Author Manuscript (Post-peer review)

Chen, Yiting (2023) Baseline and elaboration in word formation.

In F. T. Li (ed.), Handbook of Cognitive Semantics, 54-94. Leiden: Brill.

CHAPTER 28
Baseline and Elaboration in Word Formation
Yiting Chen

Abstract: This chapter explores the motivations behind various linguistic properties in
complex words, based on the “quasi-spatiotemporal relationship” in
baseline/elaboration (B/E) organization (Langacker, 2016a). According to this theory,
word formation can be viewed as an elaboration that operates on a baseline. Through
the “structured mapping in word formation”, the B/E organization can be used to
capture the dynamic process of word formation and provide a basis for analyzing the
phases before, during, and after elaboration from both spatial and temporal perspectives.
Thus, it is possible to reconsider the logical dependence of concepts as spatial
constraints and the unidirectional nature of B/E organization (or “logical necessity”) in
terms of temporal sequencing. Examining word formation phenomena such as
headedness, the order of word formation, productivity, and the dependency and manual
nature of adaptation, this chapter aims to show the utility of B/E organization in
clarifying the iconic nature of word formation.
Keywords: motivation – iconicity – quasi-spatiotemporal relationship –
structured mapping in word formation – productivity – compounding and derivation

1. Introduction

Creativity is one of the major human traits. Human cognitive ability goes beyond the
capacity to use existing ideas and concepts; it also allows us to create new concepts
based on existing ones. Adopting Langacker’s (2016a) theory of baseline and
elaboration, the creation of a new concept can be regarded as an elaboration (E) of one
or more existing concepts that serve as the baseline (B). By viewing word formation
(the creation of a new word) as an elaboration that operates on a baseline (or baselines),
the relationship between concepts represented by morphemes as material for word
formation can be captured by the “quasi‐spatiotemporal relationship” in B/E
organization. In this way, we can rethink the logical dependence of concepts in terms
of spatial constraints and the unidirectional nature of B/E organization (or “logical
necessity”) in terms of temporal sequencing. This view provides a useful account of the
iconicity 1 (see Haiman, 1980, 1983, 1985; Van Langendonck, 1995, 2007; Croft, 2003)
of the relationship between meaning and form, because spatial and temporal constraints

1
As explained in Croft (2003: 102), “the intuition behind iconicity is that the structure of language
reflects in some way the structure of experience”.
1
manifest themselves in the form of complex words. In other words, the way a concept
is created motivates the structure and meaning of that concept. Furthermore, if a
particular type of complex words exhibits a particular behavior, incorporating the
notion of B/E organization may help explain its motivation. 2 This chapter aims to show
that B/E organization can be used to reveal the motivation in the following word
formation phenomena: headedness, the order in word formation, productivity, and the
dependency and manual nature of operations in word formation.
In human language, the “word” is the most basic unit. However, it is difficult
to precisely define what a word is. As Packard (2000: 7-14) states, words can be defined
from a variety of perspectives, including orthographic, sociological, lexical, semantic,
phonological, morphological, syntactic, and psycholinguistic; moreover, different
perspectives lead to different definitions (see also Matsumoto, 1996). Considering such
ambiguity of the term “word”, Taylor (2015) proposes a prototype approach by listing
the following criteria as prototypical word properties:

… a prototypical word will

 have a stable phonological form, intolerant of interruptions and internal
variation;
 be associated with a reasonably stable semantic content (or array of related
contents, in case the word is polysemous);
 be separated in writing by spaces;
 have one main stress and be pronounceable on its own, surrounded by
pauses;
 be relatively free with regard to the items to which it can be adjacent.
(Taylor, 2015: 10)

The present study adopts this sort of prototype approach to identify the status of words
based on various aspects, as described above.
Words can be divided into simple words, which have a unitary internal
structure, and complex words, which are composed of two or more elements. Word
formation is a research field that examines how such complex words are created. Since
the end of the 1990s, there has been increasing interest in word formation from the
perspective of Cognitive Semantics (Štekauer, 1998; Dirven, 1999; Koch, 1999;
Coulson, 2001; Tuggy, 2005; Benczes, 2006; Onysko & Michel, 2010; Umbreit, 2011).
So far, however, little attention has been paid to the role of iconicity in word formation

2
“A linguistic sign (target) is motivated to the extent that some of its properties are shaped by a
linguistic or non-linguistic source and language-independent factors” (Panther & Radden, 2011: 9).
2
(cf. Ungerer, 1999, 2002, 2007; Hiraga, 2005).
In this chapter, I will show that an analysis based on the concepts of baseline
and elaboration proposed by Langacker can shed new light on the iconic nature of word
formation, with a particular focus on compounding and derivation. 3 In Section 2, I will
introduce the concepts of baseline and elaboration and present a description of the sub-
category of elaboration as a reinforcement of the theory, which supports a continuum
of derivation and compounding and yields a new classification of compounds. An
analysis to examine the spatiotemporal parallelism between B/E organization and the
linguistic properties of complex words is presented in Section 3. Finally, Section 4
concludes the chapter with a summary of the main findings and future directions for
research and development of baseline elaboration theory.

2. Baseline elaboration theory and its implication for word formation

2.1 Baseline and elaboration

According to Langacker (2016a), asymmetry can be found in many aspects of human
language and cognition, which can be understood from the concepts of “baseline” and
“elaboration” in various dimensions and levels. He thus provides an integrative theory
for thinking about the process of creating new concepts and the conceptual structure
that is formed along the way, which encompasses concepts such as reference points
(Langacker, 1993) and conceptual integration (Fauconnier & Turner, 2002).
The baseline is generally more substantive than the elaborating elements
because it is already established, in place, or under control. Its elaboration can be
abstractly characterized as a function mapping B onto the higher-level structure BE.
The structure consisting of strata (S) formed as a result of elaboration is called B/E
organization (see Fig. 28.1) and is observed in many linguistic phenomena (Langacker,
2016a: 406-407).

3
Compounding and derivation are usually distinguished in the following way: Compounding is the
combination of two or more lexemes, whereas derivation is the addition of an affix to a lexeme (Booij,
2005a).
3
Figure 28.1 B/E organization (Langacker, 2015a)

According to Langacker (2016a: 407), stratum Si comprises an array of resources

available for structure building, including both mental capacities and existing structures.
This substrate allows for the formation of more elaborate structures. Such a dynamic
process can also be seen in word formation, where a complex word is created from
existing elements. Thus, it provides a means to analyze in a comprehensive manner the
materials upon which word formation is based, how they are used, and their
interrelationships in the structure after formation.
One essential feature of baseline elaboration theory is that it differs from
compositional theories. Langacker (2016a: 418-419) explains this difference by
providing examples of English plural forms. He points out that the plural forms of
English nouns are often compositionally analyzed (see Fig. 28.2). For instance, cats can
be analyzed as the composition of the noun cat and the plural morpheme s. However,
there are some plural forms, such as geese, that are not formed by adding another
element to the singular form. Therefore, in a compositional approach, the plural form
geese is considered a combination of goose and the process morpheme that changes oo
into ee. In sheep, the plural form is morphologically identical to its singular form.
Therefore, a compositional approach must propose something that is not pronounced
but is regarded as theoretically existing, something like a zero morpheme. 4 Adopting
the concept of a zero morpheme, the plural sheep can be considered as the composition
of the singular sheep and the zero morpheme.

4
“Zero morphemes or null morphemes are those without any phonological content that add
information to the lexeme” (Fábregas & Scalise, 2012: 32).
4
Figure 28.2 Morphology (plural forms) as composition (adapted from Langacker,
2016a: 418)

However, as Langacker points out, there is nothing to prove the psychological

existence of the process morpheme or zero morpheme. 5 If a morpheme is a building
block, as assumed in compositional approaches, how can it consist of nothing at all? Or
in a process?
To address this problem, in baseline elaboration theory, examples of English
plural forms are analyzed as shown in Fig. 28.3. The upper part in the squares divided
by the horizontal line represents the semantic pole, and the lower part represents the
phonological pole. When a baseline receives the elaboration of pluralization (PL) in the
semantic pole, some cases are morphologically combined with the plural morpheme s
(e.g., cats), and in some cases, such as goose, oo changes into ee. There are also cases
such as sheep, where semantic elaboration but no phonological elaboration is observed.

5
Zero morphemes are also problematic in relation to acquisition. For instance, Fábregas and Scalise
(2012: 32) state that because zero morphemes are not directly perceived, “it is not obvious how a child
realizes that her language has zero morphemes if she never gets to hear them”.
5
Figure 28.3 Morphology as B/E organization. (adapted from Langacker, 2016a: 419)

By viewing the plural forms of English as elaboration only on the semantic

pole (e.g., sheep) or on both the semantic and phonological poles (e.g., cats and geese),
we avoid the need to suggest a zero morpheme or process morpheme.
Although this type of nonconcatenative morphology has been discussed in
other studies (Sapir, 1921; Hockett, 1958; McCarthy, 1981; Spencer, 1998), baseline
elaboration theory provides a sophisticated framework for examining the elaboration
process, especially when we focus on the iconic relationship between conceptual
structure and morphological structure.
Construction Morphology (Booij, 2010, 2013), which assumes that all lexical
items are listed in the lexicon (the full entry theory, see Jackendoff, 1975; Jackendoff
& Audring, 2020), can also explain plural forms such as geese and sheep without
positing a zero morpheme or process morpheme. As Van der Spuy (2020) shows, in
English, regular plurals (e.g., cats) can be captured by general “phonological” schemas.
Examples such as geese can be captured by a paradigmatic link of a singular noun and
its /i:/ umlaut (a singular noun of the form /C1VC2/ [V is a high back vowel /u:/ or /ʊ/]
has a corresponding plural form /C1i:C2/ [C: consonant]). Examples such as sheep can
be captured by a paradigmatic link of a singular noun and its “syncretic plural” (the

6
same form as the singular but with plural meaning). 6
However, since Construction Morphology alone cannot explain the motivation
of different behaviors in different types of elaboration, we need to consider the
elaboration process to account for these differences. As Section 2.2 will illustrate, in
complex words, the process of elaboration itself may be represented by certain lexical
elements, which cannot be captured by the compositional approach. Furthermore, if we
look at the word formation from the perspective of baseline and elaboration, we can
capture the relationship between elements before the formation, the way they are
elaborated during the formation, and their structure after it, depending on the different
types of elaboration, which allows us to better capture the nature of complex words.
Langacker’s own studies have shown the utility and breadth of the application
of baseline elaboration theory, such as Langacker (2015b), which examines the central
aspects of English clauses; Langacker (2016b), which focuses on English nominal
quantifiers; Langacker (2017), which revisits the concept of evidentiality; and
Langacker (2019), which examines the representation of reality. Recently, the
application of this theory has increased, such as Hummel (2018), which explores the
relationship between structure, construction, and baseline elaboration; Wilcox and
Martínez (2020), which examines the conceptualization of space in signed language
discourse, particularly the relationship between pointing and places, from the concept
of baseline and elaboration; and Chen (2020), which explores compound verbs 7
representing macro-events (Talmy, 2000) cross-linguistically based on baseline
elaboration theory. In this chapter, I will delve further into other word formation
processes by proposing a new classification of complex words based on different types
of elaboration.

