Journal of Second Language Teaching and Research.

Volume 6, Special Issue

An Error Analysis of the Use of Lexical Collocations in the Academic Writing of Libyan
EFL University Students

Aisha Ali Dukali, Huddersfield University, UK

Abstract

The main aim of this paper is to explore the difficulties Libyan undergraduate university
English major students have in the use of verb-noun and adjective-noun collocations by
looking at their performance in free production. Furthermore, twelve verbs and twelve
adjectives1 identified in this research were investigated in depth as part of their combinations.
To achieve the main aim, a 250-word academic writing task was used to collect data from
fourth-year university students at Tripoli University (the Department of English). The data
was analysed using AntConc 3.2.1w (Anthony, 2007). After extracting the learners’
collocations, four methods were used to determine and judge the acceptability of learners’
collocations in terms of conforming to native-like use. They are: (1) the OCD (2009), (2) the
online British National Corpus (3) consultations with two native speakers, and (4) the
acceptability-of-collocations survey was used to triangulate the above three methods
Findings from the academic writing data reveal that: (1) three broad categories of errors were
identified in the erroneously produced verb-noun and adjective-noun collocations in the LLC:
(i) grammatical errors, (ii) lexical errors and (iii) errors related to usage; and (2) Furthermore,
these categories were classified into sixteen and twelve error types in verb-noun and
adjective-noun collocations respectively such as wrong choice of verb, wrong choice of
adjective, wrong choice of noun, determiner errors, preposition errors, number errors, wrong
word order errors, word form errors, usage category errors, intensifier errors and wrong
register errors.

Keywords: verb-noun collocations; adjective-noun collocations; erroneous collocations;

acceptable collocations; collocational errors

Introduction

Libyan EFL learners majoring in English generally do not sound like a native speaker when
using the language, despite the fact that they have been learning English for about ten years

1
The twelve verbs are do, provide, acquire, gain, enhance, make, offer, take, give, get, have and
require. The twelve adjectives are good, academic, high, higher, modern, current, practical, specific,
basic, general, great and special.
55
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

by the time they graduate. One reason for this is that ESL/EFL learners encounter several
difficulties in the use of collocations within their speech or writing in English (Fan, 2009, p.
111). For example, the incorrect or inappropriate use of words and expressions in learners’
interlanguage, though they are linguistically and pragmatically correct, may still sound
‘unnatural’ or ‘strange’ (Mahmoud, 2005, p. 117). According to Selinker (as cited in Ellis,
2008, p.968), the term interlanguage refers to “the systematic knowledge of an L2 which is
independent of both these learner’s mother tongue and the target language”.

Accordingly, collocation is now considered an important aspect of foreign language learning,

necessary for knowing how to combine words to make other special meanings and essential
for all language use. Lewis (2000) highlights the importance of collocations in language use
by proving that both native speakers of a language and successful EFL advanced learners
have a high level of “collocational competence – a sufficiently large and significant phrasal
mental lexicon” (p. 177). Furthermore, Yang & O’Neill (2009) reported that “[t]his
competence plays an important role in helping them use a language fluently, accurately and
appropriately” (p. 182).

In the Libyan context, very few teachers take into consideration the importance and value of
collocations when planning their English language lessons. Hence, EFL Libyan learners often
encounter huge problems in using English lexical collocations. They cannot explain
themselves clearly in writing; for example, although perfect grammar might be used,
problems concerning lexical choice (i.e. collocational use) may still continue. On this note,
Hill (2000) explains that the language produced and used by learners “often sounds awkward
and very intermediate” (p. 50). He goes on to argue that “students with good ideas often lose
marks because they do not know the four or five most important collocates2 of a key word
that is central to what they are writing about” (p. 50). Thus, collocational violations are “an
old problem” and a frequent feature of learners’ interlanguage (Hill, 2000, p. 50). The mastery
of English collocations is consequently found to be a significant problem encountered by
EFL/ESL language learners (Granger, 1998; Howarth, 1998; Nesselhauf, 2003). As McCarthy
(1990) argues, “even very advanced learners often make inappropriate or unacceptable
collocations” (p. 13). Language learners in this case often fail to select and combine the
lexical items in native-like production and usage (see footnote 2 for definition of native

2
A collocate is a word that turns up systematically in close proximity to another word; for example,
the word murder collocates with the verb commit as in he has committed murder (for further details,
see Sinclair, 1991, p. 170).
56
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

speaker) because they are unaware of the collocational patterns and restrictions. This is
certainly the case for EFL Libyan learners. Because there are no generalizable collocational
rules that govern the construction of these appropriate combinations of words, there is,
consequently, a need for EFL learners to use conventions which have to be acquired rather
than learned.

This study investigates learners’ problems and difficulties in the use of two types of lexical
collocation, i.e. verb-noun and adjective-noun collocations, according to Benson, Benson, and
Ilson’s classification of collocation (1997). Lewis (2000) refers to the verb-noun combination
as one of the most important types of lexical collocation. Furthermore, these two types of
lexical collocation are the most commonly investigated in the literature (see the reviewed
studies below). This includes discussing, in depth, learners’ collocational performance in
academic written production, since “production data is publicly observable and is presumably
reflective of a learner's underlying competence” (Brown, 2000, p. 216). This approach –
investigating learners’ written production with regard to collocation – is supported by Lewis
(1997) who argues that by examining learners’ writing, it is possible to show that
miscollocation is a frequent source of error. In this vein, errors may be systematic or non-
systematic. The assessment of errors in this study will cover both types. According to Corder
(1967, as cited in Gass & Selinker, 2008, p. 102) non-systematic errors are mistakes which
are “akin to slips of tongue”. The speaker in this case is able to recognize the deviant forms.
Systematic errors, however, are committed out of ignorance of the grammatical system of the
target language. The learner is unaware that (s)he is committing an error and the deviant form
has been integrated into his/her interlanguage. Lexical errors (in particular, deviant
collocations / misuse of collocations), however, result from the learner’s insufficient
knowledge of appropriate word use and how words are combined or associated. Since
collocation is not determined by logic as is the case with grammar, the learner has to resort to
linguistic convention in order to produce acceptable word combinations (Lewis, 1993).

The decision to investigate learners’ collocational errors was based on the fact that error
analysis has the advantage of providing a better understanding of, and revealing valuable
information about the difficulties learners have with this linguistic phenomenon. Thus, in-
depth insights can be gained of how language is learned and acquired through examining
learners’ errors (Brown, 2000, p. 217). Possible explanations for the students’ misconstrual of
English lexical collocations will be given. This kind of study is important to all people who

57
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

are involved in the educational process, e.g., learners, teachers, syllabus designers and
coursebook writers. It is also designed to raise Libyan teachers’ awareness of the difficulties
Libyan learners encounter when using lexical collocations. Various researchers have focused
on this method to demonstrate students’ difficulties in using English collocations (e.g.
Howarth 1998; Altenberg & Granger, 2001; Namvar, Nor, Ibrahim, & Mustafa, 2012). In
addition, Bazzaz & Samad (2011) argue that “[c]ollocational knowledge is viewed as a very
important issue in writing as it is seen to discriminate (sic) native speakers from foreign
language learners” (p. 158).

Aim of the Study

The goal of this paper is to provide a better understanding of the competence of EFL
university students with regard to their use of verb-noun and adjective-noun lexical
collocations in a Libyan context. The aim is as follows:
To identify the types of collocational errors (i.e. errors that occur within phrases which
contain collocations) Libyan learners make when producing verb-noun and adjective-noun
collocations in a writing task.

Research Question

The current study aims to answer the following research question:

 What types of errors do Libyan learners make when producing verb-noun and
adjective-noun collocations?

Empirical Research on Collocation in an EFL Context

The purpose of this section is to shed light on the relevant research conducted to address EFL
learners’ use of English lexical collocations in production, with special reference to the Arab
world and the Libyan context. It also provides a critical review of the related literature in
order to address and highlight the problematic issues constituting the knowledge gap which
the current study aims to fill. The reviewed literature enabled me to shape and construct the
methodology of this research which tackled all methodological limitations found in the
relevant studies.

The reviewed studies below were classified according to their learning context in terms of
research on the use of collocations by EFL learners and Arab EFL learners while learning
English. My selection of these collocational studies is driven primarily by their relevance to

58
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

my study in terms of context; that is, they investigated EFL learners’ use of lexical
collocations in production, particularly verb-noun and adjective-noun collocations. They are
divided as follows:

1. Research on the use of collocations by EFL learners, and

2. Research on the use of collocation by Arab EFL learners while learning English.

Research on the Use of Collocations by EFL Learners

There have been a number of interesting studies in recent years focused on the collocational
knowledge of EFL learners all around the world; for example, Zarei, 2002; Wang & Shaw,
2008; Darvishi, 2011. Owing to space limitations, I am going to include a review of some
studies to show how the current study fills the knowledge gap in the collocational field.

