COLLOCATIONS AND TEACHING

Investigating word combinations in two English textbooks for

Norwegian upper secondary school students

Lívia Cerqueira de Souza Hodne

A thesis submitted for the Master‟s Degree program in English

Department of Foreign Languages
University of Bergen
Autumn 2009
ACKNOWLEDGMENTS

First of all, I would like to thank my supervisor, Professor Aud Solbjørg Skulstad, for her
support, guidance, and for valuable advice. I also thank my co-supervisor, PHD candidate
Gard Buen Jenset, for all the constructive feedback, practical tips, and help with corpus work.

I also owe much of what I know today as a teacher and as a student of English to Susan Harris
de Mello, first my teacher, then my boss, and always a friend. She was the one who first
introduced me to the phenomena of collocations and showed me their importance in English
teaching and learning. I am grateful to my students for being patient during my busy weeks
and for being positive to my new „experiments‟, especially when I wanted to try out the ideas
of the present study in my classes.

I thank my dear family and friends in Brazil, who have accompanied me and encouraged me
during the writing of this thesis even from a long distance. I would also like to thank my
Norwegian family and friends who have followed closely each step of the process. I wish to
extend a special thank you to my father, who has always shown me unconditional love and
has taught me everything I know.

Finally, I thank my husband, Arnstein, for his love and patience through the unsociable
weekends and extra housework, and for helping me „take a step back and think objectively‟
when I most needed it.

iii
iv
SUMMARY IN NORWEGIAN

Denne masteroppgaven omhandler undervisning av kollokasjoner i videregående skole i

Norge. Kollokasjoner er anerkjent som et viktig fenomen i det engelske språket, og derfor har
mer og mer oppmerksomhet blitt gitt til temaet i undervisningssammenheng. Til tross for
dette virker det som om lærebøkene i Norge enda ikke har tilpasset forskning om korpus og
kollokasjon til klasserommet.
Studien omfatter en undersøkelse av vokabularøvelser i to lærebøker som er brukt i
engelsk på videregående skoler i Norge (studieforberedende utdanningsprogram). Det første
målet er å undersøke om ord representert som isolerte ord i vokabularøvelsene er en del av
kollokasjonene i tekstene de er hentet fra. To forskningsspørsmål har blitt formulert: Hvor
mange av disse ordene dukker opp i kollokasjoner, og hvilke av disse kollokasjonene ville
være nyttige for denne elevengruppen? Det andre målet er å komme med forslag til hvordan
vokabularøvelser kan bli forbedret basert på Kunnskapsløftet og tidligere forskning. Det
tredje målet er å foreslå andre måter hvordan kollokasjoner kan bli integrert i engelsk
undervisningen.
Denne studien bruker et korpus og en kollokasjons-ordbok for å avgjøre om
ordkombinasjonene funnet i tekstene er kollokasjoner. Korpusanalyse er også brukt som en
metode for å se etter frekvensen og styrken på kollokasjoner, og for å finne kollokatorer av
ord fra øvelsene.
Resultater viser at mer enn en tredjedel av ordene fra de valgte vokabularøvelsene er
en del av kollokasjoner i tekstene de er hentet fra, og at de fleste av disse er nyttige for
undervisningen. Til slutt har eksisterende kollokasjons-øvelser og forskning på
fremmedspråklæring og korpuslingvistikk blitt brukt for å legge fram forslag til hvordan
øvelsene kan bli forbedret og komplimentert. For å oppsummere, viser studien at
kollokasjoner opptrer hyppig i skolebøker og de formidler relevant sosiokulturell informasjon.
Den viser også at korpus kan brukes til å forbedre språklige øvelser, og at det både er mulig
og hensiktsmessig å integrere kollokasjoner i språkundervisningen.

v
TABLE OF CONTENTS

Page

Acknowledgements…………………………………………………………….. iii
Summary in Norwegian………………………………………………………… v
Table of contents……………………………………………………………….. vi
List of tables …………………………………………………………………… ix
List of figures…………………………………………………………………... xi
List of abbreviations……………………………………………………………. xi

1 INTRODUCTION…………………………………………………………... 1
1.1 Relevance of the present study………………………………………… 1
1.2 Aims and scope………………………………………………………… 2
1.3 Methods………………………………………………………………… 2
1.4 Why collocations? ……………………………………………………... 3
1.5 What is a collocation? …………………………………………………. 4
1.6 Terminology……………………………………………………………. 8
1.7 Structure………………………………………………………………... 8

2 BACKGROUND…………………………………………………………….. 9
2.1 Introduction……………………………………………………………. 9
2.2 Teaching methods and the role of vocabulary………………………….. 9
2.3 Foreign language vocabulary acquisition………………………………. 12
2.4 Vocabulary in Norwegian curricula……………………………………. 15
2.5 Collocations……………………………………………………………. 17
2.5.1 Background…………………………………………………….. 17
2.5.2 Cognitive linguistics, chunking, and memory………………….. 18
2.5.3 Language teaching and learning………………………………... 19
2.5.4 Communicative language teaching and the Framework……….. 21
2.5.5 Culture and communication in word combinations……………. 25
2.5.6 Collocation dictionaries………………………………………… 27
2.6 Corpus linguistics………………………………………………………. 28
2.7 A summary: Vocabulary teaching, collocations, and corpora…………. 29

3 MATERIAL AND METHODS…………………………………………….. 31

3.1 Introduction…………………………………………………………….. 31
3.2 Material………………………………………………………………… 31
3.2.1 Criteria for selection of textbooks and exercises……………….. 31
3.2.2 @cross………………………………………………………….. 34
3.2.3 Passage…………………………………………………………. 36
3.2.4 The corpus……………………………………………………… 38
3.3 Methods………………………………………………………………… 40
3.3.1 The dictionary…………………………………………………... 41
3.3.2 MI scores and frequencies in COCA…………………………… 42
3.3.3 Description of corpus tools……………………………………... 44
3.3.4 Criteria for corpus search………………………………………. 46
3.4 Limitations of the material and methods……………………………….. 48

vi
4 RESULTS……………………………………………………………………. 51
4.1 Introduction…………………………………………………………….. 51
4.2 Presentation of the data………………………………………………… 51
4.3 Results from @cross…………………………………………………… 51
4.4 Results from Passage…………………………………………………... 70
4.5 Summary……………………………………………………………….. 86

5 DISCUSSION AND SUGGESTIONS FOR TEACHING………………... 89

5.1 Introduction…………………………………………………………….. 89
5.2 First aim: Research questions and hypothesis………………………….. 89
5.2.1 Discussion of @cross exercises………………………………… 90
5.2.2 Discussion of Passage exercises……………………………….. 92
5.2.3 Discussion of common features………………………………… 98
5.2.4 A summary: What kinds of collocations are useful for students? 104
5.3 Second aim: How the exercises can be improved……………………… 105
5.4 Third aim: Other ways of integrating collocations in language teaching 111
5.5 Summary……………………………………………………………….. 113

6 CONCLUSION……………………………………………………………… 115
6.1 Summary and conclusion………………………………………………. 115
6.2 Further research………………………………………………………… 116
6.3 Final remarks…………………………………………………………… 118

REFERENCES……………………………………………………………… 119

APPENDICES

vii
viii
LIST OF TABLES

Page

Table 2.1 Types of vocabulary learning strategies (Nation 2001:218)…………… 14

Table 2.2 An overview of the user/learner‟s competences in the Framework……. 22
Table 2.3 Classification of lexical elements compiled from the Framework……... 23
Table 2.4 Classification of lexical semantics compiled from the Framework……. 24

Table 3.1 Number and types of vocabulary exercises from each material………... 36
Table 3.2 Collocates of the adjective fancy in COCA…………………………….. 45
Table 3.3 Criteria for finding combinations in the corpus………………………... 49
Table 3.4 Types of combinations looked for in the corpus……………………….. 49
Table 3.5 Textbooks used in a selection of 60 Norwegian schools……………….. 50

Table 4.1 Collocations from exercise [1] in @cross……………………………… 55

Table 4.2 Collocates of the noun boundary in COCA……………………………. 55
Table 4.3 Collocations from exercise [2] in @cross……………………………… 56
Table 4.4 Collocates of the noun suburbs in COCA……………………………… 56
Table 4.5 Collocates of the adjective bright in COCA…………………………… 57
Table 4.6 Collocations from exercise [4] in @cross……………………………… 58
Table 4.7 Collocates of the noun disease in COCA………………………………. 58
Table 4.8 Collocations from exercise [5] in @cross……………………………… 60
Table 4.9 Collocates of the noun coverage in COCA…………………………….. 60
Table 4.10 Collocations from exercise [6] in @cross……………………………… 61
Table 4.11 Collocates of the verb pursue in COCA……………………………….. 62
Table 4.12 Collocations from exercise [7] in @cross……………………………… 62
Table 4.13 Collocates of the verb obtain in COCA………………………………... 63
Table 4.14 Collocations from exercise [8] in @cross……………………………… 63
Table 4.15 Collocates of the adverb environmentally in COCA…………………… 64
Table 4.16 Collocations from exercise [9] in @cross……………………………… 65
Table 4.17 Collocates of the noun trade in COCA………………………………… 65
Table 4.18 Collocations from exercise [10] in @cross…………………………….. 66
Table 4.19 Collocates of the noun prom in COCA………………………………… 67
Table 4.20 Collocations from exercise [11] in @cross…………………………….. 67
Table 4.21 Collocates of the noun vegetables in COCA…………………………… 68
Table 4.22 Collocations from exercise [12] in @cross…………………………….. 68
Table 4.23 Collocates of the adjective attractive in COCA………………………... 69
Table 4.24 Collocations from exercise [13] in @cross…………………………….. 70
Table 4.25 Collocates of the noun scholarship in COCA…………………………. 70
Table 4.26 Collocations from exercise [14] in @cross…………………………….. 71
Table 4.27 Collocates of the adjective constitutional in COCA…………………… 72
Table 4.28 Collocations from exercise [15] in Passage……………………………. 73
Table 4.29 Collocates of the noun slack in COCA………………………………… 73
Table 4.30 Collocations from exercise [16] in Passage……………………………. 74
Table 4.31 Collocates of the noun debate in COCA……………………………….. 74
Table 4.32 Collocations from exercise [17] in Passage……………………………. 75
Table 4.33 Collocates of the verb clasp in COCA…………………………………. 76
Table 4.34 Collocations from exercise [18] in Passage……………………………. 76

ix
Table 4.35 Collocates of the verb cast in COCA…………………………………... 77
Table 4.36 Collocations from exercise [19] from Passage………………………… 78
Table 4.37 Collocates of the verb snap in COCA………………………………….. 78
Table 4.38 Collocations from exercise [20] in Passage……………………………. 79
Table 4.39 Collocates of the adjective notable in COCA………………………….. 79
Table 4.40 Collocations from exercise [21] in Passage……………………………. 80
Table 4.41 Collocates of the verb provide in COCA………………………………. 80
Table 4.42 Collocations from exercise [22] in Passage……………………………. 81
Table 4.43 Collocates of the verb adopt in COCA………………………………… 81
Table 4.44 Collocates of the verb cleanse in COCA………………………………. 82
Table 4.45 Collocations from exercise [24] in Passage……………………………. 83
Table 4.46 Collocates of the noun prejudice in COCA……………………………. 83
Table 4.47 Collocations from exercise [25] in Passage……………………………. 84
Table 4.48 Collocates of the verb assassinate in COCA…………………………... 84
Table 4.49 Collocations from exercise [26] in Passage……………………………. 86
Table 4.50 Collocates of the noun allegiance in COCA…………………………… 86
Table 4.51 Collocations from exercise [27] in Passage……………………………. 87
Table 4.52 Collocates of the adjective totalitarian in COCA……………………… 87
Table 4.53 Collocations from exercise [28] from Passage………………………… 88
Table 4.54 Collocates of the noun outlet in COCA………………………………… 88
Table 4.55 Summary of results from the two textbooks…………………………… 89

Table 5.1 Collocates of the adjective important in COCA……………………….. 99

Table 5.2 Useful collocations for Norwegian students among the ones found in
the texts…………………………………………………………………. 107

x
LIST OF FIGURES

Page

Figure 1.1 What is involved in knowing a word (Nation 1990:31)…………………… 3

Figure 1.2 The concept of collocation as a continuum (Conzett 2000:74)……………. 6
Figure 3.1 Example of a vocabulary exercise from @cross (Rogers et al. 2006b:111) 34
Figure 3.2 Collocation exercise from Passage (Sørhus et al. 2006b)………………… 34
Figure 3.3 Gap-filling exercise from Passage (Sørhus et al. 2006b)…………………. 35
Figure 3.4 Exercise with geographical names from @cross (Rodgers et al. 2006b)…. 35
Figure 3.5 Exercise to match synonyms from @cross (Rodgers et al. 2006b:34)……. 38
Figure 3.6 Exercise to match synonyms from Passage (Sørhus et al. 2006b)………... 40
Figure 3.7 Search interface of COCA (Davies 2008)………………………………… 47
Figure 4.1 Number of combinations in COCA and the OCD………………………… 90
Figure 5.1 Suggestion on how to improve exercise [6] ………………………………. 108
Figure 5.2 Suggestion on how to improve exercise [14]……………………………… 109
Figure 5.3 Suggestion on how to improve exercise [19]……………………………… 109
Figure 5.4 Suggestion on how to improve exercise [28]……………………………… 110
Figure 5.5 Suggestion on how to improve exercise [10]……………………………… 110
Figure 5.6 Lists of collocates of the synonyms assassinate and murder in COCA....... 111
Figure 5.7 Use of the synonyms answer and retort in different sections in COCA....... 112
Figure 5.8 Concordance lines of the synonyms persist and continue in COCA............ 112
Figure 5.9 Suggestion of a grammatical collocation exercise based on exercise [21]... 113
Figure 5.10 Suggestion of a „curiosity box‟ using the adjective attractive…………….. 113

ABBREVIATIONS

BNC British National Corpus

CALD Cambridge advanced learner’s dictionary
CLT Communicative language teaching
COCA Corpus of Contemporary American English
EFL English as a foreign language
ELT English language teaching
ESP English for specific purposes
OCD Oxford collocations dictionary for students of English
SLA Second language acquisition

xi
xii
1. INTRODUCTION

Collocations appear to be widespread phenomena in English – expressions such as highly

qualified, a central feature, and receive severe criticism abound in the language and their
importance for teaching has been increasingly recognized. However, it seems that collocations
have not yet been generally integrated into teaching material, and, as a consequence, not been
given serious consideration in the English classroom. The present study intends to address
some of these issues in connection to findings in corpus linguistics.
This chapter will first situate this study in relation to previous research, then the aims
and scope will be outlined and the main methods explained. The chapter continues with a
discussion of the salience of collocations in vocabulary teaching and it provides a definition
of the term. Finally, the terminology used and the structure of the thesis will be presented.

1.1 Relevance of the present study

Since the advent of computerized corpora in the 1960s, research within corpus linguistics has
demonstrated its potential not only for lexicography (e.g. Sinclair 1987; Hunston 2002) and
language research (e.g. Biber et al. 1998; Carter & McCarthy 1999; Partington et al. 2004),
but also as a resource in language teaching (e.g. Michael Lewis 1997; Hoey 2000; Osborne
2000; Yoon 2008). However, relatively few studies have associated corpora, teaching, and
textbooks (e.g. Römer 2004), and to my knowledge there is no research which in addition
treats vocabulary or, more specifically, collocations. Meunier & Gouverneur (2007) for
example study phraseology in textbooks for English language teaching (ELT) but their focus
is on more advanced learners. In fact, it seems that the use of corpora for teaching upper
secondary school students have not yet received due consideration. According to Chambers
(2005:121), most of the studies on corpora and teaching concentrate on the context of
university education (e.g. Bernardini 2000, 2002). She also observes that there is scant
evidence confirming that corpus consultation will become a complement to course books,
hence the need for more research integrating corpora into language teaching (Cambers
2005:111).
In order to fill this gap, the present study investigates collocations in two textbooks
used in upper secondary school in Norway and offers ideas on how corpora can complement
vocabulary exercises and contribute to teaching practice.

1
1.2 Aims and scope
As a teacher of English in upper secondary school in Norway, I have come to notice that some
textbooks tend to present isolated words in vocabulary exercises. Having learned about the
importance of teaching and learning words in chunks, I pondered whether these words could
have been taught in collocations instead.
Thus, my first aim is to investigate if words taught in isolation in vocabulary exercises
are part of collocations in the texts they are taken from. Based on previous research pointing
out the pervasiveness of collocations in the English language (Sinclair 1991; Michael Lewis
1993), I hypothesize that a high proportion of these words will appear in collocations in the
texts. The following research questions will also be addressed: (a) What proportion of the
words in the exercises appears in collocations in the texts? (b) Which of these collocations
would be useful to teach Norwegian students in the first year of upper secondary school?
It has been attested by class experimentation (Michael Lewis 2000), textbooks
(McCarthy et al. 2006a, 2006b), and linguistic research (Nesselhauf 2005) that including
collocations in course material is both attainable and highly recommendable. On these
grounds, my second aim is to suggest how the vocabulary exercises selected for the present
study can be complemented based on existing teaching material on collocations and by
applying previous research on corpora and foreign language vocabulary acquisition. On the
same premises, the third aim is to suggest additional ways in which collocations can be
integrated in vocabulary teaching. Thus, this study is situated within the field of applied
linguistics. More precisely, it belongs to the field of English language education (didactics)
and corpus linguistics in the sense that it attempts to contribute to language teaching material
and practice through the use of corpus information.
The scope of this investigation is restricted to two English textbooks used by
Norwegian pupils in the first year of upper secondary school. These books are Passage
(Sørhus et al. 2006a) and @cross (Rodgers et al. 2006a), both adopted for the general studies
program. The material consists of the @cross textbook and workbook, the Passage textbook
and website, and the Corpus of Contemporary American English (COCA; Davies 2008).

1.3 Methods
Both quantitative and qualitative methods have been employed to achieve the aims outlined
above. Regarding the first aim, the words from the vocabulary exercises selected from each
textbook were initially checked in context to see if they were part of any word combinations.
To determine if the combinations found in the texts are collocations, they were looked up in

2
the Oxford collocations dictionary for students of English (OCD; McIntosh et al. 2009) and
their frequency and strength were checked in the corpus. Thereafter, the percentage of single
words from the exercises that are part of collocations in the texts was calculated. Determining
whether or not a collocation is useful for students is a more subjective matter and open to
interpretation. For the purposes of this study, I have used the competence aims of the
Norwegian national curriculum for the English subject (L06 2006b) and research on
vocabulary acquisition and collocations to help suggest what types of collocations are useful
for teaching students at the level in question. Regarding the fulfillment of the second aim, the
corpus was searched to identify the most frequent combinations with the words from each
exercise and their strength. The findings were then used together with previous research and
teaching materials based on corpora to propose ideas on how to improve the vocabulary
exercises. The same sources were used to give suggestions on how collocations can be
integrated in language teaching. In sum, this study will hopefully show that vocabulary
teaching can profit from corpus use and from placing greater emphasis on collocations.

1.4 Why collocations?

In order to situate collocations within the area of vocabulary teaching, I will take the concept
of „word‟ as a point of departure (see Carter & McCarthy 1988). What exactly does it mean to
know a word? Nation (1990:31) proposes a list of the different aspects of knowledge needed,
as presented in figure 1.1 below. The R and P stand respectively for receptive (listening and
reading) and productive (listening, reading, speaking, and writing) knowledge:

Form Spoken form R What does the word sound like?

P How is the word pronounced?
Written form R What does the word look like?
P How is the word written and spelled?

Position Grammatical Patterns R In what patterns does the word occur?

P In what patterns must we use the word?
Collocations R What words or types of words can be expected before or
after the word?
P What words or types of words must we use with this word?

Function Frequency R How common is the word?

P How often should the word be used?
Appropriateness R Where would we expect to meet this word?
P Where can this word be used?

Meaning Concept R What does the word mean?

P What word should be used to express this meaning?
Associations R What other words does this word make us think of?
P What other words could we use instead of this one?
Figure 1.1 What is involved in knowing a word (Nation 1990:31)

3
As illustrated in the table, Nation uses the four general classification criteria drawing from
George (1983): „form‟, „position‟, „function‟, and „meaning‟. Unfortunately, some of these
aspects tend to be given more prominence in the foreign language classroom, like „form‟ and
„meaning‟, while others are hardly ever mentioned, which seems to be the case of „position‟
(grammatical patterns1 and collocations). Nattinger affirms that the meaning of a word is
closely related to the word‟s associations and argues that „the whole notion of collocations is
extremely important for acquiring vocabulary and has yet to be exploited to its full potential‟
(1988:70). As the table shows, collocations, as well as all other aspects of a word, can be
taught in relation to both receptive and productive knowledge (Nattinger 1988:75), the
difference between the two being in the type of activity dealt with in class.
Thus, the main reason why the present study lays emphasis on collocations is that
although their importance is to a great extent acknowledged (Sinclair 1991; Cowie 1998a),
not much attention has been given to them in the teaching context. In order to further
elucidate the term, a literature review is provided in the next section.

1.5 What is a collocation?

McIntosh et al. define collocation as „the way words combine in a language to produce
natural-sounding speech and writing‟ (2009:v). However, there are various ways to define it
and it seems that linguists and teachers have not yet agreed upon a common description. In
line with Fontenelle, „there does not seem to be any clear-cut, non-controversial definition of
the term “collocation”‟ (1998:191).
As a means of clarifying the concept, a general distinction that needs to be made is, in
Sinclair‟s terms, between the open-choice principle and the idiom principle. As Sinclair
explains it, the open-choice principle is „the normal way of seeing and describing language‟
(1991:109), usually referred to as a „slot-and-filler‟ model where nearly any word could fill
the slots. However, this principle does not explain the numerous constraints in language
choice. For instance, there is no reason why we do not say *to put something on fire, but
English native speakers would agree that the appropriate collocation is to set something on
fire. According to Sinclair, „the principle of idiom is that a language user has available to him
or her a large number of semi-preconstructed phrases that constitute single choices, even
though they might appear to be analyzable into segments‟ (1991:110). This model of

1
It is worth observing that the grammatical patterns of a word are not to be confused with what is generally
described as „grammar‟. See Hunston (2002:137-169) for clarification and examples.

4
interpretation is illustrated not only by collocations but also by other types of fixed
expressions such as idioms, proverbs, clichés, technical terms, jargon expressions, and phrasal
verbs (Sinclair 1991:111). The idea regarding these principles is that they co-exist, but can not
be employed simultaneously.
Although the term collocation is used in different senses by a number of authors
(Halliday 1966; Sinclair 1991; Stubbs 1995; Moon 1998), two main views of the concept can
be identified: the phraseological approach and the frequency-based approach (Nesselhauf
2005:12). According to the frequency-based definition, a collocation is said to be the co-
occurrence of words at a frequency that is higher than expected if words were combined
arbitrarily in a language. Its main representatives are Firth, Halliday, and Sinclair. In the
phraseological definition, „collocations are associations of two or more lexemes (or roots)
recognized in and defined by their occurrence in a specific range of grammatical
constructions‟ (Cowie 1994:3169). This tradition emphasizes the relation between lexical and
syntactic patterning in collocations and it has its roots in the work of H. Palmer (e.g. 1933)
and Hornby (e.g. Hornby et al. 1942). This view has been strongly influenced by Russian
phraseology and some of its supporters are Hausmann (1989), Cowie (1998a), and Mel‟čuk
(1998).
Collocations are considered by both the phraseological and frequency-based
approaches as combinations of one or more lexemes or lemmas. According to Carter‟s
definition, lexemes are „the basic, contrasting units of vocabulary in a language‟ (1987:6-7).
He exemplifies this by saying that when we look up words such as bringing or brought in a
dictionary, we look for the lexeme BRING. As a way to restrict my search, I have limited
myself to collocations with only two lexemes. In the case of the collocations found in the
textbooks, however, a few longer ones have been included (e.g. get good grades) since I am
interested in the types of collocation from the texts that might be useful for students.
According to the phraseological approach, the lexemes of a collocation should be
syntactically related and these relations usually fall into ten categories. The following are the
ones set by Hausmann (1989:1010; in Nesselhauf 2005:22): adjective + noun (bright light),
(subject-) noun + verb (the wind howled), noun + noun (consumer goods), adverb + adjective
(bitterly disappointed), verb + adverb (handle carefully), and verb + (object-) noun (run the
risk) (my examples). In addition, as suggested by Benson et al. (1997:ix), lexical words plus a
preposition also constitute collocations, so that the other four categories would be: verb +
preposition (glance at), noun + preposition (interest in), preposition + noun (by accident), and
adjective + preposition (angry at) (original examples). The first six categories are usually

5
called „lexical collocations‟ and the four others „grammatical collocations‟ (Fontenelle
1998:192), the latter are sometimes also called „colligations‟ (Bartsch 2004:24). These ten
categories plus the lexical one adverb + verb (simply vanished) are used in the present study
to classify the collocations from the textbooks and corpus (see chapter 3).
Still within the phraseological approach, Cowie arranges word combinations in four
groups devised in terms of transparency and commutability. Nesselhauf (2005:14)
summarizes Cowie‟s (1981) classification as follows:

Free combinations (e.g. drink tea):

- the restriction on substitution can be specified on semantic grounds [i.e. you can
substitute tea by coffee, water, juice, etc.]
- all elements of the word combination are used in a literal sense

Restricted collocations (e.g. perform a task):

- some substitution is possible, but there are arbitrary limitations on substitution [e.g.
you can also say do a task, but not make a task]
- at least one element has a non-literal meaning, and at least one element is used in its
literal sense; the whole combination is transparent

Figurative idioms (e.g. do a U-turn, in the sense of „completely change one‟s policy
or behaviour‟):
- substitution of the elements is seldom possible
- the combination has a figurative meaning, but preserves a current literal
interpretation

Pure idioms (e.g. blow the gaff):

- substitution of the elements is impossible
- the combination has a figurative meaning and does not preserve a current literal
interpretation

Authors adopting a phraseological approach often use the term collocation to refer only to
„restricted collocations‟ and other terms to label the other kinds (c.f. Benson et al. 1997).
Although pertinent, I find Cowie‟s classification of „literal‟ and „figurative‟ senses not
sufficiently clear-cut. For example, are constitutional monarchy and customs agent (see
chapter 4) free combinations or restricted collocations? If they are restricted collocations,
which word is used in a non-literal sense? In contrast to Cowie, Conzett (2000:74) adopts a
less rigorous approach, illustrating the idea of collocation as a continuum, as pictured below:

friendly dog strong coffee sibling rivalry throw in the towel

weaker stronger
old car heavy smoker mitigating circumstances Stars and Stripes

Figure 1.2 The concept of collocation as a continuum (Conzett 2000:74)

6
She excludes the two extremities (e.g. friendly dog and throw in the towel) and treats as
collocations only the items in the middle of the continuum. Despite appearing simpler, this
approach is also problematic: Where are the boundaries in this scale? Who or what decides
where a combination should be placed? Carter (1987:55) suggests that collocational
acceptability can be decided based on statistical measures of native-speaker intuitions. Church
& Hanks (1990), on the other hand, recommend using mutual information (MI) with a score
of 3.0 as a threshold, while McEnery et al. (2006:56) report that other statistical measures of
association like z-score and t-score are also commonly used.
All the above considered, it is clear that linguists have not yet agreed upon the most
accurate measure to define what a (restricted) collocation is. Therefore, I have decided to
combine a number of elements to arrive at a definition appropriate for the present study. The
OCD (McIntosh et al. 2009) is used as one of the tools to determine whether a word
combination is a collocation or not. This dictionary was chosen due to its focus on usage (it is
based on a 2 billion word corpus), its incorporation of syntactic relations (it classifies
collocations in syntactical categories), and its concern with learners‟ needs (see McIntosh et
al. 2009:v). Thus, it combines features of the phraseological and the frequency-based
approaches and suits the pedagogical purposes of the present study. A disadvantage, however,
is that this dictionary includes too wide a range of collocations, as defined in its aim:

Give the full range of collocation – from the fairly weak (see a movie, an enjoyable
experience, extremely complicated), through the medium-strength (see a doctor, direct
equivalent, highly intelligent) to the strongest and most restricted (see reason, burning
ambition, blindingly obvious) – for around 9,000 headwords (McIntosh et al. 2009:v).

Although it is explained further that „totally free combinations are excluded and so, for the
most part, are idioms‟ (McIntosh et al. 2009:vi), the dictionary includes combinations like big
difference, very important, and is impressive. These, according to the substitution criteria set
out by Cowie (above), would fit the category of free combinations and as such are not
interesting for this study. As a means to exclude these combinations, I have used the MI
threshold of 3.0 (see chapter 3), as provided by the corpus. The dictionary also includes under
the label „phrases‟ some idiomatic expressions of the kind Cowie calls figurative idioms.
These were only included when the cultural element was considered highly important within
the chapter, as the case of member of parliament in a chapter about government in the UK in
@cross (see chapter 4). Otherwise they are not regarded as collocations in this thesis.

7
 Another criterion established to decide whether a combination is a collocation was to
check frequency. Drawing on Clear (1993:277), I have adopted the threshold of three
occurrences in COCA to classify a pair as a collocation (see chapter 3). In addition, elements
of the phraseological approach were included in that I have taken syntactic relations into
account and have adopted Cowie‟s terminology of (restricted) collocations and free
combinations. Thus, the definition of collocations adopted in this study is:

Collocations are arbitrarily restricted lexeme combinations that are syntactically fixed
to a certain degree, are included in the collocation dictionary, present an MI score
higher than 3.0, and have a raw frequency of more than three tokens in COCA.

1.6 Terminology
In addition to the term collocation, some other terms will be used on a regular basis
throughout this study. Here is a list of definitions of this terminology:
Collocate – a word which occurs in close proximity to the word under investigation without
necessarily being part of a collocation (e.g. very is a collocate of important although very
important is considered a free combination)
Corpus – a large amount of written and sometimes spoken material, usually in computerized
form, collected to show the state of a language.
Concordance – a list of the uses of a word/lexeme in context provided by a corpus.
Foreign language – the language learned by students which is neither their mother tongue nor
widely spoken in their country.
Word combinations – a general term used in this study to refer to all kinds of phraseological
chunks (collocations, free combinations, idioms, phrasal verbs, fixed expressions, etc.).
A few other technical terms are also used, but are introduced and explained where relevant for
the discussion.

1.7 Structure
This thesis is structured in six chapters. Chapter 2 provides some background for the teaching
of vocabulary and presents previous research in the areas of collocation, corpora, and
teaching. Chapter 3 describes the material and methods used, while chapter 4 presents the
results from the analyses. In chapter 5 the results are interpreted in light of the background
provided in chapter 2 and suggestions are given on how the exercises can be improved and
how collocations can be included in vocabulary teaching. Finally, chapter 6 brings the study
to a conclusion, summarizing the main findings and proposing ideas for further research.

8
2. BACKGROUND

2.1 Introduction
This chapter will first give an account of the teaching methods which have been most
influential in Norway and how these methods treat vocabulary. Aspects of foreign language
vocabulary acquisition and some of the previous Norwegian curricula will be considered.
Collocations will also be discussed in relation to other areas of study such as cognitive
linguistics, communicative language teaching, and culture. Lastly, relevant studies in corpus
linguistics will be mentioned.