2.2 Classification of elaboration

In this study, the type of elaboration was examined in more detail and classified
into subtypes. In addition to the concepts of baseline and elaboration, this study
proposes a concept called “additive”. By doing so, elaborations can be classified into

6
Relatedly, Jackendoff and Audring (2020: 78-79) explain irregular plurals (sheep and goose) by
means of “interface links” between the meaning, morphosyntax, and phonology of a word (“the parallel
structure”) and “relational links” between corresponding components of a singular noun and its plural
counterpart, which also avoids using the notions of zero morpheme and process morpheme.
7
According to Kageyama et al. (2013), compound verbs are very common in northern Indo-Aryan
languages and are less common in other Indo-Aryan languages. They are also found in Dravidian,
Turkic, Korean, Japanese, Chinese, some Tibeto-Burman languages, some Northeast Caucasian
languages, and a few South American languages.
7
three subtypes: coordination, augmentation, and adaptation. 8

2.2.1 Coordination
The first type of elaboration, coordination, is the process of combining multiple
baselines that are equally substantive in the same stratum S0 to create a new concept
(Fig. 28.4). Thus, there is no temporal difference between multiple baselines.

Figure 28.4 B/E organization of coordination

Since the coordination 9 type must be semantically and morphologically

symmetrical in the baselines, it is a process that creates a new cohesion by lining up
two (or more) concepts that are similar or opposite, as shown in (1).

(1) Examples of coordination complex words

Compounding:
singer-songwriter (B1-B2),
bittersweet (B1-B2),

8
While this chapter focuses on the most productive word formation processes of compounding and
derivation, there are also word formation processes such as duplication, backformation, conversion,
clipping, blending, and acronym (see Müller et al., 2015). It should be noted that in blending, the word
formation process of “compounding” after “ellipsis” involves two stages of elaboration.
9
The term “coordinate compound” is also called dvandva, co-compound, copulative compound, pair
word, etc.
8
 stir-fry (B1-B2),
mind-brain (B1-B2),
shoo-chiku-bai 松竹梅 (pine-bamboo-plum) ‘an auspicious grouping’
[Japanese] (B1-B2-B3),
shun-ka-shuu-too 春夏秋冬 (spring-summer-fall-winter) ‘four seasons’
[Japanese] (B1-B2-B3-B4),
daki-kakaeru 抱き抱える (embrace-hold) [Japanese] (B1-B2),
cǎi-tà 踩踏 (stamp-step.on) [Chinese] (B1-B2), etc.

Coordination is considerably more restrictive than the two types described

below. In the word formation process, the coordination type can be seen in compounds
but not in derivatives because the constituents of the coordination type need to be
equally substantive, which contradicts the definition of derivation (the addition of an
affix to a lexeme, which is inherently asymmetric). As to the morphosyntactic
restriction, both baselines need to have the same grammatical category. In the semantics
of baselines, synonyms and concepts belonging to the same semantic category are
inevitably restricted, and when it comes to antonyms, there is almost a one-to-one
correspondence.

2.2.2 Augmentation
Augmentation (Fig. 28.5) is an elaboration in which an additive (A) is embedded in the
empty slot of a baseline.

Figure 28.5 B/E organization of augmentation

(2) Examples of augmentation complex words

9
 Derivation:
ui-zan 初産 (first-child.birth) [Japanese] (A-B),
dosu-guroi どす黒い (muddy-black) [Japanese] (A-B), etc.
Compounding:
taxi driver (A-B),
headhunt (A-B),
sunrise (A-B),
stripe hat (A-B),
soo-kin (send-money) ‘remit’ [Japanese] (B-A),
mai-ochiru 舞い落ちる (dance-fall) ‘flutter down’ [Japanese] (A-B),
huá-xíng 滑行 (slide-go) ‘glide’ [Chinese] (A-B),
aruki-tsukareru 歩 き 疲 れ る (walk-be.tired) ‘get tired from walking’
[Japanese] (A-B),
dòng-sǐ 凍死 (freeze-die) ‘freeze to death’ [Chinese] (A-B), etc.

Augmentation is the embedding of a semantically compatible additive concept

in an unspecified slot in the baseline concept. 10 Thus, it is asymmetric in nature. For
instance, in the case of stripe hat, the concept of stripe is embedded in the empty slot
of [pattern] in the concept of the baseline, hat. There are other types of empty slots in
the concept of hat, such as [material] (silk hat and straw hat), [use] (construction hat
and swimming hat), and [shape] (bubble hat). There are also cases in which the concept
is embedded in a slot in the baseline verb. For instance, in the Sino-Japanese compound
noun soo-kin 送金 (send-money) ‘remit’, kin represents the concept of money, which
is embedded in a slot in the baseline concept of soo ‘send’. There are other slots in the
concept of sending, such as ways of sending, e.g., yuu-soo 郵送 (mail-send) ‘mail’
and riku-soo 陸送 (land-send) ‘land transport’, and destination: soo-ken 送検 (send-
prosecutor) ‘sending the person accused to the prosecutor’).
When combining two elements in the case of augmentation compounds, the
two concepts belong to the same stratum S0 as the material at hand to create a new
concept, such as when combining two blocks or two pieces of a jigsaw puzzle.
Langacker (2016a) provides the example of cat food and analyzes it as a dual baseline.
However, considering the existence of the “head” in a compound (see, among others,
Williams, 1981; Scalise et al., 2009), it is better to analyze cat food as the augmentation
of a baseline (food) with an additive (cat). Cat food is a kind of food, but not a kind of
cat. It is coordination compounds whose components are equally substantive and

10
In the case of augmentation compounds, whether an additive can fit into an open slot of a baseline or
not is strongly dependent on this sort of “conceptual congeniality” (see Chen, 2016; Chen &
Matsumoto, 2018).
10
constitute a dual baseline (e.g., singer-songwriter). The concept of headedness will be
explained in light of the baseline elaboration theory in Section 3.1.

2.2.3 Adaptation
In the elaboration of adaptation, E is an adaptation relating B to a higher-level structure
B’. Whereas B is the material existing in stratum S0, E is the operation performed on B
to create a new concept. Thus, in cognition, it always involves temporal ordering on a
certain time scale (Langacker, 2016a: 406).

Figure 28.6 B/E organization of adaptation

(3) Examples of adaptation complex words

Derivation:
education (educate: B, -tion: E),
awareness (aware: B, -ness: E),
quickly (quick: B, -ly: E),
entitle (en-: E, title: B),
unhappy (un-: E, happy: B),
rewrite (re-: E, write: B),
higeki-teki 悲劇的 (tragedy- adjectivizing.suffix) ‘tragic’ [Japanese] (B-E),
sai-riyoo 再利用 (re-use) ‘reuse’ [Japanese] (E-B), etc.
Compounding:
kanzen-nenshoo 完全燃焼 (perfect-combustion) ‘complete combustion’
[Japanese] (E-B),
kaki-hajimeru 書き始める (write-start) ‘start writing’ [Japanese] (B-E),

11
 khā-un ghe-ṇe (eat-CP 11 take-INF) ‘eat up’ [Marathi, Kageyama et al., 2013:
7] (B-E),
Ilk-e noh-ass-ta (read-CP put-PST-DECL) ‘read (in anticipation of …)’
[Korean, Suh, 2000: 78] (B-E),
hé-biān 合編 (be.together-edit) ‘co-edit’ [Chinese] (E-B), etc.

The type of adaptation is one in which a baseline undergoes elaboration of conceptual

operation, such as specification of the development of a situation, negation, or
pluralization. These conceptual operations are basically dependent and require the
support of autonomous concepts for their full manifestation. Thus, conceptual
operations are likely to be expressed by bound morphemes.
Semantic elaboration of adaptation often involves morphological means of
inflection and derivation. For example, the English plural cats, which semantically
expresses conceptual operation by pluralization of a concept, is expressed by the
morphological means of inflection, and in the case of derivation, as in (3), the
elaboration of adaptation is expressed using prefixes (e.g., en-, un-, re-) or suffixes (e.g.,
-tion, -ness, -ly). Adaptation can also be expressed by compounding, as in the case of
compound verbs in Japanese, Chinese, Korean, and South Asian languages.

2.3 Identification as a word

Before moving to the implications of this new typology of complex words, I would like
to highlight some of the examples mentioned in Section 2.2 and show why they qualify
as words.
Let us start with the example of taxi driver. Considering its semantic content,
a person who happens to drive a taxi (for example, a fugitive who stole a cab) does not
fit the definition of a taxi driver. From a phonological point of view, it is well-known
that compounds tend to have a main stress near the beginning of the combination (Fudge,
2015), which can be seen in the case of taxi driver as well. These unique properties
support the wordhood of taxi driver, even though there is a space between the two
elements.
As for compound verbs in Japanese, such as kaki-hajimeru (write-start) ‘start
writing’, it is known that they are very likely to be treated as accented verbs, no matter
which accent class the two element verbs belong to (Vance, 2008: 191), indicating that
the compounds behave phonologically as a single word. In addition, from a grammatical
point of view, as Matsumoto (1996: 171) states, the two compounded verbs cannot be

11
Abbreviations are as follows: 1 (first person), ACC (accusative), CAUS (causative), CP (conjunctive
participle), DECL (declarative), FUT (future), GER (gerund), IMP (imperative), INF (infinitive), M
(masculine), NPST (nonpast), PRF (perfect), PL (plural), PST (past), SG (singular), and TOP (topic).
12
separated from each other by any particles (e.g., *kaki wa hajimeru [write TOP start]),
and they participate in Renyookei Nominalization as a whole (e.g., kaki-hajime ‘the
beginning of writing’).
Regarding compound verbs in Chinese, especially resultative compound verbs
such as dòng-sǐ (freeze-die) ‘freeze to death’, there is an ongoing debate about whether
they are words or phrases. In this study, Chinese resultative V–V is treated as a
compound words, as in Li (1990), Cheng and Huang (1994), Packard (2000), Her
(2007), Ceccagno and Basciano (2009), and Lee and Ackerman (2011). As Packard
(2000) shows, the individual constituents of resultative compound verbs are not
available for syntactic operations (such as embedding, movement, deletion, insertion,
and anaphora), thus conforming to the Lexical Integrity Hypothesis (Selkirk, 1982; Di
Sciullo & Williams, 1987; Spencer, 1991; Scalise & Guevara, 2005; Booij, 2009;
Haspelmath & Sims, 2013; Kageyama, 2016).