Nesselhauf’s study (2003) is considered one of the most comprehensive studies of

collocations in learner written English to date (Martelli, 2007, p. 37). She manually
investigated the use of verb-noun collocations (for example, take a break) in free-writing by
3rd and 4th year advanced German University students of English. The data were collected
from 32 argumentative essays (500 words). Nesselhauf identified and examined various types
of mistakes that occurred in the learners’ collocations and also investigated the influence of
the learners’ L1 on the production of English collocations. In her research, she differentiated
between three main types of word combinations: free combinations, e.g. want a car,
collocations, e.g. take a picture and idioms, e.g. sweeten the pill. These distinctions were
made on the basis of what she called “restricted sense” (p. 225). Various methods were used
to judge the acceptability of learners’ collocations: a) the Oxford Advanced Learner's
Dictionary (OALD, 2000) and the Collins COBUILD English Dictionary (CCED 1995), b)
the British National Corpus (BNC), and c) consultation with two native speakers. The results
revealed that “even advanced learners have considerable difficulties in the production of
collocations” (p. 237). Regarding collocational errors it was shown that a) nine types of
mistakes appeared in the learner corpus. Among these, the wrong choice of verb was the most
frequent with 24 occurrences, and b) the greatest proportion of errors appeared in collocations
followed by free combinations and idioms, i.e. 79%, 23%, and 23% respectively. In addition,
“the learners’ L1 turns out to have a degree of influence that goes far beyond that of earlier
(small-scale) studies have predicted” (p. 223). Consequently, some suggestions were made
and discussed regarding teaching collocations.

59
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Having similar aims to Nesselhauf’s study (2003) and using similar methods to judge the
acceptability of learners’ collocations in written English, Wang & Shaw (2008) attempted to
investigate the collocational errors of 100 Swedish students in the English Department of
Stockholm University, Sweden, and 100 Chinese students from the foreign language school of
Wuhan University, China. They were asked to write a short essay of about 200 words in class
in 30 minutes. They investigated verb + noun collocations of common verbs: have, do, take
and make. They used the BBI, the CCED, the BNC and a native speaker to judge the
acceptability of learners’ collocational patterns. The results showed that the two groups of
students encountered different problems in using these common verbs, and made similar types
of error. However, the authors did not specify the register of their corpus for the benefit of the
consulted native speaker to make his/her judgement accordingly.

To measure collocational knowledge in written production, Hong, Rahim, Hua, & Salehuddin
(2011) aimed to examine the types and sources of collocational errors made in the production
of verb-noun collocations by four Malaysian learners of English from three different states in
Malaysia. The data were collected from 130 written essays (a sub-corpus of EMAS)
constituting 35,931 words. The data were analysed by using Wordsmith Tools software. In
their study, Hong et el. used two reference sources which were used to analyse the students’
erroneous collocations (i.e., those collocations do not comply with native-like production,
particularly in academic written English) to supply suggestions for correction, i.e. the Oxford
Collocations Dictionary (2009) and the BNC, to evaluate and determine the acceptability of
the learners’ produced collocations. They used Nesselhauf’s framework (2003) as a guide to
identify and classify the errors occurring in the learners’ collocations. The results reveal that
seven types of collocational errors were identified in the written essays. The most frequent
collocational error was the proposition errors with 126 (41.72%) instances in 268 erroneous
verb-noun collocations. Furthermore, three main categories of sources of collocational errors
were discovered: interlingual transfer, intralingual transfer, and paraphrase, among which
intralingual transfer was the most prominent. Hong et al.’s findings confirmed the results of
previous studies, e.g. Li, 2005; Wang & Shaw, 2008; Darvish, 2011; Phoocharoensil, 2011,
that interlingual and intralingual transfers were the key sources of learners’ collocational
errors. However a weakness of this study was that, as in Li’s study (2005), no native speakers
were consulted to evaluate the learners’ collocations.

60
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Research on the Use of Collocation by Arab EFL learners while Learning English

A number of EFL Arab researchers have examined Arab EFL learners’ use of English
collocations in production data, e.g. Elkhatib, 1984; Farghal & Obiedant, 1995; Al-Zahrani,
1998; Mahmoud, 2005; Dukali, 2010; Alsakran, 2011; Shammas, 2013.

In an early study, Elkhatib (1984) investigated the lexical errors of four undergraduate
Egyptian students. He analyzed their writing samples in order to identify their lexical
problems, discover the causes of these problems, and ascertain whether the learners were
more familiar with the material or with the language structure. The results showed that the
students made eight main lexical errors, and that they could not make appropriate lexical
collocations. He concluded that the main reason for the errors was unfamiliarity with
collocations. This caused them to make such errors shooting stones and do progress.

Similarly, Mahmoud (2005) studied the learners’ actual performance in producing English
collocations. A list of topics was given to 42 Arabic-speaking English-major university
students to enable them to write an essay as a homework assignment about one of the topics.
The results revealed that the EFL Arab learners had limited collocational competence. In
addition, the findings showed that they committed several errors. Indeed, a total of 64% of the
collocations they used were incorrect, and 61% accounted for inappropriate word
combinations. However, the main weakness of this study was to give the writing task to the
learners to do as a homework assignment. This may have had a negative impact on the
validity of the data, since the learners could have used and accessed different resources and
references to help them do the task such as dictionaries, books, the Internet or seeking help
from other people. There was also no mention of the analytical framework followed to
analyse the learners’ collocations, nor was any indication given to native-speaker consultants
regarding the register of the writing task. Another limitation of Mahmoud’s small data study
of 42 essays was that he made a large generalization of the limited results to all Arab EFL
learners. In addition, he did not specify the length of the study corpus, stating vaguely that the
length of the essays “ranged from one and half to two single-spaced pages in length” (p. 120).
Like Li (2005), Mahmoud did not specify the register of their corpus (i.e. academic or spoken
English) to the consulted native speakers to enable them to make sound judgements. Hence, it
can be said that their results are questionable since register can be a very important factor in
the process of judging the acceptability of learners’ collocational patterns. Therefore, the
current study aims to tackle those methodological weaknesses.

61
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

As can be seen from my review of the literature outlined above showed that collocations were
problematic for EFL learners, as their collocational performance in many different contexts
was consistently unsatisfactory. Furthermore, it showed that the current research is unique in
its exploration of the learners’ use of two types of lexical collocations in academic English
writing in an EFL context and particularly in the Arab world and the Libyan context. A
number of methodological issues were also revealed such as some studies did not take into
account the drawbacks of using only the BNC and/or collocational dictionaries, e.g.
collocational dictionaries are not comprehensive in the sense that they do not list every
possible collocate of a certain word. Therefore, one of the innovative natures of the current
study lies in the creation and utilisation of an acceptability-of-collocations survey to assess
the acceptability of learners’ collocational patterns. Concerning those studies which did
consult native speakers to assess the acceptability of the learners’ collocational patterns, they
did not consider indicating and specifying the register of the study corpus (i.e. academic or
spoken English) to native speakers to enable them to make sound judgements (see Wang &
Shaw, 2008; Mahmoud ,2005). Hence, it can be said that their results are questionable since
register can be a very important factor in the process of judging the acceptability of learners’
collocational patterns. For example, in reporting the Queen’s 90th birthday in a formal news
report, a newsreader might use the collocation an auspicious occasion whilst people talking
about such an event in conversation might say it was a great occasion. Therefore, in a formal
context the language user would opt for the adjective “auspicious” which is relatively more
formal than the adjective “great”. Another methodological issue in some studies is related to
giving the writing task to the students to be done as homework. As explained earlier, this
could have had a negative impact on the validity of the data, given that the learners could
have made use of different resources and references to help them do the task. Therefore, the
aim of this study was to address those methodological limitations in order to address this
knowledge gap in the collocational field, and, in so doing, establish a basis from which future
studies may follow.

Collocation: The Study Definition

According to Francis & Poole (2009, p. v), collocations may be defined as a combination of
two lexical items that frequently occur together in a language to “produce natural sounding
speech and writing”, i.e. language that would be considered natural and acceptable to a native
speaker. However, the scope of this definition needs to be expanded to incorporate a

62
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

phraseological-based perspective which distinguishes collocations from other types of word

combinations such as idioms and compound nouns. Another related area to be considered for
defining collocations in this research is the grammatical framework (i.e. verb + noun and
adjective + noun collocations; for further details see grammaticality component below). There
are four principles (i.e. grammaticality, substitutability, semantic component and
conventionality) that combine to form collocations that may be judged as acceptable and
appropriate in terms of native-like performance, which I set out (for detailed explanation, see
analytical framework for the writing task).

Methodology

Participants

The participant cohort for the writing task consisted of 186 undergraduate EFL learners
majoring in English at Tripoli University (in the Department of English). Of these, 90 were
males and 96 were females, ranging in age between 21 and 23 years old. They were in their
final year of a four year undergraduate degree programme. All of them had taken the same
academic courses in their fourth academic year. All the participants were native speakers of
Arabic, sharing the same Libyan nationality and culture. They were all studying English as a
foreign language. English is a compulsory subject within the curriculum at both preparatory
and secondary level in Libya. Hence, all of them had received classroom instruction in EFL
for a period of at least 6 years by the time they enrolled at the university. Furthermore, the
students had been assessed as being at intermediate to lower-advanced level based on their
mid-term exams.

Data Collection Method

The task was a formal written essay on the theme of education with a topic selected from the
International English Language Testing System (hereafter IELTS) test. The topic was taken
from a previous IELTS examination, which had been published in the public domain
(Cambridge ESOL, 2009, p. 102). The topic was:

How do you think universities should educate their students? Should they provide
knowledge and skills that students will need to use when they start work OR should
they simply aim to make students more knowledgeable regardless of whether it is
useful for their future careers?

