The effects of consciousness raising tasks on efl students acquisition of lexical chunks

ABSTRACT This study aimed at attesting the effect size of consciousness raising tasks CR tasks in boosting EFL learners’ acquisition of lexical chunks or multi-word units.. Among the wid

HO CHI MINH CITY, April 2014

VIETNAM NATIONAL UNIVERSITY- HO CHI MINH CITY

UNIVERSITY OF SOCIAL SCIENCES AND HUMANITIES

FACULTY OF ENGLISH LINGUISTICS & LITERATURE

THE EFFECTS OF CONSCIOUSNESS RAISING TASKS

ON EFL STUDENTS’ ACQUISITION OF LEXICAL

CHUNKS

Submitted to the Department of English Linguistics and Literature

in partial fulfillment of the Master’s degree in TESOL

By

CÙ NHỰT SƯƠNG

Supervised by NGUYỄN THU HƯƠNG, PhD

HO CHI MINH CITY, April 2014

VIETNAM NATIONAL UNIVERSITY- HO CHI MINH CITY

UNIVERSITY OF SOCIAL SCIENCES AND HUMANITIES

FACULTY OF ENGLISH LINGUISTICS & LITERATURE

THE EFFECTS OF CONSCIOUSNESS RAISING TASKS

ON EFL STUDENTS’ ACQUISITION OF LEXICAL

CHUNKS

Submitted to the Department of English Linguistics and Literature

in partial fulfillment of the Master’s degree in TESOL

By

CÙ NHỰT SƯƠNG

Supervised by NGUYỄN THU HƯƠNG, PhD

My heartfelt thanks also go to Trinh Hoang Yen, who patiently spent hours teaching an IT illiterate like me sophisticated features of the SPSS software as well

as being my demanding proofreader, giving me a much-needed sense of direction whenever I lost my heart along the way

A sincere word of thanks should then be sent to all of my students and colleagues at Viet Uc Language School and Banking University of Ho Chi Minh City, especially Perry Graf and Bryan Wise, who helped me a lot in validating the

my research instruments Throughout the various stages of this research, their participation- either enthusiastic or reluctant, and their support- either direct or indirect, have never been of less importance to its final success

Furthermore, if the much-quoted saying, “Happiness is a journey, not a destination” is trustworthy, then Bui Thi Loc, Le Hung Vu, Tran Ky Phong, Ly

Thi Kim Cuong, Pham Thi Thanh Thao, Dang Thi Doan Trang, and Le Thi Mong Tuyen… are names everlastingly capable of ringing a bell to me because m y arduous journey so far had become much more colorful and pleasant thanks to the ups and downs we shared I do appreciate them all

Finally, this thesis would have been impossible without my family members’ encouragement I owe my thanks to my father- Cu Tien Duc, my mother- Pham

Thi Be, my two elder sisters- Cu Yen Ha, Cu Khanh Thu, and my fiancé Doan Van Tuan Anh- for their tolerance of my obsessive preoccupation in writing this Master thesis Those beloved ones would barely touch upon a single page of the completed thesis and would never know I kept on fighting till the end just to guarantee that their blind trust in me had not been misplaced Let me keep them here in a place closest to my heart!

STATEMENT OF ORIGINALITY

I certify m y authorship of the thesis submitted today entitled:

THE EFFECTS OF CONSCIOUSNESS RAISING TASKS ON EFL

STUDENTS’ ACQUISITION OF LEXICAL CHUNKS

in terms of the statement of the Requirements for the Theses in Master’s Program issued by the Higher Degree Committee This thesis has not been submitted for the award of any degree or diploma in any other institutions

Ho Chi Minh City, 2014

CU NHUT SUONG

RETENTION AND USE OF THE THESIS

I hereby state that I, CU NHUT SUONG, being the candidate for the degree of Master of Arts in TESOL, accept the requirements of the University relating to the retention and use of Master’s theses deposited in the Library

In terms of these conditions, I agree that the original of my thesis deposited in the Library should be accessible for purposes of study and research, in accordance with the normal conditions established by the Library for the care, loan or reproduction of the thesis