Lexical Integrity Hypothesis: Rules of syntax can refer/apply to entire words

or the properties of entire words, but not to the internal parts of words or their
properties. (Haspelmath & Sims, 2013: 203)

As shown above, the identification of words requires a multifaceted judgment

from different types of evidence, including phonological, semantic, and grammatical.

2.4 The implications of the new typology of complex words

The new typology of complex words proposed in Section 2.2 provides new insights into
the distinction of inflection, derivation, and compounding, and the classification of
compounds.

2.4.1 The continuum of inflection, derivation, and compounding

Some argue that inflection and derivation are distinct (Kiparsky, 1985; Perlmutter,
1988; Anderson, 1992), while others argue that they are continuous (Bybee, 1985). The
basis for the distinction between inflection and derivation is that inflection occurs only
after derivation. For instance, as in the case of work-er-s, the inflection process of the
pluralization -s occurs after the derivation process of the affix addition -er (Greenberg,
1963). However, there are exceptions to this generalization. For instance, in Dutch held-
en-dom (root-PL-collective) ‘heroism’, the inflection occurs before the derivation (Booij,
2006: 659).
Questions have also been raised about the demarcation between derivation and
compounding. According to Booij (2005a), the criterion for distinguishing derivation
from compounding depends on whether the constituent can be used alone, that is,
13
whether the constituent is a lexeme. However, derivation and compounding are not
clearly distinguishable; they are continuous. For example, the Dutch words onder
‘under’ and over ‘over’ are lexemes because they can be used alone. However, despite
the rule of word stress on the first constituent in Dutch compounds, there is no word
stress on the first constituent in onder-titel (under-title) ‘to subtitle’ and over-brog
(over-bridge) ‘to bridge’ (Booij, 2005a). This shows that derivation or compounding is
not always a clear-cut decision (see also Kageyama, 1993: 13).
In Section 2.2.3, I show that the adaptation type ranges over inflection,
derivation, and compounding. This fact supports the idea that inflection, derivation, and
compounding are not distinct but form a continuum.

2.4.2 The classification of compounds

Previous studies on the classification of compound words have mainly focused on the
presence or absence of head as the primary classification criterion, such as Spencer
(1991), Bauer (2001), and Booij (2005b). Recently, aiming at a comprehensive
classification of compound words, Scalise and Bisetto (2009) proposed a classification
method based on three levels—1) the grammatical relation of constituents, 2) the
subclassification of grammatical relations, and 3) the presence/absence of a head—as
shown in Fig. 28.7.

Figure 28.7 Classification of compound words in Scalise and Bisetto (2009: 50)

According to Scalise and Bisetto (2009: 50), compound words can be

classified into three types based on the grammatical relationship of their constituents:
subordinate (SUB), attributive/appositive (ATAP), and coordinate (COORD). In the
case of subordinate compounds, the two constituents have a head-complement
14
relationship. For example, in the compound taxi driver, taxi is considered to be the
complement of the deverbal head driver. In an attributive/appositive compound, the
head of the noun is modified by the non-head part. For example, in blue cheese, the
non-head blue represents one of the attributes of the head cheese. Words such as
painter-poet are considered to be a coordinate compound, in which the two components
are linked by the semantic relation “and”.
The second level is a further subclassification of the compound words of the
subordinate and attributive/appositive types. In the second level, subordinate
compounds are further subdivided into verbal-nexus and ground, depending on whether
the head is verb-derived or not. In the verbal-nexus, the head, seller, is a deverbal noun,
as in bookseller, and the non-head part can be captured from the relation to the verb on
which the head is derived (the non-head part corresponds to the argument, complement,
or adjunct of the verb). In the case of the ground type, on the other hand, as in mushroom
soup, the head is a non-verb, and the relationship between the head and the non-head
part is captured by encyclopedic knowledge. In the case of attributive/appositive
compounds, they are subdivided into attributive, where the head is modified by a non-
head adjective, as in high school, and appositive, where the head is modified by a non-
head noun or a non-head verb, as in mushroom cloud. The difference between the
appositive-type mushroom cloud and the ground-type mushroom soup is that the non-
head part of the appositive type (mushroom) modifies the head (cloud) metaphorically.
These subclassified compounds are further distinguished by the presence or absence of
a head (endocentric or exocentric structure).
The problem with this classification is that it only focuses on compound nouns.
There are some types of compound word, especially compound verbs of the [V-V]V
type, which are not captured by this taxonomy. For example, in Japanese, yomi-
hajimeru 読み始める (read-start) ‘begin to read’ and tataki-kowasu 叩き壊す (hit-
break) ‘smash’ would both belong to the subordinate verbal-nexus type according to
the definition of this taxonomy. However, as we will see in Section 3, yomi-hajimeru
(of the adaptation type in this study) and tataki-kowasu (of the augmentation type in
this study) have many different properties and need to be treated as different types.
Some compound words, such as the compound verb yomi-hajimeru and the compound
noun kanzen-nenshoo 完全燃焼 (perfect-combustion) ‘complete combustion’, are
morphologically compositional but, semantically, many of them should not be thought
of as a composition of two concepts but in terms of certain conceptual operations to a
single underlying concept (baseline).
The aim of this chapter is not to deny the existing classifications completely,
but to present a new level of classification based on B/E organization to explain the
motivation behind the nature of compound words.
15
2.5 Structured mapping in word formation
By adopting the B/E organization, a tool that makes explicit the mental processes
involved in creating a new concept, we can account for the linguistic properties of a
complex word. According to Emmorey (2014), who applied structure-mapping theory
(Gentner, 1983) as a framework for understanding the effects of iconicity on sign
language grammar and processing, rather than understanding iconicity as a link
between linguistic form and human experience, it is better to understand it as a
structured mapping between two mental representations. Given this idea, this study
proposes a new hypothesis regarding the motivation in word formation.

(4) Structured mapping in word formation

The quasi-spatiotemporal relationship in the B/E organization that emerges
during the formation of a complex word shapes the linguistic properties of that
word.

In other words, the linguistic constraints or characteristics of a complex word arise

explicitly by structure-preserving mappings from its conceptual structure. Thus, the
linguistic properties of a complex word are motivated by its conceptual organization.
On the basis of this hypothesis, we can assume that different types of
elaboration have different linguistic properties, and we can make the following
predictions about the linguistic behavior of a complex word based on its B/E
organization.

(5) Predictions based on the structured mapping in word formation

a. Headedness:
The interrelationship of the components in combination will determine which
component is the more important one. In the coordination type, there is no
difference in importance because it is simply the juxtaposition of two similar
or opposite things. In the augmentation type, an additive is embedded in the
space of the base (baseline). Therefore, the baseline concept is more important
because it is the foundation of the additive. In the adaptation type, the baseline
is more important than the elaboration because the elaboration only reshapes
the baseline.

b. The order in word formation:

The baseline is something prior to the elaboration in the adaptation type. Thus,
based on the structured mapping in word formation, we can predict that the
16
 order of lexical elements in an adaptation complex word X–Y tends to be “X:
baseline, Y: elaboration”, which matches the temporal sequencing of B/E
organization (baseline → elaboration). In the case of augmentation complex
words, because the additive and the baseline exist in the same stratum S0, there
should be no effect of the structured mapping in word formation regarding the
order of the baseline and additive.

c. Productivity:
In the adaptation type, like kneading a lump of clay to change its initial form
and create something new, elaboration is an independent operation performed
on a baseline. Thus, the elaboration should be applicable to a variety of
baselines. In the augmentation type, by contrast, an additive is something to
fill a baseline’s empty slot, rather like assembling a jigsaw puzzle. The additive
and the baseline are mutually restricted, which is likely to cause augmentation
complex words to show limited productivity. In the coordination type, the
number of similar concepts is quite limited, and when it comes to opposite
concepts, the relationship is almost one-to-one, which means that the
combinatory pattern is fairly small.

d. The representation of the elaboration in adaptation compound verbs:

Because elaborations are a certain kind of operation performed on baselines,
verbs representing elaborations are likely to be relatively limited. Since
operations are mainly performed by hand, it is expected that there are many
verbs representing hand movements.

These predictions about the morphological behavior of a complex word based on its
conceptual structure align with the area of “Onomasiology”, which is the study of the
means of expressing a given concept (see Baldinger, 1980). In Section 3, I will show
that this hypothesis can make precise predictions about the linguistic behavior of a
complex word, focusing on headedness, order in word formation, productivity, and the
dependency and manual nature of operations in word formation.