63
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Data Collection Procedure

The academic written data were collected during the first semester of the academic year 2013-
2014 at Tripoli University, Libya. Both I and a lecturer were present at the time of conducting
the study. The participants were informed of the purpose behind the written tasks, that they
would be part of my ongoing PhD research and that their participation carried no risk to their
academic aspirations. After asking the participants to sign a consent form to show approval of
their agreement, they were told that they had the right to withdraw at any time. They were
then instructed to write an essay of 250 words within 45 minutes. The participants had no
prior preparation time to perform the writing task. Each student was given instructions to
write the essay individually, without any further discussion and without dictionaries.

The Libyan Learner Corpus (LLC)

The study corpus contained 186 academic written essays by fourth-year English major
students as indicated above. The length of the essays in the LLC varied as 74 essays were
under 150 words. Most of the essays complied with the limit of the writing task; however, a
few of the essays exceeded the limit, ranging from 260 to 320 words.

Analytical Framework for the Writing Task

In this study, Gass & Selinker’s error analysis framework (2008) was adopted to analyse the
learners’ collocational patterns. Table 1 illustrates the main steps conducted in the process of
generating and analysing the data. Furthermore, four methods were used to evaluate and
determine the acceptability of the collocations: a) the BNC, b) the OCD (2009), c)
consultations with two native speakers (a senior English Language teacher and an ordinary
native speaker), and d) the acceptability-of-collocations survey which was administered to
100 native speakers of English in order to triangulate the judgements made according to the
three methods.

Table 1: Procedure for error analysis in the present study

No. Procedure of the analysis
1. Data generation
2. Identification of collocations: extracting learners’ collocational patterns
3. Classification of collocations and collocational errors
4. Quantification of collocations and collocational errors
Triangulation methods used to judge the acceptability of the participants’
5.
collocational patterns

64
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

I further decided to classify verb-noun collocations and adjective-noun collocations according

to certain criteria which were based on native-like use of language and in particular academic
written English, which I have subsumed under what I term the ‘scale of acceptability’3. On
this note, Lewis (1997, p. 29) indicates that collocation is ‘arbitrary’. The following table
illustrates the criteria which represent three degrees of acceptability:

Table 2: The scale of acceptability

Degree of acceptability Example Error type
Acceptable make a big difference N/A
partially acceptable make big difference grammatical error: determiner (a)
Unacceptable do a difference lexical errors: verb (do)

As can be seen from the above table, acceptable refers to native-like use. Native-like use was
assessed by using the above-mentioned methods to judge the acceptability of collocations
extracted from the learner corpus in context, taking into consideration the four components,
i.e. grammaticality, substitutability, semantics and conventionality, which join to form
collocations. They are explained as follows:

1. Grammaticality refers to the syntactic relations of the components involved in a

collocation which are verb-noun and adjective-noun collocations. For example, the
following collocation does not conform to the grammaticality criterion: *He shrugged the
shoulders. This fails the grammaticality test because it includes a definite article instead of
possessive adjective (i.e. his).
2. Substitutability refers to whether certain components (i.e. verb and adjective) of
collocations can be substituted for synonyms or near-synonyms. On this note, McIntosh
(1967) proposes two kinds of collocability. The first entails the recognition of whether
certain synonyms are “mutually replaceable to produce English” (p. 310). He provides the
following synonyms as examples: short, low, small, little, and stubby. He shows that only
one of them could fit into the following sentence: He took a ______ vacation.
3. Semantic component refers to the selection of those lexical items which work best to
convey meaning and are appropriate to the context. This entails the use of the collocating
word which is included in the range or collocational set of the node. According to

3
According to Howarth (1998) and Nesselhauf (2003), the term acceptability was adopted to indicate
the degree to which a collocation conforms to native speaker usage, taking into account the context in
which it occurs. In my view, this term is more suitable than others used in earlier research (e.g.
commonness) to illustrate EFL learners’ production which is often relatively uncommon in English
language.
65
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

McIntosh (1967) the search for appropriate collocates for a given node is achieved by
applying “the test of familiarity” (p. 310), i.e. he claims that native speakers have a range
of possible collocates that go with certain nodes. A native speaker will choose a collocate
from this range with which (s)he is most familiar, i.e. the most appropriate in a certain
context. The notion of range is exemplified by the verb shrug which may collocate with
shoulders but not with other parts of the body such as stomach or arm.
4. Conventionality is another principle in defining collocations in this study. It is a cultural
phenomenon, i.e. the way in which certain words combine together as they emerge from
the collective behaviour and norms of the speech community4 which establishes a
convention that has to be memorized. For example, English native speakers use running
water and not moving or going water. For this reason, I used the intuitions of native
speakers of English as a further method for determining the acceptability of learners’
collocational patterns.

To sum up, here is an example of an unacceptable collocation which was applied in the
current study and which failed all four criteria: *He enjoyed fit educate which should read as
He enjoyed a good education. The following criteria are not met by the above collocation.
First, in terms of grammaticality there are two errors, namely, missing determiner (a) and
wrong word form (education). Second, with regard to substitutability, there are several
possibilities for the placement of an adjective to accompany the noun (education) e.g., good,
beneficial, excellent or useful. Here, the student selected the wrong adjective (fit). Regarding
semanticity, the adjective fit represents, in semantic terms, the wrong choice of adjective in
the given context. In another context the word fit might be more suitable, e.g. a(n) fit, popular,
extraordinary, excellent athlete. Fourth, native speakers would instinctively reject the choice
of adjective as well as the grammaticality of the collocation as it seems out of place in an
academic context.

It should be noted that all the above principles were taken into consideration when making
judgements using the OCD (2009), the BNC, consultations with two native speakers, and (4)
the acceptability-of-collocations survey in order to assess the acceptability of learners’
collocational patterns.

4
A speech community varies according to which part of the world English is spoken, e.g. American
speech community. In this study, I assessed the acceptability of learners’ collocations according to the
speech community of Britain.
66
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

A collocation can reside within an extended structure, e.g., a phrase. Therefore, the previous
definition of collocation is not sufficient for the purpose of the study. Hence, the analytical
framework needs to be expanded because it is essential for EFL learners to be aware of the
whole combination (i.e. lexical and grammatical elements) in order to match native-like
usage, rather than simply knowing whether the two lexical items collocate or not. Taylor
(1990) indicates that semantics and syntax are two key dimensions which constitute
collocations, i.e. “knowing the syntactic behaviour associated with the word and also knowing
the network of associations between that word and other words in the language” (p. 2). He
illustrates this point using the example of the verb ‘undertake’, which is usually a transitive
verb, followed by article + noun, or pronoun and will, more often than not, appear in the
context of transport with such words as ‘lane, ‘car’, ‘speed’. In this vein, Nesslhauf (2003, p.
231-232) argues that knowing which words combine, e.g., get + permission, fail + exam, is
insufficient for learners to produce acceptable combinations. In other words, knowing the
whole combination is important to enable them to achieve that aim e.g., get permission (to),
fail an exam). Hence, the acceptability judgment of learners’ collocations not only entails
judging whether the two lexical items (i.e. the node and collocate) combine and comply with
native-like usage, but also entails judging the acceptability of the whole combination (i.e.
verb-noun and adjective-noun phrase [NP] combination). This conforms to the grammaticality
criterion discussed above. The noun phrase includes pre-modifiers of the noun such as
articles, intensifiers and adjectives within the collocation / combination and/or in some cases a
following preposition. The component parts of the noun phrase constitute the phraseological
variations of verb-noun and adjective-noun collocations in constituency variation. For
example, some of the collocational variations of the verb-noun collocation make + difference
are make a difference and make a huge difference. Biber, Finegan, Johansson, Conrad and
Leech (1999) explain that “there are a few semantically light verbs - such as take, make, have,
and do, - that combine with noun phrases to form set verbal expressions” (p. 428). Such
combinations may include a subsequent preposition in some instances such as take care of.
Apart from the light verbs, some of the investigated verbs may at times (but not always)
require a following preposition according to the grammatical context, e.g., offer something to
someone and gain something from. This was also applied in the case of adjective noun
collocations such as a good level of.

Some components of the collocations mentioned above are grammatical words, e.g. articles
and prepositions. These grammatical associations between words are referred to as

67
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

colligations by Stefanowitsch & Gries (2003) who define them as “linear co-occurrence
preferences and restrictions holding between specific lexical items and the word-class of the
items that precede or follow them” (p. 210). In addition, Lewis (2000) defines colligations as
“the way one word regularly co-occurs with a particular (grammar) pattern” (p. 137); for
example, some verbs usually appear with a specific tense, or a noun might typically appear
preceded by a possessive adjective, instead of an article such as pass my/your driving test, It’s
my/your/our responsibility to… (c.f. I’ll take the responsibility for …).

Hence, when a collocation and a colligation co-occur and combine in a phrase, they create a
phrasal construction. According to Stubbs (2005), a phrasal construction may be defined as a
set of lexico-grammatical combinations which typically contain a stable lexical element at
their heart, accompanied by other appropriate linguistic items (p. 1). In short, it may be
described as a melange of collocations and colligations (lexical and grammatical) whose
meaning may be determined by its communicative function.

