after verb roots that
end in a vowel sequence, such as shown in (714a). It is also observed after the
distributive suffix -ahua (cf. (714b)).

(714) Negative allomorphs with glottal stop in other contexts

a) After vowel sequence:
netoe-7a-ü
be.in.mood-NEG-1 sg/E
Ί am not in the mood.'

b) After suffix:
itg,a-häku>a-7 a-ü
do-DSTR-NEG-1 sg/E
Ί did not make one by one.'

After the front vowels Id and i\! the negator -i is realised as -;/. Note that the
root vowel Id is assimilated to [i] in this combination: for instance, the
underlying form of 'want', /here/ occurs as [heri] when combined with the
negator -i.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 487

(715) Allomorph -ji

a) After root ending with /e/:
lairete-na heri-ji
pay-INF want-NEG:3ps/A
'He did not want to pay.' [Underlying form of'want' is /here/]

b) After root ending with /i/:

d$uni akano ki-ji
nobody snake eat.NEG:3ps/A
'Nobody eats snakes.' [Underlying form of'eat' is /ki/]

c) After suffix ending with /i/:

sim-si-ji
sleep-CPL-NEG:3ps/A
'He cannot sleep.'

As discussed above, further alternations occur when the negative co-occurs with
the irrealis form (cf. the paradigm in (711)).

12.2.16 Suffixal slot 16: Person

Person is the only grammatical category that is obligatorily marked on all finite
verbs. Number agreement refers to the subject of a clause of any transitivity
class. The most peculiar feature in this regard is the existence of three different
inflection classes that are related to certain syntactic conditions. The exact
occurrence of the different conjugation forms in context is discussed in §11. In
a highly simplified manner, the three classes can be characterised as follows:
1. D-class: these forms occur in dependent clause (cf. (716a).
2. Ε-class: these forms occur when a focus marker precedes the verb anywhere
in the clause (cf. (716b), or when focus or emphasis is implied in other ways.
3. Α-class: these forms occur elsewhere, particularly in short utterances (cf.
(716c)).

The occurrence of these forms is illustrated in (716). It should be noted that the
distribution of these different forms does not depend on particular verbs (i.e. the
verbs are not divided into morphological classes in this regard): instead, the
three different forms, (which correspond to three different paradigms), may be
attached to any verb, depending on the context or other conditions.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
488 Verbal morphology

(716) Person inflection types

a) D-form:
enejtQU su-a=ne hau
monkey kill-3ps/D=SUB because
'because he killed the monkey'

b) E-form:
aka=te enejtQu sw-e
3sg=FOC monkey kill-3ps/E
'He [focus] killed the monkey.'

c) A-form:
enejtQu $u-a
monkey kill-3ps/A
'He killed the monkey.'

The differences become even more evident in combination with the irrealis
form, where there is a different person suffix for each person inflection class.
Each of the verbs in (717) is composed of the root for 'do', an irrealis suffix,
and a person marker for lsg, meaning Ί would do it'.

(717) Different person inflection suffixes with irrealis form

Α-form, lsg: itQa-ri-tgäu
Ε-form, lsg: itQa-re-u
D-form, 1 sg: itga-ri-anu

Otherwise, there is some degree of overlap for the forms for 3ps/A and 3ps/D,
i.e. the number of person suffixes is not exactly tripled by the existence of the
three different classes. This is best exemplified through the paradigm in (718),
where the three forms are contrasted for the affirmative realis mood.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 489

(718) Paradigm for affirmative marking (realis)

AFF, A-form AFF, E-form AFF, D-form
lsg -am -Ü -arm
2sg -i -i -i
3sg -a -e -a
lpl/du -aka -aka -aka
lpl/in -aka-tge -aka-t$e -aka-toe
lpl/ex -akaanu -akaanu -akaanu
2pl -i-tQe -i-tQe -i-toe
3pl -uru-a -ur-e -uru-a

These paradigms allow a number of generalisations. One issue is the degree of

overlap between the three conjugation types.
1. In the affirmative, realis mood, all Α-forms and D-forms coincide, i.e. the
person markers in a dependent clause are identical to the ones for the A-class
in a main clause. Recall that this is not the case for the paradigms listed for
negation (cf. (712)), where there is a stronger relation between D-forms and
Ε-forms, e.g. the choice of the negator coincides between E/D-conjugation.
2. The 2ps forms are identical in the singular and in the plural, for all
conjugation classes. The same applies to lpl forms of any kind.

Thus, the main difference lies between the forms for lsg and 3ps (singular and
plural) of the A- and Ε-classes. In fact, these terms are based on the most
common occurrence of verb forms, i.e. the 3ps/A form is -a, whereas the 3ps/E
form is -e.58
Particularly remarkable is the composition of the forms for lpl. Normally,
one would expect a hierarchy in the composition of forms for dual, inclusive
plural, and exclusive plural. In morphological terms, the dual -aka is the least
marked form. This form refers to two participants, the listener and the speaker.
The lpl inclusive form is logically built on this through the attachment of the
plural marker for 2ps. Thus, the suffix -aka-tge refers to the speaker, the listener
and an undefined group of others 'we all'. However, the most interesting form
is the one for lpl exclusive, which again has -aka as its first element, but is then
followed by the marker for lsg, to result in -aka-anu. Semantically, it refers to a
group of people, but strictly excludes the listener. Morphologically, it is the
combination of the dual with lsg, such as to "neutralise" the involvement of the
listener by the use of the lsg. The table in (719) illustrates the distinction
between first person forms in terms of (participant) features.

58
I have adopted hese terms from P. Manus (1992) who uses them in her paper on
subordination.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
490 Verbal morphology

(719) First person distinctions

Participants: 1 2 3
1 sg -anu + — —

lpl/du -aka + + —

lpl/in -aka-t^x + + +
lpl/ex -aka-anu + - +

Probably as a recent development, the dual form is sometimes employed to refer

to a 1 pi inclusive context. This is mainly observed with younger speakers: when
referring to 'all of us', the morpheme -t$e was frequently elided. The reasons
for this could be that the principal contact language Spanish does not
distinguish between dual and 1 pl/inclusive forms. Thus, one may assume that
the imminent neutralisation between these two forms is due to the influence of
Spanish (also cf. §23.3). Alternatively, the development could be interpreted as
another instance of the general trend for simplification, as is applied to other
areas of the grammar.
The paradigm in (718) shows the prototypical forms as they occur with most
verbs. However, there is a range of phonological and morpho-phonological
alternations that apply to the combination of person markers with certain verb
stems. While these are investigated in §3.1, a brief summary is given here:
- 3sg/A is realised as -ka under the following conditions:
- if the root ends in a long vowel (except /ii/), or
- in a sequence of two vowels (but not after diphthongs).
- 3sg/A is realised as -t$a if the root ends with /ii/.
- lsg/A is realised as -käu under the following conditions:
- if the root ends in a sequence of two vowels
- with some exceptional verb roots that end in /a/
- lsg/A is realised as - f p ä « under the following conditions:
- after disyllabic verb roots ending in /i/;
- after certain trisyllabic verb roots ending in /e/ or /i/.
- All lpl forms are realised with initial /tp/ after certain trisyllabic verb roots
ending in /e/ or /i/.

There also is one irregularity for lpl marking occurring only in content
questions, which may be called a "desperate interrogative". In this case, a lpl
form such as -ri-tgaat$e (the combination of irrealis and lpl/in/A suffix) can be
replaced by the form -naare (or variant -naanere), as illustrated in (720). The
pragmatic differences between the two forms appear to be marginal, as the use
of -naare implies some degree of desperation, with reference to the future. The
presence of /re/ as a component of this form may thus point to the irrealis

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Suffixal positions 491

character of the suffix (recall that an allomorph of the irrealis suffix is -re).59
Otherwise, it has the same function as the regular lpl suffixes. Interestingly, the
distinction between inclusive, exclusive, and dual forms are neutralised by
-naare and can only be indicated by the additional use of a free pronoun.
Another peculiarity is that the desperate interrogative exists only for lpl, but
cannot be used for l s g forms (for which there is no separate suffix either).
Example (720b) shows that the form can also be preceded by other suffixes,
such as the distributive -ahua.

(720) Desperate interrogative suffix -naare

a)
d$a itg,a-naare=ta
where sleep-DSP=FRS
'What are we going to do?' [Context: 'the work is finished, so how are we going to
continue?']

b)
d^atoanei itQa-hahva-naare
how do-DSTR-DSP
'How are we going to make each of them?' [Elicited]

The suffix also has an allomorph -naanere, but the conditions for the
distribution of the two forms are unclear, as it depends on the respective verb.
For instance, with ' g o ' , the preferred variant is ku-manere, whereas ht-naare is
less acceptable. In contrast, for ' d o ' , the preferred form is itea-naare. The
examples in (721) further illustrate that the form can be used with any
interrogative pronoun. The desperate character of a situation may also be
emphasised by particles such as the pity marker tioane or the frustrative enclitic
=ta.

(721) Allomorph -naanere

a)
d$u sim-ßaanere=ta
where sleep-DSP=FRS
'Where are we going to sleep?' [Context: Arriving in the city with no place to stay]

59
However, it is not possible to attach -naa(ne)re as referring to any other person
marking than lpl. Thus, it must be assumed that /re/ is fossilised.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
492 Verbal morphology

b)
d$u noane ku-naanere
where PTY go-DSP
'Where would we go?' [Context: Arriving in the city with no place to stay]

12.2.16.1 Suffixal slot 17: Plural-2

While 3ps plural is marked by the suffix -uru (cf. §12.2.12), plural for second
person is marked by -t$e. It also functions as a plural marker for 2ps
imperatives, which confirms its function as a pluraliser for second person.
Deviating from that, -tge is also found as a component of the form for
lpl/inclusive, as mentioned above. The examples in (722) illustrate its
occurrence after person markers.

(722) 2pl with -tQe

a)
d^anu-na nitoanet itQa-ki-tQe
why-INT like.that do-2ps-PL
'Why did you (PL) do it like that?'

b)
ßäe lomaj atane nitQae turu-i-tQe=ne hana
already PSN land vicinity arrive-2ps-PL=SUB when
'when you (PL) get close to the land of Lomai'

However, the status of -toe as a suffix is only confirmed in this position, i.e.
directly attached to the 2ps form. In other positions, it is also used as a
politeness-marking enclitic (cf. §12.3.1). When several persons are to be
addressed politely (e.g. several women), -t$e is marked only once, as a plural
suffix after the person marker. Thus, one may assume that the polite form has
developed from the plural marker for 2ps.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 493

12.3 Enclitic positions

The following sections discuss further postverbal slots, with the difference that
these can be characterised as enclitics. One criterion for their distinction from
suffixes is that enclitics are typically unstressed, and that they can attach to
word classes other than verbs. While this cannot be shown for every single
clitic, the fact that the first morpheme that follows the 17 suffix slots arguably
has clitic status implies that all other elements that follow it are not verbal
suffixes (since they are not found anywhere earlier in the linear order). A more
detailed discussion of the clitic status of these morphemes is also found in
§2.11.

12.3.1 Slot 18: Politeness

The morpheme =t$e was already discussed in its function as a plural marker in
§12.2.16.1. In another position, =t$e functions as a politeness-marking enclitic,
which is most typically attached to verbs. In (723), it occurs after 3ps forms,
which dispels any doubt that it could be identical with the plural marker
discussed in §12.2.16.1.

(723) Position of politeness marker, use with verbs

a)
ßäe=he=ra sa-a=toe
already=REP=EMF end-3ps/A=PLT
'It is already finished, he says.' [Man talking to woman]

b)
raj nakana raj, ßäe huru-a=tge bitoa, na-a=ne sajhjel
POSS father-in-law for already snap-3ps/A=PLT trap say-3ps/D=SUB although
'Even though he said "the trap has already snapped" to his father-in-law,...'

However, the occurrence of =t$e is not restricted to verbs; it can even occur on
other word classes. In (724a), it is repeated on the prohibitive introducer
nihjauria, in addition to its occurrence after the verb. Note that this example is
elicited and Urarina speakers describe the presence of as being a feature of
the traditional language, where it would imply that someone is "really polite".
The example in (724b) is taken from the NT and illustrates the occurrence of
=t$e after the participle form of an auxiliary, which itself is attached to a noun.
Also note that again, the politeness marker occurs twice, being repeated on the
main verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
494 Verbal morphology

(724) Politeness marker =t$e with other word classes

a)
mhjauria=toe itQa-i-tQe-ne
don't=PLT do-2ps=PLT=NEGF
'Don't do it!' [Singular meaning, but could be interpreted as plural without context]

b)
d^anuna hurju=ne-l=tQe=ne natu, ka=kahe akau baha-i=tge
why Jew=AUX-PRT=PLT=SUB though lsg=from water ask-2ps=PLT
'Why are you, even though being a Jew, asking water from me?' [NT: John 4:9;
Woman talking to man]

Regarding the exact function of the politeness marker, a more in-depth

discussion is given in §13. As a summary it can be stated that =t$e is used in
conversation between members of the opposite sex and in certain in-law
relationships.

12.3.2 Slot 19: Assertive

The assertive form is only found in combination with the irrealis suffix (cf.
§12.2.14) in order to mark future tense. More exactly, the assertive appears to
validate an assumption expressed through the irrealis, indicating that the
possible outcome of an event is certain or very likely to be realised. The
examples in (725) illustrate some situations in which this is implied (also cf.
(709)).

(725) Future tense through the use of the assertive marker

a)
ra-naa heri-ki-tQe=ne, ku=te neda-ri-)d-tQe=i=tQa
receive-INF want-2ps-PL=CND there=FOC stay-IRR-2ps-PL=ASS=EMF
'If you (PL) want to get it [the axe], you will stay there.' [Warning by the Envoy of God
about stealing axes from magnetic soil]

b)
ii auna-he-ri-tQäu=m=tQa
2sg hear-CNT-IRR-1 sg/A=ASS=EMF
Ί will be listening to you.' [Threat by monkey that any break of promise made by
woman in birth story will be known to him]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 495

The form of the assertive is subject to a complex system of variation, depending

on the suffix that precedes it. In some cases, the clitic is realised as =1, whereas
in others, it is realised as =m. The distribution of this alternation is rather
obscure, as the following paradigm illustrates (all occurrences of =m are
shaded):

(726) Distribution of assertive allomorphs

a) Affirmative
AFF, A-form AFF, E-form AFF, D-form
lsg -ri-teau=m -re-ö=m -ri-anu=i
2sg -ri~ki=l -ri-ki=i -re-i=tti
3sg -ri-t$a=~i -re=l -ri-a=m
lpl/du -ri-tQaa=ri -ri-tg,aa=m -ri-aka=m / -ri-tgaa=m
lpl/in -ri-t$aa-tg,e=~i -ri-akatgesn -ri-akatQe=i
lpl/ex -ri-tQaanu=~i -ri-akaanu=l -ri-akaanu=i
2pl -ri-ki-tQe=~i -ri-ki-t$e=~i -ri-ki-t$e=i
3pl -uru-ri-tga=i -uru-re=l -uru-ri-a=~i

b) Negative
NEG, A-form NEG, E-form NEG, D-form
lsg -ri-ßa-ü=m -ew-re-ü=m -ene-ri-anu=l
2sg -ri-e-i=m -em-re-i-m -enc-re-i=ni
(-ri-a-i-l)
3sg -ri-jia-a=l -ene-re-i -em-ri-a=rd
lpl/du -ri-a-aka=m ~ene-n-aka=m -erte-ri-aka=m
lpl/in -ri-a-aka-tQe=l -em-ri-aka-tQe-l -ene-ri-aka-tQe=l
lpl/ex -ri-a-akaanu=l -ene-ri-akaanu=~i -ene-ri-akaam=i
2pl -ri-e-i-toe=i -ene-re-i-tg,e=l -ene-re-i-toe=i
3pl -uru-ri-ßa-a=i -uru-ne-re-l -uru-ne-ri-a=Ϊ

One can summarise that the assertive marker is realised as =rn under the
following circumstances:
- After affirmative and negative forms for 1 sg/A and 1 sg/E
- After affirmative and negative forms for 2sg/D
- After affirmative and negative forms for 3sg/D
- After affirmative and negative forms for lpl/du
- After negative form for 2sg/E and 2sg/D

These occurrences are almost impossible to group in terms of person, number,

conjugation class, or polarity, as the alternation goes through all types of these

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
496 Verbal morphology

features. One could postulate a case of metathesis as is observed in a few more

cases: assuming this, the order of segments would be reversed from /n+i/ to
/i+n/ (or vice versa), which is then realised as =7. Still, this does not account for
the conditions under which this process applies.
In combination with the politeness enclitic =t$e (e.g. used in conversation
between men and women; cf. §12.3.1), the situation is complicated further, as
illustrated with the affirmative A- and Ε-forms in (727). In this case, the
assertive is realised as =~i in all forms. Interestingly, the order between assertive
and politeness enclitics is reversed to ASS>PLT in the third person, whereas it
is in the opposite order (PLT>ASS) in all other cases. The same observation can
be made for the negative paradigm (not displayed here).

(727) Assertive form combined with politeness suffix =fpe (affirmative)

AFF, A-form AFF, E-form
lsg -ri-tg,au=tQe=~i -re-ü=tQe=~i
2sg -ri-h=tQe=~i -ri-ki=tg,e=i
3sg -ri-tQa-~i=tQe -re=i=tg,e
lpl/du -ri-tgaa=tQe=i -ri-tQaa=tQe=~i
lpl/in -ri-tQaa=tQe=i -ri-tQaa=tg,e=~i
lpl/ex -ri-tQaanu=tQe=~i -ri-tg,aanu=tg,e=~i
2pl -ri-ki=tQe=~i -ri-ki=tQe=~i
3pl -uru-ri-t$a=5=t$e -uru-re=n=t$e

12.3.3 Slot 20: Witness evidential

Urarina has two evidential markers, which refer to witness (=ra) and reportative
(=he), respectively, and are attached to the verb as enclitics. Their use is entirely
optional, as there are alternative ways to indicate the source of information. For
instance, instead of using the witness marker =ni on a verb, one could use a
construction with a control verb, such as in Ί saw/heard that he did it'.
Similarly, the reportative =he can be substituted by a construction involving the
quotative verb naa (cf. §20.2.4.2). However, the two evidential markers are not
mutually exclusive: they occupy different positional slots. While the witness
evidential occurs in slot 20 and will be discussed here, the reportative
occupies the following slot (21). Their combination will be discussed in
§12.3.5, where also their interaction with the reassurance marker =tau is
investigated.
It should be mentioned that the use of the evidentials is not restricted to any
particular verb class, nor is it limited with respect to person marking on the
verb. An evidential can occur with I s ', 2nd, or 3rd person marking and it is also

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 497

observed with stative and posture verbs. It can be used in polar questions as
well as in content questions. However, while evidential markers may occur in
subordinate clauses, neither =ni nor =he are attested in non-finite clauses.
The witness marker =ni most commonly implies eyewitness (as in (728a)),
but the source of information can also be based on acoustic evidence, or other
senses such as smell or taste, such as illustrated in (728b,c). However, an
"inferred" reading is not implied by =m.

(728) Witness evidential =rd

a) Visual:
itQa-i=m=tau
do-2ps=WIT=REAS
'You did it (earlier today).' [I know it because I was a witness]

b) Taste:
rauto-a=ni
be.tasty-3ps/A=WIT
'It was tasty.' [Referring to what I ate in the morning]

c) Hearing:
itQafiua-a=ni=tau
shoot-3ps/A=WIT=REAS
'He shot it (earlier today).' [I saw or heard him shooting]

At the same time, the witness evidential has temporal function as it refers to
"earlier today", as indicated by the glosses in (728). Apart from its occurrence
with verbs marked for 3ps, =m can also be used with other person suffixes. In
(729), = r i follows a l p s suffix: in this case, it gets assertive function, such as to
mean Ί myself am the witness that I did it'.

(729) Temporal function of witness marker =ni

itQa-käu=m
do-lsg/A=WIT
Ί [really] did it.'

The function of =m in (729) shows an interesting relation to the assertive

marker of slot 19, with which the witness evidential is not compatible. While
the assertive function of the evidential only applies to lps examples, the
assertive enclitic of slot 19 is used with any person suffix, and always co-occurs
in combination with the irrealis. In addition, it exhibits very specific order
variations with the politeness enclitic =tQe (slot 18). The evidential cannot be

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
498 Verbal morphology

used with future reference. Thus, while a common origin of assertive and
witness enclitics is very likely, the homophonous markers have evolved as
different grammatical morphemes in the present.
An extended function of the temporal use of =ni can be observed with
imperative forms (cf. §15.4). Interestingly, when =m follows a verb marked for
imperative, it has temporal function, though not referring to the recent past, but
to the near future - more exactly: later on the same day. The evidential function
of =ni does not apply in this context and it can be concluded that the non-distant
future marker is a morpheme separate from the witness marker.
Another factor is that =ni, as well as the remoteness marker =lu can be
combined with the experience evidential in the reverse order to form an adverb
(or a clitic complex that is used with adverbial function): taurti and taulu are
used to mean 'as we know from earlier today' and 'as we know from (much)
earlier', respectively.

(730) Adverb tauni

taum ii saabere u-a-u
before 2sg machete come-CAUl-IMP
'Bring me your machete "from before"!' [Referring to the one you showed me earlier
today]

Note again that the type of experience described in the above examples varies
from case to case: in (730), the phrase was the result of a situation where the
author had shown a machete to his consultant who came back later to borrow it
(visual evidence).

12.3.4 "Zone " 21: Reportative evidential; remoteness

The position after =ni can be filled by two enclitics with different functions, the
reportative =he and the remoteness marker =lu. Their order is variable, in that
each of them can precede or follow the other.
The reportative =he expresses the fact that the information given in a clause
is based on the witness of a third person and may thus be described as second
hand information. Note that the enclitic can also occur with verbs that are
marked for other persons as subject markers. In (731 b), the speaker states the
reported evidence of some one else about his own (the speaker's) action, using
the lsg form. A connotation of this example is that the speaker himself cannot
remember the incident.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 499

(731) Reportative marker =he

a)
gasona u-a-ri-tQa=l-hje hioäa
fuel come-CAU 1 -IRR-3ps/A=ASS=REP PSN
'Juan says that he will bring fuel.'

b)
it$a-kau=he
do-lsg/A=REP
'He says that I did it.'

In examples with 3ps marking plus reportative enclitic, as in (cf. (731a) or

(732a,b)), the form is ambiguous, as it can refer to the same or different subject,
depending on the context. For instance, (732b) can be read as Ά says that Β
went to his field' or Ά says he (himself) went to his field'. Example (732a)
instead, is unambiguous, since the subject surfaces as an N P ('child'). The
reportative strictly refers to a third person source that is not further specified in
the clause. In order to specify the source (e.g. 'Juan said . . . ' ) , a quotative verb
is employed.

(732) Reportative enclitic and verb with 3ps marking

a)
sini-a=he kanaanaj
sleep-3 ps/A=REP child
'He says that the child is sleeping.'

b)
n=ukwana hana=he=te ku-e
3ps=field inside=REP=FOC go-3ps/E
'He says that he went to the field.'

In (733a) the reportative evidential co-occurs with the interrogative marker =na
in a question. Note that the position of =he with content questions is after the
interrogative pronoun (733b); also cf. §2.11). As discussed there, it also occurs
in this position in clauses with an introducer. In (733c), the occurrence of =he is
illustrated as occurring after an imperative form.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
500 Verbal morphology

(733) Reportative in questions and with imperative

a) With interrogative marker:

amiane-naa hita-i=hje=na
work-NOM finish-2ps=REP=INT
'He wants to know whether you have finished the job.' [Lit. 'Have you finished? he
asks.']

b) With interrogative pronoun:

d$u=he ku-i
where=REP go-2ps
'He wants to know where you are going.' [Lit. 'Where do you go? he asks.']

c) With imperative:
turu-ni-u=he=te hjä ii huara-kuru-mu=tQa, na-ι ajto-a hau
arrive-DSTL-IMP=FOCjust 2sg see-PL-JUSS=EMF say-PRT say-3ps/D because

hi turu-a ku-e
there arrive-NTR go-3ps/E
'As he [the novice] said, "They say <He should just arrive> so that that they can see
you", he [the priest] went there.'

In example (734a), the reportative occurs with an adverb in initial position and
is again followed by a focus marker and the emphatic marker =ra. In (734b),
=he is attached to the adverb for 'already'.

(734) Reportative clitic after adverb

a)
eresi=he=te=ra uu-re=i
tomorrow=REP=FOC=EMF come-IRR:3ps/E=ASS
'He says he will come tomorrow.'

b)
ßäe=he=ra sa-a=tge
already=REP=EMF end-3ps/A=PLT
'He says it is finished already.' [Man talking to woman, about the fact that there is no
more fishing line to be distributed]

As briefly mentioned before (cf. §12.2.6, §12.2.14), the remoteness marker =lu
has a double function in that it may refer to the future under certain conditions.
However, the only context in which this is the case is in the adverb floaelu
'earlier'/'forever', where the interpretation depends on the situation and

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 5 01

accompanying words such as 'until' (i.e. jioaelu netohwei is translated as

'forever', with future reference). When =lu occurs as a verbal enclitic, it is
interpreted as past tense in all attested examples. It is only used when a
reference to remote past is intended, i.e. one would not typically use it to refer
to 'yesterday'. However, the interpretation is not restricted to a fixed time frame
and is chosen at the speaker's discretion. In most cases though, the described
event lies back a number of years. The examples in (735) give an illustration of
the different time frames where =lu is used.

(735) Use of remoteness marker =lu

a) Prehistory:
mtoanei hetau=te kaa d^aura+atane huu-hwe=lu lomaj beerte
like.that HRS=FOC this flesh+land flood-3ps/E=REM PSN female
'This is how Lomai flooded the earth (long ago).'

b) Narrator's childhood, ca. 50 years before:

itQafwa kulane=te ne-ure=lu katga-uru
rifle without=FOC be-3pl/E=REM man-PL
'The people lived without guns.'

c) Village foundation, ca. 12 years before:

atane ari-ι, eru-ι, kau turu-a u-akaanu=lu
land seek-PRT find-PRT here arrive-NTR come-lpl/ex=REM
'Looking for land, we found it and came to arrive here.'

d) Several months before:

nahari käu auauako-a=lu=na hau=te, kzuaau-he-ne
liver because.of cry-3ps/D=REM=SUB because=FOC think-CNT-NEG:3ps/E
'He can't remember because he once cried from having eaten it.' [Boy about his brother
who does not want to eat liver, because he had been sick from eating it before].

There is a peculiarity in the ordering of =lu and =he. The examples in (736)
illustrate their behaviour within a word: in (736a), the reportative precedes =/«,
whereas in (736b), the order is reversed, without any apparent difference in
meaning. However, the order in (736a) might be the unmarked variant, as it was
the one offered first when a hypothetical situation between the consultant and
his relatives was described.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
502 Verbal morphology

(736) Variable order between reportative and remoteness markers (elicited)

a) REP > REM
su-ana-i=toe=he=lu=ra
kill-ILT-NEG:3ps/A-PLT=REP=REM=EMF
'He (my in-law) is said not to have killed it (at that time).' [Man talking to woman]

b) REM > REP

su-ana-i=tge=lu=he=ra
kill-ILT-NEG: 3ps/A-PLT=REM=REP=EMF
'He (my in-law) is said not to have killed it (at that time).' [Man talking to woman]

A problem for the description of these two suffixes is that they do not occupy a
position that could be characterised as one semantic or pragmatic function,
since reportative and remoteness are two quite different features. Under
semantic considerations, the reportative would fit better into an evidential slot,
which would include the witness marker =m, but it is obvious from the example
in (737a) that they can occur in a sequence. The form in (737b) illustrates the
position of =he as following the marker for politeness (=tg,e), which was
discussed above. The order in these cases is not variable.

(737) Reportative following other clitics (elicited)

a) WIT > REP:

su-a=ni=he(=tau)
kill-3ps/A=WIT=REP(=REAS)
'He (i.e. another person) killed it earlier today (as someone said).'

b) PLT > REP:

itQa-kau-tQe-he
do-1 ps/A=PLT=REP
'He says that I (should) do it.' [Man talking to woman]

All other clitics (except for the remoteness marker, see above) follow the
reportative. This leads to the conclusion that those should be assigned to
separate positional slots. Further aspects regarding the reportative are discussed
in the following section.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 503

12.3.5 Slot 22: Reassurance

Similar to the evidential markers =m and =he, the reassurance enclitic =tau is
highly optional and rarely occurs as a single marker on a verb. More frequently,
it co-occurs with the evidentials =m and =he. The few examples where it is
attached as a single marker are characterised as a feature of the traditional
language by native speakers. In this case, where =tau is attached to the verb
without other markers, it is interpreted as referring to reassurance and past.
Even though this does not necessarily coincide with having witnessed a
situation, the evidential connotation in these cases is very strong. Judgements on
whether the speaker of a clause with =tau must be seen as a witness are
inconsistent, but this reading specifically seems to apply to examples where
there is a controversy about the truth of a situation, e.g. to convince others of
something whose rightful source they doubt. The reassuring connotation is
again neutralised when =tau is used in combination with the remoteness marker
=lu: in this case, it has the same meaning as it has when occurring by itself (cf.
(738a) and (738c)).

(738) Use of =tau without other clitics in the traditional language

a)
raana su-a=tau
w.l.peccary kill-3ps/A=REAS
'He killed a white-lipped peccary [long ago].' [Referring to the past]

b)
itQa-a=tau
do-3ps/A=REAS

'He really did it.' [Implying: 'Believe me, I know it.']

c)
itQa-a=lu=tau
do-3ps/A=REM=RJEAS
'He did it.' [Referring to the past; Implying: some one is reminding me of what
happened then]
The interaction of =tau with other enclitics leads to interesting modifications of
its functions. In a similar way as with =lu, the reassuring function of =tau is
neutralised when it co-occurs with the reportative evidential =he. The particle
hetau is a lexicalised combination of the two markers, which can be
demonstrated by its syntactic properties. It functions as a hearsay marker
particularly in traditional narratives for which no specified witness is available.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
504 Verbal morphology

The presence of =tan in this form, however, does not add any reassuring
character to it, as it solely emphasises the past character of the situation.
Examples for the occurrence of hetau are found throughout all chapters of this
grammar; an additional example is given in (739).

(739) Hearsay evidential hetau

ßoaelu hetau kwajtel m-ji=lu laano
earlier HRS again be-NEG:3ps/A=REM cassava
'Earlier, again, there was no cassava.' [Talking about the time before creation]

Interestingly, in combination with the witness marker =m, =tau loses its
temporal function and takes on reassuring function. This is illustrated in (740),
where the temporal meaning of =m as 'earlier today' has precedence over the
reference to remote past otherwise implied by =tau. In this case, the presence of
=tau has simply reassuring function. In (740), the witness function of =ni is
reinforced by the presence of =tau, especially when the speaker's statement is
opposed by the listener.

(740) Witness evidential =ni plus =tau

a)
na~i=ni=tau
say-2ps=WIT=REAS
'You told me.' [In reply to the question 'Where do you know it from?']

b)
itQa-si-e-i-ni-tau
do-CPL-NEG-2ps=WIT=REAS
'You did not do it.' [Implied: I have witnessed your failure]

It is also possible to combine the lexicalised meaning of hetau as a hearsay

marker with the witness evidential =ni. This combination describes reported
evidence with reference to 'earlier today' (i.e. A says that he himself saw that Β
did it earlier today). The function of =tau in this case is reassuring, as its
reference to the remote past is overridden by the presence of =m.

(741) Combination of =ni with hetau

itga-a=ni hetau
do-3ps/A=WIT HRS
'He told me that he (himself) saw that Β did it.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 505

The enclitic =tau may also be combined with the remoteness marker =lu. Its
function here is to reinforce a statement with past reference; i.e. evidentiality is
not involved. An example such as itQa-a=lu=tau (do'-3ps/A=RJEM=REAS)
implies that I "know" that he did it, without specifying the source of
information. However, combinations like this appear to be extremely rare and
are only attested in elicitation. Similar to its occurrence in tanni (cf. (730)),
=tau is also combined with =lu to form an adverb.

(742) Adverb taulu

munisipio kahe d^atoam-a taulu ii-tga hvaauku-i

city.council from be.how-3ps/A earlier 2sg-only think-2ps
'How is [the issue you wanted to inquire] from the mayor - you remember from earlier?
[i.e. you remember from what we discussed a while ago]

The table in (743) gives an overview of the enclitics that mark evidentiality or
that are related to it through combination with evidential markers. Recall that
=tau in the traditional language used to be specified as "1 st hand information",
i.e. with connotation to witness.

(743) Evidentials and related markers

l" hand info 2nd hand info Reassuring Same day

(=lu) unspecified unspecified unspecified no
=he no yes no unspecified
=m yes no with Isg yes
=tau unspecified unspecified yes no
=lu=tau unspecified unspecified yes no
=rd=tau yes no yes yes
=hetau no yes no no
-ni=hetau no yes yes yes

12.3.6 Slot 23: Negative, Interrogative

Slot 23 is occupied by two clitics that refer to the grammatical mood of a

sentence. These are the interrogative =na and the final negative marker =ne,
which is used with negative questions and with some introducers that imply
negativity.
The interrogative clitic =na can occur after verbs or after interrogatives such
as 'what' or 'when'. Regarding its position with respect to other enclitics, it can
be followed by attitudinal markers such as =ra and =ta (slot 24). Its occurrence

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
506 Verbal morphology

is restricted to two possible positions: =na may be attached to the first

constituent of the clause (which would be a NP in polarity questions and an
interrogative pronoun in content questions, cf. §21.1), or to the verb, which
most typically is in clause-final position. There are no attested examples in
which the verb would be followed by a subject NP that carries =na. If another
word follows the verb, the enclitic will still be attached to the verb, as shown in
(744d), where 'the ones who are here' functions as a subject NP. Note that in
most examples, there are two alternatives for the position of =na: the unmarked
variant is after the verb, whereas the position after the first constituent implies
focus.

(744) Position of interrogative marker =na

a) After interrogative pronoun:

d$a=na itQa-i ßonei
what=INT do-2ps like.that
'What are you doing like that?' [Unmarked variant: d^a itg,aa-i=ßajioneV]

b) After NP:
ate taba-j=na su-i
fish be.big-NOMSbj=INT kill-2ps
'Did you kill (any) big fish?' [Unmarked variant: ate taba-j su-i=ßa?]

c) After verb:
rai kujjia kana ra-a kana+hoaaunera=ne hau huituku-akatoe=na
that so.that lpl/in receive our.creator=SUB because know-lpl/in=INT
'Do we know that God received us for that purpose?'

d) After verb with subject following:

raj kwitpana ra-uru-a=ne kwaasi-tQuru-a=na kau
POSS blood receive-PL-3ps/D=SUB be.afraid-PL-3ps/A=INT here

ne-j=te
be-NOMSbj=RHT
'Are those who are here afraid to take its blood?'

The final negative enclitic =ne is employed to mark a certain clause type that
implies negativity, i.e. a clause that requires a negative introducer of some kind
(cf. §5.10). Interestingly, the form is not compatible with negative inflection
marking on the verb. As illustrated with negative questions in (745), the verbal
inflection follows the affirmative pattern, rather than the negative paradigm.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 507

Also note that the final negative enclitic is always attached to the last w o r d o f
the clause.

(745) Final negative marker =ne with negative questions

a)
ta nahari heri-a kaa d^uhuhe-j-ße
NEGQ liver want-3ps/A this stand-NOM Sbj =NEGF
'Does this one who is standing [here] not want any liver?'

b)
ta ate ra-uru-a katg,a-uru=ne=ta
NEGQ fish receive-PL3ps/A man-PL=NEGF=FRS
'Did the people not catch any fish?'

T h e final negative marker is also required after the clause introducers for
' d o n ' t ' , mhjauria and hjauipe, which both introduce a prohibitive clause (cf.
§15.2.1).

(746) Final negative marker = m with introducers

a) With mhjauria:
mhjauria be-i=jie
don't tell-2ps=NEGF
'Don't tell it!'

b) With jiaaoiße:
hjauijie hja-i kau=ne
don't urinate-2ps here=NEGF
'Don't urinate here!'

T h e r e are a n u m b e r of examples that s h o w the final negative marker being

followed by an attitudinal marker (slot 24). In (745b), this is illustrated by the
frustrative enclitic =ta, and in (747) by the emphatic marker =ra.

(747) Final negative marker =ne followed by other clitics

nihjauria dada-i-tQe=ne=ra
don't touch-2ps-PL=NEGF=EMF
'Don't (PL) touch him'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
508 Verbal morphology

12.3.7 "Zone " 24: Attitude/Emotion, Rhetorical question

There are four enclitics that describe the "mood" of the speaker or the
atmosphere that relates to a situation, or which describe the emotional attitude
of the speaker. These include enclitics to express "positive" emphasis (=ra),
frustrative (=ta), warning (=naare), and fear (=naate). The analysis of their
occurrence is complex in terms of frequency and actual use: Urarina speakers
characterise the markers =ra and =ta as being part of the traditional language
only and in fact, younger speakers do not make use of them. In traditional
narratives, however, the two clitics are very frequent. In contrast, the markers
=naare and =naate are characterised as being a "normal" part of speech - but
they are infrequent.
One condition for the use of the attitudinal markers is the presence of an
actively listening participant in discourse. In descriptive texts that do not have
an interactive partner in conversation, the forms do not occur.
The morpheme =ra can be characterised as an emphatic marker, such as to
stress the relevance of what has been said, (which could adequately be
represented by an exclamation mark in most examples). Its occurrence is
extremely common in imperatives, where it emphasises the urgency of a
command. In other contexts, it may convey importance with special regard to
truthfulness and positive attitude. Most typically, it occurs in clause-final
position, where it may be attached to a verb or any other word class that
happens to be in this position. The enclitic =ra cannot co-occur with negation.
Also note that it is realised as =t$a when attached to a word that end in /i/.

(748) Emphasis marker =ra

a) Urgency/Seriousness:
hja-rito-anu=ra
urinate-RAP.INT-1 ps/A=EMF
Ί (really) have to urinate!'

b) Truthfulness/Positive attitude:
nekufwelato-akaanu hau aj-akaanu=ra
wonder- lpl/ex because AUX-lpl/ex=EMF
'We did it because we were wondering.' [People in response to why they are sniffing
around the priest]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 509

c) With imperative:
amu-m-u=ra
walk-DSTL-IMP=EMF
'Go away!' [Implied: 'Get lost!']

The frustrative marker =ta can convey various kinds of negative attitudes or
emotions with respect to the mood of the speaker, in order to add an emotional
factor. "Negativity" has a fairly wide pragmatic scope here, which justifies the
term "frustrative". Similar to =ra, the enclitic is not used in the contemporary
language and its occurrences in older narratives is restricted to interactive
discourse. 60 While it frequently co-occurs with negative forms and in questions,
these contexts are not obligatory for the use of the frustrative. The range of
"negative" attitudes includes the following emotions:
- Dislike, criticising a situation
- Dissatisfaction
- Envy
- Anger
- Annoyance
- Negative expectation
- Shock
- Disbelief
- Negative surprise
- Resignation
- Disappointment
- Suspicion

This list is not exhaustive, as the use of =ta with other negative attitudes would
be possible, and some of these may exhibit a certain degree of overlap. In the
following examples, each of the emotions listed above is illustrated. Note that
the position of =ta can be described as being consistently clause-final regardless
the word class occurring at the end of a clause. This is confirmed by numerous
examples where it occurs after verbs, pronouns, or nouns.

(749) Frustrative marker =ta

a) Dislike, criticising a situation:

kauatQa-i kuraanaa ni-a=ne=ta
be.good-NEG:3ps/A chief be-3ps/D=CND=FRS
'It is not good that he is the chief.' [Context: The monkeys criticise the performance of
the spider monkey as their chief.]

60
Note that it is still used in the Corrientes dialect.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
510 Verbal morphology

b) Dissatisfaction:
kauatQa-i=ta, d$a$ri=te d$a barasane ke=te
good-NEG:3ps/A=FRS how.come=FOC what dew INST=FOC

kauatQa-1 omaro-x, kauat^a-i nukuluko-t, omaro-i,

good-PRT grow-PRT good-PRT be.nicely.green-PRT grow-PRT

naaohwa-~i ne-re=l kanaanaj-uru raj

ENUM-PRT be-IRR:3ps/E=ASS child-PL for

temule, itulere temule=ta

plant all.kinds plant=FRS
'It is not good, how, with what dew will the plants, all kinds of plants, grow and become
nice for my children?' [Context: God is unhappy with the fact that the night has not
been created yet.]

c) Envy:
nehvelaj käu, d$a kahe=te katoanei ate taba-j
envy because.of what from=FOC like.this fish be.big-NOM S bj

ru-e kaa kanaanaj manuhoaj kalanoa-ern=ta

find-3ps/E this child damned dirty-DER=FRS
'Full of envy [they said], "Where does that damned dirty kid get the fish from?'"
[Context: The people see a little child catching big fish whereas they did not catch
anything.]

d) Anger:
d^anuna kanu bua basihjau-i=ta
why lsg bag steal-2ps=FRS
'Why have you stolen my bag?' [Context: Furious kinkajou talking after the spider
monkey stole his bag. He then physically attacks the thief.]

e) Annoyance:
d$aha-l=tQa, be-u=tge=ra d$asn=te katoanei ne-i-ta
come.on-PRT-EMF tell-IMP-PLT=EMF how.come=FOC like.this be-2ps=FRS
'Come on, tell, how come you behave like this?' [Context: The people are annoyed by
the woman's behaviour as she does not want to disclose essential information of how
she gave birth.]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 511

f) Negative expectation:
mi hau ku ijiäe mihiito-oru-a hau, netQara
that because there already be.hungry-PL-3ps/D because maybe

wane ni-a kaa hana lenone-ta

PTY be-3ps/A this inside food=FRS
'Therefore, as they were hungry, [they thought], "Maybe there is food [for us poor
guys] in this [tree].'" [Context: Adam and his sons escaped from the flood by climbing
on a tree, where they have to stay for a long time.]

g) Shock:
ku bane+bani-a nii itamtga hana hau, d$a ke=te
there RED+mourn-3ps/E that mud inside because what INST=FOC

neruruhua-ure katoanei kaa katca-uru=ta

use-3pl/E like.this this man-PL=FRS
'As he was mourning in the mud, [Adam] (said), "What have these people done to
him?'" [Context: Adam finds the son of God buried in the mud.]

h) Disbelief:
ku kanaanaj kurwae-ka hau, d$asu=te katoanei itga-i=ta
there child be.happy-3ps/D because how.come=FOC like.this do-2ps=FRS
'As he enjoyed the child like this, he asked, "How is it possible that you have done
this?'" [Context: Women used to be killed when giving birth; this is the first woman to
survive.]

i) Negative surprise:
he, ta ate ra-uru-a katQa-uru=ne=ta
hey/well NEGQ fish receive-PL-3ps/A man-PL=NEGF=FRS
'Hey, have they not caught any fish?' [Context: The priest sends the people fishing and
is surprised that they were unsuccessful.]

j) Resignation:
he, d$aka+ajjiereu=ta, na-ϊ hetau ku kahjune
hey 'why.not?'=FRS say-PRT HRS there clothing

rukuiti-i hetau ku launeto-a

put.off-PRT HRS there sit.down-3ps/E
'Therefore, [Lomai said], "Well, why not?" put off her clothes and sat down.' [Context:
After a useless discussion, Lomai resigns and proceeds, being aware that she will harm
the man who insists on watching her.]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
512 Verbal morphology

k) Disappointment:
ajti-1 tQU kana letoa-j m-a amunaa+hoaaun-era
say-PRTCRTN lpl/in send-NOMSbjbe-3ps/A walk-NOM+create-AG

natu=ta
although=FRS
'[It happened], even though it is the Creator of ways who has sent us'. [Context: The
guide talking in disappointment about the failure of his team.]

1) Suspicion:
d$atoanei=te kuhiutt-ure katQa-uru=ta
how=FOC fish.with.hook-3pl/E man-PL-FRS
'Therefore, [the priest] said, "How have the people fished?'" [Context: Priest, getting
suspicious about the reason why the people did not catch any fish.]

The adversative or "warning" marker =naare is the third member in the class of
clitics that convey attitude. Its functions may indicate a warning, threat, or
advice. This can be in order to prevent an action intended by the addressee, or to
indicate an event that the listener does not expect. As illustrated in (750),
=naare can occur in contexts that imply future or non-future meaning, which is
evident from the context only, but not marked morphologically. In (750a), the
interpretation is future-related, the example in (750b) relates to the past, and the
one in (750c) implies present reading.

(750) "Warning" marker =naare

a) Future reference:
itQa-käu=naare
do-lps/A=WRN
Ί can/will do it.' [Meant as a warning]

b) Past reference:
itga-ri-a=naare
do-RAP-3 ps/A=WRN
'He (unexpectedly) did it quickly.'

c) Present reference:
sini-a-he=naare
sleep-3 ps/A=REP=WRN
'[But] he says that he [his baby] is sleeping.' [Meant as a warning not to wake the baby]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 513

The "warning" function of this enclitic also applies when it is attached to a

noun, such as in akano=mare! ('snake'=WRN) or enua=naare! ('tree'-WRN)
in order to guard someone from an accident. In contrast, when co-occurring
with the imperative form on the verb, =naare takes encouraging function, as to
animate the listener to do without worry what is ordered.

(751) Encouraging function of =naare with imperative

a)
nuhoa raj huitgana, ra-u-tQe=naare
look.here POSS blood receive-IMP-PL=WRN
'Here is its blood, just take it!' [Said passing a bowl of peccary blood, to imply 'why
don't you go ahead and take some?']

b)
lenone-u=naare
eat-IMP=WRN
'[Please] eat!' [Meant to encourage the listener not to be shy]

There is another enclitic, which is almost certainly related to =naare, having a

similar phonological shape, semantic content, and the same position as =naare.
This enclitic, =naate, is equally rare in occurrence, which makes it difficult to
work out the difference between the two markers. The database contains only
two instances of =naate. In these examples (cf. (752)), the context in which
naate is used involves a certain degree of rejection, often combined with fear or
alarm.

(752) "Fear" marker =naate

a) [Context: man arrives and finds his father-in-law in a pool of blood with both eyes
scratched out by a jaguar]:
wa, d^atoane-i=t(x=naate, na-1 turu-1 hoara-a u-e
hey be.how-2ps=PLT=FEAR say-PRT arrive-PRT see-NTR come-3ps/E

raj nakana
POSS son-in-law
'His son-in-law arrived and said: "Hey, what has happened to you?'"

b) [Context: sick man does not reply, so there is fear he might be dead]:
kunajto-a=naate
be.sick-3ps/A=FEAR
'Is he sick?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
514 Verbal morphology

In summary, one can state that the four enclitics =ra, =ta, =naare, and =naate
occupy the same postverbal slot, all having a "mood"-related function. To some
extent, the different emotional states they describe even relate to the
grammatical system, as the morphology distinguishes grammatical categories
that correspond to the ones represented by the enclitics.
1. The emphasis marker =ra does not co-occur with negation and refers to
"reality" in some way.
2. The warning marker =naare and the fear marker =naate, in contrast describe
events that can be described as unreal, in that they either predict something
to happen or transfer the feeling that something unreal or unexpected has
happened already.
3. The frustrative marker =ta deals with the negative side of situations, as it
conveys the negative mood or atmosphere of a situation.

Note that this is only an attempt to relate these clitics to the grammar. From a
morphological perspective, the division does not match: the frustrative occurs
with negative inflection, but can also occur with affirmative clauses. While the
emphasis marker is only observed with affirmative clauses, its use with irrealis
forms is not excluded. The "unreal" warning and fear markers do not coincide
with irrealis marking at all, as it is unclear whether they are ever attached to
irrealis forms.
The rhetorical question marker =te at first sight may not be explicitly related
to attitude or emotion. However, I list it in this "zone" of postverbal markers as
its position is determined by the fact that it occurs after the interrogative clitic
=na (slot 23). Apparently it cannot be followed by any other enclitic, such as
the attitudinal markers, which makes it conceivable to assign it to the same
position. In fact, the use of =te may imply a certain "attitude" in so far as it may
involve curiosity or surprise. This is illustrated in (753a), where the question 'Is
firewood missing?' with the enclitic attached implies that the speaker did not
expect this situation. Example (753b) instead, the speaker already knows that
the listener bought something, but is investigating about the exact nature of the
bought item.

(753) Rhetorical question marker =te

a) Implying surprise:
fwei lana-hee-ka=te
firewood be.missing-DIM-3ps/A=RHT
'Is firewood [really] missing?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Enclitic positions 515

b) Implying curiosity:
d$a kurete-i=te
what buy-2ps=RHT
'What did you buy?'

While occurrences of =te as illustrated in (753) are observed, its use is much
more common in co-occurrence with the interrogative marker =na. The two
enclitics may be combined to =na(a)=te when the clause contains an Ο
argument and a verb only (cf. (754a)), but when the verb occurs with other
constituents, (e.g. a subject or object noun), the two clitics are split. In (754b),
for instance, =na is cliticised to the verb whereas =te is attached to the end of
the clause. Example (754c) differs from this as the interrogative marker occurs
after the first constituent to mark focus. As the number of examples that contain
the two markers is very limited, the exact conditions for the different positions
are unclear. However, =te occurs at the clause end at all times.

(754) Rhetorical question marker co-occurring with interrogative =na

a) Ο + V=na=te:
kartutoo u-a-a=naa-te
shotgun.shells come-CAU 1 -3ps/A=INT=RHT
'Has he brought shotgun shells?' [Implying: Ί am quite sure that he has, since he
always brings some.']

b) DEP + W=na + S=te:

raj hvitQana ra-uru-a=ne ktoaasi-tQuru-a=na kau
POSS blood receive-PL-3ps/D=SUB be.afraid-PL-3ps/A=INT here

ne-j=te
be-NOM sbj =RHT
'Are those who are here afraid to take its blood?'

c) PP=na + V + PP=te:
akauru raj=ßa nekurete-i raana ke=te
3pl for=INT sell-2ps w.l.peccary VLI=RHT
'Are you selling the white-lipped peccary to them?' [Context: Ί know that you are, but
I am not happy about it.']

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
516 Verbal morphology

12.4 Other morphemes in postverbal position

There is a range of further suffixes that can be attached to verbs, as listed in

(755). As indicated in §10,1 will not attempt to distinguish these from others in
terms of "inflectional" vs. "derivational" status. A feature that all suffixes listed
here have in common is their verb-final status, i.e. they are not followed by
other suffixes. Most importantly, all forms are incompatible with person
marking.

(755) Other verbal suffixes

1) Imperative suffix -u (cf. §15.1.1)

2) Hortative suffix -e (cf. § 15.1.2)
3) Jussive suffix -i/ie (cf. § 15.1.3)
4) Jussive suffix -mil (cf. §15.1.3)
5) Distal suffix -ra (cf. §15.1.1)
6) Passive suffix -noi (cf. §16.1)
7) Suffix -Iat 'while' (cf. §20.4)
8) Privative suffix -elanaala (cf. §20.4)
9) Participle suffix -i (cf. §20.3)
10) Neutral suffix -a (cf. § 17)
11) Infinitive suffix -na (cf. §20.2.1)
12) Emphatic suffix -atQa 'only' (cf. (757))

The suffixes 1) to 5) relate to commands. As the otherwise obligatory person

marking does not apply to this clause type, these morphemes do not conform to
most regularities discussed in this chapter. While tense and aspect is unmarked
in commands, some grammatical features (e.g. causative, velocity) can be
marked with imperatives. The distal suffix -m is peculiar as it is not part of the
general inflection system and only occurs with the imperative. The details of
imperative marking are investigated in §15.
The passive forms -wi and -nohwi can be preceded only by the causative
suffixes -erate and -a. The passive is represented as a single suffix here, but it
should be mentioned that it is a morphologically complex construct out of
several components (cf. §16.1). Since the combination of these components is
lexicalised, I treat the two forms as occupying one suffixal position. Note that
-wi (and its variant) represents a nominalised form of the verb. In (756a), the
passivised form is followed by a copular auxiliary (cf. §9.6.6) and in (756b), a
noun plural is added to -wi, which further proves the noun-like status of -noi
and -whzui. In (756c), the passiviser is preceded by a causative form, which
illustrates the relative order between these two suffixes.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other morphemes in postverbal position 517

(756) Passive suffix -mi

a) Passive plus auxiliary:

mtoanei=te kweta-wi jie=lu nukue kuane
like.that=FOC clear-PASS be:3ps/E=REM creek inside
'The (inside of the) creek was cleared like that.'

b) Passive plus noun plural:

karesero kuane ßaara-atQa em-a-nohim-uru
prison inside 2pl-only enter-CAUl-PASS-PL
'the ones that were made to enter the prison by you' [NT: Acts 5:25]

c) Passive preceded by causative:

kauat$a-a=na romano-kuru kuraanaa-kuru kuraanaa efweradoro-atQa
good-3ps/A=INT PSN-PL chief-PL chief emperor-only

baha-erate-nohwi kuriki ti-akaanu=ne

ask-CAU2-PASS money give-lpl/ex=CND
'Is it good that we give the money that is "made to ask" by the emperor of the Roman's
high chief?' [NT: Matthew 22:17]

Another suffix that can be attached to verbs (but also to nouns or other word
classes, cf. §5.3.6) is the emphasiser -atQa 'only'. In (757), some examples of
its occurrence on verbs are given. It appears before the person suffix in
examples (757a), and before the imperative form in (757b). As can be observed,
clitics such as =ra and =ta may follow the emphasiser -atQa. There are few
examples in which the emphasiser is preceded by other suffixes, which makes it
difficult to determine its exact position. The only attested examples contain a
plural object suffix that precedes -atQa (cf. (757b)). A peculiarity of -atQa is
that it can be followed by person marking suffixes, which implies that its
position is to be assigned somewhere between slot 7 and 16.

(757) Emphasiser -atQa with verbs

a)
mihiterua-atQa-a=ra
make.hungry-only-3ps/A=EMF
'He only made him hungry.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
518 Verbal morphology

b)
i=tQuerehe-uru=ra beelaj-atq,a-u=ra
2sg=child-PL=EMF love-only-IMP=EMF
'Just love your children!'

Another suffix that can be attached to verbs is -lai 'while', which is attached to
the neutral form of the verb. It forms a dependent clause with same subject
reference to the main clause and is suffixed to the neutral form of the verb.
Thus, -läi may be analysed as a verb-coordinating device (cf. §20.4). Example
(758b) demonstrates that the suffix is not attached to the 3ps/A suffix but indeed
to the neutral form or the verb. Otherwise, one would expect the clause to
exhibit number concordance (i.e. ku-uru-a-lai 'go'-PL-3ps/A-while). Instead,
the suffix -a is retained, which suggests that its function is not person marking
but an instance of the neutral form.

(758) Suffix -läi 'while'

a)
akauru kaihje ini-a ku-a-lai hetau kami eori ne-t
3pl behind go.up-NTR go-NTR-while HRS sort.of termite.nest be-PRT

hetau ku-atQa neda-atca-e

HRS there-only stay-only-3ps/E
'While she was climbing behind them, she turned into a sort of termite nest and
remained right there.'

b)
hana ehipto kokzoajtQa-kuru akauru kaihje ku-a-lal satii nalu-~i
instead PLN inhabitant-PL 3pl behind go-NTR-while all fall-PRT

u-ure=lu
die-3pl/E=REM
'While they were going after the Egyptians, they all fell and died.' [NT: Hebrews
11:29]

The suffix -elanaala 'without' is attached to the verb as a privative form. The
suffix has the allomorph -lanaala occurring after the vowel /i/, such as
illustrated in (759a), and by forms such as siiri-lanaala 'without having' and
sini-lanaala 'without sleeping'. Otherwise, the initial /e/ of the suffix merges
with preceding vowel. It is one of the suffixes that absorb a preceding vowel /a/
or /e/: /it^a-elanaala/ 'without doing' and /here-elanaala/ 'without wanting'
are pronounced as [itpelanaala] and [herelanaala], respectively (cf. (759b)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other morphemes in postverbal position 519

When the preceding vowels are /«/ or /u/, they are reduced to glides as
illustrated by [kwitak H elanaala] 'without knowing' and [enwelanaala]
'without entering' (underlying /eno-elanaala/; also cf. §3.4). The form for
'without seeing' (underlyingly /kwara-elanaala/) is an exception, as it is
realised as huaa-lanaala. Note that in a few isolated cases, the privative suffix
is also attested on word classes other than verbs. For instance, it is found with
the quantifier for 'all' in satii-elanaala 'without being all'.

(759) Privative suffix -elanaala 'without'

a)
nete akauru ßasi-lanaala amu-e nii hanulari
but 3pl harm-PRV walk-3ps/E that jaguar
'But without harming them, those jaguars went away.'

b)
hjane-na here-lanaala aj/ia ku raasa-ure
leave-INF want-PRV with there dance-3pl/E
'Without letting her go, they danced with her there.'

In most cases, the privative suffix is attached directly to the root. However, it
can also be preceded by certain suffixes that form part of the positional slots
discussed in this chapter. The examples in (760) illustrate this with the causative
and possibility suffixes, respectively. In the NT, the occurrence of -elanaala is
also attested after suffixes for plural object and completive aspect, as in
hiruate-naha-elanaala ('help'-PLO-PRV) 'without helping them' and
w-rauta-si-lanaala (ITR-heal-CPL-PRV) 'without being healed'.

(760) Privative suffix -elanaala following the causative suffixes

a) With causative:
nii rihei rukne-r at-elanaala u-a-re-ü=ni
that like repair-CAU2-PRV come-CAUl-IRR-lsg/E=ASS
Ί will bring it without having it repaired.' [About a gun that was supposed to be
repaired]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
520 Verbal morphology

b) With possibility:
heräe ku-1, tihjaa-ka, heräe nijej suru-elanaala,
slowly go-PRT waIk.slowly-3ps/A slowly not.at.all run-PRV

suru-naka-helanaala
run-PSB-PRV
'They walked slowly, slowly without running at all, without being able to run.'

The functions of the participle suffix -i are further discussed in §20.3. In this
section, only its position in the verb complex is sketched. It can attach to any
verb and in most cases, -7 is attached to the root, as numerous examples
throughout this grammar already prove. However, it is also observed as
following a range of suffixes, as shown below:

(761) Participle suffix -7 following other suffixes

a) Causative-1 (slot 1): u-a-ι ('come'-CAUl-PRT) 'bringing'
b) Causative-2 (slot 2): amian-erate-i ('work'-CAU2-PRT') 'making (them) work'
c) Continuous (slot 4): ku-he-1 ('go'-CNT-PRT) 'going (continuously)'
d) Distributive (slot 7): kura-hakwa-t ('have.name'-DSTR-PRT) 'each having a name'
e) Plural object (slot 7): lurua-naha-7 ('lay.down'-PLO-PRT) 'laying them down'
f) Velocity (slot 8): itga-ri-ι ('do'-RAP-PRT) 'doing'

The neutral marker -a (used after some introducers, in serial verb constructions,
and in citation form) cannot be followed by other suffixes. However, it may be
preceded by a variety of morphemes, such as causative, passive, plural object
and even continuous aspect. This is also discussed in §17 on serial verb
constructions, which investigates the use of the neutral marker in greater detail.
A further suffix to appear on verbs is the infinitive marker -na, which is used
to mark same subject complementation (cf. §20.2.1). Similar to other suffixes, it
can be preceded by other suffixes (cf.(762)). Examples for this include the
causatives -a and -erate (slots 1, 2), the plural object suffix and the distributive
(zone 7), and the velocity suffix (slot 8). However, most commonly, -na is
found as attached directly to the root.

(762) Infinitive marker preceded by other suffixes

a) After causative:
muku-erate-na here-kuru-a=ne sajhjel
catch-CAU2-INF want-PL-3ps/D=SUB although
'even though they wanted to have him caught...' [NT: Matthew 21:46]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Other morphemes in postverbal position 521

b) After distributive:
itQa-hahoa-na najni-a
make-DSTR-INF be.able-3ps/A
'She was able to make one by one.'

In the following, the positions of the markers discussed in this section are
summarised. The table in (763) is an attempt to represent the available data in a
systematic way. However, one must be aware that some gaps still have to be
filled since not every possible combination is attested. T h e order of the
morphemes f r o m top to bottom may be understood as a tentative approach
towards a linear order.

(763) Summary of additional markers

Marker Can be preceded by Can be followed by
(slot)
Infinitive suffix -na CAUl: -a( 1) (none)
CAU2: -erate (2)
CNT -ahe (4)
PLO: -mha (7)
DSTR -ahva (7)
RAP -uri (8)
Passive suffix -nol CAUl - a ( l ) (none)
CAU2 -erate (2)
Neutral suffix -a CAUl -a{ 1) (none)
CAU2 -erate (2)
CNT: •ahe (4)
PLO: -mha (7)
Emphatic suffix -atga 'only' CAUl: -a( 1) Person markers (17)
PLO: -mha (7) (rest unclear)
Suffix -lai 'while' NTR -a (none)
Hortative suffix -e CAUl -a (1) PL -tge{ 17)
DST -ni Clitics: classes 19, 24
Jussive suffix -ipe CAUl -a (1) PL -tge (17)
PL -uru (12) Clitics: classes 19, 24
Jussive suffix -mil CAUl - a ( l ) PL -tge( 17)
PL -uru (12) Clitics: classes 19, 24
Distal suffix -ni CAUl -a (1) PL -tge{ 17)
IMP -« Clitics: classes 19, 24

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
522 Verbal morphology

Marker Can be preceded by Can be followed by

(slot)
Imperative suffix -u CAUl - a ( l ) PL -tge (17)
EMPH 'only' -at$a Clitics: classes 19, 24
RAP -uri (8)
DST -m
Privative suffix -elanaala CAU2: -erate (2) (none)
IMPOS -naka (3)
PASS -noi
PLO -mha (7)
CPL -si (8)
Participle suffix -1 CAUl -a (1) (none)
CAU2: -erate (2)
CNT: -he (4)
PLO -naha (7)
DSTR -ahoa (7)
RAP -uri (8)

12.5 Discussion

Even though person is the only obligatory marking on verbs, the large number
of suffixes that can be combined (with a few theoretical exclusions) makes
Urarina a highly polysynthetic language. To give an overview, the order of
suffixes that were presented in §12.2 is summarised in (764). The table does not
include those suffixes that exclude person marking, as was discussed in §12.4. 61
It also indicates whether the order of the respective slots may vary or not.

61
Recall that -afpa 'only' can occur with person markers, but at this stage, not enough
information is available as to determine its exact position.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 523

(764) Summary of postverbal slots: suffixes

Slot no. Suffix Function Order variability

1 -a Causative-1 No
2 -erate Causative-2 No
3 -naka Possibility No
4 -he Aspect: Continuous May vary
5 -era Impersonal passive No
6 -reheto Aspect: Habitual-1 May vary
-nahauka Aspect: Habitual-2 (unclear)
7 -akwa Distributive May vary
-naha Plural object May vary
8 -uri Velocity May vary
9 -hee Diminutive May vary
10 -si Aspect: Completive No
11 -ana In-law talk No
12 -uru Plural-1 May vary
13 -kaj Probability No
14 -ri Irrealis No
15 (Various) Negation Varies according to
paradigm
16 (Various) Person No
17 -tge Plural-2 No

Suffixes of almost half of the slots are subject to variability in their order.
Systematic deviations apply to the position of the negative suffixes (slot 15), as
the negative marker -ene systematically changes position with the irrealis
marker in certain contexts (cf. §12.2.15). The position of the plural suffix -uru
may change in combination with the habitual Suffix (slot 6). Optional changes
in order are also observed with slots 6, 8, 9, and 10, as is illustrated in table
(765), which represents the compatibility of suffixes. Note that not all possible
combinations are attested (as indicated by question marks); further data is
needed to fill the present gaps. The compatibility of enclitics is discussed
further below.
Suffixes in the leftmost column (vertical) column of the table in (765)
precede the ones listed in the first row (horizontal). In addition, each suffix class
is numbered. The shaded areas indicate the impossible combinations ("no"),
whereas the unshaded fields ("yes") confirm that a combination in the indicated
order is possible. Characteristically, a suffix of a slot that precedes another, (e.g.
class 2 precedes class 3) cannot follow it. For instance, the continuous marker
suffix (4) does not precede the causative markers (1, 2), thus, these fields are

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
524 Verbal morphology

shaded and labelled "no". If there were no variability in the entire system, the
shaded area would cover exactly half of the table, following a diagonal line.
However, there are some deviations that modify this image. In summary, the
following variations apply:
1. The habitual suffix -reheto (6) is subject to position change, but this
behaviour is likely to be a result of other suffixes that move into a position
before it, rather than moving itself. In particular, it can be preceded by the
plural marker -uru (12) and by the plural object marker -naha (7).
2. The position class that follows the habitual aspect slot (6) was characterised
as a "suffix zone" earlier in this chapter. The reasons for this are, firstly, that
the order among the two suffixes that occupy this position is free. Secondly,
both suffixes can affect the position of other suffixes outside their zone by
moving around (but note that there seem to be no semantic or pragmatic
implications involved). However, the behaviour of the two suffixes is not
totally homogeneous: the distributive marker -ahua may shift to the front
and precede the suffix in slot 4, while it is unattested for occurrence before a
slot 6 suffix. In contrast, the plural object marker -naha may occur before
both of these slots.
3. The velocity suffix -uri (8) can appear in reverse order with both suffixes of
"zone" 7.
4. The diminutive form -hee (9) is also observed before suffix zone 7, i.e. before
the plural object and distributive suffixes. Surprisingly, it does not precede
the velocity marker (8). One may either assume that this is due to an active
shift by the diminutive suffix, or a result of a repositioning of zone 7
suffixes. While the first hypothesis seems more likely since all movements
so far were shifts to the front, the latter conforms to the fact that the order
between slot 8 and 9 is stable. Therefore, I suggest that the order RAP >
DIM > PLO/DSTR is due to the shift of class 7 suffixes.

From the position changes described above, one can conclude that there is a
"variable area" between slots 4 and 9, with actively floating morphemes in the
positional slots 7, 8, and 9. The high variability is also portrayed in (765), where
the unshaded fields below the diagonal line of "no"s make this area well visible.
Apart from this, most other variations in order are of a different quality:
5. The plural marker -uru (12) has been attested before the habitual suffix
-reheto (6), but does not occur in deviating position with other suffixes. Its
occurrence with the habitual marker may be characterised as exceptional;
any semantic or pragmatic implications still require further research. It
should be noted that its shift affects the high variability area sketched above,
but does not extend to other positions as these seem to be comparatively
stable.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 525

6. The changes between the irrealis marker -ri (14) and the negative suffix -ene
(15) are of a very paradigmatic nature and cannot be compared to the
variability of other suffixes. In a systematic way, the negative suffix
precedes all irrealis forms for 3pl in the E- and D-conjugation class, whereas
for the Α-conjugation class, the order follows regular parameters.

(765) Restrictions in compatibility of suffixes

["no" = impossible combination; "?" = unattested]
(Left column: suffixes precede those displayed in horizontal order)

Slot /suffix 1 2 3 4 5 6 7 7 8 9
-a -erate -naka -he -era -reheto -akwa -naha -uri -hee
1 -a n/a yes yes yes yes yes yes yes yes yes
2 -erate no n/a yes yes yes yes yes yes yes yes
3 -naka no no n/a yes no yes ? ? 7 yes
4 -he no no no n/a yes yes yes yes yes yes
5 -era no no no no no yes ? ? 7 yes
6 -rehe to no no no no no n/a no no no 7
7 -akiva no no no no no yes n/a yes yes yes
7 -naha no no no no no yes yes n/a yes yes
8 -uri no no no no no yes yes yes n/a yes
9-hee no no no no no no yes yes no n/a
10 -si no no no no no no no no no no
11-ana no no no no no no no no no no
12-uru no no no no no yes no no no no
13 -kaj no no no no no no no no no no
14 -ri no no no no no no no no no no
15 (NEG) no no no no no no no no no no
16 (PS) no no no no no no no no no no
17-fee no no no no no no no no no no

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
526 Verbal morphology

(Table (765) continued)

Slot 10 11 12 13 14 15 16 17
/suffix -si -ana -uru -kaj -ri (NEG) (PS) -toe
1 -a yes yes yes yes yes yes yes yes
2 -erate yes yes yes yes yes yes yes yes
3 -naka ? ? yes yes yes yes yes yes
4 -he no yes yes yes yes yes yes yes
5 -era ? ? yes ? yes yes yes yes
6-reheto no yes yes yes yes yes yes yes
7 -ahoa yes yes yes yes yes yes yes yes
7 -naha yes yes yes yes yes yes yes yes
8 -uri yes yes yes yes yes yes yes yes
9 -hee yes yes yes yes yes yes yes yes
10 -si n/a yes yes yes yes yes yes yes
11 -ana no n/a yes yes yes yes yes yes
12 -urn no no n/a yes yes yes yes n/a
13 -kaj no no no n/a no yes yes yes
14 -ri no no no no n/a yes yes yes
(A)
15 (NEG) no no no no yes n/a yes yes
(E/D)
16(PS) no no no no no no n/a yes
17 -toe no no no no no no no n/a

As a tendency, the order of the clitics is well preserved according to the

numbering of slots. Most evident is the systematic order variation between slots
18 and 19 (politeness and assertive). Furthermore, the enclitics of slot 21 exhibit
some order variation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 527

(766) Summary of postverbal slots: enclitics

Slot no. Clitic Function Order variability
18 -toe Politeness Varies according to paradigm
19 =m Assertive (Future) Varies according to paradigm
20 =m Evidentiality: Witness No
21 =he Evidentiality: Reportative May vary
=lu Remoteness (past) May vary
22 =tau Reassurance No
23 =na Mood: Interrogative No
=ne Mood: Negative question; Prohibitive No
24 =ra Attitude/Emotion: (Positive) emphasis No
=ta Attitude/Emotion: Frustrative No
=naare Attitude/Emotion: Warning No
-mate Attitude/Emotion: Fear No
=te Rhetorical question No

As a general rule, all enclitics follow the suffixes discussed above. There are
seven clitic slots, some of which include several enclitics. Their order is
summarised in (767). An analysis of the relative order of these enclitics reveals
a number of peculiarities, which can be outlined as follows:
1. The distribution of the witness marker =m (20) is in so far restricted as its
occurrence with the remoteness marker =lu (21) is excluded: while =ra can
occur before the reportative =he (which also is in slot no. 21), it is
incompatible with =/«. The explanation for this is by the functions of these
two markers, which exclude each other: while =lu refers to remote past, the
clitic =rn describes an event in the recent past. Note that =rti is also excluded
from co-occurring with any clitic from slot 24.
2. The enclitics of slot 21 can occur in any order with respect to each other,
without any apparent change in meaning. Since the function of these two
enclitics does not coincide at first sight, this slot may be described as a clitic
"zone", similar to the account for some of the suffixes investigated above.
For descriptive purposes, there also is the possibility to assign these two
clitics to separate slots; I have chosen to represent them as occupying one
single slot because their exact position with respect to each other cannot be
determined.
3. The reportative marker =he (21) does not precede the final negative marker
=ne (23) on the verb itself; however, it can be attached to an introducer that
precedes the verb. Some deviations are not well attested: its occurrence
before the politeness marker (19) is observed in one elicited example only.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
528 Verbal morphology

As mentioned in the beginning of this chapter, the clitics following slots 22 are
actually clause clitics. However, since these occur in a systematic order, they
have been included in this discussion.
4. Slot 24 is represented as a single enclitic slot, as four clitics in this position
share functional features: they all describe emotion of attitude to some
degree and they cannot be combined with each other. The fifth member of
this clitic zone (rhetorical question marker =te) is infrequent in the database.
One of its peculiarities is its incompatibility with the assertive marker and
the witness evidential. For some other potential combinations, no
information is available.
5. The combination of enclitics for clause type (slot 23) and attitude (slot 24) on
the verb is subject to restrictions: while examples for the co-occurrence of
=ra or =ta with the final negative marker (which, for instance is found in
negative questions and in prohibitive clauses) are observed, no cases are
attested for their presence with the interrogative marker =na.

Overall, it is evident that the order of enclitics is less subject to variation than
the variability of suffixes. Most exclusions occurring with enclitics are based on
pragmatic incompatibilities.

(767) Restrictions in compatibility of postverbal clitics

["no" = impossible combination; "?" = unattested]
(Left column: suffixes precede those displayed in horizontal order)

Slot /clitic 18 19 20 21 21 22 23 23
=t$e =m =m =he =lu =tau =na =ne
18 =tge n/a yes yes yes yes yes yes yes
19 =«2 (ASS) yes n/a no yes no no yes yes
20 =m (WIT) no no n/a yes no yes yes yes
21 =he no no no n/a yes yes yes no
21 =/« no no no yes n/a yes yes yes
22 -tau no no no no no n/a no no
23 =m no no no no no no n/a no
23 =ne no no no no no no no n/a
24 =ra no no no no no no no no
24 =f a no no no no no no no no
24 =naare no no no no no no no no
24 =naate no no no no no no no no
24 =te no no no no no no no no

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
 Discussion 529

(Table (767) continued)

Slot /clitic 24 24 24 24 24
-ra =ta =naare =naate =te
18 -tge yes yes yes yes yes
19 =ra (ASS) yes yes yes yes no
20 =m (WIT) no no no no no
21 =he yes no yes yes ?
21 =lu yes yes yes yes ?
22 =tau yes yes no no ?
23 =na no no no no yes
23 =ne yes yes no no no
24 =ra n/a no no no no
24 =ta no n/a no no no
24 =naare no no n/a no no
24 -naale no no no n/a no
24 =te ? ? no no no

In conclusion, in the light of the number and variability of postverbal positions

Urarina can be classified as having a polysynthetic morphology. At first sight,
the variability may create a confusing impression of how Urarina suffixes are
ordered. However, the analysis in terms of a "high variability area" for suffixes
accounts for most of these changes. On the one hand it must be stressed that in
practice, the significance of such variation is only marginal for Urarina
morphology, since few verbs in naturally occurring speech combine a large
number of suffixes. The fact that only person marking is obligatory results in
most verbs being marked for a few features only. Otherwise, contexts that
require the combination of a high number of suffixes, such as for completive
aspect, distributive, diminutive - possibly talking about an in-law - are quite
rare. In most of the examples, not more than 3-4 suffixes would be
concatenated, possibly up to Six; forms longer than that would normally have to
be elicited. In addition to this, most verbs also contain enclitics of some kind, as
these cover a wide range of grammatical functions.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:33 PM
13 Politeness

Urarina has a number of ways to express politeness. Unlike many other

languages, where politeness is largely based on situational variation, most polite
forms used in Urarina follow strict patterns that depend on gender and family
relations. Since these relations are not changeable, the use of these polite forms
does not depend on the context. An overview of kinship terms and relations is
given in §13.1 in order to provide a basis for the investigation of polite forms
between in-laws.
One parameter for the use of a politeness marker regards the conversation
between members of the opposite sex and between certain in-laws, as will be
investigated in §13.2. Another factor that requires the use of a polite form is
quite rare from a typological point of view: this form is used when talking about
(but not to) a person of a certain in-law relation (cf. §13.3). In addition to these
two "fixed" parameters, Urarina also has a few strategies for expressing
politeness that are based on the situation. These forms will be discussed in
§13.4. In these cases, morphemes whose general function is to mark other
grammatical categories, such as diminutive and velocity, are employed to
convey politeness on a different level, usually associated with a request.

13.1 Kinship terms

Urarina kinship terms can be divided into descriptive terms, such as 'my
mother' (kamt neb a) and address terms, such as oma 'mother'. These are
numbered for convenience in (768). While most descriptive and address terms
are identical, some deviations occur. Differences between the two terms occur
with no. 4, 5, 10, 11, 13, 17, and 19. Apart from that, with descriptive terms, the
use of full pronouns tends to be preferred, such as in kanu daka 'my wife's
brother' (no. 15), whereas in address terms, the cliticised version is used, which
is the older form and often exhibits clear morpho-phonological differences to
the term in isolation.62 For term no. 15, for instance, the cliticised form is
ka=taa. Other address terms tend to be used without a pronoun whatsoever,
such as aua 'daughter' (no. 17), where the descriptive term is kanu kahtnu.
Note that women use different address terms than men when talking to relatives
no. 3, 6, 8, 9, and 12. For 'my husband's sister' (no. 15b), there is a term that is
only used in conversation among women. The table in (768) is based on a male

62
Some aspects of cliticised and non-cliticised nouns were discussed in §7.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Kinship terms 531

E g o ' s perspective, with the deviating forms for a f e m a l e Ego marked as

" W o m e n " (such as indicated for 3, 6, 8, 12, 21).

(768) Kinship descriptive and address terms

No. Gloss Descriptive term Address term

(Isg)
2 'my brother' (gen.) ka=itgasu (ka=)itQasu
3 'my sister' (gen.) kanu baj ka=hoaj
Women: aua
4 'my mother' kanu neba mama; uma
5 'my father' kam ßaka ofiua
6 'my father's brother' kanu nitoene ka=tgajße
Women: it gene
7 'my father's sister' kanu kaaunu, ka=käu
kanu baj
8 'my mother's brother' kanu tanaa ka=tanaa
Women: nono
9 'my mother's sister' kanu nebaene ka=nemae
Women: fofwa
10 'my brother's son' kanu kalauirihi =2
11 'my brother's daughter' kanu kakunurihi aua
12 'my sister's son' ka=banuhui banui
Women: k-akano
13 'my sister's daughter' = 12 k= akano
14, 'my husband' kanu lana (Name)
1
14 'my wife' kanu komasaj (Name)
15a 'my wife's brother' kanu daka ka=daa
15b 'my husband's sister' kanu daki ka=daki
(term only used among
women)
16 'my son' (gen.) kanu kalaui (Name)
17 'my daughter' kanu kakunu, = 11
ka=kaaunu
18 'my son's wife' kanu akano = 13
19 'my daughter's husband' kanu akana, = 12
kaana
20 'my wife's father' kanu tanaa, =8
kaana

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
532 Politeness

No. Gloss Descriptive term Address term

(Isg)
21 'my wife's mother' kanu tano, ka-tano
kaana Women: ka=nemae
22 'my grandchild' kanu itQuala ka=rtßahaa
23 'my mother's father' ka=rnhja = 22
kanu rißahaa
24 'my mother's mother' kanu daae ka=daae
25 'my father's mother' kanu aaso aaso
26 'my sibling's daughter's = 12 =12
husband'

The kinship relations are further illustrated through diagrams in (769) and
(770). Nodes that have the same kinship terms are numbered identically. Nodes
without any numbering have no corresponding terms.

(769) Kinship tree 1: Consanguineal relationships

23 Δ _ =24 0 23Δ == 25o

Δ Ο Ο Δ ο

2
2 3 2 3 U 2 3 2 3'
Δ Ο Δ Ο Δ Δ Ο Δ Ο Δ Ο
EGO

10 Δ Ι Ι Ο = = Δ 2 6 Δ 1 2 0,3 = Δ 2 6

Δ " Male relative

Ο = Female relative
= = Marriage
» Sibling
I = Descendant/Parent

As visually illustrated in (769), there is only a minimal distinction of terms for

collateral kinship relations that regard Ego's siblings and cousins (no. 2 and 3).
The only distinction made here is in terms of gender. This applies to descriptive
and address terms. Another peculiarity is that there is only one term for
'grandfather' (no. 23), but two terms for 'grandmother' (no. 24, 25), depending
on matrilineal vs. patrilineal descent.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Kinship terms 533

Within affinal relationships (cf. (770)), kinship terms are well specified for
E g o ' s direct descendants (lineal kin), but to a lesser degree for collateral kin.
E g o ' s own grandchildren are not further distinguished f r o m successive
generations of side branches.

(770) Kinship tree 2: Affinal relationships

20 A 21 Ο

EGO (1) Δ === 1 4 ό 15 Δ =|= Ο Ο Δ

18 Ο =-= 16 Δ 1 7 0 Ι9 Δ
2 Δ
11° Δ Ο

Δ 22 Ο Δ 22 Ο Δ 22 Ο Δ Ο

In addition to the relationships illustrated above, there are different terms for
younger vs. older siblings, as shown in (771). On the one hand, the descriptive
terms for younger sister vs. brother are identical; on the other, there is a match
between the address terms for younger and older sister. For male siblings, there
are several alternative address forms occurring in free variation.

(771) Sibling terms

Gloss Descriptive term (lsg) Address term

'my older brother' ka=jflana (ka=)itg,asu,
ka=jpana
'my younger brother' kanu ukvala, ukwala,
k=ukwala k= ukwala,
(ka=)itg,asu
'my younger sister' kanu ukwala, ka-kwaj
k=ukwala
'my older sister' kanu ahitpa, ka=kioaj
k-ahitQa

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
534 Politeness

13.2 Politeness with honorific enclitic

The enclitic =t$e is used in conversations between members of the opposite sex
and between male in-laws. In §12.3.1, it was pointed out that the politeness
marker =t$e is homophonous with the plural marker for second person. Since
the use of 2pl forms as polite forms is wide-spread cross-linguistically, one may
assume that there is a diachronic relation between the two markers. The main
difference between the two is that the politeness morpheme is an enclitic,
whereas the pluraliser has suffix status, which is proven by their different
position in the verb. While the plural suffix occupies slot 17 and can be
followed by the assertive enclitic, the politeness clitic occurs in slot 18 and is
followed by clitics of subsequent slots, but not normally by the assertive marker
(cf. §12.3.2 for exceptions in the paradigm). Under no circumstances, two
instances of =t<^e could co-occur in the same word, which further supports the
view that they have the same origin. The result of their mutual exclusion,
however, is a certain degree of ambiguity: When a verb is marked for 2pl,
politeness is not marked separately and the question whether the utterance is
polite is a matter of interpretation. This is illustrated in (772a), where the
woman Lomai is talking to a group of men. The fact that there are several
addressees requires plural marking; the fact that they are men and the speaker is
a woman requires marking with the politeness clitic. However, =t$e is marked
only once. In example (772b), the ambiguity is resolved by the context: here,
the speaker addresses a group of people, but since the conversation is between
men, =t$e cannot be interpreted as a polite form. Finally, when a verb that is
inflected for 2sg is followed by the politeness marker, it can be interpreted as
2pl, if the assertive marker is not present. This is shown in (772c): from the
context it becomes clear that the conversation is between two persons only and
thus, =t$e cannot be interpreted as plural. Instead, =t$e is used as a polite form
in the conversation between in-laws here.

(772) Ambiguity of =t$e

a) 2pl, polite interpretation from context:
ka=jane-u-tg£=na hauria hja-ru=tg,a
1 sg=let-IMP-PL=FOC: 1 sg first urinate-DSTL:HORT=EMF
'Let me go so that I urinate first!' [Woman talking to a group of men]

b) 2pl, non-polite interpretation from context:

hoaa=he batiri su-a-toe
don't=REP priest kill-NTR=PL
'Don't kill the priest, he says!' [Talking to a group of people]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Politeness with honorific enclitic 535

c) 2sg, polite interpretation from context:

d-z>anu=na eresi hoar a-a ku-ene-i=t£e
why=INT tomorrow see-NTR go-NEG-2ps=PLT
'Why don't you go tomorrow?' [Talking to father-in-law]

The function of =t$e may well be described as honorific, as the addressee is to

be treated with respect. However, the group of people with whom this enclitic is
used is predetermined by gender and kinship relations. As a general rule, a man
talking to a woman - and vice versa - must use =tge, which is usually attached
to the verb. This is exemplified by the examples in (773), which are taken from
a narrative where two girls are rescued from cannibals by their uncle. In (773a),
the girls address their uncle in the polite form, i.e. =t$e is attached to the verb
inflected for lpl/ex. The example in (773b) represents a verbless (copular)
construction. In this case, the politeness marker is attached to the first
constituent, which is the pronoun.

(773) Use of =t$e in woman to man speech

a) (Girls to uncle)
ku bahe mitQauri+mitga-uri-l ku lauta-1 raatiri-1
there calabash RED+fill-RAP-PRT there lay.down-PRT leave-PRT

suru-akaanu=tee=ra, na-uru-a hetau nii ranuna-uru

run-1 pl/ex=PLT=EMF say-PL-3ps/A HRS that girl-PL
"'We quickly filled the calabashes there, put them down and left them, and ran [away]",
said the girls.'

b) (Girls to uncle)
he, kanakaanu=tQe kaa=ra itgene, na-a kn-uru-a
hey lpl/ex=PLT this=EMF uncle say-NTR go-PL-3ps/A
"'Hey, uncle, this is us" they went to say.'

In the same narrative, when the uncle replies, he uses =tge on the first inflected
verb (a/a). The second verb that carries inflection for person (nekajatai) also
contains =tQe, but in this case it is not distinguishable from the plural suffix, as
the speaker addresses two persons.

(774) Use of = t i n woman to man speech

he, aj-a=t$e tQaisi=tQa, üra, d^arihjei=te=ra bakaua baaso

hey AUX-3ps/A=PLT really-EMF ERD how=FOOEMF Indio bad

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
536 Politeness

kahe=ra nekajata-i-tg,e=ta
from=EMF return-2ps-PL=FRS
"'Hey, that's right. How did you return from the bad indios?'" [Uncle to girls]

Based on the available examples in the database, the politeness marker

commonly occurs on the first inflected verb of a sentence. It is not attached to
non-finite verbs such as the participle form. In dependent clauses, its occurrence
is inconsistent and probably optional: Apparently, it tends to be marked on
verbs that are inflected for 2ps, but not on others, as is shown in (775). In
(775a), politeness is marked on both verbs, since the dependent verb directly
addresses the listener with a 2ps suffix. In contrast, =t$e is unmarked on the
dependent verb in (775b), as it contains a l s g form.

(775) Politeness marking in dependent clause

a)
siiri-ki=tQe=ne ka=raj te-u=tge
have-2ps=PLT=CND lsg=for give-IMP=PLT
'If you have [any], give [some] to me!'

b)
siiri-tQäu=ne=na d^e=te-re-u=t^
have-1 sg/D=CND=FOC: 1 sg 2sg=give-IRR-1 sg/E=PLT
'If I had [any], I would give you [some].'

Further instances for the use of =t$e are found in conversations between men
and women in the NT. In (776a), the marker occurs twice; more exactly, =tge is
attached to both inflected verbs: the first occurrence is in the dependent clause
and the second on the verb of the main clause. While this implies that =t$e can
occur on all finite verbs, its presence is not obligatory on all verbs (also cf.
(777c)). The example in (776b) exhibits only one instance of =t$e, but it
contains one finite verb only. The politeness marker here is used even though
the addressee is not mentioned in the utterance, i.e. the speaker makes a general
statement about persons not involved in the conversation ('all'). However, since
this utterance is still part of his conversation with a woman, =t$e must be
present.

(776) Occurrence of =tge in the NT

a) Woman talking to man:
d$anu=na hurju m-i=toe natu, ka=kahe akau baha-i=t$e
why=INT Jew be-2ps=PLT despite lsg=from water ask-2ps=PLT
'Why do you ask me for water, despite being a Jew?' [NT: John 4:9; Woman to Jesus]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Politeness with honorific enclitic 537

b) Man talking to woman:

satii kaa kahe akau ra-ι kuur-era=te herit$atiehei nakwaaunei
all this from water take-PRT drink-AG=FOC as.before again

kukuri-a ku-re=l=toe=ra
be.thirsty-NTR go-IRR:3ps/E=ASS=PLT=EMF
'Whoever drinks from this water shall thirst again.' [NT: John 4:3; Jesus to woman]

A s mentioned above, the use of =t$e between men and women in general is
reciprocal. Every man is expected to use the enclitic when talking to any
w o m a n . On average, a person is regarded an adult at about the age of 15. In
conversation with children, i.e. when the addressee is not regarded as an adult,
the use of =tQe does not apply. However, there are no initiation rites that would
formally mark the transition into adulthood.
A father must address his daughter using =tg,e once she is married or reaches
marital age, (which, again, would be at around 15 years of age). In asymmetry
to this, mothers do not address their sons with =tg>e. The enclitic is also not used
a m o n g spouses, reciprocally.
The honorific enclitic =tg,e is also used in a different context, which involves
conversations between in-laws. It should be explained that this relation is of
particular importance to the Urarina, due to uxorilocal marital customs: when
people get married, (which is usually arranged by the parents, but exceptions
are observed), the man leaves his parents and becomes part of his w i f e ' s family.
This involves moving into the w i f e ' s family's house. Typically, the married
couple will then build their house in close vicinity to the w i f e ' s family, i.e. next
to their house. When marrying a woman from another village, the man leaves
the village of his parents. In addition, the m a n ' s labor is assigned to his new
family and he will work on his father-in-law's field, go hunting with him, etc.
Given these circumstances, it is not surprising that the relationship between
" o l d " and " n e w " family are accompanied by a certain degree of deference,
which is expressed linguistically. Interestingly, the polite form is used only
between male relatives who are in an in-law relationship, as will be further
discussed below. The examples in (777) are part of a narrative where a man
goes hunting with his son-in-law, in accordance with tradition. While the older
man goes to check a trap he had set, his son-in-law goes hunting on the river.
A s expected, =t$e is marked on the imperative forms of the verbs in (777a,b)
and on the 3ps form of the verb in (777c). However, in (777b), the politeness
marker is only attested on the main verb, which suggests that its use on
dependent finite verbs is optional.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
538 Politeness

(777) Use of =tQe between in-laws

a) Son-in-law to father-in-law:
ßäe ii bitoa kwaa-m-u=tQe
already 2sg trap see-DSTL-IMP=PLT
'Already go to see your trap!'

b) Father-in-law to son-in-law:
m-a d$a=ne e7eo-käu=ne tQäe u-u=tge, huru-a
be-3ps/D something=CND shout-lsg/D=CND also come-IMP=PLT snap-3ps/D

bitoa-ne bi-anu=ne, u-u=t$e

trap=CND tell-lsg/D=CND come-IMP=PLT
'If there is something, if I shout, also come; if the trap has snapped, if I tell [you],
come!'

c) Father-in-law to son-in-law:
mi hau ajßa raj nakana raj ßäe huru-a=t$e bitoa, na-a=ne
that because with POSS son-in-law for already snap-3ps/A=PLT trap say-3ps/D

sajhjei, tom-m-a hau

although sound-NEG-3ps/D because
'With that, although he said to his son-in-law "The trap has snapped already", since it
did not sound [i.e. was not loud enough],... [the jaguar came]'

Interestingly, within the context of in-law relationships, women have a slightly

easier stand as they have to address fewer persons with the politeness marker
than men have to, which is already manifested by the fact they do not use =t$e
when addressing their sons. While a man (apart from using =t$e with any
woman, being his in-law or not) must address some male relatives with =t$e,
women never use this marker amongst each other, even if an in-law relationship
applies. For instance, a woman does not address her mother-in-law using =t$e .
A man is expected to use =t$e when talking to his son-in-law, which is
reciprocal, i.e. it applies to no. 19 and 20 in table (768). He is also expected to
use the marker with the husband of his sibling's wife (no. 26). In contrast, the
enclitic is not used in conversation to the son-in-law's father (i.e. father of no.
19), nor when talking to 'my sister's husband' (husband of no. 3).
It must be emphasised that =t$e is not used as a general honorific marker,
e.g. towards persons that may earn respect due to their age or special
achievements. The employment of the politeness marker is strictly limited to
persons as defined above. Some suitable examples to illustrate this fact come
from the NT, where one would expect Jesus to be addressed in a very formal

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Politeness with honorific enclitic 539

and respectful way (at least in certain given contexts). Nonetheless, the example
in (778), where Jesus is approached by a man who treats him very respectfully,
shows that this is not the case and that no honorific marker whatsoever is used
with anybody but members of the opposite sex and with in-laws.

(778) N o use of =tQe in situations that require "respect"

kuraanaa, ii ajjia=ne u-naa here-ϋ, d^unihju ku-i=jie

chief 2sg with=FOC:lsg come-INF want-lsg/E wherever go^ps^SUB

häu manatü
because even
'Master, I want to come with you wherever you go.' [NT: Matthew 8:19]

However, similar to other diachronic developments, the language is under

pressure with regard to the use of the politeness marker =tge. As stated by
younger speakers, the use of =tge is becoming less relevant for them. One
consultant reports that while he is aware of his obligations of where to employ
the marker, he does not usually address his married sister using =t$e, apart from
some exceptional situations where polite interaction may be required ("when I
give something to her"). He also mentions that he feels less obliged to use the
marker with other men, even if they are eligible in-laws. This points at a clear
degradation of the politeness marker, probably being at the initial stage.
Contextual aspects for the exceptional absence of the honorific marker are
also observed in one narrative, where part of the conversation between members
of opposite sex does not involve =tQe. This story describes the encounter of a
group of delegates with the woman Lomai. Their task is to bring the night to
their own world, and Lomai has the power to provide it (also cf. Olawsky
2002:21 Off). During their exceedingly difficult trip to the land of Lomai, they
are trapped in various ways by Lomai or her creatures, which is mainly due to
the delegates' disobedience and carelessness. Lomai herself, who usually
appears as a very polite character, addresses her visitor with =tge, as illustrated
in (779a). 63 In response, the man does not use the marker when talking to his
host. His reply as illustrated in (779a) makes things even worse through the use
of the frustrative marker, which in this case implies his annoyance about
Lomai's request. Since the way he interacts can be characterised as unwise and
arrogant, his failure to use the polite form towards Lomai can be interpreted as a
deliberate act. Thus, the absence of =t$e in this context can be interpreted as a

63
In most examples of this narrative, the presence of =fpe is ambiguous with the plural
marker, as Lomai usually addresses the whole group. This is the only instance in
which she talks to a single listener.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
540 Politeness

stylistic device to illustrate the lack of deference in his speech. (The man
ultimately pays the price for his behaviour, as he is bewitched and changed into
a heron.) The lack in deference towards Lomai is also illustrated in example
(779b), where the people address her without using the politeness marker.

(779) Absence of =t$e on contextual basis

a) Woman Lomai talks politely to man, but not vice versa:

naurutae-ri-u=toe, na-a=ra=ne, dorodoroahari=ra, na-a
move.away-RAP-IMP=PLT say-3ps/D=EMF=CND bathe:HORT=EMF say-3ps/D

lomaj=jie sajhjei hetau=ra, ku lauhua-1,

PSN=SUB although HRS=EMF there sit-PRT

he i=tQuara-ri-a-u=ni=ta
hey 2sg=see-IRR-NEG-lsg/A=ASS=FRS
'When she said "(please) move a way a little. I want to bathe", even though Lomai said
so, he [the visitor], sitting there, [said] "Hey, I won't look at you".'

b) Men do not talk politely to Lomai:

nii lomaj rem turu-Ί, Inn, ii rem=te kanakaana letoa-e
that PL place arrive-PRT hey 2ps place=FOC lpl/ex send-3ps/E

amu-naa+kiuaaun-era=ra
walk-NOM+create-AG
'Arriving at Lomai's place [they said], "Hey, the Creator of Ways has sent us to your
place.'

Further research will be necessary in order to define the exact margins for the
presence and absence of =tg>e. However, the general idea of its use in a strictly
defined, inalterable context, restricted to certain groups of people, should be
rather clear.

13.3 Politeness with in-law marker "-ana"

An additional, obligatory form of expressing politeness again regards the

relationship between in-laws, which once more emphasises the importance of
this particular kinship relation for the language. Specifically, Urarina has a
verbal suffix that is used when referring to - not talking to - certain persons to
which the speaker is linked through certain in-law relationships. The presence
of the suffix -ana (or allomorphs) in a verb automatically implies that this

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Politeness with in-law marker "-ana " 541

conditions applies to the subject of the clause. The example in (780a) is not
directly addressed to the in-law himself, but represents a permission to someone
else to let the in-law in. Note that if a woman were addressed to let the in-law
in, =fpe would be added to the verb. Example (780b) is understood accordingly.
The interpretation as 'my son-in-law' is only known through the context.

(780) In-law suffix -ana

a)
eno-ana-ipe
enter-ILT-JUSS
'He (my in-law) may enter!

b)
kwajtei kohzvamo ari-a ku-akaanu hau leijhti
again the.next.day seek-NTR go-lpl/ex because one

tabiitQa ru-hi-ana-e
finally find-DIM-ILT-3ps/E
'As we went to seek [turtles] again on the next day, he (my son-in-law) finally found
one.'

The suffix -ana is used by men and women in different ways, again with the
women having to make less use of the polite form. For a male Ego, the in-law
suffix is used in five contexts:
a) About my son-in-law (no. 19 in table (768))
b) About my father in-law (no. 20)
c) About my mother-in-law (no. 21)
d) About the husband of my sister's daughter (no. 26 in table (768))
e) About a godfather, as will be explained below.

Remarkably, the in-law suffix is not used when referring to 'my daughter-in-
law' (no. 18 in (768)), 'my wife's brother' (no. 15), 'my wife's son' (no. 12),
nor to 'my sister's husband', for which there is no specific term at all. A female
Ego uses -ana in reference to her son-in-law only, but not about persons
classified as b), c), or d). Thus, the way of application of the in-law suffix is
highly asymmetrical between the two genders.
The relationship mentioned in e) is based on godfatherhood referred to as
compadre or comadre in local Spanish. Principally, a compadre is someone
who assists a woman in giving birth, specifically by cutting the umbilical cord.
Most typically, the woman herself nominates the compadre (otherwise, the
husband may select one). In the same way as with contexts a) to d), the in-law

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
542 Politeness

suffix -ana is used to refer to persons who are in a compadre relationship to the
speaker, subject to certain conditions. From a female Ego's point of view, 'a
man who attended my child's birth' in the above described way is compadre. In
this case, the woman will use the in-law suffix when talking about him.
Secondly, a man will use this form when referring to the husband of a woman
whose baby he helped to be born. As the use is reciprocal, the husband will also
use this suffix to refer to the helper.
In addition to the use of the suffix -ana, there are also certain address terms
to be used among people with a compadre relationship (as explained above):
1. Men who are in a compadre relationship with each other address each other
as kofiua.
2. A man and a woman who are in a compadre relationship address each other
as kaana. Alternatively, the woman may address the man as köfioaere.

Based on this additional context for the use of the in-law suffix, the following
examples can be understood. In one narrative, a very special context applies, as
the relationship between the speaker and the person referred to is not strictly
based on an in-law relationship. In fact, the speaker is a monkey who teaches a
woman how to give birth, (which, according to the story, nobody knew up to
that point, as husbands used to cut their wives' belly in order to deliver a child -
which resulted in the woman's death). While the husband is away in order to
find a bamboo stick that would cut the belly, the monkey functions as the
comadre to the woman who subsequently gives birth to a child and survives. By
gaining the status of comadre, a kind of in-law type relationship is created
between the monkey and the husband. As a result, the monkey uses the in-law
suffix when talking about the woman's husband (cf. (781)).

(781) In-law suffix occurring in special context: monkey talking about his 'in-law'

a)
nuta, ßäe u-hi-ana-a, ßäe u-a raj ßakaana
look! already come-CNT-ILT-3ps/A already come-3ps/A POSS godfather
'Look, he is coming already; his [the baby's] father has already come.'

b)
tg,äe ßäe ne-molo-a raj misi=jκ amu-a ku-ana-iße,
also already ITR-cut-3ps/D POSS umbilical.cord=CND walk-NTR go-ILT-JUSS

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Alternative morphological politeness strategies 543

na-e hetau mi kati

say-3ps/E HRS that black.monkey
"'Also, when his [the baby's] umbilical cord goes off, he [the father] ought to go
hunting", said the black monkey.'

c)
mhjauria tQäe kati saate-nana-a=ne, na-e hetau
don't also black.monkey kill.by.blowgun-ILT-NTR=NEGF say-3ps/E HRS

mi eerie raj
that woman for
"'Also, he [the father] ought not to kill black monkeys", [the monkey] said to the
woman.'

Interestingly, the monkey describes the biological father of the child a s ß a k a a n a

'godfather' in (781a), rather t h a n ß a k a 'father'. The morphological structure of
ßakaana is evident: it is composed o f ß a k a 'father' plus the in-law marker -ana,
which illustrates the close relation between the two.
The notion of godfatherhood is also extended to modern customs such as
religious functions. People who function as godfathers at catholic baptisms are
integrated into the linguistic system of godfatherhood. Thus, a godfather or a
godmother in the catholic sense is referred to by the suffix -ana. Interestingly,
the politeness enclitic =t$e is not used in addressing a godfather or a
godmother, being of the traditional type or in the religious sense.

13.4 Alternative morphological politeness strategies

In addition to the two morphemes that are employed to mark politeness, Urarina
has a few strategies by which politeness can be expressed optionally. The forms
that are used with this purpose usually have other grammatical functions: the
suffix -hee is a diminutive form and the suffix -uri functions as a velocity
marker.