3. Analysis

3.1 Headedness
The concept of headedness has been discussed in many previous studies (Lieber, 1981;
William, 1981; Kiparsky, 1982; Selkirk, 1982; Zwicky, 1985). More recently, Scalise
et al. (2009) presented a typology of compound words by splitting the notion of the
17
head into three subparts: categorial head, semantic head, and morphological head. Here,
we will focus on the semantic head.
The problem here is the predictability of the head. The classifications in
previous studies have not provided a good explanation for the location of the head for
each type of compound word. By contrast, in the present classification, each type has
its unique headedness characteristics. Specifically, in the classification of this study,
the semantic head of a complex word can be determined on the basis of its conceptual
structure after elaboration (the structure of S1 in Fig. 28.4, 28.5, and 28.6).
In the coordination type, B1 and B2 are equivalent, so both are semantic heads.
For instance, if a person is an actor-author, then s/he is both an actor and an author. In
this case, one might think that there are no heads, but that would not distinguish it from
a compound word with an exocentric structure. Therefore, in this study, we consider
that multiple heads coexist in the coordination type.
In the augmentation type, the baseline is the semantic head, since baseline B
is the foundation of additive A. The component that will be the baseline is determined
by the relationship of the slots. For instance, in the case of stripe hat, the concept hat
contains information about its appearance, such as pattern and color, which becomes an
available slot (see Chen, 2015; Chen & Matsumoto, 2018). By contrast, the notion of
stripe does not seem to contain information about what it is used for directly. Therefore,
hat becomes the baseline, and the concept represented by stripe is embedded in the slot
of the pattern in hat.
As for the augmentation compound verb mai-ochiru (dance-fall) ‘flutter down’,
the notion of movement “fall” contains the relevant information about how it is
generally performed (including the information about leaves fluttering down), which is
an open slot. On the contrary, the concept of the action “dance” does not contain the
relevant information about the movement of falling. Therefore, fall becomes the
baseline, and the concept of dance is embedded as an additive in the slot of its manner.
In this case, the reason why the order is additive-baseline is probably because, in
Japanese, adverbs (or adverbial clauses) expressing manner are placed before the verb
they modify, as in mai-nagara ochiru (dance-while fall). In fact, in Chinese, where
adverbs are placed before verbs in the same way, compound verbs describing a
movement and its manner also have an additive-baseline word order, as in huá-xíng 滑
行 (slide-go) ‘glide’. In sum, we can determine which constituent is the semantic head
of a complex word based on this sort of unidirectional relationship of conceptual slots.
In the adaptation type, a single baseline B undergoes concept operation and
becomes B’. Therefore, what represents baseline B is basically considered to be the
semantic head. In the case of Japanese compound verbs of the adaptation type, V1 is
the baseline and often represents the head, as shown in (6).
18
 (6) a. フルマラソンを ｛走り切った/走った/*切った｝
furumarason o {hashiri-kit-ta/hashit-ta/*kit-ta}
full marathon ACC {run-cut-PST/run-PST/*cut-PST}
‘Ran a full marathon’ (kiru ‘cut’ adds the meaning of completion)
b. 村上春樹を ｛語り尽くす/語る/*尽くす｝
murakamiharuki o {katari-tsukusu/kataru/*tsukusu}
Murakami Haruki ACC {discuss-exhaust/discuss/*exhaust}
‘Discussing Haruki Murakami’ (tsukusu ‘exhaust’ adds the meaning of
completion)

However, in the adaptation type, the conceptual structure after elaboration is

an indivisible concept (S1 in Fig. 28.6). Therefore, when the baseline concept and the
concept after elaboration differ significantly, it is difficult to identify a specific part as
the semantic head. For instance, the meaning of the Japanese adaptation compound ii-
sokoneru (say-fail) ‘fail to say’ in (7a) is not identical to the meaning of its components
in (7b).

(7) a. 私はお礼を 言い損ねた.

watashi wa orei o ii-sokone-ta
I TOP gratitude ACC say-fail-PST
‘I failed to say thank you.’
b. 私はお礼を ｛言った/*損ねた｝.
watashi wa orei o {it-ta/*sokone-ta}
I TOP gratitude ACC {say-PST/*fail-PST}
‘I said thank you./lit. I failed gratitude.’

The expression with a V2 alone does not make sense on its own, and the expression
with a V1 alone is possible but has a different meaning (‘I said thank you’ is not the
same as ‘I failed to say thank you’). In such cases, it is difficult to determine which one
should be considered the head.
There are also so-called exocentric (headless) structures. For instance, the
compound word green beret does not denote a type of color, nor does it represent a kind
of hat. Such complex words have an exocentric structure. In terms of B/E organization,
green beret can be thought of as a conceptual elaboration of a metonymic shift added
to an original augmentation type complex concept (green: A, beret: B), which makes
green beret eventually an adaptation type compound without a head.

19
3.2 The order in word formation
The order of components in a complex word can be seen as a different constraint for
each B/E organization. To illustrate this, this section focuses on Japanese compound
verbs in which coordination, augmentation, and adaptation are found in the same form,
[V-V]V.
First, in the coordination type, B1 and B2 must be equivalent. Therefore, it is
no problem to form a coordination compound verb such as daki-kakaeru 抱き抱える
(embrace-hold) ‘embrace (in one’s arms)’ first and then add the concept to the baseline
ageru 上げる ‘raise’, as in daki-kakae-ageru 抱き抱え上げる ‘lift (a person) in one’s
arms’ (the underlined part indicates the baseline). However, if we form the
augmentation compound verb daki-ageru 抱き上げる (embrace-raise) ‘lift in one’s
arms’ first, it cannot be formed as a coordination compound verb with a simple verb,
and thus *daki-age-kakaeru 抱き上げ抱え る cannot be formed. Similarly, if the
adaptation compound verb naki-hajimeru 泣 き 始 め る (cry-start) ‘start to cry’ is
formed first, it cannot be formed as a coordination compound verb with a simple verb,
as in *naki-hajime-sakebu 泣き始 め叫ぶ (cry-start-shout); note that naki-sakebu
(cry-shout) is a possible coordination compound verb.
Next, in the augmentation type, a baseline must have some slots (something
that can specify a property, such as [color] or [shape] in a concrete object, or [cause] or
[means] in a state-change event), and the additive must fit into the slot. Compound
verbs that undergo the elaboration of coordination (e.g., daki-kakaeru [embrace-hold])
can be used as an additive to form an augmentation compound verb, such as daki-kakae-
ageru (embrace-hold-raise), but those that undergo adaptation (e.g., daki-tsuzukeru
(embrace-continue) ‘continue to embrace’) cannot be used to form an augmentation
compound verb as an additive, such as *daki-tsuzuke-ageru (embrace-continue-raise). 12
Japanese compound verbs that have undergone coordination or adaptation cannot be
the baseline of the augmentation type because of their high conceptual specificity. 13
The last type of elaboration, adaptation, only requires its baseline to be
something that can undergo elaboration. Because this type has the fewest restrictions
on the baseline, it is often compounded with coordination compound verbs (tobi-hane-
hajimeru [jump-bounce-start] ‘start jumping’) and augmentation compound verbs
(tataki-kowashi-makuru [hit-break-do.over.and.over.again] ‘smash it up’).

12
It should be noted that in the case of compound nouns, an adaptation compound can be used to form
an augmentation compound as an additive, as in ren-shoo-kiroku (consecutive-win-record) ‘record of
consecutive victories’.
13
Examples such as tochi-bai-bai (land-sell-buy) ‘land trading’ and kaimaku-ren-shoo (opening-
consecutive-win) ‘opening winning streak’, in which a coordination or adaptation compound can be
used as the baseline to form an augmentation compound, exist in compound nouns.
20
 In the case of Japanese compound verbs, these three types need to follow a
specific order of occurrence: [[[coordination]-augmentation]-adaptation]. That is,
augmentation-coordination, adaptation-coordination, and adaptation-augmentation are
all impermissible orders of occurrence. Some may argue that these phenomena can be
easily explained by grammaticalization, that is, more grammaticalized morphemes
always occur in the outer layer. However, from the point of view of grammaticalization,
there is no difference between coordination and augmentation types, which cannot
explain why the augmentation type never precedes the coordination type in Japanese
compound verbs.
The B/E organization not only predicts the order of multiple elaborations
described above but also has implications for the order of verbs in a single elaboration
type. Assuming that coordination, augmentation, and adaptation compound verbs are
morphologically distinct, we can make the following predictions about the
morphological behavior of a compound verb based on its B/E organization. The
baseline is something prior in the adaptation type. Thus, on the basis of the structured
mapping in word formation, we can predict that the order of verbs in adaptation
compound verbs tends to be “V1: baseline, V2: elaboration”, which matches the
temporal sequencing of B/E organization (baseline → elaboration). In the case of
augmentation compound verbs, because the additive and the baseline exist in the same
stratum, S0, there should be no effect of the structured mapping in word formation
regarding the order of baseline and elaboration presented in a complex word.
In a cross-linguistic study of compound verbs, Chen (2020) has shown that the
order of adaptation compounds is unified as predicted in all OV languages examined,
as demonstrated in (8). However, as (9) demonstrates, VO languages do not conform to
a unified order. The order is inconsistent even within a single language. This difference
suggests that there is competition between regular word order and order based on iconic
relations. The order of an object and a transitive verb in a language will affect the order
of verbs in adaptation compounds because they both represent an operation on
something. It is very likely for a language to adopt the same strategy to encode events
of operation. If an iconic relation exists between the word order and the temporal
sequencing of B/E organization, as in the adaptation compound verbs of OV languages
(see Fig. 28.8), the order of verbs tends to be “V1: baseline, V2: elaboration”. Otherwise,
the order of verbs will be “V1: baseline, V2: elaboration”, based on the iconicity (e.g.,
chū-jìn 出 盡 [exit-complete] in Chinese) or the reverse, “V1: elaboration, V2:
baseline”, based on the word order VO (e.g., jìn-chū 盡出 [complete-exit] in Chinese).

(8) Adaptation compound verbs in OV languages

bak-a-kal-dı (watch-GER-remain-PST) ‘keep looking’ [Turkish, Kuribayashi,
21
 2013: 291] (B-E),
jaan li-yaa (know take-PRF.M.SG) ‘know completely’ [Urdu, Butt, 1993: 116]
(B-E),
hashiri-hajimeru 走り始める (run-start) ‘start running’ [Japanese] (B-E),
yomi-kakeru 読みかける (read-put.on) ‘be about to read’ [Japanese] (B-E),
etc.

(9) Adaptation compound verbs in VO languages

jìn-chū 盡出 (complete-exit) ‘be all out’ [Chinese] (E-B),
chū-jìn 出盡 (exit-complete) ‘be all out’ [Chinese] (B-E),
af-nak-üm (stop-go.down-CAUS) ‘stop going down’ [Mapudungun, Smeets,
2007: 416] (E-B),
gana-ntuku- (earn-put.at-) ‘to continue to earn’ [Mapudungun, Smeets, 2007:
317 ] (B-E), etc.