2.2 Teaching methods and the role of vocabulary

A number of authors have given surveys on the various teaching methods employed in the last
few centuries (Larsen-Freeman 1986; Simensen 1998; Richards & Rodgers 2001; Howatt
2004) and on how vocabulary has been taught in these methods (Zimmerman 1997; Schmitt
2000). This section will give a description of the teaching methods which were most central in
Norway in recent centuries and the role of vocabulary in each of them.
The first predominant method to be implemented in Norwegian schools was the
Grammar-Translation Method. It was developed based on procedures used for teaching Latin
and evolved out of the need to systematize foreign language teaching for school children
(Howatt 2004:151). Zimmerman (1997:5) reports that this method was introduced in Prussia
at the end of the eighteenth century and its main aims were to enable students to read classical
literature and pass standardized exams. It was not expected that students would ever come to
speak the languages they were studying, but that they would profit intellectually from the
learning. Students were given extensive grammatical explanations in their own language and
bilingual vocabulary lists, which were to be learned to help them translate classical texts. In
this method vocabulary was introduced mainly as a way of illustrating a grammar point and
when vocabulary difficulties arose, they were usually addressed by means of etymology.
According to Steinberg & Sciarini, the grammar translation method „has enjoyed and
continues to enjoy acceptance in many countries around the world‟ (2006:141), principally in
countries where language teachers are non-fluent and the classes are very large. Despite its
advantages and the fact that it was widely used in Europe and in the United States well into
the twentieth century, the method has drawn harsh criticism. The main objections have been
the disregard of „realistic, oral language‟ (Zimmerman 1997:6) and the focus on language

9
analysis instead of language use (Schmitt 2000:12).
As a consequence, by the end of the nineteenth century a new movement emphasizing
listening and pronunciation appeared. This was known as the Reform Movement and one of
its great achievements was the development of phonetics and the recognition of it as a
science. Its members also defended the study of „connected texts rather than unconnected
sentences and lists of isolated words‟ (Simensen 1998:26). As a product of this movement, a
use-based method emphasizing listening was introduced towards the end of the nineteenth
century: the Direct Method. According to Zimmerman (1997:9), explicit grammar teaching
and translation were set aside, while students were expected to learn English through the same
process as native speakers do. They would listen first, and then speak, and only in later stages
would they learn to read and write. Vocabulary was thought to be generally acquired through
interactions in the classroom by asking and answering questions. Concrete words were taught
using pictures, mimic, and realia (real objects) while associations of ideas were used to teach
abstract vocabulary (Zimmerman 1997: ibid.). This method, however, also had its
weaknesses. As Schmitt (2000:12) points out, teachers were not always proficient in the target
language, failing to give students the input needed. Moreover, unlike L1 acquisition, in L2
acquisition students only have a few hours of exposure to the target language per week. This
limited instruction time was one of the factors taken into account by the 1929 Coleman Report
in the United States. This report concluded that the time of instruction was not enough to the
development of comprehensive language proficiency and recommended that secondary
students should instead be taught how to read in a foreign language (Schmitt 2000:13).
According to Richards & Rodgers, this stress on reading „led to the development of principles
of vocabulary control, which were to have a major practical impact on the teaching of English
in subsequent decades‟ (2001:37). It was at this point that lists of vocabulary based on
frequency counts began to emerge (e.g. West 1953).
During the Second World War, it became clear that the approaches above did not
manage to form competent users of the target language. The American military lacked people
who were fluent in foreign languages and needed a teaching program that could train the
soldiers rapidly in oral skills (Larsen-Freeman 1986:31). Structuralist linguist Charles Fries
developed a method based on behaviorist principles (forming language habits) and on the
Direct Method (emphasizing listening and speaking) which was first called the Army Method
and later known as Audiolingualism. In this method new words were only introduced in drills,
and vocabulary was thought to be acquired naturally through good language habits. It was
also suggested that beginners should not learn too much vocabulary to prevent a false sense of

10
security (Zimmerman 1997:11). A parallel approach was used in Britain around the 1950s
(although being developed as early as the 1920s) called Situational Approach or the Oral
Method. The name derived from the idea of teaching language in sentence patterns replicating
real situations. As Richard & Rodgers put it, „language was viewed as purposeful activity
related to goals and situations in the real world‟ (2001:40). Vocabulary was chosen to best
illustrate and practice the sentence patterns and was presented as lists in substitution tables.
In spite of their appeal in language teaching, the behaviorist ideas of habit formation
were severely attacked in the 1950s by Chomsky, who claimed that language was partly
innate and governed by abstract rules (see Chomsky‟s review of Skinner 1959). As explained
by Simensen, he „claimed that the stimulus-response and conditioning theory of the
behaviorists could not explain the creativity involved in generating all kinds of new
utterances‟ (1998:81; author‟s emphasis). Even though Chomsky brought back to attention the
creative aspect of language, he gave vocabulary scant importance. In reaction to Chomsky‟s
notion of an autonomous linguistic competence, Hymes (1972) advanced the concept of
communicative competence, emphasizing the relevance of social interaction to language
learning (Zimmerman 1997:12). A new approach was developed from this notion and became
known as Communicative Language Teaching (CLT). Elements of this approach, such as
communication and cultural knowledge, are clearly emphasized in the latest curricula in
Norway (Simensen 1998:118-119). Even though it is a meaning-based approach, vocabulary
is again given a secondary status. According to Sökmen (1997:237), in the communicative
approaches the priority is incidental vocabulary learning, so that students are encouraged to
guess from context, use monolingual dictionaries, and avoid translation. CLT will be treated
in more detail in subsection 2.5.4.
In the 1980s The Natural Approach was developed by Krashen & Terrell (1983) to aid
beginners in achieving the ability to communicate orally in the classroom. It is based on five
hypotheses, and „[its] methodology emphasizes comprehensible and meaningful input rather
than grammatically correct production‟ (Zimmerman 1997:15). It follows that vocabulary, as
relevant language input, is considered essential to the acquisition2 process. As for more
advanced students, Krashen (1989) suggests reading as the best way to develop wider
vocabulary knowledge.
Although treated separately in most teaching methods, recent evidence from corpora
has revealed that grammar and vocabulary are fundamentally linked. Lexical patterning is

2
The terms acquisition and learning as defined by Krashen are not differentiated in the present study, but used
interchangeably.

11
shown to exist on a much larger scale than ever imagined, making it difficult to separate
grammar from lexis. In line with Schmitt, „one must conceptualize them as partners in
synergy with no discrete boundary, sometimes referred to as lexicogrammar‟ (2000:14). This
idea gives vocabulary a more central place and is one of the greatest contributions corpus
studies have made to the field of language teaching. Some authors give lexis an even greater
role, and are the advocates of a lexical approach. According to Richards & Rodgers, this
approach is „derived from the belief that the building blocks of language learning and
communication are not grammar, functions, notions, or some other unit of planning and
teaching but lexis, that is, words and word combinations‟ (2001:132). Some of the attempts
to integrate this view into teaching can be seen in The lexical syllabus (Willis 1990), Lexical
phrases and language teaching (Nattinger & DeCarrico 1992), and The lexical approach
(Michael Lewis 1993). Albeit not yet diffused in Norway, these lexical approaches are closely
related to the topic of the present study, i.e. the phenomenon of collocation.

2.3 Foreign language vocabulary acquisition3

In this section, aspects of foreign language vocabulary acquisition relevant to the present
study such as implicit and explicit learning, the different steps in the vocabulary acquisition
process, vocabulary learning strategies, and the influence of L1 in L2 learning will be
presented. Although pertinent under this topic, the role of memory will be discussed in
connection with cognitive linguistics (see subsection 2.5.2).
One of the contentious issues within the field of foreign language vocabulary
acquisition is implicit (incidental) versus explicit (direct) vocabulary teaching and learning. In
line with Sökmen, „the pendulum has swung from direct teaching of vocabulary (the grammar
translation method) to incidental (the communicative approach) and now, laudably, back to
the middle: implicit and explicit learning‟ (1997:239; emphasis added). Although these
tendencies have developed in general chronologically, different views still coexist: while
some advocate the (implicit) acquisition of vocabulary mainly by guessing from context
(Nagy et al. 1985; Krashen 1989), others have realized the importance of combining the two
approaches (Coady 1993), particularly in recognition of learners‟ individual strategies
(Hulstijn 1993; Nation 2001). Despite its great impact on language teaching since the 1970s,
CLT has been criticized for prioritizing implicit vocabulary teaching. Some of the criticisms
are whether beginners know enough vocabulary to guess from context (Coady 1997:227) and

3
Although I acknowledge the difference between the terms second language (L2) and foreign language, they are
used in this section in free variation since I quote authors who do so (e.g. Swan 1997).

12
the fact that incidental learning has not shown to be more effective than other learning
strategies such as the keyword method (Moore & Surber 1992). Moreover, other studies have
proved that although reading for meaning has increased L2 vocabulary acquisition, direct
instruction has reached even better results (Paribakht & Wesche 1993; Zimmerman 1994).
These studies suggest that promoting a mixture of direct and incidental learning combined
with learning strategies is the most fruitful way of teaching vocabulary.
Concerning the way in which new words are acquired, Hatch & Brown explain the
five steps in the vocabulary acquisition process (based on Payne 1988): (1) encountering new
words, (2) getting the word form, (3) getting the word meaning, (4) consolidating word form
and meaning in memory, and (5) using the word (Hatch & Brown 1995:374). Thus, as the first
step of the acquisition process, learners will encounter new words for example on the internet,
television, music, books, movies, magazines, etc. In the case of the present study, the words
from the exercises selected are encountered by students in the course book texts or listening
activities. Hatch & Brown (1995:417) provide examples of how the other steps can be
achieved. For example, getting the word form (2) is the step when students connect the sound
with the word form, which can be done by reading the word aloud. As a way to get the word
meaning (3), students can use bilingual or monolingual dictionaries or guess from context.
Consolidating word form and meaning in memory (4) can be achieved by doing a number of
exercises such as matching or translating words, filling in gaps, crossword puzzles, memory
games, etc. Finally, to learn the uses of a word (5), students might be asked to write sentences
or texts, answer questions that require the use of the word, or even consult concordances to
check how the word is used (Hatch & Brown 1995:417-419). Most of the exercises selected
in the present study are concerned with the fourth step and only a few with the fifth.
A way of achieving all these steps is to make use of various learning strategies.
Nation (2001:218) presents a table with different kinds of vocabulary learning strategies than
can be encouraged in class, as reproduced in table 2.1 below (the explanation that follows is
also based on Nation 2001):

13
Table 2.1 Types of vocabulary learning strategies (Nation 2001:218)
General class of strategies Types of strategies

Planning: choosing what to focus on and when to Choosing words

focus on it Choosing the aspects of word knowledge
Choosing strategies
Planning repetition
Sources: finding information about words Analysing the word
Using context
Consulting a reference source in L1 or L2
Using parallels in L1 and L2
Processes: establishing knowledge Noticing
Retrieving
Generating

In table 2.1, the first strategy regarding „planning‟ is „choosing words‟. This is related to the
vocabulary goals of the learner or of the syllabus being used; consulting frequency lists for
instance might be useful to select words to be learned. The second strategy, „choosing the
aspects of word knowledge‟, means deciding which of the aspects listed in figure 1.1 (form,
position, function, and meaning) should be given prominence. Examples of „choosing
strategies‟ can be to refer to a dictionary or make word cards. The fourth is „planning
repetition‟ and this is when learners create a system for revising old material.
Under „sources‟ the first strategy is to analyze words. This can be done by looking at
word parts, for example affixes and stems. The second is that learners should use context,
drawing on background knowledge and guessing meaning from clues in the text. The third is
„consulting a reference source in L1 or L2‟, those can be dictionaries, concordances, teachers,
native speakers, etc. The fourth is using parallels from other languages to help learning a new
item (this will be developed further below).
Under „processes‟ the first strategy is noticing words to be learned according to one‟s
needs. The second is „retrieving‟, i.e. revising previously stored items by using cues (students
can for instance cover the word form and remember it by looking at its meaning). Finally, the
third strategy under „processes‟ is „generation‟. This type includes, among others, trying to
visualize examples of the word, „creating contexts, collocations and sentences containing the
word, mnemonic strategies like the keyword technique, and meeting and using the word in
new contexts across the four skills of listening, speaking, reading and writing‟ (Nation
2001:222). All of these strategies can be encouraged in class to facilitate student‟s vocabulary
acquisition. However, it is worth recalling that this taxonomy is only one of the many ways
vocabulary learning strategies can be classified; other examples are to be found in Schmitt
(1997) and Takač (2008).

14
 Another relevant aspect of vocabulary acquisition for the present study is the use of L1
in language teaching and learning. Natural methods like the Direct Method banned the use of
L1 in class due, among other reasons, to the distrust in a one-to-one correspondence between
languages (Howatt 2004:313). The idea that there are exact equivalents can indeed be
problematic when we consider the complex relationships that exist between words in different
languages. Swan (1997:157-160) lists some these relationships as follows: words in various
languages can have different grammatical contexts or collocations; equivalent concepts can be
assigned to different parts of speech; there can be false cognates and differences in style and
levels of formality; and the notion of a „word‟ might vary, i.e. a word in a language can be
translated to three or four in another. One way to avoid mistakes caused by these variations is
to follow Morgan Lewis‟ suggestion that an item (such as widely available or catch up with
the news) should be translated „not word-for-word but whole phrase to whole phrase, bearing
in mind that the structure of the expression may be very different in one language from the
equivalent expression in the other‟ (2000:16). At the same time, Takač (2008:9) defends the
importance of the L1 to the learning of the L2 in the sense that, by establishing equivalents
between the two languages, learners do not need to relearn how to categorize the world.
Hence, translation in vocabulary teaching should be used when it enhances learning and
avoided when it prevents it.

2.4 Vocabulary in Norwegian curricula

In 2006, a new national curriculum was introduced in Norway presenting revised guidelines
in different school subjects. This curriculum is called Læreplanverket for Kunnskapsløftet, or
Knowledge Promotion (2006a), and will be referred to as L06 from now on in the present
study.
Compared to the previous curriculum (L97), L06 (2006b) includes five basic skills in
the English subject to be focused on at all levels of schooling. These are:

Being able to express oneself in writing in English

Being able to express oneself orally in English
Being able to read English
Having skills in mathematics in English
Being able to use digital tools in English

The introduction of the last skill is particularly important in that it was accompanied by the
sponsoring of personal computers for all students in upper secondary school (the government

15
covers part of the cost). According to the Directorate for Education and Training
(Utdanningsdirektoratet; Udir 2007), the plan is that from the fall of 2009 every Norwegian
upper secondary school student will have a personal laptop (and internet) to be used in class.
This development has opened new possibilities for the teaching of English in many ways
since students now have access to online newspapers, blogs, encyclopedias, dictionaries,
forums, corpora websites, among other language resources.
In addition to a description of the five basic skills, L06 provides a list of competence
aims that should be achieved by the end of certain school years, among them the first year of
upper secondary school (the list of aims is provided in appendix 1). These aims are divided
under three main subject areas: „language learning‟, „communication‟, and „culture, society
and literature‟. The only reference to vocabulary in the whole curriculum is under the area of
„communication‟ and it is simply stated that students are expected to „master a wide
vocabulary‟ (L06 2006b). This is a rather general statement and gives little or no guidance to
teachers when it comes to this important part of the teaching and learning of a language. A
comparison of L06 with previous curricula reveals that they have become less and less
specific about vocabulary. Simensen reports that between 1950 and 1975, „a systematic
approach to the selection, grading, and repetition of vocabulary was […] given high priority‟
(1998:59). According to Daasvand (2001), the curricula from 1957 and 1974 contained word
lists to be learned and the curriculum from 1960 recommended the teaching of vocabulary
based on frequency lists. She also observes a decreased attention to vocabulary in the
curriculum from 1987 (2001:33-34), while Vestre (1998:9) remarks that in L97 vocabulary
work was mentioned only five times. The reason for this development might be that in a time
when English is the main language of games, movies, TV-series, music, and information,
deciding on a vocabulary list could limit this great window of possibilities. At the same time,
it is left to the teacher to decide what kind of vocabulary should be highlighted, which might
not be an easy task considering the amount of texts and new words students often encounter.
One way to tackle these problems is calling students‟ attention to useful vocabulary
appearing in the texts from textbooks. Conzett affirms that „in-context study of vocabulary
encountered in reading has the well-known advantages of point-of-need relevance to the
student, and natural, real-life examples of usage‟ (2000:72). Woolard (2000:33-34) suggests
that students should learn how to notice collocations in the texts they read, and, more
importantly, they should be able to select the ones that they need. This is a learning strategy
that can aid learners in widening their vocabulary knowledge.

16
2.5 Collocations
A definition of collocation and the reasons why it has been chosen as the topic of this thesis
were provided in chapter 1. In the present chapter collocations will be considered in relation
to other relevant fields of study.

2.5.1 Background
As pointed out in the introduction, the definition of collocation is a matter of heated dispute,
and it seems like each scholar chooses to define it either in overly general terms or according
to the specific aims of their study. What is nonetheless agreed upon is the value of
collocational knowledge for linguistic research and pedagogical purposes (Ellis 1997:128-
129).
„You shall know a word by the company it keeps‟ (Firth 1968:11). This is one of the
most common quotations used when scholars attempt to define collocation. Firth brought the
term to the linguistic arena in his paper „A synopsis of linguistic theory, 1930-1955‟ in which
he frequently alluded to „the collocational level‟ of meaning. He believed that we obtain the
meaning of a word or a text through a „mutually congruent series of levels‟ (Firth 1968:30),
these levels being contexts of situation, syntax, phonology, phonetics, and collocation.
Although the term came to be known mostly through Firth, H. Palmer had already commented
on it several years before, saying that each collocation „must or should be learnt, or is best or
most conveniently learnt as an integral whole or independent entity, rather than by the process
of piecing together their component parts‟ (1933:4; in Nation 2001:317). In fact, Bartsch
(2004:28) notes that the first recorded use of the term is a quotation from the year 1750, but
that then it simply meant a grammatical relation between words.
In spite of its early origin, the term collocation has only been given more attention
after the development of electronic corpora in the 1960s and with it the opportunity to observe
combinational patterns in the English language. Since then, research on collocations has
increased substantially. Some authors even defend the view that knowledge of language
depends on collocational knowledge (Ellis 2001) or that for a speaker to be fluent and use a
language appropriately collocational knowledge is required (Pawley & Syder 1983).
However, even though the widespread use and importance of collocations in the language is
recognized by many (Kjellmer 1984; Stubbs 1995; Nation 2001), few attempts have been
made to integrate the teaching of collocations in the English learning curriculum. Some
exceptions are Michael Lewis, who has developed the Lexical Approach (1993), and
McCarthy et al. (2006), authors of the Touchstone material that uses corpus information and

17
includes collocations in the vocabulary work. Another important contribution is from
Nesselhauf (2005), who studies collocations in a learner corpus and based on the results
provides suggestions on how to select collocations for teaching.
The next subsection presents research on why it is conceded that teaching and learning
words in chunks is more effective.

2.5.2 Cognitive linguistics, chunking, and memory

Cognitive linguistics is a large field of study with its own principles and theories. Since the
present thesis is not based on one theory but takes an eclectic approach, cognitive linguistics
will be specifically related to collocations and teaching.
A common assumption in cognitive linguistics is, according to Taylor, „the belief that
language forms an integral part of human cognition, and that any insightful analysis of
linguistic phenomena will need to be embedded in what is known about human cognitive
abilities‟ (2002:4). Some of these cognitive abilities are, as described by Schmid (2007:117),
perception, memory, and attention allocation. Among these, memory is particularly
interesting in connection with collocations, as explained by Nattinger:

Finally, one of the earliest findings from memory research was that short term memory
holds a fairly constant number of units (Miller 1956), units which later research has
shown likely to be „chunks‟ of information, composed of several rather than single
items. […] Since a great part of the learner‟s task is to chunk unfamiliar material in
meaningful ways, the teacher who makes this chunking easier increases the number of
items the learners retain. (Nattinger 1988:64)

A point worth noting here is that Nattinger has a broad interpretation of collocations,
including phrases such as How do you do?, which is different from the definition adopted in
the present study. Nonetheless, since collocations are also part of the more general idea of
„chunks‟, his discussion of storage of chunks versus single items in memory is highly
relevant.
Nation (2001) also talks about the role of chunking in the language users‟ memory. He
explains the advantages and disadvantages of chunking and affirms that, „the main advantage
of chunking is reduced processing time. […] Instead of having to refer to a rule or pattern to
comprehend or produce the chunk, it is treated as a basic existing unit‟ (Nation 2001:320).
This idea is closely related to Sinclair‟s idiom principle, which establishes that language users
do not process all of the options available when uttering a sentence; they simply use ready
chunks that have been heard and employed many times before. Nation and Sinclair‟s views

18
are directly connected to two of the various concerns of cognitive linguistics – automatization
and storage vs. computation. According to Taylor (2002:13-14), these cognitive capacities
can be perceived in language in that a great amount of what we say is automated, which is
possible because stored chunks can be recalled rapidly and effortlessly, in contrast to the slow
process of computation. He argues that the fact that rules can be applied in language to
produce the correct input (computation) does not entail that they are always used by speakers,
who often rely on ready-made language forms (Taylor 2002:14).

2.5.3 Language teaching and learning

„It is not true that our students necessarily learn what we teach them‟ (Morgan Lewis
2000:11). Morgan Lewis is categorical in his claim that teaching does not cause learning,
which is also asserted by Larsen-Freeman (1997). According to him, while teaching is linear
and step-by-step, learning is non-linear and cyclical (Morgan Lewis 2000:12). In
consequence, students might only learn a taught item many weeks, months, or even years
later, after he or she has encountered it a great number of times. Therefore, teachers should
provide opportunities in which students can repeatedly meet the target vocabulary. One way
of doing this, as suggested by Morgan Lewis and endorsed by Nesselhauf (2005:259), is that
instead of prioritizing teaching new and rare words, teachers should concentrate on expanding
knowledge of what is only „half-known‟ by teaching students new collocates of a known-
word. As he affirms, „time spent on half-known language is more likely to encourage input to
become intake than time spent on completely new input‟ (Morgan Lewis 2000:24).
Hill (2000) also promotes the idea of teaching a word with its most common
collocates, explaining that most intermediate students know for example the words hold and
conversation, but that they may not know the collocation hold a conversation (Hill 2000:64).
He argues that „the main thrust of vocabulary work in most classes should be to make students
more collocationally competent with the words with which they are already partly familiar‟
(Hill 2000:67). What would then be this „familiar language‟ and its collocations? Hill et al.
(2000) recommend teaching the collocates of nouns that do not have very specific meaning
and which are often accompanied by an adjective. These are also the nouns that present the
longest entries in a collocation dictionary. Here is the list that they propose:

Account, action, answer, approach, argument, behaviour, change, circumstances,

condition, consequences, decision, difference, discussion, effect, feature, idea,
information, interest, issue, manner, method, move, performance, plan, policy,
position, problem, programme, project, question, reason, relationship, result, scheme,

19
 situation, solution, state, story, style, system, theme, theory, use, view, vision, way,
work. (Hill et al. 2000:101)

They choose the word situation from this list and exemplify how collocations can be used in a
classroom activity presenting the following: an embarrassing situation, a bewildering
situation, a tricky situation, a unique situation, an extraordinary situation, and a tense
situation. Then they ask students to think of an example from their own life in which these
collocations would apply. This idea is in accordance with Morgan Lewis‟s argument that a
collocation „can evoke a complex situation very precisely‟, as in the cases of routine check-
up, disperse the crowd, and widely available, affirming that when we teach these items apart
they lose their power to communicate (2000:15).
In a study of phraseology in learners‟ academic writing, Howarth (1998:181) has
found that students consider de-lexicalized verbs like get, put, take, do, and make problematic
and that they avoid word combinations with these verbs because of their uncertainty as to the
appropriate collocates. This indicates that de-lexicalized verbs are another category that
should be drawn attention to in class. If this type of collocations is critical for university
students it is safe to assume that it is no less of a problem for upper secondary school learners.
In face of the various challenges in language teaching and learning, many claim that
teachers should be more concerned about providing learners with strategies (see section 2.3
above) than using all classroom time on language items. This concurs with L06, which
defines as one of the competence aims that students should „exploit and assess various
situations, work methods and strategies for learning English‟ (2006b). Michael Lewis
emphasizes the same when discussing teachers‟ roles:

Either teachers must select and teach a restricted lexicon – but on what criteria, for
students of general English? – or they must adapt classroom activities so that, rather
than teaching individual items, they provide learners with strategies which ensure that
learners get the maximum benefit from all the language they meet in and, more
importantly, outside the formal teaching situation. (Michael Lewis 2000:158)

As mentioned above, Woolard also recommends turning students‟ attention to collocations as

a learning strategy. He reports from his teaching experience that in the beginning the teacher
will have to point out useful language in texts or listening and get students to record it. In the
long-run, however, students will be trained to recognize these chunks themselves inside and
outside the classroom, thus becoming more independent learners (Woolard 2000:33-36).

20
2.5.4 Communicative language teaching and the Framework
After Hymes (1972) coined the term communicative competence a great number of methods
and approaches were developed from its premise that knowledge of a language includes the
„competence as to when to speak, when not, and as to what to talk about with whom, when,
where, in what manner‟ (Hymes 1979:15). Communicative language teaching (CLT) is the
name given to these approaches and it has had significant repercussion all over the world.
There are, however, two distinguished versions of CLT, often called the „weak‟ and
the „strong‟ versions. Simensen explains that „the “weak” version implies learning to
communicate in the target language‟ while „the “strong” version implies communicating in
the target language in order to learn it‟ (1998:113). In the strong version the texts used in class
are to be authentic, i.e. „prepared for native speakers of the language‟ (Hatch & Brown
1995:407). In fact, the use of authentic versus constructed texts has been a contentious issue
in the last decades. With an increase in research pointing to the deficiencies of prepared
materials, authentic texts have been given a more central role in language teaching (Simensen
1998:142). In Norway, authentic texts have been recommended for all levels since the 1987
curriculum. Nonetheless, this emphasis has been repeatedly criticized on various grounds.
Howatt raises for example the question of whether native-like models are the most desirable
for teaching, since:

[…] most current uses of the language are for international communication between
people who are not native speakers – a trend that looks likely to continue – and
effective communication in such lingua franca uses of English does not, it would
appear, depend on conformity to native-speaker norms. (Howatt 2004:360)

In a time when English is used in such international contexts, some might criticize the
teaching of collocations, inasmuch as it can imply that students should learn them to sound
like native speakers (see McIntosh et al. 2009:v). However, the aim of teaching collocations is
essentially to show how the language is commonly used so that students can learn to
communicate fluently and accurately (Michael Lewis 2000:177). Thus, learning collocations
should be about communicating effectively and not about sounding like a native speaker.
The use of authentic texts and their value in CLT is also questioned in relation to the
relevance of using concordance lines in the classroom. Widdowson (2000) argues that they
are decontextualized samples of the language and as such have lost their communicative
purpose. On the other hand, as shown by Stubbs (2001b), this does not undermine the value of
corpus consultation in the classroom to identify patterns and routines and show students what

21
is typical in the language. Michael Lewis (2000:196-199) defends the use of authentic
examples from corpora, but warns that they should be selected by the teacher in advance to
avoid overwhelming the learner by the great amount of data and number of unknown words.
Another important feature of CLT is the use of communicative activities in which
students are supposed to complete a certain task. The Council of Europe issued in 2001 the
Common European framework of reference for languages, which gives general guidelines for
the teaching of English as a second or foreign language in Europe (referred to from now on as
the Framework). According to these guidelines, activity is defined as „the exercise of one‟s
communicative language competence in a specific domain in processing (receptively and/or
productively) one or more texts in order to carry out a task‟ (Council of Europe 2001:10),
whereas task is defined as „any purposeful action considered by an individual as necessary in
order to achieve a given result in the context of a problem to be solved, an obligation to fulfil
[sic] or an objective to be achieved‟ (Council of Europe 2001: ibid.). Despite not being much
used in CLT, the term exercise is the one mostly used in the present study due to the type of
material selected. This word is defined by CALD as „a short piece of written work which you
do to practice something you are learning‟ (2005:432).
The Framework has widely influenced both Norwegian textbook writers and L06.
Chapter 5 of the book is particularly interesting regarding CLT and collocations, as it
describes the user/learner‟s competences and shows that collocational knowledge is included
within the linguistic competence. An overview of the main competences is given in table 2.2
below (summarized from Council of Europe 2001:101-130):

Table 2.2 An overview of the user/learner‟s competences in the Framework

User/learner’s competences
General competences Communicative language competences
knowledge of the world linguistic competence lexical lexical elements (table 2.3)
grammatical elements
sociocultural knowledge grammatical
intercultural awareness semantic lexical semantics (table 2.4)
grammatical semantics
others phonological
orthographical
orthoepic
sociolinguistic competence
pragmatic competence

As presented by table 2.2, the general competences comprehend areas like „sociocultural
knowledge‟, „intercultural awareness‟, „knowledge of the world‟, among others, while the
communicative language competences are divided into: „linguistic competence‟,

22
„sociolinguistic competence‟, and „pragmatic competence‟. It is worth recalling that the
Framework contains other subdivisions, which were not included in the table for not
constituting relevant information for the discussion. As shown in table 2.2, the linguistic
competence is subdivided into six different competences as shown in the table. The first one,
the lexical competence, is divided into „lexical elements‟ and „grammatical elements‟ and it is
defined as having knowledge of the vocabulary of a language and being able to use it
(Council of Europe 2001:110). It is within „lexical elements‟ that collocations are included.
Table 2.3 below has been compiled based on the classification of these elements in the
Framework (Council of Europe 2001:110-111):

Table 2.3 Classification of lexical elements compiled from the Framework

Lexical elements Examples

a) Fixed expressions sentential formulae How do you do?

phrasal idioms It’s a long shot
fixed frames Please may I have…?
other fixed phrases such as phrasal verbs to put up with
and compound prepositions in front of
fixed collocations to make a speech/mistake
b) Single word forms tank

Looking at this classification, we can infer that collocations are as important within the lexical
elements as idioms and phrasal verbs, which usually receive greater attention in language
teaching. Howarth, in his study of collocations in academic writing, notes the usual
disproportionate concern with idioms:

Of all categories of conventional lexical combinations, idioms have received the most
attention in linguistic theory and description […]. Idioms have also been almost the
only phraseological category to be recognized in ELT materials (though consider
Rudzka et al. 1981, 1985 and McCarthy 1990). However, they are the least frequent
category in the type of texts under discussion here, and arguably present less severe
problems to learners. Far more significant is the central area of the spectrum –
restricted collocations. (Howarth 1998:169)

Cowie (1998a:212) comments that H. Palmer showed reluctance to the type of idioms such as
put one’s shoulders to the wheel and skate on thin ice – that is, the ones that are structurally
complex and culturally marked. Parallel to Howarth‟s findings, Cowie (1998a: ibid.) reports
that computer-based studies show that restricted collocations are much more frequent in
written texts than these types of idioms.
Collocations are mentioned once more in the Framework within the „semantic
competence‟ in the category of „lexical semantics‟ (see the linguistic competences listed in

23
table 2.2). The elements of this category are compiled from the Framework (Council of
Europe 2001:115) in table 2.4 below.

Table 2.4 Classification of lexical semantics compiled from the Framework

Lexical semantics
Relation of word to general context reference
connotation
exponence of general specific notions
Interlexical relations synonymy/antonymy
hyponymy
collocation
part-whole relations
componential analysis
translation equivalence

Again the Framework shows „collocation‟ parallel to other relations that are usually given a
greater role in textbooks and in the English classroom, as it is the case of synonyms,
antonyms, and translation equivalence. These are practiced in exercises commonly used to
teach learners the meaning of a word; however, they are not the only way to do so. Teaching
students the collocates of a word can aid them in understanding the word meaning and in
communicating more effectively. Hunston claims that „one use of collocational information is
to highlight the different meanings that a word has‟ (2002:76). Earlier in her book she
explains this taking the word shed as an example:

[The meaning of shed] is something like “lose” or “give”, but the precise meaning of
each phrase depends on the collocate:
shed light (on) means “illuminate”, usually metaphorically;
shed tears means “cry” (literally) or “be sorrowful” (crying metaphorical tears);
shed blood means “suffer” or “die”, either literally or metaphorically;
shed jobs and shed staff mean “get rid of people”;
shed pounds means “lose weight”;
in shed skin and shed clothes, shed means “remove”;
shed cents is used to indicate that shares or a currency become reduced in value;
shed image means a deliberate changing of how one is perceived. (Hunston 2002:12)

As we can see from this example, simply teaching students a dictionary definition of shed, its
possible translations, or asking them to match this word with a synonym, would neglect many
of its varied and rich meanings and could give students a false idea that there is one equivalent
of this word in their language. Morgan Lewis (2000:13) uses the synonyms wound and injury
to exemplify the same view. He explains that teaching these words in collocations like stab
wound and internal injuries would be more helpful for students than simply giving them
dictionary definitions. As he affirms, „it is the collocational fields of the two words which

24
reveal the difference of meaning, or rather more precisely, the difference between the ways
the words are used‟ (Morgan Lewis 2000:13, original emphasis). What is meant by
collocational fields here is the different contexts in which a word is used; for instance, some
collocates of injury given by COCA (Davies 2008) are brain, head, knee, and shoulder and of
wound are gunshot, bullet, and stab. This indicates that one of the collocational fields of the
word injury is „parts of the body‟ while a collocational field of wound is „types of injury‟.
Although collocations are recognized as part of the communicative competence of the
learner in the Framework, little or no attention is given to it in Norwegian course material.
One reason for this neglect, as Howarth (1998) concludes in his study, might be the teachers‟
limited knowledge of the phenomenon. He also observes that learners that are unaware of
collocations might come to believe that knowing a language depends on their ability to
syntactically combine items of vocabulary and memorize idioms. However, he deduces from
his results that it is in the middle part of the spectrum, i.e. restricted collocations, where the
difficulty lies and concludes that, „such combinations are not optional stylistic adornments on
the surface of text; they are essential for effective communication […]‟ (Howarth 1998:186).