J 3.4.1 Polite function of diminutive suffix "-hee "

The diminutive suffix -hee normally refers to small size of a grammatical

subject, or to its limited physical capacity (cf. §12.2.9). Its meaning can extend
to counterexpectation, e.g. as a surprise about an unexpected action performed
by a small or weak person. The use of -hee in polite function is restricted to
three related verbs that are based on the same root. When attached to the verb

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
544 Politeness

lama 'be missing', the core meaning of the resulting form lam-hee-ka is 'a
little (of something) is missing', but this expression is commonly used to imply
Ί need', as illustrated in (782).

(782) Polite use of -hee with lanaa 'be missing'

lana-hee-ka ka=raj fioanara

be.missing-DIM-3ps/A lsg=for banana
Ί need (some) banana.'

Alternatively, the same sentence as above can be used without the diminutive
suffix (lanaa karaj fioanara) but this would be perceived as rude. However, the
interpretation as a polite function is rather contextual, as the co-occurrence of
lanaa with -hee can also have purely diminutive meaning, as shown in (783).
This sentence is again taken from the narrative where the black monkey
instructs a woman in childbearing. She is told to teach ten other women about it,
but the woman leaves out one, telling it to nine women only. As no direct
interaction between monkey and woman is involved in this situation, the
meaning of -hee is unambiguously interpreted as diminutive ('only one was
missing').

(783) Diminutive function of-hee with lanaa 'be missing'

jiäe hitarü akauru rela-t, ßäe iskö=ne-he-i-tQa itQa-t,

already all 3pl teach-PRT already nine=AUX-CNT-PRT-only do-PRT

lejhn lana-hee-ka=ne hana be

one be.missing-DIM-3ps/D=SUB when tell:3ps/E
'She taught all of them, doing nine [out of ten], [but] when only one was missing, she
told it.'

A verb related to lamia is with nelaatia 'be necessary'. In (784a), its use is
illustrated by an example from one of the Urarina flood stories, where the
woman Lomai floods the earth with her urine. Beforehand, she warns the people
about the consequences, but they disregard her advice and mock at her. This
example shows a more general meaning of 'be necessary' as 'be relevant,
useful'. The sentence in (784b) is a general statement about a need. It can, but
need not be understood as a request from a pragmatic point of view, but would
not be polite in this form; however, if the same example were accompanied by
-hee, its interpretation as a plea would be quite obvious.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Alternative morphological politeness strategies 545

(784) Use of nelaatia 'be necessary'

a)
d$a ke=te nelaate-re nii hjane kulurunoi=ta
what VLI=FOC be.necessary-IRR:3ps/E that urine dirty=FRS
'What for would that dirty urine be necessary?' [i.e. 'relevant']

b)
kuriki=te nelaate-re
money=FOC be.necessary-IRR:3ps/E
'Money would be required [in order to buy building materials for the school].'

A similar reading as a polite request can be recognised in (785), where a worker

requests to be given work. In this case, the polite connotation is implied by
modesty or self-humiliation, literally implying "I am a little bit necessary for
work".

(785) Use of nelaatia 'be necessary' with diminutive -hee to imply politeness
amiane-naa itga-na ke nelaate-hee-käu=ne
work-INF do-NOM VLI be.necessary-DIM-lsg/D=SUB
'... that I am almost worthy to work'fNT: Luke 16:3]

Another verb related to the same lexical root is the transitive verb laatia 'need',
as illustrated in (786). In this example, the verb occurs with its core meaning as
'need', without any connotation of politeness.

(786) Use of laatia 'need'

mjej enamhja laate-kuru-i=lu halai jnaelu ne-nakauru
not.at.all canoe need-PL-NEG:3ps/A=REM very earlier be-those.who
'Those who lived very much earlier did not need canoes at all.'

In (787), the use of laatia with the diminutive is exemplified and, not
surprisingly, in this combination it can refer to politeness. The suffix -hee here
does not mean that the offer of ayahuasca regards a small amount only, but it is
an instance of playing down the listeners' expectations, i.e. the speaker implies
that the listeners would not dare to make a rude request (who, in fact, did not
ask for it at all). However, the distinction between diminutive and politeness
cannot be determined formally, but is only recognised by the context.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
546 Politeness

(787) Use of laatia 'need' with diminutive -hee to imply politeness

hn, tn-a tQu hjä, laate-hee-ki-tQe-ne, hoairi-itQa, na-e

hey be-3ps/A CRTN just need-DIM-2ps-PL=CND ayahuasca-onlysay-3ps/E

hetau nii lomaj

HRS that PSN
"'Hey, there is ayahuasca, if you need some", said Lomai.'

Based on the different examples of -hee with polite functions, one can conclude
that this function only applies to interactive discourse, i.e. when directly
addressing a listener. When merely reporting facts without expecting a reaction
from the listener, the diminutive is used with its actual function.

13.4.2 Polite function of velocity suffix "-uri"

The velocity marker -uri can also take functions as a diminutive and as a
politeness marker. The latter function is restricted to its occurrence with
imperatives. However, even in this context, -uri can refer to rapid speed, as
indicated in (788a), where it describes 'one single cut'. In other contexts, it may
mean 'a little', thus having diminutive meaning.

(788) Imperative form following the velocity marker -uri

a) Velocity:
lejhii mulu-ri-ul
one cut-RAP-IMP
'Cut off one piece!'

b) Diminutive:
amu-atga-u, murutae-ri-u=ra!
walk-only-IMP move.away-RAP-IMP=EMF
'Just go, move a little over here!'

In the examples shown in (789a), the use of -uri involves a polite function in
addition to the diminutive reference. While 'stay a little' is one connotation, this
also implies that it is a cautious, polite request that does not ask too much from
the invited. In (789b), politeness is combined with the velocity function of -uri:
'Quickly try my pet' is an invitation for a meal that is not intended to delay the
guests.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Alternative lexical politeness strategies 547

(789) Combination of functions of -uri

a) Polite and diminutive function combined:

he, hauria ka=kulate-ri-u-tQe=ra, na-1 hetau kauru raj
hey first 1 sg=stay.with-RAP-IMP-PL=EMF say-PRT Η RS 3pl for

ajte mi lomaj
say:3ps/E that PSN
"'Hey, please stay a little with me first", Lomai said to them.'

b) Polite and velocity function combined:

he, iira, ka=iri aua-ri-u-toe=ra, na-1 kauru raj ti-a
hey ERD lsg=pet taste-RAP-IMP-PL=EMF say-PRT 3pl for give-3ps/D

hau
because
'As she gave it [the snail] to them, saying "Hey, well, please try my pet" ...'

As seen above, both functions of -uri are well compatible with politeness, given
an adequate contextual frame. Both imply a certain degree of modesty in
making a request, by which the imperative is significantly "softened".

13.5 Alternative lexical politeness strategies

There are two more ways to express politeness. These involve the use of lexical
items that literally have another function. A feature that is shared by both
strategies is that they occur in combination with imperatives only, where their
use corresponds to the "softening" function of -uri as suggested above.

13.5.1 Polite function of "karaj " for me'

One lexical strategy to imply politeness is to include the postpositional phrase

ka=raj ' f o r m e ' in a request, as is shown in (790). In this case, the PP is an
adjunct of 'take this fish' and has the connotation of 'kindly'.

(790) Use of ka=raj with polite function

a)
karuru-a hau, jiäe=ra, ka=raj mi hiririßo ra-1, kanu
clean-3ps/D because already=EMF lsg=for that bagre.fish take-PRT lsg

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
548 Politeness

lele tooae toku)eta-u=ra

tongue above lay-IMP=EMF
'As he [the man] had cleaned him [the envoy], [the envoy said], (kindly) take this bagre
fish and lay it on my tongue!".'

b)
eresi ka=raj k=ahima kaa enejtgu+suri enoala-u=ra
tomorrow lsg=for lsg=before this monkey+stomach cook-IMP=EMF
'(Kindly) prepare this monkey stomach for me, before I [arrive] tomorrow!'

T h e literal meaning of 'for m e ' is not neutralised in any attested example, i.e.
there is no apparent lexicalisation as a politeness marker. In fact, ka=raj is not
formally distinguishable from an adjunct that would be added in the same way.
For instance, a clause such as 'give it to m e ' involves ka=raj, which does not
necessarily have polite meaning here. However, the polite connotation of ka=raj
in combination with an imperative can be understood as asking a favour in
many situations, always depending on the context.

13.5.2 Polite function of "atu " 'nevertheless'

Another lexical strategy for making a request "softer" and thus more polite is by
the word atii, which the Urarina use to translate the Spanish word for 'please'
{por favor). Its primary meaning is 'nevertheless', which in turn has the literal
translation 'insisting'. Even though no examples for a verb atia 'insist' (whose
participle form would be atii, correspondingly), are attested at this stage, this
etymological origin would be very plausible, as the statement of a request can
involve a certain degree of insisting, especially in the light of original Urarina
culture, in which modesty is a virtue. The examples in (791) first illustrate the
most c o m m o n use of atii, meaning 'nevertheless'. In these cases, the
connotation of 'insisting' is well visible: The sentence in (791a) describes a
scene where people detain the woman Lomai even though she wants to leave.
As there is an ongoing struggle over whether they would release her, the
connotation of 'insisting' is very adequate. The scene continues in (791b),
where, despite L o m a i ' s warnings, the people do not let her go and subsequently
she floods the earth with her urine. The fact that the people continue with their
party can conceivably be described as a way of 'insisting'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Alternative lexical politeness strategies 549

(791) Use of atii as 'nevertheless'

a)
na-a hau hetau atii ku hjane-na here-kur-ene
say-3ps/D because HRS nevertheless there leave-INF want-PL-NEG: 3 ps/E
'As she said so, they still did not want to let her go.'

b)
hi atii räasa-ure ßäe akauru heruri katä uuhual m-a
there nevertheless dance-3pl/E already 3pl waist middle until be-3ps/D

akau hau
water because
'As the water already reached the height of their waists, they [still] danced
nevertheless.'

However, when atii co-occurs with an imperative, its meaning is clearly

employed to convey politeness: by indicating that a request involves 'insisting',
the speaker makes his own command so immodest that the opposite is
understood. In fact, its use in this context does not imply 'by all means', as one
may intuitively want to interpret (based on the meaning of 'insisting'). A more
precise characterisation assumes that the speaker himself takes an inferior
position by being so pretentious.

(792) Use of atii in polite function

a)
d$aha-l=tQa, hitäi kajßa-ku-tg,e ka=raj atü ka=raj=t$a,
come.on-PRT=EMF all return-IMP-PL lsg=for please lsg=for=EMF

na-a=ne asajhjei ku d$una-ure

say-3ps/D although there withhold-3pl/E
'Even though she [Lomai] said "Come on, please give everything back to me", they
withheld it there.'

b)
atü ka=raj lenone te-u
please lsg=for food give-IMP
'Please give me food!'

As both examples illustrated in (792) show, atii can be combined with ka=raj as
a polite form, which reinforces the degree of politeness implied. It should also
be mentioned that the syntactic position of both forms used as lexical strategies

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
550 Politeness

for politeness is quite variable, as atü as well as ka=raj can occur before or after
the verb.

13.6 Summary

A comparison between the different strategies that convey politeness is not

straightforward as they operate on different levels. The markers =tg,e and -ana,
which are employed to talk to or about specific groups of people can be
characterised as general honorific markers. As they reflect a rigid social system,
the speaker has practically no choice on whether to use them or not
(disregarding the fact that the social structure is in a stage of disintegration that
is already softening the system). Therefore, it is not possible to describe their
function in terms of degree of politeness.
This is in contrast to the other morphemes that are employed to convey
politeness. Here, the speaker has the freedom of choice and subsequently, a
comparison of these forms is constructive. As discussed in the previous
sections, Urarina makes use of morphological and lexical strategies in this
regard.
The suffix -hee is used only with a small set of verbs that describe the lack of
something. The diminutive is employed to downplay the actual need of the
speaker, which is conveyed as a hidden appeal. The velocity marker -uri, which
has polite function with imperative forms only, differs from this in that it is
integrated into a frank request. However, the diminutive function also expressed
by -uri is comparable to the function of -hee, as the action requested from the
listener is downplayed to an "easy" task.
The lexical politeness strategies used in Urarina have the connotation of
asking a favour, which is also implied through their use with an imperative.
While ka=raj involves the speaker himself as a recipient who will benefit from
the requested task, atü puts the speaker into the position of an intruder, which is
an expression of self-humiliation.
Further differences between the respective forms are summarised in (793):
the table compares the various strategies by a number of features, which include
the word class with which the respective morpheme occurs (or the grammatical
construction with which lexical strategies are found). Another parameter is the
productivity of the different forms with respect to the question whether a form
conveys politeness with few or many words of a class. 64 . The last comparative
feature regards the polysemy of the different forms: while some forms have

64
The exact degree of productivity for the last three strategies is not well known.
However, I assume that it is not restricted by verb type or similar factors.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Summary 551

e x c l u s i v e polite f u n c t i o n within the c o n t e x t they o c c u r in, o t h e r s are a m b i g u o u s ,

largely d e p e n d i n g on t h e contextual b a c k g r o u n d .

(793) Summary of politeness strategies

Device Primary meaning Co-occurring Productivity Ambiguity

with
=tge PLT, when Verbs, other Not restricted High with 2pl,
addressing specific word classes No otherwise
group (enclitic)
-ana ILT, when talking Verbs (suffix) Not restricted No
about in-laws
-hee Diminutive with Verbs (suffix) Low (three High (DIM)
various functions verbs only)
-uri Velocity + Verbs (suffix) Productive High (RAP)
diminutive
ka=raj 'for me' Imperative Productive High (adjunct)
atn 'nevertheless', Imperative Productive No (with IMP)
'insisting'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
14 Negation

Negativity is a grammatical feature that is articulated through a wide range of

strategies in Urarina. Apart from the prototypical way to mark negation, which
is by verbal suffixes (cf. §14.1), negative concepts can be expressed by lexical
items such as interjections, conjunctions, and introducers. These and other
inherently negative lexemes are investigated in §14.4. The special function of
the copula as an adverb related to negation is discussed in §14.2. The use of
interrogative pronouns with negation to result in indefinite pronouns or negative
quantification is discussed in §14.3. Negation can be marked in any clause type,
i.e. declarative, imperative, and interrogatives. The scope of negation is the
clause, and single constituents cannot be negated.

14.1 Clause negation

There are several allomorphs for the negative suffix, whose realisation is based
on the three different person inflection classes. Their distribution is summarised
here (for the full paradigm, see § 12.2.15):
- For all E-, D-forms, the negative suffix is -em.
- For Α-forms, the negative suffix is -a for lps, -e for 2ps, and -i for 3ps
(which probably is an allomorph of-e, see discussion in §12.2.15).
- In the Α-form, minor deviations occur when the negative suffix is preceded
by the irrealis marker -ri.

Regarding the occurrence of negation in different moods, a few minor

differences between negative forms are observed. The following is a summary
of how negative commands, which have prohibitive character, are formed, (for a
detailed analysis, see §15.2)
1. A prohibitive clause is introduced by nihjanria, hjauiße (or variant
ßaauiße), or hüa, which are clause introducers for 'don't' (also cf. §14.4.4
on the role of introducers).
2. The presence of the introducers mhjauria or hjauiße requires the same
person inflection as in declarative clause, but is only attested with forms of
the Α-conjugation class.
3. Constructions with mhjauria or hjauiße involve the clause-final enclitic =ne.
4. With the prohibitive introducer hoa, all person marking is neutralised and the
neutral suffix -a is attached to the verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Clause negation 553

In questions, further differences to declarative forms apply. This is discussed in

detail in §14.5. In the following, aspects of negativity in the declarative mood
are investigated.

14.1.1 Negation and verb classes

Regarding the distinction between different verb classes, such as transitive vs.
other verb types, no differences for negative marking occur. In the examples
(794) to (796), the occurrence of negative suffixes with all verb classes is
illustrated.
In (794a), the occurrence of negation suffixes is shown with the transitive
verb for 'drink'. Correspondingly, examples (794b-c) illustrate the negation of
active and stative intransitive verbs. Verbs describing posture, shape, or colour
behave accordingly. An example for a negated intransitivised verb is given in
(796a).

(794) Negation of different verb types

a) Transitive verb:
ipono ku-em-a katQa-ne
ayahuasca drink-NEG-3ps/D man=CND
'if a man does not drink ayahuasca'

b) Intransitive verb:
mtoani-a=na hau kuriae ktt-i ruru
be.like.that-3ps/D=SUB because distance go-NEG:3ps/A howler.monkey
'Because it is like that, the howler monkey does not go far [from his place].'

c) Stative verb:
ahißa-i rdi anofiua
be.sharp-NEG:3ps/A that knife
'That knife is not sharp.'

Also, the copula is no exception from regular negation marking. In (795a), the
negation refers to identification ('is not'), whereas in (795b), it marks negative
existential ('is not there').

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
554 Negation

(795) Negation of copula

a) Identity:
kwatia kanu i=tgene letono ni-a-ü=ta
not lsg 2sg=place envoy be-NEG-1 sg/A=FRS
Ί am not the envoy for your place.'

b) Existence:
nukue seti-aka=ne ni-ji ate taba-j
creek fish-lpl/du=CND be-NEG:3ps/A fish be.big-NOM
'When we fished in the creek, there were no big fish.'

Other verbs, which have undergone valency changing mechanisms or which

represent verbs derived from nouns through the (intransitive) verbaliser -oka (cf.
§ 10.1.2), take the same negative suffixes as all other classes.

(796) Negation of derived verbs

a) Verbs with intransitiviser ne-
ne-laati-ji anofwa, saabere ke itga-u
ITR-need-NEG:3ps/A knife machete INST do-IMP
'The knife does not serve, do it with the machete!'

b) Verbs with causativiser:

ena aj-a=ne rihihet kanaanaj hano-a-uru-i=lu
now AUX:3ps/D=SUB like child be.light-CAUl-PL-NEG:3ps/A

eene-kuru jnaelu
woman-PL earlier
'Earlier, women did not give birth to children like now.'

c) Verbs derived from nouns:

teru-k-eni-anu=ne ham ka=raj ham aj=ta
axe-VBL-NEG-1 sg/D=SUB when lsg=for inside AUX:3ps/E=FRS
'Since I do not have an axe, this one is for me.'

In general, negative forms on the verb can co-occur with almost any other suffix
or clitic. Exceptions from this are negative questions, for which specific rules
apply (cf. §14.5) and the attitudinal marker =ra (cf. §12.3.7).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Adverbial function of negative copula 555

14.1.2 Scope of negation

A negative suffix on the verb has the entire clause in its scope. This also applies
to clauses that involve serial verb constructions, as exemplified in (797): only
one negative marker may appear here (marked on the second verb), whereas the
first verb is non-finite and forms part of the same clause. It should also be noted
that NPs or other constituents cannot be in the scope of negation.

(797) Negation of SVC

kam raj bi-a u-uru-i leotca-uru

lsg for tell-NTR come-PL-NEG:3ps/A other-PL
'The others have not come to advice me.'

With participle clauses, the scope of negation can extend to structures that
involve more than a simple clause. This is the case with clause coordinations
through the participle form, as shown in (798). However, this type of negation
is rather infrequent. An alternative strategy is to express the dependent verb as a
subordinate clause with a conjugation, in which case the verb receives full
inflection for polarity and person.

(798) Negation having scope over participle

a)
nerela-t amiani-ji, nijej d$a siiri-ri-ßa=i
learn-PRT work-NEG:3ps/A not.at.all what have-IRR-NEG:3ps/A=ASS
'He does neither work nor study, he will not have anything at all.'

b)
kuriae bahi-tpuhwa kahe amu-i ne-ene ruru
distance defecate-LOC from walk-PRT be-NEG:3ps/E howler.monkey
'The howler monkey does not go far from his toilet.'

14.2 Adverbial function of negative copula

The negative form for 3ps/A of the copula is nijej, usually simplified to ttiji.65
As already shown in (795), this forms functions as the negation for identity and
existence. However, nijej also occurs with other functions, which suggests that

65
The distinction between nijej 'not at all' and ni-ji 'be'-NEG:3ps/A in the transcription
is not based on phonological differences, but in order to distinguish the two
meanings.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
556 Negation

the word is used as a lexicalised form. A very common use of nijej is as an

adverb, which adds emphasis to a negated clause and can be translated as 'not at
all'. This form always co-occurs with negation on the verb and can be used with
any verb (including a copula).

(799) Use of nijej meaning 'not at all'

a)
ipono ku-eni-a kat$a=ne nijej kana+hvaaun-era here-noi
ayahuasca drink-NEG-3ps/D man=CND not.at.all our.creator want-PASS

itQa-i
do-NEG:3ps/A
'If someone does not drink ayahuasca, he does not do God's will at all.'

b)
nii hau nijej beraj-jiaa najp-ene raj komasaj
that because not.at.all care.for-INF be.able-NEG:3ps/E POSS wife
'Therefore, his wife could not look after him at all.'

14.3 Interrogative pronouns and negation

As widely attested cross-linguistically, interrogative pronouns can also function

as indefinite pronouns (cf. Haspelmath 1997). This is also the case in Urarina,
where interrogative pronouns can occur with different functions. While they are
used as interrogatives in content questions, negative quantification is implied
when they occur in a negated declarative clause. Thus, when occurring with a
negated verb, d$a refers to 'nothing' (cf. (800a)). In (800b), negation on the
verb is expressed by the privative suffix -elanaala.

(800) Use of interrogative pronouns as negative quantifiers

a)
d$a kauatQa m-ji=lu
what good be-NEG:3ps/A=REM
'There was nothing good [for hunting].'

b)
nii häu=te d$a ke situ-elanaala turu-ure raj ne-naha
that because=FOC what VLI happen-PRV arrive-3pl/E POSS be-LPP
'Therefore, without anything happening, they arrived at their place.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Interrogative pronouns and negation 557

In the same way, other interrogative pronouns can be negated, such as

illustrated with d^abam 'when', d^u 'where' in (801a-b). In a negated clause,
these refer to 'never' and 'nowhere', respectively. Also note that in (801c), the
quantifier d$uni 'somebody' is used with negation to mean 'no one'.

(801) Other interrogatives in negative clause

a)
d$abana nitoane-j hoara-eni-a=ne baja
when be.like.that-NOMsbj see-NEG-3ps/D=SUB after
'after they had never seen something like that...'

b)
d$u ku-uru-i
where go-PL-NEG:3ps/A
'They have not gone anywhere.'

c)
d^uni eno-naa najni-ji=lu
nobody enter-INF be.able-NEG:3ps/A=REM
'Nobody could enter [the creek].'

This type of negation can be emphasised through the presence of nijej 'not at
all' in addition to the interrogative plus negation on the verb. In fact, examples
of this type occur rather frequently in the database. For instance, when nijej
precedes d$a in a negative clause, the negation is emphasised to mean
'absolutely nothing' or 'nothing at all'. This interpretation suggests that nijej
functions as an emphatic negator with adverbial function. In (802), nijej
modifies the interrogative d^a, which functions as the Ο argument of the
negated verb and is interpreted as 'nothing' through the negation on the verb.

(802) Adverbial use of nijej with negative quantification

nijej kauatQa najpehei nijej d$a hüituku-i kaa itulere

not.at.all pretty.nothing not.at.all what know-NEG:3ps/A this all.kinds

ajrirnra ne-nakauru, helej ne-rtakauru, nijej hoituku-i

outside be-those.who separate be-those.who not.at.all know-NEG:3ps/A
'They knew pretty much nothing at all about these various people who live outside,
about those who live alone; they did not know anything at all.'

The emphatic meaning of nijej is also illustrated with another expression: the
word lejhi-ße-he-ι ('one'-AUX-CNT-PRT) (with some variants in

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
558 Negation

pronunciation) is literally translated as 'being one', and referring to 'only one'

in affirmative clause. In combination with a negative form (cf. (803); (820b))
nijej conveys emphasis.

(803) Negation with lejhTpehel 'not even one'

lejhipehel kauti-ji=lu
not.even.one remain-NEG:3ps/A=REM
'Not even a single one remained [without drinking ayahuasca].'

At first sight, a few examples appear to contain rnjej as negating a noun, which
is shown in (804): In this sentence, rnjej kahjune could be translated as 'no
clothes'. However, since this is a very rare example, it is more likely to assume
that the presence of rnjej refers to the whole negative clause, thus referring to
the verb. The translation as 'they did not wear clothes at all' is therefore the
more precise one, as indicated by the square brackets in (804), which indicate
that the negated constituent involves the noun only, whereas rnjej precedes the
negated noun.

(804) nijej as constituent negator?

jioaelu hetau nijej kauatpa najpehel

earlier HRS not.at.all pretty, nothing

nijej kahjune hja-i katQa

not.at.all [clothing wear-NEG:3ps/A man]

'Earlier, people did not wear any clothes.' [Lit. 'they did not wear clothes at all']

14.4 Negation and the lexicon

Urarina has a number of inherently negative lexemes. Some of these are
interjections for 'no', others are clause introducers with negative reference. In
addition, there is one conjunction that expresses negativity.

14.4.1 Negative introducers

Urarina has a word class of clause introducers, most of which convey negativity
in one way or the other. Three of them (nihjauria, hjauipe, kwa) are used as
prohibitives, kivatia is an emphasising negative introducer for 'not'; nabana
introduces as dependent clause for negative purpose 'so that not' (cf. §14.4.4),

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Negation and the lexicon 559

and ta is an introducer for negative questions (cf. §14.5). For convenience, they
are listed again in (805). Examples to illustrate their use can be found in §5.10.

(805) Negative introducers

nihjauria Prohibitive 'don't' (with 2ps on the verb and final negative enclitic)
hjauiße Prohibitive 'don't', (with 2ps on the verb and final negative enclitic)
kwa Prohibitive 'don't' (with NTR inflection on verb)
kwatia Negative (emphasis) 'not' (with negative inflection on the verb)
nabana Adversative (optional; always co-occurring with huataa 'that not')
ta Negative question, (also cf. 14.5)

Note that with the prohibitive introducers nihjauria and hjauipe and with the
negative question introducer, the final negative enclitic =ne co-occurs in clause-
final position (also cf. §2.11).

(806) Final negative enclitic =ne

nihjauria nekajritQa-i=jie kuriki siiri-i=ße

don't suffer-2ps=NEGF money have-2ps=CND
'Do not suffer if you do not have money!' [Implied: 'but borrow some from me']

14.4.2 Negative interjections

There are several ways to respond to a question or to a command. The

interjection hjauißera is used as a negative reaction to a command or statement,
as illustrated in (807), where a dispute between the people on earth and the
woman Lomai is described. In (807a), the people criticise her intention to leave,
which represents the reaction to a statement. In (807b), her reply hjauißera is
meant as Lomai's refusal to their command to stay on earth. The morphological
structure of hjauißera probably is related to that of nihjauria, which is
discussed in §5.10.1.

(807) Negative reaction with hjauißera

a) [Preceding: "'Let me go; I have to urinate", Lomai said.']

hjauiße=ra, kau-alga aj-atg,a-u=ra, na-ure
no=EMF here-only AUX-only=EMF say-3pl/E
"'No! Do it right here" they said.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
560 Negation

b) [Lomai's reply]
hjauißß=ra, kwatia ka=kahia ka=atane ke
no=EMF not this=this.area this=land VLI

hja-eni-akaanu kanakaanu=ta
urinate-NEG-1 pl/ex lpl/ex=FRS
'No! We do not urinate on this earth [she replied].'

Another expression for ' n o ' in the traditional language is nitokwara, which still
appears to be in use in the Tigrillo dialect. In the only attested example, it is
used with a contradictory function to indicate that the speaker will not comply
with a prohibition. In this particular situation, a man refuses to depart, assuring
Lomai that he w o n ' t watch her while she is taking a bath, after she demanded
that he leave.

(808) Negative reaction with nitokwara

aj-atQa-u=ra, nitokioara, kiuatia i=tQuara-ri-a-ü=ni=ta,

AUX-only-IMP=EMF no not 2sg-see-IRR-NEG-lsg/A=ASS=FRS

na-e hetau nii lejhn

say-3ps/E HRS that one
'[As she said] "Just do so [move away]!", that one said "No! I will not watch you!'"

Apart from the use of a negative interjection, there are two other ways to give a
negative response to a question. One is by replying with the negative form of
the verb that was used in the question; the other involves the intransitive verb
aja 'do/be'. In its most typical form, it is simply given as aji, which corresponds
to the negation of 3ps/A. Literally, this means 'it does not' or 'it is not', but has
a strongly lexicalised character. Alternatively, the verb can be inflected for
negation in the same way as other verbs. In this case, it has the function of an
auxiliary. In (809a), the use of aji as an auxiliary with its literal meaning 'it
does/did not' is given; in (809b), its use as ' n o ' is illustrated, and (809c) shows
the use of the verb with inflection for negative 1 sg future tense.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Negation and the lexicon 561

(809) Response to question with forms of aja

a) Use of aji as auxiliary:

fwatoro ajßa amiane-ι, fiuatorö-tQa here-j itQa-1, naria
patron with work-PRT patron-only want-NOM do-PRT thanks

kana+kwaaun-era kurema ßäe ena aj-i

lpl/in+create-AG for already now AUX-NEG:3ps/A
'Working with the patron, doing only what the patron said, - thanks to God this is not so
any more.'

b) Use of aji with meaning 'no':

kaa ajto-huäu hau auna-i=jia

this say-lsg/D because hear-2ps=INT

aj-i, najpe-re repete-i

AUX-NEG:3ps/A be.able-IRR:3ps/E repeat-2ps
[Speaker 1]: 'Have you heard ("this") what I said?' [Speaker 2]: 'No, can you repeat it?'
(Conversation over short-wave radio) 66

c) Other forms of aja\

nihjauria kati ru-i-jie saate-i=ße,
don't black.monkey find-2ps=CND kill.by.blowgun-2ps=NEGF

na-a häu, aj-ri-a-ü=m

say-3ps/D because AUX-IRR-NEG-lsg/A=ASS
'As she said "If you find a black monkey, don't kill it!' [he replied] "I will not do so".'

T h e negative of third person f o r m s of aja can also be employed as ' n o ' , as

illustrated in (810). In this conversation, which is taken f r o m a conversation
over short-wave radio, the expected reply would involve a 1 sg such as aj-ri-a-ü
( A U X - I R R - N E G - 1 sg/A) to mean Ί would not do that'. Instead, it is given as
aj-ri-ßa-i ( A U X - I R R - N E G - 3 s g / A ) 'it would not b e ' .

66
This example exhibits a number of grammatical irregularities that are typical for
younger speakers. For details on language change, cf. §23.2.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
562 Negation

(810) Use of negative 3ps forms of aja to mean 'no'

mhjauria nekajritQa-i=jie, kuriki siiri-i=ße. - aj-ri-ßa-i,

don't suffer-2ps=NEGF money have-2ps=CND AUX-IRR-NEG:3ps/A

hit aril kaute-naa it kuriki u-a-re-ü=ni

all remain-NOM 2sg money come-CAUl-IRR-lsg/E=ASS
[Speaker 1]: 'Do not suffer [hunger] when you have money!' - [Speaker 2]: 'No, I will
bring [back] all your money that remains.' (to mean Ί will not waste your money')

14.4.3 Negative conjunction

The conjunction kwataa is the negative counterpart to the purposive

conjunction kujßa 'so that' and can optionally be accompanied by the negative
introducer nabana at the beginning of the clause (cf. §5.10.5). However, rather
than implying negative marking on the verb, this conjunction occurs with
affirmative forms only (also cf. §5.8.2). This is demonstrated in (811): In both
examples, the verb is unmarked for negation, but bears regular person inflection
for affirmative polarity. Negativity is only expressed through the presence of
hoataa.

(811) Negative conjunction hoataa

a)
ßaara ke hatal banehelanaala hoaauku-ü mihiite-na-ana
2pl VLI very very.much think-lsg/E be.hungry-NOM-inside

ne-i-tg,e=ne hoataa
be-2ps-PL=SUB so.that.not
Ί am thinking exceedingly much of you so that you would not be in hunger.'

b)
kolektivo kuane kau-a ku-i=jie huataa=na ajto-5
public.bus inside return-NTR go-2ps=SUB so.that.not=FOC:Isg say-lsg/E
Ί said it so that you would not return home by public bus.'

14.4.4 Negation and idiomatic expressions

Similar to the use of interrogatives with negation, there are a few further
expressions that slightly change their meaning when occurring in a negated
clause. The idiom kauatga najjiehei, for instance, is used to mean 'pretty much'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Negation and the lexicon 563

in an affirmative clause. Its meaning may also imply 'very well', or 'entirely',
depending on the context. Structurally, the idiom is composed of kauatQa
'good' plus the continuous participle form of najrna 'be able', whose literal
translation is opaque in terms of its relation to its function. Therefore, it is
legitimate to state that kauatQa najßehei is lexicalised. With negation, it refers
to 'pretty (much) nothing' or 'absolutely nothing', as shown in (812). The
expression is used with adverbial function. A further example is also found in
(804).

(812) Positive and negative use of kauatQa najßehei

a) Positive use:
hitarii kauatQa najpehel kam kahe kaa atane karuru-u=ra
all pretty.much lsg from this land clean-IMP=EMF
'Clean this soil all pretty well from me!'

b) Negative use:
ßäe bati-a, satii hakuru-e, ßäe ni-ji kauatQa najßehei,
already sink-3ps/A all dry-3ps/E already be-NEG:3ps/A pretty.nothing

rn-ji d$ale katQa

be-NEG:3ps/A something man
'It [the water] had already gone away, everything was dry, and there was pretty much
nothing any more, there was no man [left].'

14.4.5 'Not yet' and 'Not any more'

When the temporal adverbs ßäe 'already' and hajti 'still' occur in a negative
clause, they are translated as 'not any more' and 'not yet', respectively.

(813) Negation of hajti and ßäe

a) 'Not any more':

kuraa ke kura-hakiva-t ne-ur-ene kanaanaj-uru
name VLI name-DSTR-PRT be-PL-NEG:3ps/E child-PL

hau rihihel ßäe enene ne-ene

because like already nowadays be-NEG:3ps/E
'As the children are not called by their name, nowadays it does not exist any more.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
564 Negation

b) 'Not yet':
hajti u-em-a elo taba-j=ße hana, raj
still come-NEG-3ps/D rain be.big-NOM Sb j=SUB when POSS

bote fioau-i eno-a

boat descend-PRT enter-3ps/A
'When the big rain had not come yet, he went down and entered his boat ...' [NT:
Matthew 24:38]

14.4.6 'Without'

A s already s h o w n in e x a m p l e (800b), negativity is also expressed by the

privative s u f f i x -elanaala, w h i c h m e a n s ' w i t h o u t ' and is typically attached to a
verb. T h i s is used as a non-finite negative f o r m that m u s t be a c c o m p a n i e d by a
finite verb. In (814b), the privative s u f f i x is attached t o the quantifier f o r ' a l l ' .
A s suggested in §5.11.8, the reason f o r this is in t h e verb-like nature of this
word.

(814) Use of privative -elanaala

a)
ßäe ahe-uru-a=ne hau hoara-elanaala raatiri-ure
already get.drunk-PL-3ps/D=SUB because see-PRV leave-3pl/E
'As they were already drunk, they left her behind without seeing her.'

b)
mtoanei n-ukwana ha-ure ßäe tabauru satii-elanaala
like.that 3ps-field make-3pI/E already some all-PRV
'Already some cultivate their fields like that, [but] not all.'

T h e w o r d kulane ' w i t h o u t ' is a postposition, w h i c h f u n c t i o n s as a negator with

an N P in its scope. Additional e x a m p l e s to the ones in (815) are also f o u n d in
§5.7.4.

(815) Postposition kulane 'without'

a)
mi hau urarißa-uru nii kulane ne-na here-kuru-i
that because Urarina-PL that without be-INF want-PL-NEG:3ps/A
'Therefore, the Urarina do not want to live without this.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Negative questions 565

b)
ruenia ne-naa=te kihja kulane ne
stern be-NOM=FOC paddle without be:3ps/E
'The one who is in the stern is without a paddle.'

This past sections have illustrated that negativity can be expressed in various
ways beside the actual inflection on the verb. However, Urarina does not seem
to have any derivational processes related to negativity (such as 'un-tie' in
English), nor has it negative tags (corresponding to 'didn't he').

14.5 Negative questions

Negative questions are always marked by the clause introducer ta in initial

position. It occurs together with the enclitic =rie, which occurs in clause final
position as a final negative marker. (Note that =ne is also obligatory with the
prohibitive introducers, cf. §14.4.1). It should also be noted that negative
questions do not occur in dependent clauses.

(816) Negative questions with ta

a)
d$atoam-a rdtgae ikito-o, ta u-a elo=ne
be.how-3ps/A over.there PLN-LOC NEGQ come-3ps/A rain=NEGF
'How is it over there in Iquitos? Is it not raining?'

b)
ta ate ra-uru-a katca-uru=ne=ta
NEGQ fish receive-PL-3ps/A man-PL=NEGF=FRS
'Have the people not caught any fish?'

As is evident from the examples in (816), the verb in a negative question takes
the regular affirmative inflection rather than negative marking. Also note that
this clause type always requires the Α-form of the verb, which, however, is a
general feature of polar questions.
The answer to a negative question is given in a similar way as with positive
questions. Most typically, a form of the intransitive auxiliary aja is used. As
illustrated in (817a), the answer to the negative question that preceded ('Have
you not caught any fish?') is negative, which is marked by the negative form
ajaakaanu 'we did not'. A similar case is given in (817b), where the second part
of the reply is optional.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
566 Negation

(817) Negative reply to negative question

a) Negative reply [to question 'Have you not caught any fish?' (816b)]:
aj-a-akaanu=ra, ratato-a niki nete
AUX-NEG-1 pI/ex=EMF pull-3ps/A ADVRS but
'No (we did not) [catch any fish]. It pulled, but...'

b)
ta lenone=na here-i=jie - aj-i, (lenone heri-a-u)
NEGQ food-INF want-2ps=NEGF AUX-NEG:3ps/A food want-NEG-lsg/A
'Don't you want to eat?' - 'No, (I don't want food.)'

In contrast, a positive reply to a negative question, i.e. one that disagrees with
the proposal implied through the negative question, is introduced by eehe, an
interjection normally used in positive reply to a positive question ( ' y e s ' ) . In
(818), this is shown as a positive reply to the first part of (817b), as an answer to
the question ' D o n ' t you want to eat?'.

(818) Positive reply to negative question, using eehe

a)
ta lenone-na here-i=jie - eehe, lenone heri-tQäu
NEGQ food-INF want-2ps=NEGF yes food want-lsg/A
'Don't you want to eat?' - '[But] yes [on the contrary]: I want food.'

b)
ta hajti turu-a=ne - eehe, ßäe turu-a
NEGQ still arrive-3ps/A=NEGF yes already arrive-3ps/A
'Has he not arrived yet?' - '[But] yes, he has arrived.'

It is also possible (but less typical) to omit the interjection eehe and only reply
by repeating the verb of the question marked for the respective polarity. An
alternative way to reply positively to a negative question is by the affirmative
form of aja, as illustrated in (819). Interestingly, not the l s g form (to mean Ί
d o ' ) is chosen, but the 3ps/A form, which literally means 'it is'.

(819) Positive reply to negative question, using aja

ta haböo siiri-i=jie - (eehe), aj-a

NEGQ soap have-2ps=NEGF (yes) AUX-3ps/A
'Don't you have soap [for me]?' - '[But] yes, I do.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Multiple marking of negation 567

In summary, negative questions are answered in a similar way as positive

questions: since the verb in a negative question is inflected for affirmative
polarity, the response involves confirmation or negation of the 'statement' made
in the question by marking the verb used in the reply for polarity. This verb can
be an auxiliary or the same verb as used in the question. The same system
applies to the case where only an interjection is used, i.e. in the case of a
positive reply: originally meaning 'yes', eehe is basically used as a confirmation
of the affirmative polarity of the verb used in the question.

14.6 Multiple marking of negation

It is not possible to actually mark negation twice on the same verb. However, in
the light of different strategies presented above, several of these strategies can
be used within the same clause. The effect of this, however, is not to imply
positivity, but negativity is emphasised through the combination of several
negative strategies. In (820), a number of possible combinations of such
strategies are shown. This list is not exhaustive, as other combinations of the
different strategies discussed in this chapter may be possible.
Example (820a) contains three different strategies of negation, including the
negative introducer, 'not at all', and the negative form of aja. It represents the
statement from the "first contact" situation, where an Urarina who is sent by the
priest assures his companions that the priest is not dangerous at all and that they
should let him live. Example (820b) is taken from the same narrative; here, the
negative marking on the verb is accompanied by mjej and kwatia, preceding the
verb, and two idiomatic expressions that take negative function in a negative
clause (cf. (803), (812)). In (820c), negativity is expressed through the
postposition kulane 'without', which negates the preceding interrogative
pronoun d^a to mean 'without anything'. The negation is emphasised by the
presence of rnjej, which results in the reading as 'without anything at all'.

(820) Different strategies for negation within the same clause

a) Introducer + Adverb + Quantifier + Verbal suffix:
kwatia mjej d^alej ne-i aj-i
not not.at.all something be-PRT AUX-NEG:3ps/A
'It is absolutely nothing.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
568 Negation

b) Adverb + Introducer + Verbal suffix, Idiom + Adverb:

nijej fooatia kahjune hja-uru-i, kahjune hja-naa
not.at.all not clothing wear-PL-NEG:3ps/A clothing wear-INF

ißa-uru-ί, kauatga najpehei, lejhijxhei

know-PL-NEG:3ps/A pretty.nothing not.even.one
'They did not wear any clothes, they did not know clothes, pretty nothing, not even one
[of them].'

c) Adverb + Quantifier + postposition:

niji d$a kulane kanu bihi nehesipahei aj-anu=m hau
not.at.all what without lsg hand pure AUX-lsg/D=SUB because
'because I was without anything at all, with my bare hands'

In summary, one could say that the function of employing several negative
strategies within the same clause is to emphasise the impact of negation, i.e. to
imply "strong" negativity.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
15 Imperative

Urarina has a complex system for marking imperative, which represent a

separate clause type. It is marked by a suffix for positive imperative, which can
be followed by a plural suffix (cf. §15.1). There are special forms for first and
third person imperatives, also called hortative and jussive, respectively. The
hortative requires the presence of a clause introducer under certain conditions,
which is demonstrated in §15.1.2. §15.2 shows that clause introducers are also
used with the prohibitive form, which again incur special types of inflection on
the verb. As an additional feature, there is a distal form, which is only attested
with positive imperatives. The consecutive function of imperatives is discussed
in §15.3. Other factors related to commands are the marking of imperatives for
non-distant future (§15.4) and the suggestive form (§15.5). The differences
between declarative and imperative moods are summarised in §15.7. Intonation
does not function as a relevant feature for the distinction between imperative
and other clauses.

15.1 Positive imperative

15.1.1 2ps imperative

Commands for 2ps are marked by the suffix -u, which is attached to the verb
67
stem. With regard to the phonological properties of this morpheme, it can be
classified as a suffix that forms a separate syllable when combined with a stem
(cf. §3.3). Optional lengthening of the vowel preceding the imperative marker is
common. The realisation of -u as -o or -u is the result of general phonological
regularities that disallow sequences of /u/ and /«/ (cf. §2.2). Another alternation
occurs with verbs that take the 3ps/A suffix -ka (instead of /-a/, cf. §3.1.3): the
imperative of these verbs is realised as -ht.

(821) Imperative allomorphs

a) Allomorph -«:
bute kusipa-u enua+sakari ke
boat close-IMP tree+resin INST
'Close the boat with tree resin!'

67
1 distinguish "root" and "stem": the latter is to be understood as the root plus any
derivational extensions of the root.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
570 Imperative

b) Allomorph -u:
ahajna heloo haa-u
forest towards throw-IMP
'Throw it towards the forest!'

c) Allomorph -ku:
katga-uru kutia-ku
man-PL invite-IMP
'Invite the people!'

There are a few grammatical categories that can co-occur with the imperative.
These include the causative -a (cf. §12.2.1), the emphasiser -atga (cf. §12.4),
the velocity marker -uri (cf. §12.2.8), the distributive and plural object suffixes
(cf. §12.2.7), the plural marker -tg,e (cf. (826)), and the distal form -ni (cf.
(824)). Other categories are not attested with the imperative form. For instance,
there are no examples with the otherwise very frequent continuous form, or with
any other aspect marker. However, the imperative form is often followed by the
final emphatic marker =ra in the traditional language. The examples in (822)
illustrate the occurrence of -u after different kinds of derivational suffixes.

(822) Imperative form following other suffixes

a) After causative -a:

gasona u-a-u
fuel come-CAUl-IMP
'Bring fuel!'

b) After causative -erate:

sw-erate-u
kill-CAU2-IMP
'Send him to kill it!'

c) After emphasiser -atQa 'only':

kau-atea aj-at$a-u=ra
here-only AUX-onIy-IMP=EMF
'Just do it right here!'

d) After plural object form:

kwara-naha-u
see-PLO-IMP
'Watch them [for me]!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Positive imperative 571

e) After velocity suffix -uri:

kuhua-ri-u
give.to.drink-RAP-IMP
'Please give him to drink!' / 'Give him a little to drink!' / 'Quickly give him to drink!'

Note that the presence of the velocity suffix such as in (822e) is also employed
to mark politeness in many cases, when occurring with the imperative. As
discussed in §13.4.2, the velocity function of -uri can coincide with a
diminutive function; however, the distinction between these functions and
politeness is not formally marked but depends on the context. The following
examples involve a diminutive function with a "polite" meaning, which
typically applies in combinations with the imperative. Due to the lack of a
formal distinction, it is difficult to make a clear-cut distinction between
diminutive and polite function. However, cultural values may well be in
accordance to grammatical structure in this point, as modesty is one feature of
the (unwritten) Urarina code of behaviour. Asking for 'a little' will usually be
understood as a modest and thus polite request. In fact, most examples that
involve the suffix -uri must be seen as applications of the diminutive
functioning as a politeness strategy, whereas the interpretation as velocity in
examples such as (822e) is exceptional and only detectable from the specific
context.

(823) Politeness function of velocity suffix -uri

a)
basihti kuak-uri-u
a. while wait-RAP-IMP
'Please wait a little while!'

b)
ka=iri aua-ri-u-tge-ra
1 sg=pet hear=RAP-IMP-PL=EMF
'Please try [to eat] my pet!'

The suffix -rn, which marks the distal form, expresses a directional meaning,
implying that the intended activity will not take place at the present position of
both speaker and listener. For instance, the form lenone-ri-u ('eat'-DSTL-IMP)
'go eat' implies that the listener is not invited to eat at the speaker's place, nor
will the speaker be present at the site of the meal. This automatically excludes
the occurrence of the distal suffix with the verb for 'come', which is in fact
unattested. Further examples are given in (824), all of which imply that the
listener is sent away by the speaker.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
572 Imperative

(824) Imperative form following the distal marker -ni

a)
kauru raj ere-ni-u=ra
3pl for speak-DSTL-IMP-EMF
'Go talk to them!'

b)
ka=kajßamj muku-ni-u=ra
1 sg=older.sister catch-DSTL-IMP=EMF
'Go catch my older sister!'

Note that a combination of several suffixes to occur before the imperative form
is possible, as illustrated in (825).

(825) Combination of suffixes before imperative (elicited)

eno-a-uri-m-u
enter-CAU 1 -RAP-DSTL-IMP
'Quickly go make him enter!'

A noteworthy factor is that the distal form occurs only with the imperative,
/ o

including hortative forms, but not with any other clause type. This may be
taken as supporting evidence for the assumption that imperatives represent a
special clause type as opposed to declarative and other types of clauses.
The plural form of the imperative is identical to the plural used for 2ps forms
in declarative clauses. It is also used as a politeness marker in both imperative
and declarative clauses (but distinguishable from the plural only by context, cf.
§13.2 on politeness).

(826) Imperative plural with -toe

a)
amu-ma+kiuaaun-era raj te-u-tge=ra
walk-NOM+create-AG for give-IMP-PL=EMF
'Deliver him to the Creator of Ways!'

68
The distal form is not normally observed with the jussive, but exceptions may apply,
cf.(l)).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Positive imperative 573

b)
kaa kuane temetura-l sim-u-toe=ra
this inside stand.together-PRT sleep-IMP-PL=EMF
'Stand together in this and sleep!'

A peculiarity of the plural of imperatives was already discussed in §3.1.5: there

is a complex morphological rule for the realisation of -# in combination with
the plural -tQe and the distal -ra. Depending on the number of syllables of the
output and under consideration of possible other factors, the plural form can be
-u-tge, -ku-tQe, or -tg,u-tg,e\ in combination with the distal, the forms -ni-u-tge
and -m-tQU-tQe alternate.

(827) Imperative plural with distal marker

a) -m-u-toe:
lomaj kalaui ra-ni-u-tQe=ra
PSN son receive-DSTL-IMP-PL=EMF
'Go (PL) get the son of Lomai!'

b) -m-tQu-toe:
h-m-tQu-toe=ra
eat-DSTL-IMP-PL=EMF
'Go (PL) eat it!'

15.1.2 Hortative (lps imperative)

Urarina has a hortative form for singular and plural. These forms function as a
strategy to express self-encouragement, but since English lacks this form, the
translation is ambiguous. I will represent the gloss for lsg hortative as Met me'
and for lpl as 'let's', but it is to be understood that this is in no way related to
the verb for 'let'.
As already mentioned in §5.10.4, the hortative form requires the insertion of
the introducer hoane in clause-initial position. This marker is usually
accompanied by a focus marker, which is in concordance with person and
number marking on the verb (cf. §19 on focus). Thus, hoane=m anticipates a
lsg subject Met me'; hoam=m refers to a lpl/ex subject, and kwane=te is the
form for all other persons, including lpl/du and lpl/in. When no focus marker is
used, (which is possible under certain conditions, cf. (831)), the subject of the
clause can still be recognised from the verb. In addition, the verb receives a
hortative suffix. For lsg, the suffix is -e, which is a special form not otherwise
used for lsg in any verbal paradigm.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
574 Imperative

(828) Introducer kwane with hortative form -e

a)
kwane ra-e
let's receive-HORT
'Let me get it!' [Said by character who is about to grab an axe]

b)
kwane=na kau kuhwa-ι kiuara-e-ra
let's=FOC:lsg here wait-PRT see-HORT=EMF
'Here I will wait and see' [Said by character who is hiding in order to watch what will
happen]

All lpl forms (inclusive, exclusive, dual) take the same suffixes as in the
declarative mood (also see the table in (855) at the end of this chapter).
Subsequently, the hortative with lpl forms is only indicated through the
introducer, whereas for 1 sg, a double marking for the hortative function (using
the introducer and the verbal suffix) applies.

(829) kwane with lpl forms

a) With lpl/du form -aka:
kwane=te i=t$ej ate ti-aka
let's=FOC 2sg-for fish give-lpl/du
'Let's give you a fish' [Said by one of two persons present]

b) With lpl/in form -akatge:

kwane=te ku-akatge
let's=FOC drink-lpl/in
'Let's drink!' [Said in a group of people]

c) With lpl/ex form -akaanu:

kwane=ne ii kukwaerati-akaanu saabere ke=ra
let's=FOC:lpl 2sg give.to.drink-lpl/ex sword INST=EMF
'Let us give you to drink with the sword!' [Group of people talking to Jesus on the
cross; NT: Luke 22:49]

The hortative can also co-occur with the distal form -ni. The distal marker
precedes the suffix for person, which leads to the following surface forms:
DSTL-lpl/du: -ni +-aka -> -niaka
DSTL-lpl/in: -ni + -akatQe -nit$aat$e
DSTL- lpl/ex: -ni + -akaanu -niakaanu
DSTL-lsg:: -ni + -e -ni

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Positive imperative 575

The form for lsg exhibits a peculiarity, as the hortative suffix -e is absorbed by
the distal marker. This corresponds to general phonological regularities in the
language: recall that also the 3ps/E suffix -e is realised as zero when it is
attached to a root that ends with /i/ (cf. §3.4).
In principle, it is also possible to combine the hortative form with negation,
as illustrated in (830a). However, note that these examples are based on
elicitation. A preferred and more elegant way to express the same function is
shown in (830b), where the negative suffix is replaced by the privative form and
followed by the copula with the corresponding person marker.

(830) Hortative and negation

a)
kwane=ne itg,a-eni-akaanu
let's=FOC:lsg do-NEG-lpl/ex
'Let's not do it!'

b)
ku>ane=ne itQa-elanaala ni-akaanu
let's=FOC:lsg do-PRV be-lpl/ex
'Let's not do it!' [Lit.: 'Let's be without doing it.']

The examples in (831) illustrate the co-occurrence of the hortative with the
distal form. The directional function of the distal expresses the speaker's
intention of moving away from the site of the utterance to a place far from both
participants. The sentence in (831a), for instance, is part of a conversation
between several persons on an expedition, uttered by one man who wants to do
his necessity. It is obvious that this will happen in a good distance from site of
the utterance. The other example follows suit as it implies the action to happen
in a place distant from both speaker and listener.

(831) 1 sg hortative with distal form -ni

a)
kwane hauria ahaena-ana ku-m-0=tga
let's first forest-inside go-DSTL:HORT=EMF
'Let me go into the bush first!'

b)
huane=te akauru hoaa-ni-aka=ra
let's=FOC 3pl see-DSTL-lpl/du=EMF
'Let's go to see them!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
576 Imperative

In hortative constructions with the lsg form, the use of the introducer is
optional, as the examples in (832) illustrate. The reason for this could be that a
confusion of forms is not possible because the hortative is doubly marked. With
lpl hortative forms, the absence of the introducer is not attested, which is
plausible: without the introducer, the distinction between declarative mood and
command would be neutralised.

(832) 1 sg hortative without huane

a)
d?)aha-u=ne ii letoa-e itQa-i=jie kujßa
come.on-IMP=FOC:lsg 2sg send-HORT do-2ps=SUB so.that
'Come on, let me send you so that you might do it!'

b)
ajane tQU, ka=raaheniane fioau-t huaa-ni-0=tga
yes CRTN lsg=self descend see-DSTL:HORT=EMF
'Yes, I will go down myself and see!'

It should be mentioned that hortative forms may also be preceded by the

causative suffix -a, whereas it is unattested in co-occurrence with other suffixes.

15.1.3 Jussive (3ps imperative)

3ps imperative (or jussive) forms are marked with a special suffix -iße (or -mü
in the traditional language). One difference to 2ps commands is the use of the
plural marker -uru instead of -tg,e. This perfectly corresponds to the difference
in declarative sentences, where -tg,e is used for 2ps and -uru for 3ps. A
difference to the hortative is that the introducer huane is not used with the
jussive, whereas lps forms typically occur with it. The jussive form can be
preceded by suffixes for velocity, causative, and -at$a 'only'; however, these
are attested in elicitation only.

(833) Jussive with -mit

a) Singular:
amu-naa+kioaaun-era te-u-tge=te aka hasasa-mii=tQa
walk-NOM+create-AG give-IMP-PL=FOC 3sg untie-JUSS=EMF
'Deliver him to the Creator of Ways that he may untie him!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Positive imperative 577

b) Plural:
kwane=te lejhü kari ama-aka-te kwara-kuru-mü
let's=FOC one shell take-lpl/du=FOC see-PL-JUSS
'Let us take one shell so that they (the others) would see it!'

The fact that the jussive form -mil is quite infrequent in the database is due to
the fact that it is being replaced by the innovative form -ijie. While the suffix
-mii is a feature of the traditional language, the younger generation tends to
prefer the suffix -ijie. Both forms are less frequent in use than 2ps imperatives,
but the attested examples imply that the two jussive forms share the same
properties and restrictions. Interestingly, the form -ijie coincides with the
prohibitive form for 2sg, which is composed of a 2ps suffix and a negative
enclitic (cf. (387)). Since the relation between these two forms is less than
straightforward, the origin of the suffix -ijie is not entirely clear.

(834) Jussive form -ijie

a)
d$=uhuae tureta-u=te kwara-ijie
2sg=mouth open-IMP=FOC see-JUSS
'Open your mouth so that he may see!'

b)
eresi ka=raj ate kuhwu-a ht-uru-ipe katQa-uru=ra
tomorrow lsg=for fish fish.with.hook-NTR go-PL-JUSS man-PL=EMF
'Let the people go fishing for me tomorrow!'

In one elicited example, a jussive form is attested with the distal marker -ni. In
this case, the suffix -ipe does not surface at all, similar to the combination of
the distal suffix with the lsg hortative -e. However, there is no apparent
phonological reason for the omission of -ijw. It must be assumed that examples
such as the one in (835) are exceptions and possibly the result of elicitation. The
preferred way to express a combination of distal and jussive forms is with a
serial verb construction, such as the one in (834b). Therefore, I conclude that
normally, the combination of distal and imperative forms is restricted to first
and second person commands.

(835) Exceptional combination of distal and jussive (elicited)

d$aha-u=te ii raj ere-rd itgasu
come.on-IMP=FOC 2sg for speak-DSTL.JUSS brother
'Come on and let my brother go talk to you!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
578 Imperative

15.1.4 Lexical imperative with "d$ahaa "

There is another strategy to refer to encouragement or command, which should

be mentioned here. The verb d ^ a h a a has an inherent imperative/hortative
meaning associated with 'come on'. Its inflectional paradigm is extremely
limited as it occurs only with the 2ps imperative (d^aha-u) and with the
69
participle form (d$aha-l). Interestingly, the form of the verb following
d^aha-u is in the hortative mood, i.e. -e in the singular and the lpl forms for
plural. Thus, the use of d^aha-u in this context has the same function as the
clause introducer hvane. The difference to the declarative mood is evident from
as the use of the distal marker is possible.

(836) Use of d$ahaa

a) Singular:
d$aha-u=te ka=reruhwari-0
come.on-IMP=FOC lsg=replace-HORT
'Come on and replace me!'

b) Singular with distal form:

d^ahaa-u=te hjä kwaa-ni-0=tQa
come.on=FOC just see-DSTL:HORT=EMF
'Just come here and let me see!'

c) Plural:
mi hau hetau, hoo, d$aha-u=te lenone-tpaatQe
that because HRS oh! come.on-IMP=FOC eat-lpl/in
Therefore [he said], 'Oh, come on and let's eat!'

d) Plural with distal form:

d$aha-u-tQe=te fovajtei hodea ku-m-tgaatoe
come.on-IMP-PL=FOC again PLN go-DSTL-lpl/in
'Come on and let's go to Judea again!' [NT: John 11:7]

A possible analysis for these examples is to assume that they actually consist of
two commands: one given as a direct (2ps) command or invitation to come; the
other as an expression of the hortative which implies a promise or indication of
an action to be performed by the speaker in the presence of the listener. " C o m e "
may well be interpreted in a figurative sense.

69
There also is a derived (non-transparent) form d^aha-ma-u 'come on'-IMP, which
can be employed alternatively.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Prohibitive 579

15.2 Prohibitive

Urarina has different prohibitive forms, which are distinguished according to

the urgency or strength of the prohibition, hjauijie (with variant ßaauijie) and
mhjauria (or variant hjauria, occurring in the NT) are "strong" prohibitive
introducers, referring to an absolute prohibition, as the prohibited action will not
be allowed under any circumstances at any time. The third type of prohibition,
with the introducer hua, refers to a temporary, "weak" prohibition.
Both strong prohibitive introducers represent a very similar degree of
prohibition, but hjauijie appears to be the variant expressing a higher degree of
urgency. The difference can be illustrated through the way speakers handle
prohibitions in Bible translation: A command with mhjauria as in (837b) is
strong and not temporary, as the demon addressed in the command is strongly
ordered to leave his victim alone. However, the ten commandments, which can
be considered to have an even stronger connotation of absolute prohibition are
translated using the introducer hjauijie to express an absolute ban without
exception (cf. (837a)). The command in (837c), instead, is of a much weaker
nature as it only refers to the present situation.

(837) Different degrees of prohibition in the NT

a) With hjauijie:
hjauijie asihjau-i=jie=ra
don't steal-2ps=NEGF=EMF
'Do not steal!' [Biblical commandment quoted by consultant]

b) With hjauria:
hjauria nahoaaunel kaa enamanaa suuhua kuane ew-i=ße
don't again this young.man heart inside enter-2ps=NEGF
'Don't enter the heart of this young man again!' [NT: Mark 9:25; Command to a
demon]

c) With kwa:
kwa kanakaanu rene turu-1 kanakaanu rerueka-a
don't lpl/ex place arrive-PRT lpl/ex disturb-NTR
'Don't disturb us by coming to our place!' [NT: Matthew 8:29; Demons talking to
Jesus]

All introducers occur with a special kind of marking on the verb. However, the
ways of morphological marking differ from each other, as will become clear
below. All prohibitive forms can take plural marking on the verb. Since this is
indicated by the suffix -t$e, there is ambiguity in that this form can also refer to

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
580 Imperative

politeness; the difference is only recognizable by the context. Prohibitive forms

are mainly attested with 2ps subjects; however, there also are some examples
for mhjauria occurring with 3ps referents (cf. (840)).

15.2.1 Prohibitives with "mhjauria " and "hjauijie "

As mentioned in chapter 5.10.1, the introducer mhjauria might be derived from

the verb for 'throw away' (haua), i.e. from its detransitivised form mhjaua ('be
thrown away', with connotation to 'die'). It is plausible to assume that hjauijie
in some way refers to the same root. One of the characteristics mhjauria and
hjauijie in a prohibitive construction is the co-occurrence with the clause-final
negative enclitic =ne. With 2ps, the verb receives the same person and number
marking as declarative sentences, which results in the singular form -i followed
by =ne, (which is realised as [jie] after /i/, as a result of the general
palatalisation rule).

(838) Strong prohibitive introducers with 2sg verb marking

a)
ka=raj fwosokoro u-a-u, mhjauria ahka bate-i=jie
lsg=for matches come-CAU 1 -IMP don't 3sg forget-2ps=NEGF
'Bring me matches, and do not forget them!' [In conversation over radio]

b)
mhjauria leotga amuriti-i w-naa kioaauk-i-pe
don't other surpass-PRT be-NOM think-2ps=NEGF
'Don't believe that you are better than others!; [Given as a directive in a sermon]

In the plural, the resulting form is -i-t$£ (or -i-tQe=ne if the verb is in clause-
final position).

(839) Strong prohibitive introducers with 2pl form

a)
mhjauria hauria dada-i-t$e=ne=ra
don't first touch-2ps-PL=NEGF=EMF
'Don't (PL) touch him yet!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Prohibitive 581

b)
mhjauria hoajtei nahoaaunei netabatakahe-ϊ turu-i-t$e=ne=ra
don't again again be.reduced-PRT arrive-2ps-PL=NEGF=EMF
'Don't arrive reduced in number again!'

Similar to the verb marking for 2ps, with 3ps subjects, the forms of the
paradigm for declarative sentences are employed with the prohibitive. More
exactly, the suffix for 3ps/A is used in this context. As before, the final negative
enclitic =rie is attached to the end of the clause.

(840) Use of mhjauria with 3ps form

a)
mhjauria su-a-ne
don't kill-3ps/A=NEGF
'He ought not to kill it!'

b)
hjauria fwau-1 raj rukuele ra-naa hoaau-ka=ne
don't descend-PRT POSS things receive-INF think-3ps/A=NEGF
'He ought not to consider coming down to take his things!'

15.2.2 Prohibitive with "kwct"

Prohibition with the introducer koa involves neutral marking on the verb, as the
examples in (841) illustrate. Even though no examples for 3ps are attested, the
fact that 2ps is not explicitly marked indicates that person marking is
neutralised in this form.

(841) Use of kwa with 2sg meaning

a)
kwa kuruata-sa ti-a
don't two-times give-NTR
'Don't tell it twice!'

b)
kwa ka=su-a i=t$ej biri bi-anu=ne kujßa=ra
don't lsg=kill-NTR 2sg=for piripiri telI-lsg/D=SUB so.that-EMF
'Don't kill me and I show you piripiriV

In the plural form, -Ιςχ is suffixed to the verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
582 Imperative

(842) Use of lava with 2pl form

a)
kwa ka=su-a-tQe=ra, kanu=na hau-naka-ito-o=ra
don't 1 sg=kill-NTR-PL=EMF lsg=FOC:lsg throw-PSB-DER-1 sg/E=EMF
'Don't (PL) kill me; I am not dangerous!'

b)
kwa ubaae-ka-tge
don't be.angry-NTR-PL
'Don't (PL) get angry!'

In the table in (843), the use of different introducers and verbal inflections is
summarised again. One very obvious aspect is that the use of the distal form is
very restricted: neither does it occur in the declarative mood, nor is it attested
with 3ps forms. Its exclusive occurrence is with commands for first and second
person.
With regard to the distinction between declarative and imperative mood, the
division is not clear-cut: In the positive imperative, there is a special marking
for all forms including lsg, but not for lpl forms. 2ps commands have special
verbal inflection for the positive imperative, but not for the prohibitive. The
system is further complicated by the involvement of the neutral form with the
prohibitive introducer hoa.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Consecutive function of imperative 583

(843) Summary of imperative forms

Positive IMP Positive-IMP- Strong Weak
DSTL Prohibitive Prohibitive
Isg Introducer (huane)=na (hoane)=rw — —

Verb -e -m — —

lpl/ex Introducer hoane=ne kioane-ne — —

Verb -akaanu -m-akaam — —

lpl/du Introducer hoane=te hoane=te — —

Verb -aka -m-aka — —

lpl/in Introducer hoane=te hoane=te — —

Verb -aka-tge -m-tQaatoe — —

2sg Introducer — — mhjauria, hoa

hjauijie
Verb -« -ra'-» -i=pe (NEGF) -a (NTR)
2pl Introducer — — mhjauria, hoa
hjattiße
Verb -u-tge -m-u-tQe -i-tQe-pe -a-tge (NTR)
3sg Introducer — — mhjauria —

Verb -mit (OLD) — -a=ne —

-ijie (NEW)
3pl Introducer — — mhjauria —

Verb -urtt-mii (OLD) — -nru-a=ne —

-Hm-iße (NEW)

15.3 Consecutive function of imperative

The imperative also has an additional function in relation with syntactic

structures. As was mentioned earlier, the imperative form does not occur in
dependent clauses. However, a clause that follows an imperative form may be
interpreted as a consecutive (dependent) clause referring to a direct result of the
action expressed by the imperative. In (844a), the imperative is followed by a
subordinate clause that contains a consecutive conjunction (kujjia 'so that'); so
the consecutive function could be explained by the type of dependent clause.
Similarly, the dependent clause in (844b) contains an (archaic) form 'let there
be', which can be viewed as a command for 3ps.70 Again, this result would only
be achieved if the demand expressed by the imperative clause is fulfilled. The
sentence in (844c) follows suit, as it also contains a jussive form in the

70
The root of kulümii is not analysable. However, the ending of the word in -mti
suggests that this is a jussive form.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
584 Imperative

dependent clause, with similar propositions (only if the priest is not killed, he
could arrive).

(844) Consecutive function with imperative

a)
bute ha-u=te kuane nitQute-i-tge=ne ku\pa=ra
[boat] [make-IMP=FOC] [inside escape-2ps-PL=SUB so.that=EMF]
[Ο] [V] [DEP]
'Build a boat so that you (PL) would escape in it!'

b)
ku-rti-u=te satit kanaanaj-uru raj kulümü
[go-DSTL-IMP=FOC] [all child-PL for let.there.be
[V] ([DEP])

itulere nukue-knru kuane

various creek-PL inside]
[DEP]
'Go and let there be all kinds [of fish] in the creeks for all my children!'

c)
kwa=he batiri su-a-tge=te turu-a u-ipe=ra
[don't=REP] [priest] [kill-NTR-PL=FOC] [arrive-NTR come-JUSS]
[INTR] [Ο] [V] [DEP]
'Don't kill the priest and let him arrive, he says!'

A possible explanation for this structure lies in the nature of the imperative as
an action yet to be accomplished. Whatever is the result of the required action
depends on whether the command is obeyed or not. Instead of using the
declarative form in the dependent clause, Urarina uses a form which implies
that the action still is to be performed. There are two strategies in order to signal
this: imperative forms for first and third person, and a syntactic solution for
second person. The latter is necessary as the addition of another command with
2ps imperative would possibly not indicate the relation between the two actions,
i.e. 'build the boat, enter it!', (which is not a possible construction in Urarina),
would not suffice to mark the consecutive relation between the two clauses.
While the focus marker is employed as in the examples in (844), the same
construction is also found without focus marker, i.e. the purposive/consecutive
nature of the dependent clause is expressed by its peculiar morphological and
syntactic structure and not directly related to the focus marker.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Imperative and future reference 585

(845) Consecutive function of imperative without focus marker

a)
asi-u huara-a=ne hoataa
[hide-IMP] [see-3ps/D=SUB so.that.not]
[V] [DEP]
'Hide it so that he would not see it!'

b)
kanu hjane-ku=tQe lenone-m-0
[lsg] [let-IMP=PLT] [eat-DSTL:HORT]
[Ο] [V] [DEP]
'Let me go so that I eat!'

Note that the consecutive function does not generally apply to all dependent
clauses that follow an imperative. In (846), a conditional clause follows the
imperative.

(846) No consecutive function after imperative

d^aha-u kahia ii leotQa kuhuaerate-i=jie=ra

[come.on-IMP over.here] [2ps other defend-2ps=CND=EMF]
[V] [DEP]
'Come over here if you [want to] defend your friend!' [Context: Man talking to jaguar
after he already hurt another jaguar]

15.4 Imperative and future reference

When the imperative is followed by the enclitic =ni, the command refers to the
near future. More exactly, this form implies 'later today', as shown in (847). In
these examples, the presence of =ni implies a non-immediate command,
however, referring to the same day.

(847) Non-distant future marker =ni with imperative

a)
lenone-u-ni
eat-IMP-NDF
'Eat later!' [later today]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
586 Imperative

b)
las tres ke u-u=m
three.o'clock INST come-IMP=NDF
'Come at three o'clock!' [Implying 'around 3 o'clock', later today]

It is easily noted that the function of =ni in the above examples is in exact
correspondence with the past meaning of this morpheme as the witness enclitic
=ni (cf. §12.3.3). While an evidential function is excluded with the imperative,
the coincidence of the temporal function of this enclitic is obvious. However,
though the form may be closely related, I gloss it as a "non-distant f u t u r e "
( N D F ) form when it follows an imperative. Apart from its match with the
evidential enclitic, =m also is homophonous with the assertive marker (cf.
§12.3.2). Even though in this case, the reference to the future may somehow
coincide with the function of =m in (847), the "non-distant" feature is absent
f r o m the function of the assertive. 71 Yet another homophonous m o r p h e m e is the
distal marker -m, observed with imperatives only. On the one hand, the function
of the distal marker again appears functionally related to the N D F morpheme,
due to its meaning of "not here and now". On the other hand, this m o r p h e m e is
a suffix and occurs in a different position, i.e. before the imperative suffix. It
can even co-occur with the N D F marker, as is evident from (848a). Thus, none
of the homophonous morphemes can be proven to exactly match the functions
of the N D F marker. One could assume that the origin of all four markers may
be related, but that these have developed different functions (and even taken
different positions within the verb) over time.
Apart f r o m its compatibility with the distal form (cf. (848a)), the N D F
marker may also co-occur with strong and weak prohibitive forms, as examples
in (848b,c) illustrate. In contrast, it cannot be combined with hortative and
jussive forms.

(848) Non-distant future marker =ni with distal and prohibitive forms (elicited)

a) With distal form:

lenone-m-u=rn
eat-DSTL-IMP=NDF
'Go eat later!'

71
Recall that the assertive is only used with the irrealis in order to mark future tense.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Suggestive form 587

b) With weak prohibitive:

kwa lenoni-a=ni
don't eat-NTR=NDF
'Don't eat later!'

c) With strong prohibitive:

nihjauria lenone-i=ni=ße
don't eat-2ps=NDF=N EG F
'Don't eat later!'

When the plural suffix -tg,e is added to the prohibitive forms, the non-distant
future marker is realised as -7, which corresponds to the allomorphy observed
with its homophones as well.

(849) Non-distant future marker =rn with plural form (elicited)

ho a lenord-a-toe=~i
don't eat-NTR-PL=NDF
'Don't (PL) eater later!'

15.5 Suggestive form

The introducer d^atera is used to start a suggestive clause to mean 'what i f , or

'how about', such as in (850). It can express an unfulfilled wish or a suggestion
by the speaker. At this stage, not much is known about the exact structure and
function of this clause type as it only occurs in a few examples of the database.
The inflection of the verb depends on person marking and partly coincides with
person marking for the jussive form.

(850) Suggestive clause with 1 sg

a)
d^atera ubarlti ke ßaara kuhoa-riu
how.about liquor VLI 2pl give.to.drink-RAP-SUG
'How [would it be] if I gave you liquor to drink?!'

b)
d^atera katQa ne-1 d$e=tete-riu=te
how.about man be-PRT 2sg=make-IRR-SUG
'What if I could make you human ... [you could show me how to give birth].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
588 Imperative

While for lsg, the suffix -riu is attached to the verb, with 3ps, the jussive forms
-mii or -ipe are suffixed (cf. (851)). Examples with 2ps reference are not
attested. In most examples, the irrealis form of the transitive verb tetia 'make,
put, set' precedes the suggestive suffix to result in the form tete-riu. This can
roughly be translated as Ί would make . . . ' . The use of the jussive in (851) also
has the connotation of ' W h o could do it?' or ' W h o could go to get it?'
respectively.

(851) Suggestive clause with 3ps

a)
d^atera itga-mii
how.about do-JUSS
'How if he did it?'

b)
d^atera ra-a ku-ijw
how.about receive-NTR go-JUSS
'What if he went to get it?'

Note that negation is not compatible with the suggestive form. A sentence such
as Ί wish there were no rats' is changed as illustrated in (852), making the
copula a transitive verb (tetia) and using the postposition kulane ' w i t h o u t '
instead of negating the verb. Literally, the sentence means ' H o w if I made [the
place] without rats'.

(852) Suggestive clause and negation

d^atera saa kulatie tete-riu

how.about rat without make-SUG
Ί wish there were no rats.'

15.6 Other aspects regarding imperatives

C o m m a n d s are only marked with verbs. While the inflectional pattern involves
transitive as well as active intransitive, PSC verbs, and intransitivised verbs, it is
unclear at this stage whether stative verbs can be marked for imperative. Note,
however, that the copula is not exempted from imperative inflection.
Though rarely observed in practice, pronouns can occur as a subject in
imperative constructions, similar to declarative clauses. While preverbal as well
as postverbal position is more or less acceptable, the occurrence of a pronoun
before the verb is ambiguous, as this normally is the position of the Ο argument.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Differences between imperative and declarative mood 5 89

Thus, knowledge of the context will be essential. 72 As a further variation, the

pronoun can be suffixed with -at$a 'only'. Also note that the (elicited)
examples only involve the 2sg pronoun.

(853) Pronominal subjects in imperative constructions

a)
ii itQa-u / ii-tga itga-u / (itQa-u ii)
2sg do-IMP / 2sg-only do-IMP / do-IMP 2sg
'You do it!'

b)
ii sim-H j ii-tQa sini-u / (sini-u ii)
2sg sleep-IMP / 2sg-onIy sleep-IMP / sleep-IMP 2sg
'You sleep!'

15.7 Differences between imperative and declarative mood

The presence of suffixes that can precede any imperative marking was already
mentioned in the previous sections. Their occurrence is not limited to the
imperative, but they each have a corresponding function in the declarative
mood. The diminutive -uri may deviate from this in that its use with the
imperative almost always implies a politeness, while this is not the case when it
occurs with declarative clauses.
In opposition to number and person, other grammatical categories are
unmarked in the imperative. This includes tense, aspect, evidentiality, and
various other features, which are summarised in (854). Furthermore, a
distinction between inflection types, such as in the declarative mood (cf.
§12.2.16) is not made in the imperative mood.

72
While the syntactic peculiarity mentioned here is exceptional, it must be noted that
otherwise, imperative clauses follow the same syntactic rules as declarative clauses.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
590 Imperative

(854) Summary of differences between declarative and imperative mood

Positive Negative Imperative Strong Weak

declarative declarative Prohibitive Prohibitive
Person Inflection yes yes no no no
Types
Tense: FUT -ri -ri — — —

Tense: REM =lu =lu — — —

Aspect:CNT -he -he

HAB1 -reheto -reheto
HAB2 -nahaaoka -nahaaoka
CPL -si -si
Distal — — -m — —

Negation (n/a) -e, -i, -ene (n/a) =ne -a

Introducers — hoatia hvane mhjauria hoa
Serial verbs yes yes yes yes yes
Polite form -toe -toe -t$e -toe -toe
In-law talk -ana -ana -ana -ana -ana
Velocity -uri -uri -uri -uri -uri
SUB clauses yes yes — — —

Reportative =he =he =he =he =he

Causative 1 -a -a -a -a -a
Causative 2 -erate -erate -erate -erate -erate
Emphasis =ra =ra =ra -ra =ra
Frustrative =ta =ta =ta
<t3

II
II

Distributive -ahoa, -ahoa, -ahoa, -ahoa, -ahoa,

Plural Object -naha -naha -naha -naha -naha
'only' -atQa -atQa -atQa -atQa -atQa

In summary, the differences between declarative and imperative mood involve a

variety of morphological and syntactic peculiarities. This leads to the
conclusion that imperatives constitute a separate clause type. The table in (855)
illustrates the areas of overlap between person marking for declarative and
imperative forms, which are in fact virtually non-existent: if one disregards the
special imperative marking with introducers, the only coinciding areas between
imperative and declarative are number marking in general and person marking
for l p l only. T h e negation in strong prohibitive marking is expressed by the
enclitic = m , which does not occur in the declarative; therefore, there is n o
overlap of negators between prohibitive and negative declarative mood.
However, person marking coincides for 2ps in both declarative and the
prohibitive forms, whereas this is no the case for 3ps: interestingly, there is an

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Differences between imperative and declarative mood 591

o v e r l a p b e t w e e n 3 p s m a r k i n g b e t w e e n t h e positive declarative and the

prohibitive, w h e r e a s the negative declarative has a separate f o r m .

(855) Person marking on verb - comparison between imperative and declarative mood

Positive IMP Positive Negative Strong Weak

Declarative Declarative Prohibitive Prohibitive
lsg -e -anu / -« -a-u / -ene-H — —

lpl/ex -akaanu -akaam -a-akaanu — —

lpi/du -aka -aka -a-aka — —

lpl/in -aka-tge -aka-tge -a-aka-tge — —

2sg -H -i -e-i -i=ße -a

2pl -u-tQe -i-tQe -e-i-tQe -i-toe=ße -a-toe
3sg -mn (OLD) -a j -e -i j -em -a=ne —

-iße (NEW)
3pl -HTH-mn (OLD) -uru-a 1 -ure -Hru-ij-HrH-ene ? —

-uru-iße (NEW)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 16 Valency changing mechanisms

Valency changes in Urarina are marked by suffixes on the verb. §16.1 describes
a verbal derivation with passive function. However, more correctly, the
morpheme under investigation can be described as a nominaliser. Another form
with valency-decreasing properties is the prefix ne-, which functions as a
general intransitiviser (cf. §16.2). This morpheme may also take reflexive
meaning under certain conditions, especially when co-occurring with the adverb
raahemane 'self. §16.3 describes the reciprocal form, which uses the particle
ita. The functions of two causative forms previously discussed in §12.2.1 and
§12.2.2 are investigated in §16.4. They differ from each other in terms of
transitivity features, productivity and (partly) in function, as direct vs. indirect
causation are distinguished. When an additional argument occurs as a result of
increased valency, it is followed by the postposition he, as investigated in §16.5.

16.1 Passive

What might be called "passive" in Urarina essentially is a deverbal derivation

that takes passive-like functions: the suffix -noi (and allomorph -whwi,
pronounced [noi] and [nohwl], respectively), represent a special type of
nominaliser. The structure of this suffix can be characterised as morphologically
complex, at least from a diachronical point of view: it is plausible to say that the
suffix is composed of -no and -i. As suggested in §10.1.1, there is a
nominalising form -no, which is used as a little productive suffix on a few verbs
that refer to dimension, such as 'be deep' and 'be wide' with which the abstract
nouns 'depth' and width' are formed. In another example, it derives the word
for 'envoy' leto-no from the verb letoaa 'send'. Even though it is not attested
with similar cases, this derivation provides relevant information for the structure
of -no: An envoy is 'the one who is sent', that is, the object of a transitive verb.
The relation of this form with the "passive" marker is unquestionable.
The second component of -noi ([noi]), at first sight appears to be identical
with the participle suffix -i (cf. §12.4; §20.3.1). But there is another, more
plausible analysis: According to the phonological rule discussed in §2.8.10, a
nasal consonant spreads rightward through vowels and through /h/. Thus, the
nasalisation of the final vowel in -noi can be described as the result of /n/
spreading its nasal feature through to -i, - which is a nominaliser referring to the
object of a transitive clause or to the subject of an intransitive clause (cf.
§6.6.1). Thus, the form -noi can be characterised as a nominaliser with passive

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Passive 593

function. 73 While the underlying structure of -noi has been shown to be

complex, it must be said that the forms -whun and -noi are fossilised, as people
do not recognise the two nominalisers as active morphemes in this combination.
This is also supported by the fact that -no is quite unproductive at any rate.
Thus, one can assume that the fusion of the two nominalisers has a diachronic
background, even though the reasons for this strategy in order to create a
passive form may remain obscure at this stage. Given the fossilised status of the
suffix, I will stick to the glossing of -noi (and its allomorph), as "passive".
Despite the fact that -noi was introduced as the underlying form above, it
cannot be stated with certainty which of the variants is underlying. The
conditions of the allomorphy between -noi and -nohzoi are hardly predictable, as
is illustrated in (856). Both forms occur with many different verbs of similar or
identical phonological structure. They resemble each other in terms of number
of syllables and with respect to their final root vowel. The only prediction that
can be made is that monosyllabic roots take the allomorph -nohzoi, whereas the
choice is not predictable for other verbs. For instance, disyllabic verbs such as
amaa 'take along' take the allomorph -nohzoi, whereas the verb itQaa 'do',
which is disyllabic and ends in /a/ as well, takes the form -noi. Similar
observations can be made for most other verbs shown in (856). The "final
vowel", as given in the third column, represents the underlying phoneme
(Recall that underlying /e/ is raised to [i] before /a/; cf. §2.8.4).

(856) Allomorphs -noi and -nohzoi

Syllables Verb Final V Gloss Suffix

1 h-a /i/ 'eat' -nohzoi
1 ti-a /e/ 'give' -nohzoi
1 ha-a /a/ 'make' -nohzoi
2 ama-a /a/ 'take along' -nohzoi
2 it^a-a /a/ 'do' -noi
2 ari-a /i/ 'seek' -nohzoi
2 siiri-a /if 'have' -noi
2 muku-a /«/ 'catch' -nohzoi
2 heri-a /e/ 'want' -noi
2 kuri-a /e/ 'change' -nohwi
3 ateyia-a /a/ 'murder' -nohzoi
3 towra-a /a/ 'obey' -noi
3 raatiri-a /i/ 'leave' -nohzoi

73
The labial glide in -nohzoi results from vowel copying of /o/, which is another
automatic phonological rule (cf. §2.8.3).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
594 Valency changing mechanisms

Syllables Verb Final V Gloss Suffix

3 kureti-a /e/ 'buy' -noi
3 itahi-a /e/ 'destroy' -nohwi
3 hjunumt-a /«/ 'pull' -noi
3 kzvituku-a /«/ 'know' -nohwi
4 enoala-a /a/ 'cook' -noi
4 eseneta-a /a 'believe' -nohxui

T h e examples in (857) provide further evidence for the assumption that the
form -noi is in fact an instance of nominalisation: in these sentences, the passive
suffix is followed by the plural marker -uru, which also occurs with the
nominaliser -i (cf. §6.6.1). Even though the plural marker is optional, its
presence strongly suggests that -noi is a nominaliser.

(857) -noi followed by plural suffix

a)
rüi baja, ßäe ni-a itulere rukuele, ho at a kahe itQa-noi-kuru=ne
[that after] [already be-3ps/D all.kinds things] [iron from make-PASS-PL=SUB
[CNJ] [Main] [DEP1]

hana-te, ßäe hjane-ι, huata kahe itQa-noi katai kurete-i

when=FOC] [already leave-PRT] [iron from make-PASS hook buy-PRT]
[DEP2] [DEP3]

ke kuhwu-ure ena netohwel

[INST fish-3pl/E now until]
[DEP4]
'After that, when there were all kinds of things, made from metal, they already left [the
old customs], buying hooks made from metal and fishing with them until today.'

b)
amia-rate-noi-tQuru katQa-urn=te ena ita kuruatahane-ure
[work-CAU2-PASS-PL man-PL=FOC] [now] [REC help-3pl/E]
[A] [ADV] [V]
'The people who had been forced to work now help each other.'

As a natural restriction, -noi is combined with transitive verbs only. In most

cases, it is attached directly to the root where it is subject to the general
phonological rules. For instance, the initial nasal of he suffix is realised as [p]
when the root ends in i\l. In addition to suffixation to a verb root, -tioi can also
be preceded by a causative suffix, which corresponds to verbal slots 1 and 2,
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Passive 595

respectively (cf. §12). One example for this is also shown in (857b). In this
case, the passive has scope over the causative. On the semantic level, there are
no apparent restrictions as to which verbs can be combined with -noi. Attested
examples cover a variety of semantic types, including the verb for 'have', siiria
(cf. (858)).

16.1.1 Modifier function of "-noi"

One of the functions of the passive suffix is in an NP, where it occurs as a

modifier, as illustrated in (858). In this case, it precedes the head noun. In
(858a), the noun 'fish' is specified by the passive form 'that he had killed'; in
(858b), the noun for 'brand' is preceded by the passivised verb. Note that in this
example, the agent (i.e. the original subject) precedes the passivised verb (also
cf. §16.1.3).

(858) Modifier function of -noi

a)
neleere mi mi-tea su-nohwi ate
[become.big:3ps/E] [that that-only kill-PASS fish]
[V] [S]
'The fish that he had killed became big.' [Pangayacu dialect]

b)
ii-tQa siiri- floi maraka ne-~t=ßa
[2sg-only have-PASS brand] [be-PRT=INT]
[CC] [V]
'[Should I buy a watch] being the same brand as you have?' [Lit. 'as owned by you']

Note that -noi can also function as head of an NP, as illustrated in (859).
However, its use with modifying function is more commonly attested in the
database.

(859) Passive form as head of NP

atii mi kiitga baha-nohioi u-a-ipe
[please] [that lsg.EMPH ask-PASS] [come-CAUl-JUSS]
[ADV] [Ο] [V]
'Please, he ought to bring what I asked him for.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
596 Valency changing mechanisms

16.1.2 The role of the copular auxiliary

While the passive marker can be preceded by the causative suffixes of slot 1
and 2, respectively, no verbal suffixes can follow -not. This is conceivable in so
far as the passiviser is analysed as an instance of nominalisation, i.e. as a noun.
Therefore, in order to indicate person, number and other grammatical
categories, -mi is followed by the copula with an auxiliary function, ma, which
is attached to the passivised verb as a clitic (also cf. §9.6.6). The examples in
(860) show that person and polarity are marked as suffixes on the auxiliary. In
the same way, clitics such as the remoteness marker are attached to the
auxiliary, rather than to -mi.

(860) Copular auxiliary following the passive form

a) Marked for 3ps and remoteness:

nitoanei=te itQa-noi=jie=lu eshoela
[like.that=FOC] [make-PASS=be:3ps/E=REM] [school]
[ADV] [V] [S]
'This is how the school was made.'

b) Marked for jussive:

dede-H ßaara raj kajfla-noi=jie-mn
[heaven-LOC] [2pl for] [return-PASS=be-JUSS]
[LOC] [PP] [V]
'It shall be given back to you in heaven.' [NT: Matthew 6:20]

c) Marked for privative:

beelaj- jioi=ße-lanaala ne-naa eene=te ßäe ku-e
[love-PASS=be-PRV be-NOM woman=FOC] [already] [go-3ps/E]
[S] [ADV] [V]
'The unloved woman has already gone away.' [Lit. 'the woman that was without being
loved']

There is one reason to believe that the copula is cliticised to the passive rather
than functioning as a separate word in this context: the initial nasal of the
copula is subject to palatalisation after the final /if of the passive form. This
corresponds to the effect a final vowel /i/ of any word has on enclitics that begin
with a nasal. However, the initial nasal of a word is not subject to this
alternation. If the copular auxiliary had word status, palatalisation of its initial
nasal should not occur. Therefore, the copula can be regarded as a special kind
of clitic, which is typically followed by inflectional suffixes (cf. §2.11).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Passive 597

The examples in (860) have illustrated the use of the passive form in
combination with the auxiliary, through which the complex of verbal stem,
passive marker and cliticised auxiliary with inflectional suffixes seems to
function like any other verb. Note that many suffixes that potentially occur on
most verbs are not attested with the auxiliary attached to a passive form. It is
unclear whether this is the result of systematic exclusion or simply due to the
fact that this form is relatively rare. In fact, the most common use is with
person, negation, and tense markers, or with the participle marked on the
auxiliary. The latter is shown in (861), where the entire combination of
passivised verb and auxiliary represents a participle clause (cf. §20.3). In this
case, person and other grammatical features are expressed on the verb of the
main clause. (861a) illustrates third person marking (Ε-form) on the free copula
ne 'it is', while the passive form on 'buy' is followed by the copular auxiliary
marked with the participle suffix, literally meaning 'being bought'. A similar
example is given in (861b), where an inflected form o f ' r u n ' represents the main
verb, while 'being called' functions as a participle.

(861) Passivised verb with copular auxiliary in the participle form

a)
aj, ßäe kurete-noi=jie-i=te ne nii ii-tga
yes [already] [buy-PASS=be-PRT=FOCbe:3ps/E] [that 2sg-only
[ADV] [V] [S]

baha-ur-i
ask-PL-NOMobj]
'Yes, what you had requested already has been bought.' [Lit. 'it is being bought']

b)
mi=te kat^a kutia-mhzui-jie-l ajßa amunei ... suru-e
[that=FOC] [man call-PASS=be-PRT with as.if] [run-3ps/E]
[S] [DEP] [V]
'That [girl], as if called by some man, ran (away).' [Lit. 'as if with [the state of] being
called']

16.1.3 Argument structure

Another important question regards the role of the arguments within a passive
construction. Normally, a formal passive involves the decrease of verb valency
as its object (O) takes the function of the subject while the subject (A) is shifted
into the periphery of the clause (if surfacing at all). In Urarina, it is very

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
598 Valency changing mechanisms

common for A not to be mentioned at all in a passivised clause. The subject of a

clause is marked on the verb, which is realised through person marking, while
NPs do not receive any formal marking, e.g. for case. Thus, the syntactic
functions of the involved arguments are expressed by constituent order (but note
that this order can be modified through the use of a focus marker).
The critical point is the presence of an agent and how it is realised, as the
examples in (862) show. Remarkably, what these have in common is that the
original agent occurs in object position, but without any oblique or other
marking. For instance, in (862a), the agent 'God' simply precedes the
passivised verb, to mean 'punished by God'. The second feature these examples
share is that S of the passive clause (i.e. Ο of the active clause) does not surface
as part of the same (i.e. passivised) clause. In (862b), 'my wife' is part of a
different clause while the passivised clause 'made by eye' is a dependent,
participle clause separate from the main clause.

(862) Agent without oblique marker

a)
mtahi-a ku-re=i, kana+kioaattn-era kaseteka-wi=jx-i
[be.lost-NTR go-IRR:3ps/E=ASS] [lpl/in+create-AG punish-PASS=be-PRT]
[V] [DEP]
'He will be lost, punished by God.' [NT: John 3:36]

b)
ni-a kann komasaj nuta itQa-noi=/>e~i=tg,a
[be-3ps/A] [lsg wife] [eye make-PASS=be-PRT=EMF]
[V] [S] [DEP]
Ί have a very beautiful wife.' [Lit. "made by eye"; idiomatic expression]

In the following example, the original object (i.e. the original Ο = S of the
passivised clause) is present, in focus (= initial) position. Thus, one may assume
that the presence of a subject does not play a significant role for the fact that the
external argument is unmarked. Indeed, one may conclude that this is the
regular strategy to represent a passivised clause.

(863) Original object present

lejhii eene=te u-e anofwa itQa-noi=jK-~i

[one woman=FOC] [die-3ps/E] [knife do-PASS=be-PRT]
[S] [V] [DEP]
Ά woman died through a knife.' [Lit. "died being done by a knife"]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Passive 599

There are more examples in which the original agent occurs before a passivised
verb, again without any formal marking. In (864a), the passivised form of
' w a n t ' is preceded by the original agent ( ' G o d ' ) and in (864b), this position is
filled by a pronoun. Most interestingly, in (864c), a proclitic takes the same
function. These examples could be instances of lexicalisation, which would be
plausible since ' G o d ' s will' (864a) and the idiom in (864c) are frequently used
expressions.

(864) Original agent preceding passivised verb

a)
nijej kana+huaaun-era here-noi itQa-i
[not.at.all] [lpl/in+create-AG want-PASS] [do-NEG:3ps/A]
[ADV] [Ο] [V]
'He does not do at all what is wanted by God.'

b)
kanu here-noi
lsg want-PASS
'my will' [Lit. 'what is wanted by me']

c)
k=itQa-noi
1 sg=make-PASS
'my work' [Lit. 'what I do']

Based on the examples in (864), it is possible to derive an alternative analysis of

the passive construction: the sentences in (864) do not formally differ from
possessive constructions, which can occur with simple juxtaposition (cf.
§7.1.1). Given the fact that the passivised verbs are actually nominalisations, i.e.
a type of noun, it makes much sense to assume that these function as the
possessum of a N P that precedes. This is strongly supported by example (864c):
proclitics can generally function as object pronouns or as possessors. As they do
not have object function here, they must be regarded as possessors. This view is
further supported by examples where S of the passive clause (i.e. Ο of the active
clause) is present and follows the verb carrying the suffix -noi. In this context, it
can be said that the nominalised/passivised verb modifies the noun, as was
illustrated by the examples back in (858).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
600 Valency changing mechanisms

16.1.4 Further observations

A further observation is that the passive in combination with a participle clause

may have scope over the preceding verb as well, as is illustrated in (865a).
However, this structure is not the only possible solution, as the passive suffix,
as illustrated in (865b)) can also occur optionally on various verbs in a
sequence.

(865) Sequence of verbs with -nohiui

a) Passive having scope over participle:

ßäe sau-i muku-nohwi=ni-a ku-re=t
[already] [cut-PRT] [burn-PASS=be-NTR go-IRR:3ps/E=ASS]
[ADV] [DEP] [V]
'it will be cut and burnt' [NT: Matthew 7:19]

b) Passive marked on both verbs:

muku-nohioi=jie-l, kaseteka-mhiui=jie-l,
[catch-PASS=be-PRT] [punish-PASS=be-PRT]
[DEP1] [DEP2]

ari-nohiui=fle-l naaohiua-ι ne-nakauru

[seek-PASS=be-PRT] [ENUM-PRT] [be-those.who]
[DEP3] [DEP4] [S]
'[Blessed are] those who are being caught, punished, and sought,...' [NT: Matthew
5:10]

In summary, it is suggested that, strictly speaking, there is no separate passive

form in Urarina. The use of a special nominaliser, which can occur with two
different values for aspect, fulfils similar functions as a passive form would do.
However, from a morphological and syntactic point of view, the "passivised"
verb is a noun, which becomes even more evident through the features it
exhibits:
- It can take a noun plural -uru.
- It is not inflected like a verb, but requires an auxiliary, which takes the
"passivised" verb as a copula complement.
- It can modify another noun.
- It can be possessed.

The general meaning of the passive in discourse is not entirely clear, but
typically, it may be chosen when the agent of a clause is unknown.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Intransitiviser "ne-" 601

16.2 Intransitiviser "ne-"

In §9.3, intransitivised verbs are listed as a separate verb type. This seems
justified as the productivity of the intransitiviser ne- is limited. This prefix is
found on transitive verbs only, some of which are listed in (866). Its meaning
can be described as a general intransitiviser, as reflexivity is not its primary
function. In particular, the semantic structure of verbs derived with ne- is not
predictable. For instance, in the intransitivised for 'suffer' (nekajriteaa, derived
from kajritQaa 'make suffer'), the subject has the semantic role of a patient; in
mmitQua 'get up' (derived from mitQua 'set up'), the subject represents an
agent. In semantic terms, the S argument of the two examples plays an active
role in one case, but a passive role in the other. A reflexive interpretation does
not apply in either case. However, with some other examples, the reading '(by)
himself or '(by) itself is possible, which in fact implies a reflexive function.
This can be emphasised by the use of raaheniane ' s e l f , as will be investigated
in §16.2.2. It should also be mentioned that there is a possible diachronic
relation between ne- and the 3ps object proclitic ne=, as described in §5.5.

(866) Derived intransitive verbs with ne-

a) S argument plays passive role
Intransitive: Transitive:
ne-kajritQaa 'suffer' cf. kairitQaa 'make suffer'
n-itahia 'get lost' cf. itahia 'lose'
ne-hwaa 'increase' cf. hioaa 'increase' (tr.)
ne-relaa 'learn' cf. relaa 'teach'
ne-rutua '(be) change(d) into' cf. rutua 'transform'
ne-kwehia 'feel ashamed' cf. kioehia 'embarrass'
ne-tabatakaaeka 'be reduced' cf. tabatakaaeka 'consume/reduce'
ne-laatia 'be necessary' cf. laatia 'need'
n-itQutia 'escape, get saved' cf. itgutia 'save', 'let escape'
w-heratia 'get angry' cf. heratia 'change side' (e.g. take
paddle into the other hand)

b) S argument plays active role

Intransitive: Transitive:
ne-hoasitaa 'show off cf. kwasitaa 'praise'
ne-kuretia 'sell' cf. kuretia 'buy, change'
m-mitQua 'get up' cf. mitQua 'set up'
ne-betakaa 'relax' cf. betakaa 'cause to rest'
ne-lauhuaa 'sit down' cf. lauhuaa 'make sit down'
ne-kwatihjaa 'play' cf. kwatihjaa 'make play'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
602 Valency changing mechanisms

Even though there is a relatively large number of verbs that contain the prefix
ne- (about 160 in the lexical database), the intransitiviser is not 100%
productive. In many examples, the preflxation of ne- comes along with
unpredictable morpho-phonological changes; in some, the base verb cannot be
determined. In the examples in (867), a connection between the transitive and
the resulting intransitive form can still be recognised even though some changes
take place in the base verb. Some other verbs, as depicted in (868), appear to
inherently occur with the prefix, i.e. there is no attested transitive form for
these.

(867) Intransitive verbs with ne- that involve morpho-phonological changes

ne-kualahia 'be careful' cf. kulalahia 'care for'
ne-kumtoa 'get hurt' cf. kunaa 'hurt' (tr.)
ne-rautoa 'get well/better' cf. rautaa 'heal'
w-fwaufwajtoa 'break' cf. fioaufwaa 'beat'

(868) Lexicalised intransitive verbs with ne-

neheonoa 'circle around'
netooka 'go to sleep'
nesoonia 'be scared'
neruruhuaa 'move, function, work' (intr.)

There is no obvious difference in the argument structure between intransitivised

verbs with a "patient" subject and those that have an "agent" subject. While the
base verb is always transitive and requires an A and an Ο argument, the
resulting verb with ne- is intransitive. This is illustrated by the example in
(869a,b), where 'teach' (literally 'instruct') is contrasted with 'learn': While
'teach' takes 'woman' as its Ο argument, the form with ne- is intransitive, (with
no overt subject in this case; it is only marked by the person marker on the
verb). The content of instruction (marked as an additional argument by he) is
optional in both examples.

(869) Intransitivisation and argument structure

a) Transitive verb:
mtoanei hetau=te floaelu kanaanaj hama-naa he
[like.that HRS=FOC] [earlier] [child give.birth-NOM VLI]
[ADV] [ADV] [Ε]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Intransitiviser "ne-" 603

eene rela-e=lu kati

[woman] [teach-3ps/E=REM] [black.monkey]
[Ο] [V] [A]
'This is how the black monkey instructed the woman in child-bearing in the past.'

b) Intransitivised verb:
ita kuruatahane-l ena ne-rela-kure amiane-naa ke
[REC help-PRT] [now] [ITR-instruct-3pl/E] [work-NOM VLI]
[DEP] [ADV] [V] [Ε]
'Helping each other, they now learn to work.' [Lit. 'they learn in working']

16.2.1 Reflexive function of "ne-"

In some cases, a verb intransitivised with tie- takes a function similar to a

reflexive form, as its meaning may imply that subject and object are identical,
(or, in other terms, the object is anaphorically related to an antecedent). This is
illustrated by the example for 'wash' in (870a), which is a reflexive verb in
many other languages. The transitive counterpart in Urarina is kuleeka. A
similar example is given in (870b), where the intransitiviser again implies a
reflexive function.

(870) Reflexive function of intransitiviser ne-

a)
ne-kulehe-u
ITR-wash-IMP
'Wash yourself!'

b)
ne-sau-ri-teau=rti
ITR-cut-IRR-1 sg/A=ASS
Ί will cut myself.'