Ho Chi Minh City, 2014

CU NHUT SUONG

TABLE OF CONTENTS

Acknowledgements i

Statement of Originality iii

Retention and use of the thesis iv

Table of content v

List of abbreviations ix

List of tables xi

List of figures xii

Abstract xiii

CHAPTER 1: INTRODUCTION 1

1.1 Background and significance of the research 2

1.1.1 The rising role of chunks in modern academia and pedagogy 2

1.1.2 The intolerable situation of learners’ chunk knowledge sized up from the current study setting 3

1.1.3 The lumbering progress of chunk-oriented research and the need to triangulate conflicting research findings 4

1.2 Research aim 4

1.3 Research questions and research hypothesis 5

1.3.1 Research questions 5

1.3.2 Research hypothesis 5

1.4 Organization of the thesis 6

CHAPTER 2: LITERATURE REVIEW 7

2.1 Overview of lexical chunks 7

2.1.1 Definition 7

2.1.2 A historical review of lexical chunks 8

2.1.3 The nature of lexical chunks 10

2.1.3.1 The saliency and arbitrariness of lexical chunks………… 11

2.1.3.2 The problem of terminology……… 12

2.2 Lexical chunks as operationalized for this research 18

2.3 Aspects of lexical chunk acquisition 20

2.3.1 Chunk size 20

2.3.2 Chunk depth 21

2.4 Proposals for chunk instruction 23

2.4.1 Brief of second language vocabulary instruction 23

2.4.2 Common chunk teaching approaches 24

2.4.3 Theoretical grounds behind consciousness raising tasks 27

2.4.4 Types of CR tasks 30

2.5 Prior studies 34

2.5.1 Studies into the role of CR tasks in L2 teaching in general 35

2.5.2 Studies into the role of CR tasks in teaching lexical chunks 35

2.5.2.1 Empirical evidences for CR tasks in chunk teaching…… 36

2.5.2.2 Empirical evidences against CR tasks in chunk teaching… 37

2.6 Conceptual framework 39

2.7 Chapter summary 40

CHAPTER 3: METHODOLOGY 42

3.1 Research design 42

3.1.1 Pretest- Posttest Quasi- Experimental Design and its rationales 42

3.1.2 Convenience sampling method and risks to internal validity 44

3.1.2.1 Matching technique……… 45

3.1.2.2 Other courses of action devised to ensure internal research validity 46

3.2 Research site and participants 47

3.2.1 Research site 47

3.2.2 Participants 47

3.3 Data types 49

3.3.1 Vocabulary size test (or vocabulary proficiency test within this research situation) 49

3.3.2 Collocational competence test (chunk size parameter) 51

3.3.2.1 Construct, design and piloting……… ……52

3.3.2.2 Test reliability……… …57

3.3.2.3 Test administration and scoring method……… 57

3.3.3 Delayed written posttest (chunk depth parameter) 57

3.3.3.1 Construct, design and procedures……… ……… 57

3.3.3.2 Scoring criteria……….58

3.3.3.3 Test reliability……… ……….65

3.3.4 Questionnaire 65

3.3.4.1 Design and constructs……….……… 65

3.3.4.2 Piloting and reliability……… ……67

3.4 Research procedures 68

3.5 Data analysis 72

3.5.1 Questionnaire data analysis 73

3.5.2 Experimental data analysis 73

3.6 Chapter summary 73

CHAPTER 4: RESULTS 74

4.1 Statistical analysis of pre-test results 75

4.1.1 Vocabulary size test (placement test) 75

4.1.2 Collocational competence pre-test 76

4.2 Statistical analysis of post-test results 77

4.2.1 The collocational competence posttest 77

4.2.1.1 Paired sample T-Test for the comparison group……… 77

4.2.1.2 Paired sample T-Test for the experimental group………… ….79

4.2.1.3 Independent T-Test for the comparison group and the experimental group……….… 80

4.2.2 Delayed written posttest 82

4.3 Summary of the questionnaire analysis 82

4.3.1 Respondents’ background 82

4.3.2 The EG’s attitude towards the treatment 83

4.3.2.1 General attitude……… …….84

4.3.2.2 Preferential reasons for learning chunks with CR tasks……… …85

4.3.2.3 Unfavorable reasons against learning chunks with CR tasks……… ……85

4.3.2.4 Students’ confidence of the acquired chunks……… ……87

4.3.3 The EG’s perception of CR tasks’ efficacy on skill improvement 82

4.3.4 Students’ practice as a result of CR tasks 91

vii

4.4 Chapter summary 92

CHAPTER 5: DISCUSSION OF THE FINDINGS 93

5.1 Research question 1 93

5.1.1 From noticing to depth of processing 94

5.1.2 Higher frequency of input 97

5.1.3 Self- regulated learning resulted from a heightened awareness 98

5.2 Research question 2 99

5.3 Chapter summary 101

CHAPTER 6: CONCLUSION AND RECOMMENDATIONS 102

6.1 Summary of the key findings 102

6.2 Implications 104

6.3 Limitations of the study 105

6.4 Recommendations for further research 106

6.5 Chapter summary 107

REFERENCES 108

APPENDICES 123

LIST OF ABBREVIATIONS

BNC: British National Corpus

CCAD: Collin COBUILD Advanced Dictionary

CCPostT: Collocational Competence Posttest

CCPreT: Collocational Competence Pretest

CEFR: Common European Framework of Reference for Languages

CG: Comparison Group

COCA: Corpus of Contemporary English

COBUILD: Collins Birmingham University International Language Database CUP: Cambridge University Press CR:

Consciousness Raising DelayedWT:

Delayed Written Posttest EAP: English

for Academic Purposes EFL: English

as a Foreign Language EG:

Experimental Group

ELT: English Language Teaching

GW: Group work

IELTS: International English Language Testing System

L1: First language (mother tongue)

L2: Second language

LTM: Long-term memory

M: Mean

MCQ: Multiple Choice Question

OUP: Oxford University Press

P: Significance (two- tailed)

Sig (2- tailed): Significance (two- tailed)