Figure 28.8 Iconic relation of adaptation compound verbs (dashed lines represent the
iconic correspondence)

It is interesting to note that in Japanese, in addition to using compound verbs

such as tabe-owaru 食べ終わる (eat-finish), it is also possible to use Sino-Japanese
compound nouns such as kan-shoku 完食 (finish-eating) with the light verb suru す
る (do) to express almost the same meaning. Adaptation Sino-Japanese compound
nouns 14 such as kan-shoku are in the E-B order, contrary to the claim here that Japanese
adaptation compound verbs are in the B-E order, and it seems to be a counterexample.
However, Sino-Japanese compound nouns are arguably attributable to the patterns of
Chinese morphology that reflect the Chinese word order (Namiki & Kageyama, 2016:
210-212; see also Kobayashi et al., 2016). In fact, we can find examples of adaptation

14
In Japanese, the lexicon comprises four strata: Wago (native Japanese words), Kango (Sino-Japanese
words), Gairaigo (Foreign words), and Mimetics (Kageyama & Saito, 2016). Sino-Japanese words are
words originally borrowed from Chinese and are written in Chinese characters (Kanji), although there
are a great number of Sino-Japanese words created in Japan (Kobayashi et al., 2016).
22
Sino-Japanese compound nouns where the word order is B-E, as in ho-kan 補完
(supplement-finish) ‘complement’. Thus, the possible word orders of adaptation Sino-
Japanese compound nouns (both E-B and B-E) are the same as in Chinese adaptation
compound verbs, which further supports this study’s claim.
In Quichua-influenced Spanish (see Lipski, 2014), word order also affects the
order of compound verbs. Quichua, the predominant indigenous language spoken in the
Andean highlands of South America, has been reported to frequently exert phonological
and morphosyntactic effects on regional varieties of Spanish. Influenced by Quichua,
in which the linguistic form of a compound verb is productive, speakers living in the
Andean highlands have produced innovative compound verbs in Spanish. As (10)
demonstrates, the OV language Quichua originally had the compound verb huañu.či-
špa šita-šun (kill-GER throw-1.PL.FUT). However, in the corresponding Spanish
compound verb, the order of the contents represented by V1 and V2 was reversed
because of the influence of the word order of Spanish (subject-verb-object; SVO). Here,
the verb for ‘kill’ is the baseline and the verb for ‘throw’ is the elaboration.

(10) a. huañu.či-špa šita-šun

kill-GER throw-1.PL.FUT
‘We will kill you.’ (Quichua, Hook, 2013)
b. bota-remos mat-ándo-te
throw-1.PL.FUT kill-GER-you
‘We will kill you.’ (Quichua-influenced Spanish, Hook, 2013)

In the case of augmentation compounds, because the relation of an additive

and a baseline is not an operation, the regular word order and the temporal sequencing
of B/E organization will not affect the order of verbs. Other iconic factors, such as the
principle of temporal sequence (Tai, 1985), may cause the order of verbs in
augmentation compounds to tend to be “V1: co-event (subordinate event), V2: framing
event (main-event)” cross-linguistically (see Chen, 2020). As for the coordination type,
because the constituents are equally substantive, the order cannot be determined by B/E
organization but by other semantic, pragmatic, and phonological factors (see Kageyama,
1982; Labrune, 2006).
The difficulty of finding other VO languages that have the form of [V-V]V
compound verbs other than Chinese and Mapudungun (and maybe Quichua-influenced
Spanish) is noteworthy. 15 This phenomenon suggests that the iconic relation between

15
Although we can find compound verbs such as stir-fry in English, they are not productive (see
Lamberty & Schmid, 2013; Bagasheva, 2020).
23
word order and order of temporal sequencing in adaptation compound verbs may be the
key to the presence of productive compound verbs. Nonetheless, what matters here is
that many OV languages show a unified tendency, and only two VO languages show an
inconsistent tendency.
The B/E organization also provides useful insights into derivation. Although
the adaptation type can be found in compound words such as chī-wán 吃完 (eat-finish)
‘finish eating’ in Chinese and nomi-hajimeru 飲み始める (drink-start) ‘start drinking’
in Japanese, it often uses derivations to express elaboration by affixes. As described
above, the order of the constituents of an adaptation complex word X–Y tends to be “X:
baseline, Y: elaboration” based on the temporal sequencing of B/E organization.
Consequently, we can predict that the elaboration of adaptation is more often expressed
by suffixes than prefixes. Indeed, it is well known that suffixes are richer than prefixes
in world languages (Greenberg, 1963; Hawkins & Gilligan, 1988; Bybee et al., 1990;
Dryer, 2011; Himmelmann, 2014). Thus, we can see that the iconicity of temporal
sequencing in B/E organization supports the predominance of suffixes.

3.3 The difference in productivity

The productivity of complex words can also be explained by a classification based on
B/E organization. In other words, there is a difference in productivity depending on the
structural differences between the concepts in B/E organization. First, the productivity
of the coordination type is expected to be extremely low because only equivalent
concepts can become a complex word of the coordination type. Second, the
augmentation type has semantic restrictions imposed by the slot, much like the shape
limitations in completing a jigsaw puzzle. If the concepts are semantically compatible,
they can be formed into complex words. Thus, we can expect that there will be both
low-productivity words, such as the Japanese compound verbs V-akeru (V-open) ‘open
by Ving’, which can only be combined with a limited number of the means of opening
(e.g., keru ‘kick’, osu ‘push’, etc.), and high-productivity words, such as English
deverbal nouns -er, in which the suffix er can be combined with a variety of verbs
representing the baseline of events that an entity involved. Third, since the adaptation
type only needs to be able to undergo elaboration, it is expected that many of them will
be highly productive.
In Chen (2017), I tested these hypotheses about productivity through a corpus-
based study. When examining productivity, there are large amounts of derivates and
compound nouns. Therefore, given the data-collection possibilities, I collected
frequency information of compound verbs in Japanese based on data from the Balanced
Corpus of Contemporary Written Japanese (BCCWJ;
https://pj.ninjal.ac.jp/corpus_center/bccwj/en/). For each type, I collected 20 [[ ]V-
24
V2]V (V1 is an open slot) with different V2s in order of token frequency in the BCCWJ
corpus. Adopting the method of Asao (2007), Baayen’s (1992) productivity P is
calculated using equation (11), where N is the total number of tokens of compound
verbs with a certain V2 and n1 is the number of words that occur only once (hapax
legomena).

n1
(11) 𝑃𝑃 = N

Note that, as Booij (2005b: 69-70) states, it is problematic to use the notion of
type frequency (the number of different words [word types] of a certain morphological
type) to measure the degree of productivity. A high type frequency of a morphological
pattern does not necessarily imply a high degree of productivity since the class of words
of that type may be a closed set that is (almost) never expanded. Instead, the
productivity P proposed by Baayen provides a way to measure the possibility for
language users to coin new words.
The results of the corpus-based analysis are shown in Tables 28.1–28.3.

Table 28.1 Productivity of coordination compound verbs

[[ ]V-V2]V Total token A word that Productivity
frequency appears only (P)
(N) once (n1)
[[ ]V-matou (put.on)]V 539 0 0
[[ ]V-hibiku (sound)]V 373 0 0
[[ ]V-sigeru (grow.thick)]V 252 0 0
[[ ]V-motomeru (seek)]V 240 0 0
[[ ]V-kakaeru (hold)]V 239 0 0
[[ ]V-sakebu (shout)]V 226 0 0
[[ ]V-nozomu (hope)]V 225 0 0
[[ ]V-kagayaku (sparkle)]V 215 0 0
[[ ]V-kotaeru (bear)]V 203 0 0
[[ ]V-tatsu (be.built)]V 200 0 0
[[ ]V-haneru (bounce)]V 174 0 0
[[ ]V-tatsu (stand)]V 124 0 0
[[ ]V-kirau (dislike)]V 106 0 0
[[ ]V-wameku (yell)]V 104 0 0
[[ ]V-kanashimu (feel.sad)]V 92 0 0

25
 [[ ]V-shitagau (obey)]V 90 0 0
[[ ]V-miru (look)]V 82 0 0
[[ ]V-ononoku (shudder)]V 67 0 0
[[ ]V-tagiru (boil)]V 61 0 0
[[ ]V-kuu (eat)]V 43 0 0
Average (P): 0

Table 28.2 Productivity of augmentation compound verbs

[[ ]V-V2]V Total token A word that Productivity
frequency appears only (P)
(N) once (n1)
[[ ]V-komu (go.into)]V 64293 6 0.00009332
[[ ]V-dasu (take.out)]V 56735 3 0.00005287
[[ ]V-ageru (raise)]V 38717 3 0.00007748
[[ ]V-tsukeru (attach)]V 33861 2 0.00005906
[[ ]V-kakeru (put.on)]V 24213 1 0.00004130
[[ ]V-tsuku (stick.to)]V 22587 1 0.00004427
[[ ]V-agaru (go.up)]V 18407 3 0.00016298
[[ ]V-toru (get/remove)]V 17303 1 0.00005779
[[ ]V-ireru (put.in)]V 12246 3 0.00024497
[[ ]V-awaseru (combine)]V 9180 1 0.00010893
[[ ]V-tateru (set.up)]V 6756 0 0
[[ ]V-tatsu (stand)]V 6024 0 0
[[ ]V-deru (go.out)]V 5635 0 0
[[ ]V-kakaru (hang.on)]V 5594 1 0.00017876
[[ ]V-mawaru (go.round)]V 4123 2 0.00048508
[[ ]V-saru (leave)]V 3913 1 0.00025555
[[ ]V-mawasu (turn)]V 3856 0 0
[[ ]V-wakeru (separate)]V 3113 3 0.00096370
[[ ]V-ochiru (fall)]V 2432 1 0.00041118
[[ ]V-otosu (drop)]V 2335 0 0
Average (P): 0.00016

Table 28.3 Productivity of adaptation compound verbs

[[ ]V-V2]V Total token A word that Productivity
frequency appears only (P)
(N) once (n1)

26
 [[ ]V-au (do.each.other)]V 21191 352 0.01661082
[[ ]V-hajimeru (start)]V 18077 950 0.05255296
[[ ]V-uru (can.do)]V 13712 547 0.03989206
[[ ]V-tsuzukeru (continue)]V 13415 721 0.05374580
[[ ]V-dasu (start)]V 12529 408 0.03256445
[[ ]V-naosu (do.again)]V 8398 227 0.02703024
[[ ]V-kiru (finish)]V 8378 359 0.04285032
[[ ]V-sugiru (overdo)]V 7511 489 0.06510451
[[ ]V-kaneru (hesitate.to.do)]V 3422 285 0.08328462
[[ ]V-kakeru (begin.to.do)]V 2614 343 0.13121652
[[ ]V-tsukusu (exhaust)]V 2134 107 0.05014058
[[ ]V-owaru (finish)]V 1887 183 0.09697933
[[ ]V-makuru (do.to.excess)]V 1805 253 0.14016620
[[ ]V-nuku (do.to.completion)]V 1333 64 0.04801200
[[ ]V-oeru (finish)]V 1159 125 0.10785159
[[ ]V-hateru (be.exhausted)]V 1054 40 0.03795066
[[ ]V-nareru (be.used.to)]V 1041 23 0.02209414
[[ ]V-kakaru (almost …)]V 868 121 0.13940092
[[ ]V-wasureru (forget.to)]V 764 75 0.09816753
[[ ]V-toosu (do.to.completion)]V 516 52 0.10077519
Average (P): 0.06931

To explain these tables, we will take a closer look at each type. First, let us
take [[ ]V-matou (put.on)]V in Table 28.1 (the coordination type) as an example to
explain how its productivity P is calculated. As shown in Table 28.4, there is only one
instantiation of [[ ]V-matou (put.on)]V, which is tsuki-matou (attach-put.on) ‘follow
around’. Since tsuki-matou appears more than once in the corpus (token frequency =
539), the n1 of [[ ]V-matou (put.on)]V is zero. Therefore, the productivity of [[ ]V-
matou (put.on)]V is zero (P = n1/N = 0/539 = 0).