However, for most verbs, this function cannot be applied, as shown in (871),
where the verb has clearly intransitive status, but no reflexive meaning. Thus,
the interpretation as A = 0 is less feasible. While example (871a) might still be
interpreted as 'the arm broke itself, which would again suggest reflexivity, this
reading is not applicable in (871b) where a translation as 'we see ourselves'
does not correspond to the actual meaning. The function of ne- in this example
may therefore be described as anticausative.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
604 Valency changing mechanisms

(871) Intransitive function of m-

a)
ne-fioaao-ka kamt bihi
[ITR-break-3ps/A] [lsg arm]
[V] [S]
Ί broke my arm.' [Lit. 'My arm broke.']

b)
katpa ne-ι ni-akaanu natii=jia enejtQu ne-1 ne-huara-kaanu
[man be-PRT be-lpl/ex although=FOC: 1 pi] [monkey be-PRT] [ITR-see-lpl/ex]
[DEP1] [DEP2] [V]
'Even though we are people, we look like monkeys.' [Lit. 'we seem monkeys']

If one looks at other verbs listed in (866) and (867), a reflexive interpretation
would be possible in some cases. While some verbs might imply that the subject
is involved in performing the action, the subject does play an active part, but is
actually subject to the change without having active influence on the event. For
the three verbs in (872a), an active part may be assumed for the subject,
whereas the majority of verbs with ne- (cf. (872b)) do not allow this reading.
These are rather to be interpreted intransitively, i.e. roughly translated as 'by
itself.

(872) Re-interpretation of verbs intransitivised by ne- (cf. (866), (867))

a) Reflexive reading of intransitivised verb

Trans, verb Intrans. verb Intransitive Alternative reading
reading
kioasitaa 'praise' ne-kiuasitaa 'show off 'praise oneself'
mitQua 'set/lift up' τη-mitQua 'get up' set/lift up oneself'
koatihjaa 'make play' ne-huatia 'play' make play oneself'

b) Intransitive reading of verb with ne-

Trans. verb Intrans. verb Intransitive Alternative reading
reading
itQutia 'save' n-itQutia 'escape, get saved' save by itself'
kairitQaa 'make suffer' ne-kajritQaa 'suffer' make suffer by itself'
itahia 'lose' n-itahia 'get lost' lose by itself'
hivaa 'increase' ne-hioaa 'increase' increase by itself'
(tr.)
relaa 'teach' ne-relaa 'learn teach by itself'
ruttia 'transform' ne-rutua '(be) change(d) 'transform by itself'
into'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Intransitiviser "ne-" 605

Trans, verb Intrans. verb Intransitive Alternative reading

reading
hoehia 'embarrass' ne-koehia 'feel ashamed' "embarrass by itself'
kuretia 'buy, change' ne-hiretia 'sell' "change by itself'
tabatakaaeka 'reduce' rie-tabatakaaeka 'be reduced' "reduce by itself'
hdalahia 'care for' ne-hialahia 'be careful' "care for by itself'
kunaa 'hurt' (tr.) ne-kunitoa 'get hurt' "hurt by itself'
rautaa 'heal' ne-rautoa 'get well/better' "heal by itself'
fivaufivaa 'beat' ne-fioaufioajtoa 'break' "beat by itself'
laatia 'need' ne-laatia 'be necessary' "need by itself'

In the same way as other verbs, verbs that are intransitivised by ne- can receive
a new argument that refers to an instrument or a goal (cf. §16.5). Note that the
verbs for 'buy' and 'sell' are based on hiria 'change': hire-tia 'buy' is
composed of 'change and 'give', and in turn, ne-kuretia 'sell' is the
intransitivised form of 'buy'. As illustrated in (873), the new argument, here
corresponding to "Goal", is marked with ke. This may be somewhat
counterintuitive, since ke would not occur with the underived (transitive) base
verb. For instance, the Ο argument of kuretia 'buy' does not take the
postposition, whereas the same NP must be accompanied by the valency
increaser when occurring with 'sell' (cf. (873a)). This is entirely due to the fact
that 'sell' is intransitive and any additional argument must be marked by ke.

(873) Intransitivised verbs with additional "Goal" argument

a)
uu-re=i rukuele ke ne-kurete-nakauru
[come-IRR:3ps/E=ASS] [things VLI ITR-buy-those.who]
[V] [S]
'The traders will come.'

b)
d$a ke=te ne-laate-re
[what VLI=FOC] [ITR-need-IRR:3ps/E]
[PP] [V]
'What use would it have?' [Lit. 'what would it be necessary for?]

Other functions than goal are relatively rare with this construction. In (874a),
the instrumental function is exemplified with the verb for 'function/move',
which, however, is lexicalised as no transitive form is attested, i.e. it is only

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
606 Valency changing mechanisms

assumed that the initial m - corresponds to the intransitiviser. In (874b), the

additional argument refers to a location.

(874) Intransitivised verbs with other argument functions

a) Additional argument = instrument:

saabere ke neruruhu-e
[machete VLI] [function-3ps/E]
[PP] [V]
'He used the machete.' [Lit. 'He functioned by machete']

b) Additional argument = location:

nii kuri ke ßäe kauatga-ϊ ne-rauta-a=ne
[that jagua.tree VLI] [already] [good-PRT] [ITR-heal-3ps/D=CND]
[PP] [ADV] [ADV] [V]
'when they made themselves comfortable on that jagua tree ...' [Lit. 'When they were
getting better']

16.2.2 Functions of' 'raahen iane " 'self'

Urarina also has a lexical strategy to refer to ' s e l f : the word raahernane, which
can be characterised as an adverb, is employed in autoreflexive function to refer
to ' s e l f . Its position within the sentence is relatively free, but in all attested
examples it refers to the subject of the clause.

(875) Function of raahernane 'self

a) Initial position:
raahernane jwau-Ί kwara-a ku-e
[self descend-PRT] [see-NTR go-3ps/E]
[DEP] [V]
'He climbed down and went to see for himself.'

b) Initial position with proclitic:

ka=raaheniane=na lajjia-u=lu gobernadoro raj
[lsg=self=FOC:lsg] [write-lsg/E=REM] [governor for]
[A] [V] [PP]
Ί myself wrote to the governor.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Reciprocal form with "ita" 607

c) Final position:
jioaelu ne-nakauru hit ς, ana ha-uru-a nerajfie raaheniane
[earlier be-those.who] [blowgun] [make-PL-3ps/D must] [self]
[Α] [Ο] [V] [ADV]
'Those who lived earlier had to make blowguns themselves.'

In the a b o v e examples, raahemam (sometimes abbreviated to raahe) occurs

with intransitive and transitive clauses. Its function in combination with a verb
intransitivised by ne- is to emphasise the reflexive interpretation, as illustrated
by some examples from the NT.

(876) raaheniane with intransitivised verb

a)
turu-uru-a=ne hana hetau, raaheniane ne-huuraka-e
[arrive-PL-3ps/D=SUB when HRS] [self] [ITR-open-3ps/E]
[DEP] [ADV] [V]
'When they arrived, it [the gate] opened by itself.' [NT: Acts 12:10]

b)
raj esefioara rukuit-e, raaheniane ne-su-1
[POSS sword] [pull.out-3ps/E] [self] [ITR-kill-PRT]
[Ο] [V] [ADV] [DEP1]

m-hjau-a=ne kujßa
[ITR-throw-3ps/D=SUB so.that]
[DEP2]
'He pulled his sword in order to kill himself.' [NT: Acts 16:27] [Lit. "in order to throw
himself away, killing himself']

It can be concluded that the principal function of the intransitiviser rte- is not of
a reflexive nature. This meaning is only applicable with a minority of
intransitivised verbs. With most other verbs, the prefix functions as a general
intransitiviser without a specifically predictable meaning.

16.3 Reciprocal form with "ita"

The reciprocal is formed by the particle ita to express 'each other'. It only
occurs with the plural form of verbs and is only found with transitive verbs.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
608 Valency changing mechanisms

(877) Reciprocal it a

a)
hauna he hetau=te nii ita fiuaru-rehete-kure=lu
[bamboo.stick INST HRS=FOC] [that REC] [cut-HABl-3pl/E=REM]
[PP] [O] [V]
'They used to cut each other with a bamboo stick.'

b)
ita he kana suuhua akate-i m-aka=ne=te here,
[REC VLI lpl/in heart defend-PRT] [be-lpl/du=SUB=FOC] [want:3ps/E]
[DEP1] [DEP2] [V]

kauatQa-i ita kwara-i

[good-PRT REC see-PRT]
[DEP]
'He wants that we forgive each other and nicely see each other.'

The reciprocal particle can also be used as preceding postpositions such as ajßa
'with' and raj 'for', as is illustrated in (878a,b). With the verb rihia 'be like',
ita is employed to mean 'each in the same way' (cf. (878c)).

(878) Extended use of reciprocal form

a) With ajßa\
jietononei ita beela-ι ita ajßa rautohive-Ί kana suuhua auna-1
[always REC love-PRT] [REC with be.calm-PRT lpl/in heart hear-PRT]
[DEP1] [DEP2]
'... always loving each other, being peaceful with each other and listening to our hearts

b) With raj:
ita raj lenone te-ure
[REC for] [food] [give-3pl/E]
[PP] [Ο] [V]
'They gave food to each other.'

c) With rihihel:
amiam-akaanu ita rihihel
[work-lpl/ex] [REC like]
[V] [PP]
'We each worked the same [amount].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Causative 609

The reciprocal ita may also occur in comparative constructions that express
equality, as was mentioned in §5.4.6.

16.4 Causative

Urarina has two suffixes that mark causative, -erate and -a, which are used with
verbs of different transitivity classes. The major difference in the function of the
two suffixes regards the direct vs. indirect involvement of the causer, as will
become evident in §16.4.3. An interesting aspect is that both suffixes can occur
in combination, which typically refers to a delegated task and involves an
additional participant.

16.4.1 Causative with "-erate "

The form -erate (and allomorphs) is a common and productive morpheme that
increases the valency of a verb. It can be attached to almost any verb, which
includes transitive and active intransitive verbs, as well as some colour and
stative verbs (in elicitation only). The suffix is also observed with verbs that
contain the general intransitiviser Tie- (cf. §16.2). Furthermore, -erate can be
suffixed to the transitive verbs for 'eat' and 'drink' (ha and kua, respectively),
but does not occur with the intransitive lenoma 'eat', neither with the copula
ma. In (879), examples for the occurrence of -erate with different verb types is
given. The suffix occurs in verbal slot 2 and can be followed by almost any
suffix (with only a few unattested, but likely to be possible combinations, cf.
§12.2.2).

(879) Occurrence of -erate with different verb types

a) With intransitive verb:
kaa katga-uru amia-rate-reheto-a=lu
[this man-PL] [work-CAU2-HABl-3ps/A=REM]
[O] [V]
'He used to make these people work.'

b) With transitive verb:

arusu temu-erate-i akauru ke
[rice] [plant-CAU2-PRT] [3pl VLI]
[Ο] [V] [E]
'making them plant rice ...'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
610 Valency changing mechanisms

c) With intransitivised verb:

ne-rauta-erati-a u-u-t$e=ra
ITR-heal-CAU2-NTR come-IMP-PL=EMF
'Come to be healed!' [NT: Luke 13:14]

With stative verbs, the formation of a causative is uncommon and not attested in
natural text. However, speakers produced causativised forms of some stative
verbs in elicitation. Based on the fact that no examples are found in spontaneous
speech or other texts, one may conclude that stative verbs do not normally take
the causative suffix.

(880) Elicited causative forms for stative verbs and colour verbs

kalanohioaa 'be dirty' kalano-eratia 'send to get dirty'

ahißaa 'be sharp' ahia-eratia 'send to get sharp'
ahaatoa 'be warm' aha-heratia 'send to get warm'
somahaa 'be white' somah-eratia 'send to be white'
lanaaka 'be red' lana-heratia 'send to be red'

The causative suffix is subject to a range of morpho-phonological changes,

which depend on the shape of the root it is attached to. The underlying form is
-erate, which occurs with most verbs. Note that, similar to some other suffixes
beginning with /e/ (e.g. the 3ps/E suffix -e), the final vowel of a preceding root
is modified or elided under certain conditions. The alternations that apply are
listed here:
- When the final root vowel is /e/, it merges with the initial /e/ of the suffix,
such as in hjani-a 'leave' - hjane-ratia 'make leave', eri-a 'speak' ->
ere-ratia 'make speak' and bi-a 'tell' -> be-ratia 'make tell'. (Recall that the
alternation from /e/ to [i] is a general phonological rule occurring before /a/).
- In verb roots that end with /a/, the suffixation of the /e/-initial suffix -erate
results in the diphthong [ae]. However, in some verbs, this can be realised as
[e], as is illustrated by letoa-eratia 'make send' [letoeratia] and
kuhjuta-eratia 'make scold' [kuhjuteratia]. This rule does not apply to
all verbs of this type. The examples d^ana-eratia 'make cry'
[d3anaeratia] (rare due to existence of suppletive form) and na-eratia 'make
say' [naeratia] show that the final /a/ of the root is preserved.
- Verbs whose root ends in the diphthongs /a«/ or /au/ lose the final part of that
diphthong, i.e. the final root vowel is reduced to /a/, as in sau-a 'cut' ->
sa-eratia 'make cut', kau-a 'return' ka-eratia 'make return', and auau-ka
'cry aloud' aua-eratia 'make cry aloud'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Causative 611

- When the final root vowel is /«/, it is reduced to a velar glide before the
causative suffix: This is illustrated by the forms amu-a 'walk' -> amu-eratia
'make walk' and eru-a 'meet' eru-eratia 'make meet'.
- When the final root vowel is /u/, it is reduced to a labial glide in some verbs
before a causative suffix follows, such as in suru-a 'run' suriu-eratia
'make run'. However, this does not apply to all verbs of this shape:
ajto-a 'say' ajto-ratia 'make say' is an exception.

Similar to the causative formation of ajto-a 'say', there are a number of cases
where not the verb root itself is modified, but the suffix undergoes certain
changes:
- When the root that precedes ends in a long vowel or in vowel sequence, the
causative marker is realised as -rate, such as in kuaau-ka 'think'
hoaau-ratia 'make think' and ehoo-ka 'appear' -> ehoo-ratia 'make appear'.
- When the root that precedes ends in the vowel / i f , or in a diphthong ending
with a palatal glide, the causative marker is realised as -tQate, as exemplified
by ki-a 'eat' Ιά-tQatia 'make eat' (rare form as there is a suppletive form
for 'feed' as well) and beelaj-a 'give as a gift', love' -> beelaj-tQatia 'make
give as a gift, cause to love'.

In addition to these rules, there are some exceptions and irregular forms. These
include halaito-a 'make a hole' - halaito-rzuatia 'let make a hole', hau-a
'throw' - hwa-eratia 'make throw', and hwau-a 'descend' - hana-eratia 'make
descend'. The last two examples are based on an alternation involving the
change of initial / h w / to /hau/ and vice versa, which, however, is observed with
these two verbs only. The verb amiam-a 'work' has the irregular causative form
amia-ratia 'make work'.
With any causative, a change in the argument structure is always involved,
as the valency of the verb is increased by one argument. With intransitive verbs,
the new argument (i.e. the causer) becomes A, while the old S argument
becomes O. As could be expected, this is expressed by a change in constituent
order in Urarina, as exemplified below.

(881) Causative -erate with intransitive clause

a) Intransitive clause:
amiane-uru-a kanaanaj-uru
[work-PL-3ps/A] [child-PL]
[V] [S]
'The children worked.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
612 Valency changing mechanisms

b) Causativised clause:
kanaanaj-uru amia-rati-a katga
[child-PL] [work-CAU2-3ps/A] [man]
[Ο] [V] [A]
'The man made the children work.'

The transformation from an intransitive verb to a higher degree of valency is

unproblematic, since the resulting verb behaves in the same way as a transitive
verb. Given the fact that Urarina does not have ditransitive verbs, the following
changes apply:
- The new argument (causer) becomes A,
- The old Ο argument remains O,
- The old A argument becomes oblique.

As illustrated in (882), such an increase in valency is accounted for by marking

the now oblique argument (i.e. the old A argument, marked as "[E]" here) with
the postposition ke. Note that the new agent (i.e. the subject of the causativised
clause) does not surface as an overt NP in the examples cited here, but is only
marked by the person suffix on the verb.

(882) Causative -erate with transitive clause

a)
Transitive clause:
leotQa amiane-naa itQa-käu kanu
[other work-NOM] [do-lsg/A] [lsg]
[O] [V] [A]
Ί did other works.'

Causativised clause:
leot^a amiane-naa itQa-rati-a kanu ke
[other work-NOM] [do-CAU2-3ps/A] [lsg VLI]
[Ο] [V] [E]
'He made me do other works.'

b)
Transitive clause:
amiane-1, itulere itga-ure kaa katQa-uru
[work-PRT] [all.kinds] [do-3pI/E] [this man-PL]
[DEP] [Ο] [V] [A]
'Working, these people did all kinds of things.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Causative 613

Causativised clause:
kaa katga-uru amia-rate-1, akauru ke itulere itga-rate
[this man-PL] [work-CAU2-PRT] [3pl VLI] [all.kinds] [do-CAU2:3ps/E]
[Ol] [VI] [E2] [02] [V2]
'He made these people work and made them do all kinds of things.'

It is quite rare to see all arguments to appear as full NPs. In (883a), the object of
the causativised verb does not surface, as it is understood as a 3ps object 'it';
thus, only the external argument (= the former A) surfaces and is marked by he.
In (883b), the external argument is not stated, i.e. the causee remains
unspecified. In this example, the " n e w " subject shows up in focus position, at
the beginning of the sentence.

(883) Omission of arguments

a) Ο argument of causativised verb does not surface:

ajto-a komasaj ke itQa-rati-a=ne kujjia
[say-3ps/A] [wife VLI] [do-CAU2-3ps/D=SUB so.that]
[VI] [E2] [V2]
'He (A) spoke so that he (A/B) would have his wife do it.' [Referent for 'he' is
ambiguous]

b) Ε argument of causativised verb does not surface:

kamt=na uhoana itga-rate-ü
[lsg=FOC:lsg] [field] [do-CAU2-lsg/E]
[Α] [Ο] [V]
Ί made him make my field.' or Ί had my field made.'

In some examples, there is no overt N P in a causativised clause, as shown in

(884). Note that the verb in (884b) contains a plural object marker, which refers
to the Ο argument of the transitive verb 'teach, instruct'.

(884) No overt NP with causativised clause

a)
mi häu=te kujßadera-1 ari-tQate
[that because=FOC] [be.worried-PRT] [seek-CAU2:3ps/E]
[CNJ] [DEP] [V]
'Therefore, he [the father] had her [the child] sought [by the people].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
614 Valency changing mechanisms

b)
na-i hetau rela-erate-naha-e
[say-PRT HRS] [teach-CAU2-PLO-3ps/E]
[DEP] [V]
'Saying so, he [the priest] had them [the people] taught [by his novice].'

There are some examples in which the valency increase marker ke surfaces in a
causative construction without being accompanied by an overt N P . While it can
be present optionally in order to refer to the Ε argument (i.e. the "old" agent), it
is not obligatory in this context. In fact, there are three options in order to
represent the same content, possibly with different degrees of emphasis on the
argument in question. In (885a), ke is entirely optional, but it can also be
accompanied by the 3ps pronoun, as shown in (885b). The pragmatic
differences of these different versions are unclear, but it seems that the presence
of ke alone or ke plus pronoun is rather unusual.

(885) Occurrence of ke without overt NP

a)
(ke) kutia-rate-kure katQa-uru
[(VLI)] [call-CAU2-3pl/E] [man-PL]
[Ε] [V] [A]
'The people made him call her [his wife].'

b)
(aka ke) kutia-rate-kure katQa-uru
[(3sg VLI)] [call-CAU2-3pl/E] [man-PL]
[Ε] [V] [A]
'The people made him call her [his wife].'

Since Urarina does not have ditransitive verbs, examples such as tia ' g i v e '
behave in the same way as other transitive verbs would do with the causative:
the object remains in position, the "old" agent is marked by ke (if present), and
the recipient is followed by the postposition raj, as it is in the transitive clause.
Recall that a possible translation for tia is 'provide', rather than ' g i v e ' , which
illustrates its argument structure in a more understandable way in English. Also
note that none of the N P s in (886) is obligatory.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Causative 615

(886) Verb 'give' with causative (elicited)

äahera ke=te kiritga te-rati-a akauru raj

[angel VLI=FOC] [book] [give-CAU2-3ps/A] [3pl for]
[Ε] [Ο] [V] [Adjunct]
'He had the angels give the book to them.'

Regarding the constituent order in sentences where many NPs surface, there
appears to be only one preference: an Ε argument or a recipient typically is in
the periphery, i.e. it does not commonly occur between verb and object or verb
and subject (for exceptions see §18.4). There is no restriction as to whether a
peripheral argument occurs before or after the verb. In (886), the Ε argument is
found in initial position, as it is marked with a focus marker. The focus marker
is not obligatory in order to have Ε in this position, as is shown by (882b) and
(883a), among others. Alternatively, the ke-phrase can also occur in postverbal
position (cf. (882a).
From a semantic perspective, the causative form -erate can refer to a wide
range of meanings. The caused action may be forceful or voluntary. The causer
may be of non-human character, such as in 'this (situation) is going to make us
relax', talking about a certain work. Most typically, if the causer is specified, he
or she will trigger the action verbally, but not be directly involved with the
performed action.
Example (882b), which refers to the way patrones enslaved the Urarina,
clearly describes a forceful cause that results in the people's working. It can be
assumed that the force in most cases was not by direct physical contact (even
though possible), but through verbal threats. The same type of force can also be
interpreted for examples (879a,b) and various others. In (879c), however, the
caused action is voluntary and implies that healing is achieved by supernatural
power; obviously the patient wants to be healed. In other examples, such as
(883a), it is understood that the agent asks his wife to perform the action, which
would not involve any force. Similarly, in (885), the described situation makes
it evident that the people persuaded the man to call his wife, which does not
involve force. All examples have in common that the causer is not directly
involved in the action to be performed. In summary, it must be said that the
semantic interpretation of the causative is context-dependent. It always focuses
on the fact that the causer does not conduct the action himself, but delegates
others to do it.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
616 Valency changing mechanisms

16.4.2 Causative with "-a"

Urarina has another causative suffix, -a, which is less productive than -erate.
Besides, it can be attached to intransitive verbs only. In (887), I give a selection
of verbs that are attested for their occurrence with -a. Some of them involve
alternations which are not normally predictable from the phonological structure
of the verb root. For instance, some verbs whose root ends in /«/ or /u/ lose the
final root vowel before -a is suffixed (cf. (887b). Verb roots that end in /ni/ are
modified to result in final /JT/ before -a (cf. (887c)), and there are two examples
for which the causative suffix is -la instead of -a (cf. (887d)).

(887) Verbs that occur with causative suffix -a

a) Suffixation of -a:
atero-a 'be tired' atero-a-a [aterwaa] 'tire'
eno-a 'enter' eno-a-a [enwaa] 'make enter'
ham-a 'get light' hano-a-a [hanwaa] 'give birth'
u-a 'come' u-a-a [aaa] 'bring'
situ-a 'pass' situ-a-a [sitnaa] 'send away'
nuku-a 'wake up' (intr.) nukcua-a [nukwaa] 'wake up' (tr.)
ureri-a 'be painted' ureria-a [urerjaa] 'paint'

b) Elision of root-final /«/ or /u/ and suffixation of -a:

nalu-a 'fall' nalu-a-a [nalaa] 'drop'
turu-a 'arrive' turu-a-a [taraa] 'make arrive'
hakuru-a 'be dry hakura-a [hakuraa] 'dry'
mitu-a 'come out' mit-u-a-a [mitaa] 'get out'

c) Merging of /ni/ to /jia/ before suffixation of -a:

ini-a 'go.up' ini-a-a [ijiaa] 'make go.up'
sini-a 'sleep' sijia-a [sijiaa] 'make sleep'
toni-a 'sound' -» toni-a-a [tojiaa] 'make sound'
najni-a 'be able' najni-a-a [najjiaa] 'finish'

i) Suffixation of -la:
ahi-a 'get drunk' ahe-la-a [ahelaa] 'make drunk'
eri-a 'speak' erela-a [erelaa] 'make speak'

The relatively strong variation in these alternations plus the low number of
attested examples may imply a highly lexicalised status of the causative suffix
-a. This is also supported by forms such as the example hanoaa 'give birth' (Lit.
'make light'), as the semantic relation between 'clear up' and 'give birth' is not

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Causative 617

straightforward. The lexicalised status of verbs with the causative -a is also is

confirmed by a few more examples that exhibit morpho-phonological
irregularities. While their present structure suggests that -a probably was the
underlying causative morpheme at an earlier developmental stage, there is
another important difference to the verbs listed in (887): two of the base verbs
in (888) are transitive. Therefore, these verbs must be regarded as completely
lexicalised examples of causativised verbs.

(888) Lexicalised forms of-a

ku-a 'drink' kukioa-a [kukwaa] 'make drink'
ra-a 'receive' raera-a [raeraa] 'make receive'
mihiito-a 'be hungry' -> mihiitero-a [mihiiterwaa] 'make hungry'

In addition to these verbs, there is a verb for 'feed' that is not morphologically
related to any verb for 'eat'. While the meaning of initial part of hiraatia 'feed /
give to eat' is unknown, the second parts corresponds to the verb for 'give'
(tia).
The argument structure of verbs causativised with -a corresponds to those
that carry the suffix -erate: the new argument (causer) becomes A, whereas the
old S argument shifts into Ο position. Since the resulting clause does not exceed
the "regular" valency of two participants, the use of the postposition ke is not
required.

(889) Valency increase with causative suffix -a

a) Intransitive: ateroa 'be tired'
surutn=te atero-e katQa
[quickly=FOC] [get.tired-3ps/E] [man]
[ADV] [V] [S]
'The man became tired quickly.'

Causative: ateroaa 'tire'

kanu atero-a-a amiane-naa hatai
[lsg] [get.tired-CAU I -3ps/A] [work-NOM] [much]
[Ο] [V] [A] [ADV]
'The work has tired me a lot.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
618 Valency changing mechanisms

b) Intransitive: ewa 'enter'

mhjauria eno-a akau=ne
[don't] [enter-NTR] [water=NEGF]
[INTR] [V] [S]
'The water must not enter!.'

-> Causative: enoaa 'make enter'

tgae kanaanaj-uru eno-a-e-lu
[also] [child-PL] [enter-CAUl-3ps/E=REM]
[CNJ] [Ο] [V]
'He also made the children enter.'

16.4.3 Differences between the two causative forms and their combination

There is one significant aspect by which the function of the causative markers
-a and -erate can be distinguished. The form containing -a implies "direct
personal involvement" of the causer, while -erate points at "indirect" cause, i.e.
the causer sends someone else to give the order for the action. For instance, the
verb tur-a-a is translated as 'make arrive' and implies that the subject (A)
himself'makes arrive' the respective O. In contrast, the form turu-erati-a states
that A has someone arrive by giving the respective command to a third person,
i.e. 'he has him come'. However, there is one complication involved: it is not
uncommon to find the two causative suffixes in a sequence, occurring on the
same verb. Typically, such a form is also interpreted as 'having someone else
command somebody to do something'. This is exemplified in (890): while it is
evident that the form in (890b) implies the direct involvement of the causer, the
combination of both suffixes as in (890c) implies that the causer put someone
else in the position to make the man enter. The semantic concept to which this
double causative refers may be tagged 'indirect causation'.

(890) Two causative forms on the same verb

a) Intransitive:
eno-a / u-a
enter-3ps/A come-3ps/A
'He entered.' 'He came.'

b) Causative -a:
katQa eno-a-a / katga u-a-a
man enter-CAUl-3ps/A man come-CAUl-3ps/A
'He (himself) made the man enter.' 'He (himself) made the man come.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Causative 619

c) Both causative suffixes:

katQa eno-a-erati-a / katQa u-a-erati-a
man enter-CAUl-CAU2-3ps/A man enter-CAU 1-CAU2-3ps/A
'He had the man enter.' 'He had the man come.'

The meaning of the double causative is further confirmed by a number of

examples from the NT in (891). It is evident that Herod (cf. (891a)) did not
personally engage in binding and sending John to prison. Similarly, the
commander who sent Paul to Felix is also cited in the same verse as telling his
people tur-a-ni-tfiH-t<x=ra! ('arrive'-CAUl-DSTL-IMP-PL=EMF) 'Go and
make him arrive!'

(891) Double causative in the NT

a)
karesero kuane hwäa eno-a-erate=lu, kadena he
[prison inside] [PSN] [enter-CAU l-CAU2:3ps/E=REM] [chain VLI
[PP] [Ο] [V] [DEP]

itQuarate-1 herodesi
bind-PRT] [PSN]
[A]
'Herod had John bound with chains and had him sent to prison.' [NT: Matthew 14:3]

b)
goberenadoro fioerisi rem fwaboro turu-a-erate hetau
[governor PSN place] [PSN] [arrive-CAUl-CAU2:3ps/E HRS]
[LOC] [Ο] [V]

mi komäadäati
[that commander]
[A]
'That commander had Paul arrive at Felix's place.' [NT: Acts 23:24]

It is evident that a doubly-causativised verb involves an additional argument, as

it includes the third participant, i.e. the one who is delegated by A. However, at
this stage there are no attested examples where all participants would be
expressed by an overt NP. Therefore, in (892) the possible argument structure
of such a verb can only be sketched, by using an elicited example, similar to the
one in (891a). Similar to examples with transitive verbs causativised through
-erate, an additional argument is marked with the valency-increase marker he.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
620 Valency changing mechanisms

(892) Hypothetical example with all participants

karesero kuane hwäa eno-a-erate-lu herodesi

[prison inside] [PSN] [enter-CAUl-CAU2:3ps/E=REM] [PSN]
[PP] [Ο] [V] [A]

komäadäati he
[commander VLI]
[E]
(To mean): 'Herod made the commander send John to prison.'

In summary, the less productive suffix -a implies a direct involvement o f the

causer in the described action, whereas the highly productive suffix -erate can
imply that the causer delegates the action. Since -a is only combined with
intransitive verbs, the question remains how "direct involvement" would be
marked with a transitive verb. As a matter o f fact, the interpretation o f "indirect
involvement" is determined only when -erate is attached to an intransitive verb,
while a form that combines a transitive verb with -erate can be interpreted as
referring to direct or indirect involvement, depending on the context. This is
illustrated in (893): the causative form in example (893a), where indirect
causation is implied by the context, could also be understood as direct causation
in a different context, e.g. if he himself fed the snake to the chicken. In (893b),
the reading as direct causation is only given by the context, as the default
interpretation would be as indirect causation.

(893) Direct vs. indirect causation with transitive verbs

a) Indirect causation:
akarto h-tgati-a
[snake] [eat-CAU2-3ps/A]
[O] [V]
'He made them eat snake.' [Context: he had someone give the dead snake to the
chicken.]

b) Direct causation:
kana-atga ki-ji ki-tQati-a
[ 1 pl/in-only eat-NOM0bj] [eat-CAU2-3ps/A]
[O] [V]
'He made [us] eat what we don't [normally] eat.' [Context: He (himself) gave us canned
beef.]

The table in (894) summarises the differences between the two causative
morphemes according to parameters suggested by Dixon (2000:62). It becomes

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Causative 621

evident that Urarina distinguishes just between four of the nine criteria, which
involve state/action, transitivity, control, and directness.

(894) Differences between causative forms according to Dixon (2000)'s parameters

Parameter -a -er ate

1 Applies to state/action verbs Action verbs Action verbs, some
stative verbs
2 Transitivity With intr. verbs only With trans, and intr.
verbs
3 Causee has/lacks control over Yes Normally yes; not
the activity with stative verbs
4 Causee does it Unspecified, depends Unspecified, depends
willingly/unwillingly on context on context
5 Causee is Unspecified Unspecified
completely/partially affected
by activity
6 Causer acts Directly Depends on context
directly/indirectly
7 Causer achieves result Unspecified, but all Unspecified, but all
intentionally/accidentally examples imply examples imply
intention intention
8 Activity happens Unspecified, but all Unspecified, but all
naturally/with effort examples imply examples imply
effort effort
9 Causer is Unspecified, depends Unspecified, depends
involved/uninvolved in on context on context
activity

16.4.4 Additional strategies to express causativity

An alternative way to express causativity is by the use of the verb letoaa ' s e n d ' ,
which may imply an order, depending on the context. The structure of the
examples in (895) involves ' s e n d ' used with a subordinate clause with the
consecutive conjunction 'so that'. Since no morphological evidence for
causativity is present, this kind of construction may be called "syntactic
causativity". Semantically, this construction implies a non-direct involvement
of the causer.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
622 Valency changing mechanisms

(895) Syntactic causative

a)
kanu=na letoa-ü lenone enoala-a=ne kujyia
lsg=FOC:lsg send-lsg/E food cook-3ps/D=SUB so.that
Ί sent [someone] to cook a meal.'

b)
eene letoa-anu lenone enoala-a=ne kujßa
woman send-lsg/A food cook-3ps/D=SUB so.that
Ί sent the woman to cook a meal.'

16.5 Valency increase with "ke"

In §9.5 it was demonstrated that Urarina has no ditransitive verbs. Whenever an

additional argument is added to the clause, it is marked by the postposition ke.
This was also illustrated in §16.4, where the causative as the most common kind
of valency-increasing construction was described. In this section, I will discuss
other contexts in which ke occurs.
Any argument that occurs in addition to the lexical arguments of a verb is
marked with ke. The examples in (896) illustrate how an argument is added to
an intransitive clause. Conversely, one may argue that the added constituent is
an adjunct that is accompanied by the postposition ke, as there is no way to
formally distinguish an adjunct from an additional argument. In the same way
as the phrase with ke, adjuncts involving any other postposition may be added
into a clause. However, from a semantic point of view, a NP marked by ke is
usually similar to an Ο argument, referring to a goal, patient, or recipient,
whereas this is not the case for an adjunct that describes a location or has a
comitative function. This is illustrated by the examples in (896), where most
verbs have transitive counterparts in English and are difficult to gloss; however,
these are intransitive in Urarina, which is proven by the fact that they cannot
take an Ο argument. Instead, the additional N P is marked by ke, with the effect
of increasing the valency of the respective verb. Note that in (896c), ke occurs
without an overt NP as it refers to 'this [fish]', which functions as the object of
'take'. With the intransitive verb 'eat', the valency increase marker refer to the
fish as being the meal, which could literally be translated as 'have a meal of this
[fish]'. Alternatively, 'eat' could be present without ke, but in this case, it would
be unspecified of what the meal consists.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Valency increase with "ke" 623

(896) Valency increase of intransitive verbs with ke

a) With 'become aware':
hanulari ke nalpakuri-a
[jaguar VLI] [become.aware-3ps/A]
[Ε] [V]
'He became aware of the jaguar.'

b) With'live':
leotQa itQau-naa ke itpau-uru-a
[other live-NOM VLI] [live-PL-3ps/A]
[Ε] [V]
'They lived a different life.'

c) With 'eat / have a meal':

kaa ama-ϊ, i=tQuhe-uru ajpa ewala-~i ke lenone-m-u=ra
[this take-PRT] [2sg-child-PL with cook-PRT][VLI] [eat-DSTL-IMP=EMF]
[DEP1] [DEP2] [Ε] [V]
'Take this [fish], cook it with your children and go to have a meal of it!'

d) With 'give a name':

aj-a=m rihihei kuraa ke kura-hakwa-t
[AUX-3ps/D=SUB like] [name VLI] [give.name-DSTR-PRT
[DEP] [Ε] [V]

ne-ur-ene
be-PL-NEG:3ps/E]
'They are not giving each of them names like that.'

However, the insertion of ke can also imply other functions than the one
described above. There are examples for a range of different usages that cover
location, comitative, instrumental, and temporal functions. The major question
that arises from this is, what is the difference between an additional NP marked
with ke as compared to a NP marked by other postpositions? In fact, all
functions implied by ke in the examples in (897) to (900) are already covered by
existing locational postpositions (cf. §5.7.1), as indicated for some of the verbs
(cf. (897b,d)). As a possible interpretation, the use of ke expresses a more direct
"impact" of the verb with regard to the NP, which is to say that, for instance in
(897a), refers to the tree more as a goal, as opposed to the shed, where the PP
with asae functions as an adjunct. If this were correct, the function of ke as
introducing a new argument with close semantic relation to the verb would be
conceivable. In (897c,d), the key difference between the two forms of urinating

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
624 Valency changing mechanisms

can be described by the context: when a goal is aimed at, as in the case of
'urinate on this earth' (which could easily be replaced by a situation such as
'urinate on a flower'), ke is used; in contrast, a postposition such as htane
'inside' is preferred in order to express the unintentional character of the action.
In fact, is not specified whether this PP should be translated as locational or
directional ('into them' vs. 'inside them').

(897) Locational function of ke

a) With 'sleep' + ke:
ekuu enua ke sini-t hano-a
[above] [tree VLI] [sleep-PRT] [spend.night-3ps/A]
[ADV 1 ] [El] [VI] [V2]
'He spent the night sleeping on top of a tree.'

b) With 'sleep' + asae:

loanari asae sim
[shed under] [sleep:3ps/E]
[P] [V]
'She slept under a shed.'

c) With 'urinate' + ke:

ka=atane ke hja-em-akaanu kanakaanu=ta
[this=Iand VLI] [urinate-NEG-lpl/ex] [lpl/ex=FRS]
[Ε] [V] [S]
'We do not urinate in this world.'

d) With 'urinate' + htane:

ku hja-1 htane nianatu hja-ure
[there urinate-PRT] [inside] [even] [urinate-3pl/E]
[DEP] [P] [ADV] [V]
'Urinating there, they even urinated in them [the clothes].'

In (898), ke is used with instrumental function. In all examples, it is attached to

a NP that refers to a tool. In this context, its interpretation as "goal-like" does
not apply. However, there is only one other postposition that can take
instrumental function: htane 'inside' is used with this function only in a few
attested cases (cf. (898c)). Otherwise, ke is the regular instrumental marker.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Valency increase with "ke" 625

(898) Instrumental function of ke and knave

a) Instrumental function of ke:

saabere ke heräe sau+saune-he-u
[machete INST] [slowly] [RED+cut-CNT-IMP]
[P] [ADV] [V]
'Cut slowly with the machete!'

b) Instrumental function of ke:

raahemane hit ς, ana ha-1, ne=batohwi te-1
[self blowgun make-PRT] [3ps=bow make-PRT]
[DEP1] [DEP2]

naaohwa-l=te ke nesaritQaete-kure=lu
[ENUM-PRT=FOC] [INST] [hunt-3pl/E=REM]
[DEP3] [Ρ] [V]
'Making their blowguns and their bows all that themselves, they hunted with them.'

c) Instrumental function of kuane:

lejhii hitgana-at$a kuane=te nesaritpäete-kure kat$a-uru
[one blowgun-only inside=FOC] [hunt-3pI/E] [man-PL]
[P] [V] [S]
'The people hunted with their blowgun only.'

In a few cases, ke may also take comitative function, as illustrated with the
transitive verb hasisia ' m i x ' in (899a). In (899b), the noun for ' w i f e ' is turned
into an intransitive verb ('have a wife') through the verbaliser -ki. In order to
express with whom the protagonist of the story is married, the person is marked
with ke. However, in both examples of (899), the use of ke could also be
interpreted as a type of instrumental function.

(899) Comitative function of ke

a)
seboja ke hasiite-l ki-a lura+kuturi=jK kujjia
[onion VLI] [mix-PRT] [eat-3ps/D] [bald-head=SUB] [so.that]
[Ρ] [VI] [V2] [A] [CNJ]
'... mixing it with onions so that the bald one may eat it'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
626 Valency changing mechanisms

b)
ra'-a hetau lejhii kat$a lomaj beene ke komasaj-ki-1 tie-ma
[be-3ps/A HRS] [one man] [PSN female VLI] [wife-VBL-PRT be-NOM]
[VI] [SI] [P2] [V2]
'There was one man who had a wife [called] Lomai.' [Lit. 'who was "be-wifed" with a
wife Lomai']

In example (900), ke takes temporal function. This is an isolated example and

does not seem to characterise the typical function of ke as a valency increaser.
Other temporal postpositions such as bana 'at the time when', baja 'after', or
ahirna 'before' are normally used in this context. The use of ke here could be
due to the fact that '2002' is a loan from Spanish and thus, the valency increaser
is used as a default marker.

(900) Temporal function of ke

kaa kargo=te saatono-re=i dosmildos ke

[this mission=FOC] [end-IRR:3ps/E=ASS] [2002 VLI]
[S] [V] [P]
'This mission will end in 2002.'

Note that ke cannot occur twice in a clause, i.e. in hypothetical examples that
have more arguments than ke could introduce, other postpositions would be used
to add additional constituents. A further aspect worth mentioning is that ke is
also used with nominalisations, as shown in (901). The word for 'story' is based
on the intransitive verb for 'speak', derived by the nominaliser -naa (cf. §6.6.3).
Subsequently, the NP 'about X' takes the valency increaser ke.

(901) ke with nominalisations

hauiße-naa ke ere-naa
[get.lost-NOM VLI] [speak-NOM]
[PP] [NOM]
'Story of a lost one'

The instrumental function of ke can also apply in clauses that involve transitive
verbs. Its meaning with these can be instrumental or refer to goal.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Valency increase with "ke" 627

(902) Goal/ function of ke with transitive verbs

a) With 'make':
satu enamanaa-knru=te kam kuruatahane-re=l iglesia it$a-naa
[all young.man-PL=FOC] [lpl/in] [help-IRR:3ps/E=ASS] [church make-NOM
[Α] [Ο] [V] [PP]

ke
VLI]
'All the young people will help us (with) building the church.'

b) With 'instruct/teach':
kanaanaj hanoa-naa ke eerie rela-a kati=jie
[child give.birth-NOM VLI] [woman] [teach-3ps/D] [black.monkey=CND]
[PP] [Ο] [V] [A]
'How the black monkey instructed the woman in giving birth to a child'

c) With 'call/invite':
lejhii kat$a=te miirjka ke katga-uru kuti-hjal amia-rate
[one man=FOC] [minga VLI] [man-PL] [call-PRT work.CAU2:3ps/E]
[A] [PP] [Ο] [V]
O n e man invited the people to a minga [working session] and made them work.'

As observed in intransitive clause, the type of NP tends to be responsible for the

meaning of ke. In (903), all nouns involved here are used as tools, even when
not explicitly mentioned (but evident from the context). Example (903a)
illustrates the difference very clearly, as the verb 'make' is also used in (902a),
but with a different type of NP. However, this information alone is probably not
enough in order to determine the function of ke, as additional information by the
context may be required. Also note that various verbs can be used to express
'make', as illustrated in (903a-c): haa refers to 'make' in the sense of 'build',
but is also used to mean 'cultivate' in combination with 'field'; tia also means
'give', but can stand for 'produce', and it$aa is the general term for 'do',
'make' and can be used for any of these meanings.

(903) Instrumental function of ke with transitive verbs

a) With 'make' (itQaa):

aheri+ku-teru-atQa ke itulere itea-kure
[stone+ASC-axe-only INST] [all.kinds] [make-3pl/E]
[PP] [Ο] [V]
'They made various things with stone axes only.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
628 Valency changing mechanisms

b) With 'make' (tia):

aheri+ku-teru ke n=enanihja te-ϊ
[stone+ASC-axe INST] [3ps=canoe] [make-PRT]
[PP] [O] [V]
'making their canoes with stone axes ...'

c) With 'make' (haa):

he n=ukwana ha-uru-rehete
[INST] [3ps=field] [make-PL-HABl:3ps/E]
[PP] [Ο] [V]
'They made their field with them with [stone axes].'

d) With 'fish':
nehouuri+ofwa ke hetau=te ate kuhiuu-ttre
[garabata+thorn INST HRS=FOC] [fish] [fish-3pl/E]
[PP] [O] [V]
'They caught fish with the thorn of the garabata tree.'

e) With 'touch':
enua ra-ι ke dada-i hoara-ü
[tree] [receive-PRT] [INST] [touch-PRT] [see-lsg/E]
[Ol] [VI] [P2] [V2] [V3]
'Taking a stick and touching him with it, I looked at him [at the jaguar].'

In (903e), all three verbs of the sentence are transitive; the postposition ke adds
a phrase with instrumental function to the verb dadaa 'touch' to imply 'with it',
in addition to the (not overtly present) Ο argument 'him'.
In summary, the valency increase marker ke can be used with different
semantic functions that include reference to instrument, goal, location, time and
comitative function. The main difficulty is to distinguish between additional
arguments, for which ke is also employed in causative constructions, and
adjuncts. As no formal criteria can be applied to this, the interpretation is not
predictable without the context. Apart from that, not all semantic functions of ke
are attested with both transitive and intransitive verbs, but this disparity might
be coincidental. At any rate, the type of noun that is accompanied by ke, plus the
context in which it is used is relevant for the function of ke.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
17 Serial verb constructions

Apart from a range of other multi-verb constructions, Urarina makes use of

serial verbs. Prototypically, these constructions involve a closed class of verbs
of direction and movement. Thus, they can be called 'asymmetrical' according
to Aikhenvald (2005). While the first verb (VI) can be transitive or intransitive,
the nature of the second verb (V2) is restricted. In the majority of serial verb
constructions, V2 is represented by the intransitive verbs ua 'come' or hta 'go'.
There are a few cases in which other intransitive verbs function as V2; these are
amuemueka 'wander around' nuhtja 'do in vain', and neheria 'almost do'. In
addition, letoaa 'send' and amua 'walk' are "possible" as V2, but it must be
noted that all examples with these verbs are elicited and do not occur in natural
language. Thus, 'come' and 'go' must be regarded the actual prototypes that
occur as V2 in a serial verb construction (abbreviated to SVC here).
The primary nature of Urarina SVCs is that all verbs involved share the same
subject, which automatically excludes switch-subject serialisation. The main
feature of Urarina SVCs is that the first verb bears the neutral marker -a (or
allomorphs), which also functions as the suffix used in the citation form of
verbs. Since the neutral marker cannot be considered a linking morpheme,
SVCs in Urarina conform to the definition of serial verbs as "sequences of verbs
that act together as a single predicate without any overt marker of coordination,
subordination, or syntactic dependency of any sort" (cf. Aikhenvald 2003:1).
The only case in which object sharing would be relevant is in the elicited
examples that involve the transitive verb letoaa 'send'. With the other verbs in
second position, object sharing is not an issue, since these are intransitive.
However, even with elicited examples where letoaa is in final position, object
sharing does not apply: in sini-a letoa-a ('sleep'-NTR 'send'-3ps/A) 'he sent
him (to) sleep', the objects of the two verbs do not coincide.
In general, serial verbs refer to one event (or parts of it), which is reflected
by the fact that all morphology marked on V2 has the entire construction as its
scope. While most categories are marked on V2, some may occur on the first
verb (cf. §17.1). VI is usually less morphologically marked, but can be
characterised as the semantic head of the construction, while 'go' and 'come'
mainly describe the direction or result of the event.
The properties of the verbs involved in a SVC are discussed in §17.1 and
§17.2, while §17.3 deals with the semantic structure of these. Further properties
of SVCs, which includes questions of wordhood and contiguity, as well as some
exceptional cases of SVCs, are investigated in §17.4. A further aspect
considered in this section analyses the differences between SVCs and other

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
630 Serial verb constructions

constructions that involve more than verb. Given that Urarina SVCs exhibit a
number of untypical properties not normally predicted for this construction
type, the summary in §17.5 discusses the question in what way these deviate
from the typological criteria for SVCs according to Aikhenvald (2005).

17.1 Properties of VI

The first verb in a SVC comes from an open class. (904) lists examples in
which VI is either transitive or intransitive. 74 Stative verbs of any kind do not
occur in this position, but the copula can function as VI as well (cf. (904c)). As
an important feature, all suffixes marked on VI have scope over this verb only,
rather than over the entire SVC.

(904) Transitivity of VI in SVCs

a) Transitive VI:
obaa-hei hetau, alau hoara-a ku-e mi unee
get.angry-PRT HRS spider.monkey [see-NTR go-3ps/E] that kinkajou
'The kinkajou got angry and went to see the spider monkey.'

b) Intransitive VI:
mi rihihet ku-he-t, ßäe kahia turu-a u-akaanu=lu
that like go-CNT-PRT already here [arrive-NTR come-lpl/ex=REM]
'Going like that, we arrived here'

c) Copula:
kuraanaa m-a=te u-e
chief [be-NTR=FOC come-3ps/E]
'He came to be chief.'

The above examples illustrate that VI does not bear any inflectional suffixes;
the neutral suffix -a cannot be regarded as such as it does not specify person,
aspect or any other grammatical features. However, it can be attached to a verb
stem, i.e. to a root with certain verbal suffixes. Examples for this are illustrated
in (905), where the causative suffixes -a and -erate precede the neutral marker.

74
For clarity, all SVCs in this chapter are marked by square brackets.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Properties of VI 631

(905) Derivational suffixes in V1

a) Causative -erate:
kwitukn-erati-a ku-ure
[know-CAU2-NTR go-3pl/E]
'They went to let him know.'

b) Causative -a:
ßaara=te kann reue kaa katga turu-a-a u-i-tge
2pl=FOC lsg place this man [arrive-CAUl-NTR come-2ps-PL]
'You have come to bring this man to my place.' [NT: Luke 23:14]

It should be noted that the scope of the causative changes if it is attached to V2,
as shown in (906): In this case, the presence of -erate on 'go' implies that 'he
sent someone to go' and do the job.

(906) Causative marking on V2

ahaena su-a ku-erate
grass [kill-NTR go-CAU2:3ps/E]
'He send him to weed the grass.' [Lit. 'He made him go to weed grass.']

A further suffix that can be marked on VI is the continuous aspect marker -ahe
(or allomorphs). However, other aspect markers are not attested on V1. The two
examples in (907) illustrate the use of this suffix, but no examples are attested
for other aspect markers in this position. Note that the asymmetry in aspect
marking is significant: if the aspect were marked on the motion verb (V2), this
would imply a continuous way of 'going', which is not the case. It would be
difficult to state whether the aspectual value applies to the whole construction
or to one of its components only. In a way, this asymmetry is a challenge to the
view that the construction represents a single predicate. This issue will be
further discussed in §17.5. Morphological marking on V2 is expressed by all
other suffixes, most typically including mood/tense, person, and polarity.

(907) Aspect marking on VI

a)
ama-uru-a=ne hana situ-hi-a ku-a lejhii katQa
take-PL-3ps/D=SUB when [pass-CNT-NTR go-3ps/A] one man
'When they took him, a man went passing by.' [NT: Luke 23,26]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
632 Serial verb constructions

b)
ahaenaa su-ahi-a=ne ku-ü
grass [kill-CNT-NTR=FOC: 1 sg go-1 sg/E]
Ί went to weed grass.'

Another feature that can be marked on VI is reduplication, which may involve a

range of different semantic contents, depending on the type of reduplication and
on the context (cf. §10.3).

(908) Reduplication in VI

a)
ne-temia+temia-ka ku-ure
[ITR-RED+lean.together-NTR go-3pl/E]
'They went to lean together there.'

b)
ihja ke hetau takaaka+takaako-a ku-ure
lighthouse.tree VLI HRS [RED+sit-NTR go-3pl/E]
'They went to sit down [as birds] on a lighthouse tree.'

There are a few isolated examples in which other categories are marked.
Example (909a) shows the plural object suffix as occurring on V I . On the one
hand, this is plausible since this category cannot be marked on the motion verb
(as 'come and 'go' are intransitive); on the other, the examples listed in (909)
appear to be exceptional. (909b) illustrates the use of the velocity suffix on V I .
Note that the meaning of 'quickly' does not extend to V2 'go' here. Another
category which can be marked on VI is the passive suffix -noi followed by a
copular auxiliary (cf. §16.1.2).

(909) Other categories marked on V1

a) Plural object:
te-naha-a ku-ure
[give-PLO-NTR go-3pl/E]
'They went and delivered them.'

b) Velocity:
kwara-uri-a ku-a
[see-RAP-NTR go-3ps/A]
'She went to see him quickly' [E.g. to briefly say good-bye]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Properties of V2 633

c) Passive:
muku-nohiui-ni-a u-e
[catch-PASS=be-NTR come-3ps/E]
'It came to be caught.' [about a chicken that was easily caught]

Urarina makes use of an increasing number of borrowings from Spanish (cf.

§23.3). Interestingly, this does not only involve nouns, but also verbs. The fact
that the choice of verb for V I in SVCs is rather free is further supported by the
fact that even borrowed verbs can occur in this position. In (910), the verb
lintereniaa, which is derived from the noun Unterem 'flashlight' (from Spanish
linterna) and which can be fully inflected for person, is used with the neutral
marker in V I position.

(910) Loans as VI

hanotii lirderema-a ku-uru-a=na hau

night [hunt.with.flashlight-NTR go-PL-3ps/D=SUB] because
'... as they went hunting with a flashlight at night,...'

In fact, there are no particular semantic restrictions for verbs occurring as the
first verb in a SVC. The examples in (911) illustrate the use of verbs that
express sound (tonia) or refer to 'speak' (naa).

(911) Verbs of sound as V1

a)
jiae asaa-hi-ri-1 tom-a ku-a
already inside-CNT-RAP-PRT [sound-NTR go-3ps/A]
'It already sounded deeper [as the fruit fell into the water].'

b)
sehtee na-a ku-a
(sound) [say-NTR go-3ps/A]
'It sounded [deeper, like] "sehtee".'

17.2 Properties of V2

As mentioned above, only very few verbs can function as V2 in an SVC. Apart
from ' g o ' and ' c o m e ' , only one other verb with an independent lexical content,
amue+muee-ka 'wander around' is attested in natural text. In elicited examples,
also the verbs amua ' w a l k ' and letoaa 'send' were accepted as V2 of a serial
verb in some contexts. In the elicited cases, letoaa appears to take a purposive

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
634 Serial verb constructions

function. The example kau-a + letoa-e ('return home'-NTR + 'send'-3ps/E) was

accepted reluctantly and corrected to kau-a=ne kujjia letoa-e
('return.home'=SUB+'so.that'+'send'-3ps/E) 'he sent him home', which
literally is a purposive construction 'he sent him so that he would return home'.
Interesting is the fact that amue+muee-ka as the reduplicated (but probably
lexicalised) form of 'walk' functions as V2, but not normally the base verb
amua. Other motion verbs (e.g. surua 'run', turua 'arrive') cannot occur as V2
in a SVC.

(912) Use of amuemueeka 'wander around' as V2

katpa rela-a amuemue-kuru-a=lu

man [teach-NTR wander-PL-3ps/A=REM]
'They wandered around to teach the people.'

There are two verbs whose form in multi-verb constructions corresponds

exactly to an SVC, although their semantic structure is different from the other
verbs occurring as V2 with this function. The verb nuhija 'do in vain' (with
irregular Ε-form nuhoe), which also functions as a single full verb, can be
preceded by a verb that is marked only with a neutral form. The semantic
content of mihtja involves 'in vain', with a connotation of speaking with
offence, depreciation, shame, or "speaking as if you didn't believe in that
person" (cf. (913)).

(913) nuhija 'in vain' as V2

a)
suru-a nukuj-a=ne hana, u-a
[run-NTR in.vain-3ps/D=SUB] when die-3ps/A
'When he ran in vain, he died.'

b)
ka=nesari kurenia niririhi-a nukue-re-i=ta
1 sg=booty for [do.like.that-NTR in.vain-IRR-2ps=FRS]
'You would behave like that in vain for my booty.' [To mean: 'If you (the jaguar)
attacked me to steal my booty, you would not stand a chance.']

The other verb is neheria 'almost do', which differs from nuhija in that it does
not occur as a full verb; i.e. it is always combined with another verb that
precedes it. An interesting lexical observation is the possible diachronic relation
between neheria and neheratia 'get angry', which morphologically would be
identical to the causative form of 'almost do'. Thus, getting angry, which is one
of the worst offences in Urarina culture, would be equalled to 'make almost do'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Properties of V2 63 5

(something terrible, presumably).75 The examples in (914) illustrate the use of

neheria as V2 in a SVC.

(914) neheria 'almost do' as V2

a)
lureri-jtemaahei hetau=te komo-a nehere
house-also HRS=FOC [fall-NTR almost.do:3ps/E]
'The house also almost fell.' [Context: When the giant spirit entered the house]

b)
hanotü ku-uru-a=ne hau, hvara-elanaala tQoae bauleneto-a
at.night go-PL-3ps/D=SUB because see-PRV above [hit.down-NTR

nehere-kure enardhja kuane

almost.do-3pl/E] canoe inside
'As they were going by night, they almost ran her over with the canoe.'

One may feel tempted to analyse neheria as a combination of the subordinate

marker =ne plus the control/complement verb for 'want' (heria), which would
give the sentence in (914a) the possible reading as 'the house wanted to fall'.
However, this analysis does not account for (914b), where the neutral marker
-a, which is identical to 3ps/A, does not agree with the 3pl form of V2. In
constructions with a control verb, person must also be marked on VI, if the
subject is different. With same subject complement constructions (as in SVCs),
the infinitive suffix -na is attached to VI (cf. §20.2.1). Examples to illustrate the
difference between the three constructions are shown in (915):

(915) SVC with neheria vs. complement clause

a) SVC:
nalu-a neheri-tQäu
[fall-NTR almost.do-lsg/A]
Ί almost fell.'

b) Complement clause (same subject):

nalu-na heri-tgau
fail-INF want-lsg/A
Ί wanted to fall.'

75
The word for 'enemies' literally means 'those who get angry'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
636 Serial verb constructions

c) Complement clause (different subject):

nalu-a=ne heri-tgäu
fall-3ps/D=SUB want-lsg/A
Ί wanted him to fall.'

While a restricted range of morphemes can occur on V I , all other grammatical

categories are marked on V2. The following examples illustrate this through the
use of person/number, mood/tense, imperative, polarity, and probability. All of
these features have the entire SVC as their scope.

(916) Inflectional suffixes on V2

a) Person and number:
ßäe kahia turu-a u-akaanu=lu
already over.here [arrive-NTR come-lpl/ex=REM]
'We already came to arrive over here.'

b) Irreal is/future tense:

ßoaelu netohwel ita fiuarue+fzuaru-he-l,
forever until REC RED+cut-CNT-PRT

mi rihei itQau-a ku-ri-ki-t(%=~i

that like [live-NTR go-IRR-2ps-PL=ASS]
'You will go on living like that, cutting each other forever.'

c) In-law talk and jussive:

tQäe ßäe ne-molo-a raj misi=ße, amu-a ht-ana-ipe
also already ITR-cut-3ps/D POSS umbilical.cord [walk-NTR go-ILT-JUSS]
'When his umbilical cord comes off, he (my compadre) ought to go hunting'.

d) Negation:
heriane katga-uru beelaj-a u-ri-ßa=l
probably man-PL [give.as.gift-NTR come-IRR-NEG:3ps/A=ASS]
'Probably he will not come to make gifts to the people [but to exploit them].'

e) Probability:
kwara-a ku-kaj-a
[see-NTR go-PRB-3ps/A]
'He probably went to see.'

With the imperative marked in SVCs, which always occurs on V2, only ua
'come' can be used as V2, such as in 'come to see!' (917). The use of kua is not

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Properties of V2 637

possible: A sentence such as 'go to see!' must be expressed by use of the distal
form (cf. (917b)).

(917) Imperative in SVCs

a) With 'come':
hvara-a u-u
[see-NTR come-IMP]
'Come to see!'

b) With 'go':
hvara-ni-u
see-DSTL-IMP
'Go to see!'

c) Ungrammatical form
*hoara-a ku-u
see-NTR go-IMP
[To mean]: 'Go to see!'

The same rule applies to the hortative form, where the imperative can be used if
' c o m e ' is V2 of a SVC, but not if ' g o ' is in this position. The situation is
different for the jussive form, as it is not compatible with the distal suffix; In
this case, hta ' g o ' may be used as V2 in a serial verb (cf. (918)). However, this
is only possible with the innovative suffix -ijie, but not with the older suffix
-mti (cf. § 15.1.3). Other verbs as V2 are not attested with the imperative.

(918) Jussive in SVCs

enejtQU su-a ku-iße aka

monkey [kill-NTR go-JUSS] 3sg
'He ought to go kill monkeys!'

In a similar way, clause-level derivation is marked on the second verb of a

SVC. This includes the participle form and nominalisation, examples for which
are given below. In the same way as other morphological markings on V2, these
features apply to the whole SVC as a scope.

(919) Participle and nominalisation marked on V2

a) Nominalisation:
mi fwaru-a=ne haanuke mi-tpa ra-a ku-i
that cut-3ps/D=SUB PURP that-only [receive-NTR go-NOMSbj]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
638 Serial verb constructions

ku hau-i
there throw-PRT

'throwing away what he had brought in order to cut her ...'

b) Participle:
amu-a ku-ι aräala ru-a
[walk-NTR go-PRT] tapir find-3ps/A
'When he went hunting, he found a tapir.'

17.3 Semantic structure of SVCs

In most combinations of two verbs in a SVC, the overall semantic value will
have a directional content, i.e. the core event is specified by a motion verb for
'go' or 'come'. However, in many cases, the motion verb would not be needed
in order to express a motion, as this is already expressed by the semantics of
V I . For instance, the example in (920a) would be perfectly understandable
without the verb 'go', as 'return home' already implies motion (similar cases
are found with turua 'arrive', amua 'walk', sunia 'run', and ertoa 'enter', just
to name a few; accumulation of two identical verbs is not possible). However,
the use of a SVC in this context indicates the direction, as the people went
"away", which implies 'go home', rather than 'come home'.

(920) VI = verb involving motion

a) With 'go':
raatiri-i, kau-a ku-ure
leave-PRT [return.home-NTR go-3pl/E]
'They left him and went (away) home.'

b) With 'come':
ataw ari-1, eru-1, kau turu-a u-akaantt=lu
land seek-PRT find-PRT here [arrive-NTR come-lpl/ex=REM]
'Looking for land, we found it and arrived here.' [Lit. 'came to arrive here']

The key question which arises here is, what is the functional difference between
constructions that involve a serial verb and such that do not? While the
difference may consist of a slight nuance only, these can be characterised as
follows:
1. The presence of a motion verb in second position may indicate a point of
change with regard to a previous event, i.e. the beginning of 'going' or
'coming', which leads to the event described by the core verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Semantic structure ofSVCs 639

2. The fact that the core event requires motion of some kind is emphasised.
3. It is specified into what direction the motion expressed by VI is performed,
in terms of deixis, i.e. 'come' implies 'towards the point of reference',
whereas 'go' implies 'away from the point of reference'.
4. With other SVCs that have a non-directional interpretation (cf. (924)), the
presence of V2 is essential in order to express that function.

When V1 is not a motion verb (and V2 is), the semantic interpretation can be of
a directional nature, as illustrated in (921), but this is not necessarily the case.
The interpretation in this case overlaps with a purposive function ('go in order
to see'). Note that Urarina also has a purposive/consecutive construction (cf.
§20.2.4.1); the main difference between the SVC (as in (921)) and the purposive
construction lies in the directional component that is expressed by the motion
verb as V2. Purposive function is also implied by the example in (919a), where
the nominalisation 'what he had brought in order to cut her' involves
niitpa raa ht-i 'that what he went to take'.

(921) VI = verb not involving motion: directional/purposive interpretation

d$a baha-a u-i

what [ask-NTR come-2ps]
'What have you come to ask?'

A different function is implied by the example in (922), where a person

expresses his opinion about the amount of work that will be involved in cutting
a peccary into pieces, after the previous animal had caused problems. While a
directional interpretation can be excluded in this case, a purposive function is
possible. However, this particular example also suggests a resultative use of
'come', since 'relax' is the outcome o f ' c o m e ' .

(922) Purposive/resultative function of'come' as V2

kaa=te kana betaka-erati-a u-e

this=FOC lpl/in [relax-CAU2-NTR come-3ps/E]
'This is going to make us relax.'

In one example that involves the verb for 'wander around' as V2, the distinction
between directional, purposive, or resultative interpretation is even less
transparent. In (912) further above, the presence of V2 could emphasise on each
of the following interpretations, the function of which depends entirely on the
context:
a) The nature of teaching while wandering around, as opposed to staying in one
place,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
640 Serial verb constructions

b) The intention of wandering around in order to teach,

c) The event of teaching as a result of wandering around.

An aspectual function can be implied in cases where V2 is represented by a

future tense form of 'go', such as in (923). In this case, the form of V2 is
interpreted as a continuing action, even though continuous aspect is not marked
(also cf. example (916b)).

(923) Aspectual function of future form in V2

a)
laueka-1 raj beree kuhoalate, na-a
be.sitting-PRT POSS child have.another.child:3ps/E [say-NTR

ku-uru-re=l=tQa
go-PL-IRR:3ps/E=ASS=EMF]
'They will go on saying: "While his baby is [still] sitting, she is [already] having
another child".'

b)
ßauene kwara-hi-a ku-uru-re=l=tQa
silence [see-CNT-NTR go-PL-IRR=ASS=EMF]
'They will be watching in silence.' [NT: John 19,37]

A resultative reading can also be implied without the use of future tense, as is
the case in the example involving verbs of sound as VI (cf. (911)).
In summary, there are clues, such as the use of certain aspect or tense
markers that help to specify the type of semantic content of a SVC. In a similar
way, the nature of VI may point at specific ways of interpretation. The table in
(924) illustrates an attempt to classify different semantic functions based on the
semantic content of VI in a SVC.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Semantic structure ofSVCs 641

(924) Semantic functions of SVCs involving motion verbs as V2: summary

(Semantic functions indicated in last three columns: Dir = Directional; Res =
Resultative; Purp = Purposive

Semantic core Examples Gloss Cf. ex. Dir Res Purp

content (VI)
Motion turua 'arrive' (904b), X X
(920b)
Motion situa 'pass' (907a) X X
Motion kaua 'go home' (920a) X X X
Motion ima 'go up' (933a) X X X
Motion enoa 'enter' X X X
Motion mitua 'get o f f X X
Motion/Position d^uhuturua 'stand still' X X
Motion/Position netemiatemiaka 'stand still' (908a) X X
Motion/Position tohoeturua 'lie down' X X
Activity lemma 'eat' (904d) X X
Activity raa 'receive' (919a) X X
Activity hvaaraa 'see' (904a), X X
(916e)
Activity tia 'give' (909a) X X
Activity relaa 'teach' (912) X X
Activity sua 'kill' (907b), X X
(918)
Activity lintereniaa 'hunt with a (910) X X
flashlight'
Activity nakaauhwaa 'bathe' X X
Activity aria 'seek' X X
Activity mukua 'catch' X X
Activity bia 'tell' X X
Activity amiania 'work' X X
Activity ua 'die' X X
Activity rukua 'pull out' X X
Activity/State itpaua 'live' (916b) X
Position takaakatakaakoa 'sit in a line' (908b) X
Event nukueka 'become dark' X
Event hanoa 'become light' X
Sound tuma 'sound' (911a) X
Sound naa here: 'sound' (911b) X

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
642 Serial verb constructions

As the above table shows, most SVCs involve some degree of (directional)
motion, which is a natural consequence of the role played by V2. When VI is
also represented by a motion verb, the interpretation is resultative. For instance,
in 'he went to arrive', VI ('arrive') is the direct result of the motion expressed
by V2. While this may also involve a certain degree of purpose (i.e. the purpose
of going was to arrive), this is not the core content of the construction (and
therefore left unmarked in the table). With first verbs whose semantic core
expresses an activity not necessarily describing a directional motion, the overall
interpretation is different: these combinations typically refer to a purpose, apart
from the directional motion expressed by V2. For instance, in 'he went to see' it
is obvious that 'seeing' was the intended outcome (purpose). While it is usually
implied indirectly that the result of seeing was achieved, this does not appear to
be the core function (and is therefore unmarked in the table). This is even true
for an example such as 'he went to die', which implies a deliberate act
involving an agent (this example is repeated used in the NT when referring to
Jesus' death). However, it is evident that a certain degree of overlap between
purpose and result cannot be avoided. Other combinations may exclude an
actual motion, even though 'go' or 'come' are present as V2: This is the case
under two conditions:
1. V1 is intransitive, and
2. The subject of VI does not perform any action or activity.
For instance, when the underlying subject of naa 'say/sound' is an object
causing a noise, the interpretation in a SVC is resultative. In contrast, if the
subject of naa refers to an agent who is saying something, the interpretation
with 'go' would be 'he went to say', i.e. purposive. This illustrates that the
structure of the verbs involved in a SVC is predictable only to a certain extent,
while it also depends on the context.
Since the occurrence of final verbs other than 'come' and 'go' is marginal,
the semantic structure of those combinations will not be investigated here.
There also are a few cases of lexicalisation. In these examples, the overall
semantic content is not directly predictable from the semantics of its
components. The combination amua 'walk' plus lota 'go' is not intended to
mean 'go to walk', but refers to 'go hunting'. More transparent are the
combinations of raa 'receive' with 'go' and 'come', respectively. These are
interpreted as 'bring' or 'fetch', depending on the direction.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Further properties of SVCs 643

(925) Lexicalisation of SVCs

a) 'hunt':
ku neloanarite-i, kohzuanoo hetau amu-a ku-e
there make.shed-PRT next.day HRS [walk-NTR go-3ps/E]
'There he made a shed and went hunting the next day.'

b) 'fetch':
ßäe leot$a-uru=te eriua ra-a ku-e
already other-PL=FOC wood receive-NTR go-3ps/E
'The others already go to bring wood.'

17.4 Further properties of SVCs

This section investigates further aspects regarding the structure of SVC, which
includes general properties, such as the fact that they regularly occur in
dependent clauses (cf. §17.4.1). Contiguity and wordhood are two matters that
require some discussion, as is conducted in §17.4.2 and §17.4.5, respectively.
This is followed by the mentioning of two aspects that can be described as
being different from the prototypical structure of Urarina SVC: order variation
(cf. §17.4.3) and the exceptional occurrence of more than two verbs in a S V C
(cf. §17.4.4). §17.4.6 then discusses criteria by which SVCs can be
distinguished from other multi-verb constructions.

17.4.1 Embedding in larger constructions

Serial verbs can be part of a larger syntactic construction, as is shown in (926).

Without apparent restrictions, a S V C may form part of a dependent clause, in
which the final verb of the SVC is inflected according to the conjunction
paradigm for dependent clauses, followed by a subordinate marker and
(optionally) a conjunction (cf. §20.1) or control verb. As illustrated below,
subordinate marker and conjunction occur after the SVC. Also recall that in
(919b), the SVC itself functions as a participle clause.

(926) SVC embedded in a dependent clause

a) With hau:
turu-a ku-uru-a hau hetau=te, ke taka-i hetau,
[arrive-NTR go-PL-3ps] because HRS=FOC VLI meet-PRT HRS

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
644 Serial verb constructions

ku kurua-hei, ra-ure
there be.happy-PRT receive-3pl/E

'As they went to arrive, they met him and were happy, and received him.'

b) With hana\
enoa-to-a ku-a=ne heri-a=ne hana, hua hauria
[enter-INTS-NTR go-3ps/D=SUB] want-3ps/D=SUB when don't first
eno-a=ra
enter-NTR=EMF

'When he wanted to rush in, [she said], "Do not enter yet".'

17.4.2 Contiguity
If we define contiguity within a SVC to mean that no arguments can occur
between the verbs of a SVC, it is evident that Urarina serial verbs follow this
pattern. However, there are a few elements that can occur between V1 and V2,
even though this is fairly rare. While some of these are clitics (e.g. the focus
markers), also full adverbs can occur in this position. In detail, the presence of
raka 'yesterday' hajti 'still', and ht 'there' between the verbs in (927) was
confirmed to be acceptable in elicitation; however, there are no examples in
natural texts. In contrast, the adverb jiäe 'already' was not found to be
acceptable in this position.
(927) Occurrence of adverbs in SVC (elicited)

lenone ari-a hajti ku-e

food seek-NTR still go-3ps/E
'He went to still seek food.'

Elements that are attested between the two verbs of a SVC in natural speech
include the focus marker, the adversative marker niki, and a number of particles.
The fact that disyllabic (and tone-bearing) words such as niki can be inserted
indicates that phonological factors such as word length do not play a role with
regard to their insertability. However, it may show that only grammatical
elements are allowed in this position. (928) illustrates the use of these markers.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Further properties ofSVCs 645

(928) Other elements occurring between verbs of a SVC

a) Adversative:
mi lenone te-ra ari-a niki amuemu-he
that food give-AG [seek-NTR ADVRS wander-CNT:3ps/E]
'He was wandering around to seek the one who had provided the food [but would not
find him].'

b) Focus:
lenone ari-a=te kn-e
food [seek=FOC go-3ps/E]
'He went to seek food.'

c) Particle läe: (elicited)

kwara-a läe ku-a
[see-NTR CHNG go-3ps/A]
'But - yes - he went to see him.'

d) Particle combination t$u hja\ (elicited)

hoara-a t$u hja ku-a
[see-NTR CRTN just go-3ps/A]
'Well, he just went to see him.'

The function of the focus marker in this context is to draw attention to the
content expressed by the first verb and its argument. The insertion of =te is
optional, since the position of the VP already is initial, i.e. the clitic is not
needed to shift it to the front. Therefore, the difference in meaning as compared
to the example without =te is merely a pragmatic nuance (cf. §19).
In summary, the question whether SVCs in Urarina are contiguous depends
on the definition of contiguity. I am following the approach suggested in
Aikhenvald (2005), which regards a SVC as contiguous if no arguments can
intervene. Since this is the case in Urarina, it will be fitted into the typological
categorization as "contiguous".

17.4.3 Order variation

In combination with the imperative, an order variation for SVCs is observed in

76
rare cases. This variation, which is illustrated in (929), refers to a serial verb
with purposive/consecutive meaning. In (929a), the sentence is expressed by a

76
These findings are based on suggestions by consultants in elicitation.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
646 Serial verb constructions

subordinate clause with the purposive/consecutive conjunction kujßa; although

this version may reflect minor (unspecified) semantic differences to a SVC, it
was recommended as "almost identical" by the consultant. (929b) represents a
SVC of the regular type, as is described in this chapter. In example (929c),
however, the order of the verbs is inverted, which also includes the object of
" V I " (which now has become V2). Note that the morphological marking on
relaa 'teach' is identical to the one in a regular SVC - the neutral form.

(929) Order variation with imperative

a) With conjunction:
u-u kat^a-uru rela-i=jie kujßa
come-IMP man-PL teach-2ps=SUB so.that
'Come to teach the people!'

b) Regular SVC:
katga-uru rela-a u-u
man-PL teach-NTR come-IMP
'Come to teach the people!'

c) Order variation:
u-u katQa-uru rela-a
come-IMP man-PL teach-NTR
'Come to teach the people!'

While the exact conditions for this kind of inversion are not entirely clear, it
seems that the purposive aspect of it plays a central role. The variation is also
possible in declarative sentences such as Ί went to fish' or 'he went to build his
house', but it is not acceptable with other semantic types, e.g. such examples
with a mainly resultative or aspectual meaning.

17.4.4 Number of verbs involved in SVCs

Usually, each SVC consists of not more than two verbs, with one exception:
when the verb nuhija 'do in vain' is involved, three verbs can occur in
sequence. In this case, the first two verbs are each marked with the neutral
suffix, while mikuja receives all inflectional markers. Examples are given in
(930); however, it must be noted that these constructions are comparatively rare
and thus do not allow an in-depth study.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Further properties of S VCs 647

(930) SVCs involving more than two verbs

a)
tahia=na kuaka-a ku-a rtukuj-a-akaanu
over.there=FOC: 1 pi [wait-NTR go-NTR in.vain-NEG-lpl/ex]
'We did not go over there to wait in vain [since hiding during the hunt was successful].'

b)
ate ari-a ku-a nukuj-a=m hana, niji eru-i
fish [seek-NTR go-NTR in.vain-3ps/D=SUB] when nothing find-NEG:3ps/A
'When he went fishing in vain, he found nothing'

It should also be noted that in all examples with mihija, the second verb is kua,
which also functions as V2 in "normal" SVCs. This is merely "extended" by the
addition of nukuja, in order to add the information 'in vain'. However, by this
addition, kua itself is treated as a dependent verb, as it is marked with the
neutral suffix in the same way as V1.

17.4.5 Wordhood

The possibility of inserting grammatical elements between the verbs of a SVC

illustrates that these constructions may not be regarded as a single word in
Urarina, even more, as the elements to be inserted may have clitic or word
status themselves (such as niki).
On the grammatical level, SVCs have to be analysed as consisting of two
separate words, since each part can be used independently: VI bears the neutral
suffix, which is the minimal morphological element required to ensure the word
status of a verb (the bare root without the neutral marker is not a word). As
mentioned earlier, this suffix is also used on the verb in the citation form. V2 of
a SVC clearly is a grammatical word, as it carries all the grammatical
information encoded in its inflection. When a focus marker intervenes, it
attaches to V1, which does not change the structure of the SVC as consisting of
two separate words. With niki occurring between the two verbs, it is harder to
state how many words are involved since the adversative marker is as an
independent word. Thus, the SVC in this case is composed of three separate
grammatical words. When an enclitic is attached to V I , the SVC consists of two
grammatical words.
As example (919a) demonstrated, nominalisation can have the entire
construction as its scope, i.e. the sequence of two verbs is treated as one single
morphological unit. A similar situation applies to the negation of a SVC, which
has scope over the entire construction. In fact, all categories marked on V2 have

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
648 Serial verb constructions

scope over the preceding verb as well. Conversely, the use of grammatical
categories such as the causative, plural object marking or velocity have scope
only over the verb they are attached to (i.e. VI). While in one case, the
components of the SVC form a single grammatical unit, they represent separate
units in the other. Thus, the syntactic and grammatical features of a SVC do not
necessarily coincide with its wordhood status.
From the phonological point of view, a SVC also consists of two separate
prosodic words, since each of the verbs involved receives a Η tone (cf. §4.7.5).
When another element intervenes (i.e. =te or niki), it forms one phonological
unit with V I , as only one Η tone is assigned.
In summary, the wordhood factor points towards a tendency of both verbs of
a SVC to be characterised as separate words, which is in contrast to the finding
that the grammatical features of V2 apply to the construction as a whole.

17.4.6 Difference to other multi-verb constructions

The contrast between SVCs and complement clauses were already illustrated in
(915), where it was shown that the constructions clearly differ from each other
with respect to form and function. The differences between SVC and participle
constructions are much more subtle, as the example in (931) shows, which can
be contrasted with (912). While the SVC tends to have purposive reading ('in
order to teach'), the participle is interpreted more as referring to the manner of
teaching. In particular, this construction refers to two separate events, most
commonly occurring as a sequence, but with a range of possible functional
variations (cf. §20.3.4). In contrast, the SVC describes a single event. Another,
formal difference between the SVC and other constructions is of course that a
SVC represents a single clause.

(931) Participle construction as opposed to SVC (cf. (912))

katga rela-Ί amuemue-kuru-a=lu

man [teach-NTR wander-PL-3ps/A=REM]
'They wandered around teaching the people.'

The differences that distinguish SVCs from other constructions which involve
more than one verb are summarised in (932) and should also help in order to
classify Urarina SVCs from a cross-linguistic point of view.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Further properties of SVCs 649

(932) SVCs in comparison with other multi-verb constructions

Feature SVC Same sbj. Different Participle

COMP sbj. COMP constructions
clauses clauses
Shared subject yes yes no yes
Shared object n/a n/a n/a no
Shared yes yes no yes
morphological
features
Single action or yes yes no no
event
Contiguity yes yes no no
Linkage marker no yes yes yes
Shared transitivity no no no no
Single word no no no no

While wordhood and transitivity do not play a role for the distinction of
different multi-verb constructions, SVCs differ from all other construction types
with respect to at least one feature. The closest structural similarity exists to
same subject complement clauses (cf. §20.2.1), which mainly differ from SVCs
through a specific suffix that is attached to the first verb. While one could now
argue that the neutral marker be just a different sort of complementiser, one has
to be aware of the fact that there are significant differences on the semantic
level, as discussed above.
The question of how SVCs differ from monoverbal constructions can be
answered on semantic grounds, as is also illustrated by the examples in (933):
While the SVC refers to the purposive character of the event, the monoverbal
clause focuses on the result of the action.

(933) SVC as opposed to monoverbal constructions

a) SVC:
ema ke=te im-a ku-e
tree VLI=FOC go.up-NTR go-3ps/E
'He went to climb up the tree.'

b) Monoverbal clause:
enua ke=te irü
tree VLI=FOC go.up-3ps/E
'He climbed up the tree.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
650 Serial verb constructions

17.5 Discussion and summary

As noted above, a common feature of all SVCs is that they share the subject.
Object sharing is a different issue as it is technically not possible for an
intransitive verb to take an object - and V2 is intransitive. 77 In a similar way,
the asymmetry between V I and V2 types occurs on other levels. While V I can
be represented by a verb of either transitivity value, V2 is always intransitive.
So is it legitimate to bother about the overall transitivity value of the entire
construction? Based on the definition of SVCs as "a single predicate"
(Aikhenvald 2003:5), this question must be asked and the answer for Urarina is
not straightforward. From a morphological point of view, one could argue that
V 2 is the properly inflected verb in this construction, or the morphological
head. However, this does not have a direct impact on the transitivity of the
construction. In example (909b), plural object marking occurs on V I and it is
evident that it could not be marked on the intransitive V2. Therefore, it may be
concluded that the transitivity values of the verbs in a SVC are independent
f r o m each other. However, this puts the idea that SVCs are a "single predicate"
in doubt.
In a similar way, the fact that certain grammatical features can be marked on
V I raises questions about the status of Urarina SVCs as a single event. In
examples such as (905a), the causative is marked on V I and its scope is V I 'let
him k n o w ' , while V2 'they w e n t ' does not have a causative meaning. T h e only
w a y to regard this sort of construction as a single event is to focus on the
purposive function of V2. What remains undecided at this stage is in what way
this construction differs semantically from a "real" purposive construction with
a subordinate clause, which could easily replace some SVCs. T h e same
difficulty occurs with examples such as in (907), where aspect is marked on V I ,
but not on V2: If the construction as a whole is understood to have one aspect
value, this is not formally marked. In (907a), the entire SVC could be
interpreted to have continuous aspect value, but this is not the case for (907b).
T h e difference between these examples is that (907b) has purposive function,
while the function of (907a) implies motion/direction; yet other examples have
a resultative character. The secret of correct interpretation lies in the knowledge
about the context.
Concentrating on the predictable, formal side of SVCs, one would wonder
about the reason for the morphological asymmetry, i.e. what is the nature of
m o r p h e m e s that can be marked on V I ? If one considers the markers for
causative, continuous aspect, velocity, and plural object in the light of position
classes (cf. §12.2), it becomes evident that these suffixes occupy positions one

77
The example letoaa 'send' as V2 will not be considered here, since it is based on
isolated, elicited examples and is not representative of Urarina SVCs.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Discussion and summary 651

to seven - that is, they are comparatively close to the root, regarding the fact
that almost a dozen more suffixes may follow. In contrast, the types of
inflection typically found on V2, such as person, number, negation, mood, and
in-law talk occupy positions rather late in the linear order of suffixes. In a way,
the suffixes occurring on V1 also have a more "derivational" status (except for
aspect marking), which would indicate that VI is not actually to be considered
as "inflected". However, as considered in §10, I do not strictly split up these
suffixes into "derivational" vs. "inflectional" types here. In fact, recall that
person marking is the only obligatory grammatical category, which can
arguably be described as "inflectional". What is important to note is that the
categories which are most relevant to the grammatical interpretation of the
clause, are marked on V2. Negation, number, and mood may arguably be
described as such "essential" categories of inflection - and these are marked at
the end of a predicate in Urarina.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
18 Constituent order

Urarina constituent order represents a typological rarity that is found in very

few other languages. According to Tomlin (1986; cf. (934)), only five out of
402 selected languages exhibit the constituent order "OVS". However, in some
languages, the order is split with respect to transitive vs. intransitive clauses.
For instance, subject-final order may only apply under certain conditions. I will
therefore adopt a terminology that describes Urarina as having OVA/VS order,
where A is the subject of a transitive clause and S the subject of an intransitive
clause.

(934) Table based on 402 languages (Tomlin 1986)

Order Languages %
SOV 180 44.8
svo 168 41.8
vso 37 9.2
vos 12 3.0
OVS 5 1.2
osv 0 0
Total 402 100.00

Even for some languages listed as "OVS", this order is not entirely consistent.
Several Carib languages such as Bacairi (Wheatley 1973; DeAbreu 1895; von
den Steinen 1892), Apalai (Koehn 1974), or Arekuna (Armellada 1943; Koch-
Gruenberg 1924/1928; Edwards 1977) are subject to conflicting sources. Some
authors describe them as "OVS", but others as "SOV", which is partly related to
the Ergative system of these languages. Two other Carib languages that can be
convincingly characterised as OVA/VS are Hixkaryana (Derbyshire 1985) and
Trio (Carlin 2004). Interestingly, all of these languages belong to the Carib
family. For Urarina, the constituent order is consistently OVA/VS, with specific
deviations that are entirely predictable. This order applies to dependent and
independent clauses. In the independent (or "main") clause, constituents can be
focused. In this case they are shifted to the front and followed by a focus
marker. This is discussed in greater detail in §19. Furthermore, the structure of
all clauses is characterised by the following features:
- Subject and object are frequently omitted and understood by the context.
Note that no pronouns are required to refer to a 3ps object (cf. (938)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Discussion and summary 653

- The order for affirmative and negative sentences is identical, with possible
variation in the negative (see §18.3).
- Adjuncts and postpositional phrases usually occur in the periphery of the core
clause. Most typically, they precede the object (cf. §18.4).
- In many sentences, adverbs or other items such as 'therefore' are focused.
This does not affect the order of the other constituents. Gap fillers such as
ßäe already' and hi 'there' can occur in almost any position in the sentence.
However, they would not occur within constituents (cf. §18.4.2).

In order to give an overview of typical clause structure in Urarina, a (simplified)

table showing the statistical occurrence of order types in seven selected texts of
different genres and given by different speakers (a total of 799 clauses) is given
in (935).

(935) Summary of constituent order in selected texts in percent (dependent and

independent clauses) [Regular order types are shaded]
Type Frequency
V 58%
VS 10%
sv 2%
ov 21%
VA 1%
AV 1%
OVA 3%
AOV 4%

While details will be discussed further below, the following factors become
evident at first sight:
1. Clauses without any overt argument are extremely common. These make up
much over half of all examples.
2. Where arguments are present, most clauses follow the basic constituent order
OVA/VS. Their frequency is five times higher than the deviating order types
(35% vs. 7%).
3. There is a minority of clauses that exhibit a different order.
These can be explained by factors such as focus and a few
other special forms, which include negation and the
presence of the 'must' construction, as will be investigated
below. Other orders than the ones stated in (935) are not
attested.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
654 Constituent order

The word order within the Noun Phrase is discussed in §6. As suggested, not all
kinds of modifiers have a fixed position in the NP, but certain tendencies and
preferences can be summarised as follows:
- Demonstratives, numerals, and possessors always precede the nominal head.
Most other modifiers have variable order.
- Adjectival or noun-like modifiers typically follow the head noun.
- Other modifiers can occur in prenominal or postnominal position, especially
quantifiers.
- The typical or preferred order may be described as [DEM] - [POSSR] -
[NUM] - [N] - [ADJ],
- Urarina does not have relative clauses, but these are substituted by various
kinds of nominalisation that may occur with modifying function before or
after the nominal head. Thus, grammatical categories such as tense and
aspect are not expressed by these substitutes.

The constituent order within the dependent clause is largely identical to the
independent clause. Naturally, full NPs are even rarer in dependent clause, as
the referent of subject and object is usually mentioned in the main clause (if it is
mentioned at all). In the case of participle constructions and infinitives, the
subject is only expressed on the main verb. In subordinate clauses, the subjects
of dependent and independent clauses may be different (but they do not need to
be overtly expressed).

18.1 Transitive clause

From a typological point of view, there are six logically possible orders of
subject, verb, and object in a clause: AVO, AOV, VAO, VOA, OVA, and OAV.
Cross-linguistically, the first four types are relatively well attested, whereas the
last two types are hardly accounted for. Urarina represents one of these rare
types, since the unmarked word order is OVA. It has to be noted that, in a
language with extensive omission of overt NPs, the presence of arguments
realised as core NPs is not very frequent. However, OVA is the formally and
functionally unmarked constituent order, as is illustrated by the examples in
(936). A formal marking would involve a frontshift of the subject to the front, in
which case a focus enclitic must be attached to it. The examples are functionally
unmarked as they do not exhibit any pragmatic peculiarities in the context they
occur in. Note that in dependent clause, the presence of full NPs for both A an
Ο is rarer than it is in main clauses, which could be due to the fact that
dependent clauses mainly provide information that supplements information
given in the main clause, i.e. the participants are usually mentioned in the main

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Transitive clause 655

clause. However, (936b) is an example for the order O V A in a dependent clause

(here in a title of a narrative).

(936) Unmarked OVA order

a) Independent clause:
mtoanei hetau=te katga lemu-e=lu lomaj
[like.that HRS=FOC] [man] [sink-3ps/E=REM] [PSN]
[ADV] [Ο] [V] [A]

edara ne-1 ku-uru-a=ne kuj/ia

[water.people be-PRT go-PL-3ps/D=SUB so.that]
[DEP]
'Lomai sank the people like that so that they would become water people.'

b) Dependent clause:
unee bua basihjau-a alau=ne
[kinkajou bag] [steal-3ps/D] [spider.monkey=CND]
[Ο] [V] [A]
'When the spider monkey stole kinkajou's bag'

As was made clear through the table in (935), the omission of arguments is very
c o m m o n . In fact, subject and object are only present in 19% of all examples (S
in intransitive clauses; both A and Ο in transitive clauses), whereas in all other
cases, at least one argument does not surface as an NP. T h e examples in (937)
illustrate transitive clauses that lack an overt subject.

(937) Constituent order OV

a) Independent clause:
hitarü kaute-na ii kuriki u-a-re-ü=ni
[all remain-NOM 2sg money] [come-CAUl-IRR-lsg/E=ASS]
[O] [V]
Ί will bring all of your money that remains.'

b) Dependent clause:
ate kuhwu-uru-a häu ra-ahva-ure
[fish] [fish.with.hook-PL-3ps/D] [because] [receive-DSTR-3pl/E]
[Ο] [V] [CNJ] [Main]
'As they fished with hooks, they each caught [some].'

In the same way, the object of a transitive clause can be omitted. In this case, it
is automatically understood as 3ps (singular or plural) and its referent would be

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
656 Constituent order

known from the context. While it is not uncommon for S arguments to surface
as an NP in an independent clause, examples for the structure VA in
independent clause (such as in (938b)) are rare. Pragmatically, this is a plausible
consequence of the fact that the subject is usually expressed in the main clause
or already known from the context. In (938c), strictly speaking, the verb of the
independent clause (naa 'say (so)') is intransitive in terms of its morphological
features (cf. §9.5); however, it has a special status since it functions as a
quotative verb that is always preceded by direct speech in narratives (cf.
§20.2.4.2).

(938) VA order in transitive clause

a) Independent clause:
niriu hohvajtQa-kuru ajjia ne-rela-t, itga-kuru-a leotga-uru
[PLN inhabitant-PL with ITR-teach-PRT] [do-PL-3ps/A] [other-PL]
[DEP] [V] [A]
'Learning with the people from the Marafion River, the others learnt it [too].'

b) Dependent clause:
su-a lejhii katQa=ne hau, ßäe nalu-a alau
[kill-3ps/D] [one man=SUB] [because] [already fall-3ps/A spider.monkey]
[V] [A] [CNJ] [Main]
'As a certain man killed it, the spider monkey already fell.'

c) Dependent clause, quotative verb:

na-a ofiua häu=na niki u-ü=ra
[say-3ps/D] [father] [because=FOC:lsg] [ADVRS come-1 sg/E=EMF]
[V] [S/A] [CNJ] [Main]
'Because my father said so, I have come.'

In (939), the verb is represented by a copula with a copula complement. This

construction is similar to a transitive clause in that the copula complement
occurs in object position, i.e. in the same position as the Ο argument of a
transitive clause.

(939) Copula complement in Ο position

a) Independent clause:
kuraanaa rn-a ruru
[chief] [be-3ps/A] [howler.monkey]
[CC] [V] [CS]
'The howler monkey is the chief.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Transitive clause 657

b) Dependent clause:
kauatQa-i kuraanaa ni-a=ne=ta
[good-NEG: 3 ps/A] [chief] [be-3ps/D=CND=FRS]
[Main] [[CC] [V]]DEP
'It is not good that he is chief.'

While the figures represented in (935) represent a summary of order types, these
are shown in detail in (940). Note that copula complements are integrated into
these figures as Ο arguments, as they have syntactically identical properties.
Accordingly, copula subjects are represented as S arguments here. The table in
(940) shows the occurrence of order types in the main clause (445 examples),
(942) lists the distribution of order types in dependent clauses (354 cases).

(940) Constituent order in main clause (selected texts)

Text no.: T7 T12 T13 T16 T23A T34 T35A Total Total
Speaker: JN HN JN MA HN HN %
V 11 4 9 33 9 161 8 235 52.8
vs 5 4 9 10 2 14 7 51 11.5
sv 3 0 4 1 2 1 4 15 3.4
ov 4 0 15 6 3 53 4 85 19.1
VA 0 0 1 1 0 5 0 7 1.6
AV 0 0 2 1 1 5 1 10 2.2
OVA 0 0 5 5 1 2 3 16 3.6
AOV 2 0 6 4 0 3 10 25 5.6
OAV 1 0 0 0 0 0 0 1 0.2
Total 26 8 51 61 18 244 37 445 100.0

In Text 7, there is an exceptional occurrence of OAV, - which is based on a

speaker error corrected later to OVA. This order, as displayed in (941), is
otherwise absent from the grammar and cannot be regarded as a possible
variation.

(941) Ungrammatical OAV example

*mi=te ena itulere ne-ι komonida iglesia ena huaauku-e

[that=FOC] [now] [all.kinds be-PRT] [community church] [now] [think-3ps/E]
[O] [ADV] [ADV] [Α] [ADV] [V]
'The community and church remember that in various ways.'

As mentioned, the texts are of different genre and by different speakers. Texts 7
and 13 are told by an older speaker, however, with a sophisticated knowledge of

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
658 Constituent order

Spanish. The speaker is not used to telling stories. Text 7 is a hunting story
(personal experience) and text 13 an account of the village history. Text 16 is a
traditional narrative told by an older speaker, a trained storyteller. Texts 12 and
23A are both presented by a younger speaker with little experience in story
telling. They are narratives that describe dangerous situations experienced by
others (not the speaker himself). Text 35A is by the same speaker and is a
picture description as a given task. It is evident from the statistics that this type
of genre differs significantly from the others: the frequency of clauses with
deviational order types in this text equals the number of examples with regular
order. However, this can easily be accounted for by the fact that the speaker
shifts the subject into focus in many examples as he describes the given images.
Text 34 is a natural conversation between two speakers via a shortwave radio.
Overall, not taking into account clauses without an overt argument ("V"), 76%
(159 out of 210) of order in the main clause correspond to the regular OVA/VS
scheme, whereas 24% deviate from that.
This preference becomes even clearer in dependent clauses. In fact, 91 % of
all dependent clauses that have an overt argument follow this rule, whereas only
9% (12 out of 128) exhibit deviating orders. Again, it is observed that "V" is the
most frequent clause type.

(942) Constituent order in dependent clause

Text no.: T7 T12 T13 T16 T23A T34 T35A Total Total %
Speaker: JN HN JN ι MA HN HN
V 23 7 33 53 24 83 3 226 63.8
vs 6 1 3 5 1 10 0 26 7.3
sv 3 0 5 1 0 0 0 9 2.5
ov 14 4 21 10 1 30 4 84 23.7
VA 0 0 1 0 0 0 0 1 0.3
AV 2 0 0 0 0 1 0 3 0.8
OVA 2 0 0 1 0 2 0 5 1.4
AOV 0 0 0 0 0 0 0 0 0.0
Total 50 12 63 70 26 126 7 354 100.0

18.2 Intransitive clause

As mentioned above, the unmarked constituent order in intransitive clauses is

VS. This is illustrated in (943), where this structure is shown in independent
clause and in (944) for dependent clauses. Similar to corresponding
constructions with transitive verbs, the absence of a subject noun is very
common in dependent clauses. The examples in (943) illustrate various types of
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Intransitive clause 659

intransitive verbs, such as active intransitive verbs (943a), stative verbs (943b),
and the use of an intransitive verb as the inflected verb in a serial v e r b
construction (cf. 943c)). T h e sentence in (943d) contains a copula, which, with
existential function, behaves like an intransitive verb (cf. §9.6.4).

(943) Unmarked order VS: Independent clause

a) With intransitive verb 'get angry':

nii hau hetau=te ubaa-he mi unee
[that because HRS=FOC] [get.angry-3ps/E] [that kinkajou]
[CNJ] [V] [S]
'Therefore, the kinkajou got angry.'

b) With stative verb:

tasißohwa-a kaa lureri
[be.big-3ps/A] [this house]
[V] [S]
'This house is big.'

c) With serial verb:

ßäe n=akatia launeto-a u-e nii hanolari
[already] [3ps=next.to] [sit.down-NTR come-3ps/E] [that jaguar]
[ADV] [LOC] [V] [S]
'That jaguar came to sit down next to him.'

d) With copula:
nii baja ßäe ni-a itulere rukuele
[that after] [already] [be-3ps/A] [all.kinds things]
[CNJ] [ADV] [V] [CS]
'After that, there already were all kinds of things ... [in order to hunt].'

T h e s a m e constituent order applies to intransitive dependent clauses. In (944a),

the v e r b netooka is involved, which literally refers to ' h a n g up (a mosquito net)',
7fi

but is used intransitively to mean 'prepare to s l e e p ' / ' g o to sleep'. T h e other

sentence (944b) exemplifies V S order with a copula in dependent clause, again
with existential function.

78
The verb probably contains the intransitiviser ne-. However, a transitive base form,
which would be *tooka, is not attested.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
660 Constituent order

(944) Unmarked order VS: Dependent clause

a) With intransitive verb netooka:

netoo-ka batiri=jie hau, ... ku kiuara-kure
[hang.up-3ps/D] [priest=SUB] [because] [there see-3pl/E]
[V] [S] [CNJ] [Main]
'As the priest fixed his mosquito net,... they watched him there.'

b) With copula:
m-a akauru ufavana hana aari=jie, ßadera-ure
[be-3ps/D] [3pl field inside] [topa.tree=CND] [be.sad-3pl/E]
[V] [PP] [CS] [Main]
'If there was a topa tree in their field, they were sad ... [because they could not remove
it].'

In summary, the order of core arguments in dependent clauses does not differ
from that in independent clauses. However, there is a preference regarding the
position of non-core arguments in dependent clauses, as will become evident in
§18.4.

18.3 Deviating order types

There are several conditions that can have an impact on the unmarked word
order: focus or emphasis, negative marking, and the involvement of the 'must'
construction. The most important factor is focus (also cf. §19): any constituent
of a clause can be focused; in this case, it is followed by a focus marker and
shifted to the front. Most commonly, non-core constituents such as adverbs or
discourse-related expressions (e.g. mi hau (lit. 'that+'because') to mean
'therefore') are marked for focus. A change in the basic constituent order only
applies when the subject of a transitive or intransitive clause is focused. Typical
contexts for focus involving deviating constituent order are content questions
and the frontshifting of subjects. As illustrated in (945a), the latter results in the
order AOV, since all other constituents remain unchanged. However, the scope
of focus sometimes involves more than a single constituent: In (945b), the
subject plus the adverbial expression nianatii 'even' are focused to mean 'even
the people from the other side of the Maranon River'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Deviating order types 661

(945) AOV order through focus

a)
raj kalaui=te fwei bajhja-1 ama-e
[POSS son=FOC] [firewood] [carry.on.shoulder-PRT take-3ps/E]
[Α] [Ο] [V]
'Her son carries firewood on his shoulder and takes it along.'

b) More than a constituent focused:

nehelau ninitQU kohvajtQa-kuru ma natn=te
[other.side PLN inhabitant-PL even=FOC]
[A]

mtoanei ukivana ha-ure.

[like.that] [field] [make-3pl/E]
[ADV] [Ο] [V]
'Even the people from the other side of the Marafion River cultivate their fields like
that.'

In the same way, the subject of an intransitive clause can be shifted to clause-
initial position by the use of a focus marker. In (946a), the subject N P 'one
man' is moved to the front, followed by the focus marker =te. (946b) shows
focus of a subject pronoun. In dependent clauses, focus markers do not occur.