SPSS: Statistics Package for the Social Sciences STM: Short-term memory

VST: Vocabulary Size Test

: Mean Score

LIST OF TABLES

Table 3.2 Description of the research participants 48

Table 3.3 The criteria grading scales for the delayed written posttest 58

Table 3.4A, B, C, D Detailed grading tables for the delayed written posttest 59

Table 3.5 Summary of the questionnaire 67

Table 3.6 Description of reliability statistics of the questionnaire 68

Table 3.7 Content of the experimental lessons outlined for 8 weeks 68

Table 3.8A Summary of the Treatment for the Experimental Group 68

Table 3.8B Summary of the Treatment for the Comparison Group 71

Table 4 1 One-Sample Kolomgorov-Smirnov Test of Groups Under Study 75

Table 4 2 Summary of the T-Test Result for the VST 76

Table 4 3 Summary of the T-Test Result for the CCPreT 76

Table 4 4 Paired Sample Statistics for the CG CCPreT and CCPostT 76

Table 4 5 Results of Paired Samples Test for the CG CCPreT and CCPostT 78

Table 4 6 Paired Sample Statistics for the EG CCPreT and CCPostT 79

Table 4 7 Results of Paired Samples Test for EG CCPreT and CCPostT 79

Table 4 8 Summary of the T-Test Result for the CCPostT 80

Table 4 9 Summary of the T-Test Result for the Delayed Written Posttest 82

Table 4 10 The EG’s pre-existing knowledge of chunk notion and CR tasks 83

Table 4 11 Preferential reasons for learning lexical chunks with CR tasks 85

Table 4 12 Student’s perception of CR tasks’ efficacy on skill improvement 90

Table 4 13 Student’s after-class behaviors 91

LIST OF FIGURES AND CHARTS

Figure 2.1 Chunk categorizing criteria 15

Figure 2.2 Traditions and models of chunks 16

Figure 2.3 Chunk operationalization for the current study 19

Figure 2.4 How CR tasks foster chunk acquisition 31

Figure 3.1 Summary of the research design 44

Chart 4.1 The EG versus the CG’s performance on the CCPreT and CCPostT 81

Chart 4.2 The EG’s general attitude towards the CR task- engrossed treatment 84 Chart 4.3 The retention rate after each treatment session, as self-rated by the EG students 85

Chart 4.4 The accurate use rate after the treatment, as self-rated by the EG students 89

ABSTRACT

This study aimed at attesting the effect size of consciousness raising tasks (CR tasks) in boosting EFL learners’ acquisition of lexical chunks (or multi-word units)

To tentatively touch upon this rather confounding and multifaceted issue, a quasi-experimental pretest-posttest, non-equivalent control group design research was employed with a participant pool of forty-seven intermediate learners, testifying their level of chunk mastery within 2 major foci namely size and depth Additionally, the attitudinal scale of the treated subjects was also examined via a questionnaire to further confirm the credibility of CR tasks in chunk instruction

That the students who learned chunks via CR tasks excelled those experienced the conventional pedagogic mode (i.e teacher-fronted instruction) was statistically affirmed in both the size and depth tests Additionally, the vast majority of the experimental group avowed their positiveness towards CR tasks in their chunk learning

In brief, these yielded findings justified that the proposed instruction can enjo y our privilege It is consequently postulated that such empirical evidence will in turn challenge the deep rooted belief that the field of chunk research is nothing but

a morass and ELT practitioners rarely move beyond mere improvisation

CHAPTER 1

INTRODUCTION

Since the advent of online corpora such as the British National Corpus (BNC) or the Corpus of Contemporary English (COCA) around two decades ago, there has been an observable awakening of vocabulary research Among the wide-ranging vocabulary- oriented research topics, chunks and the need for a more structured chunk instruction grasped closer attention of lexicographers, pedagogists, and researchers Now huge databases storing real-life language samples has opened up new horizon for them because they could genuinely

explore what language actually is rather than ponder over what language should

be In fact, the issue of lexical chunks is not a new one, as linguists from the early

twentieth century had kept asking themselves about the arbitrariness of lexical

chunks (or word combinations) The query as to why strong wind and heavy rain sound perfectly well but heavy wind and strong rain do not obviously laid the

very first stone of the road leading to the current state of studies into chunks

However, the scant attention paid to these “sounds - wrong - phenomena” failed

to blossom without systematic and powerful technological aids and deeper contemplation This is probably the reason why chunk was not replaced its crown until very recently

Norbert Schmitt, in the Introduction written for the book Vocabulary:

Description, Acquisition and Pedagogy, wondered uneasily and amusingly, “Oh,

no Not another book on vocabulary What makes this one different?” (Schmitt,

1997, p i) In so saying, Schmitt seems to have understood well the nature of research Admittedly, finding a good reason for a research work to attach true value to it and sustain its readers’ interest is what a piece of scientific research is supposed to work towards Therefore, it might be worth mentioning that this thesis- level study laid its heart and soul in accomplishing such daunting tasks and was sparked off intolerant of the risks mentioned in two questions in the following page

-Is this just a lousily chosen topic no other than a quick buck of rising trends in the field of research?

-Is this merely a copycat of existing research, which eventually yields no new findings at all?