Table 28.4 The productivity of the coordination compound [[ ]V-matou (put.on)]V

[[ ]V-matou (put.on)]V Token frequency
tsuki-matou (attach-put.on) 539

Total: 539 (N)

In the same way, since n1 for all other examples of [[ ]V-V2]V in Table 28.1 is zero,
their productivity is also zero.
27
 Next, let us look at the example of [[ ]V-komu (go.into)]V in Table 28.2 (the
augmentation type). As Table 28.5 shows, there are a number of [[ ]V-komu
(go.into)]V (examples are listed in order of token frequency, and some examples have
been omitted for the sake of space). Among them, there are six words that have appeared
only once (n1 = 6). Thus, the productivity P of [[ ]V-komu (go.into)]V is 0.00009332
(P = n1/N = 6/64293).

Table 28.5 The productivity of the augmentation compound [[ ]V-komu (go.into)]V

[[ ]V-komu (go.into)]V Token frequency
mochi-komu (carry-go.into) 2136
tsuki-komu (push-go.into) 2125
… …
(omitted) (omitted)
… …
ore-komu (break-go.into) 1
eguri-komu (gouge-go.into) 1
taoshi-komu (tapple-go.into) 1
kare-komu (wither-go.into) 1
kurai-komu (eat-go.into) 1
naki-komu (cry-go.into) 1 n1 (underlined) = 6
Total: 64293 (N)

The productivity of other [[ ]V-V2]V in Table 28.2 varies from a maximum of

0.00096370 ([[ ]V-wakeru (separate)]V) to zero ([[ ]V-mawasu (turn)]V, [[ ]V-otosu
(drop)]V, etc.), depending on the semantic restrictions of the V2’s conceptual slot.
Finally, consider the example of [[ ]V-au (do.each.other)]V in Table 28.3 (the
adaptation type). As you can see in Table 28.6, there are a total of 352 words that appear
only once. Accordingly, the productivity of [[ ]V-au (do.each.other)]V is 0.01661082
(P = n1/N = 352/21191). The productivity of the other [[ ]V-V2]V in Table 28.3 is also
high overall.

Table 28.6 The productivity of the adaptation compound [[ ]V-au (do.each.other)]V

[[ ]V-au (do.each.other)]V Token frequency

28
 tsuki-au (attach-each.other) 5605
hanashi-au (talk-each.other) 2624
… …
(omitted) (omitted)
… …
ikari-au (be.angry-each.other) 1
watashi-au (hand.over-each.other) 1
… …
(omitted) (omitted)
… …
tsubuyaki-au (murmur-each.other) 1 n1 (underlined) = 352
Total: 21191 (N)

To sum up, as hypothesized on the basis of B/E organization, the coordination

type is not productive at all, while the augmentation type is quite low but not totally
unproductive (the average P = 0.00016). The adaptation type was found to have a
relatively high productivity (the average P = 0.06931). These differences in
productivity found in Japanese compound verbs support the validity of this study’s
classification. 16

3.4 The dependency and manual nature of adaptation

Classification by B/E organization can be used to identify the dependencies specific to
the elaboration of the adaptation type. As mentioned in Section 3.2, the adaptation type
often uses affixes to represent elaboration. Because elaboration depends on the baseline,
we can see an iconic relationship with the affixes that rely on the base word.
The uniqueness of the adaptation type can also be observed in the processes of
“grammaticalization”. In general, grammaticalization, the coming into being of
grammatical markers, such as for case, tense, aspect, and modality, is the creation of
grammatical categories (Lehmann, 2004: 183). In adaptation compound verbs, only
V2s of adaptation compound verbs representing elaboration form a closed class, mostly
with grammaticalized verbs that initially express concrete actions, especially manual
operations.

(12) Grammaticalized V2 in adaptation compound verbs

16
It should be noted that this comparison is limited to the case where the three types coexist in the
same linguistic form. If one type is completely or almost completely absent, there may be other factors
at play.
29
 a. khā-un ghe-ṇe [Marathi, Kageyama et al., 2013: 7]
eat-CP take-INF
‘eat up’
b. khā baiṭh-nā [Hindi, Kageyama et al., 2013: 7]
eat sit-INF
‘eat [by mistake]’
c. lel jom-me [Munda, Hook, 1991: 190]
see eat-IMP
‘see [for one’s benefit or comfort]’
d. kuppi poṭṭI pooyi [Malayalam, Abbi & Gopalakrishnan, 1991: 687]
bottle break-CP go-PST
‘(the) bottle broke (undesirably)’
e. uyu-y-u ver- [Turkish, Kuribayashi, 2013: 289]
sleep-GER give
‘sleep immediately’
f. č’wa-n rex-ana [Avar, Yamada, 2013: 10]
kill-PRF throw-PST
‘killed (emphasize the completeness)’
g. iwaːmãk-ā-chwaye [Newar, Kiryu, 2008]
forget-CP-send
‘completely forget’
h. ilk-e noh-ass-ta [Korean, Suh, 2000: 78]
read-CP put-PST-DECL
‘read (in anticipation of …)’
i. yari-nuku やり抜く [Japanese]
do-pull
‘do thoroughly’

For instance, there are thirty V2s in Japanese adaptation compound verbs, and
eight of them originally represent actions using hands by default.

Table 28.7 V2s in Japanese adaptation compounds that originally represent manual
operation
V2s originally represent Original meanings Breached meanings in
manual operation adaptation compound verbs
dasu take out start
naosu fix do again

30
 kiru cut finish
kakeru hang be about to do
makuru roll up do to excess
nuku pull do thoroughly
toosu put through do thoroughly
tsukeru put on be used to do

Although this is a process of conceptual manipulation of elaboration, which is not a

concrete action, we can see that a fairly high percentage (26.7%) of the verbs is
originally used to describe hand motions. This sort of bleaching of meaning is due to
grammaticalization; for example, dasu has changed its meaning from physically putting
an object out into making a dynamic process appear (initiation) (see Ohori, 1992: 210-
232).
Furthermore, among the V2s of adaptation compound verbs, as shown in (13),
transitive verbs are used for expressions without an object. In these cases, although the
meanings of V2s are bleached, their causative meanings are in an iconic relation with
and operating on a baseline (see Chen, 2020 for more details).

(13) a. 太郎は 怯え切っている. [Japanese]

Taro wa obie-kit-te-iru
Taro TOP be.frightened-cut-NPST
‘Taro is extremely frightened.’
b. 次郎は がんばり抜いた. [Japanese]
Jiro wa ganbari-nui-ta.
Jiro TOP work.hard-pull-PST
‘Jiro worked hard thoroughly.’

The findings in Section 3 (summarized in Table 28.8) suggest that (a)

headedness, (b) order in word formation, (c) productivity, and (d) the representation of
the elaboration in adaptation compound verbs are all related to the quasi‐spatiotemporal
relationship in B/E organization.

Table 28.8 The parallelism between the quasi‐spatiotemporal relationship in B/E

organization and linguistic properties
The quasi‐spatiotemporal Linguistic properties
relationship in B/E
organization

31
a. Headedness In the coordination type, there In the coordination type, the
is no difference in importance, constituents are semantically
as it is simply the juxtaposition equivalent, so both are
of two similar or opposite semantic heads.
things. In the augmentation type, the
In the augmentation type, the baseline is the semantic head,
baseline concept is more since a complex word X–Y
important in that it is the (X: additive, Y: baseline) is
foundation of the additive. basically a subtype of Y.
In the adaptation type, the In the adaptation type, what
baseline is more important than represents the baseline B is
the elaboration because the basically the semantic head.
elaboration only reshapes the However, when the baseline
baseline. concept and the concept after
elaboration differ
significantly, it is difficult to
identify a specific part as the
semantic head.
b. The order in The temporal order in the B/E The order of verbs in
word formation organization of adaptation is adaptation compound verbs
“baseline → elaboration”. of OV languages tends to be
“V1: baseline, V2:
elaboration”.
In the case of Japanese
compound verbs, these three
types need to follow a
specific order of occurrence,
[[[coordination]-
augmentation]-adaptation].
The elaboration of adaptation
is more often expressed by
suffixes than by prefixes.
c. Productivity In the coordination type, the In Japanese compound verbs,
number of similar concepts is the coordination type is not
quite limited, and when it productive at all, while the
comes to opposite concepts, augmentation type is quite
the relationship is almost one- low but not totally

32
 to-one, which means that the unproductive. The adaptation
pattern of combinations is type was found to have a
fairly small. relatively high productivity.
An additive is something
embedded in an empty slot of a
baseline in the augmentation
type. The additive and the
baseline are mutually
restricted, which is likely to
cause augmentation
compounds to show limited
productivity.
In adaptation compounds,
elaboration is an independent
operation performed on its
baseline. Thus, the elaboration
should be applicable to a
variety of baselines.
d. The Elaborations are a certain kind V2s of adaptation
representation of of operation performed on compounds, which represent
the elaboration in baselines. the elaboration, form a closed
adaptation class.
compound verbs Many verbs originally
representing hand movements
are used for V2s.
In some cases, transitive
verbs are used for expressions
without an object.