(946) SV order through focus

a)
lejhii katQa=te hatäi nesoonete
[one man=FOC] [very] [be.scared:3ps/E]
[S] [ADV] [V]
O n e man is very scared.' [from picture description]

b)
kaa=te hjabereonei makaak-e
[this=FOC] [entirely] [be.flat-3ps/E]
[S] [ADV] [V]
'This one is entirely flat.' [about shapes]

Focus is the most frequent cause for changes in the constituent order. As seen
above, this is marked by a focus marker. However, there are a few more
contexts that can coincide with a frontshift of the subject, even without formal
marking. In particular, this applies to clauses that involve a negated verb: in
combination with negation, focus markers are not used and subsequently, a

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
662 Constituent order

frontshift of the subject (if required to indicate focus) occurs without a focus
marker. The examples in (947) slightly differ from each other: in (947a), the
verb is active intransitive, whereas (947b) involves a copula verb, here used
with existential (=intransitive-like) function.

(947) Constituent order in negated clause: subject-initial position (SV)

a)
hat a! jioaelu ne-nakauru niji enanihja kuane
[very earlier be-those.who] [not.at.all] [canoe inside]
[S] [ADV] [PP]

amu-uru-i=lu
[walk-PL-NEG:3ps/A=REM]
[V]
'Those who lived very much earlier did not at all go by canoe.'

b)
jioaelu ena aj-a=ne tokuanei itgafiva ni-ji-lu,
[earlier] [now AUX-3ps/D=SUB compared.to] [rifle] [be-NEG:3ps/A=REM]
[ADV] [DEP] [CS1] [VI]

kartutgo m-ji=lu
[ammunition] [be-NEG:3ps/A=REM]
[CS2] [V2]
'Earlier there were no guns and no ammunition, like there are now.'

Accordingly, in a transitive clause, the constituent order can be AOV as a result

of negation, which is illustrated in (948). Again, the copula construction can be
compared to a transitive clause: in (948b), the copula occurs with a copula
complement and the subject is shifted to the front.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Deviating order types 663

(948) Constituent order in negated clause: subject-initial position (AOV)

a) With full verb:
ena tie-nakauru katQa-uru ßäe rtiji hasisi
[now be-those.who man-PL] [already] [not.at.all] [mixture]
[A] [ADV] [ADV] [Ο]

huituku-uru-i
[know-PL-NEG:3ps/A]
[V]
'Those who live now do not at all know the mixture [for making blowgun poison].'

b) With copula:
kivatia kann i=tgene letono ni-a-u=ta
not [lsg] [2sg=place envoy] [be-NEG-1 sg/E=FRS]
[INTR] [CS] [CC] [V]
Ί am not the envoy for your place!'

The function of this order variation clearly relates to focus and emphasis,
(which may coincide in Urarina, depending on the context). This becomes
especially clear in example (948b), where Adam, who is the only human left in
the world, warmly welcomes one of two women who suddenly appear. In
protest, she makes it very clear that he is embracing the wrong one: the pronoun
for lsg, which is in initial position, is in focus to mean 'it is not me (but the
other one)'. As negative clauses do not occur with focus markers, the move is
made without formal marking.
It should be noted that negation does not normally incur subject-initial order;
this only applies when focus or emphasis of the subject are implied. Examples
for the regular constituent order in combination with negation are given in
(949).

(949) Constituent order in negated clause: subject-final position (VS, OVA)

a) Transitive independent clause:
rüjej n=ere tomra-na here-kur-ene=lu katQa-uru
[not.at.all] [3ps=word] [obey-INF want-PL-NEG:3ps/E=REM] [man-PL]
[ADV] [Ο] [V] [A]
'The people did not at all want to obey his words.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
664 Constituent order

b) Intransitive independent clause:

nete akauru ißasi-lanaala amu-e nii hanulari
[but] [3pl harm-PRV] [walk-3ps/E] [that jaguar]
[CNJ] [DEP] [V] [S]
'But without doing them any harm, those jaguars went (away).'

c) Intransitive dependent clause

ne-ni-a laano-ne hau hetau=te, hvajtet
be-NEG-3ps/D cassava=SUB because HRS=FOC again

laano+bahijio siisiasiitQa-i ki suseri kalaui-tguru

cassava+roast hold.walking-PRT eat:3ps/E unchate.bird son-PL
'As there was no cassava, again, the sons of the unchale bird walked [by] holding
roasted cassava in their hands.'

Further instances of order variation in which the subject is moved to initial

position without involvement of a focus marker are found in a limited number
of contexts. Interestingly, all of these are in some way related to the concept of
focus or emphasis, which corresponds to what is suggested for the variation
with the negative form.
One construction is with the expression for 'also', -netonaj, a form that was
already discussed in §5.3.6. There are several examples where -netormj follows
a subject in initial position. Apparently, it can take the function of a focus
marker, as is illustrated in (950): In all examples, the use of 'also' puts the noun
to which it is attached in focus, as to point at the involvement of that particular
participant in addition to others. While examples (950a,b) involve intransitive
clauses, (950c) exemplifies the use of -netonaj with the transitive verb 'catch'
(here without an overt object).

(950) -netonaj with focus function: SV order

a) Dependent clause:
nete ßäe akauru-netonaj ne-rela-i ßäe enene
but [already] [3pl-also] [ITR-teach-PRT] [already nowadays
[CNJ] [ADV] [S] [V] ([Main])

nitoanei n=uhoana ha-ure

like.that 3ps=field make-3pl/E]
[Main]
'But they learnt it as well and nowadays cultivate their fields like that.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Deviating order types 665

b) Independent clause:
ausirijiem-a raj komasaj=ße hau hetau=te
[be.pregnant-3ps/D POSS wife=SUB because HRS=FOC]
[DEP]

akä-etonaj akauru kaihje ini-a ku-e

[3sg-also] [3pl behind] [go.up-NTR go-3ps/E]
[S] [PP] [V]
'As his wife was pregnant, she also climbed [up the tree] after them.'

c) Independent clause, focus marker added:

raj ßaka-mtonaj=te ajpa muku-e
[POSS father-also=FOC] [with] [catch-3ps/E]
[S] [PP] [V]
'Her father also caught her [together] with them.'

It should be noted that in (950c), a focus marker is attached to -netonaj in

addition to the fact that it already is in focus position. This excludes the
possibility of viewing -netonaj as a separate type of focus marker, which is also
illustrated by the examples in (951). In these sentences, the subject is followed
by -netonaj, but occurs in regular final position. In this case, there is no focus on
the subject. Also note that in (95 lc), -netonaj is attached to the object.

(951) -netonaj attached to subject in final position

a) Attached to S:
katpa rüa sim-a reemae-netonaj
[man side] [sleep-3ps/A] [dog-also]
[LOC] [V] [S]
'The dog also sleeps at the man's side.'

b) Attached to S, with negation:

turu-uru-i akauru-netonaj
[arrive-PL-NEG:3ps/A] [3pl-also]
[S]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
666 Constituent order

c) Attached to O:
leotQa-uru-netonaj ru-akaanu
[other-PL-also] [meet-lpl/ex]
[O] [V]
'We have also met the others.'

The same focus-like structure as illustrated with -netonaj is found with another
expression that follows a noun: manatii 'even' frequently follows a subject that
occurs in clause-initial position. Similar to -netonaj, an inherent emphatic
function is implied.

(952) manatii with focus function

a) Dependent clause:
enamanaa-kuru nianatn kana=kvaaun-era here-mi ke
[young.man-PL even] [lpl/in+create-AG want-PASS VLI]
[S] [PP]

kauatQa-i mitQu-a=ne
[good-PRT] [get.up-3ps/D=CN D]
[ADV] [V]
'If even the young people get up to [follow] what God wants, ... [the project will
succeed].'

b) Independent clause:
mjej nii nuriu kokvajtQa-kuru nianatn
[not.at.all] [that PLN inhabitant-PL even]
[ADV] [S]

nuna kohoajt^a-kuru rüa turu-i=lu

[jungle inhabitant-PL side] [arrive-NEG:3ps/A=REM]
[LOC] [V]
'Not even those people from the Marafion River (at all) had arrived close to the jungle
people.'

However, neither is the initial position an obligatory feature of manatii, nor is

the presence of a focus marker excluded, as the sentences in (953) show. This
perfectly corresponds to the function and status of -netonaj, as seen above. In
(953a), the preverbal position of the subject (followed by 'even') is marked by
the focus enclitic that is attached to the whole NP, including nianatn. In (953b),
the subject followed by 'even' is in regular final position (here occurring with a
variant natii after the clitic for 'also').

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Deviating order types 667

(953) Other positions of manatfi

a) With focus marker:

kunajtena nianatu=te amiane
[sick.person even=FOC] [work:3ps/E]
[S] [V]
'Even the sick [will] work [when we start the building project].'

b) In final position:
kana kuraanaa ne-1 kuraati-a rabiri-jtetonaj natu
[lpl/in chief be-PRT] [name-3ps/A] [Rabbi-also even]
[DEP] [V] [S]
'Even also the Rabbi [himself] called him "Lord".' [NT: Mark 12:37]

Another context which allows a subject to move to front position without the
overt presence of a focus marker is the construction with 'must'. This
construction is formed by an inflected verb followed by netene (or variants such
as mrajjie, nerajtene). While it would literally mean 'it is that', it can be
regarded as lexicalised (also cf. §20.2.2.6). With this construction, a subject
may precede the verb, as illustrated in (954).

(954) Subject fronting with 'must' construction

a)
ißono ku-a lejhii itQauena=ne=te
[ayahuasca drink-3ps/D one elder=CND=FOC]
[DEP]

satii ku-uru-a tierajße

[all] [drink-PL-3ps/D must]
[A] [V]
'If one elder drinks ayahuasca, all [people] have to drink it.'

b)
jioaelu ne-nakauru katQa-uru hitQana ha-uru-a mrajße
[earlier be-those-who man-PL] [blowgun] [make-PL-3ps/D must]
[Α] [Ο] [V]

raaheniane
[self]
[ADV]
'The people who lived earlier had to make blowguns themselves.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 668 Constituent order

Again, this order is not compulsory for the 'must' construction, as exemplified
in (955). Here, the subject is found in regular, final position despite the presence
of'must'.

(955) Unmarked constituent order with 'must' construction

a)
ßäe nehelau ne-~t amiani-a nerajtene iglesia mi komunida
[already] [other.side be-PRT] [work-3ps/D must] [church that community]
[ADV] [ADV] [V] [S]
'The community of this church must already work differently.'

b)
eskwela ke ku-uru-a nerajtene kanaanaj-uru
[school VLI] [go-PL-3ps/D must] [child-PL]
[Ε] [V] [S]
'The children must go to school.'

Parallel to the examples given with the expression for 'also' and 'even', the
fronted position of the subject in 'must'-constructions implies focus or
emphasis, whereas the regular position does not imply this. Thus, nearly all
instances of constituent order that deviate from the basic OVA/VS pattern can
be explained by focus or emphasis, which is either formally marked or
associated with particular constructions. However, there are a few isolated
examples that exhibit subject-initial order without any formal marking and
without involving any of the expressions or constructions discussed above. Two
of these are given in (956). Example (956a) was later described as "ugly but
understandable" by other speakers; the example in (956b) resembles a
construction that involves a copula complement, as is illustrated in (957).

(956) Deviating constituent order without formal marking

a)
mi hau nil eene ku helaj neda-e
[that because] [that woman] [there] [separate] [stay-3ps/E]
[CNJ] [S] [ADV] [ADV] [V]
'Therefore, that woman stayed there by herself.'

b)
ßäe itulere akauru rukuele, itulere ruria, itulere ubißa,
already [all.kinds 3pl things all.kinds masato.pot all.kinds stirring.stick
[S]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Deviating order types 669

itulere-uru jiäe amu-ahe

all.kinds-PL] already [walk-CNT:3ps/E]
([S]) [V]
'Already several of their things, various pots, stirring sticks, all kinds of things kept
going [i.e. they turned into water animals].'

In order to account for the irregular order in (956b), one could assume that the
copula form ne-7, as found with the surrounding sentences in this context was
omitted for some reason: there is a special copula construction having the
meaning 'turn into'/'become', with intransitive function. As illustrated in (957),
the copula complement, (which occurs in the same position as the Ο argument
of a transitive verb), is followed by a participle form of the copula and a motion
verb for ' g o ' or 'walk' that bear the inflection. Literally, this combination could
be translated as 'it went being a (e.g. turtle)'. All examples occur in the same
context, where the speaker lists all the household items that turned into animals
during the flood. Interestingly, in four of the clauses (all listed in enumeration),
the subject appears in initial position, followed by the object, but both preceding
the verb. This is exceptional as similar structures are not found otherwise.
Another exceptional aspect of this construction is that it occurs without any
focus marker, which would normally appear when a subject is shifted to the
front.

(957) Irregular constituent order with idiomatic copula construction

a) net + amua:
akauru ruria jiäe nuri ne-1 amu-e
[3pl masato.pot] [already] [turtle] [be-PRT walk-3ps/E]
[CS] [ADV] [CC] [V]
'Their masato pots already turned into turtles.'

b) net + amua with CNT form:

akauru raj d$ufivana kurarl ne-1 amu-ahe
[3pl POSS flute] [macane.fish] [be-PRT walk-CNT:3ps/E]
[CS] [CC] [V]
'Their flutes were already turning into macane fish.'

c) rm + lata:
kauru raj eruari ahaijx ne-1 ku-e
[3pl POSS mosquito.net] [paiche.fish] [be-PRT go-3ps/E]
[CS] [CC] [V]
'Their mosquito nets turned into paiche fish.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
670 Constituent order

Given the special meaning and its rare occurrence in other texts, the 'turn into'
construction with the copula can be regarded as idiomatic. As stated before,
variation of the basic constituent order such as illustrated in (956), which occurs
without formal or contextual marking, is extremely rare. In fact, only very few
examples of this type (including (956b)) are attested in the 799 clauses listed in
(940). Another rare example is displayed in (975a) further below, which
represents a syntactically rather complex sentence. It should also be noted that
there are no instances of order variation for the object of a transitive verb: It
always precedes the verb.
Thus, it can be concluded that Urarina has a basic constituent order
OVA/VS, which can be subject to variation in nearly entirely predictable
environments, related to focus and emphasis.

18.4 Adjuncts

18.4.1 Position ofpostpositional phrases (PPs)

In §9.5 it was stated that there is no formal distinction between adjuncts and
arguments in addition to A/S and O. As was argued there, "additional
arguments" accompanied by ÄE, e.g. as a result of valency increase, may exhibit
minor semantic differences to adjuncts (also cf. §16.5). A similar situation is
implied for the postposition raj, which may indicate a recipient or beneficiary of
verbs of saying or giving. What all postpositional phrases (PPs) have in
common is that they are optional and entirely omittable, (which is the main
argument for suggesting the absence of ditransitive verbs in Urarina, cf. §9.5).
Another feature that is shared by all PPs is their position in a clause, which is
most typically in the periphery of the core clause. The most typical position for
any PP is before the main verb. In transitive clause with an overt Ο argument,
the PP occurs before O. There are almost no attested natural examples for their
occurrence between Ο and the verb. One is cited in (958), where the
postposition raj 'for' is cliticised with a possessive proclitic; the demonstrative
which functions as the object, in turn, is realised as a proclitic attached to itg,ej
'for you'. Due to the clitic status of the elements involved, it is questionable
whether this can be characterised as a PP-insertion between Ο and V. Since this
example is marginal at any rate, one could conclude that generally, PPs do not
occur in this position.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Adjuncts 671

(958) Exceptional occurrence of OPV

na-a häu=ne ka-i=tQej be- ΰ

[say-3ps/D because=FOC:lsg] [this=2sg=for] [tell-1 sg/E]
[DEP] [0=PP] [V]
Ί am telling you this because he said so.'

Alternatively, but less commonly than before the Ο argument, a PP may occur
in postverbal position. If it does, the preferred position is after the subject.
However, there are a small number of examples in which a PP is found between
the verb and the subject (one example only in the selected texts described in
(935)). Semantic or pragmatic differences for this variation are unclear. A PP
may also be focused; in this case, it occurs with a focus marker and in clause-
initial position.
In (959), some examples for the preverbal occurrence of a PP involving ke
are shown. In (959a), the postposition has a strictly valency-increasing function,
being used in a causative construction: While the Ο argument (i.e. the recipient
of the blood) does not surface, the additional argument ("with blood") is marked
by ke. In (959b,c), the postposition has instrumental function. As seen in (959c),
the PP occurs in a dependent clause first, followed by the main clause that
contains an almost identical PP.

(959) PP with ke in preverbal position

a) Independent clause:
na-~i kwitQana ke kukwa-e nii raj ßakana
[say-PRT] [blood VLI] [give.to.drink-3ps/E] [that POSS son-in-law]
[DEP] [PP] [V] [A]
'Saying so, his son-in-law gave him [his own] blood to drink.'

b) Dependent clause:
heräe kihja ke kuati-1 ra-a
[slowly] [paddle INST] [carefully.go-PRT] [receive-3ps/A]
[ADV] [PP] [V] [Main]
'Slowly going with the paddle, he got it [the axe].'

c) Dependent and independent clause:

aheri+ku-teru ke n=enanihja te-1, aheri+ku-teru-atQa ke
[stone+ASC-axe INST] [3ps=canoe] [make-PRT] [stone+ASC-axe-only INST]
[PP1] [Ol] [VI] [PP2]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
672 Constituent order

itulere itQa-kure
[all.kinds] [make-3pl/E]
[02] [V2]
'Making canoes with stone axes, they made all kinds of things with stone axes.'

The postverbal position of a PP involving ke is illustrated in (960). In (960a), ke

has a valency-increasing function (adding an argument to the intransitivised
verb 'learn'). Example (960b) contains a PP in which ke has instrumental
function. The examples also show that the only obligatory component of a
clause is the verb, whereas subject or Ο argument need not surface as an NP.
Also note that PPs do not normally follow the verb in a dependent clause (no
examples in statistical database).

(960) PP with ke in postverbal position

a)
ita kuruatahane-1 ena ne-relaa-kure amiane-naa ke
[REC help-PRT] [now] [ITR-teach-3pl/E] [work-NOM VLI]
[DEP] [ADV] [V] [PP]
'Helping each other, they now learn (to) work.'

b)
lejhii=te teru ke fwi
[one=FOC] [axe INST] [cut:3ps/E]
[A] [PP] [V]
'One [man] cuts it with an axe.'

In the same way, the variable position of PPs is exemplified with the
postposition raj 'for/to'. Correspondingly, PPs involving raj are not obligatory.
The sentences in (961) illustrate the preverbal as well as the postverbal
occurrence of such PPs. As mentioned before, the postverbal position is less
common than the preverbal one. In (961c), two PPs involving raj occur, one in
preverbal position of the first clause, the other in postverbal position in the
second clause.

(961) Position of PPs with raj

a) Preverbal position:
ßäe koairi ra-~t akauru raj tutu-e
[already ayahuasca receive-PRT] [3pl for] [mash-3ps/E]
[DEP] [PP] [V]
'[Then] she already took ayahuasca and mashed it for them.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Adjuncts 673

b) Preverbal position:
eresi ka=raj lejhii relor kurete-u
[tomorrow] [lsg=for] [one watch] [buy-IMP]
[ADV] [PP] [Ο] [V]
'Buy a watch for me tomorrow!'

c) Preverbal, postverbal position:

eresi ka=raj ate kuhwu-a ku-uru-iße katga-uru,
[tomorrow] [lsg=for] [fish] [fish.with.hook-NTR go-JUSS] [man-PL]
[ADV 1 ] [PP1] [Ol] [VI] [Al]

na-a hetau mi raj mosaso raj

[say-3ps/A HRS] [that POSS novice for]
[V2] [PP2]
"'Tomorrow, the people ought to go fishing for me", he told his novice.'

In order to illustrate that the variable position with respect to the verb also
applies to other types of PPs, some examples with different postpositions are
given in (962) and (963). While examples are not listed for all existing
postpositions in each position, they can all occur in preverbal or postverbal
position. (962) first shows a few sentences that contain a PP in preverbal
position; (963) accounts for PPs in postverbal position. As becomes evident,
this involves PPs of different kinds, including locational, temporal, beneficiary,
and comitative postpositions (also cf. §5.7 for an overview of different
postpositions).

(962) Preverbal position of PPs involving other postpositions

a) Locational (with kuarte 'inside') (Dependent clause):

haurianehet kaa ajtQune kuane eno-akaanu-ne hana
[at. first] [this PLN inside] [enter-lpl/ex=SUB] [when]
[ADV] [PP] [V] [CNJ]

ni-ji=lu rte-naanaha
[be-NEG:3ps/A=REM be-LPP]
[Main]
'When we first entered the Espejo River, there was no place to live.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
674 Constituent order

b) Locational (with kahe 'from') (Dependent clause):

nii kahe ku-ahe-1 kau turu-akaanu, atane ari~t
[that from] [go-CNT-PRT] [here arrive-1 pl/ex] [land seek-PRT]
[PP] [V] [Main] [DEP2]
'We kept going from there and arrived here, looking for land.'

c) Comitative (with ajßa 'with') (Independent clause):

ena netohwel hajti kargo ajßa ne goberenadoro
[now until] [still] [assignment with] [be:3ps/E] [governor]
[PP] [ADV] [PP] [V] [S]

kaa komonida dirihi-a=ne kujpa

[this community lead-3ps/D=SUB so.that]
[DEP]
'Until now the governor lives with this assignment in order to lead this community.'

d) Beneficiary (with kurema 'for') (Independent clause):

he, ka=nesari kurema m-ririhi-a nukue-re-i=ta
hey [ 1 sg=booty for] [that-be.like-NTR do.in.vain-IRR-2ps=FRS]
[PP] [V]
'Hey, you are going to do like that in vain for my booty.' (Hunter talking to a jaguar
that has stolen a tapir from his trap)

e) Temporal (with netohwel 'until') (Independent clause):

ajane rihihei jioaelu netohwel kajritigi ne-ι ne-ri-ki-tQe=i
[like.that] [earlier until] [poor.person] [be-PRT be-IRR-2ps-PL=ASS]
[ADV] [PP] [CC] [V]
'You will be poor people like that forever.'

While the preverbal variant is the preferred type (in terms of frequency), there
are no apparent differences with respect to semantic or pragmatic structure.
The examples in (963) illustrate the postverbal occurrence of different PPs,
again appearing with transitive clauses (cf. (963a,b,c,f)) and intransitive clauses
(cf. (963d,e)). Examples (963d,f) contain two PPs; in this case, the locational PP
precedes the verb, whereas the comitative PP follows (and no counterexamples
are attested, but note that the co-occurrence of several PPs in a clause is quite
rare). It is noteworthy that PPs do not occur in postverbal position in dependent
clause.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Adjuncts

(963) Postverbal position of PPs involving other postpositions

a) Locational (with kuane 'inside') (Independent clause):

nit hana hetau alau rauto-hioe-i beree
[that when HRS] [spider.monkey] [be.calm-CNT-PRT] [child]
[CNJ] [A] [ADV] [Ο]

amae+mahe hetau unee bua kuane

[RED-carry.along.CNT:3ps/E HRS] [kinkajou bag inside]
[V] [PP]
'So the spider monkey easily carries his creature along in the kinkajou's bag.'

b) Locational (with kahe 'from') (Independent clause):

lejhii buruari-itQa=te kurete-kure=lu nuse kahe
[one] [poison-only=FOC] [buy-3pl/E=REM] [town from]
[ADV] [Ο] [V] [PP]
'They only bought the poison alone from town.'

c) Locational (with asae 'under') (Independent clause):

nit baja=te ruaka-ι ama-ure raj loanari asae
[that after=FOC] [carry-PRT] [take-3pl/E] [POSS shed under]
[CNJ] [DEP] [V] [PP]
'After that, they took him to his shed.'

d) Locative and comitative (with asae 'under', ajpa 'with') (Independent clause):
helaj ne-naa loanari asae sini raj beree-kuru ajpa
[separate] [be-NOM shed under] [sleep:3ps/E] [POSS child-PL with]
[ADV] [PP1] [V] [PP2]
'She slept alone under a shed with her children.'

e) Beneficiary (with kurenia 'for') (Independent clause):

ena rauto-hiue-i m-ji ahe-naa, m-ji d$a
[now be.calm-PRT] [be-NEG:3ps/A] [get.drunk-NOM][be-NEG:3ps/A] [what
[DEP] [VI] [SI] [V2] [(S2)

baaso kana+kwaaun-era kurenia

bad] [lpl/in+create-AG for]
S2] [PP2]
'Now [everything] being calm, there is no drinking, there is nothing bad for God.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
676 Constituent order

f) Instrumental and temporal (with ke, netohzuel 'until') (Independent clause):

kwata kahe itga-noi katai kurete-1 ke
[metal from make-PASS hook buy-PRT] [INST]
[DEP] [PP1]

kuhwu-ure ena netohwel

[fish.with.hook-3pl/E] [now until]
[V] [PP2]
'Until now they are buying hooks made from metal and fish with them.'

When a PP occurs in postverbal position, it usually follows the subject, if one is

present (cf. (963e)). Alternatively, but less common, is the occurrence of a PP
between verb and subject, as shown in (964). Here, PPs involving different
types of postpositions are inserted in postverbal position, but preceding a
subject NP. There are no restrictions as to which postpositions can occur in this
position. An example for a PP with the locational postposition ham 'inside' in
dependent clause was also given in (944b).

(964) PP between verb and subject

a) (With raj 'for') (Independent clause):

hi ajto-a kauru raj nii lomaj
[there] [say-3ps/A] [3pl for] [that PSN]
[ADV] [V] [PP] [A]
'[So] Lomai told them there:...'

b) (With ajpa 'with') (Independent clause):

nii häu hetau=te ku ne akauru ajpa batiri
[that because HRS=FOC] [there] [be:3ps/E] [3pl with] [priest]
[CNJ] [ADV] [V] [PP] [S]
'So the priest stayed there with them.'

c) (With hana 'inside') (Independent clause):

ku jiäe mihiito-oru-a hau, nejtQara wane
[there already be.hungry-PL-3ps/D because] [perhaps] [PTY]
[DEP] [ADV] [PRT]

m-a kaa hana lenone=ta

[be-3ps/A] [this inside] [food=FRS]
[V] [PP] [S]
'As they were hungry, (they thought), "Maybe there is food [somewhere] in here for us
poor guys.'"

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Adjuncts 677

It can be concluded that the position of PPs is fairly free, with some exceptions
and preferences. There is a strong preference for their occurrence before the
verb, but deviations from this are not uncommon. In dependent clauses, the
preverbal position o f PPs is the only possible one. PPs as occurring between Ο
and V are not normally admitted.

18.4.2 Position of adverbs

The typical position for adverbs is in clause-initial position, examples for which
are shown in (965). This regards all types of adverbs, including time, location,
manner, and epistemics (cf. §5.3). A series of several adverbs in a row is also
-JQ

quite common, as the examples (965c-e) show. Examples (965c,e), each

contain several adverbs in this position, occurring in dependent clause.

(965) Adverbs in clause-initial position

a) Temporal (Independent clause):

erta nerereta-re-ü=rü ißaahi-anu nehaa ke uahei
[now] [tell-IRR-1 sg/E=ASS] [know-lsg/D PURP VLI until]
[ADV] [V] [DEP]
'Now I will tell [stories] as far as I know.'

b) Manner (Independent clause):

akauru rela-u=ra, kauatQa-1 akauru rela-u=ra
[3pl] [teach-IMP=EMF] [good-PRT] [3pl] [teach-IMP=EMF]
[Ol] [VI] [ADV2] [02] [V2]
Teach them, teach them well!'

c) Location, time (Dependent clauses):

ku kwajtel ne-uru-a hau hetau,hajti ku sini-uru-a hau
[there][again] [be-PL-3ps/D] because HRS [still] [there][sleep-PL-3ps/D] because
[ADV1] [ADV1] [VI] [ADV2] [ADV2] [V2]
'As they were there again, as they were sleeping,... [one said ...].'

d) Time (Independent clause):

nii hau hetau=te kzuajtei ßäe amu-a ku-e
[that because HRS=FOC] [again] [already] [walk-NTR go-3ps/E]
[CNJ] [ADV] [ADV] [V]
'Therefore, he already went again.'

79
Whether there is any preferred order in sequences of adverbs is unknown.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 678 Constituent order

e) Time, manner (Dependent clauses):

itulere kurete-1 ßäe rauto-hioe-i kaitehe-lanaala
[all.kinds] [buy-PRT] [already] [be.calm-CNT-PRT] [suffer-PRV]
[Ol] [VI] [ADV2] [ADV2] [V2]

it$a-fairu-a=na hau
[do-PL-3 ps/D=SUB ] [because]
[V3] [CNJ]
'... buying all kinds of things, because they already do it easily without suffering'

Adverbs in initial position do not require a focus marker, but they can be
marked with it. In this case, the adverb is in focus, as illustrated in (966).
Interestingly, not all adverbs are attested with the focus marker, but this
combination mainly occurs with some temporal adverbs. Furthermore, the word
hoajtei 'again' and a few others are never observed with the focus marker.

(966) Adverbs in focus position

a) Time (Independent clause):

ena-te amiane kana+huaaun-era raj
[now=FOC] [work:3ps/E] [lpl/in+create-AG for]
[ADV] [V] [PP]
'Now [as opposed to before] they work for God.'

b) Time (Independent clause):

jioaelu hetau=te emjtQu-hiru kuraanaa ne-ι ne=lu
[earlier HRS=FOC] [monkey-PL chief] [be-PRT be:3ps/E=REM]
[ADV] [CC] [V]

alau
[spider.monkey]
[CS]
'Earlier, spider monkey was chief of the monkeys.'

c) Manner (Independent clause):

rauto-hwe-i=te rtiki raj kakunu ne-naha turu-e
[be.calm-CNT-PRT] [ADVRS] [POSS daughter be-LPP] [arrive-3ps/E]
[ADV] [PRT] [LOC] [S]
'Easily she arrived at her daughter's place.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Adjuncts 679

d) Intensity (Dependent clause):

mtoanei hetau=te nekajritga-ϊ hatal kau-e nii lejhtt katga
[Iike.that HRS=FOC] [suffer-PRT] [very] [return] [that one man]
[ADV] [[VI] [ADV1]] [V2] [S]
'Like that, suffering very much, that man returned home.'

There are only a few examples where adverbs occur in postverbal position.
This, however, is comparatively uncommon. In (967), the temporal adverbs
hauria 'at first' and ßäe 'already' are found in this position. It must be noted
that the word for 'already' is extremely frequent in general; sometimes there can
be various occurrences in one sentence and it is often used as a gap filler in
pauses (cf. §22.5). In (967b), the word occurs twice and it is possible that it
refers to the enumerative verb naaohwaa. Alternatively, the second occurrence
of ßäe could be interpreted as an apposition, added to the entire sentence.

(967) Adverbs in postverbal position

a) Time (Independent clause):

raahemane hitQana ha-ϊ, ne=batohioi te-i,
[self blowgun make-PRT] [3ps=bow make-PRT]
[DEP1] [DEP2]

maohzua-t=te ke nesarit Qäete-ku re=l u hauria

[ENUM-PRT=FOC] [INST] [hunt-3pl/E=REM] [at.first]
[DEP3] [PP] [V] [ADV]
'At first, they made blowguns themselves, they made their bows, and hunted with
them.'

b) Time (Independent clause):

[Preceding: 'Their weaving sticks turned into yellow-spotted river turtles']
akauru ruria ßäe nuri ne-1 amu-e naaohiua-l
[3pl masato.pot] [already] [turtle] [be-PRTwalk-3ps/E] [ENUM-PRT]
[CS] [ADV] [CC] [VI] [V2]

ßäe
[already]
[ADV]
'... and their masato pots already turned into giant water turtles, already.'

The table in (968) gives an overview of the position of 325 adverbs in the
selected texts. The figures confirm the fact that preverbal position clearly is the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
680 Constituent order

preferred one, with some adverbs never occurring after the verb. Corresponding
to the behaviour of PPs, adverbs do not occur after the verb in dependent clause.

(968) Position of adverbs

Adverb Gloss Total Preverbal Postverbal

jiäe 'already' 140 134 6
eresi 'tomorrow' 20 14 6
ena 'now' 15 13 2
hajti 'still' 12 12 0
nakwaaunel 'again' 12 12 0
hauria 'at first' 5 3 2
taulu, taum 'before' 5 5 0
enene nowadays' 4 4 0
hanotn; hanone 'at night'; 'in the
4 4 0
morning
raka 'yesterday' 3 2 1
poaelu, poaem 'earlier' 3 3 0
haurianehel 'at first' 2 2 0
kwajtei 'again' 1 1 0
rautohwei 'calmly' 15 15 0
hatal 'very' 13 12 1
kauatQäι 'nicely' 8 8 0
heritQavehei 'as before' 6 6 0
banehelanaala 'totally' 5 5 0
heriarie 'probably' 4 4 0
raahemane 'self 4 3 1
tabai 'much' 3 1 2
esißäe 'really' 3 3 0
heräe 'slowly' 2 2 0
kauatQahei 'well' 2 2 0
himikn 'together' 1 1 0
ku 'there' 33 33 0
Total 325 304 21

From the table in (968), it also seems that temporal adverbs have a higher
flexibility with regard to their position than others. In contrast, the locational
adverb lot 'there' consistently occurs in preverbal position.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Adjuncts 681

18.4.3 Position of locational adjuncts

Another type of NP that occurs in the same position as an Ο argument is a

locative. Similar to all other constituents, it is optional. However, when a
locative NP occurs with a locational verb such as 'go', 'come', or 'arrive', it is
found before the verb, i.e. in the same position as an Ο argument. All locational
verbs are intransitive and do not require an argument (and even transitive verbs
do not require an overt object NP, as was shown in §18.1). While locational
NPs follow the trend of occurring in preverbal position, there also are some
occurring in postverbal position.

(969) Locational NP in preverbal position

a) With locational NP:
rautohwei=te niki raj kakunu ne-naha turu-e
[calmly=FOC] [ADVRS] [POSS daughter be-LPP] [arrive-3ps/E]
[ADV] [PART] [LOC] [V]
'She calmly arrived at her daughter's place.'

b) With locational NP and locative marker:

satonohel=te ikito-o ama-ure
[in.the.end=FOC] [PLN-LOC] [take.along-3pl/E]
[ADV] [LOC] [V]
'In the end, they took him to Iquitos.'

18.4.4 Summary

Overall, there is a clear preference for the position of adjuncts of any type to
occur before the verb. This is summarised in (970), which illustrates that
speaker differences and the genre type hardly play a role for this preference. As
becomes evident from (970b), adjuncts occur before the verb at an average rate
of 92%.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
682 Constituent order

(970) Position of adjuncts

a) Tokens:
All clauses T7 T12 T13 T16 T23A T34 T35A Total
JN HN JN MA HN HN
ADV before V 30 5 37 53 12 159 8 304
ADV after V 1 0 5 1 1 13 0 21
PP before V 11 3 22 15 9 68 4 132
PP after V 1 0 3 0 0 8 1 13
LOC before V 5 0 6 7 0 2 0 20
LOC after V 0 1 2 0 1 2 0 6
Total 48 9 75 76 23 252 13 496

b) Summary:
Adjunct Before V After V Total (100%)
ADV 304 (94%) 21 (6%) 325
PP 132 (91%) 13 (9%) 145
LOC 20 (77%) 6 (23%) 26
Average 456 (92%) 40 (8%) 496

18.5 Position of dependent clauses

The tendency of placing syntactic material other than objects into preverbal
position is further illustrated by dependent clauses. While it is possible for a
dependent clause to follow the main clause, the more typical position is before
the main clause. This applies to all types of dependent clauses, except for
purposive/consecutive clauses with kujjia 'so that' or huataa 'so that not'.
These preferably follow the main clause. In fact, out of 34 occurrences of these
two conjunctions in the selected texts, only 9 occurred before the main clause,
whereas the majority of 25 (= 74%) followed it. This distribution is in strong
contrast to the position of dependent clauses in general, as can be seen in (971).
On average, 69% of all dependent clauses precede the main clause.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Position of dependent clauses 683

(971) Position of dependent clauses

a) Tokens:
T7 T12 T13 T16 T23A T34 T35A Total
JN HN JN MA HN HN
All clauses: 76 20 114 131 44 370 44 799
Total main clauses 26 8 51 61 18 244 37 445
Total DEP clauses 50 12 63 70 26 126 7 354
Preverbal DEP 32 9 29 59 21 89 5
clauses 244
Postverbal DEP 18 3 34 11 5 37 2
clauses 110

)) Summary in percent:
Total In %
Total main clauses 445 56%
Total DEP clauses 354 44%
Preverbal DEP clauses 244 69%
Postverbal DEP clauses 110 31%
All clauses: 799 100%

Examples for different types of dependent clauses that precede the main clause
are given in (972). Note that many sentences have more than one dependent
clause. When this is the case, again, some may follow the main clause, whereas
the majority precedes it. However, this "split" in position is not obligatory, as
there also are sentences where all dependent clauses precede or follow the main
clause.

(972) Dependent clause preceding the main clause

a) Two dependent clauses with hana 'when':

na-a näkana raj=je hana, hoara-a=ne hana
[say-3ps/D father-in-law for=SUB when] [see-3ps/D=SUB when]
[DEP1] [DEP2]

kahianei fwaururu-hwe, kahianei sal kahe

[everywhere] [be.tom-CNT:3ps/E] [everywhere all from]
[ADV1] [VI] [PP]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
684 Constituent order

m-mitQu-hua-e it$ana
[ITR-get.out-CNT-3ps/E] [blood]
[V2] [S2]
'When he said so to his father-in-law, when he saw him, he [the father-in-law] was
shredded over here, from all over here he was bleeding.'

b) Dependent clause with hau 'because':

amu-a hau ßäe huajtei tahia
[walk-3ps/D because] [already] [again] [over.there]
[DEP] [ADV] [ADV] [LOC]

turu-a u-e nii kati

[arrive-NTR come-3ps/E] [that black.monkey]
[V] [S]
'As he [the husband] walked away, again the black monkey already arrived from over
there.'

c) Various participle clauses and dependent clause with baja 'after':

rüi ham hetau=te ßäe ku rihjeoritü, ßäe ku
[that when HRS=FOC] [already] [there] [immediately] [already there
[CNJ] [ADV] [ADV] [ADV] ([DEP1])

turu-ι, ku lenone itQa-ι, ke lenone-1,

arrive-PRT] [there food make-PRT] [VLI eat-PRT]
[DEP 1 ] [DEP2] [DEP3]

ku hetau lenone-kuru-a=ne baja, ßäe hetau n=ajtoo-hwa ne-1,

[there HRS eat-PL-3ps/D after] [already HRS 3ps=say-NOM be-PRT]
[DEP4] [DEP5]

nii kuri ke inia-kure

[that jagua.tree VLI] [go.up-3pl/E]
[PP] [V]
'Then, immediately, they arrived, made food and ate it, and after they had eaten, as he
[the envoy] had said, they climbed that jagua tree.'

A dependent clause in sentence-initial position can also be focused, as is shown

in (973). The insertion of a focus marker puts special emphasis on the
dependent clause in this position, as it would do with adjuncts and adverbs (cf.
§19.3).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Position of dependent clauses 685

(973) Dependent clause with focus marker

a)
komasaj ajßa ham-a ku-a hau hetau=te
[wife with hunt.overnight-NTR go-3ps/D because HRS=FOC]
[DEP]

ku hetau ruh enejtQu saate

[there HRS] [ADVRS] [monkey] [kill.by.blowgun:3ps/E]
[ADV] [PART] [Ο] [V]
'As he was going hunting overnight with his wife, he killed a monkey with his
blowgun.'

b)
ku hetau huara-a häu=te mi itanitQa hana hetau
[there HRS see-3ps/D because=FOC] [that mud inside HRS]
[DEP] [PP]

ht bam+bane-he
there [RED+moum-CNT:3ps/E]
[ADV] [V]
'As he [Adam] was looking there, he [the envoy] was mourning there in the mud.'

As mentioned above, the occurrence of dependent clauses after the main clause is less
typical, but not exactly uncommon. In (974), I give examples for different dependent
clauses as occurring after the main clause

(974) Dependent clause following the main clause

a) Participle clause:
mtoanei hetau=te n=uhoana ha-ure-rehete jioaelu katga-uru,
[like.that HRS=FOC] [3ps=field][make-PL-HABl :3ps/E] [earlier man-PL]
[ADV] [Ο] [V] [A]

aheri+ku-teru ke kufwi-hjai
[stone+ASC-axe INST fell-PRT]
[DEP]
'The ancient made their fields like that, felling [trees] with stone axes.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
686 Constituent order

b) Dependent clause with hau 'because' following the main clause:

nii hau hetau=te mtoanei ne=lu kana raj
[that because HRS=FOC] [like.that] [be:3ps/E=REM] [lpl/in for]
[CNJ] [ADV] [V] [PP]

floaelu kana raj hawma ra-uru-a hau

[earlier lpl/in for night receive-Ρ L-3ps/D because]
[DEP]
'Therefore, it was like that for all of us, because in ancient days they received the night
for all of us.'

c) Participle and consecutive/purposive dependent clause following the main clause:

nii itQauena nahari beuri-tQU-tge, seboja ke hasiite-1,
[that elder liver] [cut.a.piece-IMP-PL] [onion INST mix-PRT]
[Ο] [V] [DEP1]

ki-a lura+kuturi=fle kujßa

[eat-3ps/D bald+head=SUB so.that]
[DEP2]
'Cut a piece from the liver of that old [peccary] and mix it with onions, so that the bald
one would eat it.'

A sentence may also involve several dependent clauses, some of which precede
the main clause, whereas others follow it. Again, there is no fixed pattern that
would determine the order in which these occur, as this is entirely based on the
context (and the order of the events described).

(975) Dependent clauses preceding and following the main clause

h* hetau netoo-hwakwa-uru-a=w hana, lejhü hetau

[there HRS go.to.sleep-DSTR-PL-3ps/D=SUB when] [one HRS]
[DEP1] [S]

ku eruari temura-u lot lauhua-t neda-e,

[there mosquito.net edge-LOC there be.sitting-PRT] [stay-3ps/E]
[DEP2] [V]

nii lomaj beene herenaka-i,

[that PSN female be.in.love-PRT]
[DEP3]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Position of dependent clauses 687

lomaj beene dada-a=rte kujßa niki

[PSN female touch-3ps/D=SUB so.that ADVRS]
[DEP4]
'When they each went to sleep there, one [man] stayed sitting [outside] at the edge of
the mosquito net, being in love with Lomai, so that he would touch her.'

As mentioned before, there is a certain degree of variation between different

speakers. The narrator of Text 7 has a clear preference for placing dependent
clauses before the main clause (cf. (971a)): in 59 out of 70 dependent clauses,
they occur in this position, while only 11 are found after the main clause. This
does not match the findings observed for the narrator of Text 13: the
distribution of dependent clauses is almost equal with 29 occurring before the
main clause and 34 after it. However, it appears that this speaker's preferences
are somewhat exceptional, as they do not match the general tendency.
There also are different preferences with regard to the number of dependent
clauses in a sentence. The table in (976) gives a statistical overview of clause
types, their number and position within the sentence from selected texts.

(976) Sentence length

a) Tokens:
Length T1 T7 T16 T12 T23A T35A T34 Total
JN JN MA HN HN HN
Main clause only 22 8 28 1 6 30 149 230
Sentences with 1 13 5 17 4 6 7 72
DEP 124
Sentences with 2 6 6 6 2 1 0 16
DEP 37
Sentences with 3 5 3 5 0 3 0 6
DEP 22
Sentences with 4 3 2 2 4 1 0 1
DEP 13
Sentences with 5 1 1 1 0 1 0 0
DEP 4
Sentences with 6 1 1 2 0 0 0 0 4
or more DEP

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
688 Constituent order

)) Summary in percent:
Length Total In %
Main clause only 230 53.0
Sentences with 1 DEP 124 28.6
Sentences with 2 DEP 37 8.5
Sentences with 3 DEP 22 5.1
Sentences with 4 DEP 13 3.0
Sentences with 5 DEP 4 0.9
Sentences with 6 or more DEP 4 0.9
Total: 434 100.0

As the figures illustrate, sentences without a dependent clause are generally

much more frequent than those that are more complex, in declining order.
While longer sentences are still common, they become rarer the more dependent
clauses are involved. Interestingly, speaker HN does not have a clear preference
for monoclausal sentence (i.e. main clause only) in the two narratives. In this
genre, the use of sentences with at least one dependent clause is preferred. A
similar tendency is also observed with speaker MA: The 33 sentences in which
he uses at least one dependent clause prevail over the 28 monoclausal sentences.
Accordingly, speaker JN uses 47 syntactically complex sentences as opposed to
only 30 monoclausal types. However, in conversation and in picture
descriptions, speakers tend to prefer shorter sentences, as becomes evident from
the figures of Texts 34 and 35A. The table in (977) specifies the position of
dependent clauses in the selected texts.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Position of dependent clauses 689

(977) Preferred position of dependent clauses

a) Tokens:
Sentence T13 T7 T16 T12 T23A T35A T34 Total
structure JN JN MA HN HN HN
Μ 22 8 28 1 6 30 149 244
DM 8 3 12 4 5 5 48 85
MD 5 2 5 0 1 2 24 39
DDM 3 3 5 0 1 0 10 22
DDDM 2 1 5 0 2 0 2 12
DMD 0 2 0 1 0 0 5 8
DDMD 2 1 0 0 1 0 3 7
MDD 3 1 1 1 0 0 1 7
DDDDM 1 2 0 1 0 0 0 4
DMDD 1 1 0 0 0 0 1 3
DDDDDM 0 0 1 0 1 0 0 2
DDDMD 0 0 1 0 0 0 1 2
MDDDD 2 0 0 0 0 0 0 2
DDDDDDDM 0 0 1 0 0 0 0 1
DDDDDMD 0 0 1 0 0 0 0 1
DDDMDD 0 1 0 0 0 0 0 1
DDMDD 0 0 1 0 0 0 0 1
DDMDDDDD 0 1 0 0 0 0 0
DDD 1
DMDDD 0 0 0 0 1 0 0 1
DMDDDD 1 0 0 0 0 0 0 1
MDDDDDD 1 0 0 0 0 0 0 1
TOTAL 51 26 61 8 18 37 244 445

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
690 Constituent order

)) Summary in percent:80
Sentence Total In %
structure
TOTAL 431 96.9
Μ 244 54.8
DM 85 19.1
MD 39 8.8
DDM 22 4.9
DDDM 12 2.7
DMD 8 1.8
DDMD 7 1.6
MDD 7 1.6
DDDDM 4 0.9
DMDD 3 0.7

The main clause may also be split by an intervening dependent clause, as is

shown in (978). In this case, the subject is shifted in front position by focus,
followed by a subordinate clause and the main verb.

(978) Main clause split by dependent clause

a) With hana 'when':
mi hetau=te ahe-uru-a=ne hana turu-a ht-e
[that HRS=FOC] [get.drunk-PL-3ps/D=SUB when] [arrive-NTR go-3ps/E]
[S] [DEP] [V]
'That [man] arrived when they [the people] were getting drunk.'

b) With hau 'because':

nii hetau=te raj rauihirii be-uru-a=na käu, eseneta-e
[that HRS=FOC] [for really telI-PL-3ps/D=SUB because] [believe-3ps/E]
[A] [DEP] [V]
'He believed it because they really told him [that his wife was lost in the swamp].'

c) With participle clause:

lejhn we-tQa hetau=te ... kauatga-1 itpau-i,
[one PSN-only HRS=FOC] [good-PRT live-PRT]
[S] [DEP1]

80
All other types, i.e. the 14 cases which make up the difference between 445 and 431,
are well below 1%.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
 Position of dependent clauses 691

heri-a kana+hoaaun-era=ne rihihel ne=lu

[want-3ps/D our.creator=SUB like] [be-3ps/E=REM]
[DEP2] [V]
'Only one (man) Noah ... [out of all people that God had created] lived well, and was
like God wanted him.'

Complement clauses generally precede the main clause, apart from very few
exceptions (cf. §23.3). Apart from that, constituent order is largely unaffected
by other clause types. The main difference is that content questions involve a
focused interrogative pronoun. Otherwise, questions have the same syntactic
structure as declaratives (cf. §21). Similarly, imperatives do not exhibit
syntactic peculiarities (cf. §15). Focus is discussed in greater detail in §19
In summary, Urarina has a constituent order that can be characterised as
pervasively OVA/VS, which is manifested both qualitatively and statistically as
the unmarked order in transitive and intransitive clauses, both as independent
and dependent, with affirmative as well as with negated verbs. The few cases in
which an alternative (AOV/SV) order is possible are confined to pragmatically
dependent contexts that involve focus or emphasis with subsequent fronting of
the subject. Adjuncts such as PPs and adverbs predominantly occur before the
verb, mostly in the periphery of the clause.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:34 PM
19 Focus

In §18.3, it was highlighted that any variation to the regular constituent order is
related to a frontshift of the subject or other constituents to mark focus. To
clarify what is meant by "focus" in this grammar, some of the terms related to
this will be defined first.
As there is a lack of uniformity in terminology within the linguistic community,
there is considerable degree of disagreement among linguists regarding the
definition of "focus" and other categories of information structure such as
"topic", "theme", "rheme", and "new vs. given information". Most authors
agree that the notion of focus and topic are related to the concept of old (or
given) information vs. new information. Old information may be characterised
as knowledge which is available to the listener at the time of utterance through
general knowledge or because it has been introduced before in the discourse. In
many languages, old information can be expressed through definiteness. New
information, in contrast, involves material that is added to the discourse for the
first time. In an oversimplified manner, topic is often characterised as "what the
sentence is about" whereas focus tend to be described as "an element of a
sentence with particular prominence". These definitions are brutal
generalisations of a concept that is too complex to be discussed in detail here.
Gundel & Fretheim (2005) distinguish between relational and referential topics
- the former referring to the theme of an utterance (which is complemented by
relational or information focus as representing the new information about the
topic), the latter as referring to topic as the current centre of attention. In a
similar way, Gundel & Fretheim distinguish between two major types of focus -
the type just mentioned (information focus) and contrastive focus, understood as
"material which the speaker calls to the addressee's attention, thereby often
evoking a contrast with other entities that might fill the same position." Both
types of focus "may be syntactically coded by placing the relevant constituent
in a syntactically prominent position" (cf. Gundel & Fretheim 2005:6). Other
languages may express such prominence prosodically, e.g. by intonation. In
fact, both topic and focus are typically represented by some formal marking,
which may be of syntactic, morphological, or phonological nature.
Urarina has only one means to indicate any sort of prominence: by attaching
an enclitic directly to the emphasised constituent, which occurs is in initial
position of the sentence (with frontshift where necessary; cf. §18.3). Focus in
the sense of contrastive focus as characterised above probably is the most
common function of this enclitic, which is the reason for glossing it as "FOC"
in this grammar. However, the functions of this marker are much more

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Position of dependent clauses 693

widespread and may also include topicality (in the sense of relational topic). A
distinction between these functions is only possible through thorough
knowledge of the discourse. Thus, topic, contrast, or relevance of a fronted or
marked constituent for the sentence may be viewed as emphasis in general. 81
One could therefore refer to the focus marker in Urarina as a general device to
mark prominence. While examples for the fine distinctions in its function are
discussed in §19.1, one has to bear in mind that its glossing as a focus marker
throughout the rest of the grammar is a terminological simplification of its
actual functions.
Emphasis or prominence, which includes the functions mentioned in the
previous paragraph, is marked by enclitics that specify co-reference of the
emphasised constituent with the subject of the clause. In negative clauses and
in dependent clauses, these cannot be employed, but constituents are fronted
without additional formal marking (cf. §19.2.4 and §19.4). With the special
expressions as indicated above ('also', 'even', 'must'), the use of focus markers
is optional. Constituents that already occur in clause-initial position, such as
objects in a transitive clause or most adverbs, can also be marked with a focus
enclitic.
In any given text, a significant number of Urarina sentences contain a focus
marker. There are three focus-marking clause enclitics that are co-referenced
with the subject of a clause. Consequently, there can only be one focus marker
per clause. Any one constituent of a clause may be focused, but sometimes,
more than one constituent occurs before a focus marker, which is discussed in
§19.4. In polar questions and in dependent clauses (cf. §19.2.3), the focus
marker is not used, whereas it does occur in other clause types, such as in
content questions (cf. §19.2.2). With imperatives, focus markers play a special
role, as they may have consecutive/purposive function (cf. §19.2.1).
As discussed in §11, the presence of a focus marker in a sentence contributes
to the choice of the verb conjugation: With a focus marker present, the E-form
of the verb must be used (cf. § 11.2). The three focus enclitics are =ne, =na, and
=te, which occur with cross-referencing function to the subject of the main
clause. While =ne and =na refer to most 1st person forms, =te marks all other
persons. In addition, the distribution of the three focus markers is subject to free
variation between the two forms =ne and =na.
- With lsg subjects, the focus marker is =na (or variant =ne)
- With lpl/ex subjects, the focus marker is =na (or variant =ne, mainly in the
NT)
- With all other subjects, the focus marker is =te.

81
As context is particularly relevant for the interpretation of focus, the background for
each example in this chapter is given in form of the gloss of the preceding sentence.
82
Out of practical considerations, I refer to these clitics as "focus markers" here.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
694 Focus

(979) Agreement of focus markers with subject marking on the verb

Person Marker
lsg =na / =ne
2sg =te
3sg -te
lpl/ex =na / (=ne)
lpl/du =te
lpl/in =te
2pl =te
3pl =te

The fact that the forms for lpl/du and lpl/in are cross-referenced with =te is
significant, as it reflects the fact that whenever the listener is involved in a form
of ' w e ' , this is expressed by the person cross-referencing system marked by the
focus enclitics: In the same way as =te refers to second and third person
subjects, it is used with the two lpl forms that include the listener, i.e. a 2ps
referent.
The variation between the two clitics =ne and =na is not homogeneous, if
one compares their occurrence in 23 selected texts of the database with their use
in the NT. In my data, =na is not only prevalent when occurring with lsg and
lpl/ex, but it is not attested at all with lpl/ex. In contrast, the much larger
database of the N T contains more examples where =tie is cross-referenced with
the subject, being either lsg or lpl/ex (cf. (980)). Since the N T text is largely
based on the Chambira dialect, the difference could be partly based on dialectal
variation. However, as variation also occurs within dialects, one might assume
that diachronic factors are at work. Speakers do not recognise any differences in
meaning between the two different markers.

83
Ron Manus (2002 p.c.) suspects that the variation occurring in the NT examples
involves a certain systematicity: regarding the six attested cases of kanakanu=na, he
suggests that 'all six of these carry a claim of distinction of speaker from others in
the context.' With respect to the 23 counts of kanakanu=ne, he assumes that they
'tend to state a feature of speaker's identity'. The variation with the lsg pronoun
'seems to reflect speaker preference'. However, Manus also mentions that the
variation occurs within the same speaker.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Position of dependent clauses 695

(980) Occurrences of =ne and =na in two databases

Person Marker Database NT
lsg =na 13 66
=ne 5 106
lpl/ex =na 3 6
=ne 0 23

Interestingly, the same variation between =ne and =m is observed for the
subordinate marker. While it consistently appears as =rte in the NT, speakers of
the Espejo dialect use these two clitics in free variation (cf. §20.1.2). Even
though the focus markers are not semantically or functionally related to the
subordinate marker (and thus must be homophones), the parallelism in variation
could be accounted for in terms of phonological variation. This however, is not
supported by any rule-based alternation of this kind. With focus markers and
with the subordinate marker, the variant =na appears to gain dominance, i.e. the
use of =ne is becoming increasingly less common.
The following sentences illustrate the use of the different markers in
examples with different lps subjects. The introducer in (981c) refers to the first
verb ('let's take') and is followed by a jussive form, which does not require an
introducer (cf. §15.1.3). In (98Id), the verb for 'say' is not explicitly mentioned,
but implied by the PP to mean 'he spoke with his father-in-law'.

(981) Subject cross-referencing with focus markers

a) lsg [Preceding: Ί took a stick and touched him [the jaguar] - he was dead.']
nii kahe=ne kwituku-it
[that from=FOC: 1 sg] [know-lsg/E]
[PP] [V]
'From that [event] I know it: [Big animals can scare you a lot]'.

b) lpl/ex [Preceding: 'But as Lomai spoke like that, they said, "No, just do it right
here." And she replied: "No, we do not urinate here.'"]
kana temura-u=na hjam-tQaanu
[lpl/in edge-LOC=FOC: 1 pi] [urinate-lpl/ex]
[LOC] [V]
'We urinate at the edge of our [world].' [Lit. 'at our edge']

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
696 Focus

c) lpl/du [Preceding: 'So Lomai said: "Return its shell to me!" Therefore, one of the
men said:']
kwane=te kaa lejhn kari ama-aka=te hoara-kuru-mu
[let's=FOC] [this one shell] [take-1 pI/du=FOC] [see-PL-JUSS]
[INTR1] [Ol] [VI] [V2]
'Let's take this one piece of shell so that they [the others] would see it!'

d) lpl/in [Preceding: 'There, at the Airico River there was a man named Raimondo.
That man had set a trap in a place called Baharikinya and as he said "I will go to look
for it the day after tomorrow" and when his son-in-law said "Why don't we go
tomorrow", [they went]')
huara-a ku-a=ne ham, d$aha-l, ajane, ama-u=tge=te
[see-NTR go-3ps/D=SUB when] [come.on-PRT yes] [take-IMP=PLT=FOC]
[DEP1] [DEP2] [VI]

kwaa-m-aka=t(x raj nakana ajjia

[see-DSTL-1 pl/du=PLT] [POSS father-in-law with]
[V2] [PP3]
'As they went to see, [he said] with his father-in-law, "Come on, yes, take it and let's go
to see!".'

The focus enclitics =ne, =na, and =te can occur after any constituent and,
subsequently, are not restricted to appear after verbs. As could be expected, the
focus marker follows other clitics such as the reportative marker (cf. (982a):
su-a=he=te 'kill'-3ps/A=REP=F0C). However, when a focus marker is
attached to verbs, its position is not necessarily in the very last slot: As shown
in (982b,c), it can be followed by the emphasis marker =ra, which itself
occupies the final slot.

(982) Position of focus marker with regard to other clitics

a) FOC in final position:

su-a-he-te ku-e
kill-NTR=REP=FOC go-3ps/E
'He (A) says that he (B) went to kill it.'

b) EMF following FOC:

eresi=he=te=ra uu-re=1
tomorrow=REP=FOC=EMF come-IRR:3ps/E=ASS
'He (A) says that he (B) will come tomorrow.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Functions of focus markers 697

c) EMF following FOC:

bi-ahe-l=te=ra basihj-e
tell-CNT-PRT=FOC=EMF steal-3ps/E
'Advising [you] he [a thief] steals from you.' [To mean the contrary: Ά thief does not
tell you when he robs you'.]

19.1 Functions of focus markers

As mentioned in the introduction to this chapter, what I call focus here can also
have functions that go beyond a narrow definition of focus. In particular, this
regards the overlap of focus and topic. A clear example for the focus function of
the enclitic is found in (983). The lsg pronoun must be marked because it is
shifted into initial position. In this example, it marks contrastive focus, which
becomes clear through the context: kamt is emphasised in opposition to 'my
sister', whose function is mentioned in the preceding sentence.

(983) Merged topic and focus function on subject

[Preceding: 'Go away and catch my older sister', she said, 'She is the one sent for you.']
kanu=na hana+tgae paara kuseera=nuke letono ne-t
[ 1 sg=FOC: 1 sg] [besides+also] [2pl assistant=PURP envoy be-PRT]
[S] [ADV] [DEP]

rte-ü=ra
[be-lsg/E=EMF]
[V]
Ί instead, am sent to be your assistant.'

An area where the enclitics clearly mark focus is in answers to content

questions. The question which precedes the answer puts one specific constituent
(also see §19.1). Correspondingly, the reply fills the respective syntactic slot
and marks it with a focus enclitic, as illustrated in (984). In (984a), a locational
argument is focused and the answer to this content question focuses on the
questioned constituent. Example (984b) shows a temporal adverb in focus, as a
reply to a time-related question. Similarly, in (984c), a copula complement
occurs with focus function, and in (984d), the subject appears with =te.
However, while the presence of a focus marker in answers to content questions
is preferred, it is not obligatory. The same observation can be made with regard
to content questions - these can contain a focus marker optionally (cf. §19.2.2;
§21.1).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
698 Focus

(984) Answers to content questions

a) [Preceding: 'Where is Medardo?']
n=uhuana hana=te ku-e
[3ps=field inside=FOC] [go-3ps/E]
[LOC] [V]
'He went to his field.'

b) [Preceding: 'When will you come?']

eresi=pe u-re-ti=ni
[tomorrow=FOC: 1 sg] [come-IRR-1 sg/E=ASS]
[ADV] [V]
Ί will come tomorrow.'

c) [Preceding: 'What is this?']

raj suri+lenow hoitgana ajjia=te ne
[POSS intestine+food blood with=FOC] [be:3ps/E]
[CC] [V]
'It's the content of his stomach [mixed] with blood.' [About a peccary that is being cut]

d) [Preceding: 'Where is Manuel's house?']

manoel lureri=te tahia ne
[PSN house=FOC] [over.there] [be:3ps/E]
[S] [LOC] [V]
'Manuel's house is over there.'

The examples (985) to (988) reflect further functions of the marker. In some
examples, focus and topic are not clearly distinguishable and the function is
determined by the context. Alternatively, both functions may merge and express
topic and focus at the same time. The examples cited below also illustrate that
the concept of "new" and "old" information is not relevant for the distinction, as
both functions (topic and focus) may occur with old or new information.
In (985a,b), all necessary background information is provided by the
repetition of the words describing the event. Thus, the clause marked with the
focus marker is referring to "old" information. The use of =te here could be
interpreted as focus, emphasising the event of going as the relevant action for
the arrival. Alternatively, one could argue that the event of going is the current
topic of the sentence. Thus, it remains undecided, with what function te is used
in this example.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Functions of focus markers 699

(985) Focus/topic, old information

[Preceding: 'Therefore, he told his novice: "Let's go to see your people!']

a) Sentence 1:
na-a hau hetau, ku-ι, ku-ure
[say-3ps/D because HRS] [go-PRT] [go-3pl/E]
[DEP1] [DEP2] [V]
'As he said so, going, they went.'

b) Sentence 2:
ku-uru-a hau hetau=te, turu-uru-a ku-ure
[go-PL-3ps/D because HRS=FOC] [arrive-NTR go-3pl/E]
[DEP1] [V]
'As they went, they got to arrive.'

The constituents marked with =te in (986a,b) clearly refer to topic, which,
however, does not implicitly exclude an interpretation as focus or emphasis. In
both examples, a subject is shifted to the front. In contrast to the examples in
(985) these constituents represent "new" information, which is evident from the
context: (986a) functions as the title of a narrative; (986b) is the first sentence
of a narrative. Thus, the participants are only introduced here and are not known
from previous references.

(986) Topic, new information:

a)
lejhii ranuna=te nitahe
[one girl=FOC] [get.lost:3ps/E]
[S] [V]
'One girl got lost' [title of story]

b)
lejhii kat$a=te miirjka ke katQa-uru kuti-hjai amia-rate
[one man=FOC] [working.session VLI man-PL invite-PRT] [work-CAU2]
[A] [DEP] [V]
O n e man invited people to a working session and made them work.'

The sentence in (987) contains a fronted pronoun, which may be characterised

as "intrinsically old information", since both participants know that the referent
is involved from the discourse (the conversation had been going on for a while
with one speaker instructing the other of what to buy for him). The function of
=te here can thus be described as focus, since 'you' is not exactly the topic of
the conversation.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
700 Focus

(987) Focus, old information

[Preceding: (Speaker Α): Ί just said that I will leave tomorrow.' (Money was
mentioned about ten sentences earlier)]
kuriki kaute ke atti, lejhti ka=raj relor
[money rest VLI] [please] [one lsg=for watch]
[PP1] [ADV1] [01]

u-a-u, ii=te ßäe hoituku-i ka=kit$a siiri-i

[come-CAUl-IMP] [2sg=FOC] [already] [know-2ps] [this=boy have-NOM
[VI] [A2] [ADV2] [V2] ([DEP2])

tokuanei maraka ne-t ne-naa

compared.to brand be-PRT be-NOM]
[DEP2]
[Speaker B]: 'With the rest of the money please bring me a watch, you already know,
the brand like the one that these boys have.'

A n o t h e r e x a m p l e that contains a constituent in focus is given in (988a). T h e

person n a m e ' L o m a i ' has this function, as it is emphasised as the (only) person
w h o w o u l d have p o w e r over the night. Since she has not been mentioned b e f o r e
in the text, this reflects " n e w " information.

(988) Focus, new information:

[Preceding: 'This isn't good. How can these plants grow nicely if there is no dew from
the night? Therefore, God said:']
lomaj=te hanonaa siiri-i-tQa ne=ra
[PSN=FOC] [night have-NOM-only] [be:3ps/E=EMF]
[CS] [CC] [V]
'Lomai is the one who has [power over] the night.'

It is evident f r o m the examples given above that the distinction between " n e w "
and " o l d " information is not encoded by the focus marker. However, also the
distinction between topic and f o c u s is not specifically marked, as it requires
background knowledge. Excerpts f r o m a conversation over short w a v e radio as
in (989) illustrate the overlap between focus and topic.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Functions of focus markers

(989) Overlap of topic and focus in conversation

a) Speaker A (Question):
d^atoarn-a mtQae ikito-o, ta u-a elo=ne
[be.how:3ps/A] [there PLN-LOC] [NEGQ] [come-3ps/A] [rain=NEGF]
[VI] [LOCI] [INTR2] [V2] [S2]
'How is it over there, in Iquitos? Does it not rain?'

b) Speaker Β (Reply from Iquitos):

raka=te u-e elo hatal tabal, hata-1 mihjaipeti-i
[yesterday=FOC] [come-3ps/E] [rain] [very much] [very be.strong-PRT]
[ADV] [V] [S] [ADV] [ADV]
'Yesterday it rained very much, very strongly.'

c) Speaker A (Reaction):
kahia-netonaj=te t$u hjä u-na heri-a elo,
[over.here-also=FOC CRTN just] [come-INF want-3ps/A] [rain]
[ADV] [V] [S]

d$atoama tQalti
["I.don't.know.for.certain"]
[idiom]
'Over here as well it looks as if it wants to rain, but I don't know for certain.'

d) Speaker Β (did not understand):

najße-re-i repete-i-jte, hoatia auna-7a-u
[be.able-IRR-2ps] [repeat-2ps=CND] [not] [hear-NEG-lsg/A]
[VI] [DEP1] [INTR2] [V2]
'Would you repeat it, I did not hear it.'

e) Speaker A (repeating the information given before)

kahia-netonaj=te u-na here elo
[over.here-also=FOC] [come-INF want-3ps/E] [rain]
[ADV] [V] [S]
'Over here as well it wants to rain.'

kahia-netonaj u-e=lu raka

[over.here-also] [come-3ps/E=REM] [yesterday]
[ADV] [V] [ADV]
'It rained over here as well yesterday.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
702 Focus

raka=te u-e=lu elo tabai

[yesterday=FOC] [come-3ps/E=REM] [rain] [much]
[ADV] [V] [S] [ADV]
'Yesterday it rained a lot.'

Specifically, the use of the focus marker in (989) is associated with contrasting
one piece of information with another. The marker =te on the adverb
'yesterday' in (989b) appears to have two functions: It introduces a new topic,
thus the utterance is about "what happened yesterday". In addition, it puts focus
on the adverb, as the time of rain is in contrast to the present situation:
"Yesterday it rained (but not now)". In (989c), Speaker A changes the topic by
marking the locational adverb 'over here' with =te. Again, this seems to refer to
topic and focus: It is topic, because it implies that the general theme of the
utterance is the situation at the speaker's place. It is in focus, because it implies
a contrast to the situation in the listener's place (it has rained in Iquitos, but
there is no rain yet in the speaker's place). When asked to repeat his statement,
Speaker A then gives two different pieces of information in (989e): First, he
repeats his previous utterance, again using the locational adverb with the focus
marker. Secondly, he adds further information in two short sentences: In the
first part ('it rained yesterday over here as well'), no marker is used, which
could be due to the fact that =te already occurred in the previous sentence and
still is understood as the general topic. In the second part ('yesterday it rained a
lot'), the temporal adverb 'yesterday' is marked with =te. Again, this implies
both topic and focus: It is in focus, as the attention is drawn to the time of the
event, rather than the place, which is different from the preceding sentence.
However, it also makes 'yesterday' the theme of the sentence.
A distinction between given and new information as defined in the
introduction to this chapter is not explicitly marked in Urarina. In fact, the
marking of focus (with all its extended meanings) does not consistently coincide
with the notion of old vs. new information. A combination between all
functions is possible, as is illustrated by the following examples:
- Focus + new information: e.g. example (988)
- Topic + old information: e.g. example (1015c)
- Focus + old information: e.g. last sentence in example (989e)
- Topic + new information: e.g. example (986a)

In conclusion, it must be repeated that the exact interpretation of the

information structure with respect to the function of the focus marker requires
background knowledge of the context. However, the scope of a focus marker
always is on the constituent it is attached to. The following sections deal with
the formal aspects of focus marking in different clause types and on different
types of constituents.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Focus marking and clause types 703

19.2 Focus marking and clause types

19.2.1 Imperative clauses

A restrictive feature of focus markers is that they do not occur within imperative
clauses with 2ps reference. However, the imperative clause itself can be focused
when it is followed by another clause, (but not when the imperative is the only
clause of a sentence). By this, the clause that follows the imperative receives
consecutive reading, as is illustrated in (991). It should be noted that this
function is not observed with prohibitive constructions, which corresponds to
the general absence of focus markers in negative clauses (cf. §14.1). The
consecutive function of the imperative when followed by certain types of
dependent clauses is also discussed in §15.3. Note that a focus marker
frequently occurs after the hortative introducer hvane, which is illustrated in
(990).

(990) Focus marker after introducer hoane

[Context: After sermon in church service]

kwane=te kana+kwaaun-era baha-akatge
[let's=FOC] [our.creator] [ask-lpl/in]
[INTR] [Ο] [V]
'Let us pray.'

The sentence in (991a) exemplifies the consecutive function after a 2ps

imperative. The hortative clause Ί hear' is interpreted as a direct consequence
of the command 'Repeat!' which precedes it. In (991b), the imperative clause
'let me go' is marked with the focus marker for lsg, followed by a hortative
clause that is meant to express the result of letting Lomai go.

(991) Focus marking of imperati ve c lause

a) [Context: Speaker did not understand what the other said]

repite-u=na auna-e
[repeat-IMP=FOC: 1 sg] [hear-HORT]
[V] [DEP]
'Repeat so that I would hear [well]!'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
704 Focus

b) [Preceding: 'As they were dancing, Lomai had to urinate and said:']
ka=jane-u-tQe=na hauria hia-ni=tg,a
[lsg=][let-IMP=FOC:lsg] [first urinate-DSTL.HORT=EMF]
[0=] [V] [DEP]
'Let me go so that I can go to urinate first!'

A consecutive meaning can also apply when the imperative is not implicitly
marked for focus, as illustrated in (992).

(992) Consecutive function of imperative without focus marker

[Context: An elder encourages a man who is cutting a peccary to share the skin
generously:]
nii barii te-u huura=nuke
[that skin] [give-IMP] [increase=PURP]
[Ο] [V] [DEP]
'Give [him] more of that skin!' [Lit. 'Give that skin in order to increase [what he has]']

While the presence of focus markers is very common in this context, their use is
not obligatory, as the examples in (993) show. In these sentences, the subject is
marked only by the hortative marker on the verb. In (993a) 'let me get it', this is
indicated by the hortative suffix -e for lsg. In (993b), the same form is included
in the distal form of the verb, again with no marking for focus on the introducer.

(993) Hortative construction without focus marker

a) [Preceding: 'There will be magnetic soil and beautiful axes that are like the ones we
have.']
mi hana d^atoanei hoane ra-e, na-t
[that when] [how] [let's] [receive-HORT] [say-PRT]
[CNJ] [DEP1] [INTR1] [VI] [DEP2]

raa-naa heri-ki-tQe=ne ku=te neda-ri-h-toe=l=tQa

[receive-INF want-2ps-PL=CND] [there=FOC] [stay-IRR-2ps-PL=ASS=EMF]
[DEP3] [ADV2] [V2]
'Then, when you say, "How? Let me get it!", when you want to grab it, you (PL) will
stay there.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Focus marking and clause types 705

b) [Preceding: 'As they were walking and listening [to the talking trees], one of them
had to go into the bush. As he had to go into the bush, [he said]:']
kwane hauria ahaena-ana ku-ni=tQa
[let's] [first] [bush-LOC] [go-DSTL.HORT=EMF]
[INTR] [ADV] [LOC] [V]
'Let me go into the bush first!'

As indicated, a focus marker does not occur within an imperative clause, which
is illustrated by (994a). An emphatic function of an imperative clause, however,
can be implied by the use of emphatic particles, such as läe in (994b); for its
function, cf. §5.15).

(994) Emphasis in imperative clauses

a) No focus marker allowed:
ena(*=te) itQa-u
now(*=FOC) do-IMP
'Do it now!'

b) With emphatic particle:

arana tooae läe te-u
[floor above] [CHNG give-IMP]
[PP] [V]
'Just put it right on the floor!'

19.2.2 Questions

Most interrogative pronouns can take a focus marker, which implies focus
function for the respective constituent. In polar questions, focus markers do not
occur. The examples in (996) illustrate the respective combinations with a
variety of different content questions. Note that there is no formal distinction
between 'who' and 'what' (cf. (996a)). The factors that influence the
interpretation of +/- human reference in this regard involve transitivity, person
marking, and contextual aspects, which is further investigated in §21.1.1.
However, focus marking also is one criterion among others that supports the
correct interpretation of Since subject focus is more typical than the
marking of Ο arguments with =te, the presence of this marker after dza usually
• ο-j

indicates that the referent is human and thus interpreted as 'who'. An elicited
example is given in (995). Theoretically, a range of interpretations could apply

84
An A argument is more likely to have a human referent than an Ο argument is.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
706 Focus

to both sentences, as grammatical function and animacy are not predetermined.

The possible readings for each sentence are:
1) ' W h o has seen it/him?'
2) 'What has seen it/him?'
3) 'What has he seen?'
4) ' W h o m has he seen?'
However, the presence or absence of a focus marker implies certain preferences:
In (995a), the absence of a focus marker results in the interpretation of d$a as
inanimate, which in turn implies its function as an Ο argument. In (995b), the
insertion of the focus marker results in a preferred reading in which d$a is
interpreted as a subject, with human reference.

(995) Focus marking with d$a

a) Preferred inanimate interpretation

d$a kwara-e
[what] [see-3ps/E]
[O] [V]
'What has he seen?'

b) Preferred animate interpretation

d$a=te hoara-e
[what] [see-3ps/E]
[A] [V]
'Who has seen it?'

In (996a), the decision regarding d ^ a ' s grammatical function as an S argument

with human reference is obvious, as najrtia 'be able' is intransitive and hence
the focus marker could not refer to an Ο argument. Thus, the interrogative
pronoun d$a is interpreted as ' w h o ' . The fact that d$a is an interrogative
pronoun with a wide semantic range is also demonstrated by (996b), where it is
interpreted as 'where from' together with kahe ' f r o m ' . The word for ' h o w ' as
exemplified in (996c) is morphologically complex, as it consists of the verb root
for ' b e h o w ' and the participle suffix. Examples for the occurrence of the focus
marker with 'where' and ' w h y ' are shown in (996d,e). The example in (996f)
shows the use of d^asü, which corresponds to 'how is it possible?', implying
' h o w ' and involving a certain degree of surprise.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Focus marking and clause types 707

(996) Interrogative pronouns with focus marker

a) 'Who'/'What' [Preceding: 'As it was clearing up, he sent one of his sons.']
d$a=te najpe-re hoar a-a ku-a=ne=ta,
[who=FOC] [be.able-IRR:3ps/E] [see-NTR go-3ps/D=CND=FRS]
[SI] [VI] [DEP1]

na-a hau, kanu=na najße-re-ü

[say-3ps/D because] [lsg=FOC:lsg] [be.able-IRR-lsg/E]
[DEP2] [S3] [V3]

raj kalaui ahitca ke taka-t hoara-a ku-e

[POSS son major VLI meet] [see-NTR go-3ps/E]
[DEP4] [V4]
'As he said: "Who could go to see?", his major son [saying] "I can" met him and went
to see.'

b) 'Where from' [Preceding" 'When he heard his children call him "Daddy", he was
happy and always asked his wife:']
d$a kahe=te kanaanaj eru-i=ta
[what from=FOC] [child] [find-2ps=FRS]
[PP] [O] [V]
'(From) where did you find the child?'

c) 'How' [Preceding: 'The people said: "Our hooks could not catch the fish". Therefore
the priest asked his novice:']
d$atoanei=te kuhiou-ure katQa-uru-ta
[how=FOC] [fish.with.hook-3pl/E] [man-PL=FRS]
[ADV] [V] [S]
'[But] how did the people fish?'

d) 'Where' [Preceding: 'He turned and saw them following and asked them "What are
you looking for?'"]
d$u=te ne-i=ta, katQa rela-era
[where=FOC] [be-2ps=FRS] [man teach-AG]
[LOC] [V] [S]
'Where do you live, Master?' [NT: John 1:38]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
708 Focus

e) 'Why' [Preceding: 'As they were molesting him, the priest said" "Hey, it seems that
the people intend to kill me.'"]
d$anu=te kaaunei k-aj-te-kure=ta
[why=FOC] [like.this] [lsg=][AUX-INT-3pl/E=FRS]
[ADV] [ADV] [0=] [V]
'Why are you doing [like] this to me? ' [Priest talking to the people]

f) 'How possibly' [Preceding: 'As he was happy about having the children, he asked his
wife:']
d$asit=te katoane-1 it$a-i=ta
[how.possibly] [be.like.this-PRT] [do-2ps=FRS]
[ADV] [ADV] [V]
'How is it possible that you did it like this?'

It must be emphasised that all interrogatives used in (996) can also occur
without a focus marker. The presence of a focus marker merely emphasises the
focus function already implied through pragmatic factors - the universal
reference to focus by any content question.
The two words for 'how much/how many',y2ataha-ϊ a n d ß a a o h w a - l are not
attested with a focus marker. Morphologically, these have the same structure as
' h o w ' , consisting of a verb and a participle suffix. The reasons for the absence
of the focus marker with these forms are unclear, but could be due to their
generally low frequency.
In polar questions, the cross-referencing enclitics are not used. However, a
special focus function can be implied by the use of the interrogative marker
=na: While its regular position is after the verb, it can also occur on the first
constituent. In this case, it marks focus or emphasis in the same way as a focus
marker would do (also cf. §21). The difference is illustrated by the (elicited)
examples in (997a,b), which contrast from each other by the position of the
interrogative enclitic.

(997) Focus function of interrogative =na in polar questions

a) Regular position:
katai kurete-i=pa
fishing.hook buy-2ps=INT
'Have you bought fishing hooks?' [General inquiry]

b) Focus position:
katai=jia kurete-i
fishing.hook=INT buy-2ps
'Is it fishing hooks that you bought?' [Focus on object]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Focus marking and clause types 709

19.2.3 Focus and dependent clauses

As mentioned in §18.3, focus in dependent clauses is not formally marked, but

indicated by constituent order only, i.e. a constituent may be fronted without the
attachment of a focus marker. Such behaviour, however, is extremely rare, as
the figures in table (942) (§18.1) indicate: only in 12 out of 354 dependent
clauses (equalling just over 3%), the subject was fronted - which is plausible
for two reasons. Firstly, there is a strong tendency for focus to be marked only
once per sentence. (Exceptions from this are discussed in §19.5.) Secondly, the
information given in a dependent clause usually is peripheral as compared to the
information given by the main clause. Prototypically, a dependent clause
comments on the action or event described in the main clause. It is common to
mark the entire dependent clause for focus, in order to convey the relevance of
the information for the main event. However, since the dependent clause is
treated like a single constituent in this case, it would be counterintuitive to
emphasise single elements within the dependent clause. The examples in (998)
illustrate the rare occurrence of SV constituent order in dependent clauses. For
(998a), it must be noted that focus is also marked through the presence of
-netomj 'also' - which is one of the exceptional expressions to predict order
variation (cf. §18.3). Note that in this example, the focused subject refers to the
main clause, as overt subjects with participle clauses are untypical. In (998b),
there is no overt marking for focus in the dependent clause. The quotation 'let
us see and do it' is embedded in the dependent clause for 'as one again said'.

(998) Focus in dependent clause

a) [Preceding: 'When they saw how others made their fields, they were surprised.']
riete ßäe akauru-netomj nerela-1, jiäe enene nitoanei
[but] [already] [3pl-also] [learn-PRT] [already] [nowadays] [like.that]
[CNJ] [ADV] [A2] [VI] [ADV2] [ADV2] [ADV2]

n=ukwana ha-ure
[3ps=field] [make-3pl/E]
[02] [V2]
'But nowadays already, they, too, make their fields like that, learning already.'

b) [Preceding: 'One man said, "What would happen if I slowly went with the paddle to
get (the axe)?'"]
rtii hau hetau=te leotga hoajtel,
[that because HRS=FOC] [other] [again]
[CNJ2] [SI] [ADV 1 ]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
710 Focus

d^aha-ι, ajane itga-l hoara-u, na-a hau hetau,...

[come.on-PRT yes do-PRT see-IMP] [say-3ps/D] [because HRS]
[QUOT] [VI] [CNJ1]

herae kihja ke ku-ati-1 ra-a

[slowly] [paddle INST] [go-DER-PRT receive-3ps/A]
[ADV3] [PP3] [V3]
'Therefore, as one [man] said, "come on, yes, let's see and do it', ... he slowly went
with the paddle to get it.'

19.2.4 Negative clauses

A further context in which focus markers do not occur is in negative clauses. As

mentioned in §18.3, focus is achieved by shifting the relevant constituent to
clause-initial position without using the focus marker. In (999), the subject is in
front position, but no focus marker is present. However, note that the verb is
obligatorily marked with the Ε-form, which otherwise co-occurs with the focus
marker.

(999) Frontshift without focus marking

[Preceding: 'Therefore, there is no such thing any more.']
enene ne-nakauru enamanaa-kuru katQa-uru huituku-ur-ene
[nowadays be-those.who young.man-PL man-PL] [know-PL-NEG:3ps/E]
[A] [V]
'The young people from nowadays do not know it.' [About traditional naming]

The pragmatic functions related to the front position of the subject in this case
again are ambiguous, since both focus and topic functions are implied here. On
the one hand, the enclitic could refer to topic, since 'the young people' have not
been previously mentioned in this text and it is plausible that the speaker now
turns to their situation. On the other hand, 'the young people' are in contrast to
'the old people', whose customs were discussed in previous sentences.
Therefore, contrastive focus is also implied.
Another example for the combined function of focus and topic is given in the
negative clause in (1000), where the subject is in preverbal position again, this
time preceded by a negative introducer, whose position can only be at the very
beginning of the clause. The 1 sg pronoun is in contrastive focus to 'the other
woman', mentioned before.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Focus marking and clause types 711

(1000) Frontshift in negative clauses:

[Preceding: 'He did not catch her [the "right" woman] any more and she [the wrong
one] said: "It is your fault.'"]
kivatia kanu i=tgene letono rü-a-ü=ta
[not] [lsg] [2sg=place envoy] [be-NEG-1 sg/E=FRS]
[INTR] [CS] [CC] [V]
Ί am not the envoy for your place.'

There is one construction that involves a focus marker and can precede a
negative clause: Expressions involving conjunction phrases, such as 'therefore',
'after that', or 'then' (cf. (1012)) frequently occur with a focus marker and can
be followed by a negative clause, as illustrated in (1001). The expression
mi häu=te in this example represents an element that has conjoining function
between this sentence and the preceding one and is not in the scope of negation.
Thus, the focus marker with this type of expression does not strictly violate
against the principle that focus markers do not occur in negative clauses.
However, subject agreement between the focus marker and the verb is still
marked and the presence of the focus marker requires the use of the Ε-form on
the verb. Thus, while the focused construction is not a constituent of the clause,
it still follows the concordance rule for cross-referencing. Since it is not in the
scope of negation, the co-occurrence of focus and negation within the same
sentence is compatible.

(1001) Focus before negative clause

a) 3ps subject concordance: [Preceding: 'Nowadays, young people don't know anything
about it.']
nii häu=te jiäe enene ne-ene mtoanei,
[that because=FOC] [already] [nowadays] [be-NEG:3ps/E] [like.that]
[CNJ] [ADV] [ADV] [V] [ADV]

ßäe sa-a=ne hau

[already end-3ps/D=SUB because]
[DEP]
'Therefore, it is not like that any more, because it has ended.' [About traditional
naming]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
712 Focus

b) lps subject concordance: ['So (the patrones) kept our parents there and they live
there until now.']
mi hau=na kuaae kuriiti kuane itgau-akaanu ene netohwei
[that because=FOC:lsg] [there] [PLN inside] [live-lpl/ex] [now until]
[CNJ] [ADV] [PP1] [V] [PP2]
'Therefore, we live on the Corrientes River until now.'

19.3 Scope of focus

In contrast to the clause types investigated above, the constituents in declarative

clause are subject to extensive marking with focus. In principle, all constituents
can be focused, i.e. they are shifted to the front if this is not already their
position. This applies to the arguments of the verb as well as to all kinds of
adjuncts. On the prosodic level, the focus marker is typically followed by a
pause; the focus enclitic carries High tone in most contexts.
The following sections illustrate focus of core arguments (S/A, O; see
§19.3.1) and adjuncts of various kinds (§19.3.2). As will become clear in §19.4,
the actual definition of the elements that are in focus position is not always
straightforward. In most cases, it will be one constituent (and not less than that),
but in some examples, a bigger "chunk" of the utterance, rather than a single
constituent is followed by the focus marker.

19.3.1 Focus-marking of core arguments

While focus of subjects (A/S arguments) is comparatively common in Urarina,

only relatively few examples for objects followed by a focus marker are
attested. A possible reason for this is that objects already occupy a position that
intrinsically implies focus. Only when an object is subject to focus, emphasis or
contrast in some way, it is marked with one of the enclitics =ne, =na, or =te.
The examples in (1002) show subject NPs of different kinds of syntactic
structure. In (1002a), the subject is a simple noun, and the verb is followed by a
dependent clause; in (1002b), the subject is a pronoun (this time separated from
the verb through a dependent clause), and in (1002c), it is a syntactically
complex NP followed by the hearsay marker hetau. In (1002d), the subject NP
involves the form for 'also', again followed by hetau. This is one of the cases
where "something more" than a constituent (namely the subject) is in focus
(also cf. §19.4).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Scope of focus 713

(1002) Subject in focus

a) Simple noun [Preceding: 'Therefore, they said: "It is not good that he [the spider
monkey] is chief.'"]
ruru=te najpe-re kuraanaa ni-a-ne=ra
[howler.monkey=FOC] [be.able-IRR:3ps/E] [chief be-3ps/D=CND=EMF]
[S] [V] [DEP]
'The howler monkey could be chief.'

b) Pronoun [Preceded by a)]

aka=te läe helaja-atga nelauria-l itQau-e=ra
[3sg=FOC] [CHNG same.place-only be.sitting-PRT] [live-3ps/E=EMF]
[S] [DEP] [V]
'He lives sitting in exactly the same place.' [About the howler monkey]

c) Complex NP [Preceding: 'Nobody knows how they found those stone axes.']
ßoaelu ne-nakauru katQa-uru hetau=te eru-rehete
[earlier be-those.who man-PL HRS=FOC] [find-HABl:3ps/E]
[A] [V]
'The people who lived earlier found them.'

d) Complex NP [Preceding: 'Therefore, they were very scared.']

nii htnajte-na-nemaahei hetau=te hatal nesoone
[that be.sick-NOM-also HRS=FOC] [very] [be.scared:3ps/E]
[S] [ADV] [V]
'That sick person also was very scared.'

In the f o l l o w i n g example, the focused subject occurs in a quotation. T h e

sentence is f r o m a narrative w h e r e the black monkey warns the w o m a n not to
disclose certain information about him.

(1003) Focused subject in quotation

[Preceding: 'He [the black monkey] arrived and removed her shed and all that and told
the woman to send her husband to get jagua fruits; [then] he applied the jagua to her
[body] and said:']
mhjauria be-i=jx, kati=te kaa tokuanei
[don't] [tell-2ps=NEGF] [black.monkey=FOC] [this compared.to]
[INTR1] [VI] [A2] [ADV2]

ka=raj kanaanaj hanoa-e na-ϊ, mhjauria

[ 1 sg=for] [child] [give.birth-3ps/E] [say-PRT] [don't]
[PP2] [02] [V2] [DEP3] [INTR4]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
714 Focus

na-Ί be-i=ße hauria-ne

[say-PRT] [tell-2ps=NEGF] [first=NEGF]
[DEP5] [V6] [ADV6]
'Don't tell [them], saying: "The monkey has given birth to the child for me", don't tell
it saying so, at first.'

Similar to the interpretation of the interrogative d^a (cf. (995)), when an object
is marked with a focus enclitic, it is distinguishable from a subject only if an
overt subject is present (in postverbal position), or from the context. For
instance, the example in (1004a), technically, could also be interpreted as 'Even
the canoes made it with stone axes.' This, of course is a rather unlikely
interpretation due to universal preferences in animacy, as human and animate
NPs or pronominals are more likely to be interpreted as agents of transitive
verbs. Example (1004b) exhibits a higher degree of ambiguity than (1004a),
since it could be interpreted as 'All kinds of [people] did it at first.' The correct
translation with 'all kinds' in object position is mainly conveyed by the context,
in which 'the ancient people' are mentioned at an earlier stage (namely, two
sentences before). Note that a focussed object can also have human reference, as
is shown in (1004c).

(1004) Object focus

a) [Preceding: 'With this, some have learned to make their fields, but not all.']
enamhja ni natii=te itQa-kure mi aheri+ku-teru ke
[canoe even=FOC] [make-3pl/E] [that stone+ASC-axe INST]
[Ο] [V] [PP]
'They made even canoes with those stone axes.'

b) [Preceding: 'Those who hunted with a bow even had to make their bows.']
itulere=te itQa-kure=lu hauria
[all.kinds=FOC] [make-3pl/E=REM] [first]
[Ο] [V] [ADV]
'They made all kinds of things at first [i.e. in those days].'

c) [Elicited, but implied 'Whom did he send?']

kanu=te letoa-e
[ 1 sg=FOC] [send-3ps/E]
[O] [V]
'Its is me whom he sent.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Scope of focus 715

19.3.2 Focus-marking of adjuncts

Apart from the core arguments of the verb, any type of adjunct can be marked
with a focus enclitic (but recall that there is no formal distinction between
adjuncts and external arguments). This includes postpositions, adverbs, and a
unit that could be called "conjunction phrase" (cf. (1012)). Focus markers can
also be attached at clause level to the last constituent of a dependent clause (cf.
(1015)). Their use with imperatives was discussed in §19.2.1. One peculiarity is
that there also are a few examples for the occurrence of a focus marker within a
serial verb construction, as will be explained further below (cf. (1016)).
As a restriction, focus markers are not cliticised to particles. Instead, the
particles hjä and t$u follow the focus marker (cf. (1017)). There is one
exception: the hearsay marker hetau, where present, always precedes the focus
marker, which could be accounted for by the fact that hetau is a (lexicalised)
combination of two enclitics (cf. §12.3.5). Focus markers do not follow
introducers, except the hortative introducer hoane, where the presence of a
focus marker is very typical (cf. §19.2.1).
In (1005), examples for focused PPs with different functions are listed.
There are no restrictions as to the type of PP that can be subject to focus. Again,
recall that the clause-initial position of PPs is formally and functionally
unmarked. Therefore, the insertion of a focus marker in this environment
implies focus, contrast, topic, or any other kind of emphasis as defined further
above. In (1005a), =te is attached to a PP involving ke, which has valency
increasing function here, licensing the additional argument 'that (thing)' for the
quotative verb naa 'say'. In (1005b), the PP with ke has instrumental function.
In example (1005c) the "recipient" (which is a maleficiary here), is marked with
raj and followed by the focus marker. Other functions of PPs, such as location,
time, or comitative, are illustrated in (1005d-f).

(1005) PP with focus marker

a) Valency increase with ke [Preceding: '... and he turned into a fallen tree. Being a tree,
this sort ofthing,...']
η» ke-te komo+enua na-ι ajte-kure
[that VLI=FOC] [fall+tree] [say-PRT] [say-3pl/E]
[PP1] [Ol] [VI] [V2]
'They call that thing "fallen tree".'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
716 Focus

b) Instrumental PP [Preceding: 'So her husband went to get a bamboo stick in order to
cut her.']
haurta ke hetau=te mi ita fiuaru-rehete-kure=lu
[bamboo.stick INST HRS=FOC] [that REC cut-HABl-3pl/E=REM]
[PP] [V]
'They used to cut each other with a bamboo stick.'

c) Beneficiary PP [Preceding: 'As the spirit said, "Come on, let's do it quickly", the
woman replied:']
hjauijiera, ka=kana raj=te nemuri kuane kalanoaae-ri-aka=ra
[no] [this=lpl/in for=FOC] [lake inside] [make.dirty-IRR-lpl/du=EMF]
[PART] [PP1] [PP2] [V]
'No, we would spoil the lake for us!' [To mean: Nobody will catch anything]

d) Locational PP [Preceding; Ά man had the people work for him, inviting them to a
minga, in order to clear [his field], [which was in a distance of] five hours by canoe.']
nii kahe=te turu-a u-e sonajpa
[that from=FOC] [arrive-NTR come-3ps/E] [afternoon]
[PP] [V] [ADV]

ßäe hauim-a raj kakunu=ne hana

[already get.lost-3ps/D POSS daughter=SUB when]
[DEP]
'He arrived from there in the afternoon, when his daughter had already got lost.'

e) Temporal PP [Preceding: 'Therefore, the Urarina do not want stay without this.']
ena mtohioei=te ne ißono ku-1 mi ajpa ne-nakauru
[now until=FOC] [be:3ps/E] [ayahuasca drink-PRT] [that with be-those.who]
[PP] [V] [DEP] [S]
'Until now there are those who live with that, drinking ayahuasca.'

f) Comitative PP [Preceding: 'Therefore, the others already returned home ahead of

her.']
lana ajjia hetau=te mki ku-e
[husband with HRS=FOC] [ADVRS] [go-3ps/E]
[PP] [PRT] [V]
'But she went with her husband.'

T h e e x a m p l e in (1006) contains the purposive enclitic =nuhe, which here

f o l l o w s the postposition raj ' f o r ' (cf. §20.2.4.1 about purposive forms). In this
case, the f o c u s marker follows the suffix.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Scope of focus 717

(1006) Focus marker after raj-ßuke

[Preceding: 'Besides, in addition to that, we build a school, if God wants, so that (for)
the children who are here can learn.']
esekwela raj=jiuke=te nelaate kalamina, material noble
[school for=PURP=FOC] [be.necessary:3ps/E] [zinc.roof material noble
[PP] [V] [S]

mrianaj
all. that]
'For the school, zinc roof, good materials, all that is necessary.'

Similar to PPs, different types of adverbs can be marked with a focus enclitic to
imply focus function. The sentences in (1007)-(1010) show examples for
adverbs that refer to location, time, manner, epistemics, and for the intensifier
hatai.

(1007) Locational adverbs with focus marker

a) Locational adverb tahia [Context: from conversation about how they killed the
peccary]
tahia=na kua-ka ku-a nukuja-a-kaanu
[over.there=FOC: lpl] [wait-NTR go-NTR in.vain-NEG-lpl/ex]
[ADV] [V]
'We did no go to wait over there in vain [but we found the peccaries there and killed
them].'

b) Locational adverb fat (cf. (993a))

raa-naa heri-ki-toe-ne ku-te neda-ri-ki-tQe=~i=tQa
[receive-INF want-2ps-PL=CND] [there=FOC] [stay-IRR-2ps-PL=ASS=EMF]
[DEP] [ADV] [V]
'If you want to get [the axe], there you will stay [forever].'

There also are some temporal adverbs that are attested in combination with a
focus marker. This includes ena 'now' as shown in (1008), and there are
examples for a few others, such as jiäe 'already', heritQanehel 'as before', and
adverbs related to daytime or temporal distance between days (e.g. raka
'yesterday'). However, there are no attested examples for a focus marker that
would follow the time-related adverbs kuajtei 'again' and hajti 'still'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
718 Focus

(1008) Temporal adverbs with focus marker

a) Temporal adverb ena [Preceding: 'How is the river?']

tabaae-ka, ena=te tabaa-he, ßäe taba-a
[be.big-CNT:3ps/A] [now=FOC] [be.big-CNT:3ps/E] [already] [be.big-3ps/A]
[VI] [ADV2] [V2] [ADV3] [V3]
'It [the river] has grown; now, it is rising, it already is high [enough].'

b) Temporal adverb eresi [Preceding: Ί haven't seen it yet. I haven't found out whether
they finished my document.']
heriane eresi=ßa hoar a-a ku-re-u=m
[maybe] [tomorrow=FOC: 1 sg] [see-NTR go-IRR-lsg/E=ASS]
[ADV1] [ADV2] [V]
'Maybe tomorrow I will go to see.'

Another type o f adjunct that can be followed by a focus marker are manner
adverbs. Attested examples are found with rautohwei 'easily', 'calmly' (cf.
(1009a)), surutii 'quickly', hiniikn 'together', and with nitoanei 'like that', for
which an example is given in (1009b). Recall that these are actually participle
forms o f verbs, (some o f which may be lexicalised; cf. §5.3). The absence o f
examples with further manner adverbs might be incidental.

(1009) Manner adverbs with focus marker

a) [Preceding: (1008a)]
rautohwei=te eno-a u-ri-ki=i
[easily=FOC] [enter-NTR come-IRR-2ps=ASS]
[ADV] [V]
'Easily you will enter.'

b) [Preceding: 'The people thought she had a baby from that spirit.']
mtoanei=te hatal eluna-i, hoaasißa käu hetau
[like.that] [very be.frightened-PRT] [fear-NOM because.of HRS]
[ADV] [DEP] [PP]

aj mi ranuna
[AUX:3ps/E] [that girl]
[V] [S]
'Like that, the girl remained very frightened and full of scare.'

The sentences in (1010) illustrate the use o f the focus marker after the epistemic
adverb heriane 'probably' and after the intensifier hatai 'very'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Scope of focus 719

(1010) Focus marker after other adverbs

a) Epistemic adverb heriane [Preceding: The jaguar did so for his booty; (besides):']
heriane=te ne raj kalaui
[probably=FOC] [be:3ps/E] [POSS son]
[ADV] [V] [S]
'Probably it had its creature.'

b) Intensifier [Preceding: 'They did not see [the spirit]; only that sick person saw him.']
hatäi hetau=te banehelanaala nesoone nii kunajte-na
[very HRS=FOC] [totally] [be.scared:3ps/E] [that be.sick-NOM]
[ADV] [ADV] [V] [S]
'That sick person was absolutely scared.'

In the same way as the adverbs illustrated above, there are non-adverbial
adjuncts for time and location that can be followed by a focus marker. A few
examples are illustrated in (1011), involving different locational adjuncts, such
as NPs marked with the locative suffix -«, NPs with rem 'place', and a
temporal expression indicating a period of time.

(1011) Focus marker after non-adverbial adjuncts

a) Location (cf. (981b)):
kana temura-u=na hjani-tQaanu
[lpl/in edge-LOC=FOC: lpl] [urinate-lpl/ex]
[LOC] [V]
'We urinate at the edge of our world.'

b) Location [Preceding: 'They went and arrived at Lomai's place. There they said:']
hn, ii rene=te kanakaana letoa-e amu-naa+kioaaun-era
hey [2sg place=FOC] [lpl/ex] [send-3ps/E] [Walk-NOM+create-AG]
[LOC] [Ο] [V] [A]
'Hey, the Creator of Ways has sent us to your place.'

c) Period of time [Preceding: 'In the end, they brought him to Iquitos in order to have
the doctors cure him.']
huusa atene-i=te nerauta-e mi kunajte-na
[eight month-PRT=FOC] [get.better-3ps/E] [that be.sick-NOM]
[TMP] [V] [S]
'Within eight months, that sick person recuperated.'

The majority of focus markers in narratives are found after expressions that can
be called "conjunction phrases", as they are composed of more than a single

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
720 Focus

conjunction. These are combinations of the demonstrative nii, followed by a

conjunction such as 'because', 'when', or 'after' to convey the meaning of
'therefore', 'then', or 'after that'. Some of these expressions also include a
postposition, which, however, cannot be translated literally. For instance,
niihäu ajßa ('that' + 'because' + 'with') is used to mean 'therefore', 'with the
situation being like that'. As a variant, nii hau raj ('that' + 'because' + 'for') is
used, conveying the same meaning. It may therefore be assumed that these
expressions are lexicalised. Their function is to relate the content of a sentence
with the preceding one. This is particularly frequent in narratives, where the
sequence of events is relevant for the structure of the text, but conjunction
phrases are also employed in other genres, such as in sermons, or in
spontaneous speech, where practical.
While all conjunction phrases are frequently accompanied by a focus
enclitic, they can occur without a focus marker. They can also involve the
hearsay marker hetau, which is placed after the conjunction (or additional
postposition, if applicable) and directly before the focus marker, if one is
present. Examples for two of the most common conjunction phrases are given in
(1012).