This initial thesis chapter was written to prove that the opposite of the questions posed had been ensured and also to provide its readers a roadmap to follow In brief, chapter 1 will centralize around four sections: (1) background and significance of the problem, (2) research aim, (3) research questions and research hypotheses, and (4) organization of the thesis

1.1 Background and significance of the research

The framework within which this research project came into shape entails the following key urgencies

1.1.1 The rising role of chunks in modern academia and pedagogy

A glance at the existing vocabulary research body reveals that chunks (or formulaic language) have lately gained due acknowledgement because they help develop fluency and are the key for learners’ advance from intermediate to more advanced level (Schmitt, 1996; Lewis, 2002) Admittedly, language learners may spend years learning a vast stock of vocabulary and grammar rules but still their speech or writing does not belong to the so-called native- like selection The reason behind this phenomenon is relatively simple In Swan’s words, “Language

of this kind (i.e formulaic language) is notoriously challenging for learners.” (2006)

So tricky as they are, lexical chunks, on the other hand, could be said “to have entered the house of language teaching and are making themselves at home”, as emphasized by the American linguist Ben Zimmer, editor- in- chief of the Visual Thesaurus Dictionary This can be obviously seen in the fact that curriculum developers and textbook writers tend to attach continuing importance

to chunks A perceived reality is that in widely-used textbook series such as New

Cutting Edge, Speak Out, New English File, Interactions, or Mosaic…to name

but a few, vocabulary is now more commonly introduced in sets rather than isolatedly Although a systematic and consistent inclusion of chunks is yet to be found, isn’t it high time asserting the luminescence of the issue under this study?

1.1.2 The intolerable situation of learners’ chunk knowledge sized up from the current study setting

The school where this study was conducted offers intensive IELTS preparation courses and the researcher had been working there as an IELTS instructor for more or less five years From her personal observations, most IELTS trainers there tend to devote much of their class time to improving learners’ skills and test-taking techniques, while overlooking the role of vocabulary teaching (not to mention chunk teaching) in their syllabus In consequence, vocabulary teaching as perceived by them simply means naming several accredited vocabulary books in the hope that learning then may take care

It was two afore-stated problems that urged for immediate rectification for this uncomfortable situation

1.1.3 The lumbering progress of chunk-oriented research and the need

to triangulate conflicting research findings

Since the pervasion and usefulness of chunks is seen to be renounced, this undeniably raises again the question of which chunk instruction modes would be the most effective in ensuring maximal acquisition and durable retention

There have been a number of chunk teaching and learning methods proposed both by ELT theorists and ELT practitioners themselves Those range from the once - enshrined method such as rote rehearsal to seemingly more innovative approaches embodied within the communicative curriculum Generally speaking, they might possibly be classified into two main divisions: the incidental (or implicit) methods and the intentional (or explicit) ones While the former mode tends to be heavily weighted in favor of self- chunk learning via extensive reading, the latter takes into account the low rate of chunk incidentally picked by learners and thus opts for a more direct and hopefully better mode The group of explicit chunk instruction is again placed into two sub-groups The first one is known as direct explicit instruction, which encourages chunk instructors to deliver direct presentation and clear explanations of new chunks (or teacher-fronted instruction) By contrast, indirect explicit teaching means that chunks are still at the heart of the lesson aims, but the way teachers introduce them is less direct, leaving students more space for self-exploration This is also the framework upon which consciousness raising (CR) tasks - the chunk instruction examined in this thesis are theoretically built

It is difficult to say whether which instruction works best because there are now conflicting findings reported by prior studies into CR tasks’ effectiveness as

a chunk instruction As such, it should be stressed that there exists an urgent need

to gain further empirical evidence to settle this inconclusive research status quo, especially in Vietnamese ELT settings

1.2 Research aim

The aim of this study is two-fold:

(1) To find out the effectiveness of consciousness-raising tasks on lexical chunks

acquisition at two levels: (a) size and (b) depth

(2) To investigate students’ perception on the use of CR tasks in lexical chunks instruction

1.3 Research questions and research hypothesis

a To what extent does the experimental group outperform the comparison one in terms of size of chunk acquisition after the treatment?

b To what extent does the experimental group outperform the comparison one in terms of depth of chunk acquisition after the treatment?

2 What are the students’ views on CR tasks as a chunk instruction mode?

1.3.2 Research hypothesis

The present study focuses on exploring the contribution (if any) of CR tasks

on EFL students’ acquisition of lexical chunks It is conducted with the assumption that CR tasks may serve as an effective way in presenting various aspects of lexical chunks to learners of English It also looks into learners’ perceptions of CR tasks as a chunk instruction That is to say, the attitudinal scale

of all those treated with CR tasks will be surveyed and then analyzed to reinforce the quantitative data collected Hence, it is hypothesized that:

1 The experimental group will significantly outperform the comparison group

in terms of their gains in (a) chunk size and (b) chunk depth

2 Students will have positive views towards the use of CR tasks

1.4 Organization of the thesis

This thesis encompasses 6 chapters Chapter 1 provides an overview of the study, including the background and significance of the research problem, research aim, research hypothesis and research questions Chapter 2 is a critical and retrospective analysis of prevalent theories about the chosen topic On this sound theoretical base, a conceptual framework is formed and clearly stated The subsequent chapter then reports on methodology-related issues such as research design, study setting, research subjects, research tools, research procedures and how the collected data was statistically analyzed Chapter 4 covers the results gathered after the treatment The findings presented therein include the T-test analyses of the pre- and post- Collocational Competence Test, the T-test analyses conducted for the delayed Writing Posttest scores, and simple descriptive statistics of the attitudinal questionnaire Next, an in- depth discussion of the reported findings could be found in chapter 5 Finally, chapter 6 is for the conclusion and some implications (taking into account the remaining limitations

of the study) to be presented

CHAPTER 2

LITERATURE REVIEW

As stated in chapter 1, this paper seeks to examine the effects of consciousness raising (CR) tasks on EFL students’ acquisition of lexical chunks The present chapter will provide a theoretical framework that serves as a basis for establishing the conceptual framework of the study It first attempts to clarify a number of key terminologies embodied in the study Then a brief review of past and more recent literature regarding CR tasks in the context of chunk teaching will

be systematically conducted This critical review aims at painting a broad picture

of the ongoing research status by highlighting achievements as well as shortcomings of previous studies Most importantly, the identified gap will be used

as justification for the chosen topic to take on added importance

2.1 Overview of lexical chunks

2.1.1 Definition

Lexical chunk is just an umbrella term In fact, there is a variety of terminologies used to refer to lexical chunks, many of which are synonymous (and could be used interchangeably) or hyponymous (with one term entailing others)