‘Partially acceptable’ means that the components of a given collocation (i.e. node and
collocate) are correct and collocate within a span which is deemed acceptable, but the
grammatical structure in which it is encased is incorrect. Unacceptable means non-native-like
use such as *high man. It is worth mentioning that this area is different from partially
acceptable in that the conventions of combining the words in a certain way are not used, e.g.
*strong smoker instead of heavy smoker. The native speaker may understand it but would not
use it.

In the following section, I present a brief explanation of the various stages of analysis of the
learners’ collocational patterns. The assessment of the participants’ collocations was executed
in relation to typical native speaker production and use (naturalness) (as judged by using the
four methods), particularly in an academic context as this study was mainly focused on
analysing the participants’ collocations in academic written English. In addition, the term
erroneous collocation refers not only to the wrong production of collocations i.e. where the
two components of collocation do not go together (which can be comprehensible, yet, still not
comply with native speaker convention), but also refers to the inappropriate usage of
collocation in this particular context (i.e. academic register) as some of the participants’
collocational patterns were deemed fairly acceptable in spoken language. In this vein,
McCarthy & O’Dell (2005) point out that learners can sound strange to the native speaker
when they say, for instance, “‘making your homework’ or ‘my uncle is a very high man’” (p.

68
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

4). Both of these phrases can be partially understood in context but they represent the kind of
language which sounds “unnatural and might perhaps confuse” (McCarthy & O’Dell, 2005, p.
4).

1. Data generation: This phase involved generating the concordance lines for each
investigated verb and adjective by using the AntConc Concordance Tool. I also investigated
the distribution of all the searched words.
2. Identification of collocations: extracting learners’ collocational patterns: The
concordance lines were checked line by line manually to search for and identify the
investigated words occurring as part of verb-noun and adjective-noun collocations, using the
BNC and the OCD (2009). Thus, the focus of the next stage of analysis was on the
concordance lines containing learners’ collocations only.

The two native speakers were then consulted. They were asked to evaluate the acceptability of
all the combinations in the LLC. They were provided with enough context to aid their
deliberations, according to the scale of acceptability above. In addition, they were also asked
to double check the work done on the basis of the collocational dictionary and BNC. They
were asked if they agreed with the judgement to put ‘yes’ and in the case of disagreement to
write their suggestions. The next phase of examination entailed comparing the similarity and
differences between native speakers’ judgements and the work done on the basis of the two
searched sources and making some modifications accordingly.

3. Classification of collocations and collocational errors: The classification of collocations

and their collocational errors were conducted at the same time to speed up the process of the
analysis. I employed two criteria in this analytical phase of classification. They were: (1)
Criteria for judging the acceptability of learners' collocational patterns. Verb-noun and
adjective-noun collocations were classified according to certain criteria which were based on
native-like use of language and in particular academic written English, which I subsumed
under what I termed the ‘scale of acceptability’ (i.e. a) acceptable; b) partially acceptable; and
c) unacceptable). As stated earlier, I used four methods to evaluate and determine the
acceptability of the learners’ collocational patterns. Secondly, criteria for classifying learners’
collocational errors were used. They were as follows: a) missing, b) superfluous, c) wrong or
d) wrong word order.
 Quantification of collocations and collocational errors: The occurrences of both
acceptable collocations and erroneous collocations were counted. In addition, different

69
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

types of collocational errors were counted. Then, the percentages and the frequencies
were calculated.
 Triangulation of the Methods Used to Judge the Acceptability of the
Participants’ Collocational Patterns: The acceptability-of-collocation survey taken
from the participants’ erroneous use of both verb-noun and adjective-noun
collocations was administered to 100 native-speakers of English in order to triangulate
the acceptability assessment of learners’ collocational patterns. The participants’
collocations were judged differently in the academic rating survey (i.e. acceptable) as
opposed to the main study. These were to be amended in the main study as acceptable.
Accordingly, the frequency of the verbs and adjectives were then amended and the
accuracy percentages were calculated. Similarly, the occurrences of the collocational
error types were amended and their percentages were calculated.

Results

The Results of the Libyan Students’ Overall Performance of Verb-Noun Collocations

Table 3 presents information about the learners’ overall performance in producing verb-noun
collocations for the twelve verbs under investigation in terms of their frequency of acceptable
collocations and erroneous collocations, their ranking of frequency of use and their accuracy
percentages. Table 4 shows that a total of 1369 collocational patterns were produced by the
participants of the study. Of these, 686 were acceptable collocations whereas 683 were
unacceptable collocations.

Table 3: The participants’ overall production of acceptable and erroneous verb-noun

collocations
Overall
No. of No. of
Frequency frequency No. of Accuracy
Verb acceptable erroneous
ranking of students percentage
collocations collocations
collocations
1. Have 278 131 147 144 47.12
2. Provide 213 117 96 128 54.9
3. Give 190 100 90 110 52.6
4. Make 181 73 108 105 40.3
5. Get 152 53 99 96 34.8
6. Gain 71 55 16 54 77.4
7. Take 67 30 37 57 44.7
8. Do 58 33 25 50 56.8
9. Acquire 47 35 12 44 74.4
10. Enhance 42 23 19 38 54.7
11. Require 37 20 17 31 54

70
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

12. Offer 32 16 17 26 48.4

Totals 1369 686 683 50.1

It is clear from the above table that the participants used 5 high-frequency verbs, i.e., ‘have’,
‘provide’, ‘give’, ‘make’, and ‘get’ in verb-noun collocations more frequently than the other
seven verbs. They had an overall high collocational frequency, at more than 150 occurrences
in every instance as used by more than half of the participants and were ranked from the first
to the fifth position respectively. Concerning accuracy percentage of collocational use, ‘gain’
was the most accurately used verb with of 77.4%, while ‘get’ was the least accurately used
verb with 34.8%.

Types of Errors Identified in Verb-Noun Collocations Produced in the LLC

As can be seen from Table 5 below, three broad categories of errors were identified when
analyzing errors made by the participants when using verb-noun collocations in their written
essays. These are (1) grammatical errors, (2) lexical errors and (3) errors related to usage (i.e.
in this study, usage errors refer to any collocation which does not exist in English). My
analysis revealed that a total of 907 errors occurred in 688 learners’ erroneous collocational
patterns. The most frequent errors were related to grammar, with a frequency of 537 (59.5%),
while lexical errors totalled 342 (37.7%) and errors associated with usage occurred at a very
low frequency of 28, constituting 3% only. Furthermore, those categories were then classified
and divided into sixteen error types which related to different parts of speech, e.g., verb, noun
and adjective, and varied in their degree of difficulty for learners.

Table 4 presents the various error types according to a hierarchy of difficulty (i.e. according to
their frequency in the learners’ erroneous collocations from the most to the least frequent
errors). The examples provided in the table below illustrate the different kinds of errors.
However, this does not mean that other types of errors do not occur in these collocational
patterns as these are covered in other sections. For example, in the erroneous collocation
*have a good knowledges (correct collocation: have good knowledge of), two types of errors
were identified, namely, (a) superfluous determiner and (b) number problems.

Table 4: Types of errors in verb-noun collocations used in the LLC

Frequency
No. Error type in the learner corpus Example of error
of errors
1. Lexical error: verb (wrong choice of verb) *give their best (do ....) 260
2. Grammatical error: Determiner (missing, or *gets a good marks (gets good 203

71
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

present albeit unacceptable or wrong) marks)

Grammatical error: Preposition (preposition *provide students opportunities
3. is missing, superfluous or is present albeit (provide students with ...) 107
unacceptable, or wrong)
Grammatical error: Number (noun used in *have a good knowledges (...
4. the singular instead of the plural or vice good knowledge) 80
versa)
*have great information (...
Lexical error: Adjective (wrong choice of good information)
5. adjective owing to: 1) wrong register or 2) *have a high education 66
adjective and noun do not collocate) (...outstanding education)

*have many doctors good

6. Grammatical error: Wrong word order 50
(have many good doctors)
Grammatical error: Parts of speech (word *have enough qualified (....
7. 39
form) enough qualifications)
Usage: implausible and irreparable
8. * get the stages 28
combination
*provide develop students’
ability (improve ....), *make
9. Grammatical error: Verb (superfluous verb) 19
enhance our society (enhance
.......)
Grammatical error: Conjunction (missing, *have provide education (have
10. 18
or superfluous) or provide ..........)
*give the right lines (give the
11. Lexical error: Noun (wrong choice of noun) 16
right guidance or guidelines)
*provided with modern way of
Grammatical error: noun (superfluous technology (provided with
12. 5
noun) modern technology)

*provide the knowledge, the

skills and also more and more
Grammatical error: Adverb (superfluous,
13. information (provide the 5
missing modifying adverb
knowledge, skills and more
information)
*get master degree (... a
14. Grammatical error: Possessive ‘s (missing) 5
master’s degree)
*have so low level education
Grammatical error: Intensifier (superfluous
15. (have such a low level of 3
or wrong intensifier)
education)
*have a good marks, good
Grammatical error: Adjective (superfluous knowledge, great information
16. 3
adjective) (have good marks, knowledge
and good information)
Total of error types 907

Lexical Collocational Errors: Wrong Choice of Verb

In terms of a hierarchy of difficulty, the results revealed that the most frequent error type
occurring in the participants’ verb-noun collocations were related to the lexical category
(wrong choice of verb) with 260 occurrences. These occurrences constitute 28.7% of the

72
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

overall frequency of errors rated as unacceptable according to the scale of acceptability used
in this study such as *take a good educate instead of have a good enough education. Below is
an example of these collocations derived from the students’ data:

 r it is harder. I guess in modern countries they got awareness to these points

and they start to make (78-02-14.txt)

Furthermore, the results revealed that the majority of learners’ unacceptable verb-noun
collocational patterns were formed by combining high-frequency verbs with various noun
collocates; for example apart from the verb ‘have’, the verbs ‘get’, ‘make’, ‘give’, ‘take’,
‘provide’, and ‘do’ were placed as the most frequent verbs in unacceptable verb-noun
collocations with 69, 68, 41, 32, 24 and 20 occurrences respectively.