(1012) Conjunction phrases with focus marker

a) [Preceding: 'When they caught her, she cried.']

nii hana=te muku-ι ra-e raj neba
[that when=FOC] [catch-PRT] [receive-3ps/E] [POSS mother]
[CNJ] [DEP] [V] [A]
'Then her mother caught her and received her.'

b) [Preceding: 'She converted the lower part (of the child) into a lizard.']
nii hau hetau=te ku kutuhja-e
[that because HRS=FOC] [there] [curse-3ps/E]
[CNJ] [ADV] [V]
'Therefore, she cursed [the child] there.'

Note that the focus marker always agrees with the subject of the verb which
follows it and that the verb takes the Ε-form. This was also demonstrated for the
examples in (1001), where it was shown that the conjunction phrase has
conjoining function, but is not part of the clause. The sentence in (1013) is
another example for this: The conjunction phrase is followed by a focus marker
for lsg with cross-referencing function to the subject marked by the personal
suffix on the verb.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Scope of focus 721

(1013) Subject agreement between conjunction phrase and verb

[Preceding: 'When I did not believe in God, 1 did not obey; I was not in the mood to
follow.']
nii hau ajpa=na araj-säi kunajte-ι ne-reheto-o=lu
[that because with=FOC:lsg][many-times] [be.sick-PRT be-HAB 1 -1 sg/E=REM]
[CNJ] [ADV] [V]
'Because of that, I used to be sick many times.'

19.3.3 Focus marking in other environments

A dependent clause can also function as a unit to which the focus marker is
attached. As discussed in §18.5, the preferred position for dependent clauses is
before the main clause, which makes them similar to adjuncts, with respect to
constituent order. Thus, when a focus marker is attached to a dependent clause,
the clause is in focus. Specifically, this emphasises the relevance of what is said
in the dependent clause for what follows. In (1014), examples are given for
dependent clauses that involve a conjunction, (which normally requires a
subordinate marker, but note that it can be omitted, cf. §20.1.2).

(1014) Focused dependent clause with conjunction

a) With häu 'because' [Preceding: 'Because the Urarinas work among themselves; they
are learning to work helping each other.']
mtoam-a=ne häu=te jiäe ne-ene enene,
[be.like.that-3ps/D=SUB because=FOC] [already] [be-NEG:3ps/E] [nowadays]
[DEP] [ADV] [V] [ADV]

mtoanei fwatorö kivitu-kur-ene

[like.that] [patron] [know-PL-NEG:3ps/E]
[ADV] [Ο] [V]
'As it is like that, [slavery] is not there any more; they do not knowpatrones like that.'

b) With hau 'because' [Preceding: 'God said: "Go from here so that there be fish in
abundance for my children there, so there be (fish) in all little creeks, even in the
waterholes.']
na-a ofwa häu-na niki u-ü=ra
[say-3ps/D father because=FOC:lsg] [ADVRS] [come-lsg/E=EMF]
[DEP] [PRT] [V]
'Because my father said so, I have come.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
722 Focus

c) With hana 'when' [Preceding: 'Looking for her, they found her, when she ran
away.']
muku-uru-a-ne hana=te kuhjuri
[catch-PL-3ps/D=SUB when=FOC] [scream:3ps/E]
[DEP] [V]
'When they caught her, she screamed.'

Similarly, focus markers are found after participle clauses, as displayed in

(1015). In (1015a), the dependent clause is cross-referenced with the subject of
the second verb ('say'), as the embedded utterance is a quotation. Note that the
quotation itself also contains a focus marker, which is attached to the object
'you' and refers to the subject of the quotation ( T ) . (1015b) represents an
utterance from a conversation between children who are discussing the way the
tape recorder functions. The sentence is meant as an explanation by one of the
boys to the (not stated) question why the wheels are moving, to mean "It is
recording, (that's) what it does". Example (1015c) is particularly interesting,
because it illustrates that the focus marker occurs only after the third dependent
clause, all being participle clauses. Though unproven, it is plausible to assume
that the focus is on the entire entity of three clauses, which is in a way
"summarised" by the focus marker, to emphasise the relevance of these actions
('making hunting tools') for the result ('hunting').

(1015) Focused dependent clause with participle

a) [Preceding: 'She arrived safely at her daughter's place.']

turu-t=te, ii=na hoara-a u-ü, na-e
[arrive-PRT=FOC] [2sg=FOC: 1 sg] [see-NTR come-lsg/E] [say-3ps/E]
[DEP2] [Ol] [VI] [V2]
'Arriving, she said: "I have come to see you.'"

b) [Context: Children commenting on recording equipment]

graba-he-t=te aj
[record-CNT-PRT=FOC] [AUX:3ps/E]
[DEP] [V]
'It is recording.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Scope offocus 723

c) [More than one DEP] [Preceding: 'They only bought the poison from town.']
raahemane hit$ana ha-l, ne-batohwi te-l,
[self blowgun make-PRT] [3ps=bow make-PRT]
[DEP1] [DEP2]

naaohwa-l=te ke nesaritQäete-kure=lu hauria

[ENUM-PRT=FOC] [INST] [hunt-3pI/E=REM] [first]
[DEP3] [PP] [V] [ADV]
'They made blowguns and they made their arrows by themselves, and hunted with
them, at first.'

Another possible occurrence of a focus marker is in serial verb constructions. In

this context, it follows the first verb, which is marked with the neutral suffix.
However, it must be mentioned that occurrences of this kind are quite rare. The
examples in (1016a) is elicited, while (1016b) is taken from the NT. The
function of the focus marker here again is one of focus. Note that the indication
of [VI] and [V2] as separate units here does not reflect the fact that a SVC
represents a single clause (as argued in §17.5), but this notation is chosen here
to illustrate the fact that only the first verb is focused. A question that cannot be
answered is whether VI should actually be characterised as a separate clause,
but one might suggest that lenone ari-a ('seek food') represents in fact a single
pragmatic constituent.

(1016) Focus marking on first verb of SVC

a) (Elicited)
lenone ari-a=te ku-e
[food] [seek=FOC] [go-3ps/E]
[Ο] [VI] [V2]
'He went to seek food.'

b) [Preceding: 'Beware therefore, lest that come upon you, which is spoken of in the
prophets:']
mtahi-a=te ku-ri-ki-tge=i, ßäe ku ti-a=ne bana
[get.lost-NTR=FOC] [go-IRR-2ps-PL=ASS] [already there give-3ps/D=SUB when]
[VI] [V2] [DEP]
'You will perish, at the time when he gives it.' [NT: Acts 13: 41]

Regarding the example in (1016b) above, it should also be taken into account
that the function of V2 is temporal/aspectual, as the form o f ' g o ' describes a
future event that could also be expressed on VI itself.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
724 Focus

Instances of focus enclitics following the hearsay marker hetau have already
been observed in various examples (e.g. (985b), (998b), (1002c), (1010b)). In
contrast, particles such as the certainty marker tgu, the form hjä for 'just', ort
the adversative marker mki always follow the focus marker, when present,
which is illustrated in (1017) (also cf. §22.5).

(1017) Particles following the focus marker

a) With tqu:
aka=te tgu m-a tenjente gobernadoro=ne kujßa projekto itg,a-e=lu
[3sg=FOC CRTN] [be-3ps/D village.leader=SUB so.that] [project] [do-3ps/E]
[A] [DEP] [O] [V]
'He is the one who made a project so that there would be a village chief.'

b) With hjä:
turu-ni-u=he=te hjä ii hoara-htru-mn=t$a
[arrive-DSTL-IMP=REP=FOCj ust] [2sg see-PL-JUSS=EMF]
[V] [DEP]
'They say that you should just arrive so that they could see you.'

19.4 More than one constituent preceding the focus marker

In the previous sections it was demonstrated what types of constituents can be

followed by a focus marker. In all cases mentioned, only one constituent was
subject to focus or emphasis. In particular, it was observed that only one focus
marker per clause may occur. However, there are examples for clauses where
the focus marker is preceded by more than one constituent. In particular, this
regards combinations of NPs and adjuncts of different types. Altogether, three
different combinations of these are observed, including different orders. A
possible analysis is to interpret the two different constituents as sharing topic
and focus functions, which is plausible in some examples. However, it is
difficult to make generalisations from this, as will be demonstrated by other
examples shown here.
In (1018), an adverb is followed by a subject noun, which in turn is followed
by the focus marker. The interpretation within the context of (1018a), is
straightforward: hauria 'at first', refers to the topic of the clause. In fact, since
the sentence represents the initial sentence of the text, it is plausible that a topic
is introduced first, which is about 'the old times'. Since the adverb can regularly
occur in initial position, no marker is required. The NP 'the patrons', instead,
has been shifted to the front from clause-final position, for which a marker is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 More than one constituent preceding the focus marker 725

required. This corresponds well with the context, as the subject receives
emphasis, i.e. the listener is intended to focus on 'the patrons'.

(1018) Focus marker preceded by adverb + subject

a) [Context: Initial sentence of a narrative]
hauria fu>ator5-kuru=te ne-rehete
[first] [patron-PL=FOC] [be-HABl:3ps/E]
[ADV] [S] [V]
'At first, thepatrones used to be there.'

b) [Preceding: 'Earlier, the night did not exist.']

rdi rihihei hiitQU nehesißahei hetau=te ne-rehete floaelu
[that like] [day pure HRS=FOC] [be-HABl:3ps/E] [earlier]
[ADV] [CC] [V] [ADV]
'Like that, it was [always] daytime, earlier.'

The example in (1018b) is from a narrative that describes how the night was
created. It is the second sentence of the story, preceded only by the introductory
description of the situation (in square brackets). In this case, the division
between focus and topic is slightly less clear than in the other example, since
'like this' is less illustrative as a general theme for a statement. However, the
N P 'pure day' is in focus, being in contrast to 'night', mentioned in the
preceding sentence. Slightly different from (1018a), the NP is a copula subject.
A similar situation as in (1018a) is observed in (1019). Here, the conjunction
phrase for 'then' is in initial position and can be interpreted as topic. The
subject N P that follows, instead, is in focus and must be marked by =te as a
result of the frontshift.

(1019) Focus marker preceded by conjunction phrase + subject

[Preceding: 'He was in great pain from that snake bite.'] [Dog was mentioned earlier]
nii hana reemae hetau=te, rüa sim himitgal
[that when] [dog HRS=FOC] [side] [s!eep:3ps/E] [together]
[CNJ] [S] [LOC] [V] [ADV]
'Then the dog slept together with him at his side.'

The situation becomes more complicated in the following examples, where an

N P occurs in initial position, followed by an adjunct, to which a focus marker is
attached. The problem in a sentence such as (1020) is that the NP represents a
subject that was shifted to the front, but it is not directly marked with a focus
enclitic. Pragmatically, the subject NP in (1020a) is to be interpreted as focus
and topic. On the one hand, it is the theme of the clause (and had not been
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
726 Focus

mentioned before), on the other, it is in focus, as describing a group of people

who lives in contrast to the agricultural methods explained in the text. The
adverbial expression 'like that', which follows the NP and occurs before the
focus marker, has contrastive function, emphasising the (different) way these
people used to work, which are then described in detail in the four dependent
clauses that follow.

(1020) Focus marker preceded by subject + adverb

a) [Preceding: 'Until today there are people in other groups who make their fields like
that.']
ajriu kuane ne-nakauru nitoanei=te
[PLN inside be-those.who] [like.that=FOC]
[A] [ADV]

n^uhoana ha-ure-rehete=lu, hauria tauhja-1,

[3ps=field] [make-PL-HABl :3ps/E=REM] [first throw-PRT]
[Ο] [V] [DEP1]

laarurna hau-ι, kuane fioanara+kuturi temu-i,

[corner throw-PRT] [inside banana+head plant-PRT]
[DEP2] [DEP3]

laano+nuhua temu-i
[cassava+trunk plant-PRT]
[DEP4]
'The people of the Airico River used to make their fields like that, first throwing [the
felled trees] to the edges [of the field], [then] planting banana plants inside, [and]
planting cassava plants.'

b) [Preceding: 'This will be "tiatiae" [name of an unidentified animal] for that the
children would say forever, "I will go hunting when tiatiae sings", - this is how she put
it, speaking like this.']
adSa kanii eba ajjia nitoanet hetau=te ne=lu
[PSN GPF PSN with] [like.that HRS=FOC] [be:3ps/E=REM]
[S] [ADV] [V]
'This is how Adam and Eve lived.' [Lit. 'they lived like that.'] [End of story]

The example in (1020b) can be analysed in a similar way. Being the last
sentence of a story that describes one Urarina version of the flood, the subject
'Adam and Eve', is shifted to the front and functions as focus. Note that in this
narrative, the two characters play a crucial role and that the final sentence in a

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 More than one focus marker in a sentence 121

way goes back to this theme. However, the adverbial 'like that' could also be
understood as being in focus; at the same time, it comments on the way Adam
and Eve lived.
In most examples discussed here, the merging of focus and topic functions
results into ambiguities or interpretations that can only be retrieved from the
context or background information. Although the examples indicate that
fronting of several constituents can be interpreted as focus being put on one
constituent and topic on another, this is not an automatic rule. While this
interpretation neatly applies to examples (1018) and (1019), the sentences in
(1020) may be interpreted differently. As becomes evident from the examples in
(1018), focus and topic function on the first constituent can also be combined,
which, however, is only interpretable from the context.

19.5 More than one focus marker in a sentence

From the previous section it is apparent that each clause has only one focus
marker. As a tendency, the rule that only one focus marker can occur per clause,
is extended to the entire sentence, which in many cases contains more than one
clause (cf. §18.5). However, there are some exceptions to this, which will be
investigated in this section. It should be realised that only few examples exist
for each context given here. Therefore, one must take caution before
generalising these cases as stiff principles.
A construction that has already been discussed in §19.2.1 regards the
occurrence of focus markers after imperative clauses, by which a consecutive
function is expressed. In this case, a sentence can contain more than one focus
marker, as illustrated in (1021). It should be noted that the two occurrences of
=te in this sentence are in different clauses: The first instance is after
'finishing', whose subject is included in the preceding imperative (i.e. 2pl); the
second occurrence is after the lpl/in form of 'let us untie him', which is to be
interpreted as a hortative.

(1021) More than one focus marker with imperative constructions

[Preceding: 'They delivered him [Lomai's son] and told him [the Creator of Ways]
[what Lomai had said].']
ku>ajtei najßa-ri-Ί lenone-ku-toe, najpa-ri-l=te
[again finish-RAP-PRT] [eat-IMP-PL] [finish-RAP-PRT=FOC]
[DEP1] [VI] [DEP1]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
728 Focus

kwajtei hasasa-kaatge=te kwara-kaatce=ra

[again] [untie-lpl/in=FOC] [see-lpl/in=EMF]
[ADV2] [V2] [V3]
'Again finishing quickly, eat (PL) so that, finishing quickly, we could again untie him
and see [what happens].'

As described in §15.1.2, the hortative is expressed through a verbal suffix in

combination with the introducer huane. In most examples, a focus marker is
attached to this introducer in order to cross-reference the subject. The sentence
in (1022) (and a previous example in (981c)) illustrates this marking, but it also
exhibits an additional focus marker that is attached to the verb as to mark the
entire clause for focus. Strictly speaking, in this case, two focus markers occur
within the same clause, as it would be implausible to describe the introducer as
a separate clause.

(1022) Two focus markers with hortative constructions

[Preceding: 'Then, when the fish became fried [on his tongue], he swallowed it and
said:']
ku)ane=te i=tgej ate ti-aka-te ajpa kau-i
[let's=FOC] [2sg=for] [fish] [give-lpl/du=FOC] [with return-PRT]
[INTR] [PP] [Ο] [V] [DEP1]

i=simhera i=tguerehe-uru ajßa ke lenoni-a ku-i=ne kußa=ra

[2sg=spouse 2sg=child-PL with VLI eat-NTR go-2ps=SUB so.that-EMF]
[DEP2]
'Let's give you a fish so that you go home and go to eat it with your wife and your
children.'

However, this is the only context in which two focus markers would occur in
one and the same clause. All other occurrences where two focus markers are
observed involve different clauses, but even this is rather unusual. Most
common among these exceptional cases is the occurrence of two focus markers
within a sentence where a quotation is embedded. In fact, the occurrence of two
focus markers in this context is unproblematic, as it can be understood as
involving two independent sentences. This is illustrated in (1023a), which
describes a situation from a narrative, where Lomai wants to bathe and sends
the by standing men away. The first marker occurs in the quotation 'it is
dangerous' (an idiomatic expression, which literally means 'want danger'); the
other occurrence of =te is after the clause in which this utterance is embedded
('she said'). The sentence continues with the reply of one bystander who refuses
to move. His reply is not embedded in a clause with naa 'say', but the focus

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 More than one focus marker in a sentence 729

marker before his reply supports the switch of situation. Again, this is not an
obligatory occurrence of =te.

(1023) Two focus markers in sentences with embedded quotation

a) [Preceding: 'As one man stayed sitting there [as she wanted to bathe], she said:']
hjauißera, d^aha-l-tQa, na-1 naurutaae-ri-u=ra,
[no come.on-PRT=EMF say-PRT] [move-RAP-IMP=EMF]
[DEP1] [VI]

eluna heri-a=te nuhue=ra, i=ßuta=ra, na-a hetau-te

[danger want-3ps/A=FOC] [in.vain=EMF] [2ps=eye=EMF] [say-3ps/A HRS=FOC]
[V2] [ADV2] [PP2] [V3]

he, aj-ri-jia-i=ta, huatia i=tg.uara-ri-a-ü=m=ta

hey [AUX-IRR-NEG-3ps/A=FRS] [not] [2ps=see-IRR-NEG-lsg/A=ASS=FRS]
[V4] [INTR5] [V5]
'"No, come on", - saying so, "Move away a little, it is dangerous for your eyes (in
vain)", she said, [and he replied]: "Hey, I won't watch you.'"

b) [Preceding: 'Then the priest told his novice: "Tomorrow let the people go fishing for
me.'"]
na-a hau hetau=te eresi hetau=te batiri raj
say-3ps/D because HRS=FOC tomorrow HRS=FOC priest for

ate kuhwu-a ku-ri-ki-tQe=ra na-a hetau

fish fish. with .hook-NTR go-IRR-2ps-PL=EMF say-3ps/A Η RS

mi kauru raj mi raj mosaso

that 3 pi for that POSS novice
'As he said so, his novice said to them: "Tomorrow you will go fishing for the priest.'"

The sentence in (1023b) is another example for the occurrence of a focus in

quotation, while another focus marker occurs in the describing clause. The first
instance of =te follows the introducing dependent clause ('as he said so'), which
is followed by a quotation. This clause marks the adverb 'tomorrow' for focus.
In (1024), focus enclitics are found as marking two subordinate clauses
(DEP2, 4). It should also be noted that they refer to different subjects: In DEP2,
'kinkajou' is the subject, whereas in DEP4, it is the spider monkey, who is
trying to defend himself. However, this kind of marking is not a general strategy
for reference tracking, nor is it obligatory in this context (cf. §22.4). The
function of the focus marker in (1024) is not entirely clear, but one possibility is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
730 Focus

that the markers are employed in order to make the change of f o c u s more
transparent.

(1024) Focus marker on dependent clauses with different subjects

[Preceding: 'Looking at him, he asked: "Why have you stolen my bag?'"]

na-1 hetau mi alau töito-a=ne ham hetau=te
[say-PRT HRS] [that spider.monkey hit=3ps/D=SUB when HRS=FOC]
[DEP1] [DEP2]

ke taka-l mi obißa muku-ri-to-a

[VLI meet-PRT] [that weaving.stick catch-RAP-INTS-NTR
[DEP3] ([DEP4])

nuku-ja-ne hana hetau=te alau bihi

in.vain-3ps/D=SUB when HRS=FOC] [spider.monkey hand]
[DEP4] [O]

beuite mi unee obißa ke

[cut.off:3ps/E] [that kinkajou] [weaving.stick INST]
[V] [A] [PP]
'Saying so, when he [the kinkajou] hit the spider monkey, (when) he [the spider
monkey] met with it and caught the weaving stick [but] in vain, [and] the kinkajou cut
off a spider monkey's finger with the weaving stick.'

In e x a m p l e (1025), each of the two occurring interrogatives is marked with a

f o c u s enclitic. T h e expressions for ' h o w is it possible' and 'with w h a t d e w ' , can
both be considered as having f o c u s function. A possible analysis is to describe
them as an enumeration of questions, i.e. d^asn=te could be compared to a
separate clause (on the function of this interrogative, cf. §21.1.6).

(1025) Two focus markers on different interrogatives

[Preceding: 'Therefore, the envoy of God said:']

kauatQa-i=ta, d$asn=te d$a barasane ke~te
[good-NEG:3ps/A=FRS] [how.possibly=FOC] [what dew INST=FOC]
[VI] [ADV2] [PP2]

kauatQa-i omaro-t, ... rte-re=1 kanaanaj-uru raj

[good-PRT grow-PRT] [be-IRR:3ps/E=ASS] [child-PL for]
[DEP2] [V2] [PP2]
'It is not good, how is it possible, with what dew [from the night] should [the plants]
grow,... for [my] children?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 More than one focus marker in a sentence 731

Again referring to different subjects, two focus markers are observed in (1026).
The first marker, attached to 'therefore', relates to the general situation, as
described in the preceding sentence. The clitic =te after 'already', which
follows several dependent clauses later, can be interpreted as having focus
function. In this case, the marker is not directly related to the change of subject:
'As he taught them' (DEP1) is not marked with =te, but followed by two
participle clauses whose subject is expressed only on the final verb of the
sentence, and the second occurrence of =te is right before the main verb.

(1026) Other cases for double occurrence of the focus marker

[Preceding: 'So he taught them to bend the needles.']
mi häu hetau=te päe ku rela-naha-a
[that because HRS=FOC] [already there teach-PLO-3ps/D
[CNJ] [DEP1]

häu, titiiritQa-ahoa-x, läe ama-i, fcw hetau

because] [bend-DSTR-PRT] [CHNG take-PRT] [there HRS]
([DEP1]) [DEP2] [DEP3] [ADV]

β äe=te kuhwu-a ku-ure

[already=FOC] [fish.with.hook-NTR go-3pl/E]
[ADV] [V]
'Therefore, as he taught them already, they bent each of them [the needles], then just
took them along and went fishing already.'

In summary, the occurrence of more than one focus marker within a sentence is
rare, optional, and the examples stated here do not really suffice in order to
formulate criteria that legitimate this use. Therefore, the following list of
contexts in which a multiple occurrence was observed, can be understood as a
summary of attested contexts, but must not be taken as a catalogue of principles.
Several focus markers within a sentence are observed:
- in imperative constructions with consecutive function
- in hortative constructions with introducer
- in clauses embedded as quotations in a descriptive clause such as 'he said'
- in separate dependent clauses with different subject
- in chaining/enumeration of interrogatives.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
732 Focus

19.6 Summary

The marking of constituents with the enclitics =te, =ney and =na is a pervasive
feature of Urarina syntax and occurs with all types of constituents in positive
declarative clauses and in content questions. These markers are cross-referenced
to the subject marked on the verb of a clause and they require the use of the E-
form for person inflection. Their function ranges between focus and topic and
their interpretation considerably depends on the context. The insertion of a
focus marker is obligatory only with a frontshifted subject, except in negative
clause. A focus marker may occur after a constituent that is in its original
position, such as an object or an adverb. In this case, it tends to take focus
function.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
20 Multi-clause constructions

From previous examples given in this grammar, it is obvious that many Urarina
sentences contain more than one verb. Since Urarina has a variety of formal
means to express different types of such structures, I will refer to these as multi-
clause constructions in general.85 These include dependent clauses, which are
subordinate to an independent clause in a variety of ways, and the coordination
of independent clauses conjoined with each other. As different types of
dependent clauses have to be distinguished, I use the term "subordinate clauses"
to refer to the specific constructions as described in §20.1. These involve the
subordinate marker =ne, whose functions extend to other clause types as well.
$6
For instance, =ne is also used as a different subject complement clause marker.
This construction, however, differs from subordinate clauses as defined in §20.1
in that complement clauses only occur with certain control verbs, whereas
subordinate clauses are not limited to any specific verb type to occur in the
main clause. Such subordinate clauses can be divided into those with or without
conjunction.
Urarina has three main types of dependent clauses, each of which has
different syntactic properties. The first section f this chapter will analyse clauses
that contain the subordinate marker =ne. Some suborcHnate clauses, which are
marked by =ne only, have temporal or conditional function, as discussed in
§20.1.1. When a conjunction follows, the meaning of the clause is determined
by the semantic content of the conjunction (cf. §20.1.2). §20.1.3 and §20.1.4
discuss further, minor strategies of subordination.
Complement clauses with same subject reference are found with a limited
number of verbs; the marker for this construction is -na and is attached to the
verb root of the non-finite verb (see §20.2.1). Complement clauses with
different subjects take the subordinate marker =tie, which will be discussed in
§20.2.2.
§20.3 deals with participle clauses, which involve a non-finite verb marked
with the suffix -7. Subsequently, person inflection is marked on the verb in the
main clause. Participles have a variety of functions, which include sequence,
overlap, or adverbial modification of events. §20.4 briefly introduces two

85 ·
Serial verbs are discussed in §17, since they are monoclausal. However, as they
contain more than one verb, they are partly included in the discussion here.
86
The distinction between same or different subject clauses applies to S and A
arguments.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
734 Multi-clause constructions

further non-finite verbal suffixes which exhibit a behaviour similar to the

participle suffix.
Clause coordination represents a multi-verb construction of a different type,
as two or more independent clauses are conjoined, usually accompanied by
some formal marking. This includes the use of coordinating conjunctions or
adverbials, or can be expressed by juxtaposition, which, however, will be
marked prosodically (cf. §20.5).

20.1 Subordination

The clitic =ne has a range of different functions. On the one hand, it functions as
a subordinate marker for temporal, conditional and complement clauses (cf.
§20.1.1). On the other, it can be combined with a conjunction, which may
express temporal, causal, concessive, comparative, or consecutive functions in a
dependent clause (cf. §20.1.2). All subordinate clauses are also distinguished
from other clause types morphologically, as the dependent verb always appears
with the D-form (cf. §11.1). As shown in §18.5, the position of the dependent
clause can vary. A subordinate clause may precede or follow the independent
clause; however, its occurrence before the main clause is clearly preferred.
With respect to the types of inflectional marking on the verb, there are minor
differences between subordinate and main clauses: Of all 17 suffixal slots that
occur in the independent clause only one is not attested in subordinate clauses:
the irrealis suffix -ri (cf. §12.2.14). In elicitation, its use in dependent clauses is
possible when the assertive enclitic =m is attached to the verb to mark future.
In general, the use of enclitics in dependent clauses is restricted, as
summarised in (1027). Only the remoteness marker =lu is freely attested in
dependent clause; the reassurance enclitic =tau only occurs as a component of
the lexicalised hearsay marker hetau in dependent clause (cf. (1027a)). There
are a number of enclitics whose occurrence in dependent clause can be viewed
as "acceptable" in elicitation (cf. (1027c)), but they do not occur elsewhere in
the text database. The enclitics of slots 18 and 23 are strictly excluded from
occurring in dependent clause (cf. (1027b)).

(1027) Distribution of enclitics in dependent clause

a) Occurring in dependent clause:
- Remoteness marker =lu (slot 21)
- Reassurance marker =tau (slot 22; in hetau only)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 73 5

b) Not occurring:
- Politeness marker =t$e (slot 18)
- Interrogative =na (slot 23)
- Final negative =ne (slot 23)

c) In elicitation only:
- Assertive =m (slot 19)
- Witness evidential =ni (slot 20)
- Reportative =he (slot 21)
- Final emphatic =ra (slot 24)
- Frustrative =ta (slot 24)
- Warning =mare (slot 24)
- Fear marker =naate (slot 24)
- Rhetorical question =te (slot 24)

A further exclusion is the absence of the 'must' construction from dependent

clauses (cf. §5.7.4; §18.3). The presence or absence of certain suffixes in
complements clauses and with the participle form is stated in §20.2.1 and
§20.3.1.

20.1.1 Subordination without conjunction

When the enclitic =ne is attached to the end of a clause without being followed
by a conjunction, its function can be conditional or temporal. However, the
distinction between these functions must be derived from the context. It occurs
with simple temporal, habitual, or conditional meaning, as will be investigated
in the following subsections.

20.1.1.1 Temporal function of "=ne "

In (1028a), the interpretation of the subordinate clause 'when they saw ...'
(DEP2) is clearly temporal and does not have any conditional reading. The
temporal function in this example, however, may refer to a period of time rather
than a single moment. Slightly different in (1028b), a punctual action at a
specific time is described by 'when he threw the fish into the canoe' (DEP3).
The fact that the event is real and happened once is evident from the context. In
this example, the punctual meaning refers to a short moment. Temporal function
is also implied by (1028c), which represents the title of a narrative. The
interpretation of =ne here is less straightforward: on the one hand, a temporal
interpretation is plausible. On the other hand, =ne could also be viewed as a
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
736 Multi-clause constructions

complement marker (assuming ellipsis, as no complement verb is present); in

this case it may be glossed to mean '[I will tell] that a man went home
swimming' (also cf. §20.2.2.5). Since =ne has different functions, the exact
function cannot be determined in all cases; besides, multiple functions can be
combined.

(1028) Temporal function of =ne

a) Time period:
mtoam-a=ne baja, nehelau ne-1 n=uhoana ha-na
[be.like.that-3ps/D=SUB after] [other.side be-PRT 3ps=field make-NOM
[DEP1] [DEP2]

kwara-kuru-a=ne kuhwaelati-l kiuara-kure

see-PL-3ps/D=CND] [be.surprised-PRT] [see-3pl/E]
([DEP2]) [ADV] [V]
'After it was like that, when they saw a different way of cultivating their fields, they
looked at it surprised.'

b) Single moment:
ku ate muku-1, kaanii aresipe muku-uri-1,
[there fish catch-PRT] [sort.of mojara. fish catch-RAP-PRT]
[DEP1] [DEP2]

emnihja kuane hauto-a=ne ahariri ne-ι nerutu-e

[canoe inside throw-3ps/D=CND] [gamitana. fish] [be-PRT turn.into-3ps/E]
[DEP3] [CC] [V]
'He caught a fish there, he quickly caught this (sort of) mojara fish, and when he threw
it into the canoe, it turned into a gamitana fish.'

c) Overlap between temporal and complementising function:

akau kuane ftuitQU-i turu-a kau-a lejhit katQa=ne
[water inside swim-PRT] [arrive-NTR return-3ps/D] [one man=CND]
[DEP] [V] [S]
'When one man returned home swimming in the water' [Story title]

An extended temporal meaning of the subordinate marker applies in contexts

that suggest a habitual temporal use. Again, this is not formally marked and the
interpretation has to be derived from the situation itself, as illustrated in (1029).
In (1029a), it is only known from the context that the event occurred repeatedly
and regularly (namely, each time when the father came home). At no point in
the sentence is habitual aspect marked. The same applies to (1029b), which

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 111

describes a habitual action in the past, though morphologically unmarked for

aspect. The sentence is taken from a narrative about the life of the ancestors and
it is obvious that not one specific tree-felling event is described, but the general
practice of doing so.

(1029) Habitual temporal function of =ne

a) [Context: The woman continuously creates children made from cotton while the man
is away hunting]:
ra'z hau, ku m-a=w, turu-a u-a
[that because] [there be-3ps/D=CND] [arrive-NTR come-3ps/D
[CNJ] [DEP1] ([DEP2])

raj ßaka-ne, mi adaa=ne, baaba, baaba na-1,

POSS father=CND] that PSN=CND [Daddy Daddy say-PRT]
[DEP2] [DEP3]

ku kujßete-1, ku kaami kuruaae-kure

[there get.near-PRT] [there GPF be.happy-3ps/E]
[DEP4] [Main]
'Therefore, as he was there, when their father came home, that [guy] Adam, they said
"Daddy, Daddy", met him, and were ... happy.' [Implied: 'every day']

b) [Preceding: 'The life of the ancestors was like that:']

enua fwi-uru-a=iie, kauati-l itQU-hwäi, ke enua fioi-ure
[tree fell-PL-3ps/D=CND] [be.good-PRT tie-PRT] [INST tree fe!l-3pl/E]
[DEP1] [DEP2] [Main]
'When they felled a tree, they tied them [the axes] well and felled the tree with them.'

20.1.1.2 Conditional function of "=ne "

In contrast to temporal clauses, there are examples for the use of =ne (without
conjunction) that illustrate its conditional function. In (1030a), the event of
telling is a condition that has not been fulfilled, which excludes any temporal
interpretation. Its meaning in this context refers to the future. In (1030b), the
event described in the dependent clause is not in the future, but refers to past or
present ('if you tell/told the truth').

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
738 Multi-clause constructions

(1030) Conditional function of =ne

a) [Preceding: 'In the beginning, do not tell anything until you have taught ten women
(how to do it), he instructed her.']
be-i=j£=te, kamt nekuehe d$anu-ri-ki=i
[tell-2ps=CND=FOC] [lsg shame make.feel-IRR-2ps=ASS]
[DEP] [Main]
'If you tell, you will embarrass me.'

b) [Preceding: 'When he said that he has a wife from heaven, (the people) said:
"Really? Is this true?'"]
esijiae ajto-i-jie, d^aha-i ajane, raa-m-u=te hvara-ka
[truth say-2ps=CND] [come.on-PRT yes] [receive-DSTL-IMP=FOC] [see-lpl/in]
[DEP1] [DEP2] [Main] [DEP3]
'If you tell the truth, come on and get her, so that we can see her!'

Since the function of =ne as temporal or conditional depends on the context, the
interpretation is not always straightforward. In fact, many examples combine
these functions, as temporal and conditional reading may overlap. One is shown
in (1031), which leaves the decision open whether =ne is to be translated as 'if
or 'when'. On the one hand, =ne may describe a condition that still has to be
fulfilled, as the sentence describes a general rule. On the other hand, a habitual
temporal function is implied if the sentence is interpreted as past.

(1031) Overlap of functions of =ne

[Preceding: 'God wants that nobody remains without drinking ayahuasca.']

ipow ku-a lejhti itQauena=ne=te satii ku-uru-a nerajjie
[ayahuasca drink-3ps/D one elder=CND=FOC] [all drink-PL-3ps/D must]
[DEP] [Main]
'If/whenever an elder drinks/drank ayahuasca, all people must/had to drink it.'

20.1.1.3 Conditional function and irrealis

As seen above, in many cases the distinction between temporal and conditional
function of =ne is interpreted from the context. However, there is one factor that
helps to distinguish these functions. In some sentences that involve =ne, the
main clause may be marked for irrealis. In these cases, a temporal reading is
excluded and the clause involving =ne can be identified as having conditional
function. Though there are four logical combinations for the co-occurrence of
realis and irrealis forms (as stated below), two are excluded because the irrealis
form does not normally occur in dependent clauses. Type C is not relevant for
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 739

this discussion because it does not involve any irrealis forms - which leaves
only Type A.

(A) Main clause = irrealis, dependent clause = realis

(B) Main clause = irrealis, dependent clause = irrealis
(C) Main clause = realis, dependent clause = realis
(D) Main clause = realis, dependent clause = irrealis

While type A clearly is the common case for an involvement of the irrealis in
conditional clauses, there are no attested examples for type B. Type C basically
corresponds to the examples discussed in the previous §(20.1.1.2), where a
conditional function applies, without any involvement of the irrealis form. Type
D is a special case: In this construction, the irrealis is not morphologically
marked in the dependent clause; thus, strictly speaking, this type does not occur.
However, the insertion of a special conjunction in the dependent clause implies
irrealis meaning, as will be demonstrated in (1033).
The examples in (1032) illustrate case (A), where the condition expressed by
the dependent clause coincides with the realis form, whereas the result as
expressed by the main clause is marked for irrealis on the verb. Semantically,
both clauses have an irrealis meaning, since both events have not taken place
yet. In (1032a), the condition that Chunski arrives is not fulfilled, but irrealis is
not marked. Thus, two translations in English are possible, as indicated. In
(1032b), the assertive enclitic =m is added to the irrealis verb to mark future.

(1032) Irrealis in main clause, realis in dependent clause (Type A)

a)
H-a tgüki=ne=te, fioosu ha-re
[come-3ps/D PSN=CND=FOC] [pond make-IRR:3ps/E]
[DEP] [Main]
'If Chunski comes, he will make a pond.' / 'If Chunski came, he would make a pond.'

b)
eresi heelaj radio kuane be-re-ü-ni ere-naa
[tomorrow same radio inside telI-IRR-lsg/E=ASS] [speak-INF
[Main]
herit-tQäu=ne
want-1 sg/D=CND]
[DEP]
'Tomorrow I will tell over the same radio if I want to speak.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
740 Multi-clause constructions

The fourth logically possible combination (Type D) combines the realis form in
the main clause with the irrealis in the dependent clause. As mentioned, this
grouping is ungrammatical, as the irrealis does not occur in the dependent
clause. Instead, the irrealis mood in the dependent clause is expressed by the
conjunction hananiane ' i f . This conjunction indicates that the condition
expressed by the dependent clause is not fulfilled, but hypothetical. In other
words, the irrealis meaning of the clause is expressed by the conjunction; a
double marking of the irrealis (i.e. through a verbal suffix in addition) is not
possible. This is illustrated in (1033), where both clauses have realis marking
(i.e. no marking for mood). Even though in (1033a,b), the events expressed by
any of the clauses lie in the past, and all are imaginary/unreal, the hypothetical
status of the situation is expressed by the conjunction. In (1033 b), the introducer
baana, which always co-occurs with hananiane, further emphasises this
function.

(1033) Use of hananiane, realis form in the main clause; past reference

a)
kau m-anu=ne hananiane kusiti-anu
[here be-lsg/D=SUB if] [tie-lsg/A]
[DEP] [Main]
'If I had been here, I would have tied it.'

b)
baana najßatihi-anu hananiane kanu su-a hanolari
[if permit-lsg/D] [if lsg kill-3ps/A jaguar]
[DEP] [Main]
'If I had let it [attack me], the jaguar would have killed me.'

Unlike (1033), the example in (1034) illustrates a sentence in which the

hypothetical status does not lie in the past, but in the present. It is taken from a
narrative where two men decide to replace the poles of a house for a while. The
sentence describes the reaction of the poles as they hear of the men's intention.
While the situation would still have the potential to take a different turn, the
hypothetical status is expressed by the hananiane, in order to indicate that this is
a proposal nobody is likely to agree to.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 741

(1034) Use of hanamane, realis form in the main clause; reference to the present

eehe, kotateto-a t$u hjä, ßaara heri-Jd-toe=ne hanamane

yes [be.tiring-3ps/A CRTN just] [2pl want-2ps-PL=SUB if]
[VI] [DEP2]

kanakaanu reroari-i-tge, na-ure hetau mi anisihja-uru

[lpl/ex] [replace-2ps-PL] [say-3pl/E HRS] [that pole-PL]
[02] [V2] [V3] [S3]
"'Yes, it is tiring indeed. If you want, you replace us", said the poles.'

Alternatively, irrealis can be marked in the main clause when the subordinate
clause contains hanamane. In all examples, the dependent clause can be
interpreted as referring to the present. The main clause in (1035a), refers to the
present, whereas in (1035b), future reference is implied by the main clause. The
presence of hanamane in the dependent clause emphasises the relevance of the
condition for the context, as transferring the money is an action entirely
dependent on the person's permission to proceed in this manner. In (1035c), the
main clause has past reference. However, the information about tense is not
formally encoded and can be derived only through the context. The examples
demonstrate that the reference to past, present or future is not formally marked
and thus has no impact on the form of the verb with respect to mood.

(1035) Use of hanamane, irrealis form in the main clause

a) Main clause with reference to present
kana ke neherati-a kana+kwaaun-era=ne hanamane,
[lpl/in VLI get.angry-3ps/D our.creator=SUB if]
[DEP]

hvatia ka=rihjei ne-ri-akatQe

[not this=like be-IRR-lpl/in]
[Main]
'If God got angry against us, we would not be [live] like this.' [From sermon]

b) Main clause with reference to future

ii orden te-i hanamane, bäako kahe ruku-Ί,
[2sg order give-2ps if] [bank from pull.out-PRT]
[DEP1] [DEP2]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
742 Multi-clause constructions

leotQa bäako kuane tee-re

[other bank inside] [put-IRR:3ps/E]
[PP] [V]
'If you gave your order, he would withdraw [the money] from the bank and put it into
another bank [account].'

c) Main clause with reference to past

ne-m-a=ne hananiane kwatia be-ri-a-u
[be-NEG:3ps/D=CND IRR] [not tell-IRR-NEG-lsg/E]
[DEP] [Main]
'If it were not so, I would not have told you." [NT: John, 14:2]

The statement that the interpretation of different tenses in this construction is

largely based on contextual knowledge is also supported by the (elicited)
example in (1036), where an insertion of the remoteness marker in the
dependent clause is possible as an optional feature (in brackets), but not
required. It would be possible to have the remoteness marker on the main clause
only, or on both clauses. Naturally, as this is an elicited example, one has to be
careful to draw conclusions from it, but the sentence should suffice in order to
demonstrate that tense distinctions are not obligatorily marked. In principle, the
past reading would apply even without the remoteness marker in the main
clause, if the time reference is provided by the context.

(1036) Main clause with remoteness marker

kuriki ti-a(=lu) hananiane ßäe motosjera kureti-tQäu =lu

[money give-3ps/D(=REM) if] [already chainsaw buy-lsg/A=REM]
[DEP] [Main]
'If he had given money, I would already have bought a chainsaw.'

Regarding the variety of functions of =ne in various combinations of clauses, a

summary might help to understand the way temporal and conditional clauses are
to be interpreted.
1. A temporal reading of the subordinate clause is always possible if both
clauses are in the realis form and the conjunction hananiane is not involved
(cf. (1028), (1029), and (1031)).
2. The interpretation as punctual or habitual in these cases is derived from the
context (cf. (1029)).
3. Given the appropriate context, a subordinate clause in a sentence with
possible temporal interpretation may also be interpreted as referring to a
condition (cf. (1030)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 743

4. If the main clause is marked for irrealis (cf. (1032)), or if the subordinate
clause contains hananiane (cf. (1033)-(1035)), a conditional reading is
obligatory.
5. If the subordinate clause contains hananiane, the time reference depends on
the context: The sentence may refer to present (cf. (1035a)), past (cf.
(1035c)), or future (cf. (1035b)).

These findings are further summarised in table (1037). It is evident that any
indication of irrealis in either clause, being through morphological marking on
the verb or through hanamane, results in a conditional interpretation of the
dependent clause. When no marking is present, a temporal (including habitual)
or conditional reading is possible.

(1037) Conditions for realis/irrealis interpretation

Main clause Dependent clause Interpretation
realis realis temporal, habitual,
conditional
realis realis + hanamane conditional
irrealis realis conditional
irrealis realis + hanamane conditional

20.1.2 Subordination with conjunction

The construction with hanamane discussed in the previous section is only one
example for subordinate clauses that involve a conjunction. Urarina has a
variety of other conjunctions, which were introduced in §5.8. The functions that
are encoded by these conjunctions can be summarised in the following way:
- Related to time
- Related to cause
- Related to effect
- Related to counter-expectation
- Related to comparison.

The position of conjunctions is always at the end of the dependent clause, i.e.
they are preceded by all other constituents. Also recall that most conjunctions
may also function as postpositions (cf. §5.8). This section points out differences
and similarities between the various conjunctions as occurring in subordinate
clauses.
It should be noted that the use of the subordinate marker =ne is optional
before conjunction. When it is present, it may also be realised as the variant

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
744 Multi-clause constructions

=na. This variation is based on individual preferences, similar to the variation

observed with the focus markers for lps (cf. §19).

20.1.2.1 Temporal function

Subordinate clauses with a time-related function represent the majority of all

conjunctions (also cf. §5.8.1). In terms of frequency, hana 'when' is the most
common temporal conjunction, followed by baja 'after' and barn 'at the time
when'. There are only few examples attested for the other conjunctions listed
below.

(1038) Time-related conjunctions

hana 'when' (cf. (1039))

bana 'at the time when' (cf. (1040))
tonoana 'while' (cf. (1041a))
najnia 'while', 'whereas' (cf. (1041b))
baja 'after' (cf. (1042a))
ahinia 'before' (cf. (1042b))
reetiai 'until' (cf. (1042c))
kureu 'after', 'in place o f (cf. (1042d))

The conjunctions ham and bana, which may both be translated into English as
'when', only slightly differ from each other in meaning. While hana describes
the punctual time of an event, bana is used to refer to a period of time.
However, the difference between the two is only relative, as becomes clear by
the examples in (1039). In (1039a), the precise time of the event 'when they
were going' cannot be indicated, but it is implied that the event described in the
main clause occurred at one specific moment. The use of hana in example
(1039b) is again punctual, since it is obvious that the first time of entering the
creek occurred at one specific moment.

(1039) hana in subordinate clause

a)
ku-he-uru-a hana=te muku-e akauru raj Unterem fiooko
[go-CNT-PL-3ps/D when=FOC] [burn-3ps/E 3pl POSS flashlight lamp]
[DEP] [Main]
'When they were going, the lamp of their flashlight went out.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 745

b)
haurianehei kaa ajtQune kuane enoa-akaanu=ne hana,
[at.first this Espejo inside enter-lpl/ex=SUB when]
[DEP]

m-ji=lu ne-naa-naha
[be-NEG:3ps/A=REM be-NOM-PURP]
[Main]
'When we entered the Espejo for the first time, there was no place to live.'

The use of bana, as shown in (1040a) differs from that of hana, as "the time
when I come" does not specify any particular moment. In the given context,
hana refers to the period of time the speaker will spend in town in order to
regulates his issues. In the same way, no specific moment is indicated in
(1040b), but 'at the time of our creation' refers to a period that is not defined in
terms of a punctual event, but represents the final statement of a narrative,
whose general time frame is briefly pointed out. Note that bana, as well as hana
may be used with any temporal reference, i.e. a dependent clause with these
conjunction may have present, past, or future reading.

(1040) bana in subordinate clause

a)
ßäe u-anu~ne bana=ne a\pa hua-re-ü=ni kanu
[already come-lsg/D=SUB when=FOC:lsg] [with see-IRR-lsg/E=ASS lsg]
[DEP] [Main]
'At the time when I come, I will see with him [what to do about the money].'

b)
jioaelu kana ne-hoaau-naa bana
earlier lpl/in ITR-create-NOM when
'Earlier, at the time of our creation' [End of a narrative]

Another "pair" of conjunctions is tonoana and najma, which both mean 'while'.
As suggested in §5.8.1, tonoana is used to refer actions or events that occur
parallelly on the timeline without implying contrast, najma, instead, is used
with contrastive function as expressed by 'whereas'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
746 Multi-clause constructions

(1041) Subordinate clauses with the meaning 'while'

a) tow ana:
hajti rautohwe-i m-aka=ne tonoana,
[still be.calm-PRT be-lpl/du=SUB while]
[DEP]

kana+kwaaun-era ajtöo-hwa ari-aka nerajtene

[Our.creator say-NOM seek-lpl/du must]
[Main]
'While we still live in peace, we must seek what God says.' [From sermon]

b) najnia:
mtoanei bi-a ku-uru-a eene-kuru=ne najnia hetau=te,...
[like.that tell-NTR go-PL-3ps/D woman-PL=SUB while HRS=FOC]
[DEP1]

nese-u ku-1 hitarii mi-tQa hoara-kur-i bi-a ku-ure

[town-LOC go-PRT] [all that-only see-PL-NOM 0b j tell-NTR go-3pl/E]
[DEP2] [Main]
'Whereas the women were going to tell it like that, they [the men], going to town, went
to tell everything they had seen.' [NT: MT 28:11]

The functions of all other time-related conjunctions are unlikely to be confused.

Examples for these are illustrated in (1042). The sentence in (1042d) exhibits
the occurrence of kureu 'after'/'in place o f , which is exceptionally rare in
frequency. It also exhibits an unusual phonological feature, as it is the only
Urarina lexeme containing the diphthong /e«/. (1042d) is taken from a narrative
where the people of a settlement express their wish to kill a slave who is a
cannibal raised by one man. Finally, the boy's owner is persuaded and instructs
the others to kill his slave during his absence.

(1042) Other time-related conjunctions in subordinate clauses

a) baja 'after':
ku uhoana ha-ι setu-ahoa-a enua=ne baja,
[there field make-PRT] [rot-DSTR-3ps/D tree=SUB after]
[DEP1] [DEP2]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 747

kau ßäe nenenahati-a u-akaanu=lu

[here already settle-NTR come-lpl/ex=REM]
[Main]
'We made our fields there, [but] after the wood rotted, we came to settle down here.'

b) ahirtia 'before':
najße-ri-a, mtahe-si-a mi kuriki ahima
[be.able-IRR-3ps/A] [get.lost-CPL-3ps/D that money before]
[Main] [DEP]
'He could [transfer the money] - before that money gets lost (completely).'

c) reetiaai 'until':
hauria hitarii letpuyka-j katQa rela-i-jie reetiat,
[first all ten-NOMsbj man teach-2ps=SUB until]
[DEP]

mhjauria be-i=jie
[don't tell-2ps=NEGF]
[Main]
'Until you have taught all ten people, do not tell it!'

d) kureu 'after'/' in place o f :

ajane, ka=nuta edaj-u itQa-ku-toe-ra,
yes [ 1 sg=eye outside-LOC do-lMP-PL=EMF]
[Mainl]

amu-anu=ne kureu hja lae itQa-ku-toe=ra

[walk-lsg/D=SUB "shortly.after" just] [CHNG do-IMP-PL=EMF]
[DEP2] [Main2]
ΌΚ, do it behind my back; do it while/after I have walked away!'

20.1.2.2 Causal function

A conjunction related to cause is häu 'as/because 1 and it can be considered to

be the most frequently used conjunction. This is mostly due to the fact that its
function does not always directly refer to cause in a strict sense. In narratives, it
is used in order to relate a sentence with the preceding one (clause chaining):
the last clause of the first sentence is repeated as a dependent clause in the
following one and marked with the conjunction häu . While this construction is
further discussed in §22.4.5, the present section focuses on the causal function
of this conjunction, which is well illustrated by the examples in (1043). In
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
748 Multi-clause constructions

(1043a), the dependent clause with hau describes the reason for the fact stated
in the main clause. This sentence functions as a summary for the entire
preceding text, which summits in an explanation as for why the kinkajou carries
its creatures with the mouth. In (1043b), the subordinate clause again represents
the direct cause for the effect described in the main clause: The only reason to
choose the howler monkey to be chief o f the monkeys is that it lives in a single
tree rather than moving about.

(1043) hau with causal function

a)
nii hau hetau=te nitoanei bua basihjau-a alau=na
[that because HRS=FOC] [like.that bag steal-3ps/D spider.monkey=SUB
[CNJ] [DEP1]

hau, nekajrÜQahe-ϊ beree ama-e unee nuhuae kuane-tga

because] [suffer-PRT] [child take-3ps/E kinkajou mouth inside-only]
[DEP2] [Main]
'Therefore, because the spider monkey stole his [kinkajou's] bag, the kinkajou has to
carry its children with the mouth.'

b)
nitoanei hetau=te enejtgu kuraanaa ne-ι ruru
[like.that HRS=FOC] [monkey chief be-PRT howler.monkey
[DEP1] [Main]

te-ure=lu, helaja-atga nesuru-1 ni-a=na hau

give-3pl/E=REM] [same.place-only relax-PRT be-3ps/D=SUB because]
[DEP2]
'That is how they made the howler monkey chief of the monkeys, because he stays
relaxing in one place.'

Manus (1992) also mention the word käu as a conjunction related to cause and
give the example as stated in (1044a):

(1044) kau with causal function

a) (Cf. Manus (1992:34)

su-a enua-ne käu =te mhjau-e
[kill-3ps/D tree=SUB because.of=FOC] [die-3ps/E]
[DEP] [Main]
'He died as a result of a tree killing him.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 749

b)
enua käu =te mhjau-e
tree because.of=FOC die-3ps/E
'He died because of a tree.'

M y consultants confirmed the sentence in (1044a) as correct. However, käu is

not attested in active use as a conjunction in my database, where its function is
restricted to postpositional use after nouns (cf. §5.7.4). In this context it is
employed with the meaning of 'because o f , indicating the cause. The sentence
in (1043a) would then be given as (1043b).

20.1.2.3 Concessive function

The conjunction sajhjel 'although' (or variant asajhje~i) is used with subordinate
clauses to mark concessive function. Sentence (1045a) is an example for this,
taken f r o m a narrative where a woman warns her husband not to watch her
while she creates children through magic. In the course of events, the man
cannot resist and subsequently, is scolded by his wife, as shown in the example,
which highlights the fact that he disobeys even though she had warned him.
(1045b) again illustrates a case of refusal, where a man insists on taking home a
cannibal boy, who is to be his servant.

(1045) Concessive dependent clause with sajhjel

a)
i=tgej ajto-kwäu -ne sajhjel, d$anu=te k=ere towra-ene-i=ta
[2sg=for say-1 sg/D=SUB although] [why=FOC lsg=word obey-NEG-2ps=FRS]
[DEP] [Main]
'Why did you not obey even though I told you to?'

b)
lot ku)unu-he-uru-a=ne sajhjel ku d^una-l ama-a
[there beg-CNT-PL-3ps/D=SUB although] [there refxise-PRT] [take.along-3ps/A]
[DEP1] [DEP2] [Main]
'Even though they were pleading there [that he leave the boy], he refused [to comply]
and took him along.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
750 Multi-clause constructions

20.1.2.4 Comparative function

Urarina has two conjunctions related to comparison of equality. Unlike other

conjunctions, both are lexicalised verbal derivations, which is demonstrated by
the fact that they co-occur with the subordinate marker. Morphologically, these
can be analysed as participle forms of verbs (cf. §5.4.6).
- rihihei 'like', 'in the same way as' (from rihia 'be like')
- tokuanei Mike', 'compared to' (from tokuania 'compare')

The conjunction rihihei (or variant rihei) is derived from the verb for 'be like'
and actually represents the continuous form of the participle. While other forms
of the verb rihia rarely occur in texts or conversations, the conjunction rihihei is
quite common. The subordinate clause in (1046a) is a direct comparison to 'the
way the peopled lived earlier'. In (1046b), a man promises his father-in-law to
seek revenge after the older man was nearly killed by a jaguar. The conjunction
is used to indicate that the revenge shall occur 'in the same way' as the jaguar
did. The function of rihihei is not strictly limited to the comparison of
situations, but may also imply an action that is to be taken as a result of the
event it is compared to (also see (1047)).

(1046) Comparison with rihihei

a) [Preceding: 'Then, those who live now cannot do such things because they do not
know them']
floaeltt aj-uru-a katQa-uru=ne rihihei itQa-ure-ni-a~na
[earlier AUX-PL-3ps/D man-PL=SUB like] [live-PL-NEG-3ps/D=SUB
[DEP1] [DEP2]

hau
because]
'... because they do not live in the same way as the people lived earlier.'

b) [Context: He sees his father-in-law bleeding all over his body from the wounds
afflicted by a jaguar]:
mtoanel i=fioitoa kurenia i=fu-a=ne rihihei
[like.that 2sg=trap for 2sg-beat-3ps/D=SUB like]
[DEP]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 751

kanu kwaa-rü=tge=ra.
[lsg see-DSTL.HORT=PLT=EMF]
[Main]
'Like he [the jaguar] has beaten you for your trap, I shall see [for revenge].' [From
Airico dialect]

The comparative function of rihihei does not exactly apply in (1047), where a
translation as 'due to the fact that' would be more appropriate. It is difficult to
account for the fact that not a causal conjunction such as hau is used in this
example. The comparative factor in this case could be characterised in a way
that "compares" the two situations following each other: "In the same way as
(or: given the fact that) they are your family, talk to them".

(1047) Other functions of rihihei

ii d$=arai ne-1 ne-uru-a=ne rihihei, akauru raj
[2sg 2sg=family be-PRT be-PL-3ps/D=SUB like] [3pl for
[DEP]
ere-m-u=ra
speak-DSTL-IMP=EMF]
[Main]
'As they are your family, go talk to them!'

For the other conjunction that has the meaning of 'like', 'in the same way as',
only few example are attested in the database, but numerous occurrences are
found in the NT. Going back to its lexical root, one could also render its
meaning as 'compared to' (as glossed here). This meaning applies well to the
example in (1048), where today's situation is compared to earlier times.

(1048) Comparison with tokuanel

jioaelu ena aj-a=ne tokuanel itQafioa ni-ji=lu
[earlier now AUX-3ps/D=SUB compared.to] [rifle be-NEG:3ps/A=REM]
[DEP] [Main]
'Earlier, there were no guns as compared to (what is) now.'

The position of the subordinate marker is fixed, as it occurs after the final
constituent of the dependent clause. In some cases, this will be after the verb
(including all verbal clitics), in others, after the subject, if expressed by an overt
NP. When a subordinate clause is in focus, the subordinate clitic precedes the
focus marker. When a conjunction is present, the focus markers occurs at the
end of the clause. If the hearsay particle hetau is part of the subordinate clause,
it follows the subordinate marker. Note that the subordinate marker is not

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
752 Multi-clause constructions

compatible with the remoteness enclitic =lu. The sentence in (1049)

exemplifies the occurrence of the focus marker after conjunction.

(1049) Subordinate clause and focus marker

anaito-a hau =te bajhja-i ama-e kaa raj ßaka

[be.heavy-3ps/D because=FOC] [carry-PRT] [take.along-3ps/E this POSS father]
[DEP1] [DEP2] [Main]
'Because it is heavy, ("this") his father takes it along, carrying it.'

20.1.2.5 Consecutive/purposive function

There are two conjunctions related to effect or purpose, positive hijßa 'so that'
and negative hoataa 'so that not'. The use of these implies intention as well as
effect, which makes it plausible to characterise them as having combined
consecutive and purposive function. As mentioned in §18.5, subordinate clauses
with these conjunctions preferably follow the main clause, whereas all other
types tend to precede it. The subordinate clause of sentence (1050a) describes
the intended result of an action, as it was Lomai's plan to turn the people into
water people and it was necessary to drown them for this purpose. Example
(1050b) is from a conversation over shortwave radio, where speaker A advises
the listener to take a motocarro (a motorised tricycle) on its way home. While
speaker Β assumes that this would be because of the distance, speaker A makes
it clear that his main concern are safety reasons, as the listener might get robbed
on the bus. Also note that the verb in a subordinate clause with huataa always
takes the affirmative form, as negation is expressed by the conjunction only
(also cf. §14.4.3). Though both examples given in (1050) refer to different
subjects in main and dependent clause, the two conjunctions can also be used in
dependent clause with a same subject interpretation.

(1050) Consecutive function of kujfla and hoataa

a)
mtoanei hetau=te katg,a lemu-e=lu lomaj,
[like.that HRS=FOC] [man sink-3ps/E=REM PSN]
[DEP1] [Main]

edara ne-~i ku-uru-a=ne kujpa

[water.people be-PRT go-PL-3ps/D=SUB so.that]
[DEP2]
'Like that Lomai sank the people so that they would become water people.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 753

b) [Preceding: (Speaker A): 'Be careful when you go home and take a motocarroV -
(Speaker B): 'Yes, I will, because it is far.']
kolektivo htane kau-a ku-i=jx kiuataa=na ajto-o
[public.bus inside return-NTR go-2ps=SUB so.that.not=FOC:lsg] [say-lsg/E]
[DEP] [Main]
Ί said it so that you would not go by public bus.'

Also note that the form haanu combines complementation with purposive
function. Even though haanu has some properties of a conjunction, its use
extends beyond that. For instance, it can also occur after a neutral form of the
verb, as is investigated in §20.2.4.1.

20.1.2.6 Innovative variants of subordination with conjunction

In the innovative language, the use of hau (lit. 'because'; cf. §20.1.2.2) in
combination with postpositions is a common feature. This construction may be
related to the function of hau as a kind of complementer used by younger
speakers (which is discussed in §20.2.4.3). Specifically, the following
postpositions can be adjoined to hau :
- ... häu raj ('for')
- ... häu ajßa ('with')
- ...häu rihihel ('like')

The sentences in (1051) from the innovative language illustrate their use. As
becomes evident, the combination of häu with the different postpositions has
different effects: With raj (cf. (1051a)), the actual meaning of häu as 'because'
is entirely retained, and possibly emphasised by the postposition ("for the
reason that ..."). In (1051b)), one could interpret the presence of ajßa in DEP1
as optional, as reason is implied ('because you want it'). In (1051c)), only the
meaning of rihihel is relevant whereas the presence of häu is obsolete and must
be regarded as optional.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
754 Multi-clause constructions

(1051) Addition of postpositions to subordinate clause

a) With raj:
motokaro kuanaj=ßa kau-a ku-re-ü=m
[rickshaw inside=FOC=lsg] [return-NTR go-IRR-lsg/E=ASS]
[PP] [V]

kuriae ni-a hau raj

[distance be-3ps/D because for]
[DEP]
Ί will return by motor rickshaw, for it is far.'

b) With ajßa\
ii=te here-1 ne-i hau ajßa kau turu-akaanu hau raj
[you=FOC want-PRT be-2ps because with] [here arrive-1 ρ 1/ex because for]
[DEP 1 ] [DEP2]

naria mi kurenia
[thanks] [that for]
[IDM] [PP]
'Thank you ("for") that we have come here ("with") because you want it.' [From
prayer]

c) With rihiheT.
ßäe taum ajto-hväu hau rihihei nihjauria nekajritQa-i=jie
[already earlier say-lsg/D because like] [don't] [suffer-2ps=NEGF]
[DEP1] [INTR] [V]

kuriki siiri-i=jie
[money have-2ps=CND]
[DEP2]
'As I already said earlier, don't suffer when you have money!'

20.1.3 Non-finite verb plus conjunction

In the innovative language, it is possible to use a verb stem without person

inflection when these are followed by the conjunctions ham or bam. In most
cases, the verb stem will be identical to the root, as is illustrated in (1052). In
(1052a), the verb for 'arrive' occurs in its root form before bam and in (1052b),
the verb for 'go up' is unmarked for person. In either case, the uninflected verb
refers to third person. Reference to any other person of the paradigm is not

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 755

possible. Though the subordinate marker is not present, the dependent clause
has subordinate function.

(1052) Verb root plus conjunction

a)
turu bana=na kwara-a ku-u
[arrive when=FOC:lsg] [see-NTR go-lsg/E]
[DEP] [V]
'When he arrived, I went to see.'

b)
enua keim-uru hana=te ausirijie raj komasaj
[tree VLI
go.up-PL when=FOC] [be.pregnant:3ps/E] [POSS wife]
[DEP] [V] [S]
'When they climbed up the tree, his wife was pregnant.'

As observed in (1052b), a dependent verb in this construction may contain the

plural suffix -uru in addition to the root, but no person suffix (also cf. DEP4 in
(1053a)). In a similar way, examples are also attested for roots followed by the
continuous aspect suffix -ahe (cf. (1053b)). Due to the low frequency of this
construction in the database, it is unclear whether other suffixes can follow the
root in this context. The function of this construction is identical to other
subordinate clauses with these conjunctions, but it is only observed among
younger speakers - who use both constructions parallelly.

(1053) Verb stem plus conjunction

a) With plural
mi hau =te amu-na najße-kur-em-a=ne hau,
[that because=FOC] [walk-INF be.able-PL-NEG-3ps/D=SUB because]
[CNJ] [DEP1]

heräe raj bihi hwue+hiuueka-l, enua ari-t

[slowly POSS hand RED+feel-PRT] [tree seek-PRT]
[DEP2] [DEP3]

ke im-uru hana torn kuruataha-1 hanulari

[VLI go.up-PL when] [sound:3ps/E] [two-PRT jaguar]
[DEP4] [V] [S]
'Therefore, as they were not able to walk [because of the darkness], they slowly felt
with their hands, seeking a tree and when they climbed up, [the voices of] two jaguars
sounded.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
756 Multi-clause constructions

b) With continuous aspect marker

ku-he-uru hana=te muku-e akauru linterono fivoko
[go-CNT-PL when=FOC] [burn-3ps/E] [3pl flashlight light.bulb]
[DEP] [V] [S]
'When they were going, the light of their flashlight went out.'

As is obvious from the sentences in (1052) and (1053), these have different
subjects for main clause and dependent clause, which is a compulsory condition
for this construction. The (elicited) example in (1054) shows that this
interpretation also holds when both clauses refer to 3ps and lack an overt
subject. A reading with same subject interpretation is not possible here.

(1054) Non-finite verb and conjunction with same subject constructions

turu hana=te ka=raj raana te-re=i

[arrive when=FOC] [lsg=for] [w.l.peccary] [give-IRR:3ps/E=ASS]
[DEP] [Ε] [Ο] [V]
'When he (A) arrives, he (B) will give me the white-lipped peccary.'

Note that the same interpretation is also possible with the inflected form of the
verb followed by subordinate marker and hana. Interestingly, hana conveys a
connotation for a different subject interpretation. Though this is not a strict rule,
it is observed in the majority of cases, as will be further analysed in §22.4.5.
This could explain the obligatory different subject interpretation in the
construction with the non-finite verb. For bana, not enough examples are
available in order to generate a similar hypothesis.

20.1.4 Alternative strategies of subordination

In §15.3, the consecutive function of the imperative with certain constructions

was mentioned. More exactly, if an imperative is followed by another clause,
the action described in this clause is interpreted as a result of the command
given in the imperative clause. As a special feature, the verb of the second
clause appears in the hortative or jussive form or contains a strategy to express a
result or a purpose (cf. (1055a)). This construction can be interpreted as a kind
of subordination of the second clause: while it can be formally described as a
juxtaposed clause, it is semantically and syntactically dependent from the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Subordination 757

imperative clause. A subordinate clause as the one in (1055b) cannot stand by

itself. 87

(1055) Subordinate function of consecutive construction with imperative

a) DEP contains lsg hortative form:

ka=jane-u-tge=na hauria hia-m=tg,a
[ 1 sg=let-IMP=FOC: 1 sg] [first urinate-DSTL.HORT=EMF]
[Main] [DEP]
'Let me go so that I can go to urinate first!'

b) DEP is a subordinate clause with kujjia

huuraka-u kanakaarm-netonaj eno-akaanu=ne kufpa=ra
[open-IMP] [lpl/ex-also enter-lpl/ex=SUB so.that]
[Main] [DEP]
'Open so that we would enter as well!'

A similar syntactic dependency is observed with the privative form -elanaala,

which is attached to the verb stem (cf. (1056); §20.4; §12.4for the kinds of
suffixes that can occur between the root and the privative). With this form,
person or number inflection cannot be marked on the dependent verb and are
expressed only by the verb forms in the main clause.

(1056) Syntactic dependency of the privative form

mi rihihei hjane-na here-lanaala a\pa ku räasa-ure

[that like let-INF want-PRV] [with there dance-3pl/E]
[DEP] [Main]
'They danced with her like that without wanting to let her go.'

As has become evident in this section, Urarina has a variety of constructions

and strategies to express subordination. The prototypical case is with the
subordinate enclitic =ne, which can occur with or without conjunction. Without
conjunction, its functions range from conditional to temporal meaning, partly
depending on accompanying elements and on the context. Possibly being based
on the fact that = m can be omitted in subordinate clauses that contain a
conjunction, the innovative language has introduced an additional type of
subordination, which combines a non-finite form of the verb with the temporal
conjunctions hana and barn. In this case, different subject interpretation is
obligatory. The use of an imperative in the main clause often has a consecutive

87
Only the insertion of the hortative introducer kwane would give this clause an
independent status.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
758 Multi-clause constructions

or purposive effect on a dependent clause, which can be expressed through a

subordinate clause with consecutive conjunction, through a non-finite purposive
form, or through a hortative or jussive form.

20.2 Complement clauses

Urarina has two basic types of complement clauses, both of which occupy the
slot of an Ο argument. One involves the suffix -na, which is used for same
subject constructions (also called infinitive here, cf. §20.2.1), the other one is
with the subordinate marker =ne, which was already discussed in §20.1. Some
verbs in combination with =ne can refer to same subject or different subject,
depending on the person inflection in the dependent clause (cf. §20.2.2). Both
complementation types require the presence of a complement-taking verb (also
called control verb here) out of a limited set of verbs. While not all verbs that
are used as control verbs with =ne can also have control over an infinitive
clause, the two partially overlap.
As a rule, any complement clause with =ne or -na precedes the main clause.
Alternative complementation strategies, as discussed in §20.2.4, include several
related constructions with the purposive form haanu (§20.2.4.1), quotative
complementation with naa (§20.2.4.2), and complementation with hau , which
only occurs in the innovative language (cf. §20.2.4.3). A summarising overview
concludes this section s illustrated by the table in (1095).

20.2.1 Same subject complement clauses

A same subject (S/A) or infinitive clause is formed by a verbal stem to which

the infinitive suffix -na is added. The verb of the infinitive clause can be from
any transitivity class (cf. §9). Some grammatical categories can be marked on
the verb before -na is attached. These are listed in (1057). Note that these
largely coincide with the suffixes that also occur before other non-finite suffixes
of the verb, as was demonstrated in §12.4. For instance, the same set of suffixes
may also precede the participle suffix -7 (cf. Table (763) in §12.4).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 759

(1057) Suffixes occurring before infinitive marker -na

CAU1: -a (slot 1)
CAU2: -erate (slot 2)
CNT -ahe (slot 4)
PLO: -naha (slot 7)
DSTR -akwa (slot 7)
RAP -uri (slot 8)

Most typically, however, the suffix -na is attached directly to the verb root. It
has a lengthened variant -naa, whose occurrence depends on a complex range of
factors, mainly involving word length. The exact conditions under which this
rule is applied are not entirely clear. The allomorphy leads to one complication
in the interpretation of the infinitive clause, as a similar alternation also occurs
with the nominalising suffix -naa, which can be shortened under equally
obscure circumstances (for an overview of the functions of this nominaliser, cf.
§6.6.3). In (1058), both suffixes have a short vowel and thus cannot be
distinguished by this feature. As nearly all control verbs for the infinitive are
transitive, the question may arise whether the suffixes for infinitive and
nominalisation could be in polysemy. The examples in (1058) illustrate the
conflict. As the two morphemes appear to be homophonous, one would wonder
whether a sentence such as (1058a) could be interpreted as Ί want (the) work' -
thus, an instance of nominalisation rather than complementation. The example
in (1058b), instead, shows the suffix -na as a nominaliser, here forming the
noun 'sick person', which functions as the Ο argument of the verb.

(1058) Infinitive or nominalisation?

a)
amiane-na heri-tgäu
work-INF want-lsg/A
Ί want to work.'

b)
kunajte-na hoara-kau
be.sick-NOM see-lsg/A
Ί saw the sick person.'

However, while the suffixes appear to be identical at first sight, there are
significant differences between the two. More exactly, the ambiguity can be
resolved in two different ways. Firstly, there are examples in which
nominalisation of the dependent verb would not make sense from a semantic
point of view, as is illustrated in (1059), where -na is attached to the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
760 Multi-clause constructions

enumerative verb naaohwaa.

(1059) Suffix -na attached to enumerative verb

ukvana ha-ι, arusu temu-t, naaohwa-na ißa-ur-em-a

field make-PRT rice plant-PRT ENUM-INF know-PL-NEG-3ps/A
'They not know [how to] make fields, and [how to] plant rice.'

Secondly and even more significantly, the two suffixes are indeed not
homophonous, as the transcription may misleadingly suggest. A difference
between the two morphemes is on the tonal level. While the entire complexity
of this difference has not been completely explored, one aspect can be
highlighted: the infinitive suffix is toneless and realised with a Low tone in
most cases, whereas the nominaliser carries a High tone. This is illustrated in
(1060): In the infinitive construction (1060a), the Η tone is on the first syllable
of the main verb (and the complement clause is toneless), whereas in (1060b), a
Η tone appears on the suffix -naa - which clearly identifies the nominalisation.

(1060) Tonal differences between nominaliser and infinitive (elicited)

a) INF:
ere-naa heri-tQäu
speak-INF want-1 sg/A
Ί want to speak.'

b) NOM:
ere-naa heri-tQäu
speak-INF want-1 sg/A
Ί want the speaker.'

However, while the above example may illustrate the difference, the situation is
more complex than this, as tonal shifts may occur with the infinitive suffix, i.e.
with some verbs, a Η tone may fall on the infinitive suffix or on another syllable
within the control verb, as illustrated with kunajtena in (1061)). 88 In examples
where this is the case, an additional alternation takes place: The suffix is
consistently realised short as -na for the infinitive, but it occurs as the
lengthened alternant -naa for nominalisations (cf. (1061b)). One could view this
as a dissimilatory rule which is employed to avoid ambiguities.

88
The regularities for this have not been identified.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 761

(1061) Dissimilation between -m and -naa

a) INF:
kunajte-na heri-tQäu
be.sick-INF want-lsg/A
Ί want to be sick.'

b) NOM:
kunajte-naa heri-tQäu
be.sick-INF want-lsg/A
Ί want the sick one.'

While some questions may remain open regarding the details of this alternation,
it becomes obvious that the two morphemes are not two polysemic realisations
of one underlying morpheme, but represent different grammatical categories
and so are to be considered as homophones. This does not exclude the
possibility that they are diachronically related - which would be plausible.
Generally, there are no restrictions for the compatibility of same subject
control verbs with any inflectional categories. Limitations of compatibility may
involve regular, morpheme-specific exclusions, such as the fact that the CAU1
causativiser is only compatible with intransitive verbs (cf. §16.4.2) - but most
complement-taking verbs are transitive. In fact, the only intransitive infinitive
control verb is najma 'be able' (cf. (1065)), whereas all others are transitive.
The verbs that can take a complement clause with -na include the following
types:
- Verbs of liking: heria 'want', hiurahia 'desire', rautaa 'like' (cf. (1062))
- Verbs of thinking or knowing: huaauka 'think', ijiaa 'know' (cf. (1063)-
(1064))
- Verb of ability: najma 'be able' (cf. (1065))
- Verbs of beginning/ending: nunuetia 'begin', najßaa 'finish' (cf. (1066))
- Verb kvaasia 'be afraid' (cf. (1067))

A few verbs that occur as infinitive control verbs in elicitation or NT examples

only are mentioned in §20.2.1.6.

20.2.1.1 Verbs of liking

Verbs of liking are typical complement-taking verbs and frequently occur in the
database. These include the verbs 'want', 'desire', and 'like'. Examples for each
of these are given in (1062).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
762 Multi-clause constructions

(1062) Verbs of wish with infinitive clause

a) heria 'want':
ku kunuhe-uru-a hau, ßäe atn be-na
[there insist-PL-3ps/D because] [already nevertheless tell-INF]
[DEP1] [DEP2]

here-ne
[want-NEG: 3 ps/E]
[V2]
'As they insisted, she still did not want to tell.'

b) kuurahia 'desire':
nakaauhwa-na kuurahi-anu
bathe-INF desire- lsg/A
Ί have the desire to bathe.'

c) rautaa 'like':
mi asajhjel=te here-lanaala, helaj ne-naa rauta-a=ne hau,
[that despite=FOC want-PRV] [separate be-INF] [like-3ps/D=SUB] [because]
[DEP1] [DEP2] [V2] [CNJ2]

helaj ne-naa loanari asae sini raj beree-kuru ajpa

[separate] [be-NOM shed under] [sleep:3ps/E] [POSS child-PL with]
[ADV4] [PP4] [V4] [PP4]
'Despite that, without wanting [to listen], as she liked to stay by herself, she slept alone
in a shed that was there, with her children.'

20.2.1.2 Verbs of thinking or knowing

There are two verbs of thinking and knowing that can take a same subject
complement clause. The verb hoaauka refers to 'thinking' and can also imply
'remember' or 'be acquainted with'.

(1063) Infinitive clause followed by hoaauka 'think'

a)
maohioa-ur-i=ßa itQa-na hoaauku-ü
[ENUM-PL-NOMSbj FOC: 1 sg] [do-INF] [think-1 sg/E]
=

[[OL] [V1]]DEP [V2]

Ί am thinking of doing all this'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 763

b)
t<^äe=na Uausura itga-na kivaauku-ü trertta wviembre ke
[also=FOC:lsg] [closure do-INF] [think-lsg/E] [30.November VLI]
[CNJ] [DEP] [V] [PP]
Ί am thinking of closing the school on 30. November.'

The verb ipaa implies "knowledge" on a more cognitive-intellectual level,

which includes the intellectual ability of how to do things. Note that another
verb of knowing, hvituhta, does not function as a control verb for infinitive
clauses, but only occurs with different subject complement clauses (cf.
(1073a)).

(1064) Infinitive clause followed by ißaa 'know'

a)
n-itga-naa itga-na ipa-h-ene aka komasaj
[3ps-do-NOM do-INF] [know-CNT-NEG:3ps/E] [3sg wife]
[DEP] [V] [A]
'His wife does not know how to do her work.' [Lit. 'her doings']

b)
hja-naa ißa-ure, bahi-i naaohwa-na ißa-ur-eri-a
[urinate-INF] [know-3pl/E] [defecate-PRT] [ENUM-INF] [know-PL-NEG-3ps/A]
[DEP 1 ] [VI] [DEP2] [DEP3] [V3]
'They did not know [how] to urinate, [how] to defecate, all this.'

20.2.1.3 Verbs of ability

There is only one verb of ability, najrtia, which also has a high frequency in the
database. Note that this is the only intransitive control verb for same subject
complement clauses. Its meaning does not strictly refer to ability only (cf.
(1065a)), but may also imply possibility, as illustrated in (1065b).

(1065) Infinitive clause followed by najnia 'be able'

a)
nii hetau=te, rüjej aj-ßa najni-si-ni-a=na häu,
[that HRS=FOC] [not.at.all AUX-INF be.able-CPL-NEG-3ps/D=SUB because]
[A] [DEP1]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
764 Multi-clause constructions

ßäe he hiniitgai nukue-ka=ne hau,

[already VLI together be.dark-3ps/D=SUB because]
[DEP2]

ukiasi he ini-ι sini

[renaco.tTQQ VLI go.up-PRT] [sleep:3ps/E]
[DEP3] [V]
'That [man], as he could not do anything at all, and as it became dark at the same time,
climbed a renaco tree and slept [there].'

b)
enene ßäe mtoanei ne-naa najni-ji
[nowadays] [already like.that be-INF] [be.able-NEG:3ps/A]
[ADV] [DEP] [V]
'Nowadays, this is not possible any more.' [Lit.: 'It cannot be like that any more'
(because they have forgotten the old customs).]

20.2.1.4 Verbs of beginning/ending

The transitive verbs for 'begin' and 'end'/'finish' can also be used to take an
infinitive clause as a complement. The verb najßaa 'finish', interestingly, is
related to najrtia 'be able' in that it represents the CAU1 form of this verb.89

(1066) Infinitive clause followed by verbs of beginning and ending

a)
itga-na najßa-a
[do-INF] [finish-3ps/A]
[DEP] [V]
'He finished doing it.'

b) nunuetia 'begin'
sittra ha-naa nunueti-a
[bag make-INF] [begin:3ps/E]
[DEP] [V]
'She has begun to make the bags.'

89
Note that najßäi, the participle form of najßaa, is lexicalised to mean 'nicely',
'thoroughly'.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 765

20.2.1.5 Verb "kwaasia " 'be afraid'

While the verb hvaasia 'be afraid' can take an infinitive clause, there is another
transitive verb with a similar meaning, which does not function as a control
verb: huaaetoa 'be scared o f can be used as a transitive verb only, i.e. it takes
an Ο argument rather than a complement clause.
In (1067), the function of Jauaasia is illustrated. Note that in (1067b), the
main clause is a prohibitive clause in which the complement clause is
embedded. Alternatively, both examples given here could also be expressed
with the different subject complementiser =ne - as it may refer to either same or
different subject of the dependent clause (cf. §20.2.2.3).

(1067) Infinitive clause followed by kiuaasia 'be afraid'

a)
nala-na kwaasi-a
[drop-INF] [be.afraid-3ps/A]
[DEP] [V]
'He is afraid of dropping it.'

b)
kwa kanu uhvana u-naa hoaasi-a
[don't] [lsg field come-INF] [be.afraid-NTR]
[INTR] [DEP] [V]
'Don't be afraid to come to my field.'

20.2.1.6 Other verbs

There are a few verbs for which it is doubtful whether these regularly function
as infinitive control verbs, as there are no attested examples in the text database,
nor in the NT. However, as these may occur in elicitation, I list them here,
leaving it open whether they are control verbs or not.
One such case is the verb transitive hiama 'leave'/'stop', which otherwise
can function as a control verb with different subject constructions (cf. (1082a)),
but it is attested with infinitives in elicitation only (cf. (1068a)). Verbs of
comparison are another type of verb that may occur with control function over
same subject complement clauses - but are only attested in elicitation.
Examples are given for tohiania 'compare' (cf. (1068b)), rihitoa 'be like' (cf.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
766 Multi-clause constructions

(1068c)), and amunia 'be as i f (cf. (1068d)). 9 0 The verb bajtia 'forget' (cf.
(1068e)) is found with this function in elicitation and in a number o f N T
examples. Note that it is also attested as a different subject control verb (cf.
(1073b)).

(1068) Infinitive control verbs in elicitation or NT examples

a) hiania 'leave':
hihwu-na hiani-a-ü
[try-INF] [leave-NEG-lsg/A]
[DEP] [V]
Ί didn't stop trying.'

b) tokuania 'compare':
kwara-na tokuarn-anu
[see-INF] [compare-1 sg/A]
[DEP] [V]
'as if I had seen it' [Lit.: Ί compared to seeing it']

c) rihitoa 'be like':

suru-naa rihito-a
[run-INF] [be.like-3ps/A]
[DEP] [V]
'He was like running away.'

d) amunej 'be as if
na-a hau hetau=te mi bajahirii jiäe suru-naa amunej.
[say-3ps/D because HRS=FOC] [that soon.after already run-INF] [as.if]
[DEP 1 ] [DEP2 [V2]
'As she spoke (like that), soon after that he seemed to run away.'

e) bajtia 'forget':
hjaaöria kauatga itQa-na bajte-i,... naaohwa-t ne-i-tge=ne
[don't] [good do-INF] [forget-PRT] [ENUM-PRT] [be-2ps-PL=SUB]
[INTR] [[DEP1] [V1]]DEP2 [DEP3] [DEP4]

90
While the example for amunej is not actually elicited, the properties of this verb are
rather unclear. In the few natural examples, it always occurs in this form (amunej)
and thus seems to be a defective verb. A citation form or a form inflected for person
could only be observed in elicitation.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 767

hiane-h-tQe=ne
[leave-2ps-PL=NEGF]
[V4]

'Don't (PL) forget to do good and don't (PL) give up ... all this!' [NT: Hebrews 13:16]

20.2.2 Different subject complement clauses

As mentioned in §20.1, =ne does not only function as a temporal and
conditional marker and as a subordinate marker; it also introduces a
complement clause ('that'). It differs from the infinitive construction in various
ways, as will be outlined in this section.
When the subject of a dependent clause is not controlled by the verb of the
independent clause (i.e. when the infinitive suffix -m cannot be used), =ne is
employed as a complementiser. This is best illustrated by a comparison of two
contrasting sentences (also cf. §17):
(1069) Complementation with same and different subject compared

a) Same subject
u-na heri-to-ann
[come-INF] [want-INTS-1 sg/A]
[DEP] [V]
Ί (really) want to come.'

b) Different subject
turu-a=ne heri-ji
[arrive-3ps=SUB] [want-NEG:3ps]
[DEP] [V]
'He (A) does not want him (B) to come.[Lit. 'arrive']

While the use of -na in (1069a) marks same subject constructions, different
subject complement clauses are formed with the subordinate marker =ne. Thus,
the interpretation of (1069b) as different subject is obligatory. In both cases, the
complement clause precedes the main clause. However, recall that =ne also has
temporal and conditional function - and that clauses with these functions may
also follow the main verb (cf. (1032b)). As a consequence, isolated examples in
which a dependent clause with =ne follows the main clause are interpreted as
temporal or conditional sentences rather than complement clauses. The example
in (1070) is thus to be interpreted as T h e howler monkey could, if he were the
chief.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
768 Multi-clause constructions

(1070) Dependent clause with =ne following the main clause

ruru=te nafpe-re kuraanaa ni-a=ne=ra

[howler.monkey=FOC] [be.able-IRR:3ps/E][chief be-3ps/D=CND=EMF]
[S] [V] [DEP]
'The howler monkey could be chief.'

Some verbs that function as control verbs for same subject constructions (with
-na) also occur as complement-taking verbs in different subject constructions
(with =ne). But while there is a certain overlap of these verbs in both
constructions, there are some verbs that function as control verbs with =ne, but
not with -na - and vice versa. For examples, some verbs of liking do not occur
with different subject constructions. In contrast, verbs of perception, such as
'hear' and 'see', and verbs of saying such as 'tell' are common with different
subject complement clauses but do not occur in same subject constructions. The
following verbs occur as control verbs over different subject complement
clauses:
- Verbs of liking: heria 'want' (cf. (1071))
- Verbs of thinking or knowing: kwaauka 'think', esenetaa 'believe' (cf.
(1072)), kuituhta 'know', bajtia 'forget' (cf. (1073))
- Verb hoaasia 'be afraid' (cf. (1074))
- Verbs of perception: aunaa 'hear', kuaraa 'see' (cf. (1075))
- Verbs of saying: bia 'tell/announce' (cf. (1076))

In addition, the construction for 'must' has control over a complement clause
(cf. §20.2.2.6). Similarly to same subject control verbs, there also are a few
different subject control verbs which are only attested in elicitation or in rare
NT examples, which are listed in §20.2.2.7. Note that I have included those
verbs that regularly and frequently occur in the NT in §20.2.2.2 to §20.2.2.5.

20.2.2.1 Verbs of liking

Different from infinitive clauses, there is only one verb of liking that functions
as a control verb with different subject complement clauses. While the use of
heria 'want' in this function is common, the verbs kuurahia 'desire' and
rautaa' like' are not attested in the same context.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 769

(1071) Different subject complement clause with heria 'want'

a)
repete-käu =ne here-i=jia
[repeat-lsg/D=CND] [want-2ps=INT]
[DEP] [V]
'Do you want me to repeat it?'

b)
ita ajßa kauatQa-t ni-aka=ne here
[REC with good-PRT be-lpl/du=SUB] [want:3ps/E]
[DEP] [V]
'He wants us to be nice with each other.'

20.2.2.2 Verbs of thinking and knowing

T h e control function of the verb hoaauka 'think' with infinitives has already
been described in the previous section. T h e example in (1072a) illustrates its
use in complement clauses with a different subject. In the same way, the verb
esenetaa 'believe' can be employed, while it is not observed in same subject
constructions (cf. (1072b)).

(1072) Different subject complement clause with verbs of thinking

a) kwaauka 'think':
asia b=atane ke nekajritQa-akaanu=lu=ne huaauk-u-tge
[PLN ASCM-land VLI suffer- lpl/ex=REM=SUB] [think-IMP-PL]
[DEP] [V]
'Remember that we have suffered in Asia!' [NT: Acts 20:7]

b) esenetaa 'believe':
nii ajßa kuruataham-a kana+hoaaun-era=ne eseneta-uru-a hau
[that with help-3ps/D our.creator=SUB] [believe-PL-3ps/D] [because]
[DEP] [V] [CNJ]
'because they believe that Gods helps them with that'

T h e verb hvitukua ' k n o w ' differs from ipaa ' k n o w ' (cf. (1064)) both
semantically and functionally. While ißaa implies knowledge in terms of
ability or ways of how to accomplish things, the meaning of kuitukua involves
knowing in a manner of 'being familiar with', 'having come to k n o w ' . Only the
latter can be used as a different subject complementising verb (cf. (1073a)),
whereas the former is only found with infinitive clauses. Another verb related to

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
770 Multi-clause constructions

thinking or knowing is bajtia 'forget', which may also be used with infinitives
(cf. (1068e)). Both verbs have in common that they do not occur in the text
database with this function, but are very frequent as complement-taking verbs in
the NT.

(1073) Different subject complement clause with verbs of knowing

a) kmitukua 'know':
na~i jnaelu ßaara rela-uru-a=lu=ne hoituku-i-tge
[say-PRT] [earlier 2pl teach-PL-3ps/D=REM=SUB] [know-2ps-PL]
[DEP1] [DEP2] [V2]
'You (PL) know that they taught you earlier, saying so.' [NT: Matthew 5:27]

b) bajtia 'forget':
kuruatahane-ki-t(x=lH=ne bajte-ne-re=t kana+kwaaun-era
[help-2ps-PL=REM=SUB] [forget-NEG-IRR:3ps/E=ASS] [our.creator]
[DEP] [V] [A]
'God will not forget that you have helped [them].' [NT: Hebrews 6:10]

20.2.2.3 Verb "kwaasia" 'be afraid'

The verb huaasia 'be afraid' can be used with same subject as well as with
different subject complement clauses. Note that ki'aasia is also used as a
transitive verb with an NP as Ο argument otherwise. The sentences in (1074)
exemplify its use after the complementiser =ne. Different from other control
verbs, constructions with hoaasia plus =ne may also refer to a same subject
complement clause, as is illustrated in (1074b) (also mentioned in §20.2.1.5). In
this example, the variant with =ne is in free variation with the infinitive
construction.

(1074) Different subject complement clause with verb kwaasia 'be afraid'

a)
hoatia kana kaseteka-a=ne hoaasi-atga-a nijej
[don't] [lpl/in punish-3ps/D=SUB] [be.afraid-only-NTR] [not.at.all]
[INTR] [DEP] [V] [ADV]
'Do not be afraid at all that he would punish us.' [NT: Romans 13:5]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 771

b)
raj kwitgam ra-uru-a=ne kwaasi-t$uru-a=na
[POSS blood receive-PL-3ps/D=SUB] [be.afraid-PL-3ps/A=INT]
[DEP] [V]

kau ne-j=te
[here be-NOMSbj=RHT]
[A]
'Are those who are here afraid of taking its blood?'

20.2.2.4 Verbs of perception

Verbs of perception do not occur with infinitive clauses, but they can be used
with different subject complement constructions in the Ο argument slot. The
t w o verbs attested in this function are hoaraa ' s e e ' and aunaa ' h e a r ' , (which
may also imply ' f e e l ' ) .

(1075) Different subject complement clause with verbs of perception

a) aunaa 'hear':
wawaho-a raana=ne auna-käu
[grunt-3ps/D w.l.peccary=SUB] [hear-lsg/A]
[DEP] [V]
Ί heard the white-lipped peccaries grunt'

b) hoar a a 'see':
ßäe nahauri-a u-a=ne hoara-1 hetau...
[already dock-NTR come-3ps/D=SUB] [see-PRT HRS]
[[DEP1] [V1]]DEP2

nit raj koseera ne-ι ne-naa muku-a ku-a

[that POSS servant be-PRT be-NOM] [catch-NTR go-3ps/A]
[02] [V2]
'As he saw that they were docking, ... he went to catch the one who was to be his
servant.'

20.2.2.5 Verbs of saying

The verb bia 'tell/announce' is the only transitive verb of saying, whereas all
others are intransitive (cf. §9.5). While bia is not attested as a complement-

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
772 Multi-clause constructions

taking verb in the text database, its use with this function is very common in the
N T and in elicitation, where it occurs as a control verb for different subject
complement clauses.

(1076) Different subject complement clause with bia 'tell'

hvaauka-u=ne be-u-tQe akauru raj

[think- lsg/D=SUB] [tell-IMP-PL] [3pl for]
[DEP] [V] [PP]
'Tell them that I remember [them].' [NT: Romans 16:3]

Remarkably, bia can also be used as a control verb without the complementiser
=ne, as illustrated in the elicited examples (1077) and (1078). This is in so far
surprising as the use of =ne is obligatory with all other control verbs. There is
one parallel to the (optional) omission of =ne as a subordinate marker before
conjunction, which is very common (cf. §20.1.2). However, in that case, the
presence of a conjunction marks the clause as dependent whereas in the
examples below, there is no overt marking of dependency for the complement
clause.

(1077) Absence of ne with bia 'tell'

a)
sini-uru-a be
[sleep-PL-3ps/D] [tell:3ps/E]
[DEP] [V]
'He told (that) they were sleeping.'

b)
obana kivara-a bi-a
[cl.peccary see-3ps/D] [tell-3ps/A]
[DEP] [V]
'He told that he has seen a collared peccary.'

Interestingly, the omission of =ne can also imply same subject meaning as
shown in (1078). Note, however, that the infinitive suffix -na cannot be inserted
in this example. Thus, from a syntactical point of view, the construction in
(1078) is identical to a serial verb. However, (1077a) shows that the
construction is not an SVC because plural is marked on Verb 1 and different
subject meaning is implied - which implies that there are two clauses. Also note
that the presence of =ne in (1078) would automatically imply a different subject
interpretation, i.e. 'that he is looking for peccaries'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 773

(1078) Omission of =ne before bia with same subject interpretation

raana ari-a be-ure
[w.l.peccary seek-NTR] [tell-3pl/E]
[DEP] [V]
'They announced that they are looking for white-lipped peccaries.'

In example (1079), the presence of =ne and the fact that dependent clause and
main clause refer to the same subject, exclude the interpretation of =ne as a
complementiser. Instead, the dependent clause is interpreted either as a
temporal or as a conditional clause. The reason for this is that =ne with
complementising function always refers to a different subject in the
complement clause.

(1079) Function of =ne with same subject dependent clause

ßäe ere-na heri-tQäu =ne be-ri-tgau=ni
[already] [speak-INF] [want-lsg/D=SUB] [tell-IRR-lsg/A=ASS]
[ADV2] [[DEP1] [V1]]DEP2 [V2]
Ί will already tell [them] if I want to talk.' [over the radio] / Ί will already tell [them]
when I want to talk.'

Another verb of saying that can take a different subject complement clause
(though only attested in elicitation and in the NT) is bahaa 'ask', as illustrated
in (1080).

(1080) Complementation with bahaa 'ask'

komasaj hiane-na najni-a katQa=ne baha-ure hetau
[wife leave-INF] [be.able-3ps/D] [man=SUB] [ask-3pl/E HRS]
[[DEP1] [VI] [S2]]DEP2 [V2]
'Is it possible that a man leaves his wife, they asked.' [NT: Mark 10:2]

20.2.2.6 'Must' construction

In §18.3 it was suggested that netene, though lexicalised, is based on a

combination of subordinate marker, followed by a focus enclitic and 'be'.
Alternative forms with the same meaning are nerajfle and nerajtene. Therefore,
if the initial [ne] in these forms really is based on the subordinate marker, the
whole construction may be interpreted as a control verb (literally 'it must (be)
that ...'). The examples in (1081) indeed demonstrate that netene is used in this
function: Though there is no morphological evidence that the dependent clause

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
774 Multi-clause constructions

is inflected with the D-form (which is identical to the Α-form in all examples
here), the sentences are structurally identical to other complement constructions.
In (1081a), the complement clause is marked for lpl/du and in (1081b) for 3sg.

(1081) Complementation with 'must'

a)
haurianehei enua ruku-aka netene
[first] [wood pull.out-lpl/du] [must]
[ADV] [DEP] [V]
'At first, we have to pull out wood.'

b)
m=hit$ana ti-a nerajjie
[3ps=blowgun give-3ps/D] [must]
[DEP] [V]
'They had to make their blowguns.'

It should be added that the 'must' construction differs in so far from other
complement constructions as 'must' is not a regular verb, but a (possibly recent)
construct. The form netene (or variants) only occurs as an impersonal form with
3ps reference. Person and number are marked on the dependent verb, i.e. in the
complement clause. If the second Iriel in this construction really is a form of the
copula (namely 3ps/E), it cannot be changed into any other person, which
implies that it is lexicalised.

20.2.2.7 Other verbs

Similar to some cases observed with the infinitive construction, there are
examples for verbs that only occur as different subject control verbs in
elicitation or in the NT translation. One example for this is hisenia 'dream',
which is only attested in elicitation (cf. (1082a)). The transitive verb hiama
'leave' has already been mentioned in the previous section. It regularly occurs
in the NT as a control verb with different subject constructions (cf. (1082b)).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 775

(1082) Different subject control verbs in elicitation or NT translation

a) kusenia 'dream':
itQa-a=ne kusem-tQäu
[do-3ps/D=SUB] [dream-lsg/A]
[DEP] [V]
Ί dreamed that he did it.'

b) hiartia 'leave':
mi=te ßäe nii ere-naa eseneta-uru-a=ne hiane-1,
[that=FOC] [already that speak-NOM believe-PL-3ps/D=SUB] [let-PRT]
[A] [[DEP1] [V1]]DEP2

leotpa-uru rela-uru-a=m hiane-t ne-ure

[other-PL teach-PL-3ps/D=SUB] [let-PRT] [be-3pl/E]
[[DEP3] [V3]]DEP4 [V4]
'Those have already given up believing in that word and given up teaching others.' [NT:
MT 13:22]

As seen in (1082b), hiania may also occur with same subject reference even
though it takes the different subject complement marker =ne. It should be noted
that this is also observed with N T examples that contain verbs of beginning or
ending.

20.2.2.8 Ambiguities in interpretation

Due to the multifunctionality of =ne, which also functions as a general

subordinate marker before conjunctions and which may have temporal or
conditional function, ambiguities may occur. This becomes particularly obvious
in the example with huaraa in (1083). Here, =ne could be interpreted as a
complementiser or as a subordinate marker with temporal function, as indicated
by the translation.

(1083) Ambiguous function of =ne with huaraa

bute itga-a noe=ne hvara-he-uru-a=lu kat$a-uru

[boat make-3ps/D Noah=SUB] [see-CNT-PL-3ps/A=REM] [man-PL]
[DEP] [V] [A]
'The people were watching "that" Noah made a boat.' / 'The people were watching
when Noah made a boat.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
776 Multi-clause constructions

When the dependent clause follows the main clause, the ambiguity is dissolved,
because in this case, temporal or conditional interpretation is preferred.

20.2.3 Special control verbs

There is a special form that is mentioned here even though it may not exactly be
characterised as a type of complementation, but which happens to occur
exclusively with such verbs that normally function as complement-taking verbs.
This form, which is marked by the suffix -erte (or allomorphs) on the verb can
be understood as a kind of nominalised irrealis form - or the irrealis form of a
nominalisation (IRN). There are three examples occurring in the database,
where this form is observed with the verb netoeka 'be in the mood'. 91 As
illustrated in (1084), the dependent clause (if it can be described as such) is
marked with -em or allomorphs to imply an unfulfilled action. In (1084a), the
irrealis form on 'follow' expresses the fact that the action of following is
unaccomplished and unreal. The dependent clause with hau in (1084b) is to be
understood as an innovative type of complementation or nominalisation 'what
he said' (discussed in §20.2.4.3). Here, it functions as an object of 'do', which
in turn is marked with the IRN suffix. In example (1084c), the presence of -ene
attached to 'throw away' again implies that the action is unaccomplished.

(1084) Irrealis nominaliser -ene with netooeka

a)
hat ät saku-ene netoe-7a-ü
[very] [follow-IRN] [be.in.the.mood-NEG-lsg/A]
[ADV] [DEP] [V]
Ί was very much not in the mood to follow [God].' [Personal experience narrative]

b)
nijej ajto-a=ne häu itQa-nene netoe-?a-u
[not.at.all] [say-3ps/D=SUB because] [do-IRN] [be.in.the.mood-NEG-lsg/A]
[ADV] [[DEP1] [V1]] DEP2 [V2]
Ί was absolutely not in the mood to do what he said.' [Personal experience narrative]

91
Note that this verb only occurs as a complement-taking verb in one elicited example
and is therefore not listed in the previous sections. It probably is not a typical control
verb, but is commonly found with the form described in this section.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 777

c)
d^anuna tQaiti rtii baaso hau-ene netohwe-m-aka
[why ever] [that bad throw-IRN] [be.in.the.mood-NEG-lpl/du]
[INT] [DEP] [V]
'For what ever reason are we not in the mood to throw away those bad things ...? [From
sermon]

The idea that the form described above is an instance of nominalisation is based
on examples that are attested with a variety of other control verbs occurring in a
similar context. While all attested examples in the database occur with netoeka,
other control verbs that incur -ene in the complement clause (in elicitation)
appear to require a sort of subject agreement on the dependent verb, as is
illustrated in (1085). Note that these can occur in the negative or affirmative
form (examples (1085a,b) as opposed to (1085c,d)). For netoeka, no affirmative
examples are attested with the IRN form.