Those include multi-word units, prefabricated patterns, formulaic speech, polywords, fixed or semi- fixed phases, sentence stems, canned phrases, lexical bundles, institutionalized expressions, phraseology, idiomatology, phrasal constraints, and collocations That too many terms had been coined referring to

this single complicated linguistic phenomenon did cause great confusion for researchers As noted by Michael Lewis (1993), “lexical chunks can usefully be sub-categorized, but any categorization will involve marginal cases, overlapping categories and fuzzy edges (p 92) Therefore, an overview of lexical chunks should probably start with a certain definition which is general enough to encompass the whole spectrum of chunks

A pure linguistic view tends to look at the predictability of word co- occurrences to distinguish between chunks and non-chunks In this sense, “lexical chunk is the readily observable phenomenon whereby certain words co-occur in

natural text with greater than random frequency” (Lewis, 1997, p 8) That is to say, in any word combination functioning as a chunk, the existence of this word will naturally call up its lexical partner Injeeli (2013) clarified this definition using

a very vivid example:” If I give you one word, you can predict the other word, with varying degree of success This predictability is not 100%, but it is always much higher than with non- collocates (or non-chunks)” (p 90)

An applied linguistic perspective, on the other hand, defines chunks in relation to a language user’s mental lexicon Alison Wray, a renowned researcher

in the field of formulaic language, commented:

A lexical chunk is a sequence, continuous or discontinuous, of words or other elements, which is, or appears to be, prefabricated: that is, stored and retrieved whole from memory at the time of use, rather than being subject to generation or analysis by the language grammar (2002, p 465)

As such, a lexical chunk, in its broadest sense, may be defined as any cluster of words whose combination is of greater than randomness and constitute just a single memory unit

2.1.2 A historical review of lexical chunks

Lexical chunks have a rich history in linguistics and related fields

The first noticeable advent was in 1956, when the term “chunking” was

coined inherent in psychologist George Miller’s Cognitive Load Theory This theory holds that human’s short term memory capacity is restricted within merely

5 to 9 units of information Therefore, to expand the existing storage space, there is always a tendency for previously stored input to gather in sets as a compensation strategy of our long term memory (Miller, 1956) The theory proved itself to be an influential one, and stimulated a bundle of follow-up research as well as derivational theories by both psychologists and psycholinguists In 1966, Michael Halliday highlighted Miller’s claim via his famous example involving the

commonness of the expression “strong tea”, as opposed to the inappropriateness of

“powerful tea” (which should have been an absolutely perfect word combination

from a pure grammarian’s view) Widdowson, a pioneer in Communicative

Language Teaching, took this point further by asserting that communicative competence is not a matter of knowing rules, but is much more a matter of knowing a stock of partially pre-assembled patterns or formulaic framework (Widdowson, 1989, p 135) Such a perspective apparently called the Chomskyan linguistic school into question and specialists in the field started to divert their attention away from Chomsky’s dominant concept of creativity to novel thought- provoking aspects such as chunks and the ways our brain handles them

In 1991, the clash with Chomskyan followers was intensified thanks to a

stronger theoretical and empirical evidence base, following the Collins

Birmingham University International Language Database (or COBUILD) project,

whose greatest achievement was the invention and analysis of a corpus of contemporary English text This marvelous creation allowed John Sinclair, the COBUILD project director, to make his voice heard in proposing the open-choice and idiom principles as two primary mechanisms of language use and persuasively concluding that the latter is of a far superior position (Sinclair, 1991, p 110) In an attempt to put this ubiquitous principle into practice, a couple of linguists rolled their sleeves up, seeking for a more applicable chunk-based pedagogy Then came Michael Lewis in 1993 with his book “The Lexical Approach”- a book which was probably mind-stimulating enough to make all of Lewis’ precursors proud This book was written focusing on 20 key assumptions, among which the claim

“language consists of grammaticalized lexis, not lexicalized grammar” was a cornerstone (Lewis, 1993, p 34) To borrow Scott Thornbury- a famous vocabulary researcher’s word, this “lexical turn” fuelled a major paradigm shift in Applied Linguistics (Thornbury, 2006, p 55) In other words, it was somewhat an awakening of vocabulary instruction over that of grammar instruction in the existing ELT institutional syllabuses

According to the so-called Lexical Approach or Chunking Approach, teachers are encouraged to value the power of lexis in aiding L2 learners’ acquisition They should also bear in mind that vocabulary is best taught in chunks rather than in single items while grammar mastery simply means the competence

to decode regularities inherent in the encountered lexical chunks On this aspect,