2.5.5 Culture and communication in word combinations

Culture has had a central place in CLT. When explaining what communicative language
teaching is, Berns stresses the cultural aspect by saying that „culture is recognized as playing
an instrumental role in shaping speakers‟ communicative competence, both in their first and
subsequent languages‟ (Berns 1990:104). Thus, culture shapes the way in which we
communicate. When advocating a communicative competence, Hymes (1979) emphasized the
sociocultural features that influence the way in which people communicate in spoken and
written discourse. Since then, sociocultural aspects have been taken more into account in
language teaching; so much so that the Framework includes these under „general
competences‟ of the user/learner (see table 2.2) and it underlines the importance of
„knowledge of the society and culture of the community or communities in which a language
is spoken‟ (Council of Europe 2001:102). This subsection will deal with the relationship
between culture and collocations and its implications for the teaching of English in Norway.
Certain collocations can reveal changes in society and, as a consequence, changes in
the language. Hunston comments on this matter affirming that „newly emerging collocations
can be used to indicate the growth of new concepts, and changes in the meaning of words‟
(2002:118). She uses Stubbs‟ examples of single parent families and unmarried mothers to
show how new social structures can reflect on the language. Stubbs also studies the change in

25
meaning of the word work from „doing something‟ to „paid social relationship‟, so that
nowadays saying that someone is a woman who works or a working woman does not indicate
that she does housework, but that she has a paid job (Stubbs 1996:177).
In the same article, Stubbs analyzes discourse in the sense of „recurrent phrases and
conventional ways of talking which circulate in the social world, and which form a
constellation of repeated meanings‟ (1996:158) and shows how collocations are used as a
means of propagating pre-fabricated ideas. Some of the collocations he studies are British
heritage, deteriorating standards, and back to basics, which were often used in the 1980s to
convey conservative values. One of these values was the idealization of education in the past
and skepticism to new teaching trends. To research how those ideas were conveyed in the
language, he checked for example the common collocates of trendy in mainstream published
language of that period and found among others: lessons, methods, and teachers. These were
usually found in negative sentences such as „Prince Charles is furious at trendy teaching
which has axed Shakespeare from many schools‟ (Stubbs 1996:162-163, original emphasis).
Thus, observing collocations in the English language today can help us identify important
aspects of British and American culture and current events. We can bring to learners‟
attention for instance the expressions commonly used after the 9/11 attacks in the USA: war
on terror, anti-terror campaign, terrorist networks, counter-terrorist operations, terrorist
attacks, etc.
The idea of identifying sociocultural aspects in word combinations is also underlined
by Teliya et al. (1998) in their study on phraseology and culture. They argue that „linguo-
cultural analysis […] is best suited for phraseology, and especially for restricted collocations.
The latter abound in cultural information and can hardly be described at all as a class of
denominations if their cultural meanings are not taken into account‟ (Teliya et al. 1998:56).
Although their study was on Russian lexical collocations, their methods could be applied to
investigations in any language since it analyzes stereotypes, cultural concepts, connotations,
and conceptual metaphors that are behind certain word combinations. They note for example
the Russian stereotype that women have low intellectual capacity and list a number of
collocations that convey this idea (1998:63). One of the few examples they give of English is
the case of the mule, saying that in English it seems that the mule is the stereotype for
„stubbornness‟, which is shown in expressions like mulish stubbornness and as stubborn as a
mule (Teliya et al. 1998:67). The results found by Teliya et al. are confirmed by some of the
findings in Bartsch‟s study of collocations in the BNC, where she has observed that certain
collocations carry „fixed and stereotyped embedded meanings which are deeply rooted in the

26
cultural background of a linguistic community (e.g. age of consent, affirmative action)‟
(2004:177).
This is an interesting and apparently underexplored area that would be worth
researching. There has been for example some studies on keywords (Said 1978; Williams
1985; Fairclough 1990) and fixed phrases (Moon 1994) in connection with culture, while not
much has been done relating collocations, culture, and language teaching. Although it is
beyond the scope of the present thesis to look into this matter in detail, I have tried to identify
in the material sociocultural aspects that could be brought up in the English classroom.
Besides being relevant regarding CLT, culture is also highlighted in L06. The curriculum sets
under the subject area of „culture, society and literature‟ the following competence aim:
„discuss social conditions and values in various cultures in a number of English-speaking
countries‟ (L06 2006b), which points out the relevance of sociocultural aspects in the teaching
of English in Norway.

2.5.6 Collocation dictionaries

According to Nation, from a vocabulary point of view, research into collocation is needed:

to tell us what the high-frequency collocations are;

to tell us what the unpredictable collocations of high-frequency words are;
to tell us what the common patterns of collocations are […];
to provide dictionaries (or information for dictionaries) that help learners deal with
low-frequency collocations. (Nation 2001: 328)

All the above is possible through the use of corpora. However, in case corpus consultation is
not possible in the classroom or students have difficulties in accessing the corpus directly,
collocation dictionaries are an option. This type of dictionary is usually based on large
corpora, including learner corpora, and provides invaluable information to learners. Three
well-known collocation dictionaries which are currently available are The BBI dictionary of
English word combinations (Benson et al. 1997), The LTP dictionary of selected collocations
(Hill & Lewis 1997) and the most recent Oxford collocations dictionary for students of
English (McIntosh et al. 2009). Such dictionaries can aid students in expressing their ideas
more accurately, both in speaking and writing. One of the competence aims of L06 is the use
of monolingual dictionaries (L06 2006b), so introducing collocation dictionaries in
Norwegian schools could help enhance students‟ vocabulary knowledge.

27
2.6 Corpus linguistics
Hidalgo et al. state that „corpus linguistics has played a powerful role in language research,
grammar construction, dictionary making, natural language processing, cognitive studies, and
language learning and teaching, among other fields‟ (2007:ix). This section will present some
research in the area of language learning and teaching, especially in relation to phraseology,
teaching materials, authenticity, English for specific purposes (ESP), and learner corpora.
Some corpus limitations will also be mentioned.
A ground-breaking effort to bring corpus use to the classroom has been the work of
Tim Johns (e.g.1994) on data-driven learning. The idea is to guide learners in using corpora
to discover features and patterns in the foreign language similarly to how corpus linguists do
(Bernardini 2004:16). This changes the focus from deductive to inductive learning and affects
students and teachers‟ roles so that teachers become more like facilitators and learners
become more responsible for their own learning. Other authors that favor this approach are
Seidlhofer (2000a), Kennedy & Miceli (2001), and Chambers & O'Sullivan (2004).
Corpus linguistics has also been widely used in phraseology to analyze lexical
patterning in language (Howarth 1998; Cowie 1998a; Stubbs 2001a). In relation to language
teaching, Meunier & Gouverneur (2007) say that many authors have pointed up how crucial it
is to address multi-word units (e.g. Michael Lewis 1993; Nation 2001), but they observe that
studies on the treatment of these units in textbooks are rare (2007:121). In face of this lack of
empirical material on the matter, they decided to investigate whether research findings in SLA
have affected the design of five British EFL textbooks and found that the editors of these
books are in fact aware of the importance of teaching phraseology. However, it seems that
more research is needed to detect how much the findings in corpus linguistics have affected
Norwegian EFL textbooks.
Another aspect of corpus linguistics related to teaching materials is on the question of
authenticity. It is argued that textbooks that present invented examples and have descriptions
based on intuition should instead use corpora as resources of „real life language data‟, which,
according to Collins (2000:52) has been happening gradually. Besides providing real life
examples, corpora findings can also be used to overrule many of the traditional grammatical
explanations presented in textbooks, as shown by a study carried out by Clear (2000:19-30).
Johansson (2007) has also described how corpora can be used in teaching, emphasizing the
role of the learner as a researcher and as such being able to have a more critical approach to
textbooks.

28
 Another area of corpus linguistics research is on the use of small corpora for teaching
ESP (see Gaviolli 2005). By using specialized corpora (e.g. medical articles or business
reports), specific expressions and technical terms commonly used in different fields can be
pointed out to students. Learner corpora are another type of corpus that has brought new
insight into language teaching and have been used to produce pedagogical material (e.g. the
OCD).
However, using corpora also has limitations, and it is essential to be aware of what a
corpus can and cannot do. Hunston lists what she sees as some of its main restrictions:

(1) A corpus will not give information about whether something is possible or not,
only whether it is frequent or not. (2) A corpus can show nothing more than its own
contents. […] Thus conclusions about language drawn from a corpus have to be
treated as deductions, not as facts. (3) A corpus can offer evidence but cannot give
information. […] The corpus simply offers the researcher plenty of examples; only
intuition can interpret them. (4) Perhaps most seriously a corpus presents language out
of its context. […] These factors all show the need for a corpus to be one tool among
many in the study of language. (Hunston 2002:22-23)

Since a corpus is used in the present study as part of the material and methods, these
constraints have been taken into consideration. Nonetheless, despite its limitations, a corpus
should not be discarded as a source of information about a language, even more importantly
when looking for what is common usage. As Sinclair states, „one of the principle uses of a
corpus is to identify what is central and typical in the language‟ (1991:17).

2.7 A summary: Vocabulary teaching, collocations, and corpora

This chapter has attempted to show how vocabulary teaching, collocations, and corpora are
related, emphasizing their importance for Norwegian students at the upper secondary school
level. Since CLT, the Framework, and cultural awareness have been central to language
teaching in Norway, they have also been given attention in the present chapter. Other topics
which give a background to the remainder of this thesis are foreign language vocabulary
acquisition and language teaching and learning.

29
30
3. MATERIAL AND METHODS

3.1 Introduction
This chapter will first introduce the primary material selected for investigation, i.e. Passage
and @cross, and the criteria for the choice of this material will be presented. Next, a detailed
description of the two textbooks and the corpus will follow. Finally, the methods that have
been used to achieve the aims will be explained and some limitations of the material and
methods will be discussed.

3.2 Material
The core material used in the present study is vocabulary exercises from @cross workbook
(Rodgers et al. 2006b) and from the Passage website (Sørhus et al. 2006b), both of which are
used in the first year of upper secondary school in Norway.
The textbooks @cross and Passage are written for pupils in the studieforberedende
program, which is translated as „programmes for general studies‟ in the English version of the
L06 (2006b). As the name indicates, the English taught in this course is more general,
different from programs such as „health and social studies‟ in which the vocabulary is more
specific to their line of work. Although I consider knowledge of collocations extremely
relevant for the teaching of ESP, I have chosen to research these in connection with the
general studies program since it is the one that I am currently teaching.
The textbooks usually adopted for this course in the upper secondary schools are the
two mentioned above in addition to Targets (Berntzen & Bårtvedt 2005) and Experience
(Heian et al. 2006). All the four books meet the requirements of the most recent English
curriculum4 (L06), which provides a list of competence aims to be achieved by the end of
certain school years; among them, the first year of upper secondary school (see appendix 1).

3.2.1 Criteria for selection of textbooks and exercises

The first criterion established for choosing the textbooks was that the vocabulary exercises in
the books should be presented separately from the grammatical ones. This can be challenging
since, as Sinclair & Renouf maintain, „it is almost impossible to teach grammar without in
passing teaching some vocabulary. Vocabulary fleshes out the structures, introduces variety
and promotes practice of the structure in question‟ (1988:143). However, considering that the

4
Although Targets’ edition is from 2005, it has been used in Norwegian schools until 2009 when a new edition
was published. It has probably been adopted for so long for meeting many of the requirements of L06.

31
present study is concerned with lexis, the exercises used should work essentially with
vocabulary. The @cross material, for example, is very clear in its categorization, placing the
lexical and grammatical exercises in different sections („building vocabulary‟ and „learning
grammar‟ respectively), which is why @cross was chosen. Below is an example of a
vocabulary exercise from @cross.

190 Write a paragraph describing American teenagers and their situation using the following
words.
Sports, prom, weird, wired, bullying, committed, frustrated, happy, shopping, trend, cliques
Add other words from the texts which you think fit in.

Figure 3.1 Example of a vocabulary exercise from @cross (Rogers et al. 2006b:111)

Passage, Experience, and Targets have vocabulary and grammar exercises mixed, which
posed a problem for deciding on the second book. The solution was to look at their websites
to see if the vocabulary exercises were clearly labeled as such there. Having found that in all
three websites they were categorized as working on vocabulary, another criterion was needed
as a means to select only one of them. Searching the three websites, it was found that Targets
and Experience have too many vocabulary exercises, and it would not be possible to analyze
all of them due to time restrictions. Therefore, the ones decided upon in the end were Passage
and @cross. It is worth noting that only two textbooks have been selected in order to limit the
scope of the study.
After choosing the two textbooks, certain criteria have been adopted to select the
vocabulary exercises. Since the first aim involves finding out if words that are usually taught
as separate items in textbook exercises are part of collocations in the texts they are taken
from, the exercises should present lists of single decontextualized words, like the one from
@cross in figure 3.1. Thus, collocation and gap-filling exercises, as exemplified respectively
in figure 3.2 and figure 3.3, were excluded.

Figure 3.2 Collocation exercise from Passage (Sørhus et al. 2006b)

32
Figure 3.3 Gap-filling exercise from Passage (Sørhus et al. 2006b)

It is necessary at this point to highlight that the two types of exercise above are highly
recommended by several authors for vocabulary practice (Brown 1974; Stevick 1976;
Nattinger 1988). Therefore, it is unfortunate that both textbooks offer remarkably few of these
in its vocabulary section and that most of its exercises work with words in isolation.
Regarding only collocation exercises, @cross includes none while Passage has only three.
Nation explains that teaching vocabulary using lists of words out of context is probably
effective for „learners with a small vocabulary who wish to go on to academic study in a few
months‟ time‟ (1990:2); which is not the case of Norwegian students at the level in question.
At the first year of upper secondary school, students have been studying English for ten years
(they usually start in first grade), so they have probably acquired a sizable vocabulary, and
they still have at least two years until entering university.
Vocabulary exercises that work with proper names and names of geographical places
or events (shown in figure 3.4 below) were also excluded. The reason is that the objective of
these activities seems to be more to remember factual information from the texts than to test
general vocabulary knowledge.

101 Word Search

Find the words in the grid.
Arsenal, Covent Garden, Oxford Street, Big Ben, Ealing, Parliament, Camden Market, Globe, Tower
of London, Chelsea, Greenwich, Tube, City, Heathrow Airport, Westminster, Downing St,
Buckingham Palace, Soho.

Figure 3.4 Exercise with geographical names from @cross (Rodgers et al. 2006b)

33
After excluding the types of activities discussed above, there were exactly fourteen single
word vocabulary exercises from each material, all of which were used in the present study.
Table 3.1 below shows the types of exercises selected and the number of each type.

Table 3.1 Number and types of vocabulary exercises from each material
Type of exercise @cross Passage Total
Matching synonyms, antonyms, or words and their definitions 4 13 17
Involving Norwegian translation 6 1 7
Using certain words to write a text 3 - 3
Finding words in grids 1 - 1
Total 14 14 28

All of these types of exercises, except the one to write a text, belong to the fourth step in the
vocabulary acquisition process mentioned by Hatch & Brown (1995; see section 2.3); namely,
consolidating word form and meaning in memory. The type in which students need to write a
text with certain words requires that students use the words, which is the last step in the
acquisition process. More examples of the types of exercise that were selected for
investigation will be given in the @cross and Passage subsections below.

3.2.2 @cross
The @cross material consists of:

textbook
workbook
teacher‟s book
8 CDs
website

For the purposes of this thesis mainly the workbook and the textbook have been used. The
material was developed by Drew Rodgers, Knut Skifjeld, Dilys Brown, & Martin Mulloy and
published by Damm. @cross targets Norwegian pupils specifically, often contrasting Norway
with the English speaking world. Regarding the variety of English used, it includes both
American and British English, depending if the text is about the USA or the UK. The book
also brings other varieties like Australian, Jamaican, and South African English.
@cross is divided into four parts: „@cross the UK‟, „@cross the English speaking
Countries‟, „@cross the USA‟, and „@cross Literature‟. Part 1 is subdivided into eleven
chapters that cover different regions of the United Kingdom, including cultural and historical
information about them. Part 2 has only three chapters, which talk about Jamaica, South
Africa, and New Zealand. Part 3 contains twelve chapters that describe various states and

34
cities in the United States, and provide historical and cultural facts. With reference to genres,
it is stated in the preface of @cross textbook that „you will find a combination of two teaching
traditions – storytelling and factual articles‟ (Rodgers et al. 2006a:5). Thus, the texts in each
of these three parts are in the form of storytelling, with young characters experiencing the
English speaking world. At the end of each chapter, there is a section called „fact files‟ which
gives extra factual information about what was presented in that chapter. The authors explain
that the idea is to give pupils the opportunity to choose the type of text that suits them best,
taking into consideration students‟ different learning styles (Rodgers et al. 2006a: ibid.). The
other kinds of genre the book includes appear in Part 4 of the textbook. This last part is
concerned with literary texts ranging from the 1500s until today, as required by the new
curriculum. Whereas the texts in the three first parts are written for the purposes of teaching
English to Norwegian students (i.e. constructed texts), the texts in the literature section are
authentic (i.e. written for English speakers) and taken from a variety of genres, including
poems, short stories, and excerpts from novels.
The textbook contains only texts, while the workbook and the website provide all the
exercises. The authors state that they „have included tasks covering communicative, cultural,
numeric and digital skills‟ (Rodgers et al. 2006a:6). These kinds of activities are grouped in
each chapter in the workbook under the headings: „geography‟, „fact file‟, „the journey‟,
„learning grammar‟, „building vocabulary‟, „reading skills‟, „writing skills‟, „speaking skills‟,
„numeric skills‟, „society and culture‟, „making a presentation‟, „listening skills‟, and
„wordbank‟. All the vocabulary exercises selected in the present study are in the „building
vocabulary‟ section, which contains from one to three exercises in each chapter.
Since all the activities are in @cross workbook, this is my main source of material. The
vocabulary exercises on the website are not clearly categorized and there are very few of
them, which is why the website has not been used. In total, fourteen vocabulary exercises
have been selected and they are mainly of the types:

1. Correlating English words with Norwegian translations

2. Translating a list of words
3. Linking English words with their synonyms or antonyms
4. Matching words with their definitions
5. Using a range of words to write a paragraph on a determined topic
6. Finding words in a grid (only those that provided a list of English words to be found
have been chosen; vocabulary exercises with grids but no lists were excluded.)

35
This list is more detailed than the one in table 3.1, where the categories were generalized. As
stated in the criteria subsection, what all of these exercises have in common is that they
present lists of single words to be worked with. A few of them had a collocation or two, in
that case the collocation was not searched in the corpus. Below is an example of type 3:

36 Match the word in the left hand column together with the word which means roughly the same,
from the right hand column. If you need help, look at pg [sic] 41 - 44 in @cross Textbook. There
you will find the words used.
1. coastal a. national costume
2. apply to b. proof
3. generally c. along the seaside
4. stunning d. normally/commonly
5. kilt e. outskirts
6. soundtrack f. 100 years
7. century g. seek admission
8. preserved h. fabulous/fantastic
9. suburbs i. taken care of/looked after
10. evidence j. music to set the mood

Figure 3.5 Exercise to match synonyms from @cross (Rodgers et al. 2006b:34)

3.2.3 Passage
Passage is, as the author expresses in the preface, „an all-in-one book with both texts and
activities‟ (Sørhus et al. 2006a:7), so the material is formed only by:

textbook
CDs
website

The authors are Theresa Bowles Sørhus, Richard Burgess, & Trond Christian Anvik and the
material is published by Cappelen. Passage is also a book written for Norwegian pupils and it
embodies both American and British varieties of English. The new curriculum has had
significant consequences for this latest edition of Passage, as stated in its preface (Sørhus et
al. 2006a:6). What is distinct from the previous edition is that to complement the material on
the USA and the UK, the authors have added literature and information about other English-
speaking countries, native peoples of the English speaking-world, and current global issues.
Parallel to what we find in @cross, the writers have incorporated literary texts from the 1500s
up to modern times; the difference is that these are included between chapters and not at the
end of the book.
The Passage material is divided not by chapters but by general themes and each of
these includes several texts. The themes/sections are:

36
 Starting Out
Global English
On the Road
We Were Here First
The Road to Freedom
The Power and the Glory
Global Village

These sections contain texts written specifically for Norwegian pupils and authentic texts
taken from novels, short stories, literature, newspapers, etc. After each text there are activities
distributed under the titles: „understanding the text/story‟, „talk about it‟, „improve your
language‟, „role play‟, „write about it‟, and „research‟, among others which work on
mathematical skills. Some texts are followed by all of these types of activities while others by
only a few. All the vocabulary exercises are listed together with the grammar exercises under
the heading „improve your language‟. The exercises on the website are listed under the title
of each text of the textbook and there can be from none to three vocabulary exercises for each
text. These are usually called „pair up words‟, „crossword‟, „find words‟, „find synonyms‟, or
„connect synonyms‟.
As explained previously, the second half of the material is taken from the Passage
website (Sørhus et al. 2006b), which offers a substantial amount of extra material, including
vocabulary exercises. Among these, fourteen were chosen to be analyzed and they are of the
following types:

1. Correlating English words with Norwegian translations

2. Linking English words with their synonyms
3. Matching up words with their definitions

Most of the exercises are to link words with synonyms (see table 3.1), as the one in figure 3.6
below.

37
Figure 3.6 Exercise to match synonyms from Passage (Sørhus et al. 2006b)

In order to learn more about the collocational behavior of the words in the exercises selected,
another kind of material was needed, which will be presented in the next subsection.

3.2.4 The corpus

The corpus chosen for this research is the Corpus of Contemporary American English
(COCA) created by Mark Davies (2008). It is the largest and most recent corpus available
online, being launched in 2008 and comprising more than 400 million words distributed
among more than 160.000 texts. As described in its introductory page, COCA includes „20
million words each year from 1990-2009, and it is equally divided among spoken, fiction,
popular magazines, newspapers, and academic texts‟ (Davies 2008). What also makes it stand
out among all the other corpora is that COCA is updated every six to nine months, and can
therefore be used to record the changes in American English over time. The access to the
corpus is completely free of charge. After trying a few query words, the user is required to
register a username and a password; when logged in, you can track your previous searches
and make your own search lists.
The main reasons for the choice of COCA are its size, its free access, and the search
tools available, which will be explained in detail in the method section. Since the present
study deals with students at upper secondary school level, a corpus that entails the latest
trends in the English language seemed the most appropriate alternative. Some might argue
that the British National Corpus (BNC) would be a better option since British English has
always exerted great influence in Norway, especially because of geographical proximity.
However, although being aware of the prestige of British English in Norway, I should point
out that Norwegian teenagers are under even greater influence of American English on a daily
basis. American bands, films, computer games, and TV-series are massively popular among
teenagers in Norway. Moreover, the BNC contains 100 million words, 300 million less than

38
COCA, and it is mostly composed of texts written in the 1980s and 1990s, which limits the
search and fails to represent English use today. Other important differences between these two
corpora are exemplified in the section „compare wordlists to BNC‟ in the „more information‟
part of the COCA webpage, as follows:

Some differences are related to culture, society, politics, or current events (e.g. Am
Republican, congressional, baseball, Iraqi; Br Tory, parliamentary, Victorian), some
are just different words for the “same” concept (e.g. store/shop, attorney/solicitor,
apartment/flat, mom/mum), and some words in COCA (1990-2009) refer to things that
are too new to have made it into the pre-1993 BNC (e.g. web, Internet, high-tech,
online). (Davies 2008)

Turning to the question of representativeness, every corpus intends to be representative of a

determined population. Large corpora like COCA and the BNC are built to represent the
English language in general, or at least American and British English respectively.
Nonetheless, to be able to make such a claim, these corpora have to follow certain criteria. In
line with McEnery et al., „the representativeness of most corpora is to a great extent
determined by two factors: the range of genres included in a corpus (i.e. balance […]) and
how the text chunks for each genre are selected (i.e. sampling […])‟ (2006:13, original
emphasis). Hunston, however, adds a third factor – its change over time. She claims that „any
corpus that is not regularly updated rapidly becomes unrepresentative‟ (Hunston 2002:30).
Davies has tried to meet these three criteria when designing COCA in that it covers a wide
variety of genres, including spoken data, so it is balanced; the 160.000 texts are evenly
divided among the five genres5 (and again in various subgenres) and among the four time-
periods, so it can be considered a sample of general English use; and it is constantly updated,
showing change over time.
Some might still question how comparable the texts in the textbooks and the texts in
COCA are. Consonant with COCA, the textbooks aim to generally represent the English
language, in particular these that are designed for the general studies program. Although
neither the textbooks nor the corpus are able to show the whole of the English language, they
can still give an idea of how English is being used through their selection of genres and
registers. Even though the textbooks include both American and British English and the
corpus comprises mainly American English, choosing a British corpus would also be
restrictive and using two corpora would be outside the scope of this study.

5
Genre is used here in the same way as it is in COCA, although it might not be the most appropriate term for
this context.

39
3.3 Methods
As described in the introduction, the first aim is to find out if single words presented in
vocabulary exercises are part of collocations in the texts they are taken from. The first
research question is how large this number is and the second is which of these are useful to
teach students. The hypothesis is related to the first research question and it predicts that a
high proportion of collocations will be found. As a way to achieve this aim and address the
research questions, first fourteen vocabulary exercises from @cross workbook and fourteen
exercises from the Passage website were selected; then the words from each of these
exercises were looked for in the texts they were taken from to see if they were part of any
word combinations. Thereafter, these were checked in the OCD (McIntosh et al. 2009) and in
the corpus and only the ones that were listed in both and which presented frequencies higher
than 3 and MI scores higher than 3.0 in COCA were classified as collocations. Next, all the
words which appeared in collocations in the texts were counted and a percentage calculated.
Finally, the competence aims of L06 and research on collocation and vocabulary acquisition
were used to discuss which types of collocations would be most useful to teach at this level.
The second aim is to suggest how the vocabulary exercises selected for the present
study can be complemented based on existing teaching material on collocations and by
applying previous research on corpora and foreign language vocabulary acquisition. Knowing
that textbooks like Touchstone have been designed with the help of a corpus (see McCarthy et
al. 2006a:iv), I have chosen to use information provided by COCA to complement the
vocabulary exercises. Thus, the method was to look up the words from the exercises in the
corpus to find their most frequent collocates with an MI above 3.0. Because of space and time
constraints, only the ten most frequent collocates are listed and only the results for one word
from each exercise are shown6.The dictionary was only used to help select collocations
appearing in the texts (related to the first aim) but not to determine if the ten combinations
found in the corpus are indeed collocations. Thus, following the definition in chapter 1
accurately, these are simply referred to as „combinations‟ or „word combinations‟. The reason
for not using the dictionary in relation to the second aim is that it would be extremely time
consuming and beyond the scope of this study.
The method used to achieve the third aim is purely qualitative. Given that
complementing vocabulary exercises is not the only way collocations can be incorporated into

6
The ten most frequent collocates of all the words from each exercise were searched; however, it was not
possible to include them in the thesis also due to space limitation.

40
teaching, previous research and existing teaching material on collocations have been
consulted to suggest other ways in which this can be carried out. All in all, the results
achieved by using these procedures will hopefully give a rough idea of the range of
collocations that appear in textbooks, show the potential of corpora for teaching, and
emphasize the importance of including collocations not only in vocabulary practice but in
English teaching in general.

3.3.1 The dictionary

One of the methods used in the present study was looking up words in a collocation
dictionary. As mentioned above, the Oxford collocations dictionary for students of English
(OCD; McIntosh et al. 2009) has been used to check whether the word combinations found in
the texts are indeed collocations. This dictionary has been chosen due to the fact that it is the
most recent and comprehensive one, containing around 9.000 headwords, 250.000 word
combinations, and 75.000 examples. This dictionary was first edited in 2001 and this revised
edition has added 100.000 collocations and 25.000 examples besides including a CD-ROM.
The OCD comprises both British and American English and these variations are labeled in the
dictionary when necessary.
The OCD covers thirteen types of syntactical combinations organized under three
general entries: noun, verb, and adjective. Under the noun entries we find: adjective + noun
(bright light), quantifier + noun (a beam of light), verb + noun (cast light), noun + verb (light
gleams), noun + noun (a light source), preposition + noun (by the light of the moon), and noun
+ preposition (the light from the window). Under the verb entries: adverb + verb (choose
carefully), verb + verb (be free to choose), and verb + preposition (choose between two
things). Under the adjective entries: verb + adjective (declare sth safe), adverb + adjective
(perfectly safe), and adjective + preposition (safe from attack) (McIntosh et al. 2009:vi-vii;
original examples). Although the dictionary only provides the type adverb + verb in the list
given in the introduction, when checking the entries it becomes clear that verb + adverb
combinations are also included. Thus, the categories are fourteen in total and they correspond
to the eleven that were listed in chapter 1 in addition to quantifier + noun, verb + verb, and
verb + adjective.
The approach adopted by this dictionary pertains to learners‟ needs rather than to a
particular theory and the editors have taken the following questions into consideration when
deciding which collocations to incorporate (McIntosh et al. 2009:v):

41
 Is this a typical use of language?
Might a student of English want to express this idea?
Would they look up this entry to find out how?

To embrace what is typical, the editors have used the Oxford English Corpus, which
comprises almost 2 billion words, to check for frequency and examples. The second question
led the editors to aim for the kind of language students might need to write, such as essays and
reports; hence, it takes account of non-specialist (science, history, sport, etc.) and specialist
(law, medicine, politics, current affairs, etc.) areas in addition to informal language of the
internet and spoken communication (McIntosh et al. 2009:vi). The third question led to the
focus on noun entries since students are more likely to look, for example, for rain to find
heavy than the other way around.
One downside of using this dictionary is that infrequent words that appear in the
exercises might not be among the 9.000 headwords. Some might argue that this does not
necessarily mean that they are not collocations, which is probably true. However, when not
found in the dictionary, the combinations were excluded from this investigation to meet the
definition established in chapter 1.

3.3.2 MI scores and frequencies in COCA

Since the aims of this study involve finding collocates of certain words and also checking
frequency and strength of collocations, a corpus turned out to be the best method to access
this kind of information. In addition to the advantages discussed in the material section, other
positive aspects of COCA, which will be particularly useful for this study, are its user-friendly
interface and its varied number of tools. As noted in the introductory text on the website, „the
corpus allows you to search for exact words or phrases, wildcards, lemmas, part of speech, or
any combinations of these‟ (Davies 2008). Some of these tools are illustrated in figure 3.7 in
subsection 3.3.3.
One of the important features of COCA is the possibility to measure Mutual
Information (MI) scores. The MI shows how closely the words in a word combination are
connected, i.e. it calculates their strength. Drawing from Fano (1961), Church & Hanks give a
more detailed definition of MI:

If two points (words), x and y, have probabilities P(x) and P(y), then their mutual
information, I(x, y), is defined to be

42
 Informally, mutual information compares the probability of observing x and y together
(the joint probability) with the probabilities of observing x and y independently
(chance). If there is a genuine association between x and y, then the joint probability
P(x, y) will be much larger than chance P(x) P(y), and consequently I(x, y) 0.
(Church & Hanks 1990:23; original emphasis)

Thus, according to this explanation, the higher the MI the stronger the combination is. Later
in the same article, Church & Hanks note that „as a very rough rule of thumb, we have
observed that pairs with I(x, y) > 3 tend to be interesting, and pairs with smaller I(x, y) are
generally not‟ (1990:24). This means that pairs with scores above 3.0 can probably be
considered collocations and below that, free combinations. Similarly, it is pointed out in the
help section of COCA that „typically, scores of about 3.0 or above shows [sic] a “semantic
bonding” between the two words‟ (Davies 2008). On these grounds, I have used the 3.0
threshold to exclude free combinations. Illustrating how the MI works in practice, if we look
for the collocates of the word fancy, we will find that fancy restaurants has an MI of 7.08,
while fancy house has an MI of 2.21, which tells us that the word restaurants has stronger
bonds to the word fancy than the word house.
However, according to Hunston (2002:72), the strength of a collocation can not always
guarantee its relevance. McEnery et al. affirm that „collocations with high MI scores tend to
include low-frequency words‟ (2006:57), and infrequent collocations do not appear
particularly interesting for language teaching. This is why the tables with the ten most
common combinations are listed according to their frequency and not MI. When the
collocates of fancy, for example, are listed in COCA in order of frequency, as displayed by
table 3.2 below, the top combinations are fancy cars, fancy restaurant, and fancy dress; on the
other hand, when shown in order of strength, the top ones are fancy air-gulper, fancy
antiseptics-soap, and fancy car-of-the-week. This clearly elucidates why MI should not be
used alone to determine what can be considered a collocation or what is relevant for teaching.