(1085) Nominalisation of -ene with other control verbs

a) Negative with 'know':

n=itea-nene ißa-uru-i
[lsg=do-IRN] [know-PL-NEG:3ps/A]
[DEP] [V]
'They do/did not know what to do.'

b) Negative with 'think':

k=itg,a-nene hvaau-a-ü
[lsg=do-IRN] [think-NEG-1 sg/A]
[DEP] [V]
Ί do/did not think of doing it.' [e.g. if somebody told me to do it but I can't because I
forgot it or I was sick]

c) Affirmative with 'begin':

n=itQa-nene nunueti-a
[ 1 sg=do-IRN] [begin-3ps/A]
[DEP] [V]
'He began (to do) what he is supposed to do.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
778 Multi-clause constructions

d) Affirmative with 'forget':

k-itQa-nene bajti-tQäu
[ 1 sg=do-IRN] [forget-lsg/A]
[DEP] [V]
Ί forgot (to do) what I was supposed to do.'

While there is not enough reliable data to formulate consistent rules, it appears
that with verbs other than netoeka, the IRN form always co-occurs with a
proclitic on the dependent verb, which provides clear evidence for
nominalisation. A similar type of nominalisation was also observed with the
(probably lexicalised) form of ' w h a t he said' n=ajtöohwa (cf. (§6.6.3). The
catalogue in (1086) lists the verbs that can control a nominalisation/clause with
the IRN suffix. Note that with this suffix occurring on the first verb, both verbs
consistently refer to the same subject.

(1086) Control verbs that can take an IRN complement

netoeka 'be in the mood' (cf. (1084))

ißaa 'know' (also cf. (1064))
kwaauka 'think' (also cf. (1063), (1072a))
esenetaa 'believe' (also cf. (1072b))
kwitukua 'know' (also cf. (1073a))
bajtia 'forget' (also cf. (1068e), (1073b))
najßaa 'finish' (also cf. (1066a))
nunuetia 'begin'(also cf. (1066b))
hoaasia 'be afraid o f (also cf. (1067), (1074))
huaraa 'see'(also cf. (1075b))
bia 'tell' (also cf. (1076))

The morpheme for irrealis nominalisation has several allomorphs: -ene, -nee,
-rtene, and -ße. While the variant -eve can be regarded as the underlying form,
-nee is attached to verbs that end in a long vowel or vowel sequence and to some
other verbs, particularly those with a root-final /a/, such as it$aa 'do'. 9 2 The
variant -pe is suffixed to verbs with a root-final I'll, such as Ma 'eat, sima
' s l e e p ' , or ima ' g o u p ' ; note that their final root vowel is lengthened before -ße
to result into [kiijie], etc.

92
Other verbs which take this allomorph include lauekoa 'be sitting', esenetaa 'believe',
räasaa 'dance', kwaasia 'be afraid o f , ajtoa 'say', and the irregular forms koaraa
'see' (to result into kwaa-nee), aunaa 'hear' (to result into aua-nee), and it$aa 'do'
(to result into itpa-nene).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 779

20.2.4 Complementation strategies

The types of complementation investigated in the previous sections involved

special control verbs in the main clause of each type and a special form to occur
in the complement clause. There are a few further strategies for
complementation which clearly differ from this. Firstly, the complementiser
haanu combines its function with the grammatical category of purpose. It can
occur after verbs of any type and it does not "control" the verb it is attached to;
it rather functions in a similar way as a conjunctions, as is analysed in §20.2.4.1.
Secondly, the quotative verb naa, as described in §20.2.4.2, takes a quotation of
any length as a complement and does not require any additional markers for this
function. Thirdly, a construction which is structurally identical to a subordinate
clause with conjunction, is used as a means of complementation by younger
speakers (cf. §20.2.4.3). As noted in §20.2.4.4, questions cannot be integrated in
a complement clause, but are juxtaposed as independent clauses.

20.2.4.1 Purposive complementiser "haanu "

Apart from the subordinate construction with kujjia/kiuataa (cf. §20.1.2.5),

Urarina has a special type of purposive clause that involves the form haanu. As
this form exhibits a high degree of variation, it is difficult to determine what
word class it belongs to. While haanu appears in the same position as a
conjunction, the use of this element is also extended to other functions. The
different ways to employ haanu can be described as follows:
1) haanu may function as a "purposive complementiser" and is usually
preceded by a subordinate marker. Due to the variation mentioned in §20.1,
the subordinate marker can be realised as =ne (such as in (1090b)) or as =na
(cf. (1087)) - or it can be omitted altogether (e.g. as in (1087c)).
2) The form haanu can also occur after the neutral form of the verb, which does
not correspond to the behaviour of any conjunction of Urarina (cf. (1089)).
While its function in this context combines complementation and purpose, it
would be problematic to describe haanu as a conjunction here.
3) It can occur in any of the functions as described in 1) and 2) and is followed
by the postposition he in addition. In this context, again, haanu could be
characterised as a special kind of complementiser (cf. (1090)).

In examples of the first type, haanu behaves similar to a conjunction: it occurs

in a subordinate clause and is preceded by the subordinate marker =ne or variant
=na, which can both be optionally omitted. The verb of this clause is inflected
in the same way as other subordinate clauses and occurs in the D-form.
However, the syntactic function of haanu is that of a complementiser: As

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
780 Multi-clause constructions

illustrated in (1087), the clause to which it is attached becomes an argument of

the main clause. In (1087a), it occurs as an Ο argument, whereas in (1087b), it
functions as the S argument of the intransitive verb of the main clause.93 In this
sentence, the structure of the complement clause exhibits one peculiarity: the
verb 'return', to which =ne haanu is attached, functions as a modifier of the
noun enanihja 'canoe'. The part kau-a=ne haanu can be translated as
'(something) to return', which in turn modifies the noun that follows and forms
the complex NP 'a canoe in order to return'. Thus, any verb to which (=ne)
haanu is attached functions as a kind of nominalisation, which can either occur
as a noun head (such as in (1087a)), or can modify another noun, as in (1087b).
However, morphologically, the form cannot be described as an instance of
nominalisation, as it is attached to a fully inflected verb - and thus should be
characterised as a complement clause. The example in (1087c) illustrates that
haanu may also occur without the subordinate marker.

(1087) Use of haanu as a complementiser

a) Ο argument:
ßäe rautohwei nesari-tQuru-a=ne haanu eru-uru-a=ne hau
[already easily hunt-PL-3ps/D PURP] [find-PL-3ps/D] [because]
[Ο] [V] [CNJ]
'... because they already found [things] to hunt easily'

b) S argument:
ne-m-a kau-a=ne haanu enanihja=na hau ,
[be-NRG-3ps/D] [return-3ps/D=SUB PURP canoe=SUB] [because]
[VI] [SI] [CNJ 1 ]

nesoonete
[be.scared:3ps/E]
[V2]
'As there was no canoe so that he could return home, he was scared.'

c) Omission of subordinate marker:

eri-anu haanu kwituku-a-ü
[speak-lsg/D PURP] [know-NEG-lsgA]
[O] [V]
Ί did not know what to say.'

93
A clause with haanu can also function as the A argument of a transitive clause.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 781

One peculiarity should be mentioned with regard to the realisation of

=ne haanu. While it is a very productive form (though not observed with stative
verbs), there is one example in which the combination of =ne + haanu is
merged and altered to -eraanu. Strangely enough, this form is a suffix and
directly attached to the root of the verb for 'cut' as shown in (1088). In this
isolated case, it can be analysed as a nominalisation, but with exactly the same
purposive function as =ne haanu. It is plausible to assume that this is a result of
diachronical changes this construction is undergoing. The example in (1088) is
from a traditional narrative. However, more data would be needed in order to
argue that suffixation is being replaced by a more analytic form, i.e. cliticisation
plus conjunction.

(1088) Variant -eraanu

kanu fivaru-eraanu ra-a ku-a kanu lana

[lsg cut-PURP] [receive-NTR go-3ps/A] [lsg husband]
[Ο] [V] [A]
'My husband has gone to get [something] to cut me.'

Another kind of use of haanu is after a verb which is marked with the neutral
suffix rather than any person inflection; thus, the purposive marker is attached
to a non-finite form of the verb. As discussed previously, the neutral suffix is
homophonous to the 3ps/A form, which is also used in serial verb constructions
(cf. §17). The example in (1089) clearly demonstrates that the clause does not
necessarily have reference to 3rd person. In this example, the reference in the
complement clause is to lsg ('what I am to say') and the presence of the neutral
form could be interpreted as a sort of impersonal marking.

(1089) Use of haanu after neutral form

ni-a eri-a=ne haanu=ne nii ajto-i hau ora ke

[be-3ps/D] [speak-NTR=SUB PURP=CND] [that say-2ps because hour VLI]
[VI] [SI] [PP2]

ii baha-re-ü=m
[2sg] [ask-IRR-1 sg/E=ASS]
[02] [V2]
'If there is anything to say, I will call you, at the time as you said.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 782 Multi-clause constructions

Another aspect is the occurrence of ke after haanu in a variety of examples. 94

Recall that the functions of ke were described twofold: firstly, as an instrumental
postposition (with some extended functions; cf. §5.7.4) and secondly, as a
valency-increaser (cf. §16.5). The following examples, where ke follows haanu
(with or without subordinate marker occurring before haanu) suggest that the
dependent clause created by the attachment of haanu is again treated like an
argument of the main verb; thus, it is to be treated like a complement clause.
What all examples in (1090) have in common is the presence of an Ο argument,
which in a way blocks the insertion of a complement clause as another Ο
argument. The only way to add an additional argument (or adjunct - recall that
there is no formal difference) is by the use of ke. Thus, in (1090a), for instance,
the Ο argument 'fuel' and the PP, which is in fact a complement clause
'(something) so that you return' are both arguments of the verb 'bring';
therefore, the valency increase marker ke is attached to the complement clause.
Note that the complement clauses in (1090a) and (1090c) follow the main
clause - which is not observed with other types of complementation. The other
examples, occurring with a different subordinate marker (or none) behave in the
same way: The complement clause is inserted as an (optional) PP with ke.

(1090) Variant with haanu ke

a) With =ne:
gasona kurete-i u-a-u, kuruata laata-he-i,
[fuel] [buy-PRT come-CAUl-IMP] [two bin-CNT-PRT]
[Ol] [VI] ([Ol])

kau-i=jie haanu ke
[return-2ps=SUB PURP VLI]
[PP]

'Buy fuel and bring it, two bins, for your return.'

b) With =na:
netei-tQa=na haanu ke ku enua te-i
[hang.up-3ps/D=SUB PURP VLI] [there] [pole] [put-PRT]
[PP] [ADV] [Ο] [V]
'... putting a pole there so that she would hang up herself

94
In combination with ke, the purposive marker also occurs as the variant haa (e.g.
(lb)).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 783

c) Without subordinate marker:

heriane enamhja siiri-ji=lu amu-a haanu ke
[maybe] [canoe] [have-NEG:3ps/A=REM] [walk-3ps/D PURP VLI]
[ADV] [Ο] [V] [PP]
'Maybe they did not have canoes with which they could go.'

There also is a way to express locational purpose, which is realised by the form
-naha (or variant -neha). This form, which is attached as a suffix to a verbal
root, refers to 'a place to ...', such as illustrated in (1091). The phonological
similarity of this suffix with haanu is evident, as is the presence of [ne] (again
alternating with [na]), which corresponds to the optional subordinate marker
also observed with haanu.
Syntactically, a verb suffixed with -naha functions as a locational noun and
thus can be characterised as a nominalised verb. Note that a nominalisation with
-naha also complies with the properties suggested for locational nouns in
§5.1.2, (cf. table in (216)). For instance, it can occur with a demonstrative or
with a modifier, it may function as a head of an NP (also of a possessive NP),
and it can take the locative suffix -u. (However, it does not function as a
modifier itself).
As shown in (1091a), the suffixation of -naha to the copula results in the
noun 'place to be', which is frequently used to refer to one's village or
temporary settlement. In example in (1091b), the suffix is attached to the
transitive verb enoalaa 'cook', which in turn is embedded in a dependent clause
with the purposive/consecutive conjunction kiijpa.

(1091) Locational purposive form -naha

a)
ßäe ne-naha kau-a ku-uru-a eene-kuru
[already] [be-LPP] [return-NTR go-PL-3ps/A] [woman-PL]
[ADV] [LOC] [V] [S]
'The woman are already going home.' [Lit. 'they return to their place']

b)
usi ha-anu m-a lertone enoala-naha=ne kujjia
[firemake-lsg/A] [be-3ps/D food cook-LPP=SUB so.that]
[Main] [DEP]
Ί made a fire so that there would be a place to cook food.'

As a variation in the use of -naha, the suffix may be preceded by the

nominaliser -naa, i.e. in this case, it is attached to a nominalised verb. There is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
784 Multi-clause constructions

not apparent difference between a verb such as ne-naha vs. ne-naa-naha 'place
to live' (cf. (365b); (546b) uhoana ha-naanaha 'place to make fields').
Another factor to be noted is the phonological resemblance of haanu to the
conjunction hau , which might be a coincidence. As a hypothesis, the two could
be diachronically related, as haanu expresses purpose and hau refers to reason,
which in a way describes a relation between cause and effect.
There also is a further purposive morpheme, =nuhe, which appears to be
phonologically related to haanu and combined with /ke/. However, while the
function of this form is clearly purposive, its distribution is less than clear.
Specifically, it can be cliticised to various different kinds of word classes. In
examples discussed in other chapters, it is found with the following word
classes, including inflected and uninfected forms. Occurrences of =nuke are
attested in the following contexts: attached to inflected verb (cf. (348a)), to a
verb root ((992)), to a noun ((983)), to the multi-function word kauatpa 'good'
in ((431a)), and following a postposition ((1006)). While no semantic
differences between haanu and =nuke seem to apply, the difference is
structural, in that haanu is more restricted with respect to its distribution. 95

20.2.4.2 Complementation with "naa" 'say'

The verb ma 'say' takes a quotation as its complement, but the quotation is not
followed by any subordinating marker such as =ne or -na. As illustrated in
(1092), the quotation always precedes the quotational verb, which perfectly
corresponds to the position of other complement clauses. While naa cannot be
described as "transitive" - it does not take noun arguments, it cannot be used
intransitively: a quotation always has to precede it. Further examples involving
naa occur throughout this grammar (e.g. (961c), (993a)).

(1092) Quotation as a complement

dr,aha~i saatono su-Ί ke lenone-ku-tQe=ra kanaanaj-uru

[come.on-PRT] [in.the.end kill-PRT] [VLI] [eat-IMP-PL] [child-PL]
[DEP1] [DEP2] [PP] [V] [vocative]

95
Yet another purposive morpheme is the (rare) irrealis purposive suffix -nenu
(occurring in sentence no. (46) of Text 53 in the appendix), which expresses 'to be
done'.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 785

na-e hetau nii kana+kwaaun-era

[say-3ps/E Η RS] [that our.creator]
[V] [S]
'And then, eventually, children, kill them and eat them, said God.' [About the peccaries
God created for the people]

An important difference between quotations and complement clauses is that

quotations may be much longer than any complement construction, i.e. they can
involve several clauses or even sentences.

20.2,4.3 Complementation with "hau" 'because'

Another strategy of complementation (or nominalisation) is found in the

contemporary language only. In fact, this construction, which employs the
conjunction hau (literally 'because'; cf. §20.1.2.2), is not attested in any of the
traditional narratives, nor is it used by older speakers. As the sentences in
(1093) exemplify, a dependent clause followed by häu (with optional insertion
of subordinate marker =ne before the conjunction) can be interpreted as an
argument, which is best translated as a nominalisation. In (1093a), for instance,
the literal translation, which would be 'Have you heard because I said this?' is
not the intended reading. Instead, the correct interpretation is as a complement
clause, as indicated by the gloss.

(1093) Complementation with häu

a)
kaa eri-anu=ne häu auna-i=jia
[this speak-lsg/D=SUB because] [hear-2ps=INT]
[DEP] [V]
'Have you ("this") heard what I said?'

b)
kuoaw=te ajto-a kana+kwaaun-era-ne hau auna-kaatQe
[let's=FOC] [say-3ps/D our.creator=SUB because] [hear-1 pl/in]
[INTR] [DEP] [V]
'Let's hear what God says!' [From sermon]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
786 Multi-clause constructions

2 0.2.4.4 Complementation and questions

When questions are combined with complement-taking verbs, there is no formal

distinction or marking of any kind. The clause which contains the question
occurs in the same way as it would without a control verb, which is illustrated
in (1094). It can occur before or after the control verb. Thus, while
semantically, the question may be understood as being subordinate to the main
verb, syntactically, the two clauses are juxtaposed without any linking
morpheme. In fact, the use of the Ε-form in the interrogative clause displayed in
(1094) shows that the question is not a dependent clause, as this would require
the use of the D-form. Note that the use of the particle tQäiti 'ever' is typical in
this type of construction, but not restricted to this context (cf. §21.5). It is not a
special marker for interrogatives in complement clauses.

(1094) Complement taking verbs and questions

d$a tQäiti hwaa letoa-1 ne=lu, hoatia faoituku-a-akaanu

[what ever PSN send-PRT be:3ps/E=REM] [not] [know-NEG-lpl/ex]
[Q] [INTR] [V]
'We don't know who (ever) sent John' [NT: Mark 11:33]

20.2.5 Summary of complementation types

The table in (1095) lists all verbs that may occur as control verbs of a
complement clause. The two major types are distinguished into verbs that can
take an infinitive clause, which is marked by the suffix -na, and those that
control a different subject complement clause, marked by =ne. Even though the
irrealis nominalisation may not strictly constitute a type of complementation,
the third column of the table states whether the respective verbs can be used
with this form.
While I have sorted the different verbs according to semantic categories (as
shaded), it becomes evident that it is impossible to categorise these in any way.
For instance, no clear classification can be made for verbs that can control an
IRN form: there are some verbs in any semantic category, except for verbs of
liking, which can control an IRN form, but others of the same semantic type
cannot. Same subject control verbs are slightly more homogeneous, as all verbs
of liking and the two verbs of beginning and ending can function as same
subject control verbs, but verbs of perception and verbs of saying cannot. For
verbs of thinking and all other semantic types, their role is not homogeneous.
Different subject complementation is possible with most verbs listed here,
though some are only attested in elicitation or in the NT. Verbs of comparison

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Complement clauses 787

do not occur in this function and for most semantic types (except verbs of
perception), there also are verbs that cannot take a different subject complement
clause. Overall, a clear classification of Urarina complement-taking verbs is not
possible, as each verb may have different preferences, not tied to its semantic
class.
(1095) Summary of complement-taking verbs96

Verb type Verb Can take -na Can take =ne Can take I R N
Liking heria 'want' yes yes no
Liking kuurahia 'desire' yes no no
Liking rautaa 'like' yes no no
Thinking kzvaauka 'think' yes yes yes
Thinking esenetaa 'believe' no yes yes
Thinking hisenia 'dream' no (in EL only) no
Thinking bajtia 'forget' in NT and EL in NT and EL yes
Thinking kohivaelatoa 'be (in EL only) in NT and EL no
surprised'
Thinking/Emotion netoeka 'be in the (in EL only) (in EL only) yes
mood'
Begin/end numetia 'begin' yes in NT, yes
with SS ref.
Begin/end najßaa 'finish' yes in NT, no
with SS ref.
Comparison amuma 'be as if yes no no
Comparison tokoartia 'compare' (in EL only) no no
Comparison rihitoa 'be like' (in EL only) no yes
Knowing/ability ijiaa 'know' yes no yes
Knowing/ability najnia 'be able' yes no no
Knowing/ability kwitttkua 'know' no yes yes
Perception annaa 'hear' no yes no
Perception kwaraa 'see' no yes yes
Saying bia 'tell' no yes yes
Saying bahaa 'ask' no in NT and EL no
Saying tiaa 'say' no no no
Fear hvaasia 'be afraid' yes yes yes
Permission hiama 'let/leave' in NT and EL in NT, yes
'Leave' with SS ref.
'Must' tietene 'must' no yes no

96
Abbreviations used in this table: EL = "in elicitation"; SS = "same subject".
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
788 Multi-clause constructions

20.3 Participle clauses

The most common type of dependent clause in Urarina is the participle

construction. This clause type can be defined as a kind of subordinate clause, as
it grammatically depends on the main clause: Urarina participle clauses
constitute non-finite same subject dependent clauses where person/subject
inflection is marked on the verb of the main clause. In contrast to its syntactic
and referential dependency, a participle clause is semantically independent from
the main clause. For instance, the sentence 'As she was beautiful, they wanted
her, caught her and danced with her', as shown in (1096), contains three
dependent clauses, the first one expressed as a subordinate clause with
conjunction, the two other ones as participle clauses. While the subject is
expressed by the personal suffixes in DEP1, the clauses DEP2 and DEP3 are
morphologically dependent from the main verb 'dance', which marks 3pl. An
alternative translation of the sentence would be '... wanting her, catching her,
they danced with her'. Semantically, however, there is no dependency relation
between the two participles and the main clause, as each of the actions
('wanting', 'catching') describes an independent event. The subordinate clause
'as she was beautiful', instead, implies some degree of semantic dependency,
expressed as a causal relation.

(1096) Subordination and participles

kauato-a hau, here-Ί, muku-l, ajpa räasa-ure
[be.nice-3ps/D because] [want-PRT] [catch-PRT] [with] [dance-3pl/E]
[DEP1] [DEP2] [DEP3] [PP] [V]
'As she was beautiful, they wanted here, caught her, and danced with her.

While the form I call "participle" here also may be characterised as "gerund",
"clause chaining", or "converb" in other terms, these all refer to very similar (if
not identical) phenomena, as is discussed in Genetti (2005). Based on the
analysis of the participle in Dolakha Newar (Tibeto-Burman), Genetti describes
the features of a participle in a way, which exactly match the characteristics of
07
the Urarina form under investigation (cf. Genetti 2005:40):
"(i) Two or more verbs or clauses occur in a sequence. Non-final verbs are
inflected with the participle; final verbs may inflect with the full range of
verb morphology.
(ii) The construction forms a chain of verbs or clauses whose status with respect
to the surrounding text is determined by the morphology on the final verb.

97
Further discussion on the typology of converbs/participles can be found in
Haspelmath & Koenig (1995).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Participle clauses 789

(iii) The participle is vague in meaning; it has no particular semantic entailment,

instead the semantic relation between the participial clause and the
surrounding clauses is determined from the context.
(iv) A participial clause is neither a nominal argument, a nominal modifier, nor
a complement of the following verb or clause."

A majority of Urarina sentences that contain more than one verb involves the
participle form, which, to put it briefly, can be described as a non-finite clause
with same subject reference. The suffix -i can be preceded by certain suffixes,
as investigated in §20.3.1, including the continuous aspect marker, whereas
most other verbal categories are marked on the main verb. Most suffixes
occurring on the main verb also have scope over the participle, but exception
apply for those that may occur on the dependent verb itself, as will be shown in
§20.3.3. The participle suffix is compatible with any verb class. A finite verb
typically occurs at the end of the sentence, but it may also precede a participle
construction. On the semantic level, the participle can refer to a sequence of
events, an overlap of states, actions or events, or it may express adverbial
modification.

20.3.1 Morphological features

The underlying form of the participle suffix is -7, which occurs after all single
final root vowels and after diphthongs. With verb roots ending in a vowel
sequence or in a long vowel, the forms are irregular and depend on the
respective verb. For instance, with kuruae-ka 'be happy', the root structure is
retained and the suffix is realised as -hei (cf. (1097)). This allomorph also
occurs with the verb obaae-ka 'be angry', but here, the final root vowel is
deleted. With itQua-ka 'tie', /h/ is inserted between the two final root vowel
(and subsequently, /h/ is labialised, according to a general rule) and the suffix is
realised as -7. The regularities for these differences, if any, are unclear.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
790 Multi-clause constructions

(1097) Participle suffix allomorphy

Context Root Gloss Participle form

After /a/ /ama/ 'take along' ama-i
After Id /te/ 'give' te-i
After Ν /ari/ 'seek' ari-ϊ
After /u/ /ruku/ 'pull out' ruku-i
After /«/ /turn/ 'arrive' turu-1
After diphthong /it^au/ 'live' itQau-1
After long vowel /huu/ 'flood' huu-hwet
After vowel sequence /kuruae/ 'be happy' kuruae-hei
/obaae/ 'be angry' obaa-hei
/itgua/ 'tie' itguhioa-t

When combined with the intensifying suffix -to (cf. §10.2.310.2.2), the vowel
of the intensifier is assimilated and the resulting participle suffix is -tri, as in
kauato-a 'be nice', which becomes kauati-J.
As already mentioned in §12.4, the participle suffix -7 can be preceded by a
range of other verbal suffixes. These include both causative forms (-a and
-erate; slots 1 and 2), the continuous aspect marker (-ahe; slot 4), the
distributive suffix -ahoa and the plural object marker -naha (slot 7), and the
velocity suffix -uri (slot 8; cf. §12.2.8).
The use of the continuous aspect suffix implies a durative function, such as
exemplified by huara-ι 'looking' vs. hvara-he-ι 'watching'. In (1098),
continuous aspect and participle suffix are combined on the verb for 'pull',
• · that this
implying · process occurred over an extended period
· of time.
· 98

(1098) Continuous form of participle

kartii-hel mi huku-nohwi-itQa tarara-he-ι ra-e

[sort.of-PRT] [that stab-PASS-only pull-CNT-PRT] [receive-3ps/E]
[ADV] [DEP] [V]
'Like that, he pulled the one that had been stabbed and took it [home].'

Note that a combination of the aspect marker with -7 is also observed in some
lexicalised derivations now used as adverbs, as illustrated in (1099). In these
words, which only occur in this form, the function of the component /he/ is not
transparent any more.

98
This form is not to be confused with the form occurring in kami-hei 'being like that'),
in which is the participle allomorph -he is used due to the final long vowel of the
root.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Participle clauses 791

(1099) Adverbs that contain /he/

kauatQahei 'nice(ly)'
satonohei 'in the end'
banetohwei 'too much'
heritganehei 'as before'
haurianehei 'at first'
herahehei 'side by side'

20.3.2 Compatibility

A participle clause can be formed with verbs of any class, without restrictions.
The examples in (1100) illustrate this with different verb types as listed in §9.
The example in (1098) already shows a transitive verb ('pull') with this form.
An intransitive active verb is shown in (1100a). In (1100b) - from a narrative
where a woman had to swim home, the stative verb seteeto-a 'be wet' occurs as
a participle clause. This example also contains a posture verb ('be sitting') with
a participle suffix. Verbs of shape and "affect" verbs may occur with the
participle in the same way as the colour verb for 'red' displayed in (1100c).

(1100) Participle form of different verb types

a) Active intransitive verb:

wa d^atoane-i=tge, rta-t, turu-1, hoara-a u-e
[Oh! be.how-2ps=PLT say-PRT] [arrive-PRT] [see-NTR come-3ps/E]
[DEP1] [DEP2] [V]

rtii raj nakana

[that POSS son-in-law]
[A]
'Saying "Oh! How are you?" his son-in-law arrived and came to see.'

b) Stative intransitive verb; posture verb:

rtii häu hetau=te akau-ana seteeti-t lauhja-1 tihja-he
[that because HRS=FOC] [water-inside be.wet-PRT] [be.sitting] [walk-3ps/E]
[CNJ] [DEP1] [DEP2] [V]

mi eerie
[that woman]
[S]
'Therefore, the woman moved swimming, become [very] wet in the water.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
792 Multi-clause constructions

c) Colour verb:
raj homaru kuane kwara-kuru-a hana
[POSS pleat.of.skirt inside see-PL-3ps/D when]
[DEP1]

kariti-i tuku-hwe lanaha-Ί raj n=usitQoe

[be.long-PRT] [be.lying-3ps/E] [be.red-PRT] [POSS fingernail]
[DEP1] [V] [DEP2] [S]
'When they looked in the pleat of her skirt, there was (lying), being [very] long and red,
his [the spirit's] fingernail.'

The participle is also found with "special" verbs that do not belong to any
specific verb class, such as the copula ma, the auxiliary aja, and the
enumerative verb naaohwaa, which are illustrated in (1101).

(1101) Participle form of "special" verbs

a) Copula:
hajti katQa ne-1 UQU+UQU+UQU+HQU na-a läe
[still man be-PRT] [(sound) say-3ps/A CHNG]
[DEP] [V]
'Still being people, they made ...' [About people who are being turned into peccaries]

b) Auxiliary aja:
aj-ahe-ί saku-e
[AUX-CNT-PRT] [follow-3ps/E]
[DEP] [V]
'She kept on doing it.' [The woman kept on urinating; from flood story]

c) Enumerative verb maohwaa·.

raj kalaui raj ßaka naaohwa-i-te kau ne-ure
[POSS son POSS father ENUM-PRT=FOC] [here] [be-3pl/E]
[DEP] [ADV] [V]
'His son and his father are here.'

20.3.3 Scope

As has become clear through the previous examples, person marking of the
subject is always expressed on the main verb, which thus has scope over the
main clause and the participle. In the same way, most suffixes occurring on the
main verb have scope over the participle clause. This is well illustrated by the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Participle clauses 793

examples in (1102), where the scope of negation expressed on the main verb
extends to the participle clause preceding it. Note that are no attested examples
in which this would apply to a participle clause that follows the main verb, but
this could also be due to the general lower frequency of dependent clauses in
this position.

(1102) Scope of negation over participle

a)
nii hau hetau=te kuriae bahitguhioa kahe
[that because HRS=FOC] [distance toilet from
[CNJ] ([DEP])

amu-1 ne-ene ruru

walk-PRT] [be-NEG:3ps/E] [howler.monkey]
[DEP] [V] [S]
'Therefore, the howler monkey does not go far away from his bathroom.'

b)
ktt hautohioern hoairi itQa-1 ne-ur-ene tabauru
[there truthfully ayahuasca do-PRT] [be-PL-NEG:3ps/E] [some]
[DEP] [V] [S]
'There some do not drink ayahuasca truthfully. *

As seen in §20.3.1, some grammatical categories can also be marked on the

participle verb itself. However, when these are marked on the main verb, they
do not have scope over the participle. For instance, it would not make sense for
a plural object suffix on the main verb to have scope over a dependent clause,
which could contain an intransitive verb. In (1103a), where the participle clause
contains a transitive verb, PLO is only marked on the main clause. However,
PLO marking occurs on both clauses in (1103b) and on the participle only in
(1103c). Thus, it can be concluded that the choice of verb where PLO is to be
marked is relatively free. In the same way, both causative forms, the velocity
marker and the continuous aspect suffix only have scope over the verb they are
attached to.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
794 Multi-clause constructions

(1103) Scope of PLO suffix

a) PLO on main and participle clause:

ku rarara-naha-~i, ku hetau hani-ßaha-a
[there lay.next.to.each.other-PLO-PRT] [there HRS] [select-PLO-3ps/A]
[DEP] [ADV] [V]
'Laying them next to each other, he counted them there.'

b) PLO on main clause and participle:

hitarii ku su-ahe-naha-1, ßäe najßäi su-ahe-naha-a
[all there kill-CNT-PLO-PRT] [already nicely] [kill-CNT-PLO-PRT-3ps/A
[DEP1] [DEP2] [V]

hetau
HRS]
'Killing them all, he nicely killed them.'

c) PLO on participle clause:

ku raralaito-naha-1, mi tgoae nii bakaua lureri d^amuiti-1
[there lay.together-PLO-PRT] [that above that Indio house destroy-PRT]
[DEP1] [DEP2]

ku raatiri
[there] [leave:3ps/E]
[ADV] [V]
'Laying them together there, he destroyed the Indios' house on top of them and left
them there.

T h e following e x a m p l e demonstrates that continuous aspect as marked by -ahe

only applies to the verb to which it is attached. In (1104), continuous aspect is
marked on ' r o w ' only, but not on the dependent verbs which follow this clause.
Its scope d o e s not extend to the main verb, as the event of arriving is not
durative, but punctual.

(1104) Scope of CNT suffix

mi rihei turaaka-he-i, n=enanihja ke tihita-ι, ama-t,

[that like row-CNT-PRT] [3ps=canoe INST tie-PRT] [take.along-PRT]
[DEP 1 ] [DEP2] [DEP3]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Participle clauses 795

turu-a ku-a mi tQu hjä nakana=ne hana, kwara-kure

[arrive-NTR go-3ps/A that CRTN just son-in-law=SUB when] [see-3pl/E]
[DEP4] [V]

sal
[all]
[A]
'Keeping on rowing like that, with [the hurt man's canoe] tied to his [own] canoe,
taking it along, when the son-in-law arrived [at the village], everybody saw him.'

20.3.4 Semantic functions

The participle construction has a range of meanings, as the joined clauses may
refer to sequences, or to overlapping actions, events, or states. Participles are
also used to express adverbial modification. The interpretation of which
function applies in the respective context is not formally marked and must be
derived from the context; in some cases, several interpretations may be possible,
or a combination of these.

20.3.4.1 Sequence of events

A very common function of participle clauses is to express a sequence of events

or actions between dependent and main clause. A causal relation between the
two clauses can, but need not be implied. Example (1105a) illustrates a
sequence of two participle clauses followed by another non-finite clause with
privative function (cf. §20.4), which each refer to subsequent actions in the
timeline. In (1105b), the dependent clause DEP2 and DEP3 each describe
events that occurred in the linear order given here. DEP2 also is a specification
(or repetition) of DEP1 and could be characterised as an overlapping action. A
causal relation is not implied in these examples.

(1105) Participle expressing sequence of events

a)
su-1, ama-i na~i rtitaa jiäe fcw hajti turn-elanaala,
[kill-PRT] [take-PRT over.there over.there] [already there still arrive-PRV]
[DEP1 ] [DEP2] [DEP3]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
796 Multi-clause constructions

ke usi muku-akaanu
[VLI] [fire] [burn-lpl/ex]
[PP] [Ο] [V]
'We killed [the crocodile], took it along over there at that place, and still not arriving yet
[at our destination], we burnt [= grilled] it on the fire.'

b)
usi itQa-l, ßae usi hoaauna-l,
[fire make-PRT] [already fire create-PRT]
[DEP1] [DEP2]

nii ke ßäe nii lenone enoala-1 ki

[that INST already that food cook-PRT] [eat:3ps/E]
[DEP3] [V]
'Making fire, creating a fire already, he cooked the food with it and ate it.'

A causal relation between the different actions or events listed as a sequence

can be interpreted in the following examples. In (1106a), 'getting angry' is the
reason for the kinkajou going to see the spider monkey. For (1106b), a causal or
possibly consecutive relation could be understood, since 'going from there' is
the basis for arriving. DEP2 in this sentence could be characterised as a
participle clause referring to overlapping actions, as 'looking for land' was the
objective during the entire expedition. However, it could also be assigned causal
function, as the search for dry land was the actual reason for the trip.

(1106) Participle expressing sequence of events with causal relation

a)
ubaa-he-i hetau alau huara-a ku-e nii ttnee
[be.angry-CNT-PRT HRS] [spider.monkey] [see-NTR go-3ps/E] [that kinkajou]
[DEP] [Ο] [V] [A]
'The kinkajou got angry and went to see the spider monkey.'

b)
nii kahe ku-ahe-i kau turu-akaanu, atane ari-l
[that from go-CNT-PRT] [here] [arrive-2pl/ex] [land seek-PRT]
[DEP1] [ADV] [V] [DEP2]
'Going [on] from there, we arrived here, looking for land.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Participle clauses 797

20.3.4.2 Overlapping events, actions or states

A further function that may apply to participle constructions involves reference

to overlapping actions, events or states. The examples in (1107) describe
overlapping states. In (1107a), this is realised through the use of the copula. In
(1107b), the posture verb 'be sitting' refers to a state (also cf. §9.2.3).

(1107) Participle referring to overlapping states

a)
ni-a hetau lejhii katQa, lomaj beene ke komasaj-ki-Ί rn-naa
[be-3ps/A HRS] [one man] [PSN female VLI wife-VBL-PRT be-NOM]
[V] [S] [DEP]
'There was one man who had Lomai as his wife.'

b)
lejhii eene=te hana nitQataha-j fioanara siitQa-1 laueke
[one woman=FOC] [instead] [three-NOMSbj banana hold-PRT] [be.sitting]
[A] [CNJ] [DEP] [V]
O n e woman, instead, is sitting [there], holding three bananas.'

In other cases, the participle clause may express an overlap between different
actions or between states and actions, as is illustrated in (1108). Overlapping
actions are described in (1108a), where 'going up' clearly coincides with
' c o m i n g ' . In (1108b), the state of sitting is parallel to speaking, which can be
interpreted as an overlap of state and action.

(1108) Participle referring to overlapping actions or states

a) Overlap of actions:
rauihirii kanakaana loanari helöo tgu hjä im-t u-akaanu
[calmly lpl/ex shed towards CRTN just go.up-PRT] [come-lpl/ex
[DEP1] [DEP2]

hau atü=β a ku turu-a u-akaanu

because] [nevertheless=FOC:lpl] [there] [arrive-NTR come-lpl/ex]
[ADV] [ADV] [V]
'As we came, just calmly going up towards our shed, we still arrived there.' [despite the
difficulties]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
798 Multi-clause constructions

b) Overlap of state and action:

tahia laueki-1 hetau hoahee na-e
[over.there be.sitting-PRT HRS] [Hoh! say-3ps/E]
[DEP] [Main]
'Sitting there, she said "Hoh" [when he surprisingly arrived].'

20.3.4.3 Adverbial modification

Adverbial modification describes the manner in which an action is performed or

an event takes place. In §5.3.3 it was stated that all adverbs of manner are
morphologically complex, usually involving the participle suffix -1 (which may
be lexicalised in some cases). Participle formation is also used productively to
refer to clauses which express adverbial modification of the main clause. This is
illustrated by the sentences in (1109), which do not only refer to an overlap of
actions, but specifically describe the manner or circumstances by which the
action of the main clause is accompanied. In (1109a), 'suffering' specifies the
circumstances in which a man lost in the jungle returned home; the participle
clause 'felling with stone axes' in (1109b) functions as an adverbial modifier to
the way people used to cultivate their fields.

(1109) Participle clause with adverbial function

a)
rntoanei hetau=te nekajritQa-1 hatai kau-e nii lejhti katg,a
[like.that HRS=FOC] [suffer-PRT very] [return-3ps/E] [that one man]
[DEP1] [DEP2] [V] [DEP3]
'That one man returned home suffering very much.'

b)
mtoanei hetau=te n=uhoana ha-ure-rehete jioaelu katga-uru,
[like.that HRS=FOC] [3ps=field] [make-PL.HABl:3ps/E] [earlier man-PL]
[DEP1] [Ο] [V] [A]

aheri+ku-teri ke kufwihja-1
[stone+ASC-axe INST fell-PRT]
[DEP2]
'The ancient people made their fields like that, felling [trees] with stone axes.'

The example in (1109b) also illustrates the use of verbs such as nitoama 'be like
that', which occur in their participle form overwhelmingly often. Further
examples for this include the forms niirihihe-i 'like that', the quotative na-i

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Participle clauses 799

'saying', and some other words which mainly occur with -7, but rarely with
other types of inflection. These may well be on their way of becoming
lexicalised.

20.3.4.4 Ambiguous function of the participle

The distinction in interpretation with regard to sequence, overlap, or

modification is not always clear-cut. In some cases, different functions may be
combined and it is only possible to determine the exact reading by the context.
Example (1110) contains eight different participle clauses with different
functions. The sentence (from a picture description) describes how a man helps
another person who was hurt by robbers.

(1110) Combination of functions

JaoitQa-l, ama-1, lureri baha-1, n=asae te-1, ku te-1,

[heal-PRT] [take-PRT][house ask-PRT] [3ps=under put-PRT] [there put-PRT]
[DEP1] [DEP2] [DEP3] [DEP4] [DEP5]

kwitga-l, kwitga-e, hano-ahe-1 bera-~i

[heal-PRT] [heal-3ps/E] [spend.night-PRT] [care.for-PRT]
[DEP6] [V] [DEP7] [DEP8]
'Healing him, he took him along, asked for a house, put him there; putting him there
and healing him, he healed him, spending the night and caring for him.'

DEP1 'healing him' may be understood as describing the first aid given to the
injured person. The subsequent clauses DEP2 to DEP4 are in a linear sequence
to the first event, which is again summarised by the main clause 'he healed
him'. The repetition o f ' h e a l i n g him' in DEP6 could be interpreted as returning
to the general theme of the sentence, also expressed by the main clause. DEP5
is a repetition or summary of DEP4. DEP7 refers to another event that follows
the actions described previously. DEP8 'caring for him', however, may be
understood as an adverbial modification, describing the way of how the healing
was performed. At the same time, DEP8 refers to an overlapping action with
'healing him'.
It is obvious that, without the knowledge of the images given in this
particular situation, it would be difficult to provide the correct interpretation of
the participle clauses in examples such as (1110).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
800 Multi-clause constructions

20.3.5 Syntactic properties

The position of a participle clause corresponds to the general tendency of

Urarina dependent clauses to occur before the main clause (which is in
correlation to the unmarked OV order as discussed above). Occurrence after the
main clause is also possible, but less common (cf. §18.5). The position of a
participle clause does not have an impact on its function or interpretation. In
(1111), this is demonstrated by participles with different functions, occurring
before the main clause, whereas (1112) shows examples in which the same
functions apply, but occurring with participles that follow the main clause.

(1111) Participle preceding the main clause

a) Sequential function:
u-u=te läe i-muku-1 i=ju-aka=ra
[come-IMP CHNG] [2sg=catch-PRT] [2sg=kill-lpl/du=EMF]
[VI] [DEP2] [V2]
'Come and I will catch you and kill you!' [said to a jaguar]

b) Overlapping actions:
itQafwa-anu=ne hana, tiririhi-i obaae-ka hatiolari=ne hana
[shoot-lsg/D when] [roar-PRT][get.angry-3ps/D jaguar=SUB when]
[DEPl] [DEP2] [DEP3]
'When I shot, [when] the jaguar was getting angry, roaring,... [I looked down].'

c) Adverbial function:
aka=te läe helaj-tQa nelaurja-i it$au-e=ra
[3ps=FOC CHNG] [separate-only be.sitting-PRT] [live-3ps/E=EMF]
[S] [DEP] [V]
'He just lives in a single place, sitting there.'

(1112) Participle following the main clause

a) Sequential function:
kauru letoa-a, kuruata+tgurjka hani-ßaha-ι, nakivaaunei sauki huhu-i
[3pl] [send-3ps/A] [two+ten select-PLO-PRT] [again five add-PRT]
[Ο] [V] [DEPl] [DEP2]
'Selecting twenty [men] and adding [another] five, he sent them.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Participle clauses 801

b) Overlapping states:
ßäe lenone-he-uru-a, heena-i ne-ti-l
[already] [eat-CNT-PL-3ps/A] [be.four-PRT] [be-INTS-PRT]
[ADV] [V] [DEP1] [DEP2]
'They are eating already, being four [people].'

c) Adverbial function:
mi hana=te näi kahe kuriae rurute, enua siitga-l
[that when=FOC] [over.there from] [far] [walk:3ps/E] [tree hold-PRT]
[CNJ] [PP] [ADV] [V] [DEP]
'Then he walks [away] far from there, holding a stick.'

It is unclear what functional difference between the two positions might apply.
Interestingly, participles that follow the main clause do not automatically
express a sequence of events or actions, where the event described by the
participle clause follows the one of the main clause. This is clearly
demonstrated by the examples in (111 2a), where the selection of people
precedes the sending, even though 'counting' follows the main clause. As a
possible explanation, participle clauses that follow the main clause could be
viewed as afterthoughts. For (1112b), no sequence of events can be determined
as the clauses refer to overlapping states.
Interestingly, when a sequential interpretation applies, the dependent clause
tends to be understood as occurring first in the timeline (with respect to the
main clause), even when it follows the main clause. This is demonstrated by
(1113), where 'leaving the booty' occurs prior to entering the canoe. However,
multiple interpretations could suggest an alternative reading here, as 'leaving
the booty' could also be understood as an overlapping action, as to imply 'he
went away while leaving the booty'.

(1113) Sequential participle clause and timeline

ßäe enamhja kuane fxoau-a ku-e,

[already] [canoe inside] [go.down-NTR go-3ps/E]
[ADV] [PP] [V]

ßäe ku mi raj bitoasiße aräala raatiri-ι

[already there that POSS booty tapir leave:3ps/E]
[DEP]
'He already went [to sit] in the canoe, already leaving his booty, the tapir.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
802 Multi-clause constructions

20.4 Other non-finite verbal suffixes

A s mentioned in §12.4, there are two further verbal suffixes that form a
dependent verb: the suffix -lat 'while' and the privative form -elanaala
'without'. In the same way as the participle suffix, the verb to which these
suffixes are attached is non-finite and refers to the same subject as the main
clause. Person is thus only marked on the main verb. The function o f the
respective clauses is according to the semantic content of the suffix. The
examples in (1114) illustrate the use of these forms. While -elanaala is attached
to the root of the verb, note that -lat follows the neutral suffix.

(1114) Other non-finite verbal suffixes

a) -lat 'while':
akau ra-a ht-a-lal bmo-a
[water receive-NTR go-NTR-while] [fall-3ps/A]
[DEP] [V]
'She fell while she was going to fetch water.'

b) -elanaala 'without':
kalaui muku-em-akaanu na-ι, kalaui muku-elanaala u-akaanu
[son catch-NEG-lpl/ex say-PRT] [son catch-PRV] [come-lpl/ex]
[DEP1] [DEP2] [V]
'Saying "we did not catch its [the jaguar's] creature", we came without catching its
creature.'

20.5 Clause coordination

In contrast to the combination of participles or other dependent clauses with a

main clause, conjoining two clauses with conjunctions t$äe and nete is a clear
instance of coordination, as both clauses are independent from a grammatical as
well as from a semantic point of view. Coordinated clauses are not normally
dependent from another and the use of a coordinator merely relates the content
of the two clauses (or sentences) on a pragmatic level (for an exception, see
(1117)). Both conjunctions discussed here have a peculiar status, as tgae may
also be defined as having adverbial function, while nete is a morphologically
complex construct with a high degree of lexicalisation.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Clause coordination 803

20.5.1 "tQäe" 'also'

The word tQäe differs from subordinate conjunctions in that is cannot co-occur
with the subordinate marker =m. Its position and function suggest that it is a
hybrid between adverb and conjunction, as it can have conjoining function
between two independent clauses or sentences, or between nouns. In both
examples of (1115), tQäe occurs at the beginning of an independent clause,
which follows another independent clause. In (1115a), tQäe refers to a similar
theme as the one mentioned in the sentence preceding it, as both sentences
involve 'buying' of different items. This also applies to (1115b), where the
actions described in the different clauses are thematically related: 'Making a
birth shed' and 'digging a hole' both are relevant preparations for giving birth.

(1115) Function of tQäe as referring to related actions

a) [Preceding: O K , I will buy [the watches] tomorrow.']

tQäe uhuee+kivi kurete-u, kuruata soru ne-1 tableta
[also] [diarrhea+medicine] [buy-IMP] [two Sol be-PRT tablet]
[CNJ] [Ο] [V] [DEP]
'Also, buy medicine against diarrhea, [the one for] two Soles' [Sol is the national
currency.]

b) [Preceding: 'As they were pregnant and had their baby, she applied magic and did
everything as the monkey had done.']
ku heritQanehel hata te-ι, tQäe atane bahe-~i,
[there as.before birth.shed make-PRT] [also] [earth dig-PRT]
[DEP1] [CNJ] [DEP2]

kuane nakaaohwa-a=ne kujßa

[inside bathe-3ps/D=SUB so.that]
[DEP3]
'[Then] she made a birth shed as before, also digging a hole, so that that she would
bathe [the baby] in there ...'

In (1116a), however, tQäe marks a change of theme: 'Bring money for my trip'
and 'tell Brother Santiago' are not thematically related; thus, 'also' refers to a
list of orders given by the speaker, as he asks the addressee to complete a
number of different tasks for him. Similarly in (1116b), tQäe does not imply that
the two sentences 'thank you for all that' and 'we must always rejoice' are
related actions or events. Instead, it marks the two sentences as both being part
of a wider theme, such as the prayer itself, in which a number of different topics

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
804 Multi-clause constructions

are addressed. These could be viewed as being related in that God is the
addressee.

(1116) Function of tQäe as referring to wider discourse

a) [Preceding: 'Bring the money for my trip.']

tQäe kartii, ermano santjago raj be-u
[also] [GPF] [brother PSN for] [tell-IMP]
[CNJ] [PP] [V]
'Also, tell Brother Santiago, ... [that I wasn't able to get his rifle repaired.]'

b) [Preceding: 'Thank you for all that.']

tQäe salmo kuaw ajte petononei kam+koaaun-era kuruaaeka
[also] [psalm inside] [say:3ps/E] [always] [our.creator] [rejoice- lpl/du
[CNJ] [PP1] [VI] [ADV] [02] [V2]

nerajjie
must]
'Also, it says in the psalm, we must always rejoice [in] God.' [From prayer]

In (1117), tQäe again refers to the wider discourse. As 'come' has not been
previously mentioned in the discourse, the use of tQäe implies that 'also' may
refer to any action by the addressee, such as to mean 'It is OK that you look for
water wolves, but also come when I call.' A crucial difference to the previous
examples is, however, that tQäe occurs between a dependent and an independent
clause. Thus, the insertion is not meant to indicate a conjoining function
between these two clauses.

(1117) Occurrence of tQäe after dependent clause

[Preceding: 'Yes, go ahead and look for water wolves, while I will check my tapir
trap.']
m-a d$a=ne e?eo-käu =ne tQäe u-u=tQe
[be-3ps/D what=CND] [shout-lsg/D-CND] [also] [come-IMP=PLT]
[DEP1] [DEP2] [CNJ] [V]
'If there is something, if I shout, also come!'

As with the examples where tQäe occurs between two independent clauses, it
has no syntactic function in (1117), i.e. its presence is not syntactically relevant
for the relation between the two clauses. Thus, it is plausible to conclude that
tQäe is in fact not a conjunction in the strict sense. Instead, it could be viewed as
having adverbial function. This view is further confirmed by the example in
(1118), where tQäe does not occur with reference to a clause, but precedes
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Clause coordination 805

nouns. Even though this is an isolated example, this sentence clearly

demonstrates that tQäe is not a genuine clause conjunction.

(1118) Occurrence of tQäe before nouns

ku ari-uru-a hau hetau ku m-a fwanara+kata kahe

[there seek-PL-3ps/D because HRS] [there] [be-3ps/A] [banana+ripe from
[DEP] [ADV] [V] [S]

ne-naa aune-bäka, tQäe erene+kutuuhue, tQäe hi saarana+kusitQuri

be-NOM plantain+soup] [also] [fat+fried.food] [also] [there] [guineapig+soup]
[CNJ] [S] [CNJ] [ADV] [S]
'As they were seeking, there was plantain soup from ripe bananas, also fried food, and
also guinea pig soup,... [all this was there].'

There also are further peculiarities regarding the position of tQäe. One aspect is
its occurrence in sentences with a clause introducer. In this case, tQäe
consistently follows the introducer, as is shown in (1119).

(1119) Position of tQäe after introducers

a) [Preceding: 'Yes, I've heard you.']
ta tQäe d^a ku)ituku-i=ße, mumsipio kahe
[NEGQ] [also] [what] [know-2ps=NEGF] [council from]
[INTR] [CNJ] [Ο] [V] [PP]
'Also, don't you know anything, from the city council [about the radio we applied for]?'

b) [Preceding: 'When the baby's umbilical cord falls off, the father should go hunting.']
mhjauria tQäe kati saate-nana-a=ne
[don't] [also] [black.monkey] [kill.by.blowgun-ILT-NTR=NEGF]
[INTR] [CNJ] [Ο] [V]
'Also, he ought not to kill black monkeys.'

In the following example, tQäe is embedded into the clause, preceding the noun,
as was observed in (1118). In this case, it cannot be shifted to the front.

(1120) tQäe embedded in clause

[Preceding: 'When you arrive in Lomai's land, there will be talking trees.']
ne-ri-tQa=i tQäe atane mamale-j
[be-IRR-3ps/A=ASS] [also] [land make.stick-NOMSbj]
[V] [CNJ] [S]
'There will also be magnetic soil... [and beautiful axes, just like we have.]'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
806 Multi-clause constructions

In some examples, tg,äe co-occurs with hana (also used as a postposition for
'inside'), such as illustrated in (1121a): The combination of both words can be
translated as 'besides', 'by the way', or 'on the other hand' and implies a kind
of contrast between the first sentence by adding new, contrasting information to
the first one. In (1121b), hana t^ae refers to related actions, as the stump
(described in the preceding sentence) also stayed at a certain place - as does the
upper part.

(1121) Combination of tg,ae and hana

a) [Preceding: 'OK, I will what you told me.']

d^atoam-a hana+tg,äe tQamö
[be.how-3ps/A] [besides] [PSN]
[V] [CNJ] [S]
'By the way, how is Salomon?

b) [Preceding: 'As he felled the giant cassava, its stump stayed at the MarafSon river.']
hana+tQäe raj haberoko hetau=te kukana ne-ι neda-e=lu
[besides] [POSS body HRS=FOC] [sea be-PRT] [stay-3ps/E=REM]
[CNJ] [S] [DEP] [V]
'Instead also, its upper part stayed at the sea.

Through the examples presented in this section, it becomes evident that tpäe
does not have a function that could be defined as syntactic, similar to
subordinate clauses or participles. While the latter clearly mark the combination
of two clauses by morphological means that result in a dependency relation
between the clauses, this is not case here. t$ae does not "conjoin" two clauses in
the strict sense, but each clause could occur without the addition of tgae. It has
pragmatic function on the discourse level, which may not exactly refer to the
actions or events described in the clauses involved, but may also have scope
over a wider context, such as was illustrated in (1116a), where it implied a
chain of commands.

20.5.2 "nete" 'but'

The word nete for 'but' can be used as a coordinator between independent
clauses. In contrast to t$äe, its position is fixed, as it can occur only as a
conjoining element between two clauses, but not within a clause. An interesting
peculiarity is that nete can be analysed as a morphologically complex form,
composed of the subordinate marker =ne and the focus marker =te - at least
from a diachronical perspective. However, this combination of two enclitics has

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Clause coordination 807

been lexicalised to a high extent, as will be demonstrated below. The examples

in (1122) illustrate the function of nete as an element that relates the content of
two independent clauses (or sentences) by expressing an opposition. The
independent status of the two clauses is manifested by the verbal inflection on
both clauses. In (1122a), the two clauses each contain a verb with 3pl marking
in the Ε-form, the second one being marked for negation. In (1122b), negation
again occurs on one of the verbs; in this case, one verb is inflected for person in
the Α-form, whereas the other exhibits marking for the Ε-form. The presence of
the adversative particle niki in both examples below is a coincidence. This
marker is further investigated in §22.5.

(1122) Use of nete 'but'

a)
tabauru=te niki kwituku-ure nete ßäe itQa-kur-ene
[some=FOC] [ADVRS] [know-3pl/E] [but] [already] [do-PL-NEG:3ps/E]
[A] [PARTI] [VI] [CNJ] [ADV2] [V2]
'Some [people] know [how to make the mixture], but they do not make it any more.'

b) [Preceding: 'Have you not caught any fish?']

aj-akaanu=ra, dadato-a niki nete muku-si-ße
[AUX-1 pl/ex=EMF] [pull-3ps/A] [ADVRS] [but] [catch-CPL-NEG:3ps/E]
[VI] [V2] [PART2] [CNJ] [V3]

kanaka ana katai=tga

[lpl/ex fishing.hook^EMF]
[A3]
'We did not. [The fish] pulled, but our fishing hooks did not catch it.'

The idea that nete is composed of the form =ne plus a focus marker is supported
by the fact that =te can be exchanged for another focus marker, according to the
subject marked on the verb (Recall that the focus marker exhibits concord with
the subject; cf. §19). In the (elicited) example (1123), the subject is lsg and the
second component of 'but' is =ne, which corresponds to the focus marker for
lsg. Accordingly with 1 pi, the form would be nena. However, the form nete
(where =te would be supposed to mark concordance with 2 nd /3 rd person) can
even be used even in contexts where the subject refers to 1 ps. It should also be
noted that an additional focus marker cannot be insert in sentences that are
preceded by 'but' - which further confirms the view that the second component
of 'but' is a focus marker (also cf. §5.8.3).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
808 Multi-clause constructions

(1123) 'But' with lps subjects

auna-kau nene/nete kwara-ene-ü

[hear-lsg/A] [but] [see-NEG-lsg/E]
[VI] [CNJ] [V2]
Ί hear it, but I did not see it.'

While it has been demonstrated that nete is composed of two different elements,
there are other aspects which support the view that nete is lexicalised:
1. If the component =te were to be defined as a focus marker, any verb
following nete would be inflected with the Ε-form for person marking. While
this is the case in most NT examples that contain 'but', the Α-form may also
be used in this context.
2. If the component =ne were to be defined as a subordinate marker, any verb
preceding nete would be inflected with the D-form for person marking - for
which there is no evidence.

Thus, it can be concluded that the forms for 'but' are at a late stage of
lexicalisation, where they still are partly transparent. In practice, however, the
combination of two morphemes is used as a single lexical unit. Interestingly,
younger speakers with an extended knowledge of Spanish also use the Spanish
word pero 'but' in their speech, sometimes even combined with nete.

20.6 Other strategies of coordination

There are two more constructions that could be regarded as alternative strategies
of clause coordination. The first one can be defined as ellipsis; it only occurs in
contexts where the quotational verb naa is elided, such as illustrated in (1124).
In this example, there is no finite verb that would refer to the subject, but a
quotation is simply inserted. The actual verb for 'say' (naa) is omitted, but
understood from the context.

(1124) Ellipsis of quotational verb

kwara-ϊ hetau, d^anuna kanu bua basihjau-i=ta

[see-PRT HRS] [why] [lsg bag] [steal-2ps=FRS]
[DEP] [INT] [Ο] [V]
'Looking [at him], [he said], "Why have you stolen my bag?'"

Another coordination strategy is the juxtaposition of finite clauses without any

morphological marker or conjunction. This construction is comparatively rare
and involves a pause between the different clauses. In (1125a), this is illustrated
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Other strategies of coordination 809

by a sentence that contains two clauses marked with the Ε-form; in (1125b), the
Α-form is used. The two forms can also be mixed, i.e. have a verb with one
form in one clause and a verb with the other in the following clause.

(1125) Juxtaposition of finite clauses

a) With E-form:
ku ku-ure, raatiri-ure
[there go-3pl/E] [leave-3pl/E]
[Mainl] [Main2]
'They went there, and they left him.'

b) With A-form:
poaelu ena aja=ne tokuanet itgafwa ni-ji=lu,
[earlier now AUX-3ps/D=SUB compared.to] [rifle be-NEG:3ps/A]
[DEP] [Mainl]

kartutoo ni-ji=lu
[shotgun.shell be-NEG:3ps/A]
[Main2]
'Earlier, comparing to what there is now, there were no rifles, there were no shotgun
shells,... [all those things for hunting].'

The function of this coordination strategy is comparable to the construction with

the enumerative verb naaohiuaa, which can be regarded as the preferred way of
expressing a chain of listed events or states. Most typically, the clauses in the
scope of naaohwaa are participle clauses, as illustrated in (1126). In these
examples, naaohwaa does not conjoin independent (finite) clauses, but follows
two participle clauses. In (1126a), naaohwaa occurs as a participle itself, thus
being dependent from the main verb ('be'). In (1126b), it functions as the main
verb, preceded by two participle clauses. The sentence in (1126c), in contrast,
shows that the enumerative verb may also refer to independent clauses. In this
example, the function of naaohwaa is to express the "listing" effect of the first
two clauses (Main 1+2), which are both independent clauses. However, through
the presence of naaohwaa, these are understood as being part of one sentence.

(1126) Coordinating function of naaohwaa

a)
äafihe-j d^a he hirete-naha-ϊ, ku-atca uahei
[be.small-NOMsbj what VLI buy-PLO-PRT] [there-only until
[DEP1 ]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
810 Multi-clause constructions

nakwaaunei itga-rate-i, naaohwa-ϊ akauru letoa-era

again do-CAU2-PRT] [ENUM-PRT] [them send-AG
[DEP2] [DEP3] ([Main])

tie-t ne-rehete-kure=lu fwatoro-kuru

be-PRT be-HAB 1 -3pI/E=REM patron-PL]
[Main]
'The patrones were those who used to send them, paying them a little and making them
do [the same works] again up to this point, all this.'

b)
Idabo kurete-1, semeto kurete-i, naaohwa-kat$e netene
[nail buy-PRT] [cement-PRT buy-PRT] [ENUM-lpl/in must]
[DEP1] [DEP2] [Main]
'We must buy nails, buy cement - all that.'

c)
akauru raj saabere ßae lerano ne-1 amu-e,
[3pl POSS machete already macana.üsh be-PRT walk-3ps/E]
[Mainl]

kauru raj eruari ahaijte ne-1 ku-e,

[3pl POSS mosquito.net paiche.fish be-PRT go-3ps/E]
[Main2]

naaohwa-t ßäe amu-ahe

[ENUM-PRT] [already walk-CNT:3ps/E]
[DEP] [Main3]
'Their machetes already turned into macana fish, their mosquito nets became patches,
all that - they already kept going.

It should be noted that the verb naaohiuaa is not only used in enumeration of
clauses, but commonly occurs in contexts where NPs are listed (cf. §5.8).
In summary, coordination in Urarina can be regarded as a relatively minor
feature of the grammar, based on the following considerations:
1) Both words described as conjunctions in this section are not genuine
conjunctions, as t$äe also has adverbial function and nete is a lexicalised
form composed out of clitics with a different function. The use of a loan for
'but' by younger people could also reflect the need for an alternative to
express coordination.
2) The occurrence of these conjunctions in the database is comparatively rare,
which implies that they might not represent popular strategies to conjoin

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Multi-clause constructions in comparison 811

clauses. In contrast, clause dependency as expressed by the various means

discussed in this chapter is an extremely common way to conjoin clauses.
3) The use of the conjunctions described above has no impact on the syntactic
or morphological structure in any way, which is different from some
dependent clauses, which may involve special inflection (in subordinate
clauses), or specific syntactic positions, such as in complement clauses.

20.7 Multi-clause constructions in comparison

As Urarina has a wide range of constructions that involve two or more verbs,
one may ask in what they these differ from each other. I will base this
comparison on participle clauses, as these are the most frequent type of multi-
clause construction in Urarina. Apart from formal and functional distinctions
between various constructions, there also are similarities. Serial verb
constructions, for instance, resemble participle clauses syntactically, in that an
SVC consists of a non-finite verb and a verb that indicates the subject and other
grammatical categories. A further feature both have in common is that they are
both same subject constructions. However, the crucial difference is that SVCs
are to be defined as a single clause, even though involving two verbs. Further
differences are to be found on the semantic level, as the function of serial verbs
strongly depends on the verbs involved. For instance, the presence of certain
verbs as VI in a SVC tends to imply directional, resultative, or puiposive
functions. Another aspect also is the restrictedness of SVCs in that only certain
verbs may occur as V2 (cf. § 17.2).
Subordinate clauses differ significantly from participles in many respects.
First of all, they can refer to a different subject than the main clause. Secondly,
they are finite clauses and require a specific type of inflection (the D-form), and
thirdly, they are always accompanied by some additional marking, which can be
a subordinate marker, a conjunction, or both. However, semantic similarities
with participle clauses may occur, as an overlap between different events,
actions, or states can be expressed by the conjunction tonoana (cf. (1041a)) The
functional difference between the two clauses is that subordinate clauses are
more explicit than participle clauses, especially as the latter do not formally
specify which of the possible functions applies. The choice between reference
to overlap, sequence, or modification entirely depends on the context, while a
subordinate clause with a conjunction specifically has the function as expressed
by the conjunction. In the example (1127), a participle form is used parallelly to
a subordinate clause, which illustrates the similarities and differences between
the two. The speaker starts with a participle clause (DEP1) and then adds a
subordinate clause (DEP2).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
812 Multi-clause constructions

(1127) Combination of participle and subordinate clause

ajßa räasa-i, ku räasa-uru-a hau, päe

[with dance-PRT] [there dance-PL-3ps/D because] [already]
[DEP1] [DEP2] [ADV]

hja-ri-te
[urinate-RAP-INTS:3sg/E]
[V]
'They danced with her, and as they were dancing there, she suddenly had to urinate.'

There are two possible reasons for using this kind of structure:
a) A semantic distinction may be implied, as the DEP1 could mark the overlap
of the actions 'dancing' and 'have to urinate' - which is not expressed by the
subordinate clause.
b) The addition of a subordinate clause is necessary, because there is a change
of subject between the dependent clauses and the main clause. Again, the
switch can only be signalled by a finite clause - DEP2. Thus, the insertion of
DEP has both semantic and functional significance.
Complement clauses differ from participles in a mainly functional manner.
Same subject complement clauses, apart from bearing a different morphological
marking than participles, have a strictly complementational function, which
excludes any interpretation as sequence, overlap, or modification. The same
functions apply to different subject complement constructions, with the
additional difference in reference.
Juxtaposition as a coordinating strategy is relatively rare and the few attested
examples all refer to the same subject in each finite clause. The emphasis here is
on the listing effect of events, which could be interpreted as implying overlap of
states or events/actions. More data would be needed in order to determine the
subtle differences between finite clause juxtaposition and participle construction
with this function. The comparison between different constructions that involve
more than one verb is summarised in table (1128).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Multi-clause constructions in comparison 813

(1128) Comparison of different multi-clause constructions

Construction type Subject Morphology Functions

reference
SVC same Non-finite, Directional, resultative,
(Note: 1 clause) citation form purposive
Participle same Non-finite, Sequence, overlap,
Participle -7 modification
SS complement same Non-finite, Complementation
clauses infinitive -na
Juxtaposition of (same) Finite, E-/A-form Enumeration of events
finite verbs
Coordination with same / Finite, any form Different
conjunctions different events/actions/states
Subordination with same / Finite, D-form Conditional, temporal
=ne different
Subordination with same / Finite, D-form As expressed by
conjunction different conjunction
DS complement different Finite, D-form Complementation
clauses

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
21 Questions

Urarina questions can be divided into several types, each of which shows
specific features. Content questions (cf. §21.1) contain an interrogative
pronoun, which is obligatorily fronted to the clause-initial position and can be
marked for emphasis in addition to that. As mentioned in § 11.2, content
questions require the use of the Ε-form on the verb. In contrast, polar questions
always occur with the Α-form. Interrogative marking for this question type is
realised through intonation and by the insertion of the enclitic =na, as is
investigated in §21.2. Negative questions, which are discussed in §21.3, are
identified through the use of a special introducer and the attachment of a final
negative marker in clause-final position. Rhetorical questions represent a further
type of question, as these are specifically marked with the enclitic =te in clause-
final position and require the use of the Α-form, as is demonstrated in §21.4. In
contrast to these different types, there is no formal way of distinguishing direct
from indirect questions, as in this case, the two clauses occur as independent
clauses in juxtaposition (cf. §21.5). - Another interesting feature related to
questions is the existence of interrogative verbs. These are investigated in
§21.6.

21.1 Content questions

As stated in §5.9, Urarina has a basic interrogative pronoun d$a from which
most other forms are derived. Note that many of the interrogative pronouns that
contain d^a in some way can also be used as indefinite pronouns (cf. §21.1.11).
The interrogatives for 'how much', 'how many' are verbs, as will be
investigated in §21.1.10. The position of all interrogatives is clause initial.
Optionally, they can be followed by a focus marker. For convenience, an
overview of the different interrogative forms is given in (1129).

(1129) Interrogative pronouns and derived forms

d$a ' w h a t ' , ' w h o ' ; cf. §21.1.1

d$a-rihi 'which' (with verb root 'be like'); cf. §21.1.3
d^a-nu 'why' (with purposive suffix); cf. §21.1.4
d$a-toane-l 'how' (with verbal derivation); cf. §21.1.5
d^a-sit 'how' (with unidentified derivation); cf. §21.1.6
d^a-bana 'when' (with postposition/conjunction 'when'); cf. §21.1.7
dyu 'where' (with LOC suffix); cf. §21.1.8

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 815

d$a-elöo 'what way' (with DIR suffix); cf. §21.1.9

paaohwaa 'how many', 'how much'; cf. §21.1.10
ßatahaa 'how many', 'how much'; cf. §21.1.10

It can be assumed that some of the interrogatives listed in (1129) are lexicalised.
For instance, this is the case with d$u 'where': in this combination of d$a with
the locative suffix -«, the vowel /a/ is elided - which does not correspond to any
phonological rule. In fact, the combination of this suffix with the demonstrative
kaa results in the form kau for 'here'. Correspondingly, a combination with the
interrogative pronouns would be expected to surface as *d$au - which does not
exist. Thus, the form d$u clearly is a lexicalised form, while for the other
combinations, the evidence is not as clear as that. In the following sections, the
use and function of each of the interrogatives will be investigated.

21.1.1 "d^a"'what', 'who'

The interrogative pronoun d$a can refer to human or non-human subjects,

meaning ' w h a t ' or ' w h o ' , respectively. There is no formal distinction for that.
Due to the fact that all interrogatives occur in initial position and since Urarina
has no morphological case marking, d$a can also refer to ' w h o m ' or 'whose'.
Some ambiguities can only be resolved by the context; however, other clues
may be provided by the transitivity of the verb involved or by the presence or
absence of a focus marker. For instance, in (1130a), the only feasible
interpretation of d$a is as a subject with human reference, for two reasons: a)
the verb 'arrive' is intransitive and thus the occurrence of an Ο argument is
excluded; b) though an interpretation as referring to a non-human subject cannot
be entirely excluded ('what has arrived?'), reference to a human subject is much
more likely, as the action of arriving is usually performed by animate, if not
human agents. Also note that in example (1130a), no focus marker is involved,
as interrogative pronouns obligatorily occur in front position. Optionally, a
focus marker may be inserted to express emphasis or contrast, such as in
(1130b) (for the functions of the focus marker, see §19.1).

(1130) Interrogative d$a with subject function and human referent

a) Without focus marker:
d$a turu-e
what arrive-3ps/E
'Who has come?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
816 Questions

b) With focus marker:

d$a=te nii kahe kana ruku-era ne-t ne=lu
what=FOC that from lpl/in pull.out-AG be-PRT be:3ps/E=REM
'Who is the one to pull us out of that?'

The example in (1131) illustrates the reference of d$a to an Ο argument with

human reference. The interpretation as human or non-human reference can only
be made through the context. Also note that the example does not contain a
focus marker; in fact, the occurrence of a focus marker on with object
function would be rather unusual, as focus marking on objects is statistically
much less frequent than focus marking on subjects. It is, however, a possible
option.

(1131) Interrogative d$a with object function

d$a kwara-i tahia
what see-2ps over.there
'Whom/what have you seen over there?'

In the examples in (1132), d$a functions as an Ο argument with non-human

reference. Most typically in this function, the interrogative pronoun occurs
without any additional marking, such as illustrated in (1132a). Emphasis can be
expressed by the attachment of either the interrogative marker =na or a focus
(such as =te), with no apparent difference in function. While the difference
between the two markings in questions is not entirely clear, the variation is
probably based on preferences in the traditional vs. innovative language.
In (1132b), emphasis is expressed by the clitic =na, and in (1132c),
emphasis is marked through the focus marker =te, here attached to the
reportative evidential =he.

(1132) Interrogative d$a with object function and non-human referent

a) Without emphasis:
d$a kurete-i na-1 baha-anu
what buy-2ps say-PRT ask-lsg/A
"'What have you bought?" I asked him.'

b) With emphasis expressed by =na:

d^a=m itca-kure nii rihihel nii ka=iri-uru
what=INT do-3pl/E that like that 1 sg=creature-PL
'What [the heck] have these creatures of mine done?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 817

c) With emphasis expressed by =na:

d$a=he=te itQa-i mtokoanel kanaanaj-uru rela-l
what=REP=FOC do-2ps like.that child-PL teach-PRT
'What [the heck] are you doing, teaching [my] children in this manner?'

Ambiguity may occur in many examples with d$a and a transitive verb; the
ultimate decision on the applicable interpretation may only be determined in the
light of the context. As mentioned in §19.3.1, the presence of a focus marker
usually gives a clue as to how a question with d$a is to be interpreted. When a
focus marker is present (cf. (1133a)), an interpretation of d$a as referring to a
subject is more likely; when there is no focus marker (cf. (1133b)), the object
reading may be preferred.

(1133) Ambiguity of d$a

a) Default = subject reading
d$a=te itga-re=~i
what=FOC do-IRR:3ps/E=ASS
'What would he do?' / 'Who would do it?'

b) Default = object reading

d$a itQa-re=t
what do-IRR:3ps/E=ASS
'What would he do?' / 'Who would do it?'

In some examples, d$a is attested with a modifying function, i.e. as a

prenominal modifier. This is illustrated in (1134), where the interrogative
pronoun has the meaning 'which', 'what kind o f . This function is otherwise
(and preferably) realised by d^arihi (cf. §21.1.3); While the lack of further
examples makes it difficult to define the functional difference between the two,
it appears that d^arihi has specific reference, i.e. 'which' out of a limited
selection of candidates, whereas d$a may imply a non-specific reference. The
examples in (1134), for instance, imply a general lack of knowledge by the
speaker about the possible choices involved in the respective questions.

(1134) d$a with modifying function

a)
d$a fwetga ke ne-ri-tg,a=! kurso na-ι baha-u
what date VLI be-IRR-3ps/A=ASS seminar say-PRT ask-IMP

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
818 Questions

herardo kahe
PSN from

'Ask Gerardo on what date will be the seminar!'

b)
d$a barasane ke=te kauatpa-i omaro-i
what dew INST=FOC good-PRT grow-PRT
'From what dew [will the plants] grow nicely ...?'

21.1.2 "d^a " with other postpositions

The interrogative d$a is also used in combinations with a variety of

postpositions, such as raj ' f o r ' , kahe ' f r o m ' , and the instrumental/valency-
increasing postposition ke. The use with raj may imply possessive function, as is
shown in the verbless clause in (1135a), where raj functions as a possessive
marker following the possessor (cf. §7.1.2). In (1135b), the function of raj is of
a beneficiary nature and cannot be defined as having possessive function.

(1135) Interrogative d$a combined with raj

a) Possessive function:
d$a raj ka=teru
what POSS this=axe
'Whose axe is this?'

b) Beneficiary function:
d$a raj aj-aka nerajjie
what for AUX-lpl/du must
'For whom must we do it?'

With kahe, a directional meaning 'where from' is expressed (cf. (1136)). Note
that in this case, not the interrogative pronoun d$u 'where' (cf. §21.1.8) co-
occurs with the postposition, but d$a, whereas the locational function is only
expressed by kahe. The use of d$a kahe in (1136b) implies a question about the
source or origin of the child, thus the locational component can also be used in a
figurative sense.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 819

(1136) Interrogative d$a combined with kahe

a)
d^a kahe=te uu-re akau
what from=FOC come-IRR:3ps/E water
'From where would the water come?'

b)
d$a kahe=te kanaanaj eru-i=ta
what from=FOC child find-2ps=FRS
'Where have you found the child?'

The use of the postposition ke with d^a can have instrumental or valency-
increasing function, depending on the context in which it is used (cf. §16.5).
Example (1137a) illustrates the instrumental function ('with what'), whereas
(1137b) refers to an increase in valency ('what for' in this example). Also note
that emphasis can be expressed through the use of the interrogative enclitic =rw
(as in (1137a)), or by inserting the focus marker, such as in (1137b). Again, the
variation is based on preferences of younger vs. older speakers.

(1137) Interrogative d$a combined with ke

a) Instrumental function:
d$a ke=na itQua-h
what INST=INT tie-2ps
'With what have you tied it?'

b) Valency-increasing function:
d$a ke=te ne-laate-re nii hjane koloromi=ta
what VLI=FOC ITR-need-IRR:3ps/E that urine useless=FRS
'What would that useless urine serve for?'

21.J.3 "d^arihi" 'which'

As indicated in §21.1.1, the form d^arihi for 'which' (probably) implies a

specific referent out of a limited selection. This is illustrated in (1138a), where
'which (one)' is clearly limited to the small number of persons present at the
scene. (The sentence is from a narrative in which a small group of Urarina

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
820 Questions

people take revenge at a group of "Indios".) 99 In (1138b), again, the choice of

people who could be candidates for chiefdom is restricted to a few.

(1138) Use of d$arihi

a)
ku akuuhua itQua-hakua-uru-a hau, he, üra d$arihi=te=ra
there entrance tie-DSTR-PL-3ps/D because hey ERD which=FOC=EMF

najße-re, bakaua rem eno-a ku-a-ne-ta

be.able.IRR:3ps/E Indio place enter-NTR go-3ps/D=CND=FRS
'As they [the Indios] tied their doors, [the others said], "Hey, which [of us] would be
able to enter the Indios' place?'

b)
nii baja hvituku-era-i d^arihi tQäiti kureram-a
that after know-IMPS-NEG:3ps/A which ever replace-NTR

eno-re=i tenjente gobernadoro ne-i

enter-1 RR: 3 ps/E=A S S village.chief be-PRT
'After that, it is unknown which [person] ever would come to be village chief.'

As is obvious, the word d$arihi is composed of the interrogative pronoun d^a

and the verbal root /rihi/ for 'be like'. Thus, the literal meaning of d^arihi could
be translated as 'be like what', referring to the specific kind of the noun in
question.

21.1.4 "d^anu" 'why'

The interrogative for 'why' is d$anu. The component / η « / also occurs in the
purposive marker (haa)nu ke (cf. §20.2.4.1). d$anu is always followed by either
the focus marker =te or the interrogative enclitic =na. These two markers are in
principle interchangeable after d^anu, the variant with the focus marker is a
feature of the traditional language, whereas younger speakers prefer the variant
with =na. However, even with older speakers, there is a high degree of variation
between the two forms. Examples for each occurrence are given in (1139).
d$anu can be used in positive as well as in negative questions (cf. (1139b)).

99
The term bakaua 'Indio' has a negative connotation in Urarina and is usually
associated with hostile and cannibalistic groups of indigenous people.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 821

(1139) Use of'why'

a) With interrogative marker:
esiflae, d$anu=na ka=rerueka-i=tQe
really why=INT 1 sg=disturb-2ps=PLT
'Really, why do you disturb me?'

b) With focus marker:

d$anu=te k=ere tonora-ene-i=ta
why=FOC lsg=word obey-NEG-2ps=FRS
'Why don't you obey me?'

21.1.5 "d$atoawi" 'how'

Corresponding to the intransitive verbs katoama and nitoama ('be like this'; 'be
like that'), which are combinations of demonstratives with the root /toane/,
there also is a verb d^atoama 'be how' (also cf. 21.6). Interestingly, the root
/toane/ itself does occur not in isolation, nor is there a verb *toania. However,
the combination of this root with the interrogative, resulting in d^atoane-ι, is a
participle form. Since the occurrence of d^atoanei is overwhelmingly frequent
as opposed to examples which involve the verb with any other kind of
inflection, one may assume that the participle form is being lexicalised and used
as an adverb, similar to many adverbs that contain a participle suffix. The
examples in (1140) illustrate the use of d^atoanei.

(1140) Use of d^atoanei

a)
d^atoanei torn kaa makina
how sound:3ps/E this machine
'How does this machine work?'

b)
d^atoanei kann auna-i=ßa
how 1 sg hear-2ps=INT
'How do you hear me?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
822 Questions

21.1.6 "d^asü" 'how'

Another interrogative referring to 'how' is d^asti, in which the meaning of

second part is not entirely clear. Note, however, that there is some phonological
resemblance with the interrogative verb ßasia and its participle form ßasii,
which is unlikely to be a coincidence (cf. §21.6). In fact, ßasii is a variant of
d^asii occurring in a few isolated examples. It is unclear whether there is any
difference in meaning between the two; however, d^asii is the more commonly
used form.
On the semantic level, there is a subtle difference between d^asii and the
other word for 'how' d^atoanei. While d^atoanei refers to the function, context,
or type of situation, d^asii specifically asks for a background or possibility,
which may be translated to 'how come?', 'why', or 'how is it possible that ...'.
It may also imply reference to the source of power which made the respective
event possible to happen. This is clearly expressed in the examples below: in
(1141a), the husband of a woman is surprised about the sudden presence of a
child and questions his wife about its origin. In (1141b), the question with
'how' again expresses astonishment about the unexpected end of a fig tree,
which had been cursed. In contrast, d^atoanei specifically asks about the
manner in which a result was achieved, as to mean 'in what manner (is it
realised)'. - Similar to d^anu for 'why', d$asü is always accompanied by =te
or =na, which both are used to express emphasis or contrast. 100

(1141) d^astt 'how'

a) With focus marker

ku kanaanaj kuruaae-ka hau, d$asü=te
there child be.happy-3ps/D because how.come=FOC

katoane-j itQa-i=ta
be.like.this-NOMsbj do-2ps=FRS
'As he was happy about the child, [he said], "How is it possible that you did something
like this?'"

b) With interrogative marker:

d$asu=β a luekutarn u-e
how.come=INT suddenly die-3ps/E
'How is it possible that [the fig tree] died so suddenly?' [NT: Matthew 21:20]

100
While it would be plausible that, parallel to the variation occurring with d$anu, the
focus marker is a feature of the traditional language, it is unclear how the two
alternative markers might differ.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 823

21.1.7 "d^abana" 'when'

The interrogative form for 'when' is a combination of d^a with the word bana,
which is otherwise used as a conjunction 'when' and as a postposition meaning
'at the time when' (cf. §5.7.3). The use of d$abana is illustrated in (1142). It
can be used with reference to past, present, or future.

(1142) Occurrence of d^abana

a) Past reference:
d^abana turu-e
when arrive-3ps/E
'When did he arrive?'

b) Future reference:
d^abana nakwaaunei kau turu-ri-ki=i
when again here arrive-IRR-2ps=ASS
'When will you come here again?' [NT: Matthew 24:3]

21.1.8 "d^u"'where'

d$u can have locational or directional meaning. The distinction depends on the
verb, as is illustrated below. In (1143a), the motion verb ' g o ' implies a
directional reference of d$u, whereas the verb 'sleep' in (1143b) assigns a
locational function to the interrogative.

(1143) Use of ^3«

a) Direction:
d$u ku-e
where go-3ps/E
'Where did he go?'

b) Location:
d$u sim-tge
where sleep-IRR:3ps/E
'Where would he sleep?'

The interrogative d^u can also be combined with tial, which has locational
function as a postposition, but temporal reference when occurring in reetiai (cf.
§5.7.2). The combination d$u tial can have either function, which is surprising,
as one might expect a locational function of d$u, which contains the locative

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
824 Questions

suffix -u (at least from a diachronical perspective). In (1144a), a locational

interpretation of d$u tiai is given indeed, whereas (1144b) exemplifies its
temporal function. There are no formal features that would predict which
interpretation is applicable; thus, it must be assumed that the distinction
depends on the context.

(1144) Interrogative d^u combined with tiäi

a) Locational:
d$u tiäi kweta-re-ü
where until clear-IRR-lsg/E
'Up to where would I clear it?' [asking about what parts of the field he has to clear]

b) Temporal:
d$u tiäi kaa iri-uru suu-re-1 ofwa
where until this creature-PL kill-IRR-2ps.ASS Dad
'Dad, ("until") when are you going to kill these creatures?'

21.1.9 "dyteloo " 'what way'

Another combination of d$a with a locational element is the combination of d$a

with the directional suffix -elöo, which was discussed in §5.6.2. Literally, the
form d^a-elöo would mean 'where to'; however, in the examples attested in the
database, the word is not used with interrogative function at all, but occurs only
in a figurative sense and always with a negative form of the verb. Thus, it does
not function as a directional interrogative pronoun, as could be expected, but
with a kind of quantifying function. For instance, in (1145a), d^aelöo means
'what(ever) way', 'in any way' and does not refer to location or direction. This
interpretation is also given in (1145b), where negation is expressed by the
privative suffix on the verb: 'without leaving it in any way'.

(1145) Use of d^aelöo

a)
fladera-eri-Ί aj-aka=m nijej d^aeloo kam-a-ka
be.sad-DIM-PRT AUX-lpl/du=CND not.at.all what.way win-NEG-lpl/du
'If we do it sadly, we cannot win in any way.' [About playing soccer; lit. "we do not
win in any direction"]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 825

nii raj d^aeloo hiane=lanaala kana=hvaaun-era baha-aka=ne kujßa

that for what.way leave-PRV our.creator ask-lpl/du=SUB so.that
'[God said so] so that we would pray without intercession.' [Lit. 'without leaving it in
any direction']

21.1.10 'How much'how many'

The terms for 'how much' do not belong to the class of interrogative pronouns.
jiatahaa and ßaaohwaa are verbs, as the examples in (1146) and (1147)
illustrate. With the participle suffix -7, they are adverbialised and function as
interrogatives. The behaviour of these verbs corresponds to that of the native
numerals one to five, which are verbs as well (cf. §5.12). The functional
difference between the two is not entirely clear, but may be related to the
amount to which they refer (see below).
The examples below show the verb ßatahaa suffixed with the participle
suffix. In (1146a), the reply to the question 'how many' contains the numeral
verb for 'three', which also occurs with the participle form. The question in
(1146b) contains the interrogative pronoun d^a in addition such as to inquire
not only about the quantity but also about the object.

(1146) Interrogative verb ßatahaa

a)
ßataha-Ί siiri-i - nitg,ataha~i siiri-ü
be.how.much-PRT have-2ps three-PRT have-lsg/E
'How many do you have? - I have three.'

b)
ßataha-Ί d$a kurete-i
be.how.much-PRT what buy-2ps
'How many of what have you bought?'

Similarly in (1147), ßaaohwaa occurs with the participle form. Note, however,
while this form is the one most commonly occurring, both verbs may also be
marked with person inflection suffixes, such as with lpl to mean 'how many are
we?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
826 Questions

(1147) Interrogative verb ßaaohwaa

a)
ßaaohwa-i arusu here-i
be.how.much-PRT rice want-2ps
'How much rice do you want?'

b)
jiaaohwa-Ί ka=raj kuriki te-naa najße-U-toe,
be.how.much=GER lsg=for money give-INF be.able-2ps-PL

β aara raj hesoso bi-anu=ne kujßa=ta

2pl for PSN announce-lsg/D=SUB so.that=FRS
'How much money can you give me so that I would deliver Jesus to you?' [NT:
Matthew 26:15]

Alternatively, as a feature of the innovative language, the roots of these verbs,

ßaaohwa and ßata(ha), can be used with interrogative function, which
indicates that the absence of quantitative interrogatives is being replaced by a
system similar to the interrogative structure of Spanish. In this form, they can be
accompanied by the copula ('being how much'; cf. (1148a)); alternatively a
participle marker is attached to the word for the measuring unit, such as 'kilo' in
(1148b). The example in (1148c) shows the occurrence of an interrogative verb
root with a finite form of the copula.

(1148) Use of interrogative verb roots

a)
ßaaohwa kilo ne-i
how.much kilo be-PRT
'how many kilo?'

b)
β ata kilo-i here-i
how.much kilo-PRT want-2ps
'How many kilos do you want?'

c)
ßaaohwa ne-i-t^e - mtgataha-akaanu
how.much be-2ps-PL three-lpl/ex
'How many are you? - We are three.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 827

Apart from the developing changes in function and use, a major challenge is the
distinction between the two terms used for 'how much'. Both words are
interchangeable for 'how much' or 'how many' and the judgements made by
different speakers appear to be contradictory. While some seem to prefer
ßaaohwaa for uncountable items, it is clearly used with countable items as
well. A possibility is that a former distinction referring to the feature [+/-
countable] or to individuality is being neutralised at the present state of the
language. Manus (2002 p.c.) thinks that ßaaohwa could be related to an
approximate quantity or value, as opposed to a more restricted way to ask for
quantity with βatahaa. As terms for 'how much' have little relevance in
traditional Urarina social life and are rarely used in natural language, it is
difficult to determine the difference at this stage. The preferred way to ask for
'how much' in terms of money is to ask for 'its exchange' (='price'), raj hire,
which can be used with d^a 'what'.

21.1.11 Indefin ite function

As mentioned in §14.3, interrogatives can also function as indefinite pronouns.

Most commonly, this occurs in negative clauses; however, the example in
(1149) illustrates that d$a may also function as an indefinite pronoun for
'something' in affirmative clause.

(1149) Interrogatives as indefinite pronouns in affirmative clause

ni-a d$a=ne, e7eo-käu=ne, tQäe u-u=tQe
be-3ps/D what=CND shout-lsg/D=CND also come-IMP=PLT
'If there is something [in the trap], when I shout, also come!' [Talking to son-in-law]

The occurrence of different interrogatives as indefinite pronouns in negative

clause is exemplified in (1150), which illustrates the occurrence of d$a 'what',
d^abana 'when', and d^u 'where' in this context. Note that in (1150c), the
adverb rdjej 'not at all' is used to emphasise the negation.

(1150) Interrogatives as indefinite pronouns in negative clause

a) With d$a:
päe d$a laate-kuru-i nesari-tQuru-a=na hau
already what need-PL-NEG:3ps/A hunt-PL-3ps/D=SUB because
'They already do not need anything to hunt.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
828 Questions

b) With d$abana:
hajti d^abana bakaua so-eni-a häu
still when Indio kill-NEG-3ps/D because
'As he [the boy] had never killed an Indio before, ... [his father was worried].'

c) With d$u:
nijej d$u ku-a-ü
not.at.all what go-NEG:lsg/A
Ί did not go anywhere.'

In the following example, d$a is used as a negative quantifier with modifying

function to mean 'nothing good' (where 'good' functions a noun).

(1151) d$a as a negative quantifier

d$a kauat^a ni-ji=lu

what good be-NEG:3ps/A=REM
'There was nothing good.'

There are two suffixes which can be combined with an interrogative pronoun to
refer to indefiniteness. The suffix -elej can be attached to to express an
indefinite pronoun for 'something', implying a non-human referent. The
resulting form d^aelej also has the variants d^aalej and d^ale. The form -ne
(which is only found in this context, including the variants -nej and -m) can be
attached to the interrogative d$u to mean 'somebody', always implying human
reference. Possible realisations of this pronoun include d$une, d^unej, and
d$um. Examples for both forms are given in (1152). Interestingly, the form
d^une does no have any connotation related to location, as one might expect
from an inherently locational interrogative ('where').

(1152) Indefinite pronouns d^aelej and d^une

a) d^aelej:
kana raj d^aelej be-ra ni-a=ne ke haaeko-aka
lpl/in for something tell-AG be-3ps/D=CND VLI look.forward.to-lpl/du
'When there is someone who promises us something, we look forward to it.' [Lit. 'when
there is an announcer of something for us']

b) d$une, d^aelej:
d^urd d$ale laati-a hana, nijej te-ni-aka=ne
somebody something need-3ps/D when not.at.all give-NEG-lpl/du^CND
'When someone needs something and we do not give [him anything] at all, ... [what
kind of behaviour is that?]'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Content questions 829

The indefinite pronoun for 'nobody' is negated in negative clause to mean

'nobody', accordingly. Optionally, the adverb nijej may be inserted for
emphasis.

(1153) Indefinite pronoun d^une in negative clause

a)
nijej d$uni kwara-?a-u
not.at.all somebody see-NEG-lsg/A
Ί did not see anybody at all.'

b)
akuuhua hwu-u eno-a d^une=ne huataa
entrance close-IMP enter somebody=SUB so.that.not
'Close the door so that nobody would enter!'

The negation of d^aelej is not attested in the same way, but it occurs in a
slightly different manner with negation. As illustrated in (1154b), this form
functions as a negative quantifier within an NP where it precedes the head noun.
Note that examples for this are attested in negative clause only.

(1154) Indefinite pronouns d^aelej as negative quantifier

a)
nijej d^alej katga
not.at.all somebody man
'No man was there.'

b)
m-ji kauatQa najjiehei, m-ji enua,
be.NEG:3ps/A pretty.nothing be.NEG:3ps/A tree

ni-ji d^aelej nuna

be.NEG:3ps/A something bush
'There was pretty much nothing, no tree, no bush.'

In addition to d^aelej and d$une, there are two related forms, expressing
indefinite pronouns for 'anything' and 'anywhere', respectively. The form
d^umhju, can be analysed as a combination of d$une followed by the locative
suffixes -hia and which is merged to -hju (in this word only; for more
detailed analysis of the locational suffixes, cf. §5.1.2 and §5.6.2, respectively).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
830 Questions

(1155) Forms d^umhju and d$alejjiehei

a)
d$umhju kuri eru-i=jie ke im-ni-u=ra
anywhere jagua.treQ find-2ps=CND VLI go.up-DSTL-IMP=EMF
'If you find a jagua tree anywhere, climb it up!'

b)
d$alejßehel ari-a ku-e, ne-ni-a lenone=ne hau
something seek-NTR go-3ps/E be-NEG-3ps/D food=SUB because
'He went to look for something, because there was no food.'

Another word that is composed of an interrogative pronoun and a suffix is the

indefinite pronoun d^unekoa (or variant d^umkoa) 'somebody'. It can be used
as head of NP (cf. (1156a)) or as a modifier, which is shown in (1156b).

(1156) Indefinite pronoun d^unekoa

a) As head of NP:
ta m-a d^unekoa kunajte-naa=ne
NEGQ be-3ps/A someone be.sick-NOM=NEGF
'Is there nobody who is sick?'

b) As modifier:
d$unekoa kanaanaj-uru ne-he-t su-ι ki-tQe=i=tQa
someone child-PL be-CNT-PRT kill-PRT eat-IRR:3ps/E=ASS=EMF
'He will kill someone of the children and eat him.'

21.1.12 Indefinite emphasiser

After some interrogatives that occur with the function of indefinite pronouns,
the word tQaiti can be inserted to mean 'ever', implying emphasis on the
indefinite meaning of the pronoun. For instance, d$a tQaiti is understood as
'whoever' or 'whatever'. This form also regularly occurs with d^u and d^arihi,
for which examples are given in (1157). Further combinations (not exemplified
here) are attested with d^abana, d^anu, d^atoarm, with d^a plus postposition,
and with äaka. Note that teiti is a variant used in the Chambira dialect.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Polar questions 831

(1157) Use of tQäiti

a) With d$a:
d$a tQäiti leotQa raj m-a=m ke kam suuhua
what ever other for be-3ps/D=CND VLI lpl/in heart

akati-tQaa nerajtene
defend-lpl/du must
'If there is anything whatsoever against another [person], we must forgive 'that
person].' [Lit. 'if there is anything "for" another']

b) With d$u\
itulere u-elanaala ne-nakauru kuruatahane-u, d$u t^aiti ne-ure
all.kinds come-PRV be-those.who help-IMP where ever be-3pl/E
'Help all those who have not come, wherever they are.'

c) With d^arihi:
d^arihi tQäiti kioara-e
which ever see-3ps/E
'Which one (whatsoever) has he seen?'

21.2 Polar questions

Polar questions are marked with the interrogative enclitic =na, which is attached
to the verb in non-focus clauses, but to the first constituent in focus
constructions. The constituent order in polar questions is identical to that of
declarative clauses. In addition, a rise in intonation towards the end of the
question can be observed. As mentioned in §19.2.2, focus markers do not occur
in polar questions. However, emphasis can be expressed by the insertion of =na
after the first constituent. The examples in (1158) illustrate the regular position
of the enclitic. In (1158a), the verb to which =na is cliticised occurs after an
adverb and the Ο argument. In (1158b), the verb follows a dependent clause.
Note that in this example (from the contemporary language), the complement
clause is realised with häu, which formally makes it identical to a subordinate
clause (cf. §20.1.2.2).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
832 Questions

(1158) Interrogative =na after verb

a)
hanone mesahe auna-i=pa
morning message hear-2ps=INT
'Did you hear the message in the morning?'

b)
mi kujßa kana ra-a kana+huaaun-era=ne hau koituku-akat<?e=na
that so.that lpl/in receive-3ps/D our.creator=SUB because know-lpl/in=INT
'Do we [all] know that God has accepted us for that [purpose]?'

A verb to which =na is cliticised does not necessarily occur in clause-final

position. The order verb - interrogative clitic is also preserved when a subject
noun follows the verb, such as in (1159a), where the intransitive verb 'come',
followed by =na, occurs before the S argument 'rain'. In (1159b), the verb is
followed by the adverb 'tomorrow'.

(1159) =na after verb with VS order

a)
u-ri-tQa=i=β a elo
come-IRR-3ps/A=ASS=INT rain
'Will it rain?'

b)
turu-ri-tQa=i=β a eresi
arrive-IRR-3ps/A=ASS=INT tomorrow
'Will he arrive tomorrow?'

When a constituent is in focus (in terms of the general emphatic function of

focus as discussed in §19), it is marked with the interrogative enclitic; focus
markers cannot occur in polar questions. Thus, example (1159b) could also be
realised as eresi=ßa tururit$äi, in order to put emphasis on the adverb ('Is it
tomorrow that he will arrive?'). Further examples are given in (1160). In
(1160a), the adverb 'really' is marked with =na; in (1160b), the enclitic follows
a conditional subordinate clause. Example (1160c) is elicited, as it is uncommon
to mark an Ο argument for emphasis in a polar question. Thus, the sentence
implies contrastive focus on 'monkey', such as to inquire whether the addressee
has this particular preference, as opposed to other food mentioned before in
discourse.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Polar questions 833

(1160) =na after first constituent

a) After adverb:
esißae=na amu-a umberto
really=INT walk-3ps/A PSN
'Did Humberto really go?'

b) After subordinate clause:

itgafwa ajßa ni-a=ne=na su-hwe-re-i modo
rifle with be-3ps/A=CND=INT kill-DIM-IRR-2ps PSN
'Modo, if you had had a gun, would you have killed it?' [Lit. 'If there had been (a
situation) with a rifle, would you have killed it easily?']

c) After Ο argument:
enejtQu=na kohwaite-i
monkey=INT desire-2ps
'Is it monkey that you want?' [Implied: 'not any other food?']

The same position of the interrogative enclitic is also observed in verbless

clauses. In (1161), =na occurs after the copula complement, which is the first
constituent of the clause. The copula subject appears in final position.

(1161) =na in verbless clause

raj nuhue=na mi
POSS excrement=INT that
'Is that its excrement?'

Mainly as a feature of colloquial language (and possibly more commonly used

in the innovative variety), omission of the interrogative marker can also be
observed in some examples. In this case, the question is marked by rising
intonation only. Both examples in (1162) are from the innovative language.

(1162) Omission of =na in polar questions

a)
radiofwonia kujßa karta letoa-a baaba=ne hau te-i
shortwave.radio so.that letter send-3ps/D Dad=SUB because give-2ps
'Have you submitted the letter Dad sent in order [to get] the shortwave radio?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
834 Questions

b)
nahari ki-i tQu mid tako
liver eat-2ps CRTN ADVRS PSN
'But did you actually eat liver, Tako?'

21.3 Negative questions

The formation of negative questions was already discussed in §14.5. This clause
type is doubly marked: firstly, by the negative question introducer ta in initial
position and secondly, by the final negative marker =ne in clause-final position
(slot 23; for the exact position of =ne, see §12.3.6). The inflection on the verb is
affirmative, i.e. negation is not marked on the verb itself, but only through the
accompanying introducer and final negative enclitic.

(1163) Negative questions with ta

a)
ta kure kivitukH-i=jie
NEGQ price know-2ps=NEGF
'Don't you know the price [of aguaje\T

b)
ta nahari heri-a ka=d^uhu-he-j=jie
NEGQ liver want-3ps/A this=be.standing-CNT-NOM sbj =NEGF
'The the one who is standing here not want any liver?'

Negative content questions are not attested in the database, but can be formed in
elicitation. In this case, the same structure as in content questions applies, but
with negative inflection on the verb. Optionally, the interrogative pronoun can
be accompanied by the adverb mjej 'not at all', which can precede or follow the
pronoun. The regular devices for negative questions, ta and =rte, do not occur in
this construction. Examples for the occurrence with 'what/who', 'where', and
'why' are given in (1164).

(1164) Negative content questions

a) With 'what':
d$a (mjej) u-ene
what not.at.all come-NEG:3ps/E
'Who has not come?'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Rhetorical questions 835

b) With 'who':
d$a (rtijej) kwara-ene-i
what not.at.all see-NEG-2ps
'What did you not see?'

c) With 'where':
(rtijej) d$u foi-ri-a-akaa=m
not.at.all where go-IRR-NEG-lpl/du=ASS
'Where will we not go?'

d) With 'why':
d^anu=m u-na here-kur-ene leotQa-uru
why=INT come-INF want-PL-NEG:3ps/E other-PL
'Why do the others not want to come?'

When an interrogative pronoun occurs in a negative question with ta, it is

automatically interpreted as an indefinite pronoun, such as in (1165). It should
also be noted that negative questions cannot occur in dependent clause.

(1165) Negative question with indefinite pronoun

ta tQäe d^a kwituku-i=jie
NEGQ also what know-2ps=NEGF
'Also, don't you know something [about the letter Dad sent to the council]?'

There also is an isolated example which may represent an innovative strategy

for negative question formation. In (1166), the (polar) negative question is not
marked with ta, but negation is expressed on the verb itself. In addition, the
clause is introduced by nijej, similar to the negative content questions described
above.