Moon (1997) distinguished between an isolationist and a collocationist view of vocabulary He posited that the former views vocabulary as a body comprising of individual lexical items while the latter accepts the significance and intricacy of the links between words, i.e their strong clustering tendencies and the patterns which are associated with them (Moon, 1997, p 40) Thus, a Lexical Approach backer is assuredly a real collocationist who is willing to hand absolute power to vocabulary teaching in general and chunk teaching in particular

This approach received mixed review from linguists and pedagogists, with keen supporters such as Prof Norbert Schmitt and Ben Zimmer (The Visual Thesaurus editor) together with doubters like Michael Swan, Brett Reynolds and several other empirical researchers Even more worryingly, Harwood (2002) suspected that “Lexical Approach is a term bandied about by many but understood

by few” (p 139) This is apparently not a casual remark if we look into the wide- ranging literature review he did He also conceded that there is still much work to

be done before the approach can hope to become more fully integrated into the mainstream ELT course- books

In brief, chunks have confirmed their importance over time as a result of Lexical Approach backers’ hard work The need to place them at the centre of the syllabus is now pledged by sound theories and strong empirical evidences Therefore, the doubt that their rising role is just a temporary “new toy” effect (which will probably not last long) could be dispelled, and the full integration Harwood mentioned hereabouts may simply mean that this “new toy” is still waiting for a more structured pedagogy, signaling an empirical research gap to be bridged

2.1.3 The nature of lexical chunks

If we look carefully into the nature of lexical chunks, there seems to be justifiable reasons why they are of such importance and why years of both separate and joint research efforts still fail to end the quest for a standard chunk pedagogy

2.1.3.1 The saliency and arbitrariness of lexical chunks

The saliency (or commonness) of lexical chunks was initially recognized based on intuition, introspection and idiolect (Moon, op cit., p 50) Following that, the advent of corpora containing millions of words allowed for more scientific verifications regarding the pervasiveness of formulaic language Zimmer (2010), for instance, contended that “native English speakers have tens or hundreds of thousands—estimates vary—of these formulae at their command” Other benchmarking statistics include the research result gained by the COBUILD project, which found that 40 per cent of running texts are lexical chunks These findings prove that such robustness is arguably inherent in lexical chunks

Another aspect worth mentioning is the arbitrariness of lexical chunks Lewis (1997) emphasized that arbitrariness lies at the very heart of a lexical understanding of language He noted, “All lexical items are arbitrary- they are simply the consensus of what has been institutionalized, the agreed language which a particular group do use, selected from what they could use, actual language as opposed to theoretically possible language” (p 17) For example, this

particular object is called a dog, while another is named a cat (and not vice versa)

We cannot usefully ask “Why?” for such a conventional way of naming things

This uncomfortable fact tends to spread its influence to formulaic language (or lexical chunks) as well Let take a closer look at this issue via the examples below:

 A relative can be close, near or distant while a friend can be close, but neither distant nor near, although a close friend may be one of your nearest and dearest

 You can want something badly (a lot), which in no way corresponds to the standard meaning of badly in other contexts

 An amusing example of the arbitrariness of language can be found in the

unlikely world of the dirty joke; such jokes are common in many languages As English- speaker will tell you, such jokes are blue- which

comes as a surprise to Spanish speakers, who are convinced that they are

green (Lewis, 1997, p 18)

It is thus generally accepted that although fields such as etymology can provide us with some tentative clues of their meaning source, lexical chunks, just like man y other linguistic phenomena, are in fact arbitrary by nature

2.1.3.2 The problem of terminology

Terminology could be considered to be the root of any chunk- oriented research problem Lewis (1997) noted that most terminologies are frequently used loosely with inadequate theoretical support and still remain half-baked (p viii)

At first glance, the term “chunk” seems rather easy to define That said, any

group of words whose constituent parts co-occur beyond mere randomness ma y

tag itself a lexical chunk or multi-word unit For example, the noun rain will automatically call up the adjective heavy in a native speaker’s lexicon Similarly, perhaps no interlocutor would say that they construct proverbs like It never rains but it pours and social pleasantries such as Long time no see or No worries from scratch, carefully analyzing the meaning of individual words: “it”/ “never”/” rain” /“but”/ “pour”; “long”/” time”/” no”/” see”; “no”/”worries” before

putting them together to serve their communicative purposes

However, there is a totally different story in the domain of applied linguistic research Any researcher, from a novice to a real professional, has probably been well-aware of the utter confusion generated by the many-sidedness of this term Admittedly, lexical chunk (or multi-word unit) is a superordinate term which entails a number of categories, dependent upon the criteria used to define or classify them

According to Moon (1997), there are three criteria by which we could distinguish chunks from non- chunks as well as, to some extent, make a clear distinction across different chunk types

One prominent criterion is the institutionalization of a lexical item (i.e the

degree to which a multi-word item is conventionalized in the language or its statistical frequency in actual language use) As such, if a search for a standard (or

native- like) word combination such as “strong tea” using Google’s search engine

yields roughly 407,000 results, it would be a safe bet to anticipate a modest several

hundred instances brought out for “powerful tea”, most of which listed as error

examples The same result is to be gained from available corpuses such as BNC or COCA This statistical approach of chunk extraction was heavily relied on by modern corpus linguists, whose job is to build electronic or written corpuses By nature, frequency information as implied by the first criterion will well suit their need in structuring dictionary entries, where word occurrences and co-occurrences require strict precision