Table 3.2 Collocates of the adjective fancy in COCA

Collocates Raw frequency MI
cars 94 7.51
restaurant 79 7.73
dress 65 7.55
footwork 63 13.74
clothes 62 7.22
name 57 5.12
way 55 3.13
restaurants 48 8.08
car 43 4.82
hotel 40 6.31

43
Hence, preferably strong but also frequent collocations should be taught or students would be
learning language items they would probably never use, which is why the raw frequency (i.e.
the actual count of elements) of the combinations in the corpus is also checked. The use of
raw frequencies might, however, be problematic and raise questions like: What should be the
bottom line for a combination to be classified as a collocation? How can one determine if a
certain number of occurrences is frequent or not? Clear affirms that „the terminology is not
well established yet, and it is by no means clear whether the observation of a single co-
occurrence in a corpus should be ignored or whether the single instance should be taken to be
one of many more which might have occurred‟ (1993:277). He decides then to adopt the
threshold value of three, discarding pairs appearing fewer than three times in the corpus. It has
been done likewise in the present study so that when a combination found in the textbook
presents a raw frequency lower than three it is not considered a collocation. The issue of what
number can be considered frequent, on the other hand, is only solved by using statistical tests
like t and z scores. Clear does not rely only on frequencies, applying also MI and t scores in
his search. In the present study, however, only the measures established in the definition of
collocations are used and the collocations have been generally commented on in terms of
usefulness for students and not in terms of frequency. The reasons why other statistical tests
have not been employed are discussed in section 3.4 at the end of this chapter.

3.3.3 Description of corpus tools

This subsection will describe how the search in COCA has been carried out. At the same time,
some of the corpus tools and technical terms will be explained.
The node is the term used to designate the word under investigation (Sinclair
1991:175) and span is the distance between the node and its collocates. The span of a word
may vary considerably, ranging from an adjacent position to a gap that might even go over
sentence boundaries (Carter & McCarthy 1988:34). The terminology used to define this exact
distance between the main items of a collocation or combination is gram. Thus, when the
node and its collocate occur as an adjacent pair they are called bigram; when there is a word
in between, they are called trigram or 3-gram; when there are two words in between they are
called 4-gram; and so forth. For instance, an adjective plus noun bigram can be illustrated by
the collocation heavy rain; a verb plus noun 3-gram would be a collocation like sips his
coffee, while sipping his decaf coffee would be a 4-gram collocation.
As to give a better idea of how the search tools of COCA work, figure 3.7 shows what
the search interface looks like followed by an explanation of each element:

44
 Figure 3.7 Search interface of COCA (Davies 2008)

Describing figure 3.7 from the top, the results of a query can be displayed in a chart or in a
list. You choose „chart‟ if you want to compare the use of a word in different genres (spoken,
fiction, magazine, newspaper, and academic) and year spans (1990-1994, 1995-1999, 2000-
2004, and 2005-2009). If you want the corpus to show only the number of occurrences (raw
frequency) of a query word, you should choose „list‟. You can also choose „compare words‟
to contrast the uses of any two words in the corpus. Moving to the „search string‟ area, there is
the „word(s)‟ field where the node word should be written. It is also possible to look for
common collocates with that word and choose from which part of speech you want these
collocates to be; to do so, you click on „POS list‟ (part of speech list). As figure 3.7 shows, the
noun collocates of fancy were looked for, so by choosing „noun.ALL‟ in the POS list, the
label (or tag) [nn*] appears automatically in the „context‟ field7. When you press „search‟ the
corpus lists the most frequent nouns used with the node or query word, in this case, fancy. (It
should be noted that for verbs I have used „verb.LEX‟, i.e. lexical verbs, to avoid auxiliary
verbs like be, have, and do). The boxes showing „0‟ and „1‟ enable you to choose the distance
between the query word and its collocates, i.e. the grams, and in which side you want the
collocates to appear (you can have up to nine words on either side). This tool has been
employed to find the ten most common combinations with the words from the vocabulary
exercises. The result of the search illustrated in figure 3.7 is shown in table 3.2 above.
Under the „POS list‟ there is a label called „random‟. This tool is meant to help those

7
The tagset used by COCA is the same as the one used by the BNC (see The BNC tagset in the reference list).

45
who are unfamiliar with the corpus, giving an idea of how an entry can be written and how the
results look like. Below „random‟ you can choose to see the results in sections, which shows
the same information as the chart, but without the graphs. You can choose to see one or two
words in different genres and subgenres (e.g. you can choose education, history, humanities,
etc. within the academic genre), year span, or a specific year. In the „sorting and limits‟ area
you can choose to sort the results by relevance (MI), frequency, or alphabetically. You can
also determine the minimum of words the corpus should display and the minimum MI score.
This tool has been used when looking for the ten most frequent combinations of a word to
select only the ones with MI scores above 3.0.
When needing to check only the frequency of a collocation found in the text, I wrote
the node word from the exercise in the „word(s)‟ field and its collocate in the „context‟ field,
so in this case the POS function was not used (to add the MI, I chose the option „relevance‟ in
the „sorting‟ field). The corpus offers other search tools which are designed for more detailed
searches. Giving that they are not relevant to this study, only the most general ones have been
described.

3.3.4 Criteria for corpus search

As a means to look for the ten most frequent combinations with the words from the exercises
in the corpus, a few criteria had to be established. Some of the questions that needed to be
addressed were: (1) If the exercise includes for example the verb look, should the corpus give
results for looking, looks, and looked or only the infinitive look? (2) How many words to the
left and to the right of the node look should the corpus find? (3) To what part of speech should
the collocates of look belong to?
Addressing question (1), the definition adopted in this study states that a collocation is
a combination of one or more lexemes so that various forms of the same word can be included
(to find lexemes in the corpus the words have to be written in square brackets). For example,
to find the most frequent collocates of boundary, the lexeme [boundary] is written in the
„word(s)‟ field and the POS for lexemes of lexical verbs (verb.LEX) is chosen. Then [vv*]
appears automatically in the context field. This search gives results like: cross the boundary,
crossed the boundaries, and crossing the boundary, which include form variations of the node
and collocates.
When looking for combinations in the corpus, something that should also be decided
upon was how many words should be allowed between the collocates, i.e. if the combination
should be a bigram, a 3-gram, a 4-gram, etc., as explained above. This choice depended

46
directly on which part of speech the node belonged to. Thus, answering questions (2) and (3),
the following criteria were established concerning grams and parts of speech:

Table 3.3 Criteria for finding combinations in the corpus

Node Category Abbreviation Gram Example
noun adjective plus noun Adj+N bigram social issues
e.g. issue verb plus noun V+N 4-gram discuss an issue
noun plus noun N+N bigram security issues
noun plus preposition N+P bigram issue concerning
preposition plus noun P+N bigram at issue
noun plus verb N+V bigram the issue arises
adjective adjective plus noun Adj+N bigram weary eyes
e.g. weary adverb plus adjective Adv+Adj bigram increasingly weary
adjective plus preposition Adj+P bigram weary of
transitive verb verb plus noun V+N 4-gram glanced at the clock
e.g. glance adverb plus verb Adv+V bigram quickly glanced
verb plus preposition V+P bigram glanced at
intransitive verb verb plus adverb V+Adv bigram vanished completely
e.g. vanish adverb plus verb Adv+V bigram simply vanished
adverb adverb plus adjective Adv+Adj bigram generally accepted
e.g. generally adverb plus verb Adv+V bigram generally speaking

The decisions relating to grams were based on how many words usually come between the
two main items, if any. Hence, looking at the table above, we can see that in most cases
bigrams were looked for since most of those lexemes are found adjacent to each other (e.g.
there are no words appearing between weary and of). In the case of verbs, a distinction had to
be made between transitive and intransitive verbs since the former takes a direct object and as
such, it is common to be found in a V+N combination (e.g. glanced at the clock); the latter,
on the other hand, is most commonly found in V+Adv combinations (e.g. vanished
completely). Explaining the gram choice, the reason why the N+V combinations are 4-gram
and not bigram like the others is that the direct object might include determiners, prepositions,
adjectives, or other types of words (e.g. discuss an important issue). Summarizing, the
following criteria have governed the search:

Table 3.4 Types of combinations looked for in the corpus

Node Types of combinations Grams
nouns Adj+N N+N bigram
adjectives Adj+P V+P
verbs P+N V+Adv
adverbs N+P Adv+V
N+V Adv+Adj
V+N 4-gram

As mentioned in chapter 1 and summarized in table 3.4, eleven categories were set up for the
search in the corpus. The criteria above were established to organize and facilitate the search

47
in the corpus and it applies mainly to the search for the ten most common collocates with the
exercise words. When encountering a collocation in the textbooks that combined some of the
elements above, fit the definition established, and seemed useful for teaching, such as get
good grades (V+Adj+N), that collocation was also included.

3.4 Limitations of the material and methods

Although the material and methods used in this study were chosen carefully and under
established criteria to achieve certain aims, they do present some limitations. Firstly, the use
of the textbooks Passage and @cross is not as widespread as that of Targets and Experience.
Granted that I could have looked for this information in other ways, I have decided to use the
Norwegian search engine Kvasir on the internet (www.kvasir.no; accessed on November 12,
2009) to find out what English textbooks are most used in the schools. The websites of the
first sixty upper secondary schools that were listed when entering the words boklista VG1
were checked, and their book lists examined to see what English textbook was adopted for the
first grade in the general studies program. All of the sixty schools adopted one of the four
books mentioned before, so I simply counted them and calculated the percentages. The results
are presented in table 3.5 below:

Table 3.5 Textbooks used in a selection of 60 Norwegian schools

English textbooks Percentage
Targets 40%
Experience 30%
Passage 21.6%
@cross 8.3%

As table 3.5 shows, Passage and @cross are the ones least used among the first sixty schools
found by Kvasir, while Targets and Experience together comprise 70% of the schools. Based
on these (approximate) numbers, some might claim that this study is not so across-the-board,
and hence not so relevant in the Norwegian upper secondary school scenario. However, this
thesis does not aim to encompass all books used in Norwegian schools, but to analyze some of
them. Thus, the number of schools adopting these books is irrelevant as long as the books are
still in use.
With regard to the corpus, there is an advantage that can turn out to be a limitation –
namely that COCA is being updated, as remarked above, every six to nine months.
Consequently, it is very likely that if someone tries to search for the same information a year
after it was retrieved frequencies and MI scores will not be exactly the same. Although the

48
data collection started in 2008, all the results have been updated after the last revision of the
corpus in the summer of 2009. However, there have probably been some adjustments a while
after since some of the MI results are now slightly different in the corpus, despite the
frequencies being the same. Another constraint pertaining to the corpus is the fact that
whereas it is taken to represent the English language, it cannot present a completely faithful
account of it, given that the corpus does not and cannot contain all the words that exist in the
language. As McEnery et al. explain, „a corpus is essentially a sample of a language or
language variety (i.e. population)‟ (2006:13; original emphasis). Thus, some words or
collocations that are used by English speakers might not be found in the corpus, which can
happen also due to the fact that COCA comprehends mostly American English.
As noted earlier in this chapter, using raw frequencies without other statistical
measures might be arguable since they are independent of the size of the corpus. McEnery et
al. (2006:56) suggests two other statistical tests besides MI to identify significant collocations:
the t and z scores. The t test takes corpus size into account but in contrast with MI, it tends to
play up high-frequency pairs. A combination of MI and t-scores would then be the ideal, so
that the intersection of the two would give good evidence for what is or is not a relevant
collocation. Be that as it may, using statistical tests would be outside the scope of this thesis
considering the great amount of calculations needed. Furthermore, applying statistical tests on
very large frequencies, such as the ones from COCA, is a non-trivial task with many pitfalls.
Bartsch (2004:103) has observed in her study of collocations in the BNC that the t-test
highlights collocations with function words (such as phrasal verbs, which are not interesting
for the present study) whereas the MI scores highlight content word collocations, also called
lexical collocations. As Gabrovšek (1998:132-133) notes, this type of collocations is more
unpredictable to students and the kind that is probably the most interesting for language
teaching. In addition, McEnery et al. say that „it is possible to use raw frequency (i.e. the
actual count) where no comparison between corpora is necessary […]‟ (2006:52). Since no
corpora comparison is required in the present study, the raw frequency results combined with
the MI can be used as an aid to determine if a combination is indeed a collocation.

49
50
4. RESULTS

4.1 Introduction
This chapter will present the results from the investigation outlined in chapter 3 and a brief
explanation will follow each exercise and table as they are introduced. These explanations
have been included for matters of layout and convenience since the results constitute a too
large amount of data to leave to comment on at the end. The interpretation of the results will
be presented in chapter 5.

4.2 Presentation of the data

In the next two sections, the exercises taken from @cross workbook and the Passage website
are presented exactly as they appear in the book and website. Each of them is followed by a
table with the exercises‟ words displayed in alphabetical order together with their respective
collocation(s) found in the texts (this type of table will be referred to as „collocation table‟).
The collocations are written exactly as found in the texts (e.g. marks the boundary in exercise
[1]) and the lexeme variations given by the corpus (e.g. marking the boundary, marked the
boundary, etc.) are listed underneath. The total frequency of the collocation is given in
parenthesis (see table 4.1 for an example). In case there were no collocations with a certain
word, this word was left out of the table.
As mentioned in chapter 3, all the words from each exercise were also entered in the
corpus to find their ten most common collocates. It should be kept in mind that these tables
list the most frequent collocates which present an MI above 3.0. Nonetheless, to avoid being
too repetitive, this information is not included in the table heading. Due to space constraints,
the results for only one of the words from each exercise will be shown. This type of table
(which will be called „collocate table‟) is placed after the collocation table, so that each
exercise will be followed by two tables. The choice of the words displayed is not arbitrary.
They have been picked depending on if they are part of interesting combinations with respect
to how corpus information can complement the exercises and to how the sociocultural aspects
that they convey can be explored in class.

4.3 Results from @cross

In this section the fourteen exercises selected from @cross workbook are presented. The
exercises are alluded to by their given number (1-14) and not their number in the textbook.

51
@cross has a total of fifty-six vocabulary exercises (listed under the „building vocabulary‟
section of the workbook). Here are the types and number of most of them:

translation (16)
gap-filling (9)
matching words to their synonyms, antonyms, or definitions (5)
unscrambling words (3)
working with other languages (Patois, French, etc.) (3)
writing a text using certain words (3)

The other various categories (crosswords, odd-one-out, find words on a grid, play with words,
etc.) contain only one or two exercises each, comprising a total of seventeen exercises. None
of the fifty-six exercises work directly with collocations and most of them present essentially
isolated words. As shown in table 3.1 and explained in chapter 3, the fourteen exercises
selected fall within three of the categories above (translation; synonyms, antonyms, or
definitions; write a text with certain words) in addition to one more in which students have to
find words in a grid. All the texts on which these exercises are based are constructed and not
authentic.
Below is the first exercise selected from @cross.

[1] 22 - Berwick pg8 36-37

Having read the text on Berwick, find the English words corresponding to the following
Norwegian words by drawing lines:
1. nylig a. attractive
2. grenselinje b. trades
3. omgi c. recently
4. forbløffende d. excellent
5. arbeidsløshet e. medieval
6. tiltrekkende f. community
7. i følge g. surround
8. middelaldersk h. unemployment
9. utsøkt i. amazing
10. næringsvei j. boundary
11. samfunn k. according to
(Rodgers et al. 2006b:28)

This is an example of an exercise matching vocabulary items in Norwegian with English

translations. The text from which these words were taken describes a town called Berwick and
it is in Part 1 of the textbook („Across the UK‟), in chapter 2, „Arrival in England‟. Table 4.1
below shows the English words from exercise [1] and, as stated above, those without any
collocations in the text were left out of the table. The words listed as not in the text do not

8
@cross seems to use pg as the abbreviation for „pages‟ when the correct use should be pp (see CALD
2005:987).

52
appear in the Berwick text, although the heading affirms that they do. As seen from the table,
the V+N (4-gram) collocations have been written like in mark_boundary to indicate that other
words appear between the node and the collocate.

Table 4.1 Collocations from exercise [1] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
boundary marks the boundary V+N (130)
- mark_boundary 42 3.35
- marks_boundary 36 5.17
- marked_boundary 28 4.56
- marking_boundary 24 6.69
community not in the text
excellent not in the text
medieval medieval art Adj+N (54)
- medieval art 49 5.52
- medieval arts 5 3.61
recently not in the text
surround not in the text

As table 4.1 shows, some collocations are stronger and more frequent depending on the form
of the verb or if the noun is plural or singular, which is the case of medieval art (MI 5.52,
frequency 49) and medieval arts (MI 3.61, frequency 5), and also the case of the forms of
mark the boundary. When any of the various forms of a collocation presents frequencies
lower than 3 and/or MI scores lower than 3.0, they are still displayed in the table as long as
one of its forms fits the category of collocation (e.g. if medieval arts appeared only twice or if
its MI was 1.5, it would still be included because medieval art presents results above the
threshold). Table 4.2 displays boundary‟s ten most frequent 4-gram verb collocates (V+N).
According to the table, cross and push are the most frequent verb collocates of boundary.

Table 4.2 Collocates of the noun boundary in COCA

Collocates Raw frequency MI
cross 157 6.23
set 113 3.16
crossed 74 6.22
crossing 66 6.61
pushing 65 5.68
push 65 5.68
transcend 60 9.02
crosses 57 7.24
define 53 5.84
draw 46 4.68

Exercise [2] below is also a matching exercise but of the type students have to correlate
words with their definitions/synonyms. The text where the words (on the left) are taken from
is in Part 1, chapter 3, „To the North – Scotland‟.

53
[2] 36 - Match the word in the left hand column together with the word which means roughly the
same, from the right hand column. If you need help, look at pg 41-44 in @cross Textbook. There
you will find the words used.
1. coastal a. national costume
2. apply to b. proof
3. generally c. along the seaside
4. stunning d. normally/commonly
5. kilt e. outskirts
6. soundtrack f. 100 years
7. century g. seek admission
8. preserved h. fabulous/fantastic
9. suburbs i. taken care of/looked after
10. evidence j. music to set the mood
(Rodgers et al. 2006b:34)

Table 4.3 is shown below with words from exercise [2]. Again there are words that do not
appear in the text indicated by the heading (pp 41-44).

Table 4.3 Collocations from exercise [2] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
coastal coastal road Adj+N (57)
- coastal road 51 5.43
- coastal roads 6 4.88
generally not in the text
preserved not in the text
suburbs not in the text

Table 4.4 lists the most frequent bigram adjective collocates with the selected noun suburbs
(Adj+N). Some collocates of this word present sociocultural aspects that could be brought up
in the classroom. The occurrence of white, affluent (the strongest one), wealthy, and middle-
class could for example be discussed in relation to the meaning of suburbs. Maybe also the
fact that eastern suburbs is not among the ten most common collocates while northern,
southern, and western are the three most frequent ones.

Table 4.4 Collocates of the noun suburbs in COCA

Collocates Raw frequency MI
northern 134 7.98
southern 118 7.35
western 102 6.85
white 90 4.58
affluent 74 10.40
wealthy 50 8.23
middle-class 44 8.72
American 43 3.25
outer 36 7.84
industrial 32 6.14

54
Exercise [3] below is in Part 1 of @cross and the text where the words are taken from is in
chapter 5, „Over the Sea Again – to North Wales‟. It is another exercise in which students
have to match words with their synonyms and explanations.

[3] 59 - Dinner and the Welsh Language pg 77-79

In the left hand column below you will find 10 words from the text. Match them with the
words/expressions in the right hand column.
1. supper a. become different
2. bright b. not paying attention to
3. neither c. small village
4. retire d. evening meal
5. changed e. talk about something else
6. afford f. clever/intelligent
7. mend g. “call it a day”
8. ignoring h. not any
9. change the subject i. have the money to
10. hamlet j. repair/keep up
(Rodgers et al. 2006b:45)

None of the words in exercise [3] appeared with collocates in the text; therefore, no
collocation table is shown for this exercise. Some of the words were found in combinations
like mend walls, cannot afford, things changed, and fine supper. Nonetheless, these pairs are
either not listed in the collocation dictionary or have MI scores below 3.0 and thus do not
constitute collocations.
Table 4.5 presents the ten most frequent bigram noun collocates of the adjective
bright (Adj+N). The table shows that different combinations depict different meanings of
bright, as the case of bright colors, bright future, and bright side.

Table 4.5 Collocates of the adjective bright in COCA

Collocates Raw frequency MI
light 545 6.42
colors 520 8.83
red 518 6.69
lights 514 8.43
side 370 5.54
future 308 5.80
spot 297 7.49
stars 268 6.96
star 256 6.54
green 235 5.91

Exercise [4] below is also taken from Part 1 of @cross (the UK) and its words were taken
from the text in chapter 6, „The North-West‟. This exercise is of the type that presents a grid
where students have to find words. Since it was not necessary to include it here, the grid was
omitted.

55
[4] 74 - Find these words on the grid below.
lesson remains
agenda similarities
amazed suburbs
commodity trendy
incredibly disease
keen instantly
pitch obviously
probably unemplyment [sic]
(Rodgers et al. 2006b:51)

Table 4.6 includes the words from exercise [4] with their respective collocations. As the table
shows, there are again words that do not appear in the text from the textbook. The collocation
incredibly busy is followed by (2x) in the table, which means that it appears twice in the text.

Table 4.6 Collocations from exercise [4] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
agenda not in the text
amazed* amazed at Adj+P 749 5.45
commodity not in the text
incredibly incredibly busy (2x) Adv+Adj 27 6.62
instantly not in the text
suburbs surrounding suburbs Adj+N 27 6.64
trendy trendy restaurants Adj+N (77)
- trendy restaurants 46 9.89
- trendy restaurant 31 8.26
unemployment high unemployment Adv+N 338 6.45

In table 4.6 amazed is marked with an asterisk because in this case the exact form amazed
was looked for in the corpus and not its lemma AMAZE. The reason is that amazed (adj) and
amaze (v) do not belong to the same part of speech and I have tried to adhere to the meaning
intended in the exercise as closely as possible.
Table 4.7 shows the most frequent bigram adjective collocates with the noun disease
in the corpus (Adj+N).

Table 4.7 Collocates of the noun disease in COCA

Collocates Raw frequency MI
infectious 1330 11.91
cardiovascular 701 11.19
transmitted 660 10.65
chronic 578 9.14
deadly 221 7.85
venereal 213 12.76
genetic 196 6.99
human 195 3.82
fatal 174 8.12
communicable 169 12.35

56
The collocates that appear in table 4.7 might be used in class to teach students the more
specific, or technical, senses of disease like venereal, chronic, and genetic.
Exercise [5] is from Part 1 and the words are taken from chapter 8, „Breakfast with the
Royal Family‟. In this exercise the answer column is not in the original activity; the answers
are added here to show the words more clearly. Note that the word Westminster from the
exercise is a proper name and therefore it was not included in the investigation.

[5] 100 - Here are a few words describing media and government. They are divided into two
columns. Figure out which words and expressions these are and translate them into Norwegian in
the 3rd column.
(Answers) Norwegian
parlia atives parliament
broad lysis broadsheet
govern archy government
cover tion coverage
West ment Westminster
ana sheet anarchy
mon ify monarchy
elec ment election
rat minster ratify
represent age representatives
(Rodgers et al. 2006b:65)

Table 4.8 contains the words from exercise [5] and their collocations. Here there is a case of
two collocations, houses of parliament and members of parliament, which fall in a different
category, namely, N+P+N and appear in the entry of „phrases‟ in the dictionary. The reason
why they are included is that, as explained in chapter 1, these constitute highly relevant
collocations for the theme of the chapter, which is about government in the UK. Due to the
presence of the preposition, they were entered in the corpus as 3-grams.
As also shown in table 4.8, general election(s) and elect_representative(s) are written
with an optional plural form. The reason is that when looking for the nodes election and
representative, the corpus only shows variations, if any, of the collocates in the result list.
However, when checking the actual sentences we find examples of election and elections,
representative and representatives. The optional plural is shown every time this is the case.
The collocation election was held was entered in COCA as hold_election, as listed in the
dictionary. The combination international coverage was found in the text and also in the
dictionary, but since its MI was only 2.44 it was omitted from table 4.8.

57
Table 4.8 Collocations from exercise [5] in @cross
Words Collocations found in the text Type of collocation Raw frequency MI
anarchy not in the text
election general election(s) Adj+N 1560 6.84
election was held V+N (502)
- hold_election 277 3.54
- holding_election 106 2.93
- holds_election 28 1.82
- held_election 91 1.71
monarchy constitutional monarchy Adj+N 73 10.44
parliament Houses of Parliament N+P+N (3-grams) (185)
- houses of parliament 67 6.04
- house of parliament 118 3.67
Members of Parliament (603)
- member of parliament 329 7.53
- members of parliament 274 6.15
representatives elect representatives V+N (129)
- elect_representative(s) 44 6.45
- electing_representative(s) 10 6.40
- elects_representative(s) 2 5.92
- elected_representative(s) 73 4.53

Table 4.9 lists the ten most frequent bigram adjective collocates of the noun coverage in the
corpus (Adj+N). Looking at the results in the table, some of the collocates of this word
involve the topic of health insurance, as the case of universal coverage, full coverage, and
medical coverage. These collocates could be used to talk about the health system in the US.

Table 4.9 Collocates of the noun coverage in COCA

Collocates Raw frequency MI
universal 528 9.75
live 373 6.25
special 177 5.59
full 171 5.48
medical 151 5.71
continuing 134 7.45
complete 116 6.04
extensive 91 7.34
political 83 3.56
additional 78 5.75

Exercise [6] is from Part 1 of @cross and the words are taken from the text „Back to School
at Eton College‟ in chapter 9. It is a matching and translation exercise9 (there was probably a
mistake by the editors in this exercise, since no word was written in number 8).

9
The two forms of the Norwegian translations in the exercise are the two official languages in Norway, bokmål
and nynorsk respectively.

58
[6] 108 - Match the words on the left with the correct translation.
1. pursue a. læreplan
2. vocational b. følge
3. curriculum c. forbedre/forbetre
4. enhancing d. yrkes-
5. evolve e. rykte
6. reputation f. fyrtårnskole
7. academically g. veiledning/ rettleiing
8. h. sette sammen/ setje saman
9. comprehensive i. teoretisk begavet/ teoretisk gaverik
10. compile j. høyt ansett/ høgt akta
11. Beacon school k. allsidig, altomfattende
12. coaching l. utvikle
(Rodgers et al. 2006b:70)

Table 4.10 shows the words from exercise [6] that present collocations in the text. All the
collocations are pertinent to the area of education, which is the theme of the chapter.
Comprehensive school seems to be especially relevant within this topic since it appears five
times in the text. Have a reputation has only just been included as a collocation, having the
highest MI as 3.19, as shown by the table. One of the variations of reputation appears with a
negative MI, which indicates that it is definitely a free combination.

Table 4.10 Collocations from exercise [6] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
comprehensive comprehensive school (5x) Adj+N (105)
- comprehensive school 95 3.79
- comprehensive schools 10 2.36
curriculum national curriculum Adj+N 170 4.42
pursue pursue an education V+N (179)
- pursue_educations 1 4.30
- pursue_education 178 3.38
reputation build (up) a reputation V+N (275)
- built_reputation 187 5.44
- builds_reputation 5 3.84
- build_reputation 35 3.20
- building_reputation 48 2.69
have a reputation (2378)
- has 978 3.19
- had 767 2.52
- have 561 1.58
- having 25 1.11
- „ve 37 0.18
- „d 10 -1.64
vocational vocational education Adj+N 303 8.95

Table 4.11 displays the ten most frequent 4-gram noun collocates of the verb pursue in the
corpus (V+N). This might be a difficult word for students and its collocates can help
clarifying its uses.

59
 Table 4.11 Collocates of the verb pursue in COCA
Collocates Raw frequency MI
career 509 6.22
interests 311 6.22
policy 279 4.53
goals 258 6.10
careers 215 7.75
policies 202 5.61
strategy 181 5.36
dream 179 5.35
education 178 3.38
goal 148 4.86

Exercise [7] is in Part 1 of @cross and the words are taken from chapter 10, „City of
Dreaming Spires‟. The words are taken from the main text about Oxford and from the Fact
File text about education. This exercise is also a matching and translation exercise.

[7] 119 - Match the words on the left with the correct translation.
1. opinion a. utlandet
2. scholarship b. klubb for kvinnelig akademikere/ klubb for kvinnelege
3. abroad akademikere
4. homesickness c. mening/ meining
5. obtain d. hjemlengsel/ heimlengt
6. fraternity e. rådgiver/ rådgivar
7. host f. stipend
8. counsellor g. oppnå
9. sorority h. student, akademiker/ akademikar
10. graduate i. vert
j. klubb for mannlige akademikere/ klubb for mannlege akademikarar
(Rodgers et al. 2006b:75)

Table 4.12 contains the words from exercise [7] with their collocations. Similarly to the
previous exercise, there is a case of a collocation which appears more than once in the texts.

Table 4.12 Collocations from exercise [7] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
abroad study abroad (8x) V+Adv (304)
- studying abroad 55 6.08
- study abroad 174 4.50
- studied abroad 23 4.37
- studies abroad 52 3.58
graduate graduate student Adj+N (3538)
- graduate student 1697 8.52
- graduate students 1841 7.11
opinion different opinions Adj+N
- different opinion(s) 326 4.60
scholarship receive a scholarship V+N (217)
- receive_scholarship 76 5.48
- received_scholarship 109 5.10
- receiving_scholarship 25 4.88
- receives_scholarship 7 4.30

60
In relation to exercise [7] above, the word combination apply for a scholarship was found in
the text and presented an MI above 3.0, but since it was not included in the dictionary it was
omitted from the table.
Table 4.13 lists the ten most frequent 4-gram noun collocates of the verb obtain in the
corpus (V+N). A discussion about the differences between get and obtain could be pertinent
here.

Table 4.13 Collocates of the verb obtain in COCA

Collocates Raw frequency MI
information 802 5.10
data 298 4.41
permission 187 6.81
results 166 3.71
consent 159 7.02
approval 124 5.76
services 124 3.41
sample 114 4.76
copy 112 5.62
degree 104 4.15

Exercise [8] is from the last chapter of Part 1 of @cross, „Glastonbury – Music, Myth and
Magic‟ and it is a matching synonyms/explanations exercise.

[8] 130 - Draw a line between words and phrases with the same meaning.
1. legendary a. mood
2. environmentally b. moon covers sun
3. charity c. offering
4. ambience d. very famous
5. eclipse e. pedigree
6. ancient f. just ahead of you
7. sacrifice g. not commercial
8. mysterious h. think back on
9. remainder i. ecological
10. descent j. very old
11. reminisce k. what‟s left
12. imminent l. incomprehensible
(Rodgers et al. 2006b:80)

Table 4.14 presents the collocations found in the text. The combination ancient ceremonies
appears in the text, has a raw frequency of 20 in the corpus, and an MI higher than 3.0, but
since it was not included in the dictionary it is not shown in table 4.14.

Table 4.14 Collocations from exercise [8] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
imminent imminent arrival Adj+N 49 9.75
sacrifice human sacrifice(s) Adj+N 330 6.70

61
Noteworthy from exercise [8] and its correspondent collocation table (4.14) is the fact that
environmentally does not appear with any collocates in the text (although environmental
issues was found), which probably contradicts our intuitions about the use of this word. Table
4.15 lists the most frequent bigram collocates with the adverb environmentally (Adv+Adj).
The collocates of environmentally shown in the table could be used to talk about what types
of products or lifestyle would help prevent damaging the environment or even to start a
discussion about environmental issues in the US and in the UK. Noticeable from table 4.15
are the high MI scores.

Table 4.15 Collocates of the adverb environmentally in COCA

Collocates Raw frequency MI
friendly 545 12.61
responsible 316 10.56
sound 286 9.25
sensitive 209 11.04
conscious 145 11.38
sustainable 138 11.73
safe 106 8.71
benign 91 12.08
destructive 52 10.82
aware 45 7.86

Exercise [9] is taken from the same chapter as exercise [8] and its words appear in the Fact
File texts. It is the type of exercise in which students need to write a text using a number of
words.

[9] 132 - Charity

Combine as many as possible of these word [sic] in a text about aid work.
Global issues, responsibility, poverty, debt, aid, trade, water, sanitation, health, dedication.
(Rodgers et al. 2006b:81)

Table 4.16 below displays the results for the words from exercise [9]. Three of the words
from the exercise were not found in the texts. The word aid was found in many combinations,
like aid projects, aid organizations, and aid initiatives. All these present MI scores higher
than 3.0 but are not included in the table for not being listed in the dictionary. They could,
however be mentioned as examples in class to help students write this specific text. The
combination living in poverty was found in the text and presents a frequency of 309 and an
MI of 4.83 in the corpus, but did not appear in the dictionary entries (only in example
sentences). On the other hand, in poverty appeared in the dictionary but its MI in the corpus
was lower than 3.0. Thus, none of them were included as collocations.

62
 The collocation the burden of debt is listed in the dictionary both as burden of debt
and as debt burden and was looked for in the corpus as the latter since it fits better in the
syntactic category of N+N. It should be noted that the results for cancel the debt include some
British variations (cancelled and cancelling), which are less frequent than the American
forms since this corpus is of American English.