4. Conclusion

The main goal of this chapter is to show the utility of B/E organization in revealing the
motivation in the word formation process. By examining various phenomena in word
formation, such as headedness, order in word formation, productivity, and dependency,
this study has shown that the linguistic properties of a complex word are motivated by
its conceptual structure and suggests that the quasi-spatiotemporal relationship in B/E
organization can serve as an essential criterion for determining the behavior of a
complex word. In addition, the new classification of elaboration proposed in this study
33
complements the original baseline elaboration theory. These findings contribute in
several ways to our understanding of the role of meaning in word formation and provide
a basis for verifying the iconic relationship between conceptual and linguistic structures.
The empirical evidence from this study also suggests that the way a concept is created
motivates the structure and meaning of that concept.
The Cognitive Linguistics enterprise has successfully demonstrated the non-
arbitrariness of the form–meaning (signifier–signified) relationship across various areas,
such as sound symbolism, gestures, sign languages, and iconic principles based on real-
world experiences (e.g., the principle of temporal sequence). The structured mapping
in word formation proposed in this study suggests that the differences in the quasi‐
spatiotemporal relationship in B/E organizations emerge iconically as explicit
differences in linguistic forms, indicating the validity of the onomasiological approach
based on B/E organization.
This chapter focuses on the differences between different types of elaboration
in word formation. A natural progression of this approach is the analysis of complex
sentences, which can be seen as creating new concepts, such as word formation,
although the size of the material varies. It would also be interesting to apply the baseline
elaboration theory to the realm of discourse analysis in the future.
Finally, the greatest challenge facing this theory is to demonstrate its
psychological reality. Thus, further experimental investigations are needed to examine
the process of baseline elaboration and the differences between each type of elaboration.