On the other hand, the verbs ‘gain’, ‘acquire’, and ‘offer’ occurred only one or two times in
unacceptable collocations where they were used instead of other verbs, e.g. *gain the
techniques instead of learn the techniques, *acquire more time instead of need or require
more time and *offer the large information instead of provide ‘valuable’ or ‘useful’
information. In contrast, there were no instances of using ‘enhance’ and ‘require’ in
unacceptable collocations.

Grammatical Collocational Errors: Determiners

Determiners, especially definite and indefinite articles, were the second most frequent
problematic error type recording 203 (22.3%) occurrences in the students’ erroneous verb-
noun collocations. The following illustrative examples demonstrate cases of: 1) redundancy as
in *acquire the knowledge skills instead of acquire knowledge and skills, 2) omission as in
*had basics instead of had the basics, and 3) substitution such as *gives an opportunity for
students instead of gives students the opportunity. The following is an illustrative example:

 now the life skills. University can make students acquire the knowledge skills.
University should be a mode (150-02-14.txt)

73
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

The Results of the Libyan Students’ Overall Performance of Adjective-Noun

Collocations

The results in table 5 below revealed that 793 adjective-noun collocational patterns were
produced by participants. Of these, 491 were acceptable collocations whereas 302 were
unacceptable or questionable collocations.

Table 5: Learners’ overall production of acceptable and erroneous adjective-noun

collocations

Overall No. of No. of

Frequency No. of Accuracy
Adjective frequency of acceptable erroneous
ranking students percentage
collocations collocations collocations
1. Good 200 124 76 103 62
2. Modern 89 63 26 64 70.7
3. Academic 87 58 29 69 66.6
4. Great 65 14 51 54 21.5
5. Higher 57 27 30 48 47.3
6. High 53 27 26 45 50.9
7. Specific 51 39 12 36 76.4
8. General 47 40 7 37 85
9. Practical 45 32 13 38 71.1
10. Basic 44 30 14 33 68.1
11. Special 36 19 17 31 52.7
12. Current 19 18 1 18 94.7
Totals 793 491 302 61.9

As can be seen above, the adjective ‘good’ was placed in the first rank according to
collocational frequency of use with 200 occurrences. It was also the best well-distributed
adjective in the LLC with 103 students using it. In terms of accuracy percentages, ‘current’
was the most accurately used adjective in the LLC with an accuracy percentage of 94.7%,
whilst ‘great’ was the lowest accurately-used adjective with 21.5%.

Types of Error Identified in Adjective-Noun Collocations Produced in the LLC

As was the case with verb-noun collocations, the analysis identified three broad categories of
error in the learners’ adjective-noun collocations. These are: (1) grammatical errors, (2)
lexical errors, and (3) errors related to usage. These contained twelve error types which were
ranked from the most frequent to the least frequent collocational errors as shown in Table 6
below. The results also indicated that errors related to grammar were the most frequent
constituting 275 (70.1%) out of an overall total of 390 errors, occurring in a total of 302
learners’ erroneous collocational patterns whereas a mere 112 (28.7%) were lexical errors. On

74
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

the other hand, errors linked with usage had very low frequency with only six occurrences,
constituting 1.5%.

Table 6: Frequency of adjective-noun error types in the learner corpus

Frequency
No. Error type in the learner corpus Example of error
of the errors
Grammatical error: Determiner (article missing *a specific information
1. 130
or present, albeit unacceptable or wrong) (specific information)
Lexical error: Adjective (wrong choice of *modern knowledge
2. 108
adjective) (recent .............)
Grammatical error: Number (noun used in the *good personality
3. 53
singular instead of the plural or vice versa) (....... personalities)
Grammatical error: Parts of speech (word *academic educate (...
4. 46
form) education)
*academic to modern
Grammatical error: Preposition (preposition is
5. language (academic 23
missing or superfluous)
modern .........)
*basic accountant stock
6. Grammatical error: Wrong word order (basic stock 16
accountancy)
7. Usage: Implausible and irreparable combination *current ages 6
*practical life (...
8. Lexical error: Noun (wrong choice of noun) 4
skills)
* a good marks, good
Grammatical error: Conjunction (conjunction
9. knowledge (good marks 3
is missing or wrong)
and knowledge)
*an academic and modern
Grammatical error: Adjective (superfluous
10. of this way (an academic 2
demonstrative adjective)
and modern way)
Grammatical error: Adverb (superfluous *more higher education
11. 1
modifying adverb) (higher education)
Grammatical error: Intensifier (superfluous or *very higher levels
12. 1
wrong intensifier) (much ..........)
Total of error types 390

Grammatical Collocational Errors: Determiners

According to the hierarchy of difficulty, the results showed that the most frequent error type
occurred when the students incorrectly produced adjective-noun collocations, namely,
determiners (i.e. definite and indefinite articles) with 130 occurrences. This error-type
constitutes 33.3% of the overall frequency of errors in the LLC. The determiner errors
included cases where the articles were missing or present, albeit wrong or superfluous. The
following examples illustrate the point: 1) article missing as in *academic aim instead of an
academic aim and *good place instead of a good place; 2) wrong choice of article such as
*the good university instead of a good university and *the specific subject instead of a

75
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

specific subject, and 3) superfluous article such as *a good experience instead of good
experience and *a good careers instead of good careers. Below is an example of these
collocations derived from the students’ data:

 xample we have 4 Arabic. Third the building isn’t good place for study. The
class isn’t large because th (64-02-14.txt)

Lexical Collocational Errors: Adjective

The ‘wrong choice of adjective’ lexical error type was the second most frequent error type in
the students’ written essays with 108 (27.7%) occurrences. This error type can be further
classified into two types:

1. Wrong choice of adjective, where the students opted for using the wrong adjective
(adjective cannot be used to modify the head nouns), resulting in unacceptable collocations
according to the study’s scale of acceptability with 83 occurrences. Examples of this error are:
*modern knowledge instead of ‘recent’/‘up-to-date’ knowledge, *a higher experience instead
of ‘better’/‘first-hand’ experience and *high way of thinking instead of a sophisticated way....
The following is an illustrative example:

 teach. But here in Libya teachers do not have a higher experience to provide
the students. If the studen (139-02-14.txt)
2. 25 instances of using the wrong adjective were errors linked with using the wrong register.
In all these instances, the students opted for using adjectives to modify the collocating nouns
which are acceptable to use in spoken language rather than in academic written English, e.g.,
*great education in place of ‘good’/’excellent’ education, *a great generation in place of
successful generation and *a great attention instead of considerable attention. The following
is an illustrative example:

 vice the managements of the universities to pay a great attention for improving
the education systems in (17-02-14.txt)
Discussion

Based on the qualitative and quantitative analysis of the obtained data from the Libyan
Learner Corpus, three broad categories of errors were identified in the learners’ erroneous
verb-noun and adjective-noun collocations. These are: (1) grammatical errors, (2) lexical
errors and (3) errors related to usage. This current study result confirms the results of previous

76
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

research investigating the errors in collocations produced by different participants all over the
world (Wang & Shaw, 2000; Nesselhauf, 2003; Li, 2005; Mahmoud, 2005; Kuo, 2009;
Miyakoshi, 2009; Darvishi, 2011; Huang, 2011).

Generally, grammatical errors were more recurrent in the two investigated types of
collocations than lexical errors. For instance, determiners especially definite and indefinite
articles were the first and second most frequent problematic error recording 130 (33.3%) and
203 occasions in the participants’ adjective-noun and verb-noun collocations respectively.
This finding demonstrates that the correct and acceptable choice of the lexical components
(i.e. node and collocate) in a collocation does not inevitably mean that the participants had no
difficulties in producing native-like and grammatically well-formed English lexical
collocations.

As indicated in the literature review studies and similar to the results of the present study,
Wang & Shaw (2008, p. 215) reported that the Chinese students committed different types of
errors in using verb-noun collocations. Examples of such errors are given below:

a) Lexical errors: (1) verb choice, e.g., *take the problem; (2) noun choice, e.g., *make
benefit, (3) adjective such as *do some protecting work.
b) Grammatical errors: (1) noun plurality, e.g., *have troubles with; (2) determiner, e.g.,
have the duty; (3) preposition, e.g., *do harm of; (4) syntactic structure, e.g., *do favor
to; (5) adverb form, e.g., *have a full functional sanitation.
c) Semantic error (where the correct collocation does not make sense) e.g., *take care of
the problem.