Table 4.16 Collocations from exercise [9] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
aid medical aid Adj+N 65 3.96
debt debt cancellation (2x) N+N 19 8.90
debt relief N+N 277 8.21
cancel the debt V+N (35)
- canceling_debt 8 6.63
- cancelling_debt 1 6.50
- cancel_debt 18 5.75
- cancels_debt 1 5.14
- cancelled_debt 2 4.55
- canceled_debt 5 3.32
the burden of debt
- debt burden N+N 24 4.59
dedication not in the text
health not in the text
responsibility not in the text
water drinking water N+N 1309 7.38
pure water Adj+N 115 4.44

Table 4.17 shows the most frequent bigram noun collocates of the noun trade (N+N). The
fact that the table shows the collocate center as the most frequent collocate can be used for a
discussion on how current events (as the attack on the World Trade Center) can have an
impact on the language.

Table 4.17 Collocates of the noun trade in COCA

Collocates Raw frequency MI
center 3615 7.63
agreement 1613 8.32
organization 822 6.99
deficit 811 8.77
commission 795 7.45
group 642 4.75
agreements 549 8.96
policy 546 5.40
unions 521 8.80
barriers 454 8.65

Exercise [10] is taken from Part 3 of @cross, called „@cross The USA‟, and its words are
taken from chapter 1, „An Invitation – Getting Ready‟. Of the fourteen exercises selected
from @cross, none was taken from Part 2, called „@cross the English speaking Countries‟

63
(Jamaica, South Africa, and New Zealand). The reason is that none of them met the criteria
established in the material and methods chapter.

[10] 190 - Write a paragraph describing American teenagers and their situation using the following
words.
Sports, prom, weird, wired, bullying, committed, frustrated, happy, shopping, trend, cliques
Add other words from the texts which you think fit in.
(Rodgers et al. 2006b:111)

Table 4.18 below shows the words and collocations of exercise [10]. The combinations suffer
from bullying and systematic bullying were found in the text, but do not appear in the corpus
or the dictionary. The word clique is not among the 9.000 entries of the dictionary and
therefore is omitted here. However, it was found in the text in combinations like member of a
clique, belong to a clique, and clique identity. The two first were found in the corpus with MI
scores higher than 3.0 while the latter is not in the corpus. The word prom is not included in
the collocation dictionary, but in the text it appears as junior prom and senior prom, which
might be interesting for students within the topic of school in the US (their frequency and MI
scores and were 33/ 9.13 and 148/ 9.00 respectively). The combination wired teenagers was
not found in the dictionary or the corpus. One last note concerning table 4.18 is the
appearance of the misspelled form sports *heros in COCA and how the MI score is affected
by low occurrence, meaning that even words that do not exist might get an MI score higher
than 3.0 (see subsection 3.3.1).

Table 4.18 Collocations from exercise [10] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
frustrated feel frustrated V+N (203)
- feel frustrated 111 5.32
- feeling frustrated 32 4.95
- feels frustrated 13 4.79
- felt frustrated 47 4.53
shopping shopping lists N+N (413)
- shopping lists 68 8.12
- shopping list 345 7.98
sports sports hero N+N (157)
- sports heros 1 6.65
- sports heroes 92 6.63
- sports hero 64 5.18
trend fashion trends N+N (82)
- fashion trend(s) 81 6.37
- fashions trend(s) 1 3.92
latest trends Adj+N 58 6.68

Table 4.19 outlines the most frequent bigram adjective collocates with the noun prom
(Adj+N). This is a case when sociocultural aspects are clearly conveyed by combinations, as

64
in white prom, integrated prom, and separate prom. The concordance lines could be checked
for more information.

Table 4.19 Collocates of the noun prom in COCA

Collocates Raw frequency MI
senior 154 9.97
junior 33 9.07
white 17 4.50
integrated 12 8.39
high-school 8 10.15
separate 7 6.04
junior-senior 4 15.73
pink-chiffoned 2 17.90
frilly 2 10.47
fancy 2 6.32

Exercise [11] is taken from Part 3 of @cross and its words are from chapter 2, „Arrival‟. It is
the type of exercise in which students have to write a short text using the given words.

[11] 203 Use as many as possible of the following words in a paragraph about travelling:
Arrival, approach, immigration agent, relatives, stay (n), purpose, stamp, customs, exit, vegetables
(Rodgers et al. 2006b:115)

Table 4.20 below presents the three words which appear with a collocate in the text. The
collocation with purpose is usually the purpose of, but it was looked in the corpus without the
article in order to follow the syntactic categories established in chapter 3.

Table 4.20 Collocations from exercise [11] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
customs customs agent N+N (180)
- customs agents 106 7.88
- customs agent 74 7.19
purpose (the) purpose of N+P 13128 3.69
stamp stamped the passport V+N (44)
- stamp_passports 17 8.34
- stamp_passport 27 7.64

Table 4.21 shows the results for the ten most frequent bigram adjective collocates of the noun
vegetable in the corpus (Adj+N). This is a word that is probably well-known by students but
whose collocates might not yet be in students‟ productive knowledge. The collocates could be
categorized into two lists: the quality of the vegetables (fresh, green, leafy, and organic) and
how you eat them (grilled, raw, frozen, cooked, steamed, and roasted).

65
 Table 4.21 Collocates of the noun vegetables in COCA
Collocates Raw frequency MI
fresh 336 7.63
green 187 5.98
leafy 140 11.14
grilled 132 9.56
raw 109 7.65
frozen 89 7.29
organic 88 7.43
cooked 87 7.89
steamed 86 10.07
roasted 86 9.11

Exercise [12] is in Part 3 of @cross and the words included are taken from chapter 3, „A Day
at Plimoth Plantation‟. This is a matching antonyms exercise and only the words from the
first column have been taken from the text in the textbook.

[12] 211 - Pair off words that have opposite meanings.

1. Attractive a. Inequality
2. Celebrate b. Expose
3. Remain c. Indulgent
4. Protect d. Awful
5. Innocent e. Ugly
6. Eager f. Molester
7. To make it g. Careless
8. Keep up h. Dishonor
9. Independence i. Guilty
10. Equality j. Fail
11. Delicious k. Stop
12. Victim l. Dependence
13. Strict m. Disappear
(Rodgers et al. 2006b:119)

Table 4.22 below lists the words which have collocations in the text. It should be noted that
to make it and keep up were not checked in the corpus for being already word combinations.

Table 4.22 Collocations from exercise [12] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
delicious delicious meal Adj+N (78)
- delicious meal 49 8.35
- delicious meals 29 8.11
victim victim of N+P 8295 3.16

Table 4.23 displays the ten most frequent bigram noun collocates of the adjective attractive
in the corpus (Adj+N). The table shows that this adjective is used over four times more
frequently to characterize women than men among the combinations with an MI higher than
3.0 (woman, women, and girl comprise 394 tokens, against 89 for man). This indicates some
sociocultural issues that could be brought to discussion. On the other hand, the table shows

66
that the strongest combinations are the ones in which the adjective attractive does not refer to
people (e.g. alternative, option, and feature).

Table 4.23 Collocates of the adjective attractive in COCA

Collocates Raw frequency MI
woman 260 6.17
alternative 99 7.25
option 97 7.77
man 89 3.67
women 85 3.86
place 62 3.55
girl 49 4.80
features 37 5.57
feature 36 6.40
investment 31 5.23

Exercise [13] belongs to Part 3 of @cross and it is in chapter 4, „Boston‟. It is a purely

translation exercise with words related to education.

[13] 227 - Education Vocabulary

Translate the following words into Norwegian.
Grades Subjects
Curriculum Classes
Scholarships Tuition
To graduate Core courses
Electives Degree
(Rodgers et al. 2006b:129)

Two of the words presented in the exercise above are not in the chapter texts, as table 4.24
shows. As explained in chapter 3, the collocation get good grades (V+Adj+N) follow the
definition of a collocation, it combines the syntactical categories devised in chapter 3, and
can be considered relevant for teaching. For all these reasons it is included in the table. This
collocation was written in the corpus as [get] good grades in the „word(s)‟ field. The
combination tuition costs was found in the text and, in the corpus, it presented a frequency of
81 and an MI of 7.51. However, it is not listed in the dictionary and so it is omitted.

Table 4.24 Collocations from exercise [13] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
classes not in the text
curriculum national curriculum Adj+N 170 4.47
electives not in the text
grades get good grades V+Adj+N (177)
- getting good grades 51 9.09
- gets good grades 12 7.99
- gotten good grades 4 7.68
- get good grades 82 7.59
- got good grades 28 6.81

67
Table 4.25 lists the ten most frequent 4-gram verb collocates of the noun scholarship in the
corpus (V+N). It shows that this word is commonly used together with the verbs win, receive,
offer, earn, award, and provide, which could be categorized as verbs for giving a scholarship
(offer, award, and provide) and verbs for receiving it (receive, win, and earn).

Table 4.25 Collocates of the noun scholarship in COCA

Collocates Raw frequency MI
won 146 5.29
received 109 5.04
offered 97 5.20
receive 76 5.42
win 74 4.32
earned 69 5.97
offer 62 4.13
awarded 58 7.09
provide 57 3.41
earn 37 5.61

Exercise [14] below is taken from Part 3, chapter 6 „Washington D.C.‟ of @cross. It is a
translation exercise and it is the last one selected from @cross workbook.

[14] 242 - Put the words in alphabetical order and translate them into Norwegian.
Words Alphabetical Order Translation
Majority
Bill
Election
Support
Constitutional
Consider
Legislative
Judicial
Executive
Law
Agenda
Minority
(Rodgers et al. 2006b:136)

Many of the words from exercise [14] have verbs and nouns as collocates in the text and for
this reason they present a great number of form variations, as provided in table 4.26 below. As
shown in the table, there are two words presented in the exercise which do not appear in the
text. It should be noted that the combination in the majority (P+Art+N) is the same case as get
good grades explained above. In the case of the word constitutional, no collocation was found
in the text and for this reason it is not included in table 4.26. On the other hand, the derivative

68
unconstitutional appeared three times in the text in the collocation law_unconstitutional (MI
6.16, frequency 112).

Table 4.26 Collocations from exercise [14] in @cross

Words Collocations found in the text Type of collocation Raw frequency MI
agenda not in the text
bill pass a bill V+N (1280)
- pass_bill 566 4.24
- passed_bill 561 4.08
- passing_bill 98 2.74
- passes_bill 55 2.49
sign a bill (725)
- signed_bill 346 4.30
- signing_bill 63 3.67
- sign_bill 266 3.05
- signs_bill 50 1.40
consider consider necessary V+N 101 3.33
election hold an election V+N (502)
- hold_election 277 3.60
- holding_election 106 2.99
- holds_election 28 1.86
- held_election 91 1.77
free election(s) Adj+N 564 5.72
executive executive branch Adj+N (1377)
- executive branch 1333 9.90
- executive branches 44 5.43
judicial judicial branch Adj+N (146)
- judicial branch 112 9.11
- judicial branches 34 7.84
law pass a law V+N (1642)
- passed_law 821 4.35
- pass_law 632 4.12
- passing_law 129 2.85
- passes_law 60 2.33
legislative legislative branch Adj+N (337)
- legislative branch 233 9.95
- legislative branches 104 9.24
majority in the majority Prep+Art+N 638 3.24
support not in the text

Table 4.27 shows the ten most frequent bigram noun collocates with the adjective
constitutional (N+N). The collocates from the table suggest that this word is restricted to
politics and law and, as such, presenting them to students might be an aid to the understanding
of the word‟s meaning.

69
 Table 4.27 Collocates of the adjective constitutional in COCA
Collocates Raw frequency MI
amendment 1112 10.32
rights 782 7.63
right 650 4.62
law 598 6.53
convention 236 7.81
court 180 4.98
crisis 161 6.52
amendments 151 9.73
protection 147 6.45
issues 143 4.93

These were all the exercises from @cross, their collocations found in the textbook, and the
combinations looked for in COCA. The total items of vocabulary included in these fourteen
exercises are 156, from which only nine are chunks and not single words (5.9%). Here are
these combinations and the exercise number in parenthesis: according to [1], apply to [2],
change the subject [3], academically gifted [6], Beacon school [6], global issues [9],
immigration agent [11], to make it [12], keep up [12], and core courses [13]. In the next
section we turn to the results of the second part of the material, the Passage website.

4.4 Results from Passage

The exercises from Passage are taken from its website, where fourteen vocabulary exercises
that present lists of single words were found. There are a total of seventy vocabulary exercises
on the website and their types and number are:

crossword (23)
matching synonyms or definitions (15)
word grid (13)
gap-filling or completing a sentence (7)
translation (6)
collocation (3)
working with compounds (2)
setting words into categories (1)

As stated in chapter 3, the first aim requires exercises that work with single words. Thirteen of
the match synonyms or definitions exercises and one involving translation (see table 3.1) have
been chosen, all of them containing single words. I did not choose more of the type involving
Norwegian translation because, as opposed to the ones selected in @cross, this type of
exercise in Passage does not present the keywords in English, but in Norwegian. As in the
previous section, the words that are already presented as combinations are not included in the

70
search, and the words which have no collocations in the text are omitted from the collocation
tables. Pertaining numbering, the exercise numbers follow the order started in @cross, going
in this section from 15 to 28 (they are not numbered on the website). Among the texts used for
the exercises seven are authentic, four are constructed and three are mixed.
Exercise [15] is a matching synonyms exercise and it refers to the first part of the
Passage textbook, which is an introduction to the other units. The text they are taken from is a
short story by Ernest Hemingway, an authentic text.

[15] Starting Out - “A Day’s Wait”

Pair up synonyms
Pair up the words on the left (as used in the text) with their synonyms on the right.
evidently distant
capsule begin
commence different
slack tablet
detached tired
various clearly
(Sørhus et al. 2006b)

Most of the words from exercise [15] have no collocations in the text they are taken from,
which is the reason why table 4.28 only contains the collocation detached from. Concerning
this collocation, the lemma was looked for in the corpus like all the other words and not the
exact form detached.

Table 4.28 Collocations from exercise [15] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
detached detached from V+P 493 6.46

Table 4.29 displays the ten most frequent 4-gram verb collocates of the noun slack in the
corpus (V+N). The table shows that the collocates of slack can demonstrate the different
meanings of this word.

Table 4.29 Collocates of the noun slack in COCA

Collocates Raw frequency MI
cut 157 5.92
pick 116 6.79
take 99 3.56
wearing 59 6.19
went 53 3.58
dressed 46 6.60
wore 40 6.47
picking 32 6.73
picked 29 5.24
taking 29 3.58

71
Exercise [16] is also from the same introductory part of Passage and its words are taken from
a newspaper article by Sarah Kershaw (an authentic text). The heading of the exercise points
out that the words are taken from the text, which, in contrast with @cross, happens indeed in
all the Passage exercises.

[16] Starting Out - “A Solitary Elephant Poses a Big Problem in Alaska”

Link the words on the left (taken from the text) with their synonyms on the right.
debate worry
anguished discussion
acknowledge worried
relocate make friends
frigid admit
consternation move
socialize suggestion
proposal cold
(Sørhus et al. 2006b)

The words from exercise [16] appear in a number of free combinations in the text (yearlong
debate, frigid afternoon, socialize with, and unusual proposal) but only in two collocations,
which are listed in table 4.30 below.

Table 4.30 Collocations from exercise [16] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
consternation cause consternation V+N (81)
- caused_consternation 48 8.78
- causing_consternation 11 8.11
- causes_consternation 10 7.09
- cause_consternation 12 5.80
debate a debate over (sth) N+P 3728 6.05

Table 4.31 exhibits the ten most frequent bigram adjective collocates of the noun debate
(Adj+N). The table shows that all the collocates of debate are associated with the topic of
politics and current issues, the strongest ones being heated, presidential, and congressional.

Table 4.31 Collocates of the noun debate in COCA

Collocates Raw frequency MI
public 805 5.43
presidential 633 7.90
political 567 5.16
national 512 4.81
heated 344 9.36
current 219 4.99
ongoing 218 4.99
great 214 3.39
congressional 160 6.60
open 139 3.47

72
Exercise [17] is a matching synonyms exercise and it practices words taken from an
autobiography called „When I was Puerto Rican‟ from „Global English‟, the second section of
Passage. This text is written by Esmeralda Santiago and it is an authentic text.

[17] Global English - “When I was Puerto Rican”

Link up the words that mean more or less the same.
notify cloudy
monologue pronounce
clasp shout
overcast accompany
yell inform
enunciate hold
escort brow
forehead speech
(Sørhus et al. 2006b)

Table 4.32 below includes the collocates of the words from exercise [17]. The combination
overcast morning was found in the corpus with an MI above 3.0 (5.98) and raw frequency of
12, but it does not appear in the dictionary.

Table 4.32 Collocations from exercise [17] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
clasp clasped my hands V+N (652)
- clasp_hands 524 8.48
- clasp_hand 128 5.76
enunciate enunciate (every) word V+N (50)
- enunciate_word 35 7.06
- enunciate_words 15 5.51
yell yell at V+P 2584 5.28

Table 4.33 contains the ten most frequent 4-gram noun collocates of the verb clasp in the
corpus (V+N). Looking at the table, we can see that the collocates of this word comprise parts
of the body (mainly hand) as its collocational field. It should be noted that the word prayer
here is part of the expression hands clasped in prayer, where the collocate hand(s) appears
again. Furthermore, many of the occurrences of knees are also together with hand(s), like in
the examples his hands clasped between his knees and his hands clasped around his knees.
Also in the case of back, the concordance lines show that back and hands appear in the same
sentences, most of them being close variations of his hands clasped behind his back. The
same happens with the collocate head, where many of the examples are of the kind hands
clasped behind/above/under his head. What becomes clear by looking at the concordance is
the strong bonding between clasp and hand (also shown by the MI of 8.42). Most of the

73
occurrences of clasp together with knife are instances of the compound word clasp knife (US
pocketknife) but there are occurrences of the collocation clasp a knife as well.

Table 4.33 Collocates of the verb clasp in COCA

Collocates Raw frequency MI
hands 524 8.42
hand 128 5.70
back 72 3.06
head 34 3.67
arms 29 5.10
knees 24 6.61
prayer 17 6.38
knife 17 6.38
shoulder 16 5.12
neck 15 5.14

Exercise [18] is also taken from the „Global English‟ unit and the words are from a short story
called „The Shining Mountain‟ by Alison Fell, an authentic text. The exercise is again of the
matching synonyms type.

[18] Global English - "The Shining Mountain"

Join words that mean the same (or nearly)!
cast slide
weep throw
pant tired
weary cry
skid burden
shin climb
load breathe heavily
(Sørhus et al. 2006b)

Table 4.34 lists the words from exercise [18] with their collocations found in the text.

Table 4.34 Collocations from exercise [18] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
load carry loads V+N (418)
- carrying_load 152 5.72
- carry_load 180 5.35
- carries_load 30 4.44
- carried_load 56 3.85
weary grow weary V+Adj (233)
- grown weary 90 9.19
- grow weary 50 7.79
- grew weary 44 7.59
- growing weary 43 6.89
- grows weary 6 6.80
weep weep _ tears V+N
- weep_tears 69 6.99
- weep_tear 3 3.84

74
Three words from exercise [18] appear in free combinations in the text and therefore are not
included in table 4.34. These are: lungs panted, skidding and sliding, and shin rocks.
Table 4.35 presents the ten most frequent 4-gram noun collocates of the verb cast in
the corpus (V+N). The collocates of cast reveal its several meanings, as for example in cast a
shadow, cast a vote, cast doubt, cast iron, and cast of characters.

Table 4.35 Collocates of the verb cast in COCA

Collocates Raw frequency MI
shadow 605 7.89
vote 574 6.22
shadows 500 8.17
characters 415 7.07
ballots 401 9.13
doubt 384 6.21
members 377 4.47
iron 374 7.11
light 308 4.06
votes 285 6.61

Exercise [19] is from the third section of Passage, „On the Road‟, and the words are taken
from the short story „Paradise‟ by Matthew Kneale, an authentic text. This is another
matching synonyms exercise.

[19] On the Road - “Paradise”

Pair up words that mean the same – or nearly the same.
snap break
push pull
sip dislike
resent look
glance drink
tug disappear
vanish shove
(Sørhus et al. 2006b)

Table 4.36 below contains the words from exercise [19] with their collocations. The
collocation glanced at, as indicated in the table, was found five times in the text. Two other
combinations were found in the text and in the corpus but not in the dictionary: tugging his
wrist and doubts vanished.

75
Table 4.36 Collocations from exercise [19] from Passage
Words Collocations found in the text Type of collocation Raw frequency MI
glance glanced at (5x) V+P 7360 5.05
push pushed through V+P 1470 4.72
sip took a sip V+N (1385)
- took_sip 820 5.67
- takes_sip 254 5.18
- taking_sip 133 3.88
- take_sip 153 2.29
- taken_sip 25 1.49
sipping tea V+N
- sip_tea 550 8.37

Table 4.37 arranges the ten most frequent 4-gram noun collocates of the verb snap in the
corpus (V+N). These collocates reveal some distinct meanings of snap, such as in snap your
fingers, snap pictures/photos, snap peas and snap beans (compounds), and snap to attention.

Table 4.37 Collocates of the verb snap in COCA

Collocates Raw frequency MI
fingers 686 7.87
pictures 205 6.01
picture 199 5.21
peas 161 8.70
attention 143 4.23
photos 131 6.15
photo 86 3.20
beans 76 6.02
neck 65 4.51
shot 56 3.22

Exercise [20] below belongs to the fourth part of Passage, called „We Were Here First‟ and
the text where the words are taken from is a factual text, as the book defines it, about the
Indians in America. Unlike all the previous exercises, this one is based on a constructed text.

[20] We Were Here First - "Native Americans - The Original Inhabitants"

Link the words that mean the same – or nearly the same.
perish important
innumerable newcomer
notable stupid
tribe nation
foolhardy many
diverse die
settler varied
(Sørhus et al. 2006b)

Noticeable from exercise [20] is that although the words in the two columns are synonyms
(e.g. foolhardy and stupid, perish and die, notable and important) their use might be distinct

76
in terms of collocations or register. One way to find out these differences is to check their
concordance lines in the corpus or the most frequent collocates of each of them.
Table 4.38 lists the words with their collocations found in the text. There is a
combination with the word foolhardy in the text, a foolhardy attack, which was not found in
the corpus or the dictionary. Another combination not found in the dictionary was Indian
tribe, although it appeared in the corpus with an MI of 8.73 and frequency of 658. The fact
that European settlers is a collocation but not African settlers for example might be
interesting for discussion in terms of historical facts.

Table 4.38 Collocations from exercise [20] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
diverse diverse group Adj+N (589)
- diverse groups 219 5.53
- diverse group 370 5.42
notable notable exception Adj+N (437)
- notable exceptions 186 11.06
- notable exception 251 10.17
settler European settlers Adj+N 145 7.75
tribe nomadic tribes Adj+N 40 10.10

Table 4.39 shows the ten most frequent bigram noun collocates of the adjective notable in the
corpus (Adj+N). What stands out in this table is the repetition of words, indicating that
notable has fewer frequent collocates than the ones seen so far so that the ten most common
collocates are variations of only six words (exception, example, feature, performance,
difference, and success).

Table 4.39 Collocates of the adjective notable in COCA

Collocates Raw frequency MI
exception 266 11.19
exceptions 190 12.02
example 79 6.00
feature 49 8.15
performances 46 9.12
examples 44 7.94
differences 40 6.59
successes 37 9.81
difference 35 6.08
features 30 6.57

Exercise [21] below is from the same part of Passage as the previous one, and the words are
taken from a text about the aborigines in Australia, a constructed text. The exercise is to
match synonyms.

77
[21] We Were Here First - "Aboriginal Australians"
The verbs on the left are all taken from the text. Link them with their synonyms on the right.
adapt hunt
emerge come out
take place change
christen name
provide give
stalk happen
(Sørhus et al. 2006b)

Only two of the words in the exercise above are part of collocations in the text, as table 4.40
shows. The combination provide (someone) with was found in the text and in the dictionary
but its MI was below 3.0.

Table 4.40 Collocations from exercise [21] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
adapt adapt to V+P 3590 3.68
emerge emerge from V+P 8252 6.10

Table 4.41 presents the ten most frequent 4-gram noun collocates of the verb provide in the
corpus (V+N). As in the case of obtain versus get (see comment about table 4.13), it would be
interesting to compare the uses of provide and give. The comparison might reveal not only
differences in collocations but also in register. In contrast with previous tables, the
frequencies in this table are higher while the MI scores are lower.

Table 4.41 Collocates of the verb provide in COCA

Collocates Raw frequency MI
information 4515 4.98
services 2785 5.28
support 2606 4.36
care 1817 3.77
evidence 1730 4.56
opportunities 1621 6.03
opportunity 1446 4.99
access 1208 4.76
service 1083 3.50
assistance 1071 5.77

Exercise [22] is from the same part as the two previous ones and it is also a matching
synonyms exercise. This is the first exercise whose words are taken from a song, „Took the
Children Away‟ by Archie Roach, an authentic text. The song lyrics are displayed as a text in
the textbook.

78
[22] We Were Here First - "Took the Children Away"
Connect the words on the left with the synonyms on the right.
origins make someone feel stupid
snatch subject, topic
adopt take away
apologize say sorry
orphanage roots
humiliate foster
issue children‟s home
(Sørhus et al. 2006b)

Table 4.42 exhibits the collocations found in the text with the words from exercise [22]. It
should be noted that unlike most of the grammatical collocations that appeared before in this
chapter, snatch_from was entered in the corpus as a 4-gram and not a bigram, since there is
usually an object between the verb and the preposition. Moreover, although the collocation
found in the text is in the past tense, the lemma SNATCH was entered in the corpus. The
combination adopted by was found in the text and in the corpus but not in the dictionary.

Table 4.42 Collocations from exercise [22] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
apologize apologize for V+P 2282 4.23
snatch snatched_from V+P (4-gram) 752 3.91

Table 4.43 below lists the ten most frequent 4-gram noun collocates of the verb adopt in the
corpus (V+N). The table reveals that the word adopt has stronger collocates when used in its
more abstract sense of „to accept or start something new‟ (CALD 2005:17), as in adopt a
resolution (MI 6.07), adopt policies (MI 5.88), and adopt an approach (MI 5.07). When the
sense is „to take a child‟, the MI scores are slightly lower. This word might be useful to talk
about politics, for example in phrases like the government has adopted (a) new
policy/approach/strategy/measures.

Table 4.43 Collocates of the verb adopt in COCA

Collocates Raw frequency MI
children 395 3.38
policy 306 4.48
approach 294 5.07
child 285 3.89
policies 278 5.88
strategy 188 5.23
resolution 182 6.07
standards 164 4.85
baby 156 4.06
measures 139 4.94

79
Still from the fourth part of the Passage material, exercise [23] is the only one based on a
listening activity. The text students listen to is „Dracula‟ by Abraham „Bram‟ Stoker and the
teaching material indicates that the text was „slightly adapted‟ (Sørhus et al. 2006b). The
transcripts are provided only in the teacher‟s resources on the website.

[23] We Were Here First - “Dracula”

The words on the left are taken from the text. Link them up with words that mean the same - or nearly
the same.
aid story
fancy faint
swoon help
narrative memory
remembrance confused
cleanse imagining
bewildered drink
draught wash
(Sørhus et al. 2006b)

Since this exercise is based on a literary text with many infrequent words, the combinations
found in the transcript are either rare or inexistent in the corpus and none of them is listed in
the dictionary. Some of these are: sleeping draught (appeared twice in the text and presented
an MI of 9.82 in the corpus), horrible fancies and awful narrative (found in neither), and in a
half swoon (found in the corpus with an MI of 4.52). Therefore, there is no collocation table
for exercise [23].
Table 4.44 presents the ten most frequent 4-gram noun collocates of the verb cleanse
in the corpus (V+N). As the table shows, the collocates are not so frequent compared to
previous tables and their MI scores are mostly close to 3.0, except for palate. It might be
interesting to contrast the uses of cleanse with the synonym wash given by the exercise.

Table 4.44 Collocates of the verb cleanse in COCA

Collocates Raw frequency MI
body 29 4.82
skin 16 5.39
face 16 3.53
blood 15 4.76
area 14 3.76
system 14 3.24
palate 13 10.37
mind 12 3.81
land 11 4.04
air 10 3.27

Exercise [24] is taken from the fifth part of Passage, called „The Road to Freedom‟, and
students are supposed to match words with their definitions. The text where the words are

80
taken from is a factual text with excerpts from autobiographies, so that it is partly constructed
and partly authentic.

[24] The Road to Freedom - "The Road to Freedom - Three Key Figures"
Which noun fits the definition?

Part of a train where the passengers sit: obligation

The feeling of bitterness at some injustice: resentment
Something you have a duty to do: cowardice
Reaching agreement by talking together: negotiation
Attitude not based on rational thinking: prejudice
Showing a lack of courage: compartment
(Sørhus et al. 2006b)

Table 4.45 displays the collocations found in the text. It stands out that the collocation color
prejudice, even though appearing twice in the text, has a raw frequency of 8 tokens and an MI
score of 3.04, barely fitting the definition of collocation provided in chapter 1.

Table 4.45 Collocations from exercise [24] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
negotiation negotiations with N+P 1717 4.04
obligation fulfill my obligation V+N (202)
- fulfill_obligation 198 9.34
- fulfills_obligation 4 6.70
prejudice color prejudice (2x) N+N 8 3.04
resentment resentment towards N+P 27 7.33

Table 4.46 lists the ten most frequent bigram noun collocates with the noun prejudice in the
corpus (N+N). It is noticeable that prejudice has a stronger bonding to racial (MI 10.02) and
ethnic (6.58) than with the collocate color, which appears in the text. The difference between
the raw frequencies is also visible, especially between racial (202) and color (8). The
question here is why this infrequent and weak collocation appeared in the text instead of the
more common, stronger, and similar in meaning, racial prejudice.

Table 4.46 Collocates of the noun prejudice in COCA

Collocates Raw frequency MI
racial 202 10.02
social 40 4.60
old 34 3.90
ethnic 27 6.58
cultural 24 5.40
religious 23 5.32
personal 22 4.70
blatant 19 10.24
subtle 17 7.41
sexual 17 5.03

81
Exercise [25] contains words from the same text as the previous one. In this exercise students
have to match synonyms.

[25] The Road to Freedom - "The Road to Freedom - Three Key Figures"
Join up the words that mean the same - or nearly the same.
murmur mutter
resentment answer
persist choose
select continue
assassinate keep separate
retort anger
segregate murder
(Sørhus et al. 2006b)

Table 4.47 exhibits the results for the words in exercise [25]. No combinations were found
with the verb segregate in the text, only with other forms, such as segregated society (37
tokens, MI 6.71), racial segregation (213 tokens, MI 11.02), and patterns of segregation (9
tokens, MI 5.13). Among these, racial segregation is listed in the OCD, but since it belongs
to another part of speech than the word in the exercise, it was not included in the table.

Table 4.47 Collocations from exercise [25] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
murmur murmured politely V+Adv 5 7.26
persist persist in V+P 1200 3.45
resentment resentment towards N+P 27 7.33

Table 4.48 presents the ten most frequent 4-gram noun collocates of the verb assassinate in
the corpus (V+N).

Table 4.48 Collocates of the verb assassinate in COCA

Collocates Raw frequency MI
president 129 5.29
leaders 32 5.58
king 22 5.02
minister 15 5.27
leader 15 4.83
officials 12 3.72
Pope 9 6.20
character 9 4.29
commander 6 5.32
militant 5 7.02

The collocates of assassinate listed in table 4.48 above confirm its definition, which is „kill
someone famous or important‟ (CALD 2005:66), as seen in collocates like president, king,
and Pope. In fact, all of the collocates could be said to be „important people‟ except from
assassinate one’s character, which is used when someone degrades somebody‟s image or

82
reputation. The collocates of the verb assassinate could be contrasted with the ones from
murder, which is given as the synonym in the exercise.
Exercise [26] takes its words from the sixth part of Passage, called „The Power and
the Glory‟ from a constructed text about the English monarchs. In this exercise students have
to match words with their Norwegian translations.

[26] The Power and the Glory - “English Monarchs”

Match the English words with their Norwegian translations:
resistance avstamning
infidelity stamme
tribe erobre
execute nymotens
allegiance knust
descent henrette
conquer blodbad
source utroskap
new-fangled motstand
devastated kilde
slaughter troskap
(Sørhus et al. 2006b)

Most of the words above are not part of any collocations in the text, as table 4.49 below
shows. The combination promise allegiance was found in the text, but it was not found in the
dictionary and, in the corpus, it is infrequent (3 tokens) and has an MI below 3.0. The
combination Germanic tribes, on the other hand, was found in the corpus with an MI of 10.45
but it is not listed in the dictionary. Thus, these were omitted from table 4.49.
The adjective new-fangled should be given some extra attention. Firstly, it is unusual
that new-fangled is presented as a single word and not a collocation in the exercise, especially
when it collocates strongly with ideas, as it was found in the text. Furthermore, when looking
for the most used nouns with new-fangled, ideas and idea are the ones on top of the list,
followed by other combinations with frequencies lower than 3. In the case of the word
resistance, the collocation found in the text was put up fierce resistance (V+Adj+N), listed
both in the corpus and the dictionary and another exception with regard to the syntactical form
(like get good grades and in the majority). If we look for put up_resistance, allowing a word
between up and resistance, and looking for the lemma PUT (entered as [put] up * resistance),
COCA shows 35 tokens of the type put up no resistance, put up little resistance, put up stiff
resistance, and put up tremendous resistance. This exemplifies other ways in which
collocations with more lexemes can be found in the corpus and shows how much variation of
results can be encountered. This is the reason why I have devised simpler criteria than the
whole spectrum available in the corpus.