The second criterion, as proposed by Moon, is the degree to which a multi-

word item is frozen as a sequence of words, or its fixedness It is thus not difficult

for us to visualize a continuum of strength of association, sliding between two

extremes namely fixed chunks (e.g on the other hand, never in the other hand or

on the other hands) and free combinations (e.g a good man; a better man; a crazy

man) This form- oriented criterion was specified by Thornbury (2002) and Wray (2010) via two much-cited terms namely fixed chunks and semi-fixed chunks, with

the former referring to non-inflected word sets while the latter indicating combinations that may inflect in a regular or predictable way Notably, this categorization is widely accepted and can be seen to be operating in most traceable chunk typologies

The third criterion set forth by Moon (op cit., p 48), on the other hand, is a

meaning- oriented one She named it non- compositionality, or the degree to which

a multi-word item cannot be interpreted on a word-by-word basis, but has a

specialized unitary meaning For instance, the phrasal combination put down has a

clear straightforward meaning because the notion it evokes is attributable to both

put and down literally Nonetheless, conducting a word-by-word exploration of chunks such as put up with, kick the bucket or battery farming in seeking for their

meaning is obviously a more daunting task See the illustrations below

Word(s) Referential meaning (related notion)

PUT

DOWN

PUT DOWN (=PUT+ DOWN)

Word(s) Referential meaning (related notion)

Idiomatic meaning (≠ KICK+THE BUCKET)

In the second example, the idiomatic meaning of the chunk “kick the bucket” has diverged too widely from its literal meaning, making it indecipherable at first sight If one attempts to establish any loose connection between the former and the

latter meaning, a number of origin theories could be reached For example, as stated in the Oxford Advanced Dictionary (p 159), a common theory is that the euphemistic meaning (i.e to kick the bucket = to die) this chunk (or idiom) takes

on today indeed originates from a method of execution common in the Middle Ages A noose is tied around the neck while standing on an overturned bucket When the bucket is kicked away, the victim is hanged However, Boers et al (2006, p 356) asserted that not all idiomatic meanings are lucky enough to have a traceable original sense, which proves that the compositionality of chunks is a stringent criterion to follow Thornbury, for instance, divides chunks into core idioms (which cannot be understood from the meaning of their parts), figuratives (which is possible to see how the parts go together or make the whole once the meaning of the multi-word unit is known) and literal sequences (whose meaning can be understood from the meaning of their part) (2006, p.121) The three criteria are summarized in the illustration below

Figure 2.1: Chunk categorizing criteria

Moon also emphasized that “the three criteria operate together, are not absolutes but variables, and are present in differing degrees in each multi-word unit (op cit.,

p 43) In reality, it must be accepted that a rigid and decisive categorization is almost impossible as regards chunk-related terminology issues Therefore, Lewis contended that researchers and ELT practitioners should bear in mind the values of grey shades and fuzziness, rather than hopelessly look for pure black, pure white

or absolute facts (Lewis, 1993, p 42) Given this fact, each study managed to form

its own grounds within specific conceptual restrictions, resting on those principles

‘compound’, or is it not: a decision has to be taken because of placement conventions in paper books (op cit., p 49)

That is to say, terminology compromises are somehow intolerable because the job

of a dictionary compiler requires definite distinction while putting a vast amount of lexical instances into different sections of a written corpus Besides, ELT practitioners, in carrying out their various pedagogy- oriented tasks such as material selecting, assessing or teaching, formed some other traditions and models

of multi-words items Those traditions and models are worth discussing because they are developed to serve specific research and teaching purposes rather than pure theoretical ones Hence, their practicality is hoped to shed light on small-scale research like the one conducted for this MA thesis The diagram below briefly summarizes three traditions and models of multi-word units that evolved up until now

WHAT TO LOOK AT IN A CHUNK

GRAMMATICAL PATTERNS

PRAGMATIC MEANING

Figure 2.2: Traditions and models of chunks

Broadly speaking, different models foreground different characteristics in a chunk

According to Moon (op cit., p 48), model 1 (i.e semantic-based) is in many respects the most traditional This model aims at identifying chunks according to their referential meaning As such, chunks are understood as the irreducible semantic building blocks of the lexicon Important work here is that of Mitchell (1971) and Makkai (1972), where issues related to collocations and idioms were respectively discussed Model 2 (i.e syntax- based) signals the next period in the development of pure as well as applied linguistics The classic four-level Streamline textbook series (OUP, 1990), for example, followed this model and chunks were loosely introduced all through this series via Note columns which were not very different from grammar notes

The next generation of English textbooks saw an emergence of notional- functional syllabuses Therefore, the way popular textbook series of this generation such as International Express (OUP, 2004) or New Cutting Edge (OUP, 2006) dealt with chunks shows a strong preference for model 3 (i.e pragmatics- based)

More particularly, situationally-bound preferred formulae such as Sorry to keep you waiting, See you later, or Have a nice day… and sentence frames such as

Would you mind if…, The thing is….were the cornerstone in constructing lessons

for these textbooks Take, for example, the content layout of International Express Intermediate (OUP, 2006) Unit 1 is entitled “WELCOMING A VISITOR”, with tasks guiding students on how to use the appropriate social formulae to perform such pragmatic function Unit 3 was devoted to another common pragmatic function, and is named “GIVING OPINION” Obviously, model 3 was heavily exploited by textbook writers of the previous decade