It is clear that these error types are a persistent problem in learning English as a foreign
language as they were also found and identified in the current study which means that many
EFL learners, whether Arab or others, commit the same collocational error types in their
writing e.g., preposition errors and noun choice errors.

Furthermore, the results of the current study confirm the results of previous research such as
Mahmoud’s study (2005) which demonstrated that Arabic learners of English have particular
difficulties in using collocations. Mahmoud identified three types of errors in the 42 students’
written essays such as: (1) word choice (where the choice of one word or both words is
incorrect), e.g., *repair his mistake and *hurts the mind, (2) word form (where the form of a
word is incorrect), e.g., *a famous musician band, and (3) contextual errors (linguistically

77
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

correct but contextually incorrect), e.g., *bring a boy instead of (give birth to a boy). He
claimed that the identified errors in the learners’ deviant collocations show that “EFL students
depend on interlingual and intralingual strategies to facilitate learning” (p. 124). The same
errors found by Mahmoud’s study were also identified in the current study, although several
other error types were identified in the collocations produced by the Libyan participants. This
further indicates that Arab EFL learners commit many of the same collocational errors.

Particularly in the case of the Libyan students of the current study, different explanations can
be provided to interpret the various types of errors committed by the participants in producing
verb-noun and adjective-noun collocations.

Wrong Choice of Verb

As the data in Tables 4 and 6 illustrated, the most frequently occurring error was related to
the lexical category (wrong choice of verb) in producing verb-noun collocations on 260
occasions in the LLC. According to Wang & Shaw (2008, p. 218), this difficulty could be due
to the participants’ unawareness of the semantic compatibility between the verb and the noun.

In addition, the results of the current study revealed that the majority of learners’ unacceptable
collocational patterns were formed by combining high-frequency verbs, for example, the
verbs ‘have’, ‘get’, ‘make’, ‘give’, ‘take’, ‘provide’, and ‘do’ were placed as the most
frequent verbs in unacceptable verb-noun collocations with 69, 68, 41, 32, 24 and 20
occurrences respectively. A possible explanation for such a finding is that highly frequent
verbs such as ‘have’, ‘make’, ‘do’, ‘take’ and ‘get’ are polysemous and can be used instead of
other verbs.

Thus, the students’ tendency to use high-frequency verbs as a substitute for other English
verbs is an important finding. This was illustrated by learners’ over-use of some of these
verbs to convey and communicate the intended meaning and to compensate for their lack of
academic vocabulary use. For example, the verb ‘get’ was over-extended as participants used
it 23 times instead of verbs such as ‘gain’, ‘acquire’ and ‘develop’ by combining it with the
nouns ‘knowledge’ and ‘skills’ to construct various verb-noun collocational patterns. The
following combination illustrates the point: *get (the) knowledge x4. Other instances of over-
extension were *get new words instead of learn new words and *get benefits instead of gain
or derive benefit. Correspondingly, the verb ‘take’ was also over-extended and was used 11
times instead of ‘obtain’ as in *take information instead of obtain information. These results

78
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

are inconsistent with Dukali (2010), who found that the participants used the wrong verb in
many instances due to overextending certain verbs such as ‘make’ “to cover a more
appropriate or concise word which may or may not have been known by the students” (p. 78).

Moreover, the participants produced high frequency verbs in their delexicalized sense
interchangeably, e.g. they used *doing the exams instead of ‘take’ or ‘sit’ the exam. It is
apparent that the participants’ awareness of collocational restrictions and use is very limited.
Also, judging by my own experience as a Libyan teacher, this reflects the practice of teaching
and learning lexical items in isolation rather than in their various collocational patterns.

Additionally, the participants’ used these two verbs (i.e. ‘do’ and ‘make’) interchangeably to
the exclusion of other potential English verbs as shown in the following examples: *make
special research instead of do special research and *do their efforts instead of make an effort.
This practice was also discovered in previous research such as in (Dukali, 2010; Ahmed,
2012). Ahmed discovered that “many students were unaware of the distinctions between make
and do and assumed that they were similar” (p. 160). Consequently, they sometimes use the
verb ‘do’ where the verb ‘make’ should be used as in: *do attempts. She mentioned that this
could be related to the translation of the two verbs into their core or original meaning in the
subjects’ mother tongue (Arabic). Consequently, the students produced collocations based on
the semantic meaning of single lexical items. Moreover, they wrongly “equated the verb ‫يقوم‬
‫ بي‬/yaqum bi/ with to do or to make because in Arabic, the verb ‘yaqum bi’ enters into a
variety of combinations meaning roughly to perform" (2012, p. 160-161). In this vein
Balhouq (1982, p. 297-298) states that these two verbs overlap together in their meaning in
the sense of ‘carry out’ which is equivalent to Libyan colloquial Arabic (LCA) /da:r/.
Therefore, they used them instead of other verbs to convey this meaning as in *make a party
instead of have a party, *make a bath instead of have, take a bath, and *do an accident
instead of have an accident. He further argued that this type of error is related to the mismatch
between L1 usage and L2 use. He stated that:

Unless the learner has acquired the L2 habitual collocation in question, he is more
likely to produce a deviant collocation because neither make nor do which have been
acquired as the equivalents of /da:r/ (or MSA /ˤamila/, /faˤala/ or /sanaˤa/) will be
acceptable (1982, p. 297-298)

79
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

The above findings can be summarised as (1) unawareness of semantic compatibility of verb-
noun, (2) polysemous verbs, (3) overextension, (4) interchangeable delexicalisation, and (5)
L1, L2 mismatch.

Determiner Errors

As the data in Table 4 and 6 demonstrates, applying the correct and acceptable choice of a
lexical component in a collocation does not inevitably mean that the participants had no
difficulties in producing correct and appropriate English lexical collocations. In this
connection, Nessulhauf (2003) explained that mistakes in non-lexical constituents shows that
“it is not sufficient for the learner to know which lexical items collocate (such as get +
permission, fail + exam), but rather in order to produce an acceptable combination, it is
essential to know the whole combination (e.g. get permission (to), fail an exam)” (p. 231-32).
It should be noted that too much emphasis is put on teaching grammatical rules and
explaining them in the learners’ L1 in the Libyan educational system while collocation is
neglected and that the Grammar-Translation method is still widely used by a number of
Libyan teachers (Saaid, 2010; Emhamed & Krishnan, 2011). However, the above finding
shows that the participants still encounter difficulties with various aspects of English
grammar. A shift in focus is therefore required by Libyan EFL teachers to bring forth a more
modern approach to teaching grammar in order to help their students to produce more native-
like utterances of English language and overcome their difficulties with basic grammar. One
way of doing this is by teaching grammar in context through teaching collocational patterns as
indicated and explained by Nessulhauf above. For example, when teaching collocational
patterns in context, e.g., within a reading text, the teacher can draw his/her students‟ attention
to the pre-modification and post-modification of the noun in the form of, e.g. articles,
prepositions and intensifiers. The following examples show how teachers might go about it:
1. pass my/your driving test instead of pass + driving test,
2. take responsibility for instead of take + responsibility,
3. good level of instead of good + level.
By so doing, the students’ overall grammatical and collocational knowledge will be improved.

For instance, determiners especially definite and indefinite articles were the first and second
most frequent problematic errors recording 130 (33.3%) and 203 occasions in the
participants’ adjective-noun and verb-noun collocations respectively. Determiner errors
include cases where the articles are missing, incorrectly selected or superfluous, e.g.

80
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

*academic aim instead of an academic aim and a specific information instead of specific
information. This could be because of the differences between the Arabic and English
grammatical systems. On this note, Tengler, Aburiaiza, Ali, & Bakarally (2013, p. 72)
explained that Arabic learners encounter difficulties in the use of English articles due to the
fact that “there is no indefinite article in Arabic, and the use of the definite article different
from the use in English, the indefinite article is routinely omitted or used incorrectly”.

The Wrong Choice of Noun

This lexical error type occurred on 16 occasions in the participants’ verb-noun collocations
and four times in adjective-noun collocations. The following example illustrates the point:
*give the right lines for instead of give the right guidance or guidelines. The participants’
unacceptable production of English collocations may be due to the fact that they tend to
produce messages by combining individual lexical items rather than taking them from
prefabricated patterns (Wray, 2002). In Sinclair’s study (1991), EFL learners have the
tendency to function more on the open choice principle rather than on the idiom principle.

Usage Category Errors

The results revealed that a total number of 28 and 6 instances occurred in verb-noun and
adjective-noun collocations. Those erroneous collocational patterns were classified as
implausible and irreparable combinations, when both I and the consulted native speakers
failed to suggest the acceptable collocations and even after looking at the whole paragraph in
which the collocation occurred to recover the intended meaning, for instance, *get take
different way, *a special speciality, and *good substances. It is worth mentioning that the
rationale for including these patterns instead of excluding them from the study is that they
provide an insight into the difficulties encountered by Libyan learners in conveying intended
meaning in their writing and using lexical verb-noun collocations in particular. The
participants clearly tend to express meaning from individual words (i.e. operate on the open
choice principle) rather than collocate lexical items. Therefore, it is necessary to demonstrate
the importance and the need to teach formulaic language including collocations.