83
Table 4.49 Collocations from exercise [26] in Passage
Words Collocations found in the text Type of collocation Raw frequency MI
descent direct descent Adj+N 8 4.63
new-fangled new-fangled ideas Adj+N (10)
- new-fangled ideas 3 9.30
- new-fangled idea 3 7.97
- newfangled ideas 3 7.32
- newfangled idea 1 4.40
resistance put up fierce resistance V+Adj+N 3 7.24

Table 4.50 below lists the ten most frequent 4-gram verb collocates of the noun allegiance in
the corpus (V+N). According to the table, the verb pledge is the most frequent and strongest
collocate of allegiance in the corpus and it appears in the dictionary. It did not, however,
appear in the text while the free combination pledge allegiance was included. Thus, the
question can be raised again: Why would the textbook writers include the weak and rare
combination promise allegiance in the constructed text instead of the more common and
stronger collocation pledge allegiance? This matter will be discussed further in chapter 5.

Table 4.50 Collocates of the noun allegiance in COCA

Collocates Raw frequency MI
pledge 82 10.04
switched 35 8.56
swear 34 8.68
pledged 33 9.21
owe 30 8.35
pledging 25 11.14
swore 21 9.18
declare 17 7.96
owed 16 8.07
sworn 15 8.22

Exercise [27] is from the same section as [26], and the text the words are taken from is an
excerpt from the novel „Animal Farm‟ by George Orwell, thus, an authentic text. It is a
matching definitions exercise.

[27] The Power and the Glory - “Animal Farm”

Match the words on the left with the correct definition in the right hand column.

oppression the practice or extending a state‟s rule over other territories

totalitarian the theory that a country‟s wealth should belong to the people as a whole
philosophy an open or armed fight against a government
socialism characteristic of a one-party state
rebellion the search for knowledge and truth, especially about the nature of
man and his behavior and beliefs
imperialism to be governed cruelly
(Sørhus et al. 2006b)

84
Only two words from the exercise above are part of collocations in the text, as displayed by
table 4.51. British imperialism, in the same way as European settlers in exercise [20], is a
collocation which shows a historical aspect that can be discussed in class.

Table 4.51 Collocations from exercise [27] in Passage

Words Collocations found in the text Type of collocation Raw frequency MI
imperialism British imperialism Adj+N 22 7.55
totalitarian totalitarian regime Adj+N (210)
- totalitarian regimes 109 12.97
- totalitarian regime 101 10.99

Table 4.52 presents the ten most frequent bigram noun collocates of the adjective totalitarian
in the corpus (Adj+N). Since this section of Passage deals with power and politics, this is one
of the words whose collocates could be explored in class. Some of the collocates show a
stronger bonding to totalitarian, especially totalitarian regime and totalitarian dictatorship.

Table 4.52 Collocates of the adjective totalitarian in COCA

Collocates Raw frequency MI
regimes 109 12.97
regime 101 10.99
state 71 5.61
system 56 6.97
government 33 5.99
states 29 5.66
society 27 6.96
rule 21 7.89
dictatorship 19 11.61
governments 18 8.45

Exercise [28] contains words taken from the last section of Passage, called „Global Village‟.
It is a matching synonyms exercise and the words are taken from a (constructed) text about
football.

[28] Global Village - “Want to buy a football club?”

Link up words that mean the same - or nearly the same.
outlet dismiss
loathe promising
revenue hate
auspicious clever
hectic income
canny payment
sack busy
fee shop
(Sørhus et al. 2006b)

Some of the words from the exercise above are quite unusual. Thus, teaching students their
collocates can help them understand their meanings, as in the case of auspicious start. Start is

85
the second most common noun collocate of auspicious in the corpus (17 tokens), preceded
only by its synonym beginning (27 tokens). The combinations canny businessman, media
revenues, and sales outlets were found in the text and corpus but not in the dictionary.

Table 4.53 Collocations from exercise [28] from Passage

Words Collocations found in the text Type of collocation Raw frequency MI
auspicious an auspicious start Adj+N 17 6.77
fee transfer fees N+N (28)
- transfer fee 16 4.73
- transfer fees 12 4.03
hectic hectic schedule Adj+N (104)
- hectic schedules 23 10.51
- hectic schedule 81 10.14

Table 4.54 presents the ten most frequent bigram adjective collocates of the noun outlet in the
corpus (Adj+N). Collocations such as retail and fast food outlet; electric and electrical outlet;
creative, social, and emotional outlet; and GFCI outlet (a boat appliance) can show the
different meanings of this noun.

Table 4.54 Collocates of the noun outlet in COCA

Collocates Raw frequency MI
retail 256 11.01
electrical 202 10.97
other 89 3.38
fast-food 59 11.49
creative 59 7.94
electric 23 6.73
social 16 3.14
emotional 15 5.59
major 15 3.51
GFCI 14 13.44

This was the last exercise from Passage. The total number of vocabulary items in the fourteen
exercises selected from the website is 102; among these only one is a collocation (0.98%):
take place from exercise [21]. The summary below will give an overview of the results.

4.5 Summary
This chapter has presented the results of the search in the textbooks, corpus, and collocation
dictionary according to the criteria established in chapters 1 and 3. Comments have been
added between the tables for matters of practicality and many of these will be discussed
further in chapter 5.
Table 4.55 below presents a summary of the results for @cross and Passage. The table
shows first the total number of exercises from each material, then the number of authentic,

86
constructed, and mixed exercises. Next, the total number of vocabulary items is listed; 156
from @cross and 102 from Passage, which comprise 258 items. In @cross, 93.6% of all the
items constitute single words, while in Passage the percentage is 99.1%; the remaining items
are word combinations. @cross presents 18 words in the exercises that are not in the texts,
which comprise 12.3% of all the single words. This means that only 128 words from @cross
(87.7% of all single words) were found in the texts. In Passage all the words have been found
in the texts.
The table also displays the total number of collocations found, which is 55 in @cross
and 36 in Passage. Since this number includes cases of one word with more than one
collocation, each word that was part of at least one collocation was counted, giving a total of
45 in @cross and 35 in Passage. When calculating the percentage of these words among all
single items (146 and 101), the result is 30.8% in @cross and 34.6% in Passage. However, in
the case of @cross, when only the words found in the texts are considered (128), the result is
35.1%. Since the first research question is to find out the percentage of words that are part of
collocations in the texts, only the second percentage (35.1%) will be taken into account.

Table 4.55 Summary of results from the two textbooks

@cross Passage Both
Type of text total 14 14 28
authentic - 7 7
constructed 14 4 18
mixed - 3 3
Items of total 156 102 258
vocabulary single words 146 93.6% 101 99.1% 247 95.7%
from the combinations 10 6.4% 1 0.9% 11 4.3%
exercises
single words not found in the texts 18 12.3% - - 18 7.3%
single words found in the texts 128 87.7% 101 100% 229 92.7%
Collocations total number of collocations 55 36 91
words that are part of collocations 45 30.8% 35 34.6% 80 32.4%
in the texts among all single words
words that are part of collocations 45 35.1% 35 34.6% 80 34.9%
among the ones found in the texts

As often commented between the tables in the previous sections, many of the combinations
found in the texts are included in COCA but not in the collocation dictionary (OCD), while
others are included in the OCD but not in COCA. As explained above, these are not
considered collocations in the present study. Figure 4.1 below shows that a total of 124
combinations from the texts were found in COCA (with frequencies above 3 and MI above
3.0); among these, only 91 were included in the OCD. Thus, 33 combinations have been
excluded by using the OCD, 17 in @cross and 16 in Passage. On the other hand, 95

87
combinations from the texts were found in the dictionary, from which 91 were listed in
COCA (with frequencies above 3 and MI above 3.0), which means that only 4 combinations
from the dictionary have been excluded by using the corpus. Thus, it seems like the corpus
has a wider range of collocations than the dictionary but that the dictionary focuses on the
most common words, probably the ones which are most pedagogically relevant.

COCA – 124
combinations

Found in both –
91 collocations

OCD – 95
combinations

Figure 4.1 Number of combinations in COCA and the OCD

Chapter 5 will give an interpretation of the results, suggest ways in which the exercises can be
improved and draw on research on collocations, corpora, and foreign language vocabulary
acquisition to propose ideas for vocabulary teaching.

88
5. DISCUSSION AND SUGGESTIONS FOR TEACHING

5.1 Introduction
The aim of this chapter is to analyze the results reported in chapter 4, interpret them in light of
the background in chapter 2, and discuss how these findings achieve the aims outlined in
chapter 1. I will offer suggestions on how the exercises can be improved and on other ways in
which collocations can be integrated into language teaching. These suggestions are not meant
to be prescriptive, but to broaden teachers and textbook designers‟ awareness of collocations.

5.2 First aim: Research questions and hypothesis

In this section the main findings of the investigation will be reported and the first research
question and the hypothesis will be addressed. Then, other interesting findings from the two
materials will be discussed. At the same time, these findings will be considered in relation to
the competence aims of the L06 in order to address the second research question.
The summary table in chapter 4 (4.55) shows that most of the items of vocabulary in
both books are single words, which was expected since these exercises have been selected
because they present lists of single words. The few items presented as combinations, 6.4% in
@cross and 0.9% in Passage, were not part of the search in the texts and corpus. What is
most interesting for the present study, however, is the percentage of words that are part of
collocations in the texts. In @cross, 35.1% (45 out of 128) of the single words in the exercises
appeared as collocations in the texts they are taken from, whereas the result for Passage is
34.6% (35 out of 101). These results enable us to achieve the first aim, which was to
investigate if words that are usually taught as separate items in textbook exercises are part of
collocations in the texts they are taken from. Addressing the first research question, these
collocations comprise more than one third of all the words taught as single items in @cross
workbook and the Passage website. Considering that the words from the exercises were not
selected with collocations in mind, one third can be regarded as a high proportion. The results
have thus corroborated the hypothesis that a large proportion of collocations would be found.
As explained in chapter 1, the corpus and dictionary have been used in this study as
methods to exclude free combinations and combinations less pedagogically relevant. Figure
4.1 shows that COCA included more collocations than the OCD, which is probably due to the
fact that the corpus contains a larger number of words than the dictionary. The latter only
includes entries for 9.000 words and it prioritizes nouns (McIntosh et al. 2009:v-vi).

89
 The next subsections will address the second research question – which of the
collocations would be useful for Norwegian students in the first year of upper secondary
school.

5.2.1 Discussion of @cross exercises

Some particular features stand out in the @cross material, two of which are that a number of
words from different exercises are in fact not in the texts and that many collocations found
were closely related to chapter themes. These and other findings will be discussed below.

Words not present in the texts

One of the first peculiarities of the @cross exercises is the fact that many of them practice
words which are not present in the texts indicated by the headings. For example, in exercise
[2] the heading is: „If you need help, look at pg [sic] 41-44 in @cross Textbook. There you
will find the words used‟ (Rodgers et al. 2006b:34), and in exercise [1], „Having read the text
on Berwick, find the English words corresponding to the following Norwegian words by
drawing lines‟ (Rodgers et al. 2006b:28). However, not all the words appear in the text, which
means that learners are expected to practice a completely decontextualized item. According to
Gass (1988; in Nation 2001:159), activities matching first or second language synonyms,
definitions or pictures (ten of the fourteen exercises in @cross are of this type), are exercises
to test if students have comprehended the input received, in this case, the text they have read.
If the word is not in the text, there is no input from which to learn.
The intention in some of the cases might have been to present to students other words
related to the same topic as those that appeared in the text. An example would be exercise [5]
which practices words related to media and government and adds the word anarchy that does
not appear in the text. The same happens in exercise [9] about aid work, where the words
responsibility, health, and dedication are added in the exercise without appearing in the text.
Exercise [13], which deals with education, adds electives and classes. Nonetheless, if the
intention of adding words in these exercises was to supply students with other items under the
same theme, this might in practice have a negative effect. Hoey talks about his experience
learning other languages using materials that worked with vocabulary in themed lists: „I can
only speak for myself as a very average learner, but I am conscious of quickly forgetting
themed lists. Because the words are learned without reference to any context in which they
might be used, they tend to get confused with each other‟ (Hoey 2000:227). Thus, although
most of the words in the exercises are in a way already decontextualized for being presented

90
as single words, it is even less productive when they are not part of a text where students can
check their meaning.

Repeated words
Another feature noticeable in @cross is that many words are repeated in different exercises,
which, according to research on language acquisition, is essential for the retrieval of
vocabulary (Schmitt & McCarthy 1997:276). Repetition is among the learning strategies
suggested by Nation (2001:218), as displayed in table 2.1, and he also argues that meeting the
word in new contexts aids in the process of generating language (2001:222). There are eight
words that have been repeated in exactly the same form in @cross in contrast with only two
in Passage. The reason why @cross seems to give more importance to recycling vocabulary
might be that the exercises in the workbook are meant as the main exercises, while in Passage
the website might be used as extra practice of what students have seen in the textbook.
Morgan Lewis attests that „learning is holistic, cyclical and evolves over time‟
(2000:12), which means that recycling is essential to internalize the meaning of a word. He
then suggests that teaching a word‟s collocates is a way of accelerating this process since it
reactivates and expands the meaning of a word by showing how it combines with new ones.
An illustration is the word election, found in the text for exercise [4] in the collocation
general election and then in the text related to exercise [14] as hold an election and as free
election. If the exercises had presented election in its collocations, students could have
widened their knowledge of this word at the same time as recycling it. Thus, teaching the
same word more than once and always providing some collocates can be a way of achieving
the competence aim of L06 which says that students should „master a wide vocabulary‟
(2006b).

Collocations with familiar words

Nation affirms that „[collocation exercises] can help learners expand their knowledge of
words that are already familiar to them‟ and that training collocations can encourage students
to widen the vocabulary they already know (1990:101). @cross contains some collocations
that fit this category, such as medieval art [1], coastal road [2], incredibly busy [3], trendy
restaurants [3], drinking water [9], pure water [9], and consider necessary [14]. The words in
bold are the ones presented by the exercise (supposedly new), and the others are the collocates
found in the text (words students probably know already). Teaching these two types of words
in a collocation might help students remember the meaning of the new ones as they learn

91
more about the collocational behavior of the ones they already know. This is confirmed by
Hill (2000:67) who argues that the main idea of vocabulary work should be to teach students
the collocations of the words they are already partly familiar with (see section 2.5.3).

Collocations related to chapter themes

In contrast with Passage, each chapter of @cross is based on a theme, which is displayed in
the table of contents of @cross textbook (Rodgers et al. 2006a:7-13). These themes can often
be recognized by looking at the collocations that have been found in the texts. For example,
exercise [5] is taken from chapter 8 where one of the themes is „British parliament and
politics‟ and some of the collocations found are: general election, constitutional monarchy,
Houses of Parliament, Members of Parliament, and elect representatives. Another case is
exercise [6] whose chapter has „education in the UK‟ as a theme and some of the collocations
found are: comprehensive school, national curriculum, pursue an education, and vocational
education. In exercise [7] the theme of the chapter is „higher education‟ and collocations like
study abroad, graduate student, and receive a scholarship have been found. Another example
is exercise [14], part of chapter 6, whose theme is „The American political system‟. The
collocations found are: pass/sign a bill, hold an election, free election, executive branch,
judicial branch, and legislative branch.
Some of the competence aims of L06 establish that students are expected to „choose an
interdisciplinary topic for in-depth studies within his or her own programme area and present
this‟ and „express himself/herself in writing and orally with subtleness, proper register,
fluency, precision and coherence‟ (2006b). To make a presentation about interdisciplinary
topics, students should first be able to select the appropriate vocabulary within these subjects.
Learning collocations related for example to politics, government, and school systems in the
US and the UK, as shown above, might help them complete their task with fluency and
accuracy. The OCD states that „a student who chooses the best collocation will express
himself or herself much more clearly and be able to convey not just a general meaning, but
something more precise‟ (McIntosh et al. 2009:v). Hence, learning collocations within subject
topics can prove convenient for students.

5.2.2 Discussion of Passage exercises

Two contrasting features that can be observed between Passage and @cross are that in
Passage all the words listed in the exercises are in fact in the texts and the units in Passage do
not contain any specific themes like in @cross. Other features which will be discussed further

92
in this subsection are: infrequent and formal words, the issue of constructed versus authentic
texts, the great number of grammatical collocations found, and the case of near synonyms.

Infrequent words
A common feature among many of the Passage exercises is the occurrence of infrequent and
more formal words (e.g. in exercise [25] retort, in [23] swoon and cleanse, in [18] skid and
shin, in [17] enunciate, in [15] commence, etc.) 10. This might be due to the fact that many of
the Passage exercises are based on literary texts, which does not happen in @cross. In any
case, this can pose advantages and disadvantages for the students. One disadvantage is to be
spending time learning words that they might not encounter again or have no need for.
Morgan Lewis (2000:14) stresses that there would be a great improvement in students‟
vocabulary if less time was spent on learning rare words and instead they would concentrate
in learning how to use the ones they already know in their collocations.
On the other hand, learning formal and rare words can be beneficial. According to
L06, students should be able to „read formal and informal texts in various genres and with
different purposes‟ and „write formal and informal texts with good structure and coherence on
personal, interdisciplinary and social topics‟ (2006b). Thus, in order to understand and write
formal texts, it is essential to have the necessary vocabulary. Nation notes that some
advantages of teaching infrequent words are: that they might contain a prefix, root, or suffix
that will help students when meeting other words; it may be a word that occurs a number of
times in a specialized text which is of interest to the students; or still „it may be a collocation
of a much more useful word‟ (1990:137). We can see some examples of the last case in the
Passage findings, such as cause consternation [16], clasped my hands [17], enunciate every
word [17], grow weary [18], took a sip [19], sipping tea [19], new-fangled ideas [26], and an
auspicious start [28]. The words in bold are the infrequent words brought up by the exercises
and the others are the collocates found in the texts. The former could be argued to be,
following Nation, collocations of other „more useful‟ words, in this case, cause, hands, word,
grow, take, tea, ideas, and start. In CALD all of these are marked as „essential‟ words, which
they define as „the most common and useful words in English‟ (2005:i).
Thus, one way or another, if infrequent or more formal words ought to be learned, it is
probably more effective when they are presented with one or more collocates.

10
In order to determine if the words mentioned in this subsection are frequent or infrequent I have looked them
up in CALD which contains a „Frequency Information system‟ (see CALD 2005:vii).

93
Authentic versus constructed texts
The issue of authentic versus constructed texts is particularly relevant in relation to some
examples found in Passage. As it can be noted from table 4.55, all the exercises selected in
@cross present words based only on constructed texts while Passage includes seven
authentic, four constructed, and three mixed. Whether this has had an influence on the results
concerning number of collocations is unclear since the proportion of collocates was
approximately the same in both materials. On the other hand, the combinations that appear in
constructed and mixed texts prompt some questions in relation to the type of input the books
provide for the learners.
A case in point is promise allegiance commented on in chapter 4, table 4.50. This
combination was found in a constructed text about English monarchs, but it is not listed in the
collocation dictionary and, in the corpus, it presents a raw frequency of 3 and an MI of 2.40
score (which is why it is not listed in the collocation table). If the corpus is searched to find
more about the uses of allegiance, pledge appears as the most frequent verb collocate, and the
collocation pledge allegiance shows a frequency of 529 and an MI of 12.52. Thus, the
question is: why would the textbook writers include such an infrequent and weak combination
(promise allegiance) instead of the stronger and more common collocation (pledge
allegiance)? The reason might be that allegiance is a new word for students and the writers
attempted to combine it with a word that students probably already knew (promise). Be that as
it may, pledge allegiance is what students will most likely encounter, as in the American oath
of loyalty to the republic: „I pledge allegiance to the flag of the United States of America
[…]‟, which is used in schools, government meetings, and even in sports events. These
findings could indicate that constructed texts tend to include weaker combinations than
authentic ones.
However, another example should be considered before suggesting any tendencies. In
the text related to exercise [24], the infrequent and weak combination color prejudice was
used instead of the stronger and more common racial prejudice. The former was included in
the dictionary and in the corpus, but its raw frequency was 8 and MI score 3.04; racial
prejudice, on the other hand, presented a frequency of 202 and an MI of 10.02 (see table
4.46). Then the question arises again: Why was color prejudice used instead of racial
prejudice? Unlike the case of exercise [26], this exercise is based on a mixed text so that parts
of autobiographies are combined with the constructed text and color prejudice was used by
Ghandi when talking about his experiences in America, which is the authentic part of the text.
It could still be argued that Ghandi was not a native speaker of English. Nevertheless, the

94
main point here is that there may be no assurance that authentic texts have stronger and more
common collocates than constructed ones. These, however, are very preliminary results and
should be investigated further to draw any definite conclusions.

Grammatical collocations
Another particular feature of the Passage results is the large number of grammatical
collocations found in the texts. Calculating the proportion of this type of collocation in both
materials, I found that 5.6% in @cross are grammatical collocations in contrast to 34.3% in
Passage. The ones from Passage are (exercise number indicated in brackets): detached from
[15], yell at [17], glanced at [19], pushed through [19], adapt to [20], emerge from [20],
apologize for [22], snatched from [22], negotiations with [24], resentment towards (which
appeared twice) [24] and [25], and persist in [25]. At first glance, it might not be clear why
Passage contains a larger proportion of grammatical collocations. However, it is observable
from the examples that most of these collocations are of the type V+P; thus, having a
suspicion that the presence of verbs in the exercises might have influenced the results, I
counted all the verbs appearing in the exercises from both materials and found that 10.5% of
the words in @cross are verbs while in Passage the proportion is 44.5%. Hence, it seems like
verbs tend to be in grammatical collocations more often than nouns and adjectives.
Grammatical collocations might not be so relevant to teach when they are congruent to
Norwegian, as for example adapt to (tilpasse til) and snatched from (tatt fra). On the other
hand, those which show a different grammatical structure should probably be taught, such as
apologize (to someone) for (something/doing something), (beklage seg for noe, literally
„apologize oneself for something‟) or persist in (fortsette med, „continue with‟). Emphasizing
collocations which are not obvious to students and highlighting the ones that might be
structurally different from the L1 can, as a matter of fact, be recommended regarding any
collocation and not only grammatical ones.
A note of caution is pertinent here relating to the percentages of verbs provided above.
The number of verbs given is only approximate since the distinction between verbs and nouns
are not always straightforward in the exercises, even when checking the synonyms given. An
example is exercise [23], where aid is the keyword and the synonym given is help; another is
[26] where the English word descent should be linked to the Norwegian translation stamme,
which can be a verb or a noun. In fact, some of the exercises, especially the two mentioned,
seem confusing due to the lack of context or clues. In exercise [23], the words are infrequent
and the synonyms do not give any indication of the part of speech they belong to. Moreover,

95
in one case the keyword and its synonym are written in different forms: fancy (supposedly a
verb) is to be matched with imagining. One of the competence aims of L06 is that students
should be able to „use relevant and precise terminology to describe the forms and structures of
English‟ (2006b). Therefore, it would probably be advantageous to include at least the part of
speech of the words.
A final comment on grammatical collocations is that their frequency tends to be higher
than the lexical type (see collocation tables from Passage), which is owing to the fact that
prepositions are a generally frequent class in the corpus.

Collocations of near synonyms

The great majority of the exercises selected in the Passage website are of the matching
synonyms type. While this kind of exercise can help students relate the new words with
known ones, the way they have been designed might pose some challenges to students. The
reason is that, different from @cross, which gives definitions, explanations, or collocations as
matching options, Passage provides mostly single words. This type of exercise might lead to
the same misconception as the translation ones (these are discussed in subsection 5.2.3), i.e.
that students may think that there is a one-to-one correspondence between synonyms. What is
clear from the exercises, however, is that the synonymous words provided have very distinct
uses and that corpus information can reveal their different meanings, as previous studies on
near-synonyms have found (see Kennedy 1991 and Partington 1998).
An example is the verb assassinate [25] that should be matched with murder. When
looking at table 4.48, which lists the ten most common collocates of assassinate, it becomes
clear that this word is used with „important people‟, as shown by the collocates president,
king, minister, and Pope. Indeed, as mentioned in chapter 4, the definition of assassinate
given by CALD is „kill someone important or famous‟ (2005:66). To attest the differences
between these two verbs, the ten most frequent noun collocates of murder were checked in the
corpus and the following nouns have been found (in order of frequency): wife, husband,
father, daughter, parents, son, thousands, brother, civilians, and sister. These results contrast
with the ones from table 4.48 in that only „common people‟ are listed, many of which are
family members. Hence, an important distinction between these two verbs is hidden by the
simple „match the synonyms‟ association required by the exercise. Although assassinate the
president or murder his wife are not considered collocations by this study, showing students
the collocates of the synonyms is certainly beneficial.

96
 Another example is the word notable, which is to be matched with the synonym
important in exercise [20]. The collocate table of notable, 4.39, can be compared to the table
of collocates of important (5.1) shown below:

Table 5.1 Collocates of the adjective important in COCA

Collocates Raw frequency MI
thing 4219 6.01
role 2563 6.58
part 2431 5.11
things 1268 4.01
factor 1105 6.88
point 1073 4.28
issue 1031 5.04
issues 960 5.16
question 776 4.10
aspect 682 7.29

Comparing both tables it becomes evident that none of the ten most used collocates of these
words is the same. In addition, the frequencies with the adjective important are much higher
while the MI scores are lower, probably for being a more frequent word than notable in the
corpus. The MI scores and frequencies displayed by table 4.39 indicate that the collocation
notable exception found in the text is highly relevant (most frequent and strongest in the
corpus) and could have been taught as a chunk in the exercise. It should be noted that in case
the words were to be taught in collocations, the model of the exercise would probably have to
be altered, which will be discussed further in section 5.3 below. The difference between these
two words could also be debated in terms of register or genre. A notable example, for
instance, could be taught as a more formal or more academic variation of an important point.
This could be brought up in class in relation to spoken and written language. One of the aims
specified in L06 is that students should „select appropriate listening, speaking, reading and
writing strategies adapted to the purpose, situation and genre‟ (2006b).
Other examples related to the issue of genre and register are provide and give in
exercise [21], where give is presented as the synonym of provide. Table 4.41 shows that the
collocates of provide seem neutral, such as information, support, care, and assistance. On the
other hand, when looking at the most common noun collocates of get and checking their
concordance lines in the corpus, we find informal expressions like get a chance, get in
trouble, get the impression, get a grip, and get your ass out of here.
Frank Palmer (1976) proposes five types of synonymous relations: register relation
(e.g. nasty smell vs. obnoxious effluvium), collocation relation (e.g. rancid butter vs. addled
eggs), connotation relation (e.g. thrifty vs. stingy), dialectal relation (e.g. autumn vs. fall), and

97
the category of partial synonyms (e.g. promise vs. pledge). The ones that are present in the
Passage exercises seem to fit in the categories of register (commence vs. begin [15], answer
vs. retort [25], loathe vs. hate [28]), collocation (cast vs. throw [18], push vs. shove [19],
notable vs. important [20]), and partial synonyms relations (sip vs. drink [19], tribe vs. nation
[20], assassinate vs. murder [25]). Pointing out these relations to students or asking them to
discuss what kinds of relations the synonyms in the exercise bear can help them more fully
understand the meanings of the words.
My contention here is that teaching the collocates of a word can help students
differentiate the meanings of near synonyms. As Carter affirms, „we should […] note that
synonymic relations between words can be usefully distinguished with reference to the
different collocational ranges of the synonyms involved‟ (1987:53-54).

5.2.3 Discussion of common features

Although @cross and Passage presented individual peculiarities, there are central aspects
which are common to both materials. The issues discussed in this subsection are connected to
translation exercises, unexpectedly uncollocated words, collocations which convey
sociocultural aspects, and collocations which evoke communicative situations.

Translation exercises
Among the twenty-eight exercises selected, one fourth is translation exercises: six in @cross
and one in Passage. Although making use of the L1 has its advantages when learning another
language (when for example drawing parallels or for true cognates), it might also be
problematic, as discussed in chapter 2 (section 2.3). Translation exercises of the type „match
the words with the correct translation‟ used in these two materials might give students the
false idea that there is a one-to-one correspondence between words in the two languages,
especially when only one translation is given for the English words, as in exercises [1], [6],
and [7] in @cross and [26] in Passage.
Using translation in class should not be strictly banned as it was in the Direct Method,
but a more effective way to do it would be to translate words as chunks into the students‟ L1,
particularly when a term is not congruent in the languages. Nesselhauf (2005:260) concludes
in her study of collocations used by German learners that more emphasis should be given to
non-congruent than to congruent collocations in teaching. This seems to be the case of the
collocations: general election [5], pursue an education [6], comprehensive school [6],
graduate student [7], different opinions [7], consider something necessary [14], and pass a

98
law or a bill [14]. These would probably be translated respectively as stortingsvalg
(„parliament election‟), ta en utdanning („take an education‟), grunnskole („base school‟),
akademiker/student („scholar‟, „student‟), ulike meninger („different opinions‟), se noe som
nødvendig („see something as necessary‟), and opprette en ny lov („establish a new law‟).
These collocations might seem simple in English but their translations are not literal and
many of them illustrate Swan‟s (1997) list of aspects that might interfere when translating
from one language into another (see section 2.3). For instance, the case of graduate student
would be an example of the different notions of a „word‟, since university or college students
are called simply student in Norwegian (the English word student is elev in Norwegian). So in
English the concept is realized in two words while in Norwegian it is realized in only one.
Another aspect that might generate confusion is false cognates, which is the case of different
opinions. This collocation seems fairly obvious in English and some might even question if it
is indeed a collocation. However, the translation to Norwegian is mening, which is a false
cognate, and students tend to utter sentences like my meaning is that... („My opinion is
that...‟). Swan (1997:159) also notes that words might have different grammatical contexts,
which is the case of consider necessary. The Norwegian version of this collocation needs the
preposition som („as‟), whereas in English the collocation has no preposition. Morgan Lewis
says that only after an item is shown to students in its bigger context is it safe to translate it
into the learners‟ L1, and „not word-for-word but whole phrase to whole phrase, bearing in
mind that the structure of the expression may be very different in one language from the
equivalent expression in the other‟ (2000:16). In the case of the collocations put up fierce
resistance in exercise [26], the literal translation in Norwegian would be „make hard
resistance‟ (gjøre hard motstand), which is not a collocation in English.
However, there are examples which are probably obvious for Norwegian students for
being congruent and for having one of the collocates as a true cognate, such as hold an
election (holde valg) [14], direct descent (direkte nedstamming) [26], and mark the boundary
(markere grensen) [1]. In such cases the L1 can be helpful. On the other hand, Hill et al.
(2000:94-97) report about a study on teaching collocations to business English students that
concludes that learners profited from positive feedback on their right use of collocations,
which enabled them to fully internalize what was only half-known. Thus, transparent
collocations can probably be omitted from exercises so that more unpredictable ones can be
given priority. However, when students use them correctly in their writing, a positive
feedback can help them internalize what might have been just a successful guess.

99
Unexpectedly uncollocated words
As seen in the summary table in chapter 4 (table 4.55), more than one third of the single
words presented in the exercises were part of collocations in the texts. Although my claim is
that many of these single items should have been presented in collocations instead, I do not
intend to argue that absolutely all need to be shown with a collocate. Cases like broadsheet
[5], anarchy [5], happy [10], evidently [15], forehead [17], orphanage [22], humiliate [22],
and bewildered [23], among others, are high information words (see Woolard 2000:33) and
not usually part of strong and frequent collocations11. It would be awkward for instance to
present happy with birthday, which is shown as the strongest collocate of happy in the corpus;
or present Russian orphanage, where Russian is the strongest and most frequent adjective
collocate of the noun orphanage.
On the other hand, there are words which occur repeatedly as part of strong and
frequent collocations but that for some reason are presented in the exercises as single items.
This is the case of environmentally in exercise [8], which quite unexpectedly appears without
any collocates. By looking at the collocate table (4.15), it is apparent that the weakest
combination, environmentally aware, is still strong (MI 7.86) compared to collocates in other
tables. In fact, after the word allegiance, this is the word with the strongest collocates among
all the twenty-eight investigated (see collocate tables). The word environmentally appears
3052 times in the corpus; in 545 of the cases it appears with friendly, 316 with responsible,
and 286 with sound, which means that more than one third of its occurrences are with only
these three collocates. Thus, I argue that there is sufficient evidence that this word should
have been taught as a collocation and not as a single item.
Another example that should be given attention is the word new-fangled in exercise
[26] of Passage. This word only appears 54 times in the corpus and in all of them followed by
a noun. Among these, it appears 6 times with the collocate idea(s), 2 with the word glass and
only once with all the other nouns. Taking into consideration that in the text where this word
is taken from the collocation new-fangled idea was found, it is surprising that the adjective is
presented as an individual item in the exercise. Michael Lewis maintains that „collocations are
not words which we, in some sense, “put together”, they co-occur naturally, and the first task
of the language teacher is to ensure that they are not unnecessarily taken apart in the
classroom‟ (2000:132).