(1166) Negative question with nijej

nijej kaa baja eresi mesahe itQa-na najße-ri-e-i=m
not.at.all this after tomorrow message do-INF be.able-IRR-NEG-2ps=ASS
'After this [conversation], will you not be able to send a message tomorrow?'

21.4 Rhetorical questions

The rhetorical question marker =te (slot 24) has already been discussed in
§12.3.7. As stated there, it is related to surprise or curiosity in a question. The
clitic =te can be attached at the end of polar or content questions. Two examples

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
836 Questions

are given in (1167), in addition to the ones listed in §12.3.7. The co-occurrence
with the interrogative marker =na is very common. As in other questions, =na
may be attached to the verb (1167b), or to the first constituent (such as in
(1167a)). Note that in content questions, =na is obligatorily attached to the
interrogative pronoun. In polar questions, the insertion of =na is optional. The
polar question in (1167b) occurs without the interrogative marker, whereas it is
inserted in (1167c). Note that in this example, the verb is also marked for
negation.

(1167) Rhetorical questions with =te

a) With content question:

d$a=na kaa kuane ne=te
what=INT this inside be:3ps/E=RHT
'What would be inside this?' [To imply: 'what kind of interesting thing']

b) With polar question:

ruru kwaasi-i=te
howler.monkey be.afraid.of-2ps=RHT
'Are you afraid of the howler monkey?'

c) With polar question and =na\

kaa kwaasi-ji=jiaa=te
this be.afraid.of-NEG: 3ps/A=INT=RHT
'Is he not afraid of this one?'

As a further regularity, polar rhetorical questions always co-occur with the A-

form, whereas rhetorical content questions take the Ε-form of the verb. The use
in dependent clauses is excluded.

21.5 Indirect questions

Urarina has no formal strategy to create indirect questions. In constructions that

involve two clauses where one is a question, the two are simply juxtaposed.
Thus, the question is not actually "indirect" or dependent from another clause.
The constituent order in a clause which functions as a question is the same as in
other questions. The position of the dependent clause in this context can vary.
However, the position before the main clause is preferred, as with dependent
clauses except subordinate clause with purposive function. Some examples are
given in (1168). These include questions with 'who', 'when', and 'where'. In
example (1168c), the question is negative.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Interrogative verbs 837

(1168) Questions with indirect reference

a) With 'who':
kwituku-a-ü d$a tQäiti kurerani-a eno-re=l
know-NEG-1 ps/E what ever replace-NTR enter-IRR:3ps/E=ASS
Ί don't know who is going to replace him.' [Lit. Ί don't know. Who would replace
him?']

b) With 'when':
ka=raj be-u - d$abana u-ri-ki=~i
1 sg=for tell-IMP when come-IRR-2ps=ASS
'Tell me when you will come!' [Lit. Tell me! When will you come?']

c) With 'where':
d$u ne huliä - ta hoituku-i=ße
where be:3ps/E PSN NEGQ know-2ps-NEGF
'Don't you know where Julian is?' [Lit, 'Where is Julian? Don't' you know (it)?']

21.6 Interrogative verbs

Apart from the various question types, Urarina has a few verbs with
interrogative function. These are pasia 'do how', d^atoama 'be how', tonaa
'sound how', and äaka 'be where'. The use of either of these verbs
automatically implies a question. While d$atoania and jiasia mainly occur in
the participle form where they are used as interrogatives for 'how', tonaa does
not occur in this form. The examples in (1169) illustrate the use of these verbs.
Note that different forms of person marking are possible with all three verbs;
however, occurrences with forms other than 3ps are extremely rare. While an
irrealis form can be observed with flasia in (1169a), it is unclear which other
grammatical categories can be marked on any of the verbs.

(1169) Interrogative verbs

a) jiasia:
flasi-ri-akatQe=ta
do.how-IRR-1 pl/in=FRS
'What will we [possibly] do?' ['This man does miracles and we cannot blame him.' NT:
John 11:47]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
838 Questions

b) d$atoama:
d^atoam-a nukue
be.how-3ps/A river
'How is the river?' [Referring to the water level]

c) tonaa:
tona-a kaa kuraa
sound.how-3ps/A this name
'What is the name of this?'

The intransitive interrogative verb äaka occurs only in the 3ps form to mean
'where is' and precedes the S argument (where present), which can be animate
(as in (1170a)) or inanimate (cf. (1170b)). Note that example (1170c) occurs
with a plural suffix on the verb to mark 3pl. However, this form is optional, as
3pl can also be referred to by using the singular form (cf. §8.7).

(1170) Interrogative verb äaka

a)
äaka ririau
where.is PSN
'Where is Ririau?'

b)
äaka kanii ka=saabere
where.is GPF lsg=machete
'Where is that machete of mine?'

c)
äaka-uru tQäe leotQa-uru isköo ne-ur-i=ta
where.is-PL also other-PL nine be-PL-NOMSbj=FRS
'[But] where are the other nine?' [NT: Luke 17:17]

The verb äaka has the same function as any construction with d^u plus copula
(cf. §21.1.8) and there are no known differences between the two constructions.
In questions with 2ps reference, the only way to express 'Where are you?' is
with the construction involving d$u.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Summary 839

21.7 Summary

A summary of Urarina interrogatives is given by the table in (1171). It becomes

obvious that there are different strategies to mark content and polar questions,
with separate devices to mark affirmative and negative clauses. However, most
question types make use of morphemes that also have other functions in the
grammar. For instance, negative content questions exhibit no special marker
that would be inherent to this clause type: the addition of mjej is optional and
this word is also used with other functions (cf. §14.2). As for affirmative
content questions, the interrogative pronouns also function as indefinite
pronouns (cf. §21.1.11). The final negative marker in negative polar questions is
not exclusive to questions either, as it also co-occurs after some other negative
introducers (cf. §5.10). Rhetorical questions do not constitute a special clause
type, as the verbal inflection applied with these corresponds to the conditions
for polar and content questions in general. Thus, the enclitic =te for rhetorical
questions may not be an element to mark a question as such, but could be
regarded as a morpheme to express emotion or attitude. In fact, =te occupies the
same slot as other attitudinal markers (slot 24), as was suggested in §12.3.7. In
conclusion, the interrogative marker =na and the introducer ta are the only two
morphemes that are exclusively related to questions. These are complemented
by some interrogative verbs, which have an inherent interrogative function.

(1171) Question types

Type Device Additional features

Affirmative content Interrogative pronouns Always fronted
With Ε-form on verb
Optionally marked for emphasis
Interrogative verbs Usually with participle form
Alternatively use of short forms
Affirmative polar Enclitic =na on verb Emphasis through shift of na
Negative polar Introducer ta, enclitic Affirmative marking on verb
=ne With Α-form on verb
Negative content Interrogative pronouns Always fronted
With negation and Ε-form on verb
Optionally marked for emphasis by
mjej
Rhetorical Enclitic =te Optional co-occurrence with =na
Α-form with polar questions
Ε-form with content questions

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
22 Discourse strategies and expressive elements

The organisation of discourse in Urarina texts may slightly vary according to

text genre. The main types on which this work is based are narratives and
conversations. Descriptions or instructions are a further type of genre
represented in the database (cf. §1.6). As narratives constitute the major part of
the text database, I will focus my analysis of discourse to this genre. A few
features of non-narrative genres are mentioned in §22.6.
Narratives can be further divided into traditional narratives and "modern"
type narratives, which include self-experienced hunting stories (such as "When
I was attacked by a jaguar") or reports about dangerous events experienced by
others (e.g. the dramatic rescue of someone bitten by a snake). The differences
between the two narrative types are minor and mainly regard variations between
traditional vs. innovative language, as is further investigated in §23.2. This may
involve the use of special vocabulary in traditional narratives, plus the frequent
use of the hearsay evidential hetau. This marker, however, is also used in
narratives that describe events not experienced by the story-teller himself,
especially when these lie in the distant past. Conversations exhibit a much more
frequent use of particles than narratives. This is plausible, as most particles
express some kind of information about the speaker's emotional state - which
may be less relevant in a third person narrative. This is similar to the use of the
attitudinal markers =ra and =ta, which only occur in direct discourse (in the
traditional language only; cf. §12.3.7).
Features with specific functions for the discourse occur on all grammatical
levels, as will be demonstrated in the following sections. These include
phonological peculiarities (cf. §22.1), morphological features (§22.2), and
particular syntactic structures (cf. §22.3). As Urarina makes ample use of
ellipsis, reference tracking is another important aspect, which is investigated in
§22.4. A few further features are then discussed in §22.5.

22.1 Phonological peculiarities

Features that are not normally part of the phonological system are commonly
used to emphasise dramatic events, or to "spice" the description of actions or
events, especially in narratives. Most typically, such exceptional sound patterns
occur in ideophones. This is illustrated in (1172a), where the speaker intends to
dramaticise the description of the way people hunt. The ideophone Ρεεε is
unusual, as it contains a glottal stop after consonant and a vowel not occurring

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Phonological peculiarities 841

otherwise in the system, in order to imitate the sound of a rifle. The sentence in
(1172b), from a traditional narrative, reproduces the alarm noise made by a
certain bird whose beak was stuffed in order to silence him. In this case, the
muted attempt of croaking is realised with a final glottal stop.

(1172) Non-standard sounds in ideophones

a)
ahaena-ana amu-akaanu, nesari-akaanu, obana su-akaanu,
jungle-inside walk-lpl/ex hunt-lpl/ex cl.peccary kill-lpl/ex

m-a eru-akaanu=ne itgafioa, Ρεεε

be-3ps/D find-lpl/ex=CND rifle (shooting)
'We walk in the jungle, we hunt, we kill collared peccaries; when we find [one], if there
is a rifle - boom! [we shoot it].'

b)
nii hau hetau=te ku hee+hee+hee+hee7 na-~i, kani
that because HRS=FOC there (muted.croak) say-PRT sort.of:3ps/E
'Therefore, [as his beak was stuffed], he made "hee ...".'

Another type of deviation from the regular phonology is the gemination of

consonants (here marked by a repetition of the consonant in brackets). Again,
this device is used in order to express emphasis on the sudden or abrupt manner
in which an action or event occurs; however, in 1st person narratives, it may also
be understood as a sort of boasting by the speaker, as some consider it
inappropriate to exaggerate one's own achievements. It is unclear whether the
feature of consonant gemination occurs in all dialects, as the example in (1173)
is from the Chambira dialect of Pionero (cf. §23.1.8).

(1173) Gemination in narratives

itQaf(f)a-anu hana=te lejhn nii eene-etga nala-t(t)o-a=ne hau

shoot-lsg/D when=FOC one that woman-only fall-INTS-3ps/D—SUB because
'When I shot it [the monkey], as the female fell down ... [the male tried to escape].'

Extreme vowel lengthening, which involves a lengthening clearly beyond the

normal length of long vowels (indicated by multiple vowels in brackets in the
transcription below), is another characteristic observed especially in narratives.
The example in (1174a) (from the Corrientes dialect) illustrates extreme vowel
lengthening in the adverb 'earlier', where it expresses intensity, implying 'very
long time ago'. Lengthening such as illustrated here can potentially occur with
any vowel, but is most common with /a/, as this is the most frequent vowel. The
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
842 Discourse strategies and expressive elements

feature is much more widespread in ideophones, as extreme vowel length may

be used to imitate sounds. This is illustrated in (1174b), where the speaker
describes the sound of desire made by a man who is watching the woman
Lomai taking a bath. It is intensified in the second part, where his desire and
groaning becomes so strong that he turns into a heron (which happens to make a
similar sound to the one described by the ideophone.

(1174) Extreme vowel lengthening

a)
baaba-kuru siitQura-uru-a hetau okäfiuo-oru ßoa(aa)elu
Dad-PL deport-PL-3ps/D HRS PSN-PL earlier
'In the old days, the Ocampo clan deported [our] fathers.'

b)
ku kuurahe-t ßaaau ßaaau ßaaau na-e,
there desire-PRT (desire) (desire) (desire) say-3ps/E

ku kuurahi-a läe hetau ku ßaaa ßaaaaa na-Ί hetau,

there desire-3ps/A just HRS there (strong.desire) say-PRT HRS

kaa nehuuri ne-1 amu-e

this heron be-PRT walk-3ps/E
'He desired her and made "nyaaau ...". He just desired her [so much], and made
"nyaaaa ..." - and turned into this heron.'

22.2 Morphological features

The most significant feature of morphological marking related to the

organisation of discourse is discussed in a separate chapter on person inflection
types (§11). As pointed out there, the Ε-form is not only the obligatory
inflection type for clauses that contain a focus marker, but it also is the typical
inflection type for a particular text genre - narratives. In contrast, the Α-form is
the preferred type used in conversation (cf. § 11.3).
Another characteristic of narratives, which may be described as reflecting
phonological and morphological factors, is reduplication. A detailed analysis of
reduplication is given in §10.3, where five main types are identified. While
these can be assigned well-defined grammatical functions, they also create a
more lively effect in the narration of stories. In a few examples, this is
expressed by triplication of the verbal root, for which the function cannot be
determined in grammatical terms. Thus, it can be assumed that this feature

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Morphological features 843

(such as illustrated in (1175)) adds to the dramaticising effect of the event's

portrayal. In (1175a), the first reduplication is of Type 4, marking repeated
action; the triplication which follows is based on Type 2a reduplication to
indicate 'run in a certain state or position' (cf. §10.3 on different reduplication
types). In this case, it is used to express total disorientation in the jaguar's
movements. A further case of triplication is shown in (1175b), where the
reduplication Type l a is employed to indicate the woman's motion as she
embraces her husband. The use of the triple repetition of the verbal root adds
further emphasis on the intensity of the action.

(1175) Reduplication in discourse

a)
n=asi kahe itQana tutuara+tutuara-ι, habaa+habaa+habaa-oki-i,
3ps=nose from blood RED+make.flow-PRTRED+RED+be.open.above-DER-PRT

situhuaitu suru-e nit hanolari

rapidly run-3ps/E that jaguar
'[As he struck the jaguar's nose] the jaguar was splashing out blood from its nose,
running around [with its head] upward, and rapidly ran away.'

b)
nii hau hetau ke balu+balu+balu-uki-t, kanaanaj rihitii
that because HRS VLI RED+RED+hold.on.to-DER-PRT child like

auauako-a nii raj komasaj=fle hau

cry-3ps/D that POSS wife=SUB because
'As his wife was crying and holding on to him like a child ... [he said "Don't behave
like that!"]'

One could also speak of reduplication in ideophones. In fact, most ideophones

inherently occur with a number of repetitions of the same syllable, which may
represent a plain imitation of a sound, but can also be viewed as a stylistic
device to describe a situation more realistically. The use of such ideophones in
narratives is also considered as good style, as it tends to make a narrative more
interesting to listen to. An example is given in (1176), where the ideophone
represents the sound of an owl, which is "talking" in this way, in order to advice
the son of a man deported by cannibals as of how to trace their tracks.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
844 Discourse strategies and expressive elements

(1176) Reduplication in ideophones

tu tu tu tu tu tu tu tu tu tu na-i, aräala beru tokuani-a

(sound.of.owl) (sound.of.owl) say-PRT tapir way compare-3ps/A

kasarena bakaua beru=ra, tu tu tu tu tu na-ϊ, ku u-a

TRN Indio way=EMF (sound.of.owl) say-PRT there come-3ps/A
'[From far it sounded like an owl] "tu tu tu tu tu" - the path of the Casarena Indio is
like the path of the tapir - making " t u tu tu tu tu", [the owl] came there.'

22.3 Syntactic structures and discourse

There are a number of syntactic properties that are typically observed in Urarina
narratives. While omission of constituents may occur, the insertion of syntactic
material is much more predominant, as demonstrated by the following list of
features commonly found in narration:
- Ellipsis
- Iconic repetition
- Tail - head style clause chaining

- Concatenation of several dependent clauses to very long sentences

22.3.1 Ellipsis
Ellipsis mainly regards two aspects: firstly, the omission of core constituents
such as subject or object NPs, where this is acceptable in terms of grammatical
structure (cf. §18 to review the possible types of argument structure). Such
omissions can be observed in any text genre. Secondly, and more specific to
narratives, the quotative verb naa, which marks direct speech embedded in a
sentence, can be elided. In this case, there usually is some other marking to
mark a quotation. Most typically, this is achieved by the discourse particles he
and üra, which may mark a change of speaker and occur utterance-initially (cf.
§22.4.6). However, a quotation may also be introduced by other ideophonic
markers that indicate the beginning of a quotation. One example is given with
ha7 a in (1177a), which is not found elsewhere in the database, but clearly
introduces a quotation in this context. Example (1177b) contains a more
common ideophone, wa, which may express disagreement, disgust, or repulsion
towards a given situation. (Example (1177c) differs from the other two cases, as
no ideophone is used, but the recipient of the omitted quotational verb is
indicated as 'to his family'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Syntactic structures and discourse 845

(1177) Omission of naa in discourse

a)
satii amu-uru-a itQauena-uru=ne kureu, kanaanaj-uru-atQa ajßa
all walk-PL-3ps/D elder-PL=SUB in.place.of child-PL-only with

ne-uri-a=ne, ha7a rauto-a d^ane kaa ßaara bihi alarihja

be-RAP-3ps/D=CND hmm be.tasty-3ps/A certainly this 2pl hand palm
'After all the elders had gone [hunting], when [the cannibal] was [alone] with the
children, [he said], "Hmm, these palms of your hands are certainly tasty.'

b)
kuruataha-uru-a hau, hatai ßäe baneheteuruni-a häu,
be.two-3ps/D because very already be.in.trouble-3ps/D because

10a ßäe ka=jane-u-toe=ra

(repulsion) already lsg=leave-IMP-PL=EMF
'As they [the jaguars] were two, as he already was in big trouble, [he said], hey, leave
me [alone]!'

c)
mi häu hetau turu-ϊ, mi n=arai-tg,uru raj,
that because HRS arrive-PRT that 3ps=family-PL for

tau=te loawaake raka turato-a u-naa ritcuri=tg,a

over.there=FOC grunt:3ps/E yesterday arrive-NTR come-NOM herd=EMF
'Then, as he arrived, [he said] to his family, "Over there the herd that arrived yesterday
is grunting.'"

As a third, but less common alternative, a quotation may occur without any
segmental marking in place of the quotational verb. However, in the few
examples recorded, this would usually involve a change in intonation or voice
as to indicate the switch from narration to quotation. The sentence in (1178) is
an example for this.

22.3.2 Iconic repetition

The intentional exact repetition of constituents is a relatively rare feature in

discourse. More exactly, it is only observed with verbs and it emphasises the
continuing duration of an action or event. The iconic repetition as illustrated in

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
846 Discourse strategies and expressive elements

(1178) occurred several times in the narrative in order to stress the exceedingly
long travel two men had to undertake in order to take revenge at their enemy.

(1178) Iconic repetition

mi hau hvajtei ku-he-ure, ku-he-ure, ku-he-ure,

that because again go-CNT-3pl/E go-CNT-3pI/E go-CNT-3pl/E

kwajtei ekwajjieherii lauhu-e, hoajtel paaohwa ahaane kane-i

again a.little.low be.sitting-3ps/E again how.many year pass-PRT

kau situ-e
here pass-3ps/E
'Therefore, again they went, and went, and went - it [the grass] was a little lower [and
so they asked] again, "How long ago did [the tapir] pass here?'"

22.3.3 Long sentences

Another typical feature of narratives is the occurrence of sentences with a large

number of dependent clauses - which does not normally occur in conversation.
The dependent clauses that occur may be of different types, such as illustrated
in (1179): three clauses are subordinate clauses with a conjunction, two are
participle clauses, and one is a non-finite privative form of the verb. The
concatenation of several clauses probably has the function of emphasising the
close relation of events in a particular context and tends to increase the tension
of the listener.

(1179) Long sentences

jiäe akauru tahitQu ke uhuäi ni-a akau hana,

[already 3pl ankle VLI until be-3ps/D water when]
[DEP1]

ku atii auno tute-ι, töoto tute-ϊ, ku raasa-uru-a

[there nevertheless flute play-PRT] [drum play-PRT] [there dance-PL-3ps/D
[DEP2] [DEP3] [DEP4]

hau, ßäe akauru tihja+suri ke uhua1 ni-a akau=ne hana

because] [already 3pl foot+belly VLI until be-3ps/D water=SUB when]
[DEP5]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Syntactic structures and discourse 847

mi häu atii ku raasa-ure hajti hiane-na here-lanaala

[that because nevertheless there dance-3ps/E] [still leave-INF want-PRV]
[Main] [DEP6]
'When the water already was [rising] up to their ankles, they kept on playing the flute
and playing the drum, and as they danced - therefore, when the water was up to their
calves, they danced [on], still not wanting to release her.'

22.3.4 Clause chaining

Clause chaining is another extremely common tool of relating events in a

narrative to each other. This is achieved through tail-head clause linkage, i.e.
the end of the previous sentence is repeated as a dependent clause. Most
typically, the dependent clause is realised as a subordinate clause with a
conjunction (usually hau 'because' or hana 'when'), or, less frequently, with a
participle clause. Example (1180) illustrates how the final clause of (1180a) is
repeated as a dependent clause in (1180b). In this example, the repeated clause
is a serial verb.

(1180) Clause chaining with subordinate clause

a) Sentence 1 :
raj kalaui ahitQa ke taka-1 hoara-a ku-e
[POSS son major VLI meet-PRT] [see-NTR go-3ps/E]
[DEP] [Main]
'Meeting with her eldest son, he went to see.'

b) Sentence 2:
hoara-a ku-a häu hetau=te ßäe hetau mki
[see-NTR go-3ps/D because HRS=FOC] [already HRS ADVRS
[DEP 1 ] [DEP2]

ßäe bati-a hau, ßäe bati-a(aaa) na-a

already sink-3ps/D because] [already sink-3ps/A.(fading.out)] [say-3ps/D
([DEP2]) [QUOT] [DEP3]

häu hetau=te kami komokomo ne-Ί amu-e

because HRS=FOC] [sort.of comocomo.bird be-PRT wa!k-3ps/E]
([DEP3]) [Main]
'As he went to see, as [the water] had dropped, he said, "It has suuu(nk)!" and turned
into that sort of comocomo bird.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
848 Discourse strategies and expressive elements

The sentence in (1181b) exemplifies the repetition of the final clause of (1181a)
as a participle. Note, however, that the hearsay marker hetau is inserted in
addition, in order to emphasise the reported character of the situation. The
repetition of the last "chunk" of (1181b) in (1181c) demonstrates that clause
chaining does not only involve the verb, but may repeat structures up to a
substantial length. The structures that are glossed as "Main" in (1181b), strictly
speaking, consist of two further participle clauses and a main verb, which I have
summarised as "main clause" in order to mark the repeated part as a unit.
However, one simplification can be noted in the repetition: in (1181c), the
words ml and nanaj are omitted. Such omissions, however, are optional.

(1181) Clause chaining with participle clause

a) Sentence 1:
na-1 hetau hi kuruae-kure
[say HRS] [there be.happy-3pl/E]
[DEP] [Main]
'Saying so, they were happy.'

b) Sentence 2:
ku hetau kurua-hei ku ra-1 ita rihiana-ϊ urwari
[there HRS be.happy-PRT] [there receive-PRT] [REC be.like-PRT urarixoot
[DEP1] [DEP2] [Main]

ne-1 nanaj ke belaj-ure

be-PRT each VLI give.as.gift-3pl/E]
'Being happy (there), they received him and each of them in the same way gave him
urari roots as a gift.'

c) Sentence 3:
it a rihiana-i urwari ke belaj-uru-a hau hetau=te
[REC be.like-PRT urari.root VLI give.as.gift-PL-3ps/A because HRS=FOC]
[DEP]

ku hetau kauru hueena-e

[there HRS 3pl laugh-3ps/E]
[Main]
'As each of them gave him urari roots in the same way, he laughed at them.'

Example (1182) again involves clause chaining with subordinate clauses. The
sentence in (1182b), however, does not pick up the final main verb of (1182a),
but repeats the quotative verb naa, embedded in a subordinate clause with hau.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Reference tracking 849

This is unusual, as normally, the final clause is repeated. Sentence (1182c), in

turn repeats the final verb of (1182b).

(1182) Clause chaining involving the penultimate clause

a) Sentence 1:
hoane=te akauru hoaa-ni-aka=ra, na-ι hetau ajte
[let's=FOC 3pl see-DSTL-lpl/du=EMF] [say-PRT] [HRS say:3ps/E]
[QUOT] [DEP1] [V]
"'Let's go see them!" he said.'

b) Sentence 2:
na-a hau hetau, ku~i, ku-ure
[say-3ps/D because HRS] [go-PRT] [go-3pI/E]
[DEP1] [DEP2] [V]
'As he said so, going, they went.'

c) Sentence 3:
ku-uru-a hau hetau=te, turu-uru-a ku-ure
[go-PL-3ps/D because HRS=FOC] [arrive-NTR go-3pl/E]
[DEP1] [V]
'As they went, they got to arrive.'

22.4 Reference tracking

Reference tracking in Urarina discourse can be problematic, which is especially

noticeable in narratives. One reason for this is that Urarina has no syntactic
pivot and no formal marking for switch-reference; in combination with omitted
arguments, thus, the interpretation of a clause can be ambiguous and much
depends on the context and general knowledge of the listener. There are several
ways in order to indicate a switch of subject reference in discourse:
a) by insertion of the respective NP (cf. §22.4.1)
b) by use of the demonstrative nii in place of an NP (cf. §22.4.2)
c) by subject marking on the verb (cf. §22.4.3)
d) through the context (cf. §22.4.4)
The last case involves no formal marking at all and may result in ambiguities.
An additional case is discussed in §22.4.5, regarding the use of the conjunctions
hana and hau, which may have had a discourse function at an earlier stage. In
the present language, these do not seem to distinguish same/different subject
reference any more. A change of speakers in quotations can marked by
ideophonic discourse particles, as is investigated in §22.4.6.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
850 Discourse strategies and expressive elements

22.4.1 Overt NPs

T h e presence of N P s clearly is the most unambiguous way to signal reference,

as is the case in (1183), where both clauses contain Ο and A arguments.
However, as noted before, this situation is comparatively rare, as more
frequently, arguments are omitted.

(1183) Overt NPs for reference tracking

nii hau hetau=te nitoanei bua basihjau-a alau=na

[that because HRS=FOC] [like.that] [bag] [steal-3ps/D] [spider.monkey=SUB]
[CNJ] [ADV1] [Ol] [VI] [Al]

hau, nekajritQa-he-ι beree ama-e unee nuhuae knanaj-tg,a

[because] [suffer-CNT-PRT] [child] [take-3ps/E] [kinkajou] [mouth inside-only]
[CNJ 1 ] [DEP2] [02] [V] [A2] [PP2]
'Therefore, because the spider monkey stole kinkajou's bag like that, the kinkajou
carries her children in her mouth, suffering.'

22.4.2 Demonstrative "nii"

As an alternative to the insertion of full NPs, the demonstrative nii 'that'

(meaning 'close to the addressee; cf. §5.6.1) is used as a tracking device in
discourse. However, while the use of nii helps to identify a participant as
previous mentioned or unmentioned, it does support the distinction between S/A
and O. When a participant is introduced for the first time, it is usually preceded
by lejhii ' o n e ' , as to mean 'a certain . . . ' . I n any sentence after this introduction,
the character is referred to by nii; however, as this is an optional feature,
exceptions are possible. In fact, in most sentences, no N P at all is present as to
refer to the subject. An example for the use of mi and lejhii is given below. In
(1184a), ' o n e m a n ' is introduced with lejhii. In (1184b), reference to this
character is expressed by nii. The sentence in (1184c), in contrast, does not
contain any overt marking for reference, as 'his branch' is ambiguous; thus, it
could be 'that m a n ' s branch' or the branch of another participant.

(1184) Discourse function of nii and lejhii

a) Sentence 1:
asaru tajala-l=te nalu-e lejhii katQa, alberto na-a raj kuraa
avocado harvest-PRT=FOC fall-3ps/E one man PSN say-3ps/A POSS name
'[When he was] harvesting avocado, a certain man fell [down]; his name is Alberto.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Reference tracking 851

b) Sentence 2:
mi=te asaru taj-a ku-e raj daka ajßa
that=FOC avocado pluck-NTR go-3ps/E POSS brother-in-law with
'He went to pluck avocado together with his brother-in-law.'

c) Sentence 3:
taj-a=ne hana=te ne-fiuaufiuajte raj nune
pluck-3ps/D=SUB when=FOC ITR-break:3ps/E POSS branch
'When he was plucking [avocado], his branch broke off.'

Another example is based on the narrative Text 8 as displayed in Appendix A.

In sentence (2), 'one m a n ' again is introduced as a new character. From
sentence (3), it can be derived through the context that 'his daughter' refers to
this man. Sentence (4) still is referring to the man, but without any overt
marking. Sentence (5) contains t w o clauses; the subject of the first clause is
marked as 3pi through person and number suffixes on the verb, thus referring to
' t h e y ' (= the people). The dependent clause 'when [she] ran a w a y ' is not
marked and is entirely ambiguous; however, it can be concluded from the
context of the story that it refers to the daughter. 101 Similarly, no N P s are
present in sentence (6). Sentence (7) again introduces 'her mother' as a new
participant. Note, however, that lejhii is not compatible with the possessive
clause, possibly because a possessive implies definiteness whereas lejhii refers
to indefiniteness.

22.4.3 Verbal marking

Inflection on verbs may help to track a referent as long as it refers to different

values for number and person. In example (1185), the subject is changed several
times, but in some cases this is signalled by a finite verb form (and subordinate
marker + conjunction). Even though the different referents are not expressed by
N P s , the identification of most referents is straightforward, as they are
distinguished for number, which is marked on several verbs in the sentence.
DEP1 does not refer to a participant, but to the situation described in the
previous sentence of the narrative. D E P 2 is marked for 3pl and hereby indicates
a change of subject, DEP3 is potentially ambiguous, as it could be understood
as referring to either DEP2 or to DEP4; however, as most participle clauses
precede the clause in which the subject is marked, and through knowledge of

101
The construction with hana as directly following the verbal root tends to imply a
different subject from the main clause (cf. §22.4.5). However, this could also refer to
'the man', as the subject of the main clause is 'they'.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
852 Discourse strategies and expressive elements

the context, one can retrieve that 'going by canoe like that' refers to DEP4,
which describes a fishing scene. DEP5, however, is critical, as it is marked for
3sg and thus could be understood, as referring to the same subject as DEP4 -
which it is not. Instead, this is a case of absence of number agreement, which is
then corrected/specified in DEP6, which contains the plural form. The sentence
does not have a main clause, as the quotative verb naa, which would be
expected here, is omitted. Therefore, O u t of envy' implies that "they spoke out
of envy", followed by what they said.

(1185) Reference tracking by verbal marking

nitoanei ni-a-ne hau huara-he-uru-a-ne hau,

[like.that be-3ps/D=SUB because] [see-CNT-PL-3ps/D=SUB because]
[DEP1] [DEP2]

ku nirihei laulauelau-he-1, ate taba-j muku-a hau

[there like.that go.by.canoe-PRT] [fish be.big-NOMSbj catch-3ps/D because]
[DEP3] [DEP4]

Jauara-hi-a hau huara-uru-a häu nehvelaj kau

[see-CNT-3ps/D because] [see-PL-3ps/D because] [envy because.of]
[DEP5] [DEP6] [PP]

d$a kahe=te katoanei ate taba-j eru-e

[what from=FOC like.this fish be.big-NOMSbj find-3ps/E
[QUOT]

kaa kanaanaj manuhoaj kalanoaeri-l=ta

this child damned be.dirty-PRT=FRS]
'As is was like that, as they were watching [him], as he was going in a canoe like that,
as he caught big fish, as they were watching, as they saw him, [they said], because of
envy, "Where does that damned dirty kid get the fish from?'"

22.4.4 No marking

In contrast to the different ways of reference marking as sketched above, many

examples have no formal marking at all, which easily results in
misinterpretation if the context is not well-known. Native speakers usually have
no difficulties to retrieve the omitted information from the context or from their
general knowledge of how things work. However, on some occasions,
consultants who were assisting in text analysis of narratives by other speakers

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Reference tracking 853

agreed that ambiguities were evident. The sentence in (1186) is one example in
which the reference for some participants can only be recovered from the
context and through carefully making conclusions from rather subtle clues. For
instance, the subject of DEP1 can be determined with relative certainty as it is
likely to be identical to the one of DEP2 which must be 'kinkajou', given that
'spider monkey' already functions as the object. The same implications apply to
DEP3, where the subject can be assumed to be identical to that of DEP4. To
identify the subject of DEP4 as 'spider monkey', however, appears to be
entirely dependent on the thorough understanding of the ongoing story, as the
catching of the weaving stick could have been by either participant. The main
clause, in contrast, is unambiguous, as A and Ο arguments surface.

(1186) Clue and context-based reference tracking

na-ϊ hetau mi alau tolto-a=ne hana hetau=te

[say-PRT HRS] [that spider.monkey hit-3ps/D=SUB when HRS=FOC]
[DEP1] [DEP2]

he taka-ϊ nii obipa muku-rito-a nukuj-a=ne

[VLI meet-PRT] [that weaving.stick catch-DER-NTR in.vain-3ps/D=SUB
[DEP3] [DEP4]

hana hetau=te alau bihi beu-ite

when HRS=FOC] [spider.monkey hand] [cut-DER:3ps/E]
([DEP4]) [O] [V]

nii unee obipa he

[that kinkajou] [weaving.stick INST]
[A] [PP]
'Saying so, when he [the kinkajou] hit the spider monkey, he [the kinkajou] quickly
caught a weaving stick, [and] when [spider monkey] tried to quickly catch it in vain,
[the kinkajou] cut off the spider monkey's finger.'

Similarly, in (1187a), the interpretation of the two different subjects involved in

the sentence completely depends on the context. Analysing this example from a
purely grammatical point of view, one could also conclude that the man who
caught the bagre fish turned into a dorado fish himself. The same problem
occurs in (1187b): whereas the subject concord between DEP1 and DEP2 can
be derived through their relation as participle followed by a finite clause, there
is no indication whatsoever that DEP3 and DEP4 refer to a different subject.
Thus, one could hypothetically understand this section to mean that the son
himself caught another fish and put it on his own tongue. The actual meaning,

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
854 Discourse strategies and expressive elements

however, becomes clear only through knowing that this procedure occurred
before in discourse and that Adam is the one who catches the fish and puts them
onto the son's tongue.

(1187) Context-based reference tracking

a)
hana t$äe hirirüßo ma natu muku-1, enamhja kuane
[inside also] [bagreS]ish even catch-PRT] [canoe inside
[CNJ] [DEP1] [DEP2]

hauto-a=ne siria ne-ι nerutu-e

throw-3ps/D=CND] [dorado.fxsh. be-PRT] [turn.into-3ps/E]
([DEP2]) [DEP3] [V]
'Also, even when [the man] caught a bagre fish, when [the man] pulled it into his
canoe, it [the fish] turned into a dorado fish.'

b)
mi hau hetau=te ku kuru-ri-i nakwaaunei
[that because HRS=FOC] [there swallow-RAP-PRT] [again
[CNJ] [DEP1] ([QUOT])

hvajte-u=ra, na-a hau, ku hetau hoajtei muku-uri-l,

repeat-IMP=EMF] [say-3ps/D because] [there HRS again catch-RAP-PRT]
([QUOT]) [DEP2] [DEP3]

lele tohvete
[tongue lay:3ps/E]
[Main]
'Therefore, he [the son] quickly swallowed [the fish] and as he [the son] said, "Do it
again!" he [Adam] again caught a fish quickly and put it on his [the son's] tongue.'

The interpretation of some clauses can be even more complicated when the
participants have to be derived through general knowledge of situations as the
one depicted in (1188), where at no point it is mentioned who or what was left -
neither in the preceding sentences of the stoiy, nor in the title. Even the
sentence following the displayed example gives no clue as to what could be the
subject: 'As they left [him], [he] got scared, because there was no canoe in
order to return home.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Reference tracking 855

(1188) Reference tracking through general knowledge

lejhü miirjka he ku-uru-a hau hetau=te,

one working.session VLI go-PL-3ps/D because HRS=FOC

päe ahe-uru-a=ne hau, ku>ara-elanaala raatiri-ure

already get.drunk-PL-3ps/D=SUB because see-PRV leave-3pl/E
'As they went to a work session, they left behind [one man], without seeing [him],
because they were drunk.'

22.4.5 Conjunctions "hana" and "hau"

T o account for the semantic and functional difference between the two most
frequently used conjunctions hau 'because' and hana ' w h e n ' , is a challenge.
The problem lies in the fact that these two do not always match the meaning as
given above. For instance, hau cannot be strictly translated as 'because' in the
following examples. In (1189b), a more adequate translation of hau is with
'even though'. The actual function of hau in this example is one of clause
chaining, i.e. to relate the sentence to the previous one (1189a) by repeating the
last clause. The same conditions apply to (1189d), where hau is the used in a
clause chaining construction based on (1189c). Again, in this context, a
counterfactual translation as 'even though' is more adequate.

(1189) Use of häu with other meanings

a) Sentence 1:
d$asu=te katoami kanaanaj hanoa-i=ta,
how.come=FOC like.this child give.birth-2ps=FRS

na-ϊ hetau ku kunuhe-ure

say-PRT HRS there insist-3pl/E
'Saying, "How is it possible that you gave birth to a child like this?" they insisted.'

b) Sentence 2:
hi kunuhe-uru-a häu, ßäe atii be-na her-ene
there insist-PL-3ps/D because already nevertheless tell-INF want-NEG:3ps/E
'Even though ["because"] they insisted, she did not want to tell.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
856 Discourse strategies and expressive elements

c) Sentence 1:
ßäe kana bihi+kuturi ke nisisiri-ahe
already lpl/in hand+head VLI clean-CNT:3ps/E

na-1 hetau ku hveena-e

say-PRT HRS there laugh-3ps/E
'Saying, "he is already cleaning [his bottom] with our fingers", [the trees] laughed.'

d) Sentence 2:
kwena-a hau hetau=te ku atii abaata-he
laugh-3ps/D because HRS=FOC there nevertheless resist-CNT:3ps/E
'Even though ["because"] it [the talking tree] laughed there, he still resisted [to laugh as
well].'

In a similar way, the use of hana in discourse can have alternative

interpretations. For instance, the translation of hana as 'when' is not the only
possible one in some examples of Text 53 (cf. Appendix A). In sentence 3, it
could be interpreted as meaning 'after'.
Based on these variations in meaning, one might assume that the actual
function of hau and hana extends to the wider discourse and is not restricted to
clause level. According to an idea mentioned by Manus (2002, p.c.), the use of
hana may involve an opposition related to what is the subject of the new
sentence, as compared to the previous sentence. This hypothesis can be
confirmed by data from some text passages, but it cannot hold as a general
hypothesis, as is evident from other passages. Let us assume for a moment that
hana has the discourse function as described above. This is partly confirmed by
the Text 8 in Appendix A, looking at all sentences that contain hana as a
conjunction: in sentences 3, 5, 6, hana in fact coincides with a change of the
subject between main clause and dependent clause. The same observation is
valid in sentence 6 of Text 10 (Appendix A). However, looking at Text 53, the
comparison between hana and hau is not entirely conclusive: on the one hand it
is correct that hana indeed seems to mark a switch in subject (in 7 out of nine
occurrences); on the other hand, hau does not consistently mark a same subject
relation between clauses: this is the case in ten examples from Text 53, but in
eight other examples, hau occurs in a context that implies a switch of subject.
As a modification to the original hypothesis that hana expresses a change of
subject, however, one could assume that hana marks a significant "change of
situation", which can but does not necessarily involve a change of the subject.
This interpretation seems to match a much larger number of examples. Thus,
while there might be some clue as of the function of hana, it could be suspected
that the distinction between the two conjunctions has been lost over time (if
there ever was one). There also is a high degree in variation among different

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Reference tracking 857

speakers, which explains the fact that the speaker of T8 and T10 seems to make
systematic use of the "discourse" function of hana in order to mark a change of
situation. In contrast, this is less the case for the speaker of T53. A further
complication is that the distinction between mi hau 'therefore' and nii hana
'then' is even less clear, as a systematic switch in reference cannot be observed
with these.

22.4.6 Introduction of quotations

As briefly mentioned in §22.3.1, the introduction of a quotation in traditional

narratives can be accompanied optionally by certain ideophonic discourse
markers. These are especially important when the quotative verb naa is omitted;
however, introducing discourse markers such as he are also used when naa is
present. Also note that naa always follows a quotation, whereas the markers
discussed here precede it.
Occurring only in traditional narratives, he is a very frequent discourse
introducer, indicating that someone is starting to say something. This can also
mark the change of speaker in quotations, when no quotative verb is present. In
addition, in most examples, he has a connotation that expresses a negative
feeling, such as annoyance. The use of he also tends to be understood as an
expression of superiority towards the addressee. Thus, it can be assumed that he
has a double function - as a quotation introducer and as an ideophone referring
to an emotional factor (negative feeling). In (1190), the emotional side of he is
demonstrated: the sentence is from a traditional narrative in which a spirit
bewitches someone with a curse. The (more powerful) husband of the woman is
annoyed and sends the curse back to the spirit, ordering him to come. Thus, the
use of he below expresses a) annoyance, because of the spirit's shameless
behaviour, and b) superiority towards his envoy that is addressed.

(1190) he with connotation of annoyance and superiority

[Preceding: 'Saying "He has taught my children too much" he put the curse into the
container.']
sarara-ι, he kaa ama-l hoaa-m-u=ra, kutia-m-u=ra,
put.into-PRT hey this take-PRT see-DSTL-IMP=EMF call-DSTL-IMPHEMF

na-ι hetau raj letono raj

say-PRT HRS POSS envoy for
'Putting [the curse] into [the container], saying to his envoy "Hey, take this and go to
see him! Go call him!" ... [he sent him].'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
858 Discourse strategies and expressive elements

The sentence in (1191) exemplifies that the negative connotation does not apply
in all examples. In this case, two girls who had been abducted by cannibals,
suddenly see their uncle who has come to save them. The first use of he is in
conversation of the girls amongst each other; the second occurrence is towards
the uncle. In both situations, the context strictly excludes any connotation of
annoyance or superiority. Thus, the function of he in this example is purely as
an introducer of a quotation.

(1191) he as a quotation introducer

mi hau hetau=te he kana n=itg,ene=ra, na-l rene turu-1,
that because HRS=FOC hey lpl/in 3ps=uncle=EMF say-PRT place arrive-PRT

he kanakaanu=tQe kaa=ra itoene, rta-a ku-ttru-a

hey lpL/ex=PLT this=ERD uncle say-NTR go-PL-3ps/A
'Therefore, saying "Hey, our uncle [is there]!" they went towards him and went to say
"Hey, uncle, this is us!'"

Another extremely common marker occurring only in traditional narratives is

üra (also realised as hrtra, nra, or ra). Younger speakers are unable to define its
meaning and even older speakers have difficulty in describing its function,
simply stating that "this is how they spoke in the old days". However, one older
consultant characterised it as a way of expressing sincerity and respect in
conversation. This function is not entirely consistent, but as it matches the
meaning of many examples, I have glossed this marker as "Emphatic Respect
marker in Discourse" (ERD). While üra may occur as an introducer at the
beginning of a quotation, its use is not limited to that. In fact, multiple
occurrences of üra, often realised as =ra when attached to another word, are
very common. A problem is, when this marker surfaces as =ra, it is
homophonous to the final emphatic marker =ra (slot 24; cf. §12.3.7) and it is
difficult to distinguish their functions. In fact, a diachronic relation between the
two is likely. One occurrence of the ERD marker is on the demonstrative 'this'
in (1191); whereas the occurrence on 'our uncle' in the previous clause could
also be interpreted as a final emphatic marker, as glossed. Further occurrences
of üra are illustrated in (1192), in two sentences from a traditional narrative
about two men who stole a piece of shell from Lomai. Most commonly, it co-
occurs with he, which precedes üra as to introduce the quotation and its
"negative" function may be neutralised. Example (1192a) illustrates the use of
üra after he as implying respect, emphasising the sincerity of the utterance
(even though the audience of the story knows it is a lie). Sincerity is also
emphasised in (1192b), where the accused man tries to defend himself against
the accusation brought forward by Lomai.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Reference tracking 859

(1192) Multiple occurrence of ra

a)
jiae tg,u hjä=ra hitarii kaipa-kaanu=ra, kurura, na-ure
already CRTN just=ERDall return-lpl/ex=EMF Milord say-3pl/E
" O h Milord, we really have returned all (pieces)", they said.'

b)
he, ma mtohuara ma aj-i-ta na-a hetau
hey ERD no EMF AUX-NEG:3ps/A=FRS say-3ps/A HRS
"'No! absolutely not!" he said.' (as response to accusation of having stolen something)

The respectful function of üra is less evident in (1193), where Lomai comments
on the consequences o f her visitor's disobedient behaviour.

(1193) üra with connotation of respect

mi hau hetau=te he üra mi-tea here tQu-ra

that because HRS=FOC hey ERD that-only want:3ps/E CRTN=EMF

ka=hoedaj, üra naaohwa-i hetau=ra, lomaj kiuedaj na-uru-a=ne

lsg=visitor ERD ENUM-PRT HRS=EMF PSN visitor say-PL-3ps/D=SUB

hana hetau=ra, he nauruta-eri-u=toe, na-a=ra=ne

when HRS=EMF hey move.over-DIM-IMP=PLT say-3ps/D=EMF=CND

dorodoroahari=tQa na-a lomaj=jie sajhjel hetau=ra, ku

bathe:HORT=EMF say-3ps/D PSN=SUB although HRS=EMF there

lauhua-ϊ, he i=t$uara-ri-a-ü=m=ta
be.sitting-PRT hey 2sg=see-IRR-NEG-lsg/A=ASS=FRS
'Therefore, (Lomai said), "Well, hm, that's what he wanted, my visitor", when she said
to Lomai's visitors, "Go over there", when she said it, although she said "I want to
bath", he sat there, saying "I won't watch you.'"

Different from the occurrence o f =ra as a final emphatic marker (cf. §12.3.7),
the position o f the ERD marker is virtually unrestricted, as it is found in any
position o f the sentence. This can be observed in (1193), where üra occurs
clause-initially, clause-finally, or within a clause. Multiple occurrences o f üra
are also observed in (1194), which contains five instances of the ERD marker,
some even in a sequence.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
860 Discourse strategies and expressive elements

(1194) Excessive occurrence of üra

na-a hau hetau=te he üra ajatQaisi=tQa=ra hauria=ra

say-3ps/D because HRS-FOC hey ERD exactly=EMF=ERD first=ERD

üra ka=lae=ra raa-m=tQa, na-e hetau

ERD 1 sg=penis=ERD receive-DST=EMF say-3ps/E HRS
'The spirit said "Exactly, first let me go and get my penis!'"

22.5 Other features

A typical characteristic of any kind of narrative is the insertion of gap fillers,

such as in pause or hesitation. However, these can also occur when no pause is
made as they may be used in order to avoid a pause. There also appears to be a
tendency of avoiding isolated verbs as forming a clause. As a result, an adverb
is inserted in order to "spice" the utterance. There are three adverbs that occur
with this function most commonly: ßäe 'already', ku 'there', and hvajtei
'again'. The preference for any of these may differ from speaker to speaker or
even from sentence to sentence within the same text. Examples for each are
given in (1195). It is obvious that the repeated occurrence of the adverbs does
not have a semantic value. For instance, the occurrence of ku before each verb
in (1195b) is entirely optional and other speakers suggested that at least the
second and the fourth occurrence of ku may be removed in transcription. The
excessive repetition of huajtei in (1195c) was perceived as bad style by others.
Yet, this is a feature occurring to some extent in most narratives.

(1195) Repetition of adverbs as gap fillers

a) ßäe:
ku ne-uru-a hau, ßäe tabiit$a näe huajtei ßäe
there be-PL-3ps/D because already finally already again already

nareti-a bihi häu, ßäe kuri+ißaa ari-l huajtei ßäe

miss-3ps/D hand because already y'agwa.tree+fruit seek-PRT again already

nala-a, hajti ÜQÜU ne-he-1 tohiooj na-a

drop-3ps/D still near be-CNT-PRT "plump" say-3ps/A
'As they were [staying] there, again finally, as hind missed [the water], when he again
looked for a jagua fruit and dropped it, it sounded "plump", still being near.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Other features 861

b) ku:
fa lenone-Ί, ku ne-uru-a, ht karahai ht ne-ure
there eat-PRT there be-PL-3ps/A there long-time there be-3pl/E
'They ate there and staid there; they staid there for a long time.'

c) kwajtev.
nii baja kwajtel bataneheti na-akaanu hau aj
that after again PLN say-lpl/ex because AUX:3ps/E

kwajtel lejhii katga hoajtei

again one man again
'After that, again, [something] happened [to] a certain man again in [a place] we call
Bataneheti [= mosquito forest].'

Another gap filler commonly used is the word kaami (or variant kardi) which
can be characterised as a multi-class word. As briefly described in §5.16, it can
be used as a full verb to mean 'do like that', 'do this sort of thing', but it is also
used as a noun 'this sort of thing' - and as a gap filler, as exemplified in (1196).
In (1196a), kardi may be analysed as a noun to bridge the temporary memory
loss of the word for the beverage described here. This is evident from the fact
that it fills an argument slot as a copula complement for the first occurrence of
rda. In (1196b), kardi cannot be characterised as functioning as an argument, nor
is it a verb. In this example, it simply functions as a gap filler similar to the
other words described in (1195).

(1196) kanii as a gap filler

a) Head of NP:
ku ni-a kami, rd-a fioanara+kata kahe ne-naa aaune+baka
there be-3ps/A sort.of be-3ps/A banana+ripe from be-NOM plantain+soup
'There was this sort of thing, plantain soup [made] from ripe bananas ...'

b) Noun modifier:
amuemue-ka hau hetau=te ku hetau kardi banaao ke
wander-3ps/D because HRS=FOC there HRS sort.of leaf.shelter INST

asara-hei hetau
cover-PRT HRS
'As he wandered around, [there was food for him], covered with (kanii) - leaves as a
shelter.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
862 Discourse strategies and expressive elements

Another word which is overwhelmingly frequent in narratives is the hearsay

marker hetan. For instance, in Text 25 (a narrative on the creation of the
peccaries; see Appendix A), hetau occur 58 times in a total of 44 sentences.
However, its occurrence and position does not seem to be exactly prescribed
and may also vary among speakers. Interestingly, the teller of texts 8 and 10
does not use the hearsay marker at all. A possible explanation for this is that
hetau occurs much more frequently in narratives that refer to a distant past
(though this is not definable in absolute terms). In yet other narratives of a
similar kind (e.g. hunting stories), but told by other speakers (not represented in
the text selection), hetau occurs once every two to three sentences, on average,
which demonstrates the high degree of speaker variation. As text 25 offers the
widest range of use for hetau, I will focus on this narrative here, selecting the
most relevant occurrences for the analysis.
In the first sentence, hetau occurs twice after finite verbs, but not in each
clause. The occurrence after finite verbs is very frequent in general, but not
restricted to this. In sentence 7, hetau occurs after a participle (i.e. non-finite)
form of the verb. In sentence 6, the position of hetau is not adjacent to any verb,
as it occurs between n=obana nuke and the gap filler kanii. It is also found in
most sentences that begin with 'therefore' or 'then', where hetau occurs after
the conjunction, often followed by a focus marker. In sentence 11, hetau is
observed as following a dependent clause with conjunction. Beside that, when it
occurs adjacent to a verb, the position of hetau is not necessarily postverbal: in
sentence 12, it occurs after the adverb ku 'there' and before a participle verb
(ku hetau nesarakata-i). In sentence 14, hetau occurs after the emphatic particle
lae 'just' and before ku. In summary, there are few positions in which hetau
cannot occur (e.g. in sentence-initial position) and there are no apparent
restrictions as to the word classes that occur in its vicinity. These factors justify
the status of hetau as a particle, as opposed to the reportative enclitic =he,
whose occurrence is more strictly defined (cf. §12.3.4).
The adversative particle niki is relevant for discourse structure as it does not
directly refer to the clause in which it occurs. Instead, it marks in anticipation
that something negative is going to happen as a result of the action described,
such as to add 'but at the end of a clause. For instance in (1180b) further
above, the presence of niki indicates that the fact of the water having dropped is
not entirely positive, since the person who announces it is going to turn into a
bird. However, the position of niki in this example is at the beginning of the
clause that describes the sinking of water levels. The function of niki is even
better illustrated by the example in (1197), which is the initial sentence of a
narrative. The presence of the adversative marker in this sentence already gives
the listener an idea that the described hunting trip is going to be abnormal.
However, the course of the story goes on normally for a long time before the
actual negative event (the woman is abused by an evil spirit) happens, niki is
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Other features 863

used twice more in other sentences before this scene. Thus, the use of niki can
be described as having a long-distance effect across a wider discourse.

(1197) Adversative marker niki in discourse

βoaelu hetau=te niki hano-a ku-e raj komasaj

earlier HRS=FOC ADVRS hunt.overnight-NTR go-3ps/E POSS wife

ajßa lejhii katga

with one man
Once a man went hunting overnight with his wife.'

The function of the adverb/conjunction t$äe 'also' has been discussed in

§20.5.1. It should be repeated here that in some contexts, t$äe has scope over a
wider discourse in that it does not necessarily conjoin two adjacent sentences
but can refer to a pragmatic level, such as the repeated use of commands during
discourse.
Another feature typical to narratives are the common expressions that
c o m m e n c e and end a story. Though there are no fixed literal formulae in this
regard, the expressions used in this context very much resemble each other.
Traditional narratives typically begin with jioaelu hetau=te, such as in (1197),
which could also be translated as 'once upon a time'. Modern narratives
sometimes have a title such as example (1198), which constitutes a dependent
clause with subordinate marker and could be interpreted as temporal clause. The
title in (1198a) contains the word for 'story' (a nominalised form of ' s p e a k ' ) ,
whereas (1198b) simply occurs as a temporal subordinate clause.

(1198) Initial sentences of narratives

a) With noun for 'story':

leihii eene su-a ema=ne ke ere-naa
one woman kill-3ps/D tree=CND VLI speak-NOM
'Story of when a tree killed a woman'

b) Without noun:
ekuu enua ke sini-ι hano-a kuruataha-j katQa=ne
top tree VLI sleep-PRT spend.night-3ps/D be.two-NOMSbj man=CND
'When two men spent the night sleeping on a tree'

T h e end of a narrative is most commonly concluded by using the verb saa

' e n d ' , which can occur with or without an intensifying suffix.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
864 Discourse strategies and expressive elements

(1199) Concluding sentence in narratives

ßäe sa-a / ßäe sa-to-a
already end-3ps/A already end-INTS-3ps/A
'[The story] has ended.'

22.6 Observations on non-narrative genres

While narratives form the majority of the texts contained in the database, a few
observations can be made on other types of texts. One regards instructions, for
which a sample text is also given in the text selection in Appendix A. Different
from English, where instructions are typically impersonal, this is not the case in
Urarina, which does not favour "abstract" description in general. Instead, a first
person plural form is used, which may in some cases include the listener, with
some variation. Different examples - all from Text 86B in Appendix A - are
given in (1200), where the speaker uses different lpl forms in the course of the
text. Using the lpl/ex form in sentences 1 through 5, she switches to the dual
form in sentence 6 and sticks to this form through the end of the instruction. In
fact, at no stage are passive or impersonal forms used in this text.

(1200) Use of lpl forms in instructions

a) lpl/ex (sentence 3):
haurianehei ßäe turu-Ί, uhoana hana turu-ι, n=asahi
first already arrive-PRT field inside arrive-PRT 3ps=under

hveeta-kaanu
clear-lpl/ex
'First, arriving already, arriving at the field, we clear the underground.'

b) lpl/du (sentence 6):

ruku-1, ßäe kodorja-ϊ, saauhi-ι ßäe nii baja ruee-kaa
pull.out-PRT already heap.up-PRT cut.off-PRT already that after peel-lpl/du
'Pulling them out, heaping them up, cutting them off - after that we already peel them.'

It should be mentioned that in Text 86A (not listed here), which precedes Text
86B, the lpl/ex form is used from sentence 1 through 27 and switches to the
dual form in sentence 28, from it is used through the end (sentence 38), with
one exception occurring in the exclusive form in sentence 35. In an instruction
about how to fish, by another speaker, the dual form is used throughout. Thus,
both forms are possible options in instructions and one can only speculate about
the different preferences. One possibility is that, strictly speaking, the listener

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Observations on non-narrative genres 865

should not be involved by using the dual form, as he or she does not participate
in the described activity. From this point of view, the lpl exclusive form seems
more appropriate. However, the use of the dual form may be understood as a
friendly gesture as to include the listener in the described activity, rather than
excluding him from an action that he or she could potentially participate in.
This view would be consistent with my experience with the Urarina on the
social level, as outsiders may be invited to participate in some activities once a
relationship of trust is reached.
As a (marginal) feature in conversation, it may be noted that people do not
typically greet each other. As mentioned in §11.3, the greeting formulae
hartoara and soara for 'good morning' and 'good afternoon' are likely to be of
recent origin and could be roughly translated as 'it is light' and 'it is dark',
respectively. More commonly, if occurring at all, people would briefly ask on
an encounter with another person ßae=na, which is the word for 'already'
followed by the interrogative enclitic, to mean 'How's it?'. The reply can occur
simply by saying eehe 'yes'. In more elaborate interactions, people would greet
a relative by calling his kinship term, such as aua 'sister', possibly adding a
question as to enquire to her task (e.g. 'Are you going to the field?'). When
receiving the answer, one may optionally express a sign of good will by adding
something such as 'Beware of (the) snakes!' However, these utterances cannot
be characterised as standardised forms.
More common is an expression used after a conversation with a visit to
somebody's house. In this case, the sentence ßäe kauanu Ί am already going
home' could be described as a common greeting formula, to which the
addressee replies eehe 'yes'. Slightly less common is the related formula
ßäe amuanu Ί am already going (away)', which may be used when a person
leaves his or her own house.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
23 Variation

Urarina exhibits two important types of variation. One is based on geographical

divisions and results in dialect differences; the other is the distinction between
contemporary and traditional language, i.e. between younger and older
speakers. Due to dialect mixing and individual preferences, the differences
between the dialects can be transitional. In some cases it is difficult to define
whether a grammatical feature is part of the traditional language or not, because
some younger speakers may make use of a feature and others not. The further
separation into dialects makes a definition even more complicated. To give an
example - the frustrative form is described as a feature of the traditional
language throughout this grammar. However, it is still actively used even by
younger speakers in the Corrientes dialect. Therefore, I restrict the division into
traditional and innovative varieties of Urarina to what is known from the Espejo
dialect. Furthermore, it should be noted that the information on dialectal
differences gathered for this study is relatively superficial and based on just one
or a few texts in most cases. It was also not possible to cover all dialect areas, as
most information was collected from settlements in the vicinity of the Chambira
River. The details for this study are discussed in §23.1, whereas I investigate
some features of the traditional vs. innovative language in §23.2. The contact
with Spanish has also had some impact of Spanish on the development of
Urarina - this is discussed in §23.3.

23.1 Dialectal differences

As mentioned in § 1.2, Urarina is spoken in an area between the Urituyacu and

Corrientes rivers and their tributaries (cf. map in (1201)). While the language is
close to extinction on the Urituyacu river, it is still actively spoken in the
Chambira river system and on the Corrientes river, though reportedly, the
monolingual use of Spanish is much more on the advance in the latter area as
some people start abandoning their language. While some data could be
gathered from two speakers of the Corrientes dialect, the bulk of texts come
from the Chambira river and some of its tributaries. Note that Urarina is not
spoken on the Tigre River in the East.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 867

(1201) Urarina dialect zones with names of rivers and tributaries

While each tributary (and probably each village) reflects minor dialectal
differences on its own, a division can be made into four major dialect areas:
- Zone A: the Western dialects, including Tigrillo and Espejo dialects
- Zone B: the Lower Chambira dialects, including Asna and Airico rivers and
possibly Pangayacu river and the villages below its mouth (note that no
information is available about the Patoyacu dialect)
- Zone C: the Upper Chambira with all its tributaries, thus forming the largest
dialect area
- Zone D: the Corrientes River, with a large number of villages and settlements
(at least 20). However, these include mixed communities with a majority of
Spanish-speaking people.
No data is available on the Urituyacu dialect, except for what is reported by
speakers of other dialects. According to several accounts, those people who still
know Urarina in that area are reluctant to use it.
The division into these dialect areas is based on a number of features that
were compared during a dialect survey in the Chambira area and its tributaries
(zone C). The findings on zone A dialects are based on the author's fieldwork in
the Espejo river and work with visitors from the Tigrillo river. Zone Β and D
were not investigated within the areas themselves, but with invited speakers
from these places.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
868 Variation

Another important aspect to be considered is that there are few speakers with
a "pure" dialect of a certain area, as the Urarina traditionally are a semi-
nomadic people and used to change places frequently. Over the past decades,
regional politics has enhanced the installation of fixed settlements, which
includes assigning land titles to native communities, thereby reducing the
tendency of short-term movements. However, as discussed in §13, marriage is
uxorilocal, i.e. the man moves to the wife's location, which may or may not be
within the same community. If no bride is available in the own community, men
move to other places for marriage, which makes dialect mixing a common
feature. Also note that due to cultural taboos, the author has very few recordings
from female speakers, who would be more likely to stay at their place of birth.
The majority of texts were obtained from male speakers, some of whom,
naturally, live in a community other than their parents' community. Fortunately
for this study, most speakers who were recorded stayed in their dialect zone
even when moving to another village.
The differences between the dialects under investigation are mainly of a
phonological and lexical nature. All dialects are mutually intelligible and no
major grammatical differences on the morphological and syntactic level were
observed. In the following, a range of distinctive features is discussed to justify
the division into four zones as suggested above. The discussions starts with a
detailed investigation of the feature of aspiration in §23.1.1, which permits to
assign specific dialects to dialect zones.

23.1.1 Postvocalic aspiration

All Urarina dialects exhibit some degree of postvocalic aspiration as described

in §2.1.10 for the Espejo dialect. As the presence of the vowel which precedes
aspiration is of minor significance (aspiration may occur after any vowel, with
few exceptions), I will focus on the types of consonants before which this
feature occurs. As mentioned in §2.1.10, aspiration is an optional feature -
which is valid for all dialects. However, interestingly, the dialects slightly differ
from each other with respect to the consonants before which aspiration typically
occurs. While the feature is observed before /t/, /k/, and /kw/ in all dialects and
before in /tp/ in most, the occurrence of postvocalic aspiration is not equally
observed before /I/ and /s/. As a tendency, aspiration occurs in more contexts in
the South and West, whereas there are more restrictions in the North and East.
For a more detailed map, including the location of the communities mentioned
in this chapter, see (2) in §1.2. In zone A (Espejo/Tigrillo), aspiration occurs

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 869

before /t/, /k/, /kw/, /tp/, and /s/, but not before /!/ (cf. (1202). As an exception 1 ΑΛ

from this, the speaker from Tagual (Tigrillo) also lacks aspiration before /tp/.

(1202) Postvocalic aspiration in the Western dialects (zone A)

Text / Dialect Before Before Before Before Before Before

/t/ ikJ /kw/ /tp/ Ν Isl
T9 Espejo yes yes yes yes no weak
T53 Espejo yes yes yes weak no yes
T16 Espejo yes yes yes yes no yes
T83 Tagual / yes yes yes no no yes
Espejo
T70 Puerto Rico yes yes yes yes no ?

Below is a short speaker profile for the speakers of the zone A dialects (cf.
(1203)). All speakers lived in Nueva Union de Espejo at the time of recording;
HN grew up there, while VA and MA were among the founders of the
settlement in the Espejo swamp ca. in 1986. JO and CI grew up in different
parts of the Tigrillo river and according to other Espejo dialect speakers have
preserved most of their dialectal peculiarities from that area. It should also be
noted that the entire group of founders of the Espejo community came from the
Tigrillo River. Yet, these have developed peculiarities that can be easily
identified by speakers of other dialects.

(1203) Speaker profile zone A

Text/ Age Living in Grown up in Parents

Speaker ca.
T9: HN 23 Espejo, 18 years Espejo Mother: Tigrillo
Father: nomadic
T53: VA 42 Espejo, 18 years Tigrillo/Espejo Mother: Tigrillo
Father: Pucayacu
T16: 48 Espejo, 18 years Tigrillo (San Mother: Tigrillo
MA Lorenzo) Father: Pucayacu

102
It is unclear whether this is speaker-specific or a general property of the Tagual
dialect. However, aspiration before /tg/ is generally less intense as compared to other
phonological environments. The notation as "weak" used in the table indicates that
aspiration might not occur in all examples and that it also is less intense as compared
with other speakers. The notation "?" indicates that no examples for the specific
context were available.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
870 Variation

Text/ Age Living in Grown up in Parents

Speaker ca.
T83: JO 22 Espejo, 5 years Tigrillo (Tagual) Tigrillo (San
Lorenzo)
T70: CI 21 Espejo, 1 year Tigrillo (Puerto Rico) Tigrillo (Puerto
Rico)

The sample from zone Β dialects involves one speaker from the Asna River (cf.
(1204)). However, it should be mentioned that this also is a rather young
community, probably less than 10 years old. The speaker (MR) is ca. 22 years
old and has been living in the Asna for 5 years. He grew up in Sta. Teresa
(Patoyacu), where both his parents come from. Thus, the information on which
the characterisation of the Asna dialect are based might not be solid enough and
be subject to other factors. The speaker from the Airico dialect (Sta. Rosa de
Airico), however, has always lived in that area and both his parents are speakers
of the Airico dialect. The other Airico speaker (MM), is WM's brother but has
lived in the Espejo for ca. 12 years. According to other Espejo speakers, he has
preserved part of his dialectal peculiarities and can still be easily identified as an
Airico speaker. With regard to postvocalic aspiration, all three speakers have in
common that they exhibit aspiration before any of the consonants in question.

(1204) Postvocalic aspiration in the Lower Chambira dialects (zone B)

Text/ Before Before Before Before Before Before

Speaker It/ /k/ /kw/ /tß/ l\l /s/
T69: MR yes yes yes yes yes yes
T71: WM yes yes yes weak yes yes
T22: MM yes yes yes yes yes yes

The largest group of language samples are from the Upper Chambira area (zone
C). The table in (1205) provides an overview of the different speakers. Note that
the first speaker (PA) originally is from the Urituyacu River but has been living
in Nva. Esperanza for 35 years - her parents were among the founding members
of the community. It is unclear whether her dialect is typical for the Chambira
area or whether she has preserved traits of the Urituyacu dialect. MC and MV
have not lived on the Chambira for long, but the Pucayacu River, from which
they moved, is part of the Upper Chambira area as well. All other speakers grew
up in zone C or moved to their current living place from nearby villages.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 871

(1205) Speaker profile zone C

Text/ Age Living in Grown up in Parents

Speaker ca.
T72: PA 60 Nva. Esperanza Urituyacu Urituyacu
(Chambira), 35 years
T73: MC 30 28 de Julio Sta. Cecilia Sta. Cecilia
(Chambira), 3 years (Pucayacu) (Pucayacu)
T74: MV 29 28 de Julio Sta. Cecilia Sta. Cecilia
(Chambira), 5 years (Pucayacu) (Pucayacu)
T75: PN 41 Pionero (Chambira), Pionero Mother: Chambira
32 years (Chambira) Father: Pucayacu
T80: JQ 70 Sta. Cruz (Chambira), Sta. Cruz Sta. Cruz
since birth (Chambira) (Chambira)
T76: JC 40 Sta. Silvia (Chambira), Sta. Silvia Sta. Silvia
since birth (Chambira) (Chambira)
T77: BM 21 Copal (Chambira), San Marco San Marco
2 years (Chambira) (Chambira)
T78: AC 28 Buenavista (Pucuna), Sta. Silvia Sta. Silvia
10 years (Chambira) (Chambira)
T79: MI 40 Buenavista (Pucuna), Buenavista Buenavista
since birth (Pucuna) (Pucuna)

Different from zone A dialects, zone C speakers exhibit aspiration before the
consonant /!/ (cf. (1206)). In contrast, aspiration does not occur before /s/,
except for a few isolated examples in Pionero. As an exception, the Pionero
speaker also has no aspiration before ΙΧς,Ι. However, note that aspiration in this
environment is also especially weak in a number of other dialects of zone C.

(1206) Postvocalic aspiration in the Upper Chambira dialects (zone C)

Text / Dialect Before Before Before Before Before Before

/t/ DU /kw/ /tp/ Ν /s/
T72: Esperanza / yes yes yes weak yes no
Urituyacu
T73: 28 Julio/ yes yes yes yes ? no
Pucayacu
T74: 28 Julio / yes yes yes weak yes no
Pucayacu
T75: Pionero yes yes yes no yes weak
T80: Sta. Cruz yes yes yes yes yes no
T76: Sta. Silvia yes yes yes yes yes no

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
872 Variation

Text / Dialect Before Before Before Before Before Before

Ν /k/ /kw/ /t<j/ m /s/
T77: Copal / San yes yes yes weak yes no
Marco
T78: Pucuna / Sta. yes yes yes yes yes ?
Silvia
T79: Pucuna yes yes ? yes yes no

There are two text samples from the Corrientes River. Both speakers come from
villages in the central area of this region and exhibit a higher influence of
Spanish than other speakers. Both speakers have a similar history, as their
ancestors moved to the Corrientes River from the Chambira/Tigrillo region in
their youth. Speaker CF moved back to the Chambira 5 years ago.

(1207) Speaker profile zone D

Text/ Age Living in Grown up in Parents

Speaker ca.
T85: CF 24 Nva. Esperanza Las Palmeras Mother: Patoyacu
(Chambira), (Corrientes) Father: Tigrillo
5 years
T84: HA 23 Nvo. Porvenir Nvo. Porvenir Mother: Hormiga
(Corrientes), (Corrientes) Father: Tigrillo
since birth

With respect to aspiration, the Corrientes dialect differs from the others as
aspiration does not occur before /I/ and /s/ (or at least to a much lesser degree -
a few examples contain aspiration before /s/ with speaker HA).

(1208) Postvocalic aspiration in the Upper Chambira dialects (zone C)

Text/ Before Before Before Before Before Before

Speaker /t/ fkJ /kw/ /t<;/ Ν /s/
T85: CF yes yes yes yes no no
T84: HA yes yes ? weak no weak

For the texts collected from the Pangayacu dialect, the findings are
inconclusive. Even though the area is part of the Upper Chambira region, it is
the southernmost settlement of this area. A problematic aspect also is that the
speaker profile is not homogeneous: speaker AN, aged 35, has been living in
Sta. Beatriz (Pangayacu) for about 10 years; her parents are from the Pucayacu

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 873

river. For speaker JB, it is unknown for how long he lived on the Pangayacu
River. As a tendency, their aspiration pattern resembles that of zone A.

(1209) Pangayacu postvocalic aspiration

Text / Before Before Before Before Before Before
Speaker /tf IYJ /kw/ /t<j/ Ν Ν
T81: AN yes yes yes weak no yes
T82: JB yes yes yes no weak yes

Based on a comparison of other features, it can be stated that the data from
Pangayacu dialect does not match any of the four given groups. A detailed
summary is given in (1215) further below.

23.1.2 Vowel copying

Another phonological feature already discussed in §2.8.3 is vowel copying,

which occurs with the vowels /u/, /i/, and /«/ in the Espejo dialect. (Recall that
these are copied through the consonants /h/, Irl, and /k/). The same observation
can be made for other dialects in zone A. In contrast, there are limitations to
vowel copying in the other three dialect zones: in these, no copying of /u/ to
result in [w] is observed. 103 For instance, the numeral verb 'two' is realised as
[ k u r w a t a h a a ] in the Western dialects, but [ k u r a t a h a a ] in other dialects. In
addition, the Esperanza and Pangayacu dialects (T81 only) do not exhibit vowel
copying of /«/.

23.1.3 Very long vowels

Extreme vowel lengthening is a feature that was already described in §22.1 as a

way to mark emphasis. However, some dialects have the same feature as a
regular feature in lps forms: most dialects of the northernmost Upper Chambira
area exhibit a significant additional lengthening of the long vowel /aa/ in the
lps suffixes -am (lsg/A) and -akaanu (lpl/ex). In addition, the place of
articulation of this vowel is further back than the corresponding forms in other
dialects. For instance, the form it$a-akaanu ('do'-lpl/ex) would be transcribed
as [ i . t p a a . k a a . n a ] in a narrow transcription. This phenomenon is strongest in

103
No statements can be made as to the role of different consonants through which a
vowel can be copied, as not sufficient data is available for all dialects.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
874 Variation

the Nva. Esperanza, Pionero, Sta. Silvia, Pucuna and the Corrientes dialect of
Nvo. Porvenir and to a lesser degree in other Northern dialects of zone C.
23.1.4 Diphthongisation

Another vowel-related feature is the diphthongisation of vowels, especially in

emphasis with the intention of dramaticising. 104 Most typically, this involves the
insertion of an /a/ before an existing single vowel, to form a diphthong with it.
Thus, /u/ becomes [au], /«/ turn into [a«], and /e/ or /i/ are realised as [ae]. In
the Copal/San Marco dialect, /e/ can be pronounced as [a«]. The occurrence of
diphthongisation is strongest in zone C. It is observed to a much lesser degree in
zones A and B, whereas it is absent from the Corrientes dialects (zone D). More
specifically, as displayed in (1210), zone A only exhibits diphthongisation of
one type in one dialect (Tagual), whereas it is absent from others. Similarly,
only one type, again, the formation of [ae], is observed in one dialect of zone Β
(Airico). In zone C, several kinds of diphthongisation are attested in different
dialects.

(1210) Types of diphthongisation

a) Distribution:
Zone Type
A Tagual: /e/ [ae]
Β Airico: /i/, /e/ -> [ae]
C Esperanza: /o/ [au], /i/ [ej]
Sta. Cruz: /a/ [a«]
Copal: /e/ [aa]
Pucuna: /u/ [au], /i/ [ae],
/a/ [a«], Ν [ej]
D none

b) Examples:
/e/ -> [aa] /tihje-ara/ 'squirrel monkey'-PL -> [ti.hjaa.a.ra]
/e/ [ae] /leotpa/ 'other' [lae.o.tQa]
/i/ [ae] /ahinia/ 'before' [a.hae.ni.a]
Ν [ej] /aji/ 'no' -> [a.jej]
/u/ [au] /su-!/ 'kill'-PRT [sau.l]
/a/ [aa] /kutari/ 'head' [ku.taa.ri]

104
This phenomenon is most common in hunting stories and other self-experienced
narratives as to express intensity, velocity, or danger in an exceptional manner.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 875

23.1.5 Monophthongisation and vowel elision

A process opposite to diphthongisation may also occur in some dialects

mentioned in (1210) and in a few others. In this case, a diphthong or a sequence
of two vowels is simplified to a single vowel. There are several types of such
simplification: In the diphthongs /aj/ and /au/, the first element (/a/) is elided
and the resulting vowels are /i/ and /u/, respectively. When the second vowel of
the diphthong / a t * / is elided, the resulting vowel is [äa], with nasalisation
being retained. In vowel sequences, the second element is elided and
nasalisation is retained, which results in a long nasalised vowel. The different
types are listed in detail in (1211b). As shown in (1211a), the distribution of
these rules is different in the various dialect areas. The smallest variety of
alternations are found in zones A and D; in zone B, vowel elision of this type
does not occur (but note that apocope in other contexts occurs in all dialects).
Zone C again exhibits the largest variety of alternations, being especially
prominent in the Sta. Silvia data. It should be noted that the alternations listed
here are not obligatory changes that would apply to all words, but they are
possible alternations - which occur in some dialects, but not in others.

(1211) Types of vowel elision in diphthong or vowel sequence

a) Distribution:
Zone Type
A Tagual: /a«/ [äa]
Pto. Rico: /a«/ [aa]
Β none
C 28 de Julio: /a«/ -> [äa]
Sta. Cruz: /a«/ [aa]
Sta. Silvia: /au1 [o], /äu/ [äa], /e.«/ -» [ee]
Copal: /e.ö/ -> [ee]
Pucuna: /ätt/ [äa]
D Both locations: /aj/ [i]; /ä«/ [äa]

b) Examples:
/aj/ [i] /hitQuhwa-j/ 'be black'-NOMsbj [hi.tQu.hwi]
/ a u / -» [u] /iteau-akaanu/ 'live'-lpl/ex [i.tQo.a.kaa.n«]
/a«/ [aa] /laaeko-a/ 'sit'-3ps/A [laa.e.ko.a]
/äa/ [äa] /lenoni-tpä«/ 'eat'-lsg/A [le.no.ni.tcäa]
/ a . u / - > [aa] /laauhiri/ 'small' [laa.hi.ri] (Pangayacu)
/ a . ! / -> [äa] /nai/ 'over there' [näa] (Pangayacu)
/e.l/ [ee] /kale-he!/ 'wash'-PRT [ktt.le.hee]
/e.ö/ [ee] /su-ene-ä/ 'kill'-NEG-lsg/E [so.e.nee]
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
876 Variation

23.1.6 Lowering of[eeü] to [aaü]

Most dialects have specific words in which a diphthong or vowel sequence is

replaced by another, such as the example käu 'because o f , which becomes kai
in the Puerto Rico dialect. However, there is one alternation of this kind that is
more general than such specific alternations. Some dialects exhibit a change of
[ee«] to [aaü], especially in lsg/E forms of the verb. For example, /here-«/
('want'-lsg/E; normally [he.ree.«]) is pronounced as [he.raa.ä]. This
alternation is found in the southern regions, such as in the Asna dialect (zone
B), in Tagual (zone A), and in the Pangayacu dialect, but not in the Chambira
dialects. In the Corrientes dialects, the rule also occurs, but with a slight
variation: the resulting vowel is further simplified to [äa].

23.1.7 Variations between /r/, Id/, and /t/

Variations that involve changes between Irl, /d/, and It! are difficult to describe,
as it is unclear which consonant is the underlying one (cf. §2.1.3 on differences
between the Espejo and Chambira dialects). However, there is one tendency that
can be observed as a dialectal difference. While dialects of zone A and D (i.e.
the outermost areas) exhibit variations that involve Irl and Id/ only, the other
two dialect zones are characteristic for variations that involve Idl and /t/.
Specifically, Irl and Id! are hardly distinguished in the Western dialects; on the
Corrientes River many words can be pronounced with either [r] or [d].
Especially before III, the vowel [d] is preferred. The situation is different in
dialects of zones Β and C: here, Ir and Idl are not interchangeable, i.e. the
distinction between the two consonants is retained. However, Idl is usually
replaced by [t] in these dialects: for instance, occurrences of Idl in the Nva.
Esperanza dialect are attested in two words only (darue 'pot' and lidiane 'type
of leaf), as all other words that contain Idl in other dialects are pronounced with
[t]. The Pangayacu dialect (T82) stands out, as there is not only variation
between Idl and It/, but also between Irl and /t/, such as in /raatiri-1/ 'leave'-
PRT pronounced as [taatiru].

23.1.8 Consonant gemination

Another variation mainly occurring in dramaticised, emphatic contexts is the

gemination of consonants. This phenomenon is most widespread in zone C,
whereas it is absent or a marginal feature in other dialect areas. In some cases
(Pionero, Airico), it can co-occur with postvocalic aspiration, whereas in others
it replaces aspiration where it occurs. Consonant gemination is absent from the
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 877

Western dialects. In the Corrientes dialects, only gemination of Ν to [tt] is

observed, which can coincide with additional glottalisation in the syllable
preceding the gemination, such as in mlato-a 'he fell down' being realised as
[na.la?.tto.a]. Glottalisation before gemination is also observed in the Pionero
dialect. While word-internal gemination is the most common type, it is also
observed word-initially in some dialects. In the Pionero dialect, gemination also
occurs across word boundaries, where it is confined to the word-level in other
dialects. The Pionero variety is also exceptional in that consonant gemination
may occur after long vowels, whereas it is observed only after short vowels in
other dialects. Another peculiarity is the gemination of transitional glides (cf.
§2.8.1), as observed with [iquj] and [ww] in Pionero and with [ww] in the
Pucuna dialect. In the Copal dialect, consonant gemination may also occur
word-initially, even when not preceded by another word. Gemination to [dd]
may be derived from gemination of /d/, such as in the 28 de Julio dialect
(marked "dd-1" here), or from gemination of /r/, as attested in the Airico and
Copal dialects. 105 Gemination of /t$/ results in the first component being
reduplicated, thus surfacing as [tt(j]. The Asna dialect is peculiar in that it is the
only one that does not exhibit gemination of /t/, but the otherwise much rarer
gemination of /s/.

(1212) Consonant gemination

a) Distribution:
Zone Type
A none
Β Asna: [ss]
Airico: [tt], [dd-2], [hh]
C 28 de Julio: [tt], [tte], [bb], [dd-1]
Pionero: [tt], [ff], [iquj], [ww], [ss], [tt<?]
Sta. Cruz: [tt]
Sta. Silvia: [tt]
Copal: [tt], [dd-2], [ss], [ff]
Pucuna: [tt], [kk], [bb], [nn], [ss], [ww]
D Las Palmeras: [tt]

b) Examples:
[tt] /satii/ 'all' [sattfl]
[tte] /kat<;a/ 'man' -» [kattea]
[ss] /tasinoha-a/ 'be big-3ps/A' [tassijiohwaa]
[dd-1 ] /ku-darue/ ASC-'pot' [kwaddaruhwel

105
The variation between /d/ and /t/ is not observed in the Copal data.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
878 Variation

[dd-2] /heräe/ 'slowly' [heddäe]

[hh] /bahe-V 'sink'-PRT [bahheei]
[bb] /obana/ 'collared peccary' [obbana]
[kk] /kanu+akatia/ 'lsg + 'opposite' [kanakkatia]
m /itcafwa-an«/ 'shoot'-lsg/A [tcaffwaanH]
[UJU}] /mane-a/ 'walk in herd'-3ps/A [rutqiqania]
[ww] /wawako-a/ 'grunt'-3ps/A [wawwakoa]
[nn] /r«-naka-ene-a/'fmd'-PSB-NEG-3ps/A [mnnakaenia]

23.1.9 Plural "-kuru " instead of "-tg,uru "

There are a few more marginal characteristics by which differences among

dialects can be identified. One is the plural form -tpuru for trisyllabic words
ending in /i/, such as arai 'family' (PL araitpuru; this word is the only example
attested in several dialects for the differences investigated here; for rules on
plural formation, see §3.1.1). This suffix is also attached to the demonstrative
mi 'that' in most dialects (PL mitQuru). However, there are some dialects in
which the suffix for these words can be -kuru (i.e. araihtru; mikuru). This is
observed in zone A with the Tagual dialect and in zone C (28 de Julio), but not
in any other dialects.

23.1.10 Pronunciation of 'a little'

There is some degree of variation for the pronunciation of the fricative in the
word 'a little', which is a verb usually occurring in the participle form or with a
nominalising suffix (cf. §5.4.1). The most widespread variant is with [tp], as in
[äatpihn], which is common in zone A (Tigrillo). However, some speakers of
the Espejo dialect pronounce the word with [(], as in [äa/ihii], while other
follow the variant with [t$]. This pronunciation is also observed in the
Corrientes dialect (zone D). In zone C, however, only the pronunciation with [s]
is attested, such as in 28 de Julio, where the word is pronounced [äasihii] (and
[äasihjei] in the Pucuna dialect). In Sta. Cruz (Upper Chambira), the word is
realised as [aasajjiehei]. For zone B, no occurrences of the word are attested
(which is coincidental). However, note that in the Airico dialect, [/] regularly
occurs as a variant of /s/ after /i/, which might indicate that [/] is also the variant
in the word for 'a little' (cf. §2.1.8).

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 879

23.1.11 Summary of common dialectal features

In the past section, ten almost exclusively phonological features were

investigated that are attested in a range of different dialects, with different ways
of realisation. Some of these features are of a binary nature in that they are
either present or absent in a given variety. Others occur to varying degrees,
when they regard certain processes that may involve specific groups of
phonemes or allophones. The ten distinctive features discussed above are listed
again in the overview in (1213).

(1213) Overview of some dialectal distinctive features

1) Aspiration
2) Vowel copying
3) Very long vowels
4) Diphthongisation
5) Monophthongisation and vowel elision
6) Diphthong modification
7) Variations between Irl, /d/, and It/
8) Consonant gemination
9) Plural -kuru vs. -tQuru
10) Pronunciation of'a little'

The table in (1214) summarises the relation of each feature with respect to the
four given dialect areas. It is evident that there is some degree of overlap
between features of different dialect zones. On the one hand, it is possible to
find similarities between most dialects of a specific group, which justifies the
division into four zones as proposed here. However, this does not imply that the
four dialect zones can be neatly distinguished with regard to all the distinctive
features. For instance, the fact that long vowels do not occur in zones A and Β
does not imply that the two groups share many further features. In fact, the
differences between the dialect zones are so extensive that it would be difficult
to join any of these as being more related than others.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
880 Variation

(1214) Summary: Distribution of dialectal features

No. West Lower Upper Chambira Corrientes

Chambira
1) Not before /I/; Before all Not before /s/; Not before /l/, Isl
Tagual: not consonants Pionero: not before ltg.1
before /t<p/
2) Copy /i/, /u/, /a/ Airico: no copy Esperanza; No copy of lul
of/u/ Copal, Pucuna: no copy of /u/
3) no no Esperanza; Pionero; Sta. Nvo. Porvenir
Silvia; Pucuna
4) Tagual: Airico: Esperanza: to! [au], none
/e/ [ae] /i/, /e/ [ae] /i/^[ej]
Sta. Cruz: /«/ [aa]
Copal: /e/ [an]
Pucuna: /u/ [au], Iii [ae],
/a/ [aa], /i/ [ej]
5) Tagual: none 28 de Julio: /äa/ [äa] Both: /aj/ -> [i]
/äa/ [äa]; Sta. Cruz: /aa/ [aa]
Pto. Rico: Sta. Silvia:
/a»/ [aa] /au/ [o], /äa/ [äa],
/ee.ä/ -> [ee]
Copal: /ee.ä/ -> [ee]
Pucuna: /äa/ [äa]
6) Tagual Asna None /eeä/ -> [aaä]
7) Tagual:/r/~/d/ Airico: Esperanza only Nvo. Porvenir:
/d/~/t/ Μ ~ /d/
8) none Asna: [ss] 28 de Julio: Las Palmeras: [tt]
Airico: [tt], [tt], [tte], [bb], [dd-1]
[dd-2], [hh] Pionero:
[tt], [ff], [iqvq], [ww], [ss],
[tt£]
Sta. Cruz: [tt]
Sta. Silvia: [tt]
Copal: [tt], [dd-2], [ss], [ff]
Pucuna:
[tt], [kk], [bb], [nn], [ss], [ww]
9) Tagual no 28 Julio no
10) Tagual: [te] ? 28 de Julio, Sta. Cruz, Pucuna: [te]
Espejo: [tg] / [/] [s]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 881

The data recorded in the Pangayacu dialect does not well fit into the scheme
given above. It can be stated with certainty that it does not belong to the Upper
Chambira dialect zone (C), as it does not share any of the features occurring in
this area. As the table in (1215) clarifies, three of the features resemble the
findings from zone A dialects, four match zone B, and four correspond to zone
D. With respect to vowel copying, the Pangayacu dialect is different from all
other varieties. Geographically, this dialect is located quite centrally, which
makes the mix of different features very plausible. However, as the speakers on
which these findings are based may exhibit influence from other dialect areas
(Patoyacu, Pucayacu) due to family history, these conclusions are preliminary.

(1215) Summary of features of Pangayacu data

Feature Details Matching

dialect zone
1) Aspiration Some not before IV A
Some note before ΙΧς,Ι (A, C)
2) Vowel copying No copy of /a/ (unique)
3) Very long vowels No A, Β
4) Diphthongisation No D
5) Monophthongisation /aj/ -> [i], D
/a.u/ [aa],
/a.V [äa]
6) Diphthong modification Yes A, Β
7) Variations between Irl, /d/, Idl ~ /t/, Β
and/t/ Irl ~ III (none)
8) Consonant gemination [tt] D, (C)
9) Plural - k a m vs. - t ^ a r a No B, D
10) Pronunciation o f ' a little' ? —

23.1.12 Some dialect-specific features

In addition to the differences pointed out in the previous section, each dialect
has some features that are unique to this particular variety. As the amount of
data is quite extensive, I will only sketch some of the most interesting
peculiarities here, using keywords for a brief overview. Note that all features
described below are optional.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
882 Variation

Puerto Rico (zone A):

- Loss of /«/ in some forms, as in /saaahi-l/ 'cut'-PRT [saahu];
/ b a r a e - k a r a - a / 'make masato'-PL-3ps/D [baraek«raa]; /tura-a+ka-e/
' arrive'-NTR + go-3ps/E [tarake]
- Lexicon: omari 'basket' -> amari\ bakusia 'beat' -> fivaufwaa; soaitQa
'break' soaaeka
- Postposition käu 'because o f

Tagual (zone A):

- /«/ [u] in /enejtp«/ 'monkey' [enejtpu]

It should be noted that Espejo speakers occasionally pointed out some further
peculiarities of the Tigrillo dialects (which do not occur in the texts from Puerto
Rico and Tagual). These include the replacement of /e/ or /i/ by [«] in some
words, mostly word-initially as in /edaa/ outside' -> [ttdaa]; /isi/ 'fire' ->
[usi], but also word-internally as in /nese/ 'village' [n«se]. These
pronunciations are also adopted by most Espejo speakers, as there is frequent
contact with Tigrillo communities.

Asna (zone B):

- Focus marker =te sometimes realised as =ti
- Insertion of initial /of in /btia/ 'bag' [obua]
- nuhoa 'look here' realised as kö
- Use of stative plural subject marker -ana with transitive verbs, as in
elele baitQa-ana-ϊ ('palm branch' + 'carry'-STPL-PRT) 'carrying palm
branches' (1 example only)
- Use of ajar a for yes' (instead of eehe)

Airico (zone B):

- Diphthong modification as in /soaja/ 'break' [swoja]
- /s/ /// before /i/, as in /i=su-a/ (2sg='kill'-3ps/A) 'he killed you' -> [ijua]
- /reetial/ 'until' -> [räatiaäl]
- Lexicon: rerneka 'disturb' -> tiarania\ nauania 'cliff/edge' neloa
- Frequent use of mihve (3ps/E form of 'be in vain') as a particle
- Final emphatic marker =ra still in use in some examples (only occurring in
the traditional variety of other dialects)
- Introducer baana for negative condition baane

Pangayacu (T81):
- Frequent elision of initial /a/: /aka«/ 'water' [kau]; /ahaena/ 'jungle'
[haena]

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Dialectal differences 883

- Other simplifications/reductions: /naaohwa-a/ (enumerative verb-3ps/A)

[nahöo]; elision of final -7, as in /tura-l/ [t«rö]
- ITR prefix tie- realised as na- in nehawaa 'be born' -> nahanoaa
- Insertion of initial /o/ in /fwei/ 'firewood' -> [o.fwee.i]

(T82):
- /i/ -> [e] in /itene/ 'beach' -> [etene]
- Occasional vowel copy across word boundary: /banaao + ha-~iI 'shed' +
'make'-PRT ('making a shed') [banaaohwaal]

Nueva Esperanza (zone C):

- Long vowel /ii/ can be realised as [eei]: /hasiite-V 'mix'-PRT ->
[haseeiteei]

28 de Julio/Pucayacu (zone C):

- Tendency of voicing /k/ to /g/, such as in /katpa/ -> [gatpa]
- nii hau 'therefore' realised as mi na hau, inserting na (with unclear meaning)
- häu 'because' and hana 'when' often replaced by haanu (normally a
purposive marker)

Pionero (zone C):

- Lexicon: tiaka 'be standing' tihjaa·, ruhuaria 'carry' lauhuaria, nesari
'booty' nesario
- Strong tendency for procliticisation: kanu aaso 'my stupidity' -> k=aaso;
kanu enamanaka 'my youth' -> ka=namanaka; Hitga su-nohwi (lsg.EMPH +
'kill'-PASS) 'what I killed' -» M=so-nohwi
- Devoicing of initial ΛΙ3/ in / d 3 « h a e k i - l / 'be standing'-PRT
[tp«h«eki-l]/ee.o/ /aa.u/ in /leotpa-ttiW 'other'-PL [laa.u.t$aa.ra]
(also note assimilation of a+« to [aa] here)

Santa Cruz (zone C):

- /t$/ [dj] in /kat^a/ 'man [ka.dja]
- /laahiri/ 'small' [laauhwiri], like some Espejo speakers
- Elision of final /a/ for 3ps/A in some examples: /itpa-a/ 'do'-3ps/A ->
[i.tpa]; /letoa-a/ 'send'-3ps/A [leh.twa]

Sta. Silvia (zone C):

- /e/ [i] in some contexts, as in /ajte/ 'say':3ps/E -> [ajti], /tonit-eni-a/
'sound'-NEG-3ps/A -> [tonitinia]
- Imperative suffix -h -> [o], as in /ama-n/ 'take'-IMP [amaao]
- Simplification of HRS marker with FOC enclitic /hetau=te/ to [toote] or
[hetote]
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
884 Variation

- Loss of nasalisation in some words: /hja=ra/ 'just'=EMF [hjaara]

- Lexicon: kutiaka 'call' kutihjaa
- lpl/ex suffix -akaanu after /i/ (normally -tQaanu), as in /nesari-t^aanu/
'hunt'-lpl/ex [nesariakaanu]

Copal (zone C):

- Transitional glide [g] more intense than in other dialects
- Negative condition introducer baana bana
- Two examples with VO order, as in nena kanu tQafwäu mi hanolari
('but':lsg + lsg + 'shoot'-lsg/E + 'that' + 'jaguar') 'but I shot that jaguar'
(AVO)
- lsg suffix -arm after l\l (normally -t$au), as in /saati-t$a«/ 'kill with
blowgun'-lsg/A [saati-anu] (both variants occur)
- Assertive emphatic enclitic ha (not observed elsewhere), as in /raauto-a=ha/
'it was tasty, I assure you!'
- /«/ [u] in /btirari/ 'poison' -> [burari]
- Lexicon: /lana-i/ ('be missing'-NEG:3ps/A) 'there is no shortage of ...' (to
imply there is sth. in abundance) banihi-ji (root unknown; suffix is
NEG:3ps/A)
- Suffix -nakauru 'those who' -> -nakuri

Buenavista de Pucuna (zone C):

- Use of stative plural subject marker -ana with posture/motion verbs, as in
/tihja+tihja-ana-a/ (RED+'walk'-STPL-3ps/A) 'they walked', realised as
[tihjatihjaaonaa]
- Suffix -at$a -e, as in /aka-atpa/ 'he only' [ahkaae]
- Vowel change /ii/ [ee], as in /hanori-1/ 'turn its back'-PRT [hanoree]
- Lexicon: obale 'soup' obaa; idiom mi kahe hoaaunei 'then', 'so', 'from
there' nii kahe fatal ('creating from that' 'going from that')
- Loss of /h/ as in /tasinoha-a/ ([tasijiohwaa]) 'be big'-3ps/A) [tasijto-a]
- lsg suffix -arm after /i/ (normally -tpäu), as in /kwaasi-t^a»/ 'be afraid'-
1 sg/A [kwaasi-ana] (both variants occur)
- Diphthong [ow] (not observed elsewhere), as in /kuhjuri-a/ 'curse'-3ps/A ->
[kohowdja]

Nuevo Porvenir (zone D):

- Voicing of ftg,/ to [dj], as in /hanii-t£«re/ 'select'-3pl/E [haniid3«re];
/amiane-i-tpe=ne=ra/ 'work'-2ps-PL=SUB=EMF [amianeeidjenera]
- Tendency of voicing /k/ to fgi, as in /bakaua-kura/ 'Indio'-PL
[bahka»a-göra]
- Regular use of the frustrative marker =ta by younger speakers (cf. § 12.3.7)
- Elision of final -z, as in /satii/ 'all' [sati]; /su-1/ 'kill'-PRT [so]
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Traditional and innovative language 885

- Insertion of initial /«/ in /kuaae/ 'over there' -> [ttkuaae]

- Privative suffix -elanaala often realised as -laa
- Postposition hilane 'without' -> [kulaae]
- Elision of /h/ in forms such as /tee-he/ 'give'-CNT:3ps/E -> [tee]
- Conjunction saihjei 'although' -> [sahn]

Las Palmeras (zone D):

- Hortative fauane -> hvana
-Tendency of voicing /k/ to /g/, as in /lia-akaan«/ 'move'-lpl/ex
[ljaagaanu]
- lsg/A suffix -t$äu -käu as in hjam-käu Ί leave it'

According to memories of the speaker JN, whose parents were from the
Urituyacu River, this dialect exhibits a range of words that have an initial /e/
where it is absent in other dialects. Examples for this are erua 'find' (otherwise
rua) and ektt 'there' (cf. ku).
There certainly are many more, though minor, distinctive features between
the various dialects. The data presented here are based on a short survey which
tried to cover as many different dialects as possible, on the expense of a more
detailed analysis. More in-depth studies will be required in order to document
the existing differences more accurately.

23.2 Traditional and innovative language

Differences between traditional and innovative language can be observed in

many areas of the grammar. The definition of this distinction is not clear-cut
and is best described as a continuum. Some features of the traditional variety
can be characterised as extinct or close to extinct, as they are only found in
traditional narratives, and not used in everyday speech (not even by older
speakers). Other features are still used by older speakers, but not by the younger
generation (and naturally, the transition between these is fluent). As a
simplification, I classify any forms that are not used by younger speakers any
more as traditional language.
Distinctive aspects include phonological factors, morphological peculiarities,
and the lexicon. Variations are also found with regard to possessive
constructions, and on the syntactic level, such as with regard to
complementation and subordination strategies. While syntactic innovations may
demonstrate that the grammar is enriched by the introduction of additional
constructions, other grammatical areas largely exhibit the loss of features, such
as phonemic distinctions or the disappearance of certain morphemes.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
886 Variation

23.2.1 Changes in phonology

The relation between [hw] and [fw] was discussed in §2.1.11. This is one
example for the loss of distinctive features in phonology. While the two sounds
were separate phonemes in the traditional language, [fw] and [hw] now function
as allophones in the innovative language. A similar situation applies to [r] and
[d], as the former phonemic status of [d] may be on the verge of being
neutralised in the innovative language, with only few distinctive examples
remaining (cf. §2.1.3). A possible factor for the loss of this distinction could lie
in the variation between [r], [d], and [t] in the various dialects, which may have
lead to different ways of pronunciation in some words. In addition, the
influence of Spanish may have contributed to an increased degree of confusion
between these consonants. For instance, the Spanish word lampadin 'oil light' is
realised as rafwari (cf. (1229)).
As a trend opposite to the loss of segments, loans are likely to support an
expansion of the phoneme inventory. While no new phonemes seem to have
been added in the past, the use of words borrowed from Spanish has a certain
impact. The details of this development are discussed in §23.3. Further
variations related to phonology are observed on the morpho-phonological level
- these are included in the following section.

23.2.2 Changes in morphology

As mentioned in §3.5.2, there is a trend towards a more analytic morphology, to

some degree. This is demonstrated by the gradual loss of proclitics, which are
being replaced by free pronominal forms. This applies to possessive pronouns
as well as to object pronouns, as is illustrated in (1216). In (1216a,b), the
occurrence of the lsg proclitic ka= in traditional narratives is shown, where the
storyteller imitates the speech of people in ancient times when he quotes them.
In contrast, (1216c,d) exemplify the use of full pronouns in the innovative
language ((1216c) is from a modern narrative and (1216d) from everyday
speech). For both varieties, the pronominal function in possessives and as an Ο
argument is given. Note however, that while these forms represent typical
examples for the use of bound vs. free pronouns, both types occasionally occur
in either of the varieties, which can be considered proof for the gradual change.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Traditional and innovative language 887

(1216) Loss of proclitics

a) Traditional language: possessive proclitic
ßäe hoereneto-nana-a ka=iri-uru ofioa
already be.fat-STPL-3ps/A lsg=creature-PL Dad
'Dad, my creatures are already fat [enough to be eaten].'

b) Traditional language: object proclitic

kwa ka=so-a-tge=ra
don't 1 sg=kill-NTR-PL=EMF
'Don't kill me!'

c) Innovative language: possessive proclitic

kamt atane ajriu kuane
lsg land PLN inside
'in my land, the Airico River'

d) Innovative language: object proclitic

kanu auna-i=ßa
lsg hear-2ps=INT
'Did you hear me?'