Intensifier Errors

These are the lowest and second lowest frequency errors with only 3 and 2 occurrences in
verb-noun and adjective-noun collocations, e.g., *has a very big role instead of has a(n) vital/
key/important role. The low frequency of intensifiers was due to the fact that learners, in

81
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

general, did not use intensifiers extensively in their verb-noun collocations. A possible
explanation is learners are not confident in using intensifiers, thus, they avoid using them and
tend to produce simple verb-noun collocational patterns in the form of either verb + noun
collocations or verb + adjective + noun collocations.

Conclusion

This study represents, to my knowledge, the first large-scale investigation of university

learners’ difficulties in terms of the grammatical and lexical errors typically made by EFL
learners in the use of collocation in academic written English in the Arab EFL context.
Furthermore, it is the first large exploratory study conducted in a specifically Libyan EFL
context. I aim to fill this gap in knowledge and, in so doing, establish a basis from which
future studies may follow.

My study is especially important to EFL teachers and learners in general. It may provide some
help in solving the problems that learners encounter in the process of language acquisition and
in the learning of this particular linguistic phenomenon. In addition, this study contributes to
the enrichment of collocational studies and the difficulties encountered in this area in all
teaching/learning contexts. The results of this current research confirm the findings of other
studies in the area of the use of collocations in English language learning.

In general, the obtained results from the current research support the claim that EFL learners
have insufficient knowledge of English lexical collocations as revealed by their error-strewn
performance in producing them (e.g. Wang & Shaw, 2000; Nesselhauf, 2003; Mahmoud,
2005). According to this research and others, EFL learners tend to commit the same
collocational error types in their writing e.g. preposition errors and noun choice errors which
means these error types are a persistent problem in learning English as a foreign language
worldwide. In addition, other collocational errors, for example, determiner errors, number
errors and word order errors may occur as a result of the differences between the Arabic and
English grammatical systems.

All in all, it is hoped that this research will contribute to a greater understanding of the
difficulties encountered by EFL learners when producing verb-noun and adjective-noun
collocations as well as the various types of collocational errors made. Therefore, it is hoped
that EFL language instructors will be informed of such difficulties and will then be more able
to implement lessons centred on English collocations which address these difficulties and

82
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

target their students’ specific needs in order to improve their collocational knowledge to
eventually achieve native-like competence. In addition, it is hoped that the obtained results
from the current study will encourage language instructors to recognize the importance and the
benefits of incorporating various types of English collocations into their teaching instructions
in the classrooms.

Recommendations

Based on the obtained results, a variety of recommendations are made, and suggestions given,
for Libyan EFL English language instructors to take into consideration when teaching and
introducing English collocations to their students.
1. While teaching lexical collocations, particular attention should be given to teaching
verb-noun collocations as the results confirmed that this type was more problematic
for the participants than adjective-noun collocations.
2. Special attention should be paid to teaching the verbs ‘make’ and ‘get’ along with the
adjectives; ‘great’ and ‘higher’ by highlighting their various noun collocates in verb-
noun and adjective-noun collocations respectively. Hence, these verbs and adjectives
have the lowest accuracy percentages in the LLC, in spite of the fact that ‘make’ and
‘great’ were placed in the relatively high position of fourth and ‘get’ and ‘higher’
were ranked in the fifth position in terms of overall collocational frequency.
3. Particular focus should be given to the teaching of delexicalized verbs in collocations
by introducing their different noun collocates since the results showed that the
participants produced high frequency verbs in their delexicalized sense
interchangeably. In addition, it is vital to attract students’ attention to the commonly
mistaken collocations and in particular ‘make’ and ‘do’ such as “make a mistake” and
not “do a mistake” and “do a research” and not “make a research” as the participants
(Libyan learners) used these two verbs (i.e. ‘do’ and ‘make’) interchangeably or
similarly and instead of other English verbs. This can be done by making use of native
speaker corpus data such as British National Corpus (BNC) and COBUILD Bank of
English corpus which are excellent resources of common and typical English
collocations. The teacher would need to identify appropriate collocations and then
bring them to the attention of the students by means of concordance lines. Thornbury
(2002) explained the benefits of recommending the use of corpus data to EFL teachers
and learners as “it provides them with easily accessible information about real

83
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

language use, frequency and collocation” (p. 68). In addition, those two corpora
represent different types of English collocations in their most standard structures and
offer a variety of collocations in both written and spoken language.
4. The study shows that the grammatical errors were more frequent than lexical errors in
the participants‟ collocational patterns when producing verb-noun and adjective-noun
collocations. Therefore, introducing and teaching the whole collocational pattern (i.e.
not only the node and the collocate, but also the pre-modification and post-
modification of the noun) to the students is vitally important in order to overcome
students’ difficulties in terms of the grammatical perspective of language. In addition,
the students’ overall English language proficiency would be improved and they would
be more capable of producing native-like utterances. This would also ease the process
of communication in terms of communicating and conveying the intended meaning.
This is particularly so when it comes to pre-modifiers of the noun such as articles,
intensifiers and adjectives within the collocation. On this note, I reiterate Nesselhauf’s
call (2003) for a more comprehensive approach. She suggests that it is not enough to
“merely teach the lexical elements that go together, but it is necessary to teach entire
combinations including prepositions, articles, etc” (p. 238).
5. In respect of using a wrong register, the study revealed that the participants did not
comply with the requirements of the task of writing an academic essay, since on many
occasions they chose the wrong verb and adjective when producing verb-noun and
adjective-noun collocations owing to wrong register although they may be considered
acceptable in other contexts, e.g., spoken language. Thus, it is recommended that
teachers can raise awareness in students of how register affects choice of collocation.
6. The results revealed that some of the participants’ collocational errors were, in my
opinion, due to the influence of their mother tongue (Libyan colloquial Arabic)
because some of their erroneous collocational patterns had equivalents in Libyan
colloquial Arabic. It is therefore recommended that Libyan EFL instructors raise
greater awareness in their students of such difficulty and point out that transferring
language from their mother tongue does not always result in acceptable production of
English collocations. This could be done by compiling a bilingual list of collocations
and bringing it to their students’ attention.
7. The current study results show that L1 interference, overgeneralization, the use of
synonymy and the use of de-lexicalized verbs were shown to be common difficulties

84
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

for the participants in the production of English collocations. It is therefore strongly

recommended that Libyan EFL teachers should design and use collocational activities
which address these difficulties in order to raise awareness in their learners. The
activities should particularly aim to draw attention to areas of difficulty described
below.
1) L1 interference in collocation production, which includes coverage of: a)
lexical confusion, for example, *make (idier) the curriculum (LCA) instead of
design the curriculum, and b) grammatical elements such as: (i) determiner
errors as in *academic aim instead of an academic aim, (ii) singular/plural
errors as in *have a good knowledges instead of have good knowledge and (iii)
wrong word order errors as in *make a student good instead of make a good
student. These grammatical error types were identified as arising from the
differences between the Arabic and English grammatical systems. Thus, it is
also recommended that when teaching collocational patterns, emphasis should
be given to the differences between the grammatical systems between the
learners’ L1 (Arabic language) and L2 (English language).
2) Synonym problems as in *modern knowledge in place of “recent” / “up-to-
date” knowledge.
3) The use of de-lexicalized verbs as in *got awareness instead of have an
awareness of and *make modern activities instead of do modern activities.
4) Overgeneralization, such as the verb “take”, which was over-extended 11
times, instead of using a verb such as “obtain” as in *take information instead
of obtain information.

Limitation of the Study

There are, of course, caveats and limitations to all research. There are several limitations to
this particularly study and these are highlighted below to pinpoint areas where future research
is required.
1. One limitation of the study was that the data was confined to one university in Libya
(Tripoli University) and it was also collected from a relatively small number of
participants (186 fourth year English major students). Therefore, due to these two
limitations, marginally different results may have been obtained if the study had included
English major students from other Libyan universities. However, I would argue that the

85
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

results can be generalised to all Libyan EFL learners due to the following reasons. Libyan
students majoring in English are all native speakers of Arabic and studying English as a
foreign language. They all have a similar background (i.e. the same Libyan nationality and
culture) and they learn English according to one curriculum which is the same for both
private and state schools. However, it should be noted that in some cases there may be
differences between the students such as their age and English proficiency level. These
limitations arose out of difficulties encountered in collecting the data such as the ongoing
conflict in other cities, political wrangling and the unstable situation in Libya arising from
the recent revolution. This limitation would have been overcome if it had been possible to
collect data from various universities in Libya on a larger scale.
2. The focus of the investigation was on two types of English lexical collocations (verb–
noun collocations and adjective-noun collocations). Furthermore within each type, the
study examined a limited number of verbs and adjectives (twelve each). However, an in-
depth investigation and analysis were conducted to determine the overall frequency of
every investigated verb and adjective. The frequency of acceptable as well as erroneous
occurrences in verb-noun and adjective-noun collocations were also identified and
counted.
3. A third limitation of the study was related to the participants’ English language
proficiency level. As there was no standardised way of assessing the participants’ level of
English Language proficiency such as the TOFEL or IELTS tests due to financial and
political constraints, there was a lack of distinction between the participants’ level of
proficiency. They were assessed according to their mid-term exams from the University.
They were rated as intermediate to lower advanced as indicated by their writing professor.
This application of less reliable means of testing language proficiency represents a
limitation since more reliable testing services such as the IELTS proved too difficult to
administer in terms of obtaining permission from the relevant authorities and funding.