11
As mentioned above, all of the words from the exercises were entered in the corpus to find their ten most
common collocates, so the assertion here that they are not part of strong and frequent collocations is based on
corpus information.

100
 A final interesting case is of auspicious start in exercise [28]. As commented in
chapter 4, auspicious start and auspicious beginning are the most frequent Adj+N
collocations in the corpus and the second and third strongest. Taking into consideration that
the meaning of these noun collocates is very close, that this adjective is uncommon, and the
fact that auspicious appears together with start in the text, my claim is that they should have
been presented as a collocation in the exercise.

Collocations evoking communicative situations

Hill et al. (2000:101) and Morgan Lewis (2000:15) emphasize that the communicative
potential of many collocations can be explored in the classroom. As Morgan Lewis affirms:

Many collocations have immediate pragmatic force or are situationally evocative. For
example, it is hard to think in which situation someone might say: This is a corner.
But if I say to you: This is a dangerous corner, it immediately suggests two people in
a car as they approach a corner where lots of accidents have happened. (Morgan Lewis
2000:15)

Morgan Lewis gives other examples of this kind, like routine check-up, disperse the crowd,
and widely available (see section 2.5.3), which can suggest, respectively, a visit to the doctor,
a police action after an incident, and talking about a new product. He argues that these real-
world situations must be brought to students‟ attention in class (Morgan Lewis 2000: ibid.).
Similarly, students could be asked in what kind of situation they would use or see the
following collocations found in the texts: incredibly busy [4], feel frustrated [10], delicious
meal [12], have a reputation [6], and hectic schedule [28]. Some ideas could be to use
incredibly busy, feel frustrated, and hectic schedule to talk about a busy lifestyle; a delicious
meal when complimenting someone on the food or thanking for the food; and have a
reputation when commenting on someone‟s behavior (e.g. He has a reputation for being
rude). These are only suggestions and students could use their creativity to imagine other real
life situations. They could also be asked to write a dialogue using these collocations and then
practice it with a partner. This kind of activity could be an aid to accomplish L06‟s
communicative aim of „take initiatives to start, finish and keep a conversation going‟ (2006b).
The central point is that, as research in vocabulary acquisition confirms (see section 2.3),
visualizing examples of the word and „creating contexts, collocations and sentences
containing the word‟ (Nation 2001:222) can foster retention and production of new
vocabulary. More ideas on how collocations can be integrated in teaching will be proposed in
section 5.4 below.

101
Collocations to talk about news and current events
L06 determines that students should be able to „present and discuss international news and
current events‟ (2006b) and some of the collocations found in the texts could be helpful to
accomplish this aim. When preparing for a news presentation, students could use high
unemployment [4] to talk about the job market; cause consternation [16] to talk about some
polemic event or issue; or put up fierce resistance [26] to talk about a rebellion against the
government, for example. Certain collocations are very restricted to the language of news and
advertizing and they should also be given some attention in class, such as hit the headlines
and innovative features respectively (McCarthy & O‟Dell 2002:124-126).

Collocations revealing various meanings of a word

Some of the collocate tables with words from the exercises reveal that different collocates can
show a word‟s various meanings. For instance, the collocates of bright (starts, red, future,
etc.) listed in table 4.5 show that this word has the meanings „full of light‟, „strong in color‟,
and „full of hope for success or happiness‟, besides the „intelligent‟ meaning like in bright
student (CALD 2005:152). Thus, teaching students some of the collocational fields of bright
(„things that shine‟, „colors‟, „people‟, etc.) can help them use this word.
Another example is seen in table 4.29, where the verb collocates of slack show the
different meanings of this noun. The concordance lines of the collocates show that this word
is used in expressions such as pick/take up the slack, cut him/her some slack, his/her
face/jaw/fingers…went slack, and dressed in black/dark/pressed…slacks. Another example is
the collocates of snap from exercise [19] shown in table 4.37: snap your fingers, snap
pictures/photos, snap peas and beans (which are compounds), and snap to attention. All these
examples are parallel to Hunston‟s list of collocates of the word shed (2002:12; see section
2.5.4) and to her argument that collocational information can bring forward the different
meanings of a word (2002:76). Hence, teaching the collocates of a word and their
collocational fields can aid students in using the words they are taught.

Collocations conveying sociocultural aspects

Maybe one of the most interesting findings in this study has been how collocations can reflect
sociocultural aspects of the American society. In this study I use the term „culture‟ in a broad
sense to mean behaviors, ways of living and thinking, traditions, historical facts, and current
events in a society. Investigating how these aspects are reflected in the language is a vast field

102
of research. However, as discussed in chapter 2, not so much has been done connected to how
collocations in English can reflect the culture of its native speakers.
A first example is the word suburbs from @cross. Some of its collocates listed in table
4.4 are especially interesting: white, affluent, wealthy, and middle-class. These reveal that
people living in the suburbs are probably white, rich, and middle-class. The collocations
conveying economic status are the strongest ones – affluent with an MI of 10.40, middle-class
with 8.72, and wealthy with 8.23. If we look for the collocates of neighborhood for example,
we find among the ten most common ones black, working-class, poor, and tough. This kind of
information provided by the corpus can foment a discussion about the class system in the US
and how the different classes are spread around a city. Another example is coverage shown in
table 4.9 in exercise [5]. The collocates of this noun indicate that this word is commonly used
to talk about health insurance, like full coverage and medical coverage. These collocations
can incite a discussion about the health system in the US.
Another thought-provoking example is the collocates of the word prom, displayed in
table 4.19 of exercise [10]. Firstly, the prom tradition might be interesting to discuss since
Norwegian students have nothing of the type, so collocates like senior, junior, and high
school could be useful in this context. Secondly, what might be striking to Norwegian
students are the collocates white, integrated, and separate, which show that in many places in
the USA blacks and whites still have separate proms. Looking at the concordance lines of
these three words, the following sentences are found: „Saltzman's fine documentary "Prom
Night in Mississippi" tells about the first integrated prom in the history of Charleston high
school - in 2008‟; „White students plan the white prom, and black students plan their own‟;
and „[…] a black boy tried to get into the white prom to see his friend, and they told him he
had to leave, and they had him escorted out‟ (Davies 2008). These would certainly be relevant
to bring to students‟ attention in the context of school and racial issues in the US.
Another example worth mentioning is the collocates of the adjective attractive in table
4.23, exercise [12]. What these collocates indicate is that women are referred to as attractive
much more than men: woman has a raw frequency of 260, women 85, and girl 49, totaling 349
hits, against 89 applying to men. This aspect is probably not only related to the American
society but to the general belief that women‟s appearance is more important than men‟s. This
could be compared for example with the collocates of smart, which show guy as the second
most frequent collocate while the first word designating a female person (girl) only appears in
the fourteenth place.

103
 A final example is the high frequency of the collocate center with the node trade in
table 4.17, exercise [9]. This is an indication of how events, like the attack on the World
Trade Center, can influence the language. As commented in chapter 2, it would be interesting
to investigate in the corpus the concepts that were often used after the attack, such as war on
terror, anti-terror campaign, terrorist attacks, terrorist networks, etc. (see section 2.5.5). Due
to space constraints I have looked only at the most frequent noun collocates of the word
terrorist and found the following collocates (raw frequency and MI score indicated
respectively in parenthesis): attacks (2207, 11.59), groups (755, 8.18), organizations (402,
8.85), and threat (267, 8.29). Also interesting is that the raw frequency of the word terrorist
from the year 2000 to 2002 has risen from 243 to 2001 tokens.
In sum, it seems like the corpus is a good asset for teachers when discussing
sociocultural differences between English speaking countries and Norway. However, since
this corpus is of American English, most of the information reveals features related to the US.
Using the British National Corpus (BNC) would probably be a better option when discussing
characteristics of the British culture. For teachers wanting to use the BNC, it is available
online with the same interface as COCA, so all the corpus tools explained in the present study
can be employed to search the BNC. It should be noted that knowledge of culture and society
is central in L06, as some of the competence aims reveal: „the pupil shall be able to discuss
social conditions and values in various cultures in a number of English-speaking countries
[and] present and discuss international news and current events‟. Also under the „main subject
areas‟ it is explained that „working with various types of texts and other cultural expressions
is important for developing linguistic skills and understanding how others live, and their
cultures and views on life‟ (L06 2006b). Calling attention to culturally-loaded collocations in
these texts would be a way of achieving some of the objectives of the Norwegian curriculum.

5.2.4 A summary: What kinds of collocations are useful for students?

This section summarizes what types of collocation might be useful to teach Norwegian
students in the first year of upper secondary school. The suggestions are based on the findings
described above, research on collocations and foreign language vocabulary acquisition, and
on the competence aims of L06. Thus, answering the second research question, the types of
collocations found in the books which would be useful for Norwegian students are displayed
in table 5.2 below:

104
Table 5.2 Useful collocations for Norwegian students among the ones found in the texts
Types of collocation Examples
with words that appear repeatedly in the exercises general election, free election, hold an election
with words that are already familiar to students but incredibly busy, trendy restaurants
which expand the knowledge of the known word
related to chapter themes constitutional monarchy (government in the UK),
vocational education (education)
when teaching infrequent words, those containing grow weary and auspicious start
the infrequent word and a known word
frequent and strong in corpora pledge allegiance, environmentally friendly
collocations of near synonyms to avoid using them notable example and important point
interchangeably independent of context
non-congruent with the Norwegian equivalent put up fierce resistance, graduate student
evoking communicative situations hectic schedule, feel frustrated
useful to talk about news and current events high unemployment, cause consternation
revealing various meanings of a word bright future and bright light, cast a vote and cast
doubt
conveying sociocultural aspects (These do not need affluent suburbs, white prom
to be directly taught, but brought to students‟
attention in order to foment discussions about
cultural issues.)

Thus, it seems like most of the collocations found in the textbooks would be useful to teach
Norwegian students and, most importantly, that there is an even greater variety yet to be
explored in class. My claim, however, is not that traditional vocabulary exercises should be
merely turned into collocation exercises. Although they can be improved by collocational
information, as will be discussed in the next section, I argue for a change in the way we view
language. As Hill observes, language is „a predominantly lexical phenomenon‟ (2000:47), and
this view of language should be reflected in foreign language teaching and learning. This does
not mean that we should teach students all the collocations of English, but that awareness of
collocations should be encouraged as a learning strategy. Woolard maintains that it is
„essential […] that the teacher equips the students with search skills which will enable them to
discover significant collocations for themselves, in both the language they meet in the
classroom and, more importantly, in the language they meet outside the classroom‟ (2000:34).
This statement is particularly relevant for the context of Norwegian upper secondary students
who probably meet English words more often outside than inside the classroom.

5.3 Second aim: How the exercises can be improved

It is paramount that research on corpus linguistics and particularly on collocations finds its
way into the classroom. The second aim of the present study is to suggest how the vocabulary
exercises can be improved and complemented by previous research and ideas from existing
collocation exercises. As mentioned in section 2.3, the exercises selected from @cross and

105
Passage are the type used to consolidate word form and meaning in memory, which is the
fourth step in the vocabulary acquisition process (Hatch & Brown 1995). This section will
give suggestions on how the vocabulary exercises discussed above can be improved using
corpus information so that they will still work at consolidating form and meaning at the same
time as they will take students a step further in the learning process: how to use the words.
The suggestions are also based on previous research and teaching materials on collocations.
Woolard affirms that „students with limited time available for study will not learn high
priority lexis if it is not deliberately selected and incorporated into learning materials.
Collocations, then, must become part of that planned language input‟ (2000:32). One way to
turn collocations into part of the input is including more common and stronger collocations in
the texts constructed for the textbooks. As discussed above, the collocation pledge allegiance
could have been used in the text instead of promise allegiance.
Relevant to how the exercises can be improved, it is generally accepted that students
learn best when words are practiced in context (Paribakht & Wesche 1993); what I would like
to add is that teaching the collocates of a word is also a form of contextualization and should
not be regarded as less important. On these grounds, and inspired by the example of
McCarthy et al. (2006a:48; see appendix 2) exercise [6] in @cross could be redesigned as
shown by figure 5.1:

[6] Find these words and expressions in the text. Match them with the correct translation.
1. pursue an education 7. academically giftet
2. vocational education 8. comprehensive school
3. national curriculum 9. compile (v)
4. enhance (v) 10. Beacon school
5. evolve (v) 11. coaching (n)
6. build up a reputation
Figure 5.1 Suggestion of how to improve exercise [6]

Using translation in this exercise is not a problem as long as it is a translation of the whole
expression or collocation (see Morgan Lewis 2000:16, section 2.3), but another option would
be to list definitions in English. What has been changed here is that the collocations found in
the text were added, the combinations that were there from before were kept, and the single
words which did not have collocates were given parts of speech.
Another suggestion is to turn a simple translation exercise like [14] into a collocation
exercise where the collocates are to be matched (if in doubt students can check them in the
text) and then translated to Norwegian. This kind of exercise practices collocations, uses the

106
context of the text to help students guess meanings, and makes them think about the concepts
as a unit in Norwegian. The suggestion is presented in figure 5.2 below:

[14] Match the words on the left with their collocates on the right (some of the collocates can be used with
more than one word). Then write them down as in the text and then translate them into Norwegian.
1. pass a. election
2. consider b. branch
3. hold c. law
4. executive d. necessary
5. legislative e. bill
6. free
7. judicial
8. sign
Figure 5.2 Suggestion on how to improve exercise [14]

In the exercise in figure 5.2, only the collocations found in the text were used; however, this
exercise would be more challenging if other words and collocations related to government
were also added.
In exercises like [19] where some of the key words have different meanings
determined by their collocates (e.g. snap, push, and tug), these collocates can be presented
and students are supposed to find the words in the text, as shown in figure 5.3 below. As a
follow up, students can discuss in pairs the different meanings conveyed by the collocates.
This exercise is inspired by one in Michael Lewis (2000:112; example 2 in appendix 2).

[19] Try to find in the text the verbs that complete the following collocations. The same verb completes all
three examples. Thereafter, discuss with a partner the different meanings of each word.

1. ……………. pictures 2. ……….…. the button 3. ……………… of war

…………….. to attention ……………. up princes ……………… at somebody‟s heart
………..…… your fingers ……………. the limits an emotional ………….
Figure 5.3 Suggestion on how to improve exercise [19]

Another suggestion is one related to exercise [28] in Passage. Three words from this exercise
were part of collocations in the text (hectic schedule, auspicious start, and transfer fees) and
the word outlet was checked in the corpus for its most frequent collocates (see table 4.54).
One way to adapt this exercise to include collocations could be to group their collocates and
ask students to match the key word with the group of collocates, as in an exercise made by
Michael Lewis (2000:115; see example 3 in appendix 2). In case students need to check the
meaning of the words first, they can look at the glossary provided next to the text in Passage
(Sørhus et al. 2006a:311-314). Checking the meaning would be part of the third step in the
vocabulary acquisition process according to Hatch & Brown (1995, section 2.3) and the

107
collocation exercise suggested in figure 5.4 would aid in achieving the fourth step, which is
„consolidating form and meaning to memory‟. In order to reach the fifth step, „using the
words‟, students could be asked to write sentences or short dialogues including the
collocations. The suggestion is shown in figure 5.4 below:

[28] Match the following words with their group of collocates. One of the words does not match any groups.
auspicious canny sack fee hectic revenue loathe

a. entry, registration, membership, admission, school

b. beginning, start, opportunity, event, day
c. tax, oil, advertising, government, export, sales
d. schedule, day, pace, life, work, lifestyle
e. businessman, politician, marketing, move, eyes, smile
Figure 5.4 Suggestion on how to improve exercise [28]

The collocates listed in figure 5.4 above were all taken from the corpus. They were not
checked in the dictionary since many of the node words are not included in OCD.
Some of the vocabulary exercises have a potential to help students internalize new
collocations and discuss cultural differences. Nesselhauf says that „the topic of an essay could
serve as a trigger for teaching a number of collocations […] that occur frequently in
connection with this topic‟ (2005:265). She also recommends that the collocations are
introduced before students write the text. Hence, exercise [10], in which students have to
write a text about American teenagers, could be supplemented so that certain collocations are
discussed first in relation to the American teenage culture and then students can use them to
write a text about the topic, as illustrated in figure 5.5.

[10] The following collocations have been found in an American corpus. What do they tell you about
teenagers in the US and the American culture?

a. sports hero d. shopping lists

b. senior prom, white prom e. fashion trends, latest trends
c. feel frustrated f. bullying prevention

Now write a paragraph describing American teenagers and their situation using the collocations
above and the following words:
weird, wired, committed, happy, cliques
Add other words from the texts which you think fit in.

Figure 5.5 Suggestion on how to improve exercise [10]

In the text related to exercise [10], the following combinations were also found: wired
teenagers, member of a clique, and clique identity. They could also be discussed since they
convey the close relationship of the new generation with technology and the importance of

108
belonging to a clique in American high schools. Something similar to the suggestion in figure
5.5 could be done in exercise [9] so that students would have to use collocations like medical
aid, cancel de debt, drinking water, etc. (see table 4.16) to write the text about aid work.
Exercise [25] asks students to match synonyms. The heading is „Join up the words that
mean the same – or nearly the same‟ (Sørhus et al. 2006b) and some of the synonyms are:
murmur and mutter, persist and continue, select and choose, assassinate and murder, retort
and answer. This exercise could be complemented so that after students match the synonyms,
they could try to find out the differences between them. Nesselhauf recommends that in more
advanced levels the simple matching synonyms activities should „be replaced by exercises
actually focusing on the differences in meaning and usage of near-synonyms‟ (2005:266).
Whether or not students at the first year of upper secondary school in Norway are advanced is
disputed. What is certain is that changing the focus of the exercises would add some challenge
to the activity and give students a deeper understanding of the uses of the words they are
learning. Since this exercise is on the website, one way to redesign it would be to hyperlink
the words so that students could click on them and see some information taken from corpora
to be analyzed. They could discuss in pairs and draw conclusions about the various uses of the
two words. Hatch & Brown show that this type of activity is possible using programs like
Hypercard or Toolbook. They explain that by clicking or touching a word, grammatical or
cultural information, a translation, or the meaning of the word is provided (Hatch & Brown
1995:408). It should also be possible to include corpus data so that information as shown in
figures 5.6, 5.7, and 5.8, could pop up when students clicked on the words. The heading of the
exercise could be: „Look at the corpus information and discuss the differences between the
following synonyms‟. Figure 5.6 below show the most common collocates of the words
assassinate and murder.

assassinate murder

Figure 5.6 Lists of collocates of the synonyms assassinate and murder in COCA

109
Based on this information students will probably conclude that the difference between
assassinate and murder is mainly the type of people killed and the motivations for killing.

answer retort

Figure 5.7 Use of the synonyms answer and retort in different sections in COCA

Looking at the charts in figure 5.7, students might conclude that answer is more used in
spoken English while retort is mostly used in fiction, and that answer is much more frequent.

persist

continue

Figure 5.8 Concordance lines of the synonyms persist and continue in COCA

With reference to persist and continue in figure 5.8, students will probably see that persist is
more commonly followed by the preposition in and continue by to. These are only some
examples of how corpus can help students discover language patterns. L06 states that pupils
must „be allowed to choose tasks which will challenge them and give them the opportunity to
explore, both alone and together with others‟ (2006b). Obviously, the textbooks do not need
to present the raw data as displayed here, the information could be summarized and provided
in other ways. The difference is that when observing raw data to draw conclusions about how
language works, students learn inductively (discovery learning) and when the data is
interpreted for them the learning is deductive (see section 2.6). In case students find working
with raw data overwhelming, the exercise can be constructed simply informed by corpus. The
main idea is that corpus information is used.

110
 Turning now to grammatical collocations, McCarthy et al. (2006b:60) include an
exercise in which students have to choose the correct preposition to complete a statement and
then write if the statement is true of false (see example 4 in appendix 2). Thus, it works with
collocations and relates them to familiar experiences. This type of exercise could be a follow
up of a matching definitions exercise containing verbs that are part of grammatical
collocations. Since this was not the case of any of the exercises investigated, I have used two
verbs from exercise [21] to exemplify how this could be done, as shown in figure 5.9 below:

[21] Circle the correct prepositions to complete each sentence. Then write T (true) or F (false) depending on
whether the sentence is true or false for you. Correct the false statements.

1. ________ I always adapt well (in / to / at) new environments.

2. ________ I think that the parents should provide (with / to / for) their children until they can get a
job, no matter their age.
Figure 5.9 Suggestion of a grammatical collocation exercise based on exercise [21]

Textbooks can also supply tips about common usage based on corpus information. In
Touchstone (McCarthy et al. 2006a), there is usually a „curiosity‟ box in the vocabulary
section where some data from corpora is provided (see example 5 in appendix 2). This could
be done for example in relation to exercise [1] from @cross. In this exercise, the first word to
be matched with its translation is attractive and the exercise is about a town in the UK. A
curiosity box could be displayed next to the exercise, as follows:

[1] In the text about Berwick, the adjective attractive is used to describe a town. Here are the words most used
with attractive in a corpus in order of frequency:

1. attractive woman 5. attractive women

2. attractive alternative 6. attractive place
3. attractive option 7. attractive girl
4. attractive man 8. attractive feature
Figure 5.10 Suggestion of a „curiosity box‟ using the adjective attractive

This approach might be effective in getting students more interested in corpora without
overwhelming them with frequencies and MI scores. Since the text is about a town in the UK,
information from a British corpus might be more appropriate here. Many other suggestions
could be given exploring all the potential of the corpus.

5.4 Third aim: Other ways of integrating collocations in language teaching

In addition to the suggestions given above, other activities can be prepared by teachers or
offered by textbooks in order to include collocations in language teaching. Some options are,

111
as provided by Conzett (2000; see appendix 2, examples 7 and 8), make an „odd one out‟
exercise with collocations; give a few collocates of a word and ask students to add others;
make thematic collocation lists; and play collocation dominoes (writing part of a collocation
in each end). Hill et al. (2000) recommend brainstorming collocations on a topic before
students write a text about it; bringing to class an extra text on the topic of the essay where
students can find useful collocations; exercises that teach alternatives to very (e.g. highly
qualified and bitterly disappointed); matching collocates and using them to fill in gaps; etc.
One of my suggestions is to give students a list of collocations that convey cultural
aspects to be discussed when appropriate to the theme of the lesson or the chapter. Another
idea is to ask students to write down confusable words like do and make and dictate collocates
to be written under the correct verb. Howarth says that learners tend to avoid collocations
with delexical verbs (take, make, give, do, etc.) „due to uncertainty over appropriate
collocability‟ (1998:181); thus, working with these verbs can be a way of addressing the
problem (see example 6 in appendix 2). Students can also be given sentences from their own
writings which contain collocational mistakes to be corrected. In this case they might need to
consult a corpus or collocation dictionary, which is a good opportunity to teach how to use
them.
Teachers can also call attention to important collocations that appear in the texts from
the textbook, as suggested by Conzett (2000:72; see section 2.4). The relevance can be
according to, as discussed above, lesson or chapter topic, a grammatical point, a distinction
from a false cognate in Norwegian, or the need of the collocations to write a text. Another
option is to suggest collocates on the margin of students‟ essays when they have written an
odd word combination. Woolard says that „one obvious way of finding out which words our
students do not expect to find together is through the mis-collocations they make in their
production of language‟ (2000:30). If a student has written for example high poverty in an
essay, the teacher can write on the margin extreme/severe/absolute poverty. The corpus can
also be used to check students‟ doubts about word uses or to illustrate a grammatical point.
Dealing with corpus examples can encourage discovery learning, so that students can infer the
rules by looking at usage-based examples. If the teacher finds the concordance lines too
demanding for students due to the amount of unknown words, he or she can previously edit
the concordance. In a study by Chambers, one of the students affirms: „I discovered that
achieving results from my concordance was a highly motivating and enriching experience.
I've never encountered such an experience from a textbook‟ (2005:120). Hence, the corpus
can be used as an asset by teachers to improve the classes and to motivate students.

112
 An idea suggested by Nation (2001:302) is to encourage students to keep word cards:
the word to be learned can be written on one side and the meaning on the other and drawings
can be included. Although Nation suggests that the cards should not carry too much
information, I argue that collocates and example sentences should also be added to lead
students from the fourth step in vocabulary acquisition – consolidating form and meaning to
memory – to the fifth, knowing how to use the word (see Hatch & Brown 1995). Woolard
(2000:43) suggests that students keep a lexical notebook which can be divided by situations
(e.g. at the bank), functions (e.g. complaining), and topics (e.g. occupations). When students
encounter the words again, they can add a sentence with the new context in their notebooks as
a way of recycling and expanding their vocabulary knowledge.
McCarthy & O‟Dell (2005) have compiled a resourceful book called English
collocations in use that can be used by teachers to complement vocabulary work. It includes
collocations taken from the Cambridge International Corpus which are taught grouped by
themes such as „travel‟, „relationships‟, „sport‟, „music‟, „study and learning‟, „work‟, etc.
They are first presented in context and can then be practiced through exercises. Learning tips
are also given based on analysis of learner errors from the Cambridge Learner Corpus
(McCarthy & O‟Dell 2005:3).
In conclusion, course material and teachers should help students perceive the lexical
patterns existing in the English language and point out that learning these patterns can
develop fluency. Conzett states that what is most important for teachers, „more than worrying
whether or not something is a collocation, is to shift their and their students‟ focus away from
individual words to chunks of language. These chunks improve the fluency and accuracy of
the English students produce‟ (Conzett 2000:80). Teaching students to notice collocations in
the language input they receive and to have a system to record them are some of the best
learning strategies that teachers can equip learners with.

5.5 Summary
This chapter has attempted to discuss and interpret the results of the investigation taking into
account the background in chapter 2 and the three aims outlined in chapter 1. Concerning the
first aim and first research question, the percentages from table 4.55 reveal that more than one
third of the single words from the exercises which appeared in the texts are part of
collocations. This confirms the hypothesis that a high proportion of collocations would be
found in the texts. Some peculiarities from each material have been discussed separately
while the features observed both in @cross and Passage have been interpreted together. The

113
findings have shown that collocations can be useful to widen students‟ vocabulary and also to
provoke discussions about sociocultural matters. As a means to address the second research
question, a list of useful collocations for Norwegian students in the first year of upper
secondary school has been suggested. As for the second aim, ideas for improving the
exercises have been given based on previous research and on existing collocation exercises.
Finally, achieving the third aim, ideas have been given on other ways in which collocations
can be integrated into language teaching.
In the present study a measurable definition of collocation was needed in order to
make quantitative claims. Consequently, combinations that were potentially useful for
Norwegian students have been excluded either because of the information provided by the
corpus or the dictionary. As shown in figure 4.1 in the results chapter, 33 combinations have
been excluded by the use of the dictionary and 4 have been left out by the use of the corpus.
In practice, the theme being discussed, students‟ L1, and cultural relevance are probably more
important than whether a combination is or is not a collocation.
Many of the ideas discussed in the present study, and especially in this chapter, are
connected to Beck et al.‟s concept of „rich instruction‟ for vocabulary teaching. They
emphasize the importance of relating words to one another and to personal experiences, and
say that discussing words is crucial for the learning process. This is expressed by their
quotation below:

Although the instruction would need to include associating words with definitions, it
would need to go well beyond that. […] Students should be required to manipulate
words in varied and rich ways, for example by describing how they relate to other
words and to their own familiar experiences. To promote and reinforce deep
processing, activities should include much discussion of the words and require
students to create justifications for the relationships and associations that they
discover. This feature we labeled rich instruction. (Beck et al. 1987:149, added
emphasis)

All in all, this study has attempted to give an idea of the range of collocations that appear in
textbooks and show their relevance for vocabulary teaching. Furthermore, it has sought to
stress the potential of corpora for teaching and to show how it can be used to complement
vocabulary exercises, which is especially relevant in the Norwegian context where computers
and internet are readily available. It has also tried to emphasize the importance of integrating
collocations not only in vocabulary practice but also in English teaching in general.

114
6. CONCLUSION

This chapter will summarize the findings of the present study and draw conclusions based on
what has been discussed above. Suggestions for further research will be offered and the
chapter will conclude with the main findings of the investigation.

6.1 Summary and conclusion

This thesis has attempted to show how pervasive collocations are in the English language and
their importance for the teaching of English as a foreign language. In order to raise awareness
of the phenomenon, the first aim was to investigate two textbooks for Norwegian upper
secondary students to find out if the words presented in isolation in the vocabulary exercises
were part of collocations in the texts they were taken from. Recapitulating, the first research
question was:

(a) What proportion of these words appears in collocations in the texts?

Based on the results presented in chapter 4 and discussed in chapter 5, it has been found that
34.9% of all the isolated words in the selected exercises from @cross and Passage were part
of collocations in the texts they were taken from. The individual result for each textbook was
35.1% in @cross and 34.6% in Passage. This shows that more than one third of all single
words from the exercises appeared in collocations in the texts, which confirms the hypothesis
that a high proportion of collocations would be found. Although what is a high or low
proportion is difficult to define, we can conclude that above one third is more than one would
normally expect given that the choice of words for the exercises did not take collocations into
consideration. The next research question was:

(b) Which of these collocations are useful for Norwegian students in the first year of upper
secondary school?

The collocations found in the texts were considered in relation to previous research combined
with the competence aims of the Norwegian curriculum (L06). A list was provided in chapter
5 indicating which kinds of collocations would be useful for teaching. Some of these are:
collocations related to the themes of the textbook, collocations of near synonyms, collocations
to talk about news and current events, and collocations conveying sociocultural aspects. The

115
findings related to the first aim can be taken as evidence to claim that at least some of them
should have been listed as chunks in the exercises.
Fulfilling the second aim, findings on corpora, collocations, and foreign language
vocabulary acquisition have been used together with existing material on collocations to
improve and complement the exercises selected from the books. The third aim has been
achieved by suggesting other ways in teachers and textbook writers can integrate collocations
into language teaching, particularly in the area of vocabulary.
Although this study deals with material designed for Norwegian students in the first
year of upper secondary school, I argue that some of the implications can be transferred to
students of other nationalities and at other levels. For example, many of the types of
collocations that are recommended for teaching can be extended to any learner of English:
collocations which are frequent and strong in the corpus, collocations revealing various
meanings of a word, collocations that convey sociocultural aspects of language use,
collocations of near synonyms, collocations of words that are already familiar to students but
which expand the knowledge of the known word, and the ones related to themes being
discussed in class (e.g. politics, environment, education, etc.). In addition, the conclusion that
vocabulary exercises can be complemented by corpus information can also be applied to any
English teaching context.
In sum, this thesis has detected a large amount of unexplored collocations in texts
from textbooks that have the potential to improve textbook vocabulary exercises and to
expand student‟s knowledge of single words. The present study has also attempted to show
that corpora can be used as a resource to complement language teaching in general and that
collocational knowledge can help students communicate more successfully and fluently.