References

Abbi, Anvita and Devi Gopalakrishnan. 1991. Semantics of explicator compound verbs
in south Asian languages. Language Sciences 13: 161-180.
Anderson, Stephen R. 1992. A-morphous Morphology. Cambridge: Cambridge
University Press.
Asao, Yoshihiko. 2007. Hukugō-go no seisansei to bunpō-teki seisitu. [Productivity and
grammatical properties of compounds]. Papers from the The 134th meeting of the
Linguistic Society of Japan: 416-421.
Baayen, Harald. 1992. Quantitative aspects of morphological productivity. In G. Booij
and J. van Marle (eds.), Yearbook of Morphology 1991, 109-149. Dordrecht:
Kluwer.
Bagasheva, Alexandra. 2020. Paradigmaticity in compounding. In J. Fernández-
Domínguez, A. Bagasheva and C. Lara-Clares (eds.), Paradigmatic Relations in
Word Formation, 21-48. Leiden: Brill.
Baldinger, Kurt. 1980. Semantic Theory: Towards a Modern Semantics. Oxford:
34
 Blackwell.
Bauer, Laurie. 2001. Compounding. In M. Haspelmath, E. König, W. Oesterreicher and
W. Raible (eds.), Language Typology and Language Universals, 695-707. Berlin:
De Gruyter.
Benczes, Réka. 2006. Creative Compounding in English: The Semantics of
Metaphorical and Metonymical Noun–Noun Combinations. Amsterdam: John
Benjamins.
Booij, Geert. 2005a. Compounding and derivation: Evidence for Construction
Morphology. In W. U. Dressler, F. Rainer, D. Kastovsky and O. Pfeiffer (eds.),
Morphology and its Demarcations, 109-132. Amsterdam: John Benjamins.
Booij, Geert. 2005b. The Grammar of Words. An Introduction to Linguistic Morphology
(1st ed.). Oxford: Oxford University Press.
Booij, Geert. 2006. Inflection and derivation. In K. Brown (ed.), Encyclopedia of
Language and Linguistics, 654-666. (2nd ed.). Oxford: Elsevier.
Booij, Geert. 2009. Lexical integrity as a morphological universal, a constructionist
view. In S. Scalise, E. Magni and A. Bisetto (eds.), Universals of Language Today,
83-100. Berlin: Springer.
Booij, Geert. 2010. Construction Morphology. Oxford: Oxford University Press.
Booij, Geert. 2013. Morphology in CxG. In T. Hoffmann and G. Trousdale (eds.), The
Oxford Handbook of Construction Grammar, 255-273. Oxford: Oxford University
Press.
Butt, Miriam. 1993. The Structure of Complex Predicates in Urdu. PhD dissertation,
Stanford University.
Bybee, Joan. 1985. Morphology: A Study of the Relation between Meaning and Form.
Amsterdam: John Benjamins.
Bybee, Joan L., William Pagliuca and Revere D. Perkins. 1990. On the asymmetries in
the affixation of grammatical material. In W. Croft, K. Denning and S. Kemmer
(eds.), Studies in Typology and Diachrony: Papers Presented to Joseph H.
Greenberg on his 75th Birthday, 1-42. Amsterdam: John Benjamins.
Ceccagno, Antonella and Bianca Basciano. 2009. Sino-Tibetan: Mandarin Chinese. In
R. Lieber and P. Štekauer (eds.), The Oxford Handbook of Compounding, 478-490.
Oxford: Oxford University Press.
Chen, Yiting. 2015. Nihongo no Goiteki-fukugō-dōshi no keisei-mekanizumu ni tsuite:
Cyugokugo tono hikaku-taisyo to awasete [The Formation of Japanese Lexical
Compound Verbs: A Comparison with Chinese]. PhD dissertation, Kobe University.
Chen, Yiting. 2016. A frame-semantic approach to verb-verb compound verbs in
Japanese: A case study of –toru. BLS 39: 16-30.
Chen, Yiting. 2017. Kitei to seichika kara mita fukugō-go no bunrui: Nihongo fukugō-
35
 dōshi o chushin ni [The classification of compounds in baseline/elaboration theory:
Focusing on Japanese compound verbs]. NINJAL Research Papers 13: 25-50.
Chen, Yiting. 2020. Macro-events in verb–verb compounds from the perspective of
baseline and elaboration: Iconicity in typology and grammaticalization. Cognitive
Semantics 6: 1-28.
Chen, Yiting and Yo Matsumoto. 2018. Goiteki-fukugō-dōshi no imi to taikei:
Konsutorakushon-keitairon to furēmu-imiron [The Semantics and Organization of
Japanese Lexical Compound Verbs: Construction Morphology and Frame
Semantics]. Tokyo: Hituzi Syobo.
Cheng, Lisa.-S.L. and Huang, C.-T. James. 1994. On the argument structure of
resultative compounds, in honor of William S.-Y. Wang. In M. Chen and O. Tzeng
(eds.), Interdisciplinary Studies on Language and Language Change, 187-221.
Taipei: Pyramid.
Coulson, Seana. 2001. Semantic Leaps: Frame-Shifting and Conceptual Blending in
Meaning Construction. Cambridge: Cambridge University Press.
Croft, William. 2003. Typology and Universals (2nd ed.). Cambridge: Cambridge
University Press.
Di Sciullo, Anna-Maria and Edwin Williams. 1987. On the Definition of Word.
Cambridge, MA: MIT Press.
Dirven, René. 1999. Conversion as a conceptual metonymy of event schemata. In G.
Radden and K.-U. Panther (eds.), Metonymy in Language and Thought, 275−287.
Amsterdam: John Benjamins.
Dryer, Matthew S. 2011. The evidence for word order correlations. Linguistic Typology
15: 335-380.
Emmorey, Karen. 2014. Iconicity as structure mapping. Philosophical Transactions of
the Royal Society B: Biological Sciences 369 (1651).
Fábregas, Antonio and Sergio Scalise. 2012. Morphology: From Data to Theories.
Edinburgh: Edinburgh University Press.
Fauconnier, Gilles and Mark Turner. 2002. The Way We Think: Conceptual Blending
and the Mind’s Hidden Complexities. New York: Basic Books.
Gentner, Dedre. 1983. Structure-mapping: A theoretical framework for analogy.
Cognitive Science 7: 155-170.
Greenberg, Joseph. 1963. Some universals of grammar with particular reference to the
order of meaningful elements. In J. Greenberg (ed.), Universals of language, 73-
113. Cambridge, MA: MIT Press.
Haiman, John. 1980. The iconicity of grammar: Isomorphism and motivation.
Language 56: 515-540.
Haiman, John. 1983. Iconic and economic motivation. Language 59: 781-819.
36
Haiman, John. 1985. Natural Syntax: Iconicity and Erosion. Cambridge: Cambridge
University Press.
Haspelmath, Martin and Andrea D. Sims. 2013. Understanding morphology (2nd ed.).
London and New York: Routledge
Hawkins, John A. and Anne Cutler. 1988. Psycholinguistic factors in morphological
asymmetry. In J. A. Hawkins (ed.), Explaining Language Universals, 280-317.
Oxford: Blackwell.
Her, One-Soon. 2007. Argument-function mismatches in Mandarin resultatives: A
lexical mapping account. Lingua 117: 221-246.
Himmelmann, Nikolaus P. 2014. Asymmetries in the prosodic phrasing of function
words: Another look at the suffixing preference. Language 90: 927-960.
Hiraga, Masako. 2005. Metaphor and Iconicity: A Cognitive Approach to Analyzing
Texts. Houndmills: Palgrave Macmillan.
Hockett, Charles F. 1954. Two models of grammatical description. Word 10: 210-234.
Hook, Peter E. 1991. The compound verb in Munda: An areal and typological overview.
Language Sciences 13: 181-195.
Hook, Peter E. 2013. A descriptive account of South Asian verbal extension: Inferring
the semantic conditions on the alternation of Vx with V. Paper presented at NINJAL
International Symposium 2013: Mysteries of Verb-verb Complexes in Asian
Languages.
Hummel, Martin. 2018. Baseline elaboration and echo-sounding at the adjective adverb
interface. Cognitive Linguistics 29: 407-452.
Jackendoff, Ray. 1975. Morphological and semantic regularities in the lexicon.
Language 51: 639-71.
Jackendoff, Ray and Jenny Audring. 2020. The Texture of the Lexicon: Relational
Morphology and the Parallel Architecture. Oxford: Oxford University Press.
Kageyama, Taro. 1982. Word formation in Japanese. Lingua 57: 215-258.
Kageyama, Taro. 1993. Bunpō to go keisei [Grammar and Word Formation]. Tokyo:
Hituzi Syobo.
Kageyama, Taro. 2016. Verb-compounding and verb-incorporation. In T. Kageyama
and H. Kishimoto (eds.), Handbook of Japanese Lexicon and Word Formation,
273-310. Berlin: Mouton de Gruyter.
Kageyama, Taro, Prashant Pardeshi and Peter E. Hook. 2013. The position paper of
NINJAL International Symposium 2013: Mysteries of Verb-verb Complexes in
Asian Languages.
Kageyama, Taro and Michiaki Saito. 2016. Vocabulary strata and word formation
processes. In T. Kageyama and H. Kishimoto (eds.), Handbook of Japanese
Lexicon and Word Formation, 11-50. Berlin: Mouton de Gruyter.
37
Kiparsky, Paul. 1982. Lexical phonology and morphology. In Linguistic Society of
Korea (ed.), Linguistics in the Morning Calm, 3-91. Seoul: Hanshin.
Kiparsky, Paul. 1985. Some consequences of lexical phonology. Phonology Yearbook
2: 85-138.
Kiryu, Kazuyuki. 2008. Compound verbs in Newar: Aspect and modality. Paper
presented at Workshop on Compound Verbs in Indo‐Turanian Linguistic Area.
Kobayashi, Hideki, Kiyo Yamashita and Taro Kageyama. 2016. Sino-Japanese words.
In T. Kageyama and H. Kishimoto (eds.), Handbook of Japanese Lexicon and Word
Formation, 93-131. Berlin: Mouton de Gruyter.
Koch, Peter. 1999. Frame and contiguity: On the cognitive bases of metonymy and
certain types of word formation. In K.-U. Panther and G. Radden (eds.), Metonymy
in Language and Thought, 139-167. Amsterdam: John Benjamins.
Kuribayashi, Yu. 2013. V + V gata fukugōdōshi to gokeisei: Torukogo kara mita
nihongo. [V–V compound verbs and word formation: Japanese from the
perspective of Turkish] In T. Kageyama (ed.), Fukugōdōshi-kenkyu no saisentan
[New Exploration into the Mysteries of Compound Verbs], 273-300. Tokyo: Hituzi
Syobo.
Labrune, Laurence. 2006. Patterns of phonemic preferences in Japanese non-headed
binary compounds: What waa-puro, are-kore and mecha-kucha have in common.
Gengo Kenkyu 129: 3-41.
Lamberty, Angela and Hans-Jörg Schmid. 2013. Verbal compounding in English: A
challenge for usage-based models of word-formation? Anglia 131: 591-626.
Langacker, Ronald W. 1993. Reference-point constructions. Cognitive Linguistics 4: 1-
38.
Langacker, Ronald W. 2015a. Baseline and elaboration. Handout of Plenary lecture.
International Cognitive Linguistics Conference. Northumbria University. July 24,
2015.
Langacker, Ronald W. 2015b. How to build an English clause. Journal of Foreign
Language Teaching and Applied Linguistics 2:1-45.
Langacker, Ronald W. 2016a. Baseline and elaboration. Cognitive Linguistics 27: 405-
439.
Langacker, Ronald W. 2016b. Nominal grounding and English quantifiers. Cognitive
Linguistic Studies 3: 1-31.
Langacker, Ronald W. 2017. Evidentiality in Cognitive Grammar. In J. I. Marín-Arrese,
G. Haßler and M. Carretero (eds.), Evidentiality Revisited. Cognitive Grammar,
Functional and Discourse-Pragmatic Perspectives, 13-55. Amsterdam: John
Benjamins.
Langacker, Ronald W. 2019. Levels of reality. Languages 4, 22.
38
Lee, Leslie and Farrell Ackerman. 2011. Mandarin resultative compounds: A family of
lexical constructions. In M. Butt and T. Holloway King (eds.), Proceedings of the
LFG11 Conference. Hong Kong: University of Hong Kong.
Lehmann, Christian. 2004. Theory and method in grammaticalization. Zeitschrift für
Germanistische Linguistik 32: 152-187.
Li, Yafei 1990. On V-V compounds in Chinese. Natural Language and Linguistic
Theory 8: 177-207.
Lieber, Rochelle. 1981. On the Organization of the Lexicon. Indiana: Indiana University
Linguistics Club [reprinted 1991: New York: Garland].
Lipski, John M. 2014. Syncretic discourse markers in Kichwa-influenced Spanish:
transfer vs. emergence. Lingua 151: 216-239.
Matsumoto, Yo. 1996. Complex Predicates in Japanese: A Syntactic and Semantic
Study of the Notion 'Word’. Stanford: CSLI Publications and Tokyo: Kurosio
Publishers.
McCarthy, John J. 1981. A prosodic theory of nonconcatenative morphology. Linguistic
Inquiry 12: 373-418.
Müller, Peter O., Ingeborg Ohnheiser, Susan Olsen and Franz Rainer, eds. 2015. Volume
1 Word-Formation: An International Handbook of the Languages of Europe.
Berlin: Mouton de Gruyter.
Namiki, Takayasu and Taro Kageyama. 2016. Word structure and headedness. In T.
Kageyama and H. Kishimoto (eds.), Handbook of Japanese Lexicon and Word
Formation, 201-235. Berlin: Mouton de Gruyter.
Ohori, Toshio. 1992. Diachrony in clause linkage and related issues. PhD dissertation,
University of California at Berkeley.
Onysko, Alexander and Sascha Michel, eds. 2010. Cognitive Perspectives on Word
Formation. Berlin: Mouton de Gruyter.
Packard, Jerome L. 2000. The Morphology of Chinese: A Linguistic and Cognitive
Approach. Cambridge: Cambridge University Press,
Panther, Klaus-Uwe and Günter Radden. 2011. Introduction: Reflections on motivation
revisited. In K.-U. Panther and G. Radden (eds.), Motivation in Grammar and the
Lexicon, 1-26. Amsterdam: John Benjamins.
Perlmutter, David. 1988. The split morphology hypothesis: Evidence from Yiddish. In
M. Ham-mond and M. Noonan (eds.), Theoretical Morphology: Approaches in
Modern Linguistics, 79-100. San Diego: Academic.
Sapir, Edward. 1921. Language. New York: Harcourt, Brace and World.
Scalise, Sergio and Antonietta Bisetto 2009. The classification of compounds. In R.
Lieber and P. Stekauer (eds.), The Oxford Handbook of Compounding, 34-53.
Oxford: Oxford University Press.
39
Scalise, Sergio and Emiliano Guevara. 2005. The lexicalist approach to word-formation
and the notion of the lexicon. In P. Štekauer and R. Lieber (eds.), Handbook of
Word-Formation, 147-187. Dordrecht: Springer.
Scalise, Sergio, Antonio Fábregas and Francesca Forza. 2009. Exocentricity in
compounding. Gengo Kenkyu 135: 49-83.
Selkirk, Elizabeth. 1982. The Syntax of Words. Cambridge, MA: MIT Press.
Smeets, Ineke. 2007. A Grammar of Mapuche. Berlin: Mouton de Gruyter.
Spencer, Andrew. 1991. Morphological Theory: An Introduction to Word Structure in
Generative Grammar. Oxford: Blackwell.
Spencer, Andrew. 1998. Morphophonological operations. In A. Spencer and A. Zwicky
(eds.), Handbook of Morphology, 123-143. Oxford: Blackwell.
Štekauer, Pavol. 1998. An Onomasiological Theory of English Word-Formation.
Amsterdam: John Benjamins.
Suh, Yongwhan. 2000. Study of the Auxiliary Verb Construction and Verb Serialization
in Korean. Seoul: Hankuk.
Tai, James H-Y. 1985. Temporal sequence and Chinese word order. In J. Haiman (ed.),
Iconicity in Syntax, 49-72. Amsterdam: John Benjamins.
Talmy, Leonard. 2000. Toward a Cognitive Semantics. Volume II: Typology and Process
in Concept Structuring. Cambridge, MA: MIT Press.
Taylor, John R. 2015. Introduction. In J. R. Taylor (ed.), The Oxford Handbook of the
Word, 1-20. Oxford: Oxford University Press.
Tuggy, David. 2005. Cognitive approach to word-formation. In: P. Štekauer and R.
Lieber (eds.), Handbook of Word-Formation, 233-265. Dordrecht: Springer.
Umbreit, Birgit. 2011. Motivational networks: An empirically supported cognitive
phenomenon. In: K.-U. Panther and G. Radden (eds.), 269-286. Motivation in
Grammar and the Lexicon. Amsterdam: John Benjamins.
Ungerer, Friedrich. 1999. Diagrammatic iconicity in word-formation. In M. Nänny and
O. Fischer (eds.), Form Miming Meaning, 307-324. Amsterdam: John Benjamins.
Ungerer, Friedrich. 2002. The conceptual function of derivational word-formation in
English. Anglia 120: 534-567.
Ungerer, Friedrich. 2007. Word-formation. In D. Geeraerts and H. Cuyckens (eds.), The
Oxford Handbook of Cognitive Linguistics, 650-675. Oxford: Oxford University
Press.
Van der Spuy, Andrew. 2020. English plurals in Construction Morphology. Language
Sciences 77: 101240.
Van Langendonck, Willy. 1995. Categories of word order iconicity. In M. E. Landsberg
(ed.), Syntactic Iconicity and Linguistic Freezes: The Human Dimension, 79-90.
Berlin: Mouton de Gruyter.
40
Van Langendonck, Willy. 2007. Iconicity. In D. Geeraerts and H. Cuyckens (eds.), The
Oxford Handbook of Cognitive Linguistics, 394-418. Oxford: Oxford University
Press.
Vance, Timothy J. 2008. The Sounds of Japanese. Cambridge: Cambridge University
Press.
Wilcox, Sherman and Rocío Martínez. 2020. The conceptualization of space: Places in
signed language discourse. Frontiers in Psychology 11: 1406.
Williams, Edwin. 1981. On the notions ‘lexically related’ and ‘head of a word’.
Linguistic Inquiry 12: 245-274.
Yamada, Hisanari. 2013. V1 (perfective converb) + V2 compound verbs in Standard
Avar. Paper presented at NINJAL International Symposium 2013: Mysteries of
Verb-verb Complexes in Asian Languages.
Zwicky, Arnold. 1985. Heads. Journal of Linguistics 21: 1-29.

41