Very recently, model 2 seems to gradually fall out of interest The argument against this model claims that structural peculiarities of multi- word items is found

to be unruly and as such should not function as a defining characteristic Likewise, pragmatics- based models can work well in a Communicative Language Teaching framework, but may mismatch other teaching approaches, which proves the

limitation of such models Consequently, after a long period of overreliance on discourse-structuring frames, ELT practitioners of recent years show a tendency to favor model 1 again as they realize that referential and pragmatic meanings deserve equal attention This resurging interest could be apparently seen in latest vocabulary books, where chunks are also introduced thematically based on their referential meaning

In brief, lexical chunks, or the real linguistic instantiations generated by the idiomaticity or collocability of a language could be said to be very salient but complicated by nature Even though they are considered an arbitrary linguistic phenomenon and diverged widely in types and characteristics, there are still a number of categorizing criteria as well as specific traditions and models, all of which attempt to account for a certain kind of multi-word unit (in a certain way) This issue, thus, requires any researcher’s careful attention before defining their own theoretical framework

2.2 Lexical chunks as operationalized for this research

To exercise the caution advised earlier, this section sets out to define the range of chunks used as the target of the current research To put it briefly, the lexical chunks falling into the scope of this research were chosen adopting a semantics- based model That means referential meanings worked as a primary defining factor in this selection stage while all pragmatics or grammar- related features were overlooked

According to Lewis (1997), chunks carrying pragmatic meanings include

conventional speech routines such as sentence frames (Would you mind if…, the thing is…), social formulae (see you later, have a nice day, yours sincerely) and discourse markers (frankly speaking, on the other hand…) Referential meanings,

on the other hand, calls up chunks which are collocations (widely travel, set the table), phrasal verbs (get up, log on, run out of), and idioms or catchphrases (get cold feet, mind your own business…) The chosen frame of reference for this stud y

will thus result in the exclusion of sentence frames, social formulae, or discourse

markers, leaving collocations, phrasal verbs and idioms in the candidate list Such

a focus also eschewed grammatical collocations (i.e constructions where a verb or

an adjective must be followed by a particular preposition or a noun must be

followed by a particular verb form such as depend on, afraid of, avoid+ v-ing…)

In addition, Paul Meara (1997, p 110) also commented that many empirical studies in this field seem to base its claim on just a handful of chunks and as such there exists a lack of generalizability from those research findings One other perceived shortcoming of those studies is the seemingly random list of chunks targeted in each Corpus-based or exploratory studies may handle this thorny issue easily using information on the lemmatized frequency of target chunks (Durrant and Schmitt, 2010) However, with empirical studies examining the effectiveness

of an instruction on learners’ chunk acquisition, a pool of target chunks collected based on their lemmatized frequency may seem too chaotic a list to compile meaningful and natural lesson plans for the treatment (and as such treated students may be taught under “an innovative learning condition” some stuff which is too artificial and mechanical) Therefore, this study attempted to avoid such confusion

by using thematic (or topic-based) features as a sub- selection policy That said, the chunks targeted were chosen out of particular topics and bear only referential meaning Even though lemmatized frequency is still one of the aspects considered

in the chunk sampling stage, the fact that those chunks are intrinsically topic-based

is hoped to ease the burden of material development for the intended treatment The diagram below illustrates how the pool of target chunks for this study was selected based on the operationalization adopted

THE POOL OF TARGET CHUNKS

Figure 2.3: Chunk operationalization for the current study

2.3 Aspects of lexical chunk acquisition

The human mental lexicon is sometimes called a “black box” because of its

psychological intricacy In the new heyday of vocabulary research, greater interest has been geared towards tapping on this complicated system and the mechanism

by which it functions Read (2000) also noted that there have been numerous efforts to specify the components of vocabulary knowledge and to propose what the key variables are (p 36) Yet, a definite vocabulary acquisition model and lexical knowledge dimension has not been found yet

As regards lexical chunk acquisition (i.e the way our brain deals with those multi-word units), there is a general consensus amongst psycholinguists that chunks behave more like individual words than like separately constructed sequences That is to say, chunks should be treated via a lexical rather than grammarian view, or at best, somewhere in between If this is the case, it is likely

to hypothesize that those aspects of vocabulary acquisition and vocabular y knowledge could then dictate aspects of lexical chunk acquisition as well

Generally speaking, chunk acquisition, just like vocabulary acquisition, includes the storage, processing, and retrieval of lexis This ongoing and incremental process would normally result in final products of varying qualities, which is no other than a spectrum of knowledge from shallower to deeper degree,

or from “recognition” to “mastery” (Ellis, 1993) However, for research purposes

and for convenience’s sake, two constructs namely size and depth of chunk

knowledge are habitually used as two quantifiable factors, dividing the formulaic lexicon into two general domains, with each consisting of several aspects

2.3.1 Chunk size

According to Schmitt (2010, p 38), size (or breath) of vocabulary knowledge in general and chunk knowledge in particular is by definition the number of lexical items “known” In the same context, depth implies some criterion of mastery, or how well a learner knows those items He also remarked that size and depth are in fact not an all-or-nothing concept That is to say, there will always be some measure of depth in a size