6.2 Further research

Some interesting findings have been made in the course of this investigation which could be
further explored. A future study could be to use the corpus to find the most common
collocates of all near synonyms provided by the exercises and check the concordance lines for
each of them. This would certainly lead to new findings about the use of these words. Other
words that could be researched are synonyms that are problematic for students, such as wound
and injury as exemplified by Morgan Lewis (2000:13; see subsection 2.5.4).
A study that would be valuable for language teaching would be to investigate
collocations common to different topics or themes usually discussed in the classroom. Lists of

116
salient collocations could be compiled (see example 7 in the appendix) and recommendations
could be given regarding how these collocations can be used in class.
The collocations given most attention in the present study were the ones found in the
texts of the textbooks and it has not been discussed extensively what other specific
collocations would have been more relevant to include in the exercises. The reason why the
ones found in the texts were given prominence is the fact that students have encountered them
already in the input, and including them in the exercises would help the input become intake.
However, there are probably other useful collocations to be taught, as the ones suggested by
Nesselhauf (2005:237-273) in her study of deviant collocations produced by German students.
How these collocations can be integrated into the syllabus can be a topic for further research.
Another finding that could be studied in more depth is the sociocultural aspects
conveyed by certain collocations. This could be carried out in connection with the general
studies program for students of the second and third year of upper secondary school. Their
main subject areas in English are „international English‟, „social English‟, and „English
literature and culture‟. The textbooks for this course and/or various corpora could be used to
find examples of culturally-loaded collocations that can serve as starting points for
discussions in the classroom about other English speaking countries. It would also be
interesting to conduct a study similar to the one by Teliya et al. (1998) discussed in chapter 2.
Studying stereotypes, cultural concepts, connotations, and conceptual metaphors in English
collocations seems to be a large and underexplored topic of research.
Yet another area that could be pursued in further studies is the teaching of English for
the specialized courses in Norway – the vocational education programs. Although the
competence aims established for these courses are the same as for the general studies,
teaching specific collocations for the areas of restaurant and cooking, health and social
studies, media and communication, among others, would be of great benefit to these students.
Small corpora could be created by scanning texts within these fields and concordance
programs available online could be used to look at concordance lines, collocations,
frequencies, etc. (see Gavioli 2005). There is also a program online called „Sketch Engine‟
(http://www. sketchengine.co.uk/; accessed on November 16, 2009) which can be used to
build your own corpora. Using certain genres and subgenres of COCA is another option.
An investigation could also be conducted applying the ideas proposed by the present
study. If the exercises suggested are further developed, they could be used in class and
students‟ vocabulary knowledge could be tested. Corpus consultation can also be
implemented. Testing students taught in more traditional ways and then comparing the results

117
can reveal if collocations indeed contribute to vocabulary learning. In addition, classroom
procedure could be recorded and teachers and students could be interviewed regarding the use
of corpora. In line with Yoon, „we need an empirical report from actual teaching that uses
easily accessible general corpora to encourage teachers and students to use the new corpus
approach‟ (2008:32-33).
Finally, the treatment of vocabulary in the textbooks analyzed should not be judged
only by the findings of the present study since the material selected comprises a small amount
of all the vocabulary work provided by the book and website. A topic for further research
would be to include other types of exercises in the investigation. Furthermore, the new
editions of @cross and Passage could be investigated to find out if the vocabulary exercises
have been changed and if collocations have been given a more central place.

6.3 Final remarks

The interest in collocations has increased in the last decades and considerable research has
been done to find out about the collocational behavior of words. However, there seems to be a
gap between what has been discovered and the application of the findings in language
teaching. According to Meunier & Gouverneur, „studies on the actual treatment of
phraseology in ELT material are rare. Information on the selection of learning and teaching-
prone formulaic sequences is nowhere to be found, and precise guidelines on how to teach
formulaic sequences is just as scarce‟ (2007:121). This thesis attempts to fill this gap by
providing new insight into the treatment of collocations in language material, by
recommending types of collocations that can be taught, and by showing how these can be
integrated in vocabulary exercises and teaching.
The findings of the present study can be summarized in four main points: (1) single
words taught in the textbooks have a high proportion of collocates in the texts they are
selected from, (2) these collocations convey relevant sociocultural information encoded in
language use that can foment discussions in the classroom, (3) corpora can be used to improve
vocabulary exercises, and (4) the integration of collocations in vocabulary teaching is feasible
and it is beneficial for learners. Thus, I believe that this study has contributed to the field of
applied linguistics, and more specifically, to the fields of corpus linguistics and English
language education by trying to fill the gap between research findings on collocations and the
applications of these findings in language teaching.

118
REFERENCES

Bartsch, Sabine. 2004. Structural and functional properties of collocations in English: A

corpus study of lexical and pragmatic constraints on lexical co-occurrence. Tübingen:
Gunter Narr Verlag.

Beck, Isabel L., Margaret G. McKeown, & Richard C. Omanson. 1987. The effects and use of
diverse vocabulary instruction techniques. In The nature of vocabulary acquisition, eds.
Margaret G. McKewon & Mary E. Curtis, 147-163. Hillsdale, N.J.: Lawrence Erlbaum.

Benson, Morton, Evelyn Benson, & Robert F. Ilson. 1997. The BBI dictionary of English
word combinations. 3rd ed. Amsterdam: John Benjamins.

Bernardini, Silvia. 2000. Competence, capacity, corpora: A study in corpus-aided language

learning. Bologna: CLUEB.

Bernardini, Silvia. 2002. Exploring new directions for discovery learning. In Teaching and
learning by doing corpus linguistics, ed. Bernhard Kettemann & Georg Marko, 165-182.
Atlanta, Georgia: Rodopi.

Bernardini, Silvia. 2004. Corpora in the classroom: An overview and some reflections on
future developments. In How to use corpora in language teaching, ed. John M. Sinclair,
15-36. Amsterdam: John Benjamins.

Berns, Margie. 1990. Contexts of competence: Social and cultural considerations in

communicative language teaching. New York: Plenum Press.

Berntzen, Jorunn & Alf Bårtvedt, eds. 2005. Targets. Oslo: Aschehoug.

Biber, Douglas, Susan Conrad, & Randi Reppen. 1998. Corpus linguistics: Investigating
language structure and use. Cambridge: Cambridge University Press.

British National Corpus. http://www.natcorp.ox.ac.uk/ or http://corpus.byu.edu/bnc/ (same

interface as COCA; accessed November 16, 2009).

Brown, Dorothy F. 1974. Advanced vocabulary teaching: The problem of collocation. RELC
Journal 5(2): 1-11.

Cambridge advanced learner’s dictionary (CALD). 2nd ed. 2005. Cambridge: Cambridge
University Press.

Carter, Ronald. 1987. Vocabulary: Applied linguistics perspectives. London: Allen & Unwin.

Carter, Ronald & Michael McCarthy, eds. 1988. Vocabulary and language teaching. London:
Longman.

Carter, Ronald & Michael McCarthy. 1999. The English get passive in spoken discourse:
Description and implications for an interpersonal grammar. English Language and
Linguistics 3: 41-58.

119
Chambers, Angela. 2005. Integrating corpus consultation in language studies. Language
Learning & Technology 9(2): 111-125. http://llt.msu.edu/vol9num2/chambers/ (accessed
November 6, 2009).

Chambers, Angela & Íde O‟Sullivan. 2004. Corpus consultation and advanced learners‟
writing skills in French. ReCALL 16(1): 158-172.

Chomsky, Noam. 1959. Review of Verbal behaviour by B.F. Skinner. Language 35: 26-58.

Church, Kenneth & Patrick Hanks. 1990. Word association norms, mutual information, and
lexicography. Computational Linguistics 16(1): 22-29.

Clear, Jeremy. 1993. From Firth principles – Computational tools for the study of collocation.
In Text and technology: In honour of John Sinclair, eds. Mona Baker, Gill Francis, &
Elena Tognini-Bonelli, 271-292. Amsterdam: John Benjamins.

Clear, Jeremy. 2000. Do you believe in grammar? In Rethinking language pedagogy from a
corpus perspective: Papers from the third international conference on teaching and
language corpora, eds. Lou Burnard & Tony McEnery, 19-30. Frankfurt: Peter Lang.

Coady, James. 1993. Research on ESL/EFL vocabulary acquisition: Putting it all in context.
In Second language reading and vocabulary, eds. Thomas Huckin, Margot Haynes, &
James Coady, 46-64. Norwood, N.J.: Ablex.

Coady, James. 1997. L2 vocabulary acquisition through extensive reading. In Second

language vocabulary acquisition, eds. James Coady & Thomas Huckin, 225-237.
Cambridge: Cambridge University Press.

Collins, Heloisa. 2000. Materials design and language corpora: A report in the context of
distance education. In Rethinking language pedagogy from a corpus perspective, eds.
Lou Burnard & Tony McEnery, 51-63. Frankfurt: Peter Lang.

Conzett, Jane. 2000. Integrating collocation into a reading & writing course. In Teaching
collocation, ed. Michael Lewis, 70-87. London: Heinle, Cengage Learning.

Council of Europe. 2001. The user/learner‟s competences. In Common European framework

of reference for languages: Learning, teaching, assessment, 101-130. Strasbourg:
Cambridge University Press.

Cowie, Anthony P. 1981. The treatment of collocations and idioms in learners‟ dictionaries.
Applied Linguistics 2: 223-235.

Cowie, Anthony P. 1994. Phraseology. In The encyclopedia of language and linguistics, eds.
R.E. Asher & J. Simpson, 3168-71. Oxford: Pergamon Press.

Cowie, Anthony P., ed. 1998a. Phraseology. Theory, analysis, and applications. Oxford:
Clarendon Press.

120
Daasvand, Siri. 2001. Some aspects of teaching and learning vocabulary in English as a
foreign language. Unpublished thesis. Department of British and American Studies and
Department of Teacher Education and School Development. University of Oslo.

Davies, Mark. 2008. Corpus of Contemporary American English (COCA).

http://www.americancorpus.org (accessed October 23, 2009).

Ellis, Nick C. 1997. Vocabulary acquisition: Word structure, collocation, word-class, and
meaning. In Vocabulary: Description, acquisition and pedagogy, eds. Norbert Schmitt &
Michael McCarthy, 122-139. Cambridge: Cambridge University Press.

Ellis, Nick C. 2001. Memory for language. In Cognition and second language instruction, ed.
P. Robinson, 33-68. Cambridge: Cambridge University Press.

Fairclough, Norman. 1990. What might we mean by „enterprise discourse‟? In Enterprise

culture, eds. R. Keat & N. Abercrombie, 38-57. London: Routledge.

Fano, Robert M. 1961. Transmission of information: A statistical theory of communications.

Cambridge, MA: MIT Press.

Firth, John R. 1968. A synopsis of linguistic theory, 1930-1955. In Studies in linguistic

analysis, eds. J.R. Firth et al., 1-32. Special volume of the Philological Society. 3rd ed.
Oxford: Blackwell.

Fontenelle, Thierry. 1998. Lexical functions in dictionary entries. In Phraseology, ed.

Anthony P. Cowie, 189-207. Oxford: Clarendon Press.

Gabrovšek, Dušan. 1998. Toward an interlingual collocational dictionary of English and

Slovene. In Language policy and language education in emerging nations: Focus on
Slovenia and Croatia, eds. Robert de Beaugrande, Meta Grosman, & Barbara Seidlhofer,
127-142. Stamford, CT: Ablex.

Gass, Susan M. 1988. Integrating research areas: A framework for second language studies.
Applied Linguistics 9(2): 198-217.

Gavioli, Laura. 2005. Exploring corpora for ESP learning. Amsterdam: John Benjamins.

George, Horace V. 1983. Classification, communication, teaching and learning. English

Language Institute, Victoria University of Wellington.

Halliday, Michael A.K. 1966. Lexis as a linguistic level. In In memory of J.R. Firth, eds. C.E.
Bazell, J.C. Catford, M.A.K. Halliday, & R.H. Robins. 148-162. London: Longmans.

Hatch, Evelyn & Cheryl Brown. 1995. Vocabulary, semantics, and language education.
Cambridge: Cambridge University Press.

Hausmann, Franz J. 1989. Le dictionnaire de collocations. In Wörterbücher, Dictionaries,

Dictionnaires. Ein internationales Handbuch zur Lexikographie, eds. Franz J. Hausmann,
Herbert E. Wiegand, & Ladislav Zgusta, 1010-1019. Berlin: Walter de Gruyter.

121
Heian, Bente, Gro Lokøy, Brynjulf Ankerheim, & Ion Drew. 2006. Experience. Oslo:
Gyldendal.

Hidalgo, Encarnación, Luis Quereda, & Juan Santana, eds. 2007. Corpora in the foreign
language classroom. New York: Rodopi.

Hill, Jimmie. 2000. Revising priorities: From grammatical failure to collocational success. In
Teaching collocation, ed. Michael Lewis, 47-69. London: Heinle, Cengage Learning.

Hill, Jimmie & Michael Lewis, eds. 1997. The LTP dictionary of selected collocations. Hove:
Language Teaching Publications.

Hill, Jimmie, Morgan Lewis, & Michael Lewis. 2000. Classroom strategies, activities and
exercises. In Teaching collocation, ed. Michael Lewis, 88-117. London: Heinle, Cengage
Learning.

Hoey, Michael. 2000. A world beyond collocation: New perspectives on vocabulary teaching.
In Teaching collocation, ed. Michael Lewis, 224-243. London: Heinle, Cengage
Learning.

Hornby, Albert S., Edward V. Gatenby, & Harold Wakefield. 1942. Idiomatic and syntactic
English dictionary. Tokyo: Kaitakusha.

Howarth, Peter. 1998. The phraseology of learners‟ academic writing. In Phraseology, ed.
Anthony P. Cowie, 161-186. Oxford: Clarendon Press.

Howatt, Anthony P.R. (with H.D. Widdowson). 2004. A history of English language teaching.
2nd ed. Oxford: Oxford University Press.

Hulstijn, Jan H. 1993. When do foreign-language readers look up the meaning of unfamiliar
words? The influence of task and learner variables. Modern Language Journal 77: 139-
147.

Hunston, Susan. 2002. Corpora in applied linguistics. Cambridge: Cambridge University

Press.

Hymes, Dell. 1972. On communicative competence. In Sociolinguistics, eds. J.B. Pride & J.
Holmes, 269-293. New York: Penguin.

Hymes, Dell. 1979. On communicative competence (extracts). In The communicative

approach to language teaching, eds. C.J. Brumfit & K. Johnson, 5-26. Oxford: Oxford
University Press.

Johansson, Stig. 2007. Using corpora: From learning to research. In Corpora in the foreign
language classroom, eds. Encarnación Hidalgo, Luis Quereda, & Juan Santana, 17-28.
New York: Rodopi.

122
Johns, Tim. 1994. From printout to handout: Grammar and vocabulary teaching in the
context of data-driven learning. In Perspectives on pedagogical grammar, ed. T. Odlin,
293-313. Cambridge: Cambridge University Press.

Kennedy, Graeme. 1991. Between and through: The company they keep and the functions
they serve. In English corpus linguistics: Studies in honour of Jan Svartvik, eds. Karin
Aijmer & Bengt Alternberg, 95-110. London: Longman.

Kennedy, Claire & Tiziana Miceli. 2001. An evaluation of intermediate students‟ approaches
to corpus investigation. Language Learning and Technology 5(3): 77-90.

Kjellmer, Göran. 1984. Some problems relating to the study of collocations in the Brown
corpus. In Computer corpora in English language research, ed. Stig Johansson, 25-33.
Bergen: Norwegian Computing Center for the Humanities.

Knowledge Promotion (L06). 2006a. Knowledge promotion – Kunnskapsløftet.

http://udir.no/Artikler/_Lareplaner/_english/Knowledge-promotion---Kunnskapsloftet/
(accessed October 7, 2009).

Knowledge Promotion (L06). 2006b. Common core subjects in primary and secondary
education. http://www.utdanningsdirektoratet.no/Artikler/_Lareplaner/_english/
Common-core-subjects-in-primary-and-secondary-education/ (assessed November 3,
2009).

Knowledge Promotion (L06). 2006c. Programme for specialization in general studies.

http://www.utdanningsdirektoratet.no/Artikler/_Lareplaner/_english/Programmes-for-
general-studies/Languages-Social-Sciences-and-Economics-Studies-/ (accessed October
7, 2009).

Krashen, Stephen. 1989. We acquire vocabulary and spelling by reading: Additional evidence
for the input hypothesis. Modern Language Journal 73(4): 440-464

Krashen, Steven & Tracy Terrell. 1983. The natural approach: Language acquisition in the
classroom. Oxford: Pergamon Press.

Larsen-Freeman, Diane. 1986. Techniques and principles in language teaching. Oxford:

Oxford University Press.

Larsen-Freeman, Diane. 1997. Chaos, complexity, science and second language acquisition.
In Applied Linguistics 18(2): 141-165.

Lewis, Michael. 1993. The lexical approach: The state of ELT and a way forward. Hove:
Language Teaching Publications.

Lewis, Michael. 1997. Implementing the lexical approach: Putting theory into practice. Hove:
Language Teaching Publications.

Lewis, Michael, ed. 2000. Teaching collocation. London: Heinle, Cengage Learning.

123
Lewis, Morgan. 2000. There is nothing as practical as a good theory. In Teaching collocation,
ed. Michael Lewis, 10-27. London: Heinle, Cengage Learning.

McCarthy, Michael & Felicity O‟Dell. 2002. English vocabulary in use. 2nd ed. Cambridge:
Cambridge University Press.

McCarthy, Michael & Felicity O‟Dell. 2005. English collocations in use. Cambridge:
Cambridge University Press.

McCarthy, Michael, Jeanne McCarten, & Helen Sandiford. 2006a. Touchstone 3 student’s
book. New York: Cambridge University Press.

McCarthy, Michael, Jeanne McCarten, & Helen Sandiford. 2006b. Touchstone 3 workbook.
New York: Cambridge University Press.

McEnery, Tony, Richard Xiao, & Yukio Tono. 2006. Corpus-based language studies: An
advanced resource book. New York: Routledge.

McIntosh, Colin, Ben Francis, & Richard Poole, eds. 2nd ed. 2009. Oxford collocations
dictionary for students of English (OCD). Oxford: Oxford University Press.

Mel‟čuk, Igor. 1998. Collocations and lexical functions. In Phraseology, ed. Anthony P.
Cowie, 23-53. Oxford: Clarendon Press.

Meunier, Fanny & Céline Gouverneur. 2007. The treatment of phraseology in ELT textbooks.
In Corpora in the foreign language classroom, eds. Encarnación Hidalgo, Luis Quereda,
& Juan Santana, 119-139. New York: Rodopi.

Moon, Rosamund. 1994. The analysis of fixed expressions in text. In Advances in written text
analysis, ed. Malcolm Coulthard, 117-135. London: Routledge.

Moon, Rosamund. 1998. Frequencies and forms in phrasal lexemes in English. In

Phraseology, ed. Anthony P. Cowie, 79-100. Oxford: Clarendon Press.

Moore, Johanna C. & John R. Surber. 1992. Effects of context and keyword methods on
second language vocabulary acquisition. Contemporary Education Psychology 17: 286-
292.

Nagy, William E., Patricia A. Herman, & Richard C. Anderson. 1985. Learning words from
context. Reading Research Quarterly 20: 233-253.

Nation, Paul. 1990. Teaching and learning vocabulary. New York: Newbury House
Publishers.

Nation, Paul. 2001. Learning vocabulary in another language. Cambridge: Cambridge

University Press.

Nattinger, James R. 1988. Some current trends in vocabulary teaching. In Vocabulary and
language teaching, eds. Ronald Carter & Michael McCarthy, 62-82. New York:
Longman.

124
Nattinger, James R. & Jeanette S. DeCarrico. 1992. Lexical phrases and language teaching.
Oxford: Oxford University Press.

Nesselhauf, Nadja. 2005. Collocations in a learner corpus. Amsterdam: John Benjamins.

Osborne, John. 2000. What can students learn from a corpus?: Building bridges between data
and explanation. In Rethinking language pedagogy from a corpus perspective: Papers
from the third international conference on teaching and language corpora, eds. Lou
Burnard & Tony McEnery, 165-172. Frankfurt: Peter Lang.

Palmer, Frank. R. 1976. Semantics: A new introduction. Cambridge: Cambridge University

Press.

Palmer, Harold E. 1933. Second interim report on English collocations. Tokyo: Kaitakusha.

Paribakht Tahereh S. & Marjorie B. Wesche. 1993. Reading comprehension and second
language development in a comprehension-based ESL programme. TESL Canada
Journal 11: 9-27.

Partington, Alan, John Morley, & Louann Haarman, eds. 2004. Corpora and discourse. Bern:
Peter Lang.

Partington, Alan. 1998. Patterns and meanings. Amsterdam: John Benjamins.

Pawley, Andrew & Frances H. Syder. 1983. Two puzzles for linguistic theory: Nativelike
selection and nativelike fluency. In Language and communication, eds. Jack C. Richards
& Richard W. Schmidt, 191-225. London: Longman.

Payne, M. E. 1988. Vocabulary learning strategies used by ESL students and their relationship
to perceptual learning style preferences. Unpublished thesis. Linguistic Department,
Brigham Young University, Provo, Utah.

Richards, Jack C. & Theodore S. Rodgers. 2001. Approaches and methods in language
teaching. 2nd. ed. New York: Cambridge University Press.

Rodgers, Drew, Knut Skifjeld, Dilys Brown, & Martin Mulloy. 2006a. @cross textbook.
Oslo: N.W. Damm & Søn.

Rodgers, Drew, Knut Skifjeld, Dilys Brown, & Martin Mulloy. 2006b. @cross workbook.
Oslo: N.W. Damm & Søn.

Römer, Ute. 2004. Comparing real and ideal language learner input: The use of an EFL
textbook corpus in corpus linguistics and language teaching. In Corpora and language
learners, eds. Guy Aston, Silvia Bernardini, & Dominic Stewart, 151-168. Amsterdam:
John Benjamins.

Said, Edward. 1978. Orientalism. London: Routledge & Kegan Paul.

125
Schmid, Hans-Jörg. 2007. Entrenchment, salience, and basic levels. In The Oxford handbook
of cognitive linguistics, eds. Dirk Geeraerts & Hubert Cuyckens, 117-138. Oxford:
Oxford University Press.

Schmitt, Norbert. 1997. Vocabulary learning strategies. In Vocabulary: Description,

acquisition and pedagogy, eds. Norbert Schmitt & Michael McCarthy, 199-227.
Cambridge: Cambridge University Press.

Schmitt, Norbert. 2000. Vocabulary in language teaching. New York: Cambridge University
Press.

Schmitt, Norbert & Michael McCarthy. 1997. Vocabulary: Description, acquisition and
pedagogy. Cambridge: Cambridge University Press.

Seidlhofer, Barbara. 2000a. Operationalizing intertextuality: Using learner corpora for

learning. In Rethinking language pedagogy from a corpus perspective, eds. Lou Burnard
& Tony McEnery, 207-233. Frankfurt: Peter Lang.

Simensen, Aud Marit. 1998. Teaching a foreign language. Principles and procedures.
Bergen: Fagbokforlaget.

Sinclair, John M., ed. 1987. Looking up: An account of the COBUILD project. London:
HarperCollins.

Sinclair, John M. 1991. Corpus, concordance, collocation. Oxford: Oxford University Press.

Sinclair, John M. & Antoinette Renouf. 1988. A lexical syllabus for language learning. In
Vocabulary and language teaching, eds. Ronald Carter & Michael McCarthy, 140-160.
London: Longman.

Steinberg, Danny D. & Natalia V. Sciarini. 2006. An introduction to psycholinguistics. 2nd

ed. Harlow: Pearson Longman.

Stevick, Earl W. 1976. Memory, meaning and method. Cambridge, MA: Newbury House
Publishers.

Stubbs, Michael. 1995. Collocations and semantic profiles: On the cause of the trouble with
quantitative studies. In Functions of Language 2(1): 1-33.

Stubbs, Michael 1996. Text and corpus analysis. Computer-assisted studies of language and
culture. Oxford: Blackwell.

Stubbs, Michael. 2001a. Words and phrases: Corpus studies of lexical semantics. Oxford:
Blackwell.

Stubbs, Michael. 2001b. Texts, corpora, and problems of interpretation: A response to

Widdowson. Applied Linguistics 22(2): 149-172.

126
Swan, Michael. 1997. The influence of the mother tongue on second language vocabulary
acquisition and use. In Vocabulary: Description, acquisition and pedagogy, eds. Norbert
Schmitt & Michael McCarthy, 156-180. Cambridge: Cambridge University Press.

Sørhus, Theresa Bowles, Richard Burgess, & Trond Christian Anvik. 2006a. Passage. Oslo:
Cappelens Forlag.

Sørhus, Theresa Bowles, Richard Burgess, & Trond Christian Anvik. 2006b. Passage
website. Oslo: Cappelens Forlag. http://passage-old.cappelendamm.no/ (accessed October
28, 2009).

Sökmen, Anita J. 1997. Current trends in teaching second language vocabulary. In

Vocabulary: Description, acquisition and pedagogy, eds. Norbert Schmitt & Michael
McCarthy, 237-257. Cambridge: Cambridge University Press.

Takač, Višnja P. 2008. Vocabulary learning strategies and foreign language acquisition.
Clevedon: Multilingual Matters LTD.

Taylor, John R. 2002. Cognitive grammar. Oxford: Oxford University Press.

Teliya, Veronika, Natalya Bragina, Elena Oparina, & Irina Sandomirskaya. 1998.
Phraseology as a language of culture: Its role in the representation of a collective
mentality. In Phraseology, ed. Anthony P. Cowie, 55-75. Oxford: Clarendon Press.

The BNC basic (C5) tagset. <http://www.natcorp.ox.ac.uk/docs/c5spec.html> (accessed

October 27, 2009).

Utdanningsdirektoratet (Udir). 2007. Gratis læremidler i videregående opplæring.

http://udir.no/Brev/_tilskott/Gratis-laremidler-i-videregaende-opplaring/ (accessed
October 7, 2009).

Vestre, Sissel Kværne. 1998. Coping with words. A study of vocabulary acquisition and
Norwegian learners of English. Unpublished thesis. Department of British and American
Studies. University of Oslo.

West, Michael P. 1953. A general service list of English words. London: Longman.

Widdowson, Henry. 2000. The limitations of linguistics applied. Applied Linguistics 21(1): 3-
25.

Williams, Raymond. 1985. Mining the meaning: Key words in the miners‟ strike. New
Socialist 25: 6-9.

Willis, Dave. 1990. The lexical syllabus. London: Collins-COBUILD.

Woolard, George. 2000. Collocation – encouraging learner independence. In Teaching

collocation, ed. Michael Lewis, 28-46. London: Heinle, Cengage Learning.

127
Yoon, Hyunsook. 2008. More than a linguistic reference: The influence of corpus technology
on L2 academic writing. Language Learning & Technology 12(2): 31-48.
http://llt.msu.edu/vol12num2/yoon.pdf (accessed November 6, 2009).

Zimmerman, Cheryl Boyd. 1994. Self-selected reading and interactive vocabulary instruction:
Knowledge and perceptions of word learning among L2 learners. Ph.D. diss., University
of Southern California, Los Angeles.

Zimmerman, Cheryl Boyd. 1997. Historical trends in second language vocabulary instruction.
In Second language vocabulary acquisition, eds. James Coady & Thomas Huckin, 5-19.
Cambridge: Cambridge University Press.

Curricula

Læreplan for forsøk med 9-årig skole. 1960. Forsøksrådet for skoleverket. Oslo: Aschehoug.

Læreplanverket for den 10-årige grunnskolen (L97). 1996. Kirke-, utdannings- og

forskningsdepartementet. Oslo: Nasjonalt Læremiddelsenter.

Mønsterplanen for grunnskolen. 1974. Kirke- og undervisningsdepartementet. Oslo:

Aschehoug.

Mønsterplanen for grunnskolen. 1987. Kirke- og undervisningsdepartementet. Oslo:

Aschehoug.

Normalplan for byfolkeskolen. 1957. Normalplankomitéen, Kirke- og

undervisningsdepartementet. Oslo: Aschehoug.

128
 APPENDIX 1

This appendix lists only the competence aims for the programs for general studies (which are
the same as the vocational education programs). For the objectives of the subject, main
subject areas, teaching hours, and basic skills see the whole document (L06 2006b).

Competence aims after Vg1 – programmes for general studies

Competence aims after Vg2 – vocational education programmes

Language learning
The aims are that the pupil shall be able to
• exploit and assess various situations, work methods and strategies for learning English
• discuss similarities and differences between English and other foreign languages and use
this knowledge when learning English
• use relevant and precise terminology to describe the forms and structures of English
• describe and assess own progress when learning English
• use a wide selection of digital and other aids independently, including monolingual
dictionaries

Communication
The aims are that the pupil shall be able to
• master a wide vocabulary
• use the forms and structures of the language in spoken and written presentations
• understand extended written and oral presentations on different personal, literary,
interdisciplinary and social topics
• extract essential information from spoken and written texts and discuss the author's attitudes
and point of view
• express himself/herself in writing and orally with subtleness, proper register, fluency,
precision and coherence
• select appropriate listening, speaking, reading and writing strategies adapted to the purpose,
situation and genre
• take initiatives to start, finish and keep a conversation going
• read formal and informal texts in various genres and with different purposes
• write formal and informal texts with good structure and coherence on personal,
interdisciplinary and social topics
• select and use content from different sources independently, critically and responsibly
• use technical and mathematical information in communication
• produce texts with complex content using digital media
• choose an interdisciplinary topic for in-depth studies within his or her own programme area
and present this

Culture, society and literature

The aims are that the pupil shall be able to
• discuss social conditions and values in various cultures in a number of English-speaking
countries
• present and discuss international news and current events
• explain the main characteristics of the development of English from an Anglo-Saxon
language to an international world language
• analyse and discuss a film and a representative selection of literary texts in English from the
genres poetry, short story, novel and drama
• discuss a selection of literary texts in English from various regions of the world and different
periods from the 1500s up to the present
• discuss literature by and about indigenous peoples in the English-speaking world
• prepare and assess his or her own written or oral texts inspired by literature and art
 APPENDIX 2

This appendix lists the collocation exercises that were used as examples for the suggestions
on how to improve the textbook exercises and for giving ideas on how collocations can be
integrated in teaching.

Example 1

Find these words and expressions in the article. Match them with the definitions.
1. do the trick ___g___ a. make use of
2. is right at your fingertips _______ b. a time when you need money
3. take advantage of _______ c. not having
4. a rainy day _______ d. a lot of money
5. count _______ e. is where you need it
6. doing without _______ f. make a difference
7. a bundle ________ g. be a solution

McCarthy et al. 2006b:48

Example 2

The missing verb

What are the missing verbs in the following collocations? The same verb completes all three
examples. If in doubt, check the nouns in a collocation dictionary. Notice how important it is
so learn words in phrases rather than single words.

1. ……………… a mistake 4. ……………… panic

a statement a problem
an observation embarrassment

2. ……………… to a complete standstill 5. ……………… danger

to an understanding an accident
to a decision a question

3. ……………… concern
embarrassment
fear

Michael Lewis 2000:112

Example 3

Words into groups

Match each of these nouns to one of the groups of verbs. Remember, all the verbs in the group
must collocate with the noun.

attack battle dispute fight struggle war

Group A: avoid, get into, pick, provoke, start, win

Group B: declare, go to, lead to, prolong, wage, win
Group C: be engaged in, continue, face, give up, join, take up
Group D: be vulnerable to, carry out, launch, mount, resist, step up
Group E: fight, force, go into, lose, win
Group F: aggravate, get involved in, intervene in, put an end to, resolve

Now do the same with these:

fine penalty punishment sentence discipline

Group A: heavy, lenient, suspended, life, long, reduced

Group B: harsh, heavy, severe, death, stiff, huge
Group C: heavy, hefty, immediate, on-the-spot, stiff, token
Group D: effective, firm, strict, slack, poor, excessive
Group E: appropriate, brutal, capital, fit, lenient, harsh

Michael Lewis 2000:115

Example 4

1. If I had a problem...

Circle the correct words to complete each sentence. Then decide if each sentence is true or

false for you. Write T (true) or F (false). Correct the false statements.

1. ______ If I had a problem, I would talk (for / to / about) my best friend.

If I had a problem, I would talk to my Aunt Lisa.
2. ______ I always apologize (about / to / for) my mistakes.
________________________________________________________________________

3. ______ My friends never thank me (for / about / with) helping them.

________________________________________________________________________

4. _______ I always worry (for / about / from) taking tests!

________________________________________________________________________

5. _______ I usually share CDs (to / with / about) my neighbors.

________________________________________________________________________

6. _______ If I forgot (with / from / about) a friend‟s birthday, I would feel bad.
________________________________________________________________________

McCarthy et al. 2004b:60

Example 5

What do we make?

The most common collocations with the verb make are:

1. make sure 4. make a decision
2. make sense 5. make a mistake
3. make a difference 6. make money

McCarthy et al. 2006a:62

Example 6

Make up your mind

Complete the make and do expressions. Use the definitions to help you.
1. make a _____________ = work to earn money
2. make a good _____________ = make someone think of you positively
3. make a ___________ = make a positive change
4. do your ____________ = try your hardest
5. make up your __________ = decide
6. make ____________ of = make jokes about and laugh at
7. make ___________ = make certain
8. make a ___________ = get something wrong
9. do the ____________ = figure out the numbers
10. make ___________ = seem logical

McCarthy et al. 2006b:44

Example 7

Thematic collocation lists

Prisons The workplace
prison sentence mental challenge
corrections office prospective employees
prison-issue clothing job autonomy
self-help courses hourly wage
kill time straight salary
re-entry into society employee turnover
doing time incentive schemes
alternative sentence unskilled workers
prison capacity external recognition

Conzett 2000:80
Example 8

Odd one out

Cross out the word which does not belong in the group:
potent car, potent drink, potent drug, potent weapon

Advance learners
For the following words, add one or more words with which it might be expected to occur:
potent – potent drug, potent weapon, ....
reality – face reality, harsh reality, virtual reality, ...

Conzett 2000:84