A similar tendency can be observed for the use of the pronouns with the
postposition raj, such as in ka=raj vs. kanu raj 'for me', 'to me'. In this context,
the variant with the proclitic is much more stable and thus more commonly used
by younger people, possibly because these forms tend to be lexicalised, as they
are rather frequent (for instance, karaj is also used as a politeness strategy; cf.
§13.5.1). Also recall that with kinship terms, body parts, and other formerly
inalienably possessed nouns, the degree of retaining the proclitic form is much
higher than with other possessives (cf. §7.4). Though less imminent, the gradual
loss of the feature non-alienability is a further example of syntheticisation in the
morphology. During the discussion of formerly inalienably possessed nouns it
was also indicated that words in the innovative language become longer through
the insertion of additional material - which might be a trigger for subsequent
splitting of words into free pronouns and nouns (cf. §7.1.5).
Other features related to possession involve the associative modifier b-.
However, for this morpheme it is unclear whether its presence is typical for the
traditional variety, as speakers make contradicting statements about this. On the
one hand, there is no sign that the grammatical function of b- is lost in the
contemporary language; on the other, its presence vs. absence in some lexical
items is inconclusive, as is discussed in §7.3.2.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
888 Variation

A further example for the loss of morphological features is the use of the
attitudinal markers =ra and =ta, as pointed out in §12.3.7, While these regularly
occur in the traditional language, they are not part of the contemporary language
any more, apart from a few dialects that have retained them (=ra in the Airico
dialect and =ta in the Corrientes dialect, see §23.1.12). Another enclitic that is
not used any more as a single marker of the innovative language is the
reassurance enclitic =tau - which now mainly occurs in combination with the
evidential markers =m and =he (cf. §12.3.3; §12.3.4).
A different kind of change is observed with the jussive form, where the
suffix -mii is lost in a manner of speaking, but replaced by another suffix with
the same function: in the innovative language, the jussive is expressed by the
suffix -ijie. The reasons for this are entirely unclear and even the comparison of
dialects is inconclusive since both forms are rather rare in occurrence and hardly
attested at all in the dialect database.
Another example of variation that does not involve a loss of morphological
features is the use of the interrogative enclitic =na in place of the focus marker
=te after the interrogative pronoun d^anu 'why' (cf. §21.1.4). However, while
younger speakers prefer =na as a tendency, there is a high degree of variation
for both forms. Also investigated in §21.1.10 is the formation of the
interrogatives ßaaohwa and ßata(ha) from the respective verbs. These new
forms can be used in the same way as the quantifying interrogative pronoun
cucmto(s) in Spanish and their formation may be interpreted as a result of
Spanish influence. In a similar way, the impact of Spanish on Urarina could be
made responsible for the introduction of a new question type, in which a
negative question may occur without the introducer ta, but with negative
inflection on the verb (illustrated in example (1166) in §21.3)
The reasons for the less frequent use of the politeness marker =t$e by
younger speakers, as reported in §13.2, can be characterised as a cultural
phenomenon, i.e. the absence of this morpheme in conversations with some
family members may reflect the fact that different cultural values are being
adopted. As these also include a degradation of respect and reverence towards
others, its impact on the use of polite forms is conceivable.

23.2.3 Changes in the lexicon

There is a range of lexicon entries that are attributed to the traditional language
by younger speakers. More precisely, most of these words appear not to be used
even by older speakers, whereas those terms were apparently used by earlier
generations. Thus, while they are still understood and occur in traditional
narratives, even older speakers do not actively employ these in everyday
speech. A list of examples of such words is given in (1217), each one with the
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Traditional and innovative language 889

old and the new version. As demonstrated in the list below, lexical pairs of
identical semantic meanings for the traditional and innovative variety are found
throughout almost all word classes. Many words do no exhibit any phonological
similarities to each other (cf. (1217a)), whereas others arguably have undergone
some phonological change but are still based on the same lexical item (cf.
(1217b)).

(1217) Lexical changes in different word classes

a) No similarity:
Traditional Innovative Gloss Word class
buunuutaa abaataa 'resist' Verb
erele atanahe 'mosquito' Noun
kuloronoi baaso 'useless', 'worthless' Adjectival noun
hautohwerti rauihirü 'really', 'in truth' Adverb
mtokwara hjauipera; aji 'no' Interjection
heeku arahü 'many' Quantifier

b) With phonological similarities:

Traditional Innovative Gloss Word class
kauatia hoataa 'so that not' Conjunction
rdhjaaoria hjauijie 'don't' Introducer
tQaisi tQuisi 'indeed' Particle
tQuahi tQoae 'on top', 'above' Postposition

More examples of words that were subject to some phonological or

morphological change are found in the class of verbs, as illustrated in (1218).
These changes, however, are not based on regular or transparent rules.

(1218) Non-regular phonological changes

haerakaa -> hakuraa 'dry well' (tr.)

buhuaenia -> buhuatia 'protect' (tr.)
netoaaeka wtohweeka 'disgrace oneself (intr.)

Other words have simply been "forgotten", i.e. they have not been replaced by
other words as the items they relate to are not in use any more, clearly due to
the abandoning of some cultural practices. Younger speakers of Urarina are not
familiar with the meaning of two examples given in (1219), while some older
speakers may still remember these items. In traditional narratives, these words
may still occur.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
890 Variation

(1219) "Forgotten" words

däato 'container for magic items'
uhuaanaa 'breast shield'

Yet another kind of change is the substitution of originally descriptive terms

with native words by Spanish loans. Traditionally, new items were expressed in
Urarina terms, i.e. through compounds. For instance, the word for 'shoe'
(Spanish ζαραίο) was described as "foot container" - a few more examples are
given in (1220). Out of motivation for simplicity, younger speakers nowadays
only know the Spanish terms and are unfamiliar with the traditional
expressions.

(1220) Descriptive terms replaced by loans

tihja kttru safiuato 'shoe' ('foot' + 'container')
enoto kuhwunaka relo 'watch','clock'('sun'+'signalise')
katQa kuhiottnaka ruhoera kamara 'photo camera' ('man' + 'signalise' +
'pull out'-AG (= "which pulls out
people's sign")

A particularly fascinating area are traditional animal names, which, similar to

the compounds shown in (1220), are morphologically complex terms with a
literal meaning that reveals how the respective animals' role or characteristics
were perceived. For instance, the two words for 'tapir' are entirely different, as
can also be seen in (1221): the traditional terms contain the no longer existing
term ate (roughly to be translated as 'big mammal'), which is homophonous
with the still current word ate 'fish'. Apparently, the former ate functioned as a
classifying noun for big, four-legged mammals. The term aräala 'tapir' is
entirely unrelated to this word - and the same observation applies to all other
animal terms.

(1221) Animal names with ate ('fish')

Traditional Innovative Gloss
ate kutaabahi aräala 'tapir' ("big fish")
ate kulareri uhuaae 'deer' ("red fish")
ate kuratua obana 'collared peccary' ("smallest of big fish")
ate mululutie raana 'white-lipped peccary' ("fish who walks in herd")

Also note that most traditional terms contain the associative prefix hi- (cf.
§7.3.3). Apart from this affix, the morphology of the traditional terms is not
always transparent, but their root is usually recognised by native speakers and

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Traditional and innovative language 891

thus can be assigned a meaning. For instance, in the term ate hitaabahi 'tapir'
the second word clearly contains the associative prefix followed by elements
that resemble the root / t a b a / for ' b i g ' . The example atemululune in (1221)
lacks the prefix hi-\ in (1222), one example occurs with the associative
modifying prefix b- (for no known reason).
Another group consists of big water-dwelling animals, including fish and
reptiles. These start with the class noun edara 'water people', for which some
background information is required. According to one traditional narrative, the
w o m a n Lomai once came to visit a group of Urarina people and danced with
them. When she had to urinate, they refused to let her go and so she flooded the
earth. As a consequence, all their possessions turned into water-dwelling
animals - for instance, the mosquito nets became paiche fish. The people
themselves became edara - water people. Exactly this is reflected by the
traditional terms displayed in (1222).

(1222) Animal names with edara ('water people')

Traditional Innovative Gloss

edara benamhja d^akari 'crocodile'
('water people' + ASCM-'canoe')
edara kivurukari ahaiße 'paiche fish'
('water people' + ASC-'mosquito net')
edara kwaamaa akauißo 'anaconda'
('water people' + ASC-'hammock')
edara kururihja nuri 'giant river turtle'
('water people' + ASC-'masato tub')
edara (ko)fwauafwa tariat$a 'yeHow-spotted river turtle'
('water people' + (ASC)-'bench')

Another class noun is found with two jungle-dwelling animals - a land turtle
whose name is associated with its shape ( ' b o w l ' ) and the two largest-sized
monkeys - therefore called 'chief of the j u n g l e dwellers'. Interestingly, 'howler
m o n k e y ' and 'spider m o n k e y ' , for which now separate (and quite distinct)
terms exist, were described by a single traditional term. The class noun itself
contains the current noun for 'jungle' nuna, with a suffix that could possibly be
identified as the agentive marker -era (which, however can only be attached to
verbs otherwise). In the traditional language, the word nunera referred to
' m o n k e y ' as a generic term.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
892 Variation

(1223) Animal names with nunera ('jungle dweller')

Traditional Innovative Gloss

nunera enejtQu 'monkey'
nunera kuraanaa ruru; 'howler monkey';
alau 'spider monkey' ("jungle dweller's chief')
nunera murae ahaaori 'turtle' ("jungle dweller's bowl")

The database contains a few more terms which do not contain any of the three
classifiers discussed above. These are listed in (1224). The traditional term
tahaaelijiaa (literally meaning 'small bird') was a generic term for 'parrot' and
is now represented by different names for each kind of parrot. The word for
'snake' (generic) is morphologically not entirely transparent but means "danger
of the jungle" according to one consultant. The term atajbißaae formerly
referred to 'jaguar', but now is used as a generic term for any animal, which can
include insects. The word for 'white monkey' is not transparent, but evidently
exhibits some phonological similarities between the old and the new form.

(1224) Other animal names

Traditional Innovative Gloss

tahaaelipaa e.g. soisoi 'parrot' ("small bird")
kabelaraj akano 'snake' ("danger of the jungle")
atajbißaae hanolari 'jaguar'
taurutau karitau 'white monkey'

Older native speakers declare that Urarina once had traditional terms for all
animals, but the ones listed here are the only ones they remembered. It is
unclear whether these are still used in ayahuasca rituals - a ceremonial occasion
in which apparently much of the traditional language is still used. However, the
data presented above brings up some challenging questions: Did the Urarina
have a "secret" language, or were the animal names used nowadays taboo and
not to be used? Otherwise it would be difficult to explain why the present terms
were introduced at all and where they came from. Further studies on these and
related questions should lead to interesting insights on some spiritual aspects of
Urarina culture which at this stage are not well-known.

23.2.4 Changes in syntax

In the previous chapters, also a few syntactic features were mentioned, which
differ from each other in the traditional vs. innovative language. One of these is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Language endangerment and Hispanicisation 893

the use of hau as a complementiser, as briefly repeated in (1225); also cf.

§20.2.4.3). This different subject complementation strategy is in addition to the
existing ones and may be the result of Spanish influence, as it largely
corresponds to complement clauses with (lo) que.

(1225) hau as a complementiser

ajto-a hau itQa-käu
say-3ps/D because do-lsg/A
Ί did what he said.' (Spanish Ηice lo que dijo.)

Related to that, hau is also commonly used in combination with the

postpositions raj ('for'), ajßa ('with'), and rihihei ('like') in the innovative
language (cf. Conjunction phrases in §19.3.2).
A further syntactic peculiarity observed only in the contemporary language
is the use of a verb stem without person inflection when followed by the
conjunctions ham or bana - as described in §20.1.2. This functions as an
alternative strategy to subordination with a conjunction, however, strictly
having different subject reference. As there is no apparent reason for the
addition of this construction to an already fairly wide inventory of multi-clause
constructions, this could be accounted for as another instance of simplification,
as the person suffix is elided.

23.3 Language endangerment and Hispanicisation

The results of language contact with Spanish has been previously mentioned in
a number of passages. Here I will summarise what impact Spanish as a
language with a high status has had on Urarina so far and stipulate which
further effects it may have in the future. Apparently, the influence of Quechua
was limited as not many traces from the period of contact are visible. Urarina
has borrowed a number of Quechua terms, including the numerals from six
onward. In a way, these have been integrated into the Urarina lexicon through
some phonological assimilation. However, their status as alien elements of the
language is still recognisable in that they are treated as nouns - different from
native numerals, which are verbs (cf. §5.12). The impact of Spanish - which
may also have had more time to take influence - is more serious. Worst of all,
the Spanish-speaking population has always conveyed the understanding that it
is superior to the speech of the indigenous people, whose languages are simply
described as dialectos, with the connotation that they are not real languages.
Even though the Urarina have managed to keep their language alive over the
past centuries despite the pressure of Spanish, the grammar has conceded to the

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
894 Variation

influence in certain ways. Instances of change under the influence of the former
colonial language can be observed in many areas of the grammar. Language
contact at the present stage largely occurs through the increased number of
mestizo settlers in Urarina-speaking areas and through the intrusion of traders
and other groups with commercial interests.
The strongest impact of Spanish undoubtedly is on the lexical level, as an
increasing number of items are named according to the local Spanish term.
However, borrowings are not restricted to nouns, as other word classes are open
to new words as well. (1226) gives some examples of loans from different word
classes. Sometimes loans involve more than single lexemes, as entire
expressions or idioms are copied.

(1226) Loans in different word classes

Urarina Spanish Gloss Word class
arusu arroz 'rice' Noun
säato santo 'holy' Adjective
kanaa ganar 'earn' (money) Verb
sikjera, heriane quisiera 'perhaps' Adverb
pero, nete pero 'but' Conjunction
dosmildos dosmil dos '2002' (year) Numeral
metoro metro 'meter' Measuring unit
a las mvebe a las nueve '(at) nine o'clock' Temporal expression
sotnaarja Jesus Maria 'Oh God!' Idiom
pormaskesia por mas que sea 'it should rather be' Expression/Idiom

The adoption of loans covers a range of semantic areas. Naturally, the majority
of these refer to innovations of some kind. New words were thus adopted to
describe items that did not exist earlier. This typically includes terms to
technical innovations, clothing, household items, or tools and manufactured
materials. However, also words that describe abstract concepts such as 'colour'
or 'size' are adopted (cf. (1227)). Remarkably, the word for 'size' is formed
from the root /taba/ for 'big' - which probably is a coincidence. Other terms
include words for plants not known before contact with Spanish was made, plus
some nature-related terms such as 'field' and 'meadow' - a geophysical
appearance originally unknown to the Urarina. Further loans refer to social and
spiritual aspects of life. For instance, neither the event of 'party' as such is
common to traditional lifestyle, nor the idea of keeping a boy as a helper - as
the former patrones did. A range of further new words are adopted through
contact with Christianity. While the NT translation attempts to use Urarina
terms to describe new concepts, this is not possible in all cases; this is

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Language endangerment and Hispanicisation 895

particularly difficult with some terms referring to concepts of Judaic ancient

history, as exemplified by the words for 'king' and 'angel'. The table in (1227a)
lists nouns for each semantic type. Borrowed verbs cover a similar semantic
range, as is illustrated in (1227b). 106

(1227) Semantic types of loans

a) Nouns
Urarina Spanish Gloss Semantic area
shenela shinela 'sandal' Clothing
torusa truza (trousers) 'short pants' Clothing
motosjera motosierra 'chainsaw' Technical
relo, relö, relor reloj 'watch', 'clock' Technical
kalamina calamina 'zinc roof Innovations
d^abi llave 'key' Innovations
kulu color 'colour' Abstract
tabajpe tamano 'size' Abstract
pasto pasto 'meadow' Nature
käafwo campo 'field' Nature
kasju casho 'cashew' Plants
seboja cebolla 'onion' Plants
besta fiesta 'party' Social
mosaso muchacho 'assistant', 'helper' Social
iglesia iglesia 'church' (institution) Christian
teefwolo templo 'church' (building) Christian
rej rey 'king' Christian/Historic
äahera angel 'angel' Christian/Spiritual

5) Verbs:
Urarina Spanish Gloss Type
grabaa engravar 'record' Technical
kastaa gastar 'spend' (money) Innovations
kasetekaa castigar 'punish' Social
bautisaa bautisar 'baptise' Christian

It is plausible to assume that the terms listed in (1227) were introduced because
no equivalent Urarina expressions were available. The examples shown in
(1228) differ from this, as these terms were adopted from Spanish even though

106
Note that the local version of Spanish makes use of the phoneme /// in words such as
shinela 'sandal' and casho 'cashew'.
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
896 Variation

corresponding expressions exist in Urarina. As the Spanish ones are becoming

increasingly more popular, the Urarina terms are likely to disappear.

(1228) Alternative terms

Urarina term Borrowed term Spanish Gloss
laenaa eskribia escribir 'write'
bitoa träafwa trampa 'trap'
itgasu ermano hermano 'brother' (spiritual)
kwerene asejte aceite 'oil'
atura borosa bolsa 'bag'
amäa amaka hamaca 'hammock'
tabake abese a veces 'sometimes'
riete sino sino 'but'
sajhjet aaüke aunque 'although'
letQurjka lejhia dose once 'eleven'

The introduction of words from Spanish comes along with a range of

phonological peculiarities which have an impact on the realisation of the
realisation of these words in Urarina. Naturally, there is a high degree of
variation between different speakers, depending on their level of proficiency in
Spanish. Urarina speakers with only basic knowledge of Spanish and with
relatively little contact to Spanish speakers are likely to adapt Spanish words to
the Urarina phonology. This is usually also true for terms that have been
borrowed a long time ago, such as saabere 'machete' and hoata 'metal' (also
'money'), which were probably borrowed centuries ago. In these words,
phonological differences are replaced by elements that exist in the native
phonology. For instance, the consonant cluster in [sa.ble] is not retained, but an
additional syllable is formed through epenthesis of /e/, in order to comply with
the CV-type syllabic requirements of Urarina. Similarly, the initial cluster in
[pla.ta] is assimilated to /kw/ - an existing phoneme of Urarina. Interestingly,
the consonant /!/ in Spanish is often replaced by Irl in the Urarina form, as is
also demonstrated by the examples for 'lemon' and 'oil light' in (1229). The
reason for this is unclear, as Urarina has the phoneme /l/, and its use as syllable
onset is not restricted in native Urarina words (cf. §2.1.13). A possible factor for
the alternation is that the occurrence of /I/ is comparatively rare in native words
- about 5% as opposed to /r/, which is the most frequent consonant (17%).
However, there also are loans in which /l/ is retained, as the example Unterem in
(1229) shows. Further modifications that occur quite regularly include the
realisation of /p/ as /fw/ and the pronunciation of nasal consonants in coda
position as nasalisation on the syllable core, such as in lampadin -> [rafivari].

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Language endangerment and Hispanicisation 897

The Spanish voiced velar /g/ is usually devoiced to /k/ in Urarina. Furthermore,
syllable final consonants other than nasals are typically elided, such as the /r/ in
ganar 'earn'.

(1229) Phonological adjustments of borrowed terms

Urarina Spanish Gloss Adaptation process

saabere sable, machete 'machete' /ble/ -» /be.re/
hvata plata 1. 'metal', 'iron'; /pla/ /kwa/
2. 'money'
remöo limon 'lemon' /I i/ /re/; /mon/ -> /mö/
räfwafi lampadin 'oil light' /lam/ -> /rä/; /paJ -> /fwa/;
/din/ /rV
Unterem Iinterna 'flashlight' /ter/ -> /te.re/; /na/ -> /no/
kanaa ganar 'earn' (money) /g/ -> /k/; /nar/ -> /naa/

Most younger speakers become increasingly fluent in Spanish and quickly adapt
its phoneme inventory, which is manifested in their realisation of Spanish
borrowings. In fact, the pronunciation of proficient Urarina speakers of Spanish
hardly differs from that of the local Spanish. This implies significant
phonological changes with respect to Urarina phonology, as is illustrated by the
examples in (1230). Some of the most noteworthy factors are the use of coda
consonants and the formation of consonant clusters. In addition, " n e w "
consonants such as /p/ and /g/ are being imported into Urarina phonology.

(1230) Phonological innovations

Urarina Spanish Gloss Innovation

linter ma a linterniar 'hunt with flashlight' /n/ in coda
kartutQo cartucho 'rifle shell' Irl in coda
plato plato 'plate' cluster /pi/
projekto proyecto 'project' cluster /pr/

Also note that to each borrowed noun, a tone pattern is assigned. It appears that
tonal class A (cf. §4.1) is a kind of default class for new words, as the majority
of loans exhibit the tonal features of this class. Another large proportion of
loans behave according to Type D, often copying the intonational pattern of
Spanish words. For instance, the loan posta ('medical post') has penultimate
stress in Spanish, which is interpreted as a Η tone on the penultimate syllable in
Urarina. Before a modifier, the tone shifts to final position, thus behaving
according to tonal Type D (subclass DP-f). However, other Type D loans may
behave differently with many variations and irregularities, such as preserving

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
898 Variation

the penultimate Η tone even in noun-adjective constructions. Other tonal types

(B, C) are extremely unusual.
The introduction of new phonemes and additional syllable types into Urarina
will become more significant as more speakers adopt these phonological
structures. A particularly interesting development may be expected for the
"striving" consonant pairs [d] vs. [r] and [fw] vs. [hw], as Spanish distinguishes
/d/ and Irl as well as /f/ and /hw/ (the latter in words with /hu/). It is conceivable
that this factor has actually contributed to the development by which the
distinction of these sounds in Urarina is lost.
Hispanicisation has had little if any direct impact on Urarina morphology up
to this stage. Nonetheless, the loss of certain grammatical categories that tend
not to be marked any more could be accounted for by the fact that Spanish lacks
these categories. As described in §23.2.2, the forms not used any more by
younger speakers, such as procliticisation and attitudinal markers, also happen
not to be marked by Spanish. However, given such a complex polysynthetic
morphology as Urarina, one might expect more drastic changes to appear -
which is not the case at this stage. An example for the introduction of new
features - here manifested by a change of word class (along with morphological
reduction) - is the formation of the interrogative pronouns ßaaohwa and
βata(ha). As mentioned already, these are not loans but their use corresponds
to the way cuanto(s) is used in Spanish.
The impact of Spanish as a possible trigger for some syntactic changes has
been mentioned in §23.2.4. These include the introduction of new strategies for
complementation and subordination. A further innovation to be mentioned here
is the use of some coordinating conjunctions from Spanish. This is plausible
since Urarina for the most part lacks any words to fulfil this function (apart
from t$ae 'also', which is more of an adverb rather than a conjunction and nete
'but', which is morphologically complex, at least from a diachronical point of
view; cf. §20.5.2). As a consequence, the Spanish clause coordinatorspero and
sino (both meaning 'but'; sino in a counterfactual function: 'not this but that')
are sometimes observed in younger speakers' speech. This is exemplified in
(1231). Note that in (1231a), pero is combined with nete - which is an optional
solution, as pero can also occur by itself. The use of aaüke in this sentence
appears to take the function of a kind of clause introducer in this case. Example
(1231b) illustrates the use of sino 'but', which is used between two independent
clauses where the first clause is negative.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Language endangerment and Hispanicisation 899

(1231) Innovative forms of 'but'

a) Pero\
aaüke nijej heriane tabauru u-i, pero riete
though not.at.all maybe some come-NEG:3ps/A but but

kana+hoaaun-era kana kuruatahani-a nerajjie

Our.creator lpl/in help-3ps/D must
'Even though some [people] maybe have not come at all, ("but") God must help us.'

b) Sino:
alasjete ke ma natii ere-i=ße hoatia kauatQa-i, sino
seven.o'clock VLI even speak-2ps=CND not good-NEG:3ps/A but

alasjeteimedia ke ere-i=ße=te rautohwe-ι ere-ri-tQaaka=m

half.past.seven VLI speak-2ps=CND=FOC be.calm-PRT speak-IRR-lpl/du=ASS
'If you talk [over the radio] at seven o'clock, it is not good, but if you talk at half past
seven, we can talk in tranquillity.'

One might expect that the peculiar constituent order of Urarina would also be
subject to pressure from Spanish (a notorious A V O / S V language), but
significant changes to constituent order in Urarina are not observed. As
mentioned in §18.3, there are a few isolated examples of an S or A argument
occurring in preverbal position that cannot be accounted for in terms of the
predicted features (focus, emphasis, negation). Beside that, in one of the dialects
investigated further above (Copal), two examples with an Ο argument in
postverbal position were observed. While such examples are extremely rare,
one could of course attribute these to the influence of Spanish.
O n e innovation in constituent order can be observed with the position of
complement clauses produced by younger speakers. Example (1232) illustrates
a complement clause type not found with speakers of the older generation. In
this case, t w o deviations from other complement clauses occur: a) The
complement clause follows the control verb, rather than preceding it. b) The
control verb is impersonal in that it is marked for 3ps. Thus, the literal
translation of the example could be given as 'Would it be possible that you
repeat it?' - which could be interpreted as a caique from Spanish.

(1232) Order variation in innovative complement clause

kaa ajto-huau hau auna-i=ßa

this say-lsg/D because hear-2ps=INT

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
900 Variation

aj-i, najße-re repete-i

AUX-NEG:3ps/A be.able-IRR:3ps/E repeat-2ps
[Speaker 1]: 'Have you heard ("this") what I said?' [Speaker 2]: 'No, can you repeat it?'

Further analysis of texts by young speakers will be required to determine

whether an actual trend towards an AVO/SV constituent order can be detected.
Other grammatical features already discussed elsewhere include the gradual
loss of marking for inalienably possessed nouns, which could well be related to
the absence of this feature in Spanish (cf. §23.2.2). The changes with respect to
politeness marking, as mentioned in the same section, may be interpreted as a
reflection of the loss of certain cultural values in the younger generation. It
appears that such changes based on the assimilation of Urarina culture are
unavoidable in the not too distant future. While the foundations for this study
have still been laid on time in order to acquire a basic understanding of Urarina
grammar, a lot of irretrievable information has already been lost. It is my hope
that this book provides a stimulus to readers as to get engaged in further
research on this fascinating language.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
Appendix A

Selection of texts

The following five texts were selected as to exemplify the use of words and
sentences in a wider context. Recordings for each text can be found on the CD
accompanying this book. The texts included here are of different genres and
involve two short third person experience narratives (i.e. non-self experienced
events), one first person (self-experienced) narrative, one traditional narrative,
and one instruction. Note that the numbering of the texts is according to their
number in the database. The speakers of these texts can be characterised as
follows:
Track 01+02 - Texts 8+10 107 : Humberto Nuribe Arahuata (23) has been living
in the Espejo for 18 years. His father is the village elder (Julian Nuribe,
mentioned in §1.6) and his mother is from San Lorenzo (Tigrillo). He is
bilingual and has ambitions to become a storyteller, which is why he
collects stories told by others and has written down some of them. His
narratives were chosen as representative examples for narratives told by the
younger generation.
Track 03 - Text 53: Vicente Arahuata Manizari (42) grew up in San Lorenzo
(Tigrillo) and moved to the Espejo with the team of founders 18 years ago.
His father was from the Pucayacu, his mother from San Lorenzo. Vicente
now is the mayor of Nueva Union and one of my regular consultants. His
text represents a narrative told by a speaker "between" the younger and
older generations. It is particularly interesting with regard to the forms used
to refer to his in-law (cf. § 13.3).
Track 04 - Text 25: Medardo Arahuata Manizari (Vicente's older brother), is
the main storyteller and was mentioned in §1.6. This text is one of the
shorter traditional narratives.
Track 05 - Text 86B: Anita Macusi Nuribe (ca. 32) was born in Santa Rosa de
Airico by parents from that area and has been living in the Espejo for about
12 years, since she got married to one of the local residents. The recording,
which was conducted by Michaela Olawsky in 2004 is a typical example
for an instruction, here on how to make cassava beer. Being one of the few
texts collected from women, it also demonstrates that there are no relevant
differences between the speech of either sexes.

107
The text numbers are based on their occurrence in the Urarina database (cf. §1).
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
902 Appendix A: Texts

Text 8 (Track 01: Third person experience narrative),

by Humberto Nuribe Arahuata, 2001

(1) hauißenaa he ere-naa

lost.person VLI speak-NOM
'Story of a lost person'

(2) lejhii kat$a=te miirjka ke katga-uru kuti-hjai amia-rate,

one man=FOC minga VLI man-PL call-PRT work.CAU2:3ps/E

utene-kuru-a=ne kujjia kuriae sauki ora-ϊ

clear.ground-PL-3ps/D=SUB so.that distance five hour-PRT

amu-naanaha rn-i enamhja kuatie

walk-LPP be-PRT canoe Inside
One man invited the people to a minga [working session] and made them work,
so that they would clear the ground [for his field], [which was] about five hours
of travel by canoe.'

(3) mi kahe=te turu-a u-e sonajßa ßäe hauim-a

that from=FOC arrive-NTR come-3ps/E afternoon already get.lost

raj kakunu=ne hana

POSS daughter=SUB when
'He arrived [back] from there in the afternoon, when his daughter had got lost.'

(4) nii hau=te kujßadera-ι ari-tgate

that because=FOC be.worried-PRT seek-CAU2:3ps/E
'Therefore, he was worried and had her sought.'

(5) mi häu=te ari-ι muht-ure suru hana

that because=FOC seek-PRT catch-3pl/E run when
'Therefore, they sought her and caught her when she ran away.'

(6) muku-uru-a=ne hana=te kuhjuri

catch-PL-3ps/D=SUB when=FOC scream:3ps/E
'When they caught her, she screamed.'

(7) nii hana=te muku-i ra-e raj neba

that when=FOC catch-PRT receive-3ps/E POSS mother
'Then her mother caught her and took her.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 903

(8) raj paka-netonaj=te ajßa muku-e

POSS father-also=FOC with catch-3ps/E
'Her father also caught her together with [the mother].1

(9) aladino na-a raj ßaka kuraa

PSN say-3ps/A POSS father name
'Her father's name is Aladino.'

(10) marga na-a mi hauijienaa kuraa

PSN say-3ps/A that lost.person name
'The lost person's name is Marga.'

(11) ßäe sa-a

already end-3ps/A
'[The story] is over.'

Text 10 (Track 02: Third person experience narrative),

by Humberto Nuribe Arahuata, 2001

(1) n-itQafioa-a katg,a=ne ke ere-naa

ITR-shoot-3ps/D man=CND VLI speak-NOM
'Story of a man who shot himself

(2) hano-a ku=na hau=te n-itgafwa-e lejhti katQa,

hunt.overnight-NTR go=SUB because=FOC ITR-shoot-3ps/E one man

santjago na-a raj kuraa

PSN say-3ps/A POSS name
Ά certain man shot himself as he went hunting overnight; his name is Santiago.'

(3) nii=te raana ari-a ku-e kuriae kuruataha-j

that=FOC w.l.peccary seek-NTR go-3ps/E distance be.two-NOMsbj

katQa ajßa, beru kuane ku-he-i

man with way inside go-CNT-PRT
'He went to seek white-lipped peccaries far [from here], with two men, [and they
were] walking on a path.'

(4) nii baja=te beru edaa eno-1 hißori kari

that after=FOC way outside enter-PRT hinyoriXxtt bark

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
904 Appendix A: Texts

rulu-e, raj itQafioa d^uhuta-1

strip.off-3ps/E POSS rifle lean-PRT
'After that, he left the path and stripped off the bark of a hinyori tree [in order to
make a basket from it], leaning the rifle [against the tree].'

(5) nii hana=te raj n=itg,afioa tgoae kodoonete

that when=FOC POSS 3ps=rifle above heap.up:3ps/E
'Then [the bark] was heaping up on his rifle.'

(6) nii häu=te nofoite raj n=itQafiua rüa

that because=FOC go.off:3ps/E POSS 3ps=rifle side

raj tihja d$uhueka-a=ne hana

POSS foot make.stand-3ps/D=SUB when
'Therefore, his rifle went off when he put his foot next to it.'

(7) nii häu=te raj tihja halalti-1 situ-e monisiö

that because=FOC POSS foot make.hole-PRT pass-3ps/E ammunition
'Therefore, the bullet went through his foot, making a hole.'

(8) nii hau=te auake

that because=FOC shout:3ps/E
'Therefore he cried.'

(9) nii häu=te surunei ku-t huara-a ku-ure

that because=FOC quickly go-PRT see-NTR go-3pl/E
'Therefore, [the others] ran quickly and went to see.'

(10) kwara-l=te nesoone-to-ore

see-PRT=FOC be.scared-INTS-3pl/E
'They saw him and were very scared.'

(11) nii baja=te ruaka-1 ama-ure raj loanari asae

that after=FOC carry-PRT take-3pl/E POSS shed under
After that, they took him along carrying him to their shed.'

(12) nii häu=te hatäi naahetokurü nerauta-e

that because=FOC very barely get.better-3ps/E
'Therefore, he was barely getting better.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 905

(13) kahianei=te ama-i hoitga-rate-kure

everywhere=FOC take-PRT heaI-CAU2-3pl/E
'Taking him everywhere, they had him healed.'

(14) haurianehe~i=te konkordia ama-ure

first=FOC PLN take-3pl/E
'At first they took him to Concordia.'

(15) ra'i baja=te ama-ure majpuku-u

that after=FOC take-3pl/E PLN-LOC
'After that, they took him to Maipuco.'

(16) satonohei=te ikito-o ama-ure, huitga-kuru-a dokotoro-kuru-ne kujpa

at.Iast=FOC PLN-LOC take-3pl/E heal-PL-3ps/D doctor-PL=SUB so.that
'At last, they took him to Iquitos, so that the doctors would heal him.'

(17) huusa atene-i=te ne-rauta-e mi kunajte-na

eight month-PRT=FOC get.better-3ps/E that be.sick-NOM
'That sick person got well in eight months.'

(18) ßäe sa-a

already end-3ps/A
'[The story] is over.'

Text 53 (Track 03: First person experience narrative),

by Vicente Arahuata Manizari, 2003

(1) kaami kaana ajßa le=sahi kuane amu-a ku-akaanu

GPF son-in-law with one=time inside walk-NTR go-lpl/ex

kaana ajßa kuniniko b-atane ke, ahäaori ari-a

son-in-law with PLN ASCM-land VLI turtle seek-NTR

ku-akaanu hau
go-lpl/ex because
'Once, with our son-in-law, we went hunting at Cuninico hill with our son-in-
law, because we were going to look for turtles.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
906 Appendix A: Texts

(2) haurianehei ku-akaanu, ahäaori ari-a ke nekureti-a

first go-1 pl/ex turtle seek-NTR VLI sell-NTR

ku-akaanu ot$äata=ne kujßa

go-1 pl/ex PLN=SUB so.that
'At first we went, we went to seek turtles in order to sell them in Ollanta.'

(3) mi hana ku-Ί, ßäe kaa kahe ku-akaanu=ne ham ßäe,

that when go-PRT already this from go-lpI/ex=SUB when already

kuniniku b-atane ku-fioauana turu-Ί sim-akaanu

PLN ASCM-land ASC-harbour arrive-PRT sleep-1 pl/ex
'Then, going, when we went from here, we arrived at the harbour of Cuninico
hill, we slept.'

(4) ku hanotü hanoahel ku-1, sim-akaanu

there early.morning all.night go-PRT sleep-1 pl/ex
'Travelling all night, we slept there in the early morning.'

(5) mi hana=te ku-hi-akaanu hana kanakaanu ahinia

that when=FOC go-CNT-1 pl/ex when 1 pl/ex before

lunuako-a d$akari hau, su-akaanu

stick.out.of.water-3ps/D crocodile because kill-1 pl/ex
'Then, when we went, as a crocodile was sticking [its head] out of the water, we
killed.it.'

(6) su-i, ama-1, nal nitaa ßäe ku hajti

kill-PRT take-PRT over.there over.there already there still

turu-elanaala, ke isi moko-akaanu

arrive-PRV INST fire burn-1 pl/ex
We killed it and took it along over there, to that area, and still not arriving [at our
destination] we grilled it on the fire.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 907

(7) isi moko-1 ku haree-hel, ki-1, najpai he lewne-1,

fire burn-PRT there make.soup-PRT eat-PRT nicely VLI eat-PRT

ßäe kahe kwaaune-Ί ku-akaanu hanotii hanoahei ku-akaanu

already from create-PRT go-lpl/ex early.morning all.night go-lpl/ex
'We made a fire and prepared a soup there and ate [the crocodile], making a nice
meal of it, we went from there in the early morning, going all night.'

(8) hanoahei ku-1 ßäe kunimku b-atane ku-fioauana turu-1

all.night go-PRT already PLN ASCM-land harbour arrive-PRT

ku hanotii turu-1 ßäe ku sini-akaanu hioaaerel,

there early.morning arrive already there sleep- lpl/ex somehow

kani
sort.of
'Going all night, already arriving at the harbour of Cuninico hill, arriving there in
the early morning, we slept there somehow, like that.'

(9) haurane tgoae sini-akaanu

leaves above sleep-1 pl/ex
'We slept on leaves.'

(10) sini-1 kohioanoo, neloerite-1, kahe hoaaune-1 amu-a

sleep-PRT next.day build.shed-PRT from create-PRT walk-NTR

ku-akaanu
go-lpl/ex
'We slept and on the next day we built a shed and went hunting, starting from
there.'

(11) amu-a ku-akaanu ßäe, ku ßäe kumniku b-atane ke

walk-NTR go-lpl/ex already there already PLN ASCM-land VLI

ini-1, kwara-kaanu=ne hana ku kureote ahäaori

go.up-PRT see-lpl/ex=SUB when there leave.track:3ps/E turtle
'We went hunting already, and when we looked, going up Cuninico hill turtles
had left their tracks there.'

(12) kureoto-a ahäaori=ße hau ku ari-akaanu

leave.track-3ps/D turtle=SUB because there seek-lpl/ex
'As turtles had left their tracks, we sought there.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
908 Appendix A: Texts

(13) kanakaana ita ke herahehe ni-tQaanu=ne päe ahäaori

lpl/ex REC VLI in.two be-lpl/ex=CND already turtle

ru-hi-akaanu
find-DIM-lpl/ex
'When we each were in pairs, we already found a few turtles.'

(14) nii hana=te kaana hana tQäe nijej ru-si-pana-ene

thatwhen=FOC son-in-law inside also not.at.all find-CPL-ILT-NEG:3ps/E
'But then, my son-in-law did not find any at all.'

(15) ru-si-pana-em-a=ne hau, nii rihihe-1 kohwanoo kzvajtei

find-CPL-ILT-NEG-3ps/D=SUBbecause that be.like-PRT next.day again

ari-a ku-akaanu
seek-NTR go-lpl/ex
'As my son-in-law did not find any, like that, we went to seek again the next
day.'

(16) kwajtei kohwamo ari-a ku-akaanu hau tabiitga lejhü

again next.day seek-NTR go-lpl/ex because finally one

ru-hi-ana-e
find-DIM-ILT-3ps/E
'As we went to seek again the next day, he finally found one.'

(17) lejhii ru-hi-ana-e, ru-hi-ana-l, ku nit$ata hano-t

one find-DIM-ILT-3ps/E find-DIM-ILT-PRT there three day-PRT

hajti ku ari-akaanu
still there seek-1 ρ 1/ex
'He found one; finding it, we still sought there for three days.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 909

(18) ku ari-akaanu nitgata hano-1, hana ßäe, ßäe tg,u hjä

there seek-lpl/ex three day-PRT inside already already CRTN just

itQuneheori-a eresi ne-hi-a kau-akaanu=ne raj,

be.pretty.close-3ps/D tomorrow be-DIM-NTR return-lpl/ex=SUB for

kwajtel rtahvaaunei ari-a ku-akaanu

again again seek-NTR go-lpl/ex
'Seeking for three days, with this, the next day already was pretty close for us to
return, [but] again we went to seek.'

(19) mi hana=te kuaae näi kanakaanu hiuuite, näi

that when=FOC over.there over.there lpl/ex cover:3ps/E over.there

atane kataa ni-akaanu=ne hana, elo

land centre be-lpl/ex=SUB when rain
'Then, over there in that place, when we were in the centre of that area over
there, the rain came over us.'

(20) nii hau ßäe nii ora hvituku-eni-akaanu=ne hau,

that because already that hour know-NEG-lpI/ex=SUB because

kuaae sonafwa-akaanu, kuaae sonafioa-akaanu

over.there be. late-lpl/ex over.there be. late-lpl/ex
'Therefore, because we did not know the time, we ran late over there, we ran late
over there.'

(21) nii hau d^atQU ßäe nii enoto hvara-eni-akaanu=ne

that because suddenly already that sun see-NEG-lpl/ex=SUB

hau kuaae sonafwa-akaanu

because over.there be.late-lpl/ex
'Therefore, as we suddenly did not see the sun, we ran late over there.'

(22) kuaae sonafioa-l ßäe, hajti hetga ekuu ne-te enoto

over.there be.late-PRT already still maybe top be-INTS:3ps/E sun

na-1, kuaae hajti ahäaori ari-akaanu

say-PRT over.there still turtle seek-lpl/ex
'Running late over there, saying, "maybe the sun is still up there", we were still
seeking turtles.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
910 Appendix A: Texts

(23) mi hana=te ßäe d^atgu heriane su-e ßäe

that when=FOC already suddenly maybe dawn-3ps/E already
'Then, suddenly it already it dawned indeed.'

(24) ßäe hajti ßäe nal atane katäa hajti, itgiiuti-Ί hajti
already still already over.there land centre still be.close-PRT still

ne-ni-akaanu, hiiri ku buhua-u nai atane ke

be-NEG-lpl/ex swamp there behind-LOC over.there land VLI

ni-akaanu=ne hana ßäe, ßäe nukue-na here

be-lpl/ex=SUB when already already be.dark-INF want:3ps/E
'Already, still in the centre of that area, we were still not near [our shed]; when
we were behind the swamp, it already wanted to become dark.'

(25) hana ßäe, d$ahama-u=te kau-ni-tQaatge heriane su-a,

inside already come.on-IMP=FOCreturn-DSTL-lpl/exmaybe dawn-3ps/A

na-akaanu, na-a=ne häu, ßäe kau-a u-akaanu

say-lpl/ex say-3ps/D=SUB because already return-NTR come-lpl/ex
'Then, already, we said, "let's go home, maybe it's becoming dark" and as we
said so, we went home.'

(26) kau-a u-akaanu hana=te ßäe nii ßäe hiiri hana

return-NTR come-lpl/ex when=FOC already that already swamp inside

fioau-ahi-akaanu=ne hana ßäe, ßäe nalaakoe-nu tokoane-1

descend-CNT-lpl/ex=SUB when already already fall-PRIR compare-PRT

ßäe ku-e sona

already go-3ps/E evening
'When we went home, when we already were going down that swamp, it already
seemed as if the evening [i.e. the darkness] fell [over us].'

(27) nii häu=na ßäe nesoone-to-akaanu eruhioel

that because=FOC:lpl already be.scared-INTS-lpl/ex quickly

kau-a u-akaanu
return-NTR come-lpl/ex
'Therefore, we were very scared; we quickly went home.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 911

(28) nii hana=te ßäe banetohwelanaala kukoo-kaanu ßäe,

that when=FOC already totally be.exhausted-lpl/ex already

amuneto-hwelanaala ßäe kuna-ahua-a kanakaanu tihja

walk-INTS-PRV already hurt-DSTR-3ps/A lpl/ex foot
'Then we already were totally exhausted, our feet hurting [so much that we]
could not walk.'

(29) nii hana ßäe, nii hana=te ßäe t$äe mihiite-na

that when already that when=FOC already also be.hungry-NOM

käu nihianaj ßäe ateruna-ana-e ßäe hatai ßäe kanii

because.of all.that already be.tired-ILT-3ps/E already very already GPF

d$a ama-ene netohwe-ni-akaanu

what take-IRN be.in.the.mood-NEG-lpl/ex
'Then already, then because of hunger and all that, my son-in-law already was
very tired; we were not in the mood to take along anything.'

(30) mi häu ßäe hiiri katäa kanakaanu-at$a koene raatiri-1

that because already swamp centre lpl/ex-only drink leave-PRT

atgano raatiti-1 aj-akaanu häu, ßäe atii ku

travel.food leave-PRT AUX-lpl/ex because already nevertheless there
turu-1 akau ku-a u-akaanu
arrive-PRT water drink-NTR come-lpl/ex
'Then, as we had left our drink and travel supplies in the middle of the swamp,
we arrived there nevertheless and came to drink chapo.,i01

(31) nii häu=te ßäe hatäi ßäe nitahe-naa here enoto

that because=FOC already very already get.lost-INF want:3ps/E sun
'Then the sun wanted to disappear very much.'; [To mean 'it really got almost
dark']

107
Chapo is a drink made from cooked plantains. The word akau, literally meaning
'water' can be used to refer to any kind of drink.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
912 Appendix A: Texts

(32) ßäe hatai nunaa asae tihja najßehe-ι ßäe, hatäi ßäe
already very jungle under foot finish-PRT already very already

situna-a ku-hi-a enoto=ne hau, ßäe kahe hoaaum-i

sink-NTR go-CNT-3ps/D sun=SUB because already from create-PRT

kau-a u-akaanu hana=te hoaraha-eni-a hau

return-NTR come-lpl/ex when=FOC bridge-NEG-3ps/A because

tgauhiva-a hiiri hau, amu-si-ni-akaanu tonoana ßäe

be.ugly-3ps/D swamp because walk-CPL-NEG-lpl/ex while already

nukue-kaanu ßäe hiiri katäa

stay.overnight already swamp centre
'As the sun was going down very much, sinking below the trees [lit' "finishing
her foot under the jungle"], when we went home from then on, because there
was no bridge [i.e. no poles to step on] and because the swamp was ugly, while
we could not walk, we already spend the night in the middle of the swamp.'

(33) hiiri katäa ku-hi-akaanu hana, kanaka ana ahinia alaa

swamp centre go-CNT-lpl/ex when lpl/ex before aguaje.tree

hatgu kuane tutatahane kaanii anuri raj kalaui ig,oae

hole inside stick.out:3ps/E sort.of parrot POSS son above
'As we went in the middle of the swamp, , in front of us was this sort of parrot,
sitting on its creatures in the hole of an aguaje tree.'

(34) nii hana=te hu-a ku-ana-e atii mi sorta

that when=FOC fell-NTR go-ILT-3ps/E nevertheless that evening
kwara-hi-ana-a=ne asajhjei, hu-a ku-ana-e
see-CNT-ILT-3ps/D=SUB although fell-NTR go-ILT-3ps/E
'Then my son-in-law went to fell it even though he was seeing that it was late
[lit. 'he saw the evening']; he went to fell it.'

(35) nii hana=te ßäe nukue-hei ßäe kuu-he

that when=FOC already be.night-PRT already go-CNT:3ps/E
'Then it was already going to become night.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 913

(36) ku-hi-a hana, he d$aha-ohiua-u, d$aha-ohwa-u=te

go-CNT-3ps/D when hey come.on-DER-IMP come.on-DER-IMP=FOC

d$aha-i kau-ni-tQaatge na+na-a kaana

come.on-PRT return-DSTL-lpl/in RED+say-3ps/D son-in-law

ukwala=ne sajhjet, ku hajti hu-ahi-ana-e

younger.brother=SUB although there still fell-CNT-ILT-3ps/E
'When it was going [to be night], even though my son-in-law's younger brother
was saying "Come on, come on, let's go home!" he still kept on felling.'

(37) mi hau ßäe najßa-ι, hu-Ί atii ßäe nii

that because already nicely fell-PRT nevertheless already that

ke=ne ßäe nukue-tg>a-kaanu ßäe hatal

INST=FOC:lpl already be.night-only-lpl/ex already very

kauatga najßehel
pretty, much
'Therefore, felling [that tree] thoroughly by all means - with that went pretty
much into the night.'

(38) ßäe nijej ißaka-i beru kuane

already not.at.all know-IMPS-NEG:3ps/A way inside
'It was already impossible to know the way.'

(39) mi hau atü, somaraha+somaraha-a karni-uru hau

that because nevertheless RED+be.white-3ps/D sort.of-PL because

kahe kwara-he-i, atii kau-a u-akaanu

from see-PRT-PRT nevertheless return-NTR come-lpl/ex
'Therefore, nevertheless, because we were seeing some sort of white things [to
mean 'fluorescent'], we went home nevertheless.'

(40) mi hana=te kanaka ana ajßa u-e kaami lejhn kanaanaj

that when=FOC lpl/ex with come-3ps/E GPF one child
'Then, there [also] had come with us a certain child.'

(41) nii häu=te ßäe äkä=etonaj ßäe aka ßäe wane

that because=FOC already 3sg=also already 3sg already PTY

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
914 Appendix A: Texts

ßäe nijej amu-si-ße

already not.at.all walk-CPL-NEG:3ps/E
'Therefore, he also, he already could not walk any more at all.'

(42) mi hana kanaka ana safwato ajßa aj-akaanu, hana aka tihja
that when lpl/ex shoe with AUX-lpl/ex instead 3sg foot

nehesißa ne-Ί aj
pure be-PRT AUX:3ps/E
'Then, we did it [i.e. walked] with our shoes, but he instead did it barefoot.'

(43) mi hana=te ßäe ke kuaaeti-i akano jaaeto-a

that when=FOC already VLI be.afraid-PRT snake step.on-3ps/D

kwataa, ku kwaaeti-i, ßäe kau-a u-akaanu

so.that.not there be.afraid-PRT already return-NTR come-lpl/ex
'Then, already being afraid of that, so that he would not step on a snake, being
afraid we already went home.'

(44) atii ita ßäe amu-si-lanaala hwaaerei-kuru nesuru-t

nevertheless REC already walk-CPL-PRV somehow-PL relax-PRT

atii, kau-a u-akaanu

nevertheless return-NTR come-lpl/ex
'Nevertheless, (each of us) without being able to walk any more, we relaxed
somehow and [then] went home.'

(45) ßäe hatäi nuna-ana eno-ι mi hiiri kahe ßäe hatäi

already very jungle-inside enter-PRT that swamp from already very

ßäe d$a hoara-eni-akaanu hatäi rtukue-na ne-1 m-a=na

already what see-NEG-lpl/ex very be.night-NOM be-PRT be-3ps/D

häu beru hatQU ku kwara-eni-akaanu häu ßäe

because way hole there see-NEG-1 pl/ex because already
nesoone-to-akaanu
be.scared-INTS-1 pl/ex
'Already having entered very [deep] into the jungle, we did not see anything
from that swamp; as it was very dark and because we did not see a hole in the
way, we were very scared.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 915

(46) ne-m-a kanakaanu ne-naha loanari ne-ti-l e?eo-nenu

be-NEG-3ps/A lpl/ex be-LPP shed be-INTS-PRT shout-PRIR

katga=ne häu, ßäe hatäi nesoone-ohwa-akaanu ßäe,

man=SUB because already very be.scared-DSTR-lpl/ex already

d^arihjet kau-naa najße-m-akaanu=ne hau, ahaeto-a

which return-INFbe.able-NEG-lpl/ex=SUB because be.overgrown-3ps/D

beru kuane hau

way inside because
'As there was no one to shout [from] our place, the shed, we each were already
very scared, because we were not able to return in what ever way, because the
way was overgrown.'

(47) mi hau ajßa amuriti-ι hatäi enoto kwara-em-akat^e=ne hau,

that because with exceeding very sun see-NEG-lpl/in=SUB because

hatal nukue-kaatge, na-akaanu

very be.night-lpl/in say-lpl/ex
'Therefore, we said: "Even worse as we have not seen the sun, we have
encountered the night.'"

(48) ßäe läe ena kwaaune-Ί huajte-ri-a-akatQe=i, na-akaanu

already just now create-PRT repeat-IRR-NEG-lpl/in=ASS say-lpl/ex

ßäe, mtoani-akaanu häu, ena=ra kanakaana nesoone käu

already be.like.that-lpl/ex because now=EMF 1 pl/ex fear because.of
'"From now on we won't do [this] again", we said already because of our fear,
because we were in this situation.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
916 Appendix A : Texts

(49) mi hana atii päe heritQanehei päe kanaka am bihi

that when nevertheless already as.before already lpl/ex hand

hivue+hwueka-i, atii kanakaana ahinia kaanii enua sau-ή-ϊ,

RED+feel-PRT nevertheless lpl/ex before GPF tree cut-RAP-PRT

kanakaana ahlitQa jwaua-hei-l, akano sita-ϊ päe,

lpl/ex before beat-CNT-PRTsnake make.pass-PRT already

kanakaana ahinia ni-a=ne, kanakaana kala-a=ne hoataa

lpl/ex before be-3ps/D=CND lpl/ex bite-3ps/D=SUB so.that.not

akano sita-ι, atii kau-a u-akaanu, päe

snake make.pass-PRT nevertheless return-NTR come-lpl/ex already

turu-a u-akaanu kanakaanu ne-naha

arrive-NTR come-lpl/ex lpl/ex be-LPP
'Then, nevertheless, feeling [the trees] with our hands in the same way as before,
cutting the branches in front of us, beating [with our hands] in front of us,
making the snakes pass if they were in front of us, so that they would not bite us
- making the snakes pass, we went home; we already arrived at our place.'

(50) rauihirü kanakaana loanari helöo tQU hjä ini-t

really lpl/ex shed direction CRTN just go.up-PRT

u-akaanu hau atii päe ku turu-a u-akaanu

come-lpl/ex because nevertheless already there arrive-NTR come-lpl/ex
'As we really came going up towards our shed indeed, we arrived there
nevertheless.'

(51) mtoanei=pa päe turu-akaanu kanakaana ne-naha

like.that=FOC:lpl already arrive-lpl/ex lpl/ex be-LPP
'That's how we arrived at our place.'

(52) nii kohiüawo=na läe päe kau-a u-akaanu

that next.day=FOC:lpl just already return-NTR come-lpl/ex
'On the next day we went home.'

(53) nitoam-a
be.like.that-3ps/A
'It was like that.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 917

(54) jiäe sa-to-a

already end-INTS-3ps/A
'[The story] is over.'

Text 25 (Track 04: Traditional narrative),

by Medardo Arahuata Manizari, 2001

(1) aj-a hetau jioaelu kana+hoaaun-era, m-a hetau mi

AUX-3ps/A HRS earlier our.creator be-3ps/D because that

raj fluke mi-tea huaauna-t jie-1 ne-naa heelej katga

for PURP that-only create-PRT be-PRT be-NOM same man

ne-Ί mki
be-PRT ADVRS
'Earlier, God made [something ] equal to humans. [Lit. 'he made that there
would be [something ] equal to humans that he created']

(2) mi hetau=te akäaorja neko ne

that HRS=FOC TRN be:3ps/E
'Those were [called] Acaaonria Neko[meaning of name unknown]

(3) nii hau hetau=te mi-tguru raj ajte kana+kwaaun-era,

that because HRS=FOC that-PL for say:3ps/E our.creator

enene ka=raj temu-m-tQu-tge=ra, na-e hetau

today 1 sg=for plant-DSTL-IMP-PL=EMF say-3ps/E HRS

katia+kwaaun-era
our.creator
'Therefore, God said to them, "Today go planting [peanuts] for me!" (said God).'

(4) katoe temu-m-tQU-tge, jiaara beree-kuru ama-1, hitarii jiaara

peanuts plant-DSTL-IMP-PL 2pl child-PL take-PRT all 2pl

beree-kuru ama-i, ka=raj katoe temu-ni-QU-tge

child-PL take-PRT lsg=for peanuts plant-DSTL-IMP-PL
'"Go to plant peanuts, taking along your children, taking along all your children
- go to plant peanuts for me!'"

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
918 Appendix A: Texts

(5) ita kweena-l, ita mtQaahe-l, temu-m-tgu-tQe=ra, na-e

REC laugh-PRT REC joke-PRT plant-DSTL-IMP-PL=EMF say-3ps/E

hetau
HRS
"'Go to plant them, laughing and joking with each other!" he said.'

(6) nii hau hetau=te helaj raj n=obana nuke

that because HRS=FOC separate POSS 3ps=cl.peccary PURP

hetau kanii, obana heena-het hana heena-het kwajtel letoa-e

HRS GPF cl.peccary four-PRT inside four-PRT again send-3ps/E
'Therefore, he sent them, four to become his collared peccaries and again four
out of these collared peccaries [to become white-lipped peccaries].'

(7) βaara hana kaa temu-ni-tQU=toe=ra, na-1 hetau helaj

2pl instead this plant-DSTL-IMP-PL=EMF say-PRT HRS separate

kwajtel temu-erate
again plant-CAU2:3ps/E
'Saying, "Go to plant this in your group, he sent them to plant separately again.'

(8) nii hau hetau=te ku-ure, arahü le=lauri-l,

that because HRS=FOC go-3pl/E many one=group-PRT

(akäaorja neko) akäaorja neko na-1 kuraate-1

TRN TRN say-PRT name-PRT
'Therefore, they went, being one group, being called Acaaonria Neko.'

(9) mi häu hetau=te temu-a ku-ure

that because HRS=FOC plant-NTR go-3pl/E
'Therefore, they went to plant.'

(10) nii ajto-a häu rihihe-l hetau ku ita kweena-1,

that say-3ps/D because be.like-PRT HRS there REC laugh-PRT

ita nitQaahe-1 hetau ku nii katoe temu-ure

REC joke-PRT HRS there that peanuts plant-3pl/E
'As [God] had said, they planted peanuts, laughing and joking with each other.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 919

(11) katoe temu-uru-a hau hetau=te ku hetau (hjaram

peanuts plant-PL-3ps/D because HRS=FOC there HRS is ana.cane
baaune-t) hjarana baaune-Ί akauru tgoae hetau hwa-e
bewitch-PRT isana.cane bewitch-PRT 3pl above HRS throw-3ps/E

mi hirikuri, kam+kwaaun-era letono tie-1 ne-naa hirikuri

that PSN our.creator envoy be-PRT be-NOM PSN
'As they were planting peanuts, "that" Hirikuri, the envoy of God, Hirikuri,
bewitched the isana cane and launched it over them.'

(12) rtii hau hetau=te (aau) mi hetau mi hjaram

that because HRS=FOC (magic.item) that HRS that isana.cane

aau sit-a-a hau hetau, ku hetau (ku hetau)

(magic.item) pass-CAUl-3ps/D because HRS there HRS there HRS

oooooo na-ι hetau, ku hetau nesarakata-1

(start.grunting) say-PRT HRS there HRS start.moving-PRT

ka=enene aj-a=ne kujfla nuke, ku hetau

this=nowadays AUX-3ps/D=SUB so.that PURP there HRS

nesarakata-i, ku kami-tQuru-a
start.moving-PRT there sort.of-PL-3ps/A
'Therefore, as he made pass the isana cane's curse [over their heads], they did
that sort of thing, starting to move [like peccaries] so that it would be as it is
now, saying"ooo...".'

(13) nii hau hetau hajti katpa ne-1, hajti kat$a ne-1,
that because HRS still man be-PRT still man be-PRT

itQu itQU ÜQU itQU na-a läe ooooo na-t hetau=te

(frightful.sound) say-3ps/A just (start.grunting) say-PRT HRS=FOC

nii hana heena-hei ne-naa

that when four-PRT be-NOM
'Therefore, still being humans, those who were in the group of four made "it$u
. . . " a n d "ooo...".'

(14) ku nii katoe kujßaera ke ku temerihiati-1, he ah

there that peanuts digging.tool INST there lean-PRT hey oh

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
920 Appendix A; Texts

d$a=he=na na-a, wa wa na+na-a läe hetau,

what=REP=INT say-3ps/A (almost.grunt) RED+say-3ps/A just HRS

ku hetau hwa hwa na+na-1 hetau, amu-a akä-etonaj

there HRS (grunting.sound) RED+say-PRT HRS walk-3ps/A 3sg-also
'Leaning against the digging tool for peanuts, they said, "Hey! What?", - they
just started to grunt again and again and walked [away] as well.'

(15) mi hau hetau=te mi ku kuhjuta-l hetau, run kami

that because HRS=FOC that there conjure-PRT HRS there sort.of

risiße+ohwaa ke mriutu-m-u-toe=ra
cAamfe/ra.palm+thorn VLI turn.into-DSTL-IMP-PL=EMF
'Therefore, cursing them there, [the spirit said], "Go and turn into [this sort of]
chambira thorns!'"

(16) na-x hetau kuhjuta-ι itQa-a hau, ku hetau msarara-i,

say-PRT HRS conjure-PRT do-3ps/D because there HRS go-away-PRT

ku risijte+ohwa ke nerjutu-a ku-ure

there chambira.palm+thorn VLI turn.into-NTR go-3pl/E
'Saying so, as he did it cursing them, they went away and turned into chambira
thorns there.' [Intended meaning: their skin changed into thorns to resemble
peccary bristles]

(17) ke nuhuihja-a ku-ure, ßäe kauru uhuiße-1

INST get.bristles-NTR go-3pl/E already 3pl have.bristles-PRT

ku ne
there be:3ps/E
'They went to get bristles, and already having bristles, they were there.'

(18) mi hau hetau=te nii raana hana (ohzoa) ohwa

that because HRS=FOC that w.l.peccary inside thorn thorn

muurana ke nirjutu-hi-a ku-e

thorny.part.of.trunk VLI turn.into-CNT-NTR go-3ps/E
'Then, out of the group of white-lipped peccaries, [their skin] turned into the
thorny part of a trunk.'

(19) mi hau hetau=te nitoanel kani-hei raana

that because HRS=FOC like.that sort.of-PRT w.l.peccary
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 921

kwaauna-ι ne=lu kana+hvaaun-era

create-PRT be:3ps/E=REM our.creator
'Therefore, God created this sort of white-lipped peccary like that.'

(20) raana kwaauna-ι, obana hoaauna-i ne-lu mtoanei

w.l.peccary create-PRT cl.peccary create-PRT be:3ps/E=REM like.that

kana+kwaaun-era, nehelau helaj katga ne-1 ne-naa kahe

our.creator other.side separate man be-PRT be-NOM from

niki nete, aka ne-ι ne-ene, netQu hatal mi raj

ADVRS but 3sg be-PRT be-NEG:3ps/E absolutely that for

ßuke ne-naa kahe itg,a-e=lu kana+kwaaun-era

PURP be-NOM from do-3ps/E=REM our.creator
'God created white-lipped peccaries, and created collared peccaries like that
from other, separate people, but they were not identical [to us]; God made them
from [such] that were for that purpose only.'

(21) mi hau hetau=te nitoam-a hau, ku

that because HRS=FOC be.like.that-3ps/A because there

ni-a=ne wawaako-rehete
be-3ps/D=CND grunt-HAB 1:3ps/E
'Therefore, as it was like that, when [the peccaries] were there, they used to
grunt.'

(22) ku ioawaako-reheto-a hau, nii hau hetau=te ku mi

there grunt-HAB l-3ps/D because that because HRS=FOC there that

auetie su-ti-ι, ki-ure leotQa-uru, mi kana+kwaaun-era

secretly kill-INTS-PRT eat-3pl/E other-PL that our.creator

ajfla ne-lanaala ne-nakauru, helaj ne-nakauru

with be-PRV be-those.who separate be-those.who
'As they used to grunt there, therefore, the other secretly killed [some] and ate
them - those who live without God, those who live separately.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
922 Appendix A: Texts

(23) mi häu hetau=te wa kauatQa-i=ta kahia

that because HRS=FOC (disgust) good-NEG:3ps/A=FRS this.way

m-a=ne, ka=hitQihi he ni-a=ne, kzvatia

be-3ps/D=CND this=side VLI be-3ps/D=CND not

kauatQa-ri-ji=ta
good-IRR-NEG: 3 ps/A=FRS
'Therefore, [the envoy of God said], "It is not good if it is this way, if it is in this
manner; it would not be good!'"

(24) na-ι hetau, d^aha-i lejhii so-o-toe=te hoara-kaat<^e=ra

say-PRT HRS come.on-PRT one kill-IMP-PL=FOC see-lpl/in=EMF
'Saying so, [God] said, let's kill one and see!'"

(25) na-a häu hetau=te ku leijhn so-ore

say-3ps/D because HRS=FOC there one kill-3pl/E
'As he said so, they killed one there.'

(26) so-oru-a hau (hajti)hajti katga bihi hajti raj tihja-uru

kill-PL-3ps/D because still still man hand still POSS foot-PL

hajti mi neneofl-ene
still that be.transformed-NEG:3ps/E
'As they killed it, the human hands and the feet still had not been transformed.'

(27) nii häu hetau=te wa hajti-tQa, hajti

that because HRS=FOC (disgust) still-only still

laana-hee-he=ra
be.missing-DIM-CNT:3ps/E=EMF
'Therefore, [the envoy of God said], "hey, a little bit is still missing.'"

(28) na-1 hetau, na~i hekuri temure-u

say-PRT HRS say-PRT part.of.tree.root edge-LOC

luruari-m-u-toe=ra na-e hetau

lay.down-DSTL-IMP-PL=EMF say-3ps/E HRS
'Saying so, saying so, he said, "Go and lay them down [between] the tree upper
roots!'"

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 923

(29) mi häu hetau=te ene netohivel raana

that because HRS=FOC now until w.l.peccary

su-era-rn-a=ne, suru=a=ne (nijej) eru-naka-i,

ki!l-IMPS-NEG-3ps/D=CND run-3ps/D=CND not.at.allfmd-PSB-NEG:3ps/A

asi-a hitQutari hau

hide-3ps/D vulture because
'Therefore, [it is like that] until today, when it is impossible to kill white-lipped
peccaries, when they run away, when it is impossible to find them, [this is],
because the vulture hides them.'

(30) nitoani-a häu, mi hau hetau=te, mi baja ßäe

be.like.that-3ps/D because that because HRS=FOC that after already

kwajtei karahairii m-ori-i, hoajtei d^aha-ι läe

again not.long be-RAP-PRT again come.on-PRT just

so-o-tge=te Jauara-kaatQe=ra
kill-IMP-PL=FOC see-lpl/in=EMF
'As it is like that, therefore, after that, not long [after that, the envoy again said],
"Come on, let's kill [one] again!'"

(31) na-a häu, ßäe so-oru-a häu, ßäe läe

say-3ps/D because already kill-PL-3ps/D because already just

kauatQa najßehei nartißo-a kat^a bihi-uru ßäe,

pretty.much be.transformed-3ps/A man hand-PL already

tihja-uru
foot-PL
'As he said so, as they killed it already, the human hands and the legs had
already been transformed just in the right way.'

(32) mi häu hetau=te ßäe läe=ra

that because HRS=FOC already just=EMF
'Therefore, [he said], "Well done.'"

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
924 Appendix A : Texts

(33) d$aha-i saatono su-ι ke lenone-ku-toe=ra kanaanaj-uru,

come.on-PRT in.the.end kill-PRT VLI eat-IMP-PL=EMF child-PL

na-e hetau nii kana+kwaaun-era letono

say-3ps/E HRS that our.creator envoy
"'Come on children, for the last time let's kill [peccaries] and eat them!" said the
envoy of God.'

(34) mi hau hetau=te ßäe ku su-t, ki-ure

that because HRS=FOC already there kill-PRT eat-3pl/E
'Therefore, killing [some], [the people] ate them.'

(35) ku su-t, h-uru-a häu, hitarii hetau kauru raj

there kill-PRT eat-PL-3ps/D because all HRS 3pl for

edaatj-e
be.tame-3ps/E
'Killing them there, as they ate them, they were all tame for them.'

(36) ku hetau kuhjuta-i, emto+su+heloo hetau saru-e nii

there HRS conjure-PRT sun+dawn+towards HRS push-3ps/E that

kana+hoaaun-era letono
our.creator envoy
'[Then], conjuring there, the envoy of God drove them towards the West.'

(37) saru-1, ajti-i letoa-a

push-PRT say-PRT send-3ps/A
'Pushing them, he send them saying [that curse].'

(38) nii häu hetau=te ku nesararahi-t hetau ku heräe

that because HRS=FOC there go.away-PRT HRS there slowly

ku-t, tihjaa-ka, heräe nijej suru-elanaala,

go-PRT walk.s!owIy-3ps/A slowly not.at.all run-PRV

suru-naka-helanaala
run-PSB-PRV
'Therefore, they went away, walking slowly, they walked, slowly without
running at all, without being able to run.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A : Texts 925

(39) mi hau hetau=te so-ahi-a ku najßäi, mi ajto-a

that because HRS=FOC kill-CNT-3ps/A there nicely that say-3ps/D

hau rihihei, nihjauria lae kunajta-1, hau-1 itQa-ki-tQe=ra

because like don't just hurt-PRT throw-PRT do-2ps-PL=EMF
'Therefore, they continued to kill nicely, as [God] had told: "Don't do it hurting
[the herd] throwing them [all] away!'"

(40) na-1, ajto-a hau rihihei hetau ku kulaalahii

say-PRT say-3ps/D because like HRS there carefully

(so-ahe-t, najßäi) so-ore

(kill-CNT-PRT) nicely kill-3ps/E
'Saying so, according to the way he spoke, they killed them carefully

(41) mi hau hetau=te ßäe nai kana tQoae

that because HRS=FOC already over.there lpl/in above

ku-e=lu ßäe kaa raana

go-3ps/E=REM already this w.l.peccary
'Therefore, these [remaining] white-lipped peccaries already went up there, to
the sky.' [Lit. 'above us']

(42) mi hau hetau=te (mi) mtoanei te-uru-a mi

that because HRS=FOC that like.that give-PL-3ps/D that

kana+hoaaun-era ajßa m-ji-tguru hau, amuurani-a

our.creator with be-NEG:3ps/D-PL because continue-3ps/A

(hetau=te) (itulere ne-1) ena netohwei leotga-uru ne-nakauru

HRS=FOC all.kinds be-PRT now until other-PL be-those.who

kana hau-ι raana so-1 ria natu ki-ure

lpl/in throw-PRT w.l.peccary kill-PRT even eat-3pl/E
'Therefore, because those who do not live with God put it that way, those other
[people] continue like that, taking advantage of us until now, even killing white-
lipped peccaries and throwing them away, they eat them.'

(43) mtoanei hetau kana raj raana hoaauna-i ne=lu

like.that HRS lpl/in for w.l.peccary create-PRT be:3ps/E=REM

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
926 Appendix A: Texts

kana+kwaaun-era
our. creator
This is [how] God created the white-lipped peccaries for us.'

(44) ßäe
already
'[The story is finished] already.'

Text 86B (Track 05: Instruction),

by Anita Macusi Nuribe, 2004

(1) naahei barue ha-akaamt

like.this masato make-lpl/ex
'This is how we make masato.'

(2) haurianehei (barue) laano ra-a ku-akaanu

first masato cassava receive-NTR go-lpl/ex
'First we go to get cassava.'

(3) haurianehei ßäe turu-ι, ukwana hana turu-1,

first already arrive-PRT field inside arrive-PRT

n=asahi kweeta-kaanu
3ps=under clear-lpl/ex
'First, arriving already, arriving at the field, we clear the underground.'

(4) n=asahi kweeta-ι, ßäe najßäi kweeta-kaanu=ne ßäe, nuhua

3ps=under clear-PRT already nicely clear-lpl/ex=CND already stem

saauhi-ι nai ahaena-ana hau-akaanu, ukwana laaruma

cut.off-PRT over.there forest-inside throw-lpl/ex field edge
'Clearing the underground, finishing already, when we have cleared it, we cut off
the trunks and throw them away into the woods at the edges of the field.'

(5) ukwana laarunia nuhua hau-ι, mi baja saauhi-akaanu=ne

field edge stem throw-PRT that after cut.off-lpl/ex

baja ruku-akaanu
after pull.out-lpl/ex
'Throwing the trunks away to the edges of the field, after that, after we have cut
[the trunks] off, we pull them out.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
 Appendix A: Texts 927

(6) ruku-J, ßäe kodorja-ΐ, saauhi-ι ßäe mi baja

pull.out-PRT already heap.up-PRT cut.off-PRT already that after

ruee-kaa
peel-1 pl/du
'Pulling them out, heaping them up, cutting them off - after that we already peel
them.'

(7) rue-hei, saku kuatie sarara-t u-a-ϊ, nii baja

peel-PRT sack inside put.into-PRT come-CAUl-PRT that after

tur-a-Ί, ßäe kule-het, isi-ana ßäe laurja-kaa

arrive-CAUl-PRT already wash-PRT fire-inside already make.sit-1 pl/du
'Peeling them, putting them into a sack, bringing them [home], after that making
them arrive and already washing them, we make them sit on the fire [in a pot].'

(8) ßäe enoa-a hana, ßäe tutu-aka ruria kuatie

already cook-3ps/D when already mash-1 pl/du tub inside
'When they are cooked, we already mash them in a [special mashing] tub.'

(9) tutu-i, ßäe tgu najßäi tutu-aka=ne baja, (akii)

mash-PRT already CRTN nicely mash-lpl/du=SUB after camote.fruit

akii ruhua-aka kimtgu ke

camote. fruit grate-1 pl/du grater INST
'Mashing them, after we have mashed them nicely, we grate camote fruit with a
grater.'

(10) kirntQU ke akii ruhua-~i, ßäe tonono-Ί ßäe

grater INST camote.fruit grate-PRT already squeeze.out-PRT already

kuane te-ι, ßäe te-ϊ, ßäe (kuane) kulia

inside give-PRT already give-PRT already inside mass

d^urH+d^uru-ri-l ßäe kuane nal-a-aka

RED+mix.in-RAP-PRT already inside fall-CAUl-lpl/du
'We grate camote fruit with a grater and squeeze out [the juice], put [the juice]
inside [the cassava mass], putting it inside, we mix [the juice] into the mass
again and again and drop [the juice] inside the mass.'

(Π) kuane d$uru-l tutu-si-tgaanu baja, tutu-si-tgaa

inside pour.out-PRT mash-CPL-lpl/ex after mash-CPL-1 pl/du
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
928 Appendix A: Texts

netuaka-a=ne kujßa
become.soft-3ps/D=SUB so.that
'After we mashed it completely, pouring [the camote fruit] out, we mash it
completely so that [the mass] become soft.'

(12) ßäe netuaka-a=ne ßäe, barude kuane ti-aka

already become.soft-3ps/D=CND already bucket inside give-lpl/du
'When it has become soft, we put it into a bucket.'

(13) barude kuane ti-aka=ne baja, nii rihe-t

bucket inside give-lpl/du=SUB after that be.like-PRT

ßäe kohzoanoo ahaha-Ί ßäe ku-aka

already next.day heat-PRT already drink-lpl/du
'After we have put [the mass] into a bucket, like that, on the next day, we heat it
and drink already [in order to try].'

(14) mi baja heena hano-1 mi rihe-t laueka-aka

that after four day-PRT that be.like-PRT make.sit-lpl/du

(kwaaere) kwaaera-a kujßa

(ferment) ferment-3ps/D so.that
'After that, we make it "sit" like that for four days, so that it would ferment.'

(15) kwaaera-a hana, ke katQa kuhva-aka

ferment-3ps/D when VLI man give.to.drink-lpl/du
'When it has fermented, we give it to drink to the people.'

(16) kat^a kuhva-Ί, nii kana rüa ne-nakauru kukwa-aka

man give.to.drink-PRT that lpl/in side be-those.who give.to.drink-lpl/du
'Giving the people to drink, we give to drink those on our side.'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:35 PM
Appendix Β

List of scientific names for plant and animal terms (in alphabetical order)

The determination of Latin and names is largely based on David Fleck's

wordlist of Matses (ms., RCLT 2004). Other terms, including English glosses
are based on various internet sources. While I have aimed at providing
definitions thoroughly, some terms may be controversial, as many of the plant
and animal species of the Amazon rainforest are little explored.

Urarina Local Spanish Scientific name English

anew sacha culantro Eryngium foetidum L. (coriander-like herb)
(Fam. Apiaceae)
aari topa Ochroma pyramidale 'balsa wood tree'
(Fam. Bombacaceae)
aari arambaza (unknown) (type of bee)
aal otorongo Felis yagouaroundi 'jaguar'
aerana pucalupuna Ceiba pentrada (type) 'lighthouse tree' (type)
ahaipe paiche Arapaima gigas 'giant arapaima' (fish)
(Fam. Arapaimidae)
ahariri gamitana Colossoma macropomum 'tambaqui' (fish)
(Fam. Characidae)
ahjaui uvilla Pourouma cecropiifolia (type of tree)
(Fam. Moraceae)
ahjaaone santa maria Piper peltatum (type of pepper plant)
(Fam. Piperaceae)
ajt&ihiri armadillo grande Priodontes maximus 'giant armadillo'
aknaraa pona Dictyocaryum ptarianum 'white nist palm'
(Fam. Arecaceae)
akaa toe Brugmansia sp. 'angel trumpet vine',
(Fam. Solanaceae) 'datura'
akii camote Ipomoea sp. 'morning glory' (vine)
(Fam. Convolvulaceae)
alaa aguaje Mauritia flexuosa L. f. 'swamp palm'
(Fam. Palmae)
alaaeri caimitu Pouteria caimito 'yellow star apple'
(Fam. Sapotaceae) (tree)
alajtgöo jergon Bothrops atrox 'South American
lancehead snake'
alau maquisapa Ateles paniscus 'spider monkey'
alauihja pero caspi (unknown) (type of tree)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
930 Appendix Β: List of scientific names

Urarina Local Spanish Scientific name English

alöori panguana Cypturellus undulatus 'tinamou' (partridge)
anaahe uvos Spondias mombin 'hog plum', 'yellow
(Fam. Anacardiaceae) mombin' (tree)
anajsihje mucura Petivera Alliacea 'anamu' (herb)
(Fam. Phytolaccaceae)
anane pichico Saguinus spp. 'tamarin' (monkey)
arauata choro Lagothrix lagothricha 'woolly monkey'
aresiße mojara Acestrocephalus boehlkei type of 'tetra' (fish)
(Fam. Characidae)
atari paufil Crax mitu 'razor-billed curassow'
(bird)
auri paucar Psarocolius sp. 'oriole' type (bird)
(Fam. Icteridae)
baka isitQii teta de vaca Solanum mammosuni (type of plant)
d^aruba vaca marina Trichechus inunguis 'Amazonian manatee'
darane mojara (related to Acestrocephalus (type of fish)
boehlkei
(Fam. Characidae))
duhxua cedro masha Cabralea sp. 'cancharana' (tree)
(Fam. Meliaceae)
ejtQii cashapona Socratea exorrhiza (Mart.) 'stilt palm'
(Fam. Palmae)
elelia, elele chevon (unknown) (type of tree)
enuasoone nobia Ageneiosus atronasus (type of catfish)
(Fam. Auchenipteridae
esu shimbillo; guaba Inga edulis 'ice-cream bean' (tree)
(Fam. Fabaceae)
etoe musmuqui Aotus nigriceps 'owl monkey'
fioafioafioa katu sapucho; sapo Musa sp. (type of banana)
platano
fioaftoafioa huapapa bird Cochlearius cochlearius 'boat-billed heron'
fioanara lanahaj guineo Musa sp. (type of banana)
haja mullaca caspi (mullaca: Physalis angulata) (type of tree)
hiflori (unknown) (unknown) (type of tree)
hiririßo bagre Pinirampus pirinampu (type of catfish)
(Fam. Pimelodidae)
hjaane achiote Bixa orellana L. 'annato' (tree, herb)
(Fam. Bixaceae)
hjarana isana, carta brava Gynerium sagittatum 'arrow cane'
(Fam. Graminae)
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
 Appendix Β: List of scientific names 931

Urarina Local Spanish Scientific name English

itgai sacha papa Dioscorea trifida (type of tuber)
(Fam. Dioscoreaceae)
kahjausi cortadera Scleria (Fam. Cyperaceae) (type of herb)
kajahuri cumala (Fam. Myristicaceae - any (type of tree)
type)
kakuri trompetero Psophia crepitans 'grey-winged
trumpeter' (bird)
kameranati guineo pindorito (unknown) (type of banana)
kati mono negro Cebus apella 'dark capuchin
monkey'
Mrimata boquichico Prochilodus 'tilapia' (fish)
(Fam. Prochilodontidae)
komokomo comocomo, garza Casmerodius alba 'great egret'
blanca
hihiri carachupa Cabassous unicinctus 'Southern naked-tailed
armadillo'
kiirari palometa; Mylossoma/Metynnis/Myleus 'silver dollar' (fish)
macane spp. (Fam. Characidae)
hiri jagua, sacha Tocoyena williamsii Standley (type of tree)
huito (Fam. Rubiaceae)
hoairi ayahuasca Banisteriopsis caapi (type of liana)
(Fam. Malpighiaceae)
laaw yuca Manihot esculenta Crantz 'cassava'; 'manioc'
(Fam. Euphorbiaceae)
lerano macana Adontostemarchus balaenops (type of knifefish)
meseri cocona Solanum sessiliflorum 'peach tomato' (bush)
(Fam. Solanaceae)
nehuuri puma garza Botaurus pinnatus, 'rufescent tiger-heron'
Tigrisoma lineatum
nekivuuri garabata (similar to Uncaria spp. (type of liana)
(Fam. Rubiaceae))
obana sajino Pecari tajacu 'collared peccaiy'
ohxoa huicungo Astrocaryum murumuru Mart. (type of tree)
(Fam. Palmae)
raana huangana Tayassu pecari 'white-lipped peccary'
rihje pijuayo, pifayo Bactris gasipaes 'peach palm'
(Fam. Palmae)
risijie chambira Astrocaryum chambira 'chambira palm'
(Fam. Arecaceae)
ruru coto Alouatta seniculus 'red howler monkey'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
932 Appendix Β: List of scientific names

Urarina Local Spanish Scientific name English

siria dorado Brachyplatystoma (type of catfish)
filamentosum
(Fam. Pimelodidae)
suseri unchala Aramides cajanea 'grey-necked wood-rail'
(bird)
tariatga taricaya Podocnemis unifilis 'yellow spotted river
turtle'
tiirnri llanchama Poulsenia armata (type of fig tree)
(Fam. Moraceae)
unoari huitina Xanthosoma sagittifolium 'arrowleaf (tuber)
(Fam. Araceae)
fabeeto guayaba Psidium guajava L. (type of tree)
(Fam. Myrtaceae)
feremia pandisho, pan de Artocarpus altilis (Park.) Fosb. 'breadfruit tree'
ärbol (Fam. Moraceae)
ukuari oso hormiguero Myrmecophaga tridactyla 'giant anteater'
ufoasi renaco Ficus spp., Coussapoa spp. 'fig tree'
(Fam. Moraceae)
unee kinkajou Bassaricyon gabbii 'kinkajou'
urerej otorongo Panthera onca 'panther'

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
 Appendix C - Plates

Plate 1: Urarina houses

Plate 2: Medardo Arahuata Manizari

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
934 Appendix C: Plates

Plate 3: Urarina group including

Vicente Arahuata Manizari (back
left) and Julian Nuribe Vela (back
right)

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
References

Abbott, Miriam
1991 Macushi. In: Handbook of Amazonian Languages, Volume 3,
Desmond C. Derbyshire and Geoffrey K. Pullum (eds.): 23-160.
Berlin: Mouton de Gruyter.
Aikhenvald, Alexandra Y.
1995 Person-marking and discourse in North-Arawak languages. Studia
Linguistica 49: 152-195.
2003 Imperatives and other commands: a cross-linguistic typology. Ms.,
Research Centre for Linguistic Typology.
2005 Serial verb construction in typological perspective. In: Serial verb
constructions: a cross-linguistic typology, Alexandra Y. Aikhenvald
and R.M.W. Dixon (eds.): Oxford: Oxford University Press. In press.
Cajas Rojas, Judith
1989 Nasalizaciön en urarina. In: Temas de linguistica amerinha, R. and
G.S.F. Palomino, Lima: CONCYTEC.
Cajas Rojas, Judith, A. Corbera, B. Gualdieri, and G. Solis
1987 Bibliografia etnolingiiistica urarina. Lima: Instituto de Lingilistica
Aplicada (CILA), Universidad Nacional Mayor de San Marcos.
Carlin, Eithne
2004 A Grammar Of Trio: A Cariban Language Of Suriname. Duisburger
Arbeiten Zur Sprach- Und Kulturwissenschaft, Bd. 55. Duisburg:
Peter Lang Publishing Company, Volume 1.
Chappell, Hillary and William McGregor (eds.)
1996 The Grammar of Inalienability. A typological perspective on body
part terms and the part-whole relation. Berlin: Mouton de Gruyter.
Dean, Bartholomew
1994a The Poetics of Creation: Urarina Cosmogony and Historical
Consciousness. Latin American Indian Literatures Journal. Review of
American Indian Texts & Studies Spring 1994, Vol. 10(1): 22-45.
1994b: Multiple Regimes of Value: Unequal Exchange and the Circulation of
Urarina Palm-Fiber Wealth. Museum Anthropology. Journal of the
Council for Museum Anthropology 18(1): 3-20.
1995 Forbidden fruit: infidelity, affinity and brideservice among the
Urarina of Peruvian Amazonia. The Journal of the Royal
Anthropological Institute 1(1): 87-110
1999 Language, Culture & Power: Intercultural Bilingual Education among
the Urarina of Peruvian Amazonia. Practicing Anthropology Vol.
20(2) (Special Issue: Reversing Language Shift in Indigenous
America.), L. Watahomigie, T. McCarty & A. Yamamoto (eds.): 39-
43.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
936 References

Derbyshire, Desmond C.
1979 Hixkaryana. Lingua Descriptive Studies 1. Amsterdam: North-
Holland.
1985 Hixkaryana and linguistic typology. Summer Institute of Linguistics
Publications in Linguistics, 76. Dallas: Summer Institute of
Linguistics and the University of Texas at Arlington.
Derbyshire, Desmond C., and Geoffrey K. Pullum (eds.)
1998 Handbook of Amazonian Languages, Amsterdam: Walter de Gruyter.
Dixon, R.M.W.
2000 A typology of causatives: form, syntax, and meaning. In: Changing
valency: case studies in transitivity, R.M.W. Dixon, & Alexandra Y.
Aikhenvald: 30-83. Cambridge: Cambridge University Press.
Dixon, R.M.W.
2002 Word: a typological framework. In: Word. A cross-linguistic typology,
R.M.W. Dixon and Alexandra Y. Aikhenvald (eds.): 1-41.
Cambridge: Cambridge University Press.
2004 Adjective classes in typological perspective. In: Adjectives classes. A
cross-linguistic typology, R.M.W. Dixon and Alexandra Y.
Aikhenvald (eds.): l-49.0xford University Press.
2004 Comparative constructions. Ms., Research Centre for Linguistic
Typology, La Trobe University.
Dryer, M.
2002 Order of subject, object, and verb. Ms., to appear in: World atlas of
language structures. Matthew Dryer, Martin Haspelmath, David Gild
& Bernard Comrie (eds). To be published by Oxford University Press.
Figueroa, Francisco de, Cristobal de Acufia, et al.
1986 Informes de Jesuitas en el Amazonas 1660-1684. Monumenta
Amazönica. Iquitos: IIAP-CETA.
Fleck, David
2004 Matses-English analytical and etymological dictionary. In
preparation.
Frajzyngier, Zygmunt
2004 Tense and aspect as coding means. In: Journal of West African
Languages, XXX, 2: 53-67. Special issue, edited by Bernard Comrie
and Ekkehard Wolff.
Genetti, Carol
2005 The participial construction of Dolakha Newar: Syntactic implications
of an Asian converb. Studies in Language 29:1: 35-87.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
 References 937

Gildea, Spike
2000 On the genesis of the Verb Phrase in Cariban languages: diversity
through reanalysis. In: Reconstructing grammar: comparative
linguistics and grammaticalization, Spike Gildea (ed.): 65-106.
Amsterdam/Philadelphia: John Benjamins.
Givön, T.
2001 Syntax: An Introduction. Amsterdam/Philadelphia: John Benjamins.
Greenberg, Joseph H.
1960 The General Classification of Central and South American
Languages. In: Men and Cultures, A.F.C. Wallace (ed.). Selected
Papers of the 5th International Congress of Anthropological and
Ethnological Sciences. Philadelphia: University of Philadelphia Press.
1963 Some universals of grammar with particular reference to the order of
meaningful elements. In: Universals of language, Joseph H.
Greenberg (ed.). 2 ed.: 73-113. Cambridge, MA: MIT Press.
Gualdieri, Beatriz C., and Judith .1. Cajas Rojas
1987 Kaöa eje (lengua urarina). Aspectos de la fonologia. Tesis de
Licenciado en Lingüistica. Lima: Universidad Nacional Mayor de San
Marcos.
Gundel, Jeanette, and Thorsten Fretheim
2005 Topic and focus. To appear in: The handbook of pragmatic theory.
Laurence Horn and Gregory Ward (eds). In Press. Blackwell
Haspelmath, Martin
1997 Indefinite pronouns. (Oxford Studies in Typology and Linguistic
Theory) Oxford: Oxford University Press.
Haspelmath, Martin and Ekkehard König (eds.)
1995 Converbs in Cross-Linguistic Perspective. Structure and meaning of
adverbial verb forms - adverbial participles, gerunds. Berlin: Mouton
de Gruyter
Keenan, Edward L.
1978 The syntax of subject-final languages. In: Syntactic Typology: Studies
in the Phenomenology of Language, Winfred P. Lehmann (ed.): 267-
328. Austin: University of Texas Press.
Lambrecht, Knud
1994 Information Structure. Cambrige: Cambridge University Press.
Loukotka, Cestmin
1968 Classification of South American Indian Languages. Reference Series
Vol. 7, Johannes Wilbert (ed.). Los Angeles: University of California.
Manus, Phyllis
1992 Proposiciones de desarrollo y de apoyo y su expresion en las
oraciones del urarina. Revista Latinoamericana de Estudios
Etnolingüisticosl'. 15-45.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
938 References

Matthews, Peter H.
1997 Dictionary of Linguistics. Oxford/New York: Oxford University
Press.
Ministerio de Educaciön, Educaciön Bilingüe de la Selva, con la colaboraciön del
Instituto Lingüistico de Verano
1981 Icha. majas (y otros cuentos de animales). Coleccion literaria
Oranna. Libro No. 1. Yarinacocha, Pucallpa: Centro Amazönico de
Lenguas Autöctonas Peruanas "Hugo Pesce".
Montag, Susan
1981 Diccionario Cashinahua. Serie linguistica Peruana No. 9. Mary Ruth
Wise (ed.) Yarinacocha: Ministerio de Educaciön, Instituto
Lingüistico de Verano.
Olawsky, Knut J.
2002 örarina texts. Languages of the World/Text Collections. Munich
2002: LINCOM.
2003 Urarina word classes. In: Proceedings from the fifth Workshop on
American Indigenous Languages, Santa Barbara Papers in Linguistics,
vol. 13, Jeanie Castillo (ed.). Santa Barbara: University of California:
38-52.
2005 Urarina - Evidence for OVS Constituent Order. In: Leiden Papers in
Linguistics 1.2, Boban Arsenijevic, Noureddine Elouazizi, Martin
Salzmann & Mark de Vos (eds.): 97-123.
Olawsky, Knut J., and Medardo Arahuata Manizari
2004 Orarinaauru Nereretaao (1). Cuentos Urarinas (1). Iquitos: Su
Camino Internacional, Red Educativa de la Etnia Urarina.
Payne, Doris L.
1990 The pragmatics of word order: Typological dimensions of verb initial
languages. Empirical approaches to language typology 7. Berlin,
New York: Mouton de Gruyter.
Prince, Alan, and Paul Smolensky
1993 Optimality Theory: Constraint Interaction in Generative Grammar.
RuCCS-TR-2. ROA-537.
Shell, Olive Α., and Mary Ruth Wise
1971 Grupos idiomäticos del Peril. Lima: Universidad Nacional Mayor de
San Marcos, Instituto Lingüistico de Verano.
Tessmann, Günter
1930 Die Indianer Nordost-Perus. Hamburg: Friederichsen, de Gruyter and
Co.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
 References 939

Tomlin, R.
1986 Basic word order: Functional principles. London: Croom Helm.
Velasco, Juan de
1960 Historia del reino de Quito en la America Meridional. Segunda
Parte: La colonia y la republica, 1842-44. 2 vols. Puebla: Editorial
Jungle.
Whitehead, Carl R.
1987 The category 'relevance' in Menya verbal morphology. Language and
Linguistics in Melanesia 16(1-2): 41-53.

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
Index
Adjectives, 134, 139, 187, 203, 308, Consecutive, 583, 704
309 Construction with 'must', 241, 667,
Adjuncts, 670, 715 773
Adverbs, 145, 169, 498, 505, 555, 644, Content questions, 438, 698, 814
680, 798, 860 Continuous aspect, 96, 173,276, 277,
Adversative, 645, 862 462,464, 570, 631,755,790
Affect verbs, 384,410 Contrastive emphasiser, 287
Affricatisation, 71 Coordination, 256, 802, 898
A-form, 258, 441, 488, 565, 807, 814, Copula, 89,391,656, 659
842 Copular auxiliary, 73, 277, 397, 516,
Agentive, 112, 162, 322, 359, 429, 891 596
Airico dialect, 38, 115, 867, 882 Corrientes dialect, 2, 841, 867
Ambitransitivity, 166, 385 Demonstratives, 216, 299,401,430,
Apprehensive, 265 850
Assertive, 482,494, 586, 734 Denasalisation, 117
Assimilation, 33, 63, 66, 101, 107, 115, Dependent clause, 434, 536, 643, 658,
183,883 682, 689, 709, 730
Associative ku-, 37, 347, 426, 890 Derivation, 143, 343, 380,405,412,
Associative modifier, 345 457
Associative modifier b-, 165, 887 Desperate interrogative, 491
Attitude, 508, 888 D-form, 247, 434,453, 774, 786
Auxiliary, 399, 560, 565, 792 Diminutive, 237, 412,474, 543, 571
Beneficiary, 239, 333, 386, 674, 675, Diphthongisation, 113, 118, 874
818 Diphthongs, 60, 93, 99, 107, 112
Causative, 112, 459,460, 609 Distal form, 102, 132, 149,264, 574,
Certainty, 284, 724 637,704
Chambira dialect, 21, 32, 38,47, 69, Distributive, 198, 362,409, 418, 468,
165, 223,461,867 486
Citation form, 79, 152, 291, 348, 441 Ditransitivity, 168, 386
Classifying nouns, 890 Double causative, 619
Clause chaining, 747, 847 Dual, 213, 364, 409,489, 574, 864
Clause introducers, 259, 558, 574, 805 E-form, 258, 435, 481, 693, 786, 807,
Clitics, 63, 69, 74, 80, 84, 107, 182, 814,842
284, 350, 493, 596, 734 Ellipsis, 403, 736, 808, 844
Colour terms, 188, 204, 309, 609, 792 Emphasiser 'just1, 286
Comitative, 625, 674 Emphatic Respect marker, 858
Comparison, 207, 750 Enumerative verb, 293, 760, 809
Complement clause, 209, 434, 635, Epistemics, 179,480
758, 780,812, 831,893,899 Fear marker, 81, 513, 735
Completive aspect, 477 Final emphasis, 441, 508
Compounds, 81, 292, 347, 425 First person forms, 99, 130, 490
Conditional, 73, 256, 266, 585, 735, Focus, 146, 436, 438, 645, 661, 692,
737, 832 815
Conjunction phrase, 436, 715 Frustrative, 81, 266, 509, 539, 735,
Conjunctions, 245, 562, 721, 743, 855 866,884
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
942 Index

Future tense, 481, 585 Monophthongisation, 71, 875

Gap fillers, 172, 653, 679, 860 Naming, 14, 393
Gemination, 841, 876 Narratives, 439, 840, 864
Glottal stop, 50, 59, 72,485, 840 Nasal spreading, 74, 82,476
Grammatical word, 81,647 Nasalisation, 21, 63
Grammatical Word, 77 Negation, 263,484, 505, 552, 662,
Greater plural, 409 663,710
Habitual aspect, 105, 248, 360,465, Negative questions, 266, 565, 834
467, 736 Neutral form, 79, 148, 262, 432, 441,
Hearsay evidential, 504, 862 518, 520, 629,723,779, 802
Hortative, 264, 573, 704, 728 New Testament, 17, 536, 579
Iconic repetition, 845 Nominaliser-i, 162, 185, 190,321,637
Ideophones, 283, 841 Nominaliser -naa, 161, 323, 626, 759
Idioms, 24, 227,297, 399, 429, 562, Non-finite verb plus conjunction, 754
599, 669, 728, 884, 894 Non-singular, 355
Imperative, 102, 131, 513, 516, 569, Nouns, 79, 92, 121, 123, 152, 193, 246,
583, 637, 703, 727, 757 298
Impersonal passive, 465 Number agreement, 365, 370
Impossibility, 461 Numerals, 275, 300, 367, 850
Inalienability, 350 Omission of arguments, 376, 613,656
Indefinite emphasiser, 830 Order variation, 645, 660, 668, 899
Indefinite pronouns, 827 Palatalisation, 72, 89, 107, 114, 115,
Indirect questions, 836 183,596
Infinitive, 141,520, 654, 758 Pangayacu dialect, 881, 882
In-law talk, 478, 540, 636 Paradigms, 433,483,484,485, 489,
Innovative language, 30, 34,44, 337, 495,496
353, 365, 509, 753,785, 826, 831, Participle, 291, 373,440, 520, 555,
885 637, 648, 722, 788, 848
Instrumental, 625, 627 Particles, 186, 284, 491, 645, 705, 715,
Intensity, 379, 416 724,840
Interjections, 282, 559 Passive, 516, 592, 633
Interrogative, 73, 87, 258, 336, 401, Paucal, 409
430, 505, 730, 888 Person inflection types, 431, 487
Interrogative pronouns, 556 Phonological simplification, 70, 71,
Intransitive verbs, 142, 167 117, 118
Intransitiviser, 114, 116, 601 Phonological word, 42, 88, 148
Irrealis, 482, 738 Phonological Word, 77, 89, 139
Irrealis nominaliser, 776 Phonotactics, 50
Jussive, 576, 637, 888 Pity marker, 290,491
Juxtaposition, 333, 809 Plural, 91, 97, 103, 113, 152, 294, 295,
Kinship, 353, 530 355, 378, 379, 479, 492, 572, 878
Loans, 275, 633, 890, 894 Plural object, 360, 468, 632, 793
Location, 154, 176, 219, 225, 394, 624, Polar questions, 442, 831
673, 675, 681 Politeness, 357,472,493, 530, 534,
Medium degree, 143, 189,414 571,888
Modifiers, 161, 206,277,298, 315, Possession, 137, 145, 215, 301, 331,
327, 369, 595 397, 887

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
 Index 943

Postpositions, 136,224,244, 294, 670, Suffix 'while', 518, 802

818 Suffix order variation, 523
Postvocalic aspiration, 41, 868 Suffix position classes, 456
Prefixes, 459 Suggestive, 267, 587
Privative, 112, 519, 564, 757, 802 Superlative, 210
Probability, 480, 636 Syllable structure, 60, 75, 94, 110, 120,
Proclitics, 83, 114, 115, 116, 139, 153, 123, 125, 128, 139
214,337,351,886 Third person forms, 98
Prohibitive, 260, 262, 579 Tigrillo dialect, 3, 157, 180, 560, 867
Pronouns, 213, 589 Tone, 21, 87, 120
PSC verbs, 379,407, 553, 791 Tone classes, 124, 149
Purposive, 253,404, 558, 639, 716, Topic, 692, 697
752,779,820 Traditional language, 43, 107, 115,
Quantifiers, 267,275, 300, 303, 368 153, 174,214, 253, 264,290, 337,
Questions, 705, 786, 814 341,481,493, 503, 560, 820, 885
Quotations, 729, 779, 784, 808, 857 Transitional glides, 64, 877
Reassurance, 503 Transitivity, 141, 166, 167,376, 604,
Reciprocal, 211, 607 654
Reduplication, 418, 632, 842 Valency increase, 377, 614, 622, 671
Reference tracking, 849 Velocity, 415,471, 546, 571,632
Reflexivity, 168, 385,603,606 Verbalisation, 291, 343,406, 554, 625
Relativisation, 320, 371 Verbless clauses, 401, 833
Remoteness, 174, 465,481, 498, 596, Verbs, 79, 95, 127, 166, 188, 341,375,
734, 742,752 553
Reportative evidential, 498 Vowel copying, 36,44, 61, 67, 463,
Resultative, 640 873
Rhetorical questions, 508, 514, 835 Vowel length, 41, 55, 93, 98, 121, 123,
Second person forms, 104 128, 137, 143, 345,463,476,611,
Sentence length, 687, 846 789
Serial verbs, 148, 555,629,659, 723 Vowel lengthening, 109, 842, 873
Special verbs, 259, 403, 792, 837 Vowel merging, 110, 113
Stative verbs, 142, 167, 191, 192, 362, Vowel raising, 70
378,659, 791 Vowel sequences, 58, 98, 106
Stem extension, 106 Warning marker, 85, 512, 735
Subordinate clause, 141,246,434,482, Witness evidential, 496, 586, 735
497, 690, 729, 733,734, 831, 847 Word classes, 151,291,296
Suffix 'also', 182, 664 Word length, 77, 91, 94, 125, 138, 152,
Suffix 'only', 184,517 473,644

Brought to you by | Max-Planck-Gesellschaft - WIB6417

Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM
Brought to you by | Max-Planck-Gesellschaft - WIB6417
Authenticated | 194.94.96.194
Download Date | 9/19/13 5:36 PM