In conclusion, it is important to point out that I anticipated some of the above-mentioned

limitations. However, due to the instability of the political situation in Libya, I could not
address them. Nevertheless, the limitations of this study provide opportunities for improved
future research. Suggestions for future research are introduced in the following section.

86
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Suggestions for Further Research

Researching English collocations is still in its infancy particularly in the Arab context.
Therefore, considerable attention is required from researchers and linguists to conduct more
research to examine the nature of this linguistic phenomenon in-depth.

1. Future studies need to include a wide range of homogeneous participants from

different universities and institutions in Libya in an attempt to enhance the reliability
and validity of the findings.
2. It would also be of interest to assess the Libyan learners’ knowledge of collocations at
varied language proficiency levels along with a range of learning stages to further
investigate their difficulties with different types of English collocation in written
production.
3. Furthermore, more research is needed to investigate other types of lexical collocations.
Further studies are needed to examine the learners’ ability to use various types of
grammatical collocations as well.

Biodata

Dr. Aisha A Dukali is a corpus linguist, who joined Research Group in Applied Linguistics in
2013. She obtained her PhD degree from the School of Music Humanities and Media at the
University of Huddersfield in 2017. Her research focuses on exploring and analyzing the use
of English collocation in the academic writings of fourth-year university students at Tripoli
University in Libya (the Department of English, Faculty of Arts). She is interested in topics
related to TEFL, language acquisition, corpus studies, corpus linguistics and applied
linguistics in EFL environment in particular.

References

Ahmed, A. (2012). English lexical collocation knowledge of Libyan university students.

Unpublished PhD thesis, Bangor University.

AL-Amro, M. (2006). Saudi learners’ knowledge and its relationship to their vocabulary size
and writing quality. Unpublished MA thesis, Colorado State University.

Al-Zahrani, M. (1998). Knowledge of English lexical collocations among male Saudi college
students majoring in English at a Saudi university. Unpublished PhD thesis, Indiana
University of Pennsylvania.

87
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Alsakran, R. (2011). The productive and receptive knowledge of collocations by advanced

Arabic-speaking ESL/EFL learners. Unpublished MA thesis, Colorado State
University.

Altenberg, B. & Granger, S. (2001). The grammatical and lexical patterning of MAKE in
native and non-native student writing. Applied Linguistics, 22(2), 173-195.

Bahns, J. (1993). Lexical collocations: A contrastive view. English Language Teaching

Journal, 47(1), 56-63.

Balhouq, S. (1982). Problems encountered by Libyan learners of English with special

reference to the lexicon. Unpublished PhD thesis, University of Sheffield.

Bazzaz, F. & Samad, A. (2011). The use of verb-noun collocations in writing stories
among Iranian EFL learners. English Language Teaching, 4(3), 158-163.

Benson, M., Benson, E. & Ilson, R. (1997). The BBI dictionary of English word combinations
[Rev. ed. of: The BBI combinatory dictionary of English, 1986]. Amsterdam:
Benjamins.

Biber, D., Conrad, S. & Reppen, R. (1998). Corpus linguistics: Investigating language
structure and use. Cambridge: Cambridge University Press.

Biber, D., Finegan, E., Johansson, S., Conrad, S. & Leech, G. (1999). Longman grammar of
spoken and written English. London: Longman.

Brown, H. (2000). Principles of language learning and teaching (4th ed.). Englewood Cliffs:
Prentice-Hal.

Corder, S. P. (1967). The significance of learners' errors. IRAL, 5, 161-170.

Crystal, D. (1997). A dictionary of linguistics and phonetics (4th ed.). Oxford: Blackwell.

Darvishi, S. (2011). The investigation of collocational errors in university students’ writing

majoring in English. IPEDR, 18, 52-56.

Dukali, A. (2010). Collocations in learners’ English: A case study of collocation with respect
to delexicalized verbs in Libyan university students’ English writing. Unpublished
MA thesis, Manchester Metropolitan University.

88
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Elkhatib, A. S. A. (1984). A classification of the lexical problems of EFL/ESL students.

Retrieved from http://eric.ed.gov/?id=ED246691.

Ellis, R. (2008). The study of second language acquisition (2nd ed.). Oxford: Oxford
University Press.

Fan, M. (2009). An exploratory study of collocational use by ESL students: A task-based

approach. System, 37(1), 110-123.

Farghal, M. & Obiedat, H. (1995). Collocations: A neglected variable in EFL. International

Review of Applied Linguistics, 33(4), 315-31.

Francis, B. & Poole, R. (2009). Oxford collocations dictionary for students of English.
Oxford: Oxford University Press.

Gass, S. M. & Selinker, L. (2008). Second language acquisition: An introductory course (3rd
ed.). New York: Routledge.

Granger, S. (1998). Prefabricated patterns in advanced EFL writing: collocations and

formulae. In: Cowie, A. P. (Ed.), Phraseology: Theory, analysis and application (pp.
145-160). Oxford: Oxford University Press.

Hill, J. (2000). Revising priorities: from grammatical failure to collocational success. In:
Lewis, M. (Ed.), Teaching collocation: Further developments in the lexical approach
(pp. 47-69). London: Language Teaching Publications.

Hong, A. L., Rahim, H., Hua, T. & Salehuddin, K. (2011). Collocations in Malaysian English
learners’ writing: A corpus-based error analysis, 3L; Language, Linguistics,
Literature, The Southeast Asian Journal of English Language Studies, 17(special
issue), 31-44.

Howarth, P. (1998). The phraseology of learners’ academic writing. In: Cowie, A. P. (ed.)
Phraseology: Theory, analysis and application (pp. 161-186). Oxford: Oxford
University Press.

Kuo, C. (2009). An analysis of the use of collocation by intermediate EFL college students in
Taiwan. ARECLS, 6, 141-155.

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

89
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Lewis M. (2000). Teaching collocation: Further developments in lexical approach. London:

Language Teaching Publications.

Li, C. (2005). A study of collocational error types in ESL/EFL college learners' writing.
Unpublished M.A. thesis, Ming Chuan University.

Mahmoud, A. (2005). Collocation errors made by Arab learners of English. Asian EFL
Journal, 5(2), 117-126.

Martelli, A. (2007). Lexical collocation in learner English: A corpus-based approach.

Alessandria: Edizioni dell’Orso.

McCarthy, M. (1990). Vocabulary. Oxford: Oxford University Press.

McCarthy, M. & O’Dell, F. (2005). English collocations in intermediate use: How words
work together for fluent and natural English self-study and classroom use. Cambridge:
Cambridge University Press.

Miyakoshi, T. (2009). Investigating ESL learners' lexical collocations: The acquisition of

verb+ noun collocations by Japanese learners of English. Unpublished PhD thesis,
University of Hawai'I, Manoa. Retrieved from http://eric.ed.gov/?id=ED513117.

Namvar, F., Nor, N., Ibrahim, N. & Mustafa, J. (2012). Analysis of collocations in the Iranian
postgraduate students’ writings. The Southeast Asian Journal of English Language
Studies, 18(1), 11-22.

Nesselhauf, N. (2003). The use of collocations by advanced learners of English and some
implications for teaching. Applied Linguistics, 24(2), 223-242.

Phoocharoensil, S. (2011). Collocational errors in EFL learners’ interlanguage. Journal of

Education and Practice, 2(3), 103-120.

Sinclair, J. (1991). Corpus, concordance, collocation. Oxford: Oxford University Press.

Stefanowitsch, A. & Gries, S. (2003). Collostructions: Investigating the interaction of words

and constructions. International Journal of Corpus Linguistics, 8(2), 201-243.

Stubbs, M. (2005). The most natural thing in the world: Quantitative data on multi-word
sequences in English. In: Conference presentation at Phraseology. Louvain-la-Neuve,
Belgium, 13-15 October 2005.

90
 Journal of Second Language Teaching and Research. Volume 6, Special Issue

Taylor, L. (1990). Teaching and learning vocabulary. Hemel Hempstead: Prentice Hall.

Tengler, H., Aburiaiza, O., Ali, M. & Bakarally, B. (2013). Common mistakes made by ESL
learners using Arabic as reference language. King Abdulaziz University.

Thornbury, S. (2002). How to teach vocabulary. London: Longman.

Wang, Y. & Shaw, P. (2008). Transfer and universality: Collocation use in advanced Chinese
and Swedish learner English. ICAME Journal, 32, 201-232.

Wray, A. (2002). Formulaic language and the lexicon. Cambridge: Cambridge University
Press.

Yang, H. (2002). An introduction to corpus linguistics. Shanghai: Shanghai Foreign

Language Education Press.

Ying, Y. & O’Neill, M. (2009). Collocation learning through and ‘AWARE’ approach:
learner perspectives and learning process’. In: Barfield, A. & Gyllstad, H. (eds.)
Researching collocations in another language: Multiple interpretations. Basingstoke:
Palgrave Macmillan, 181-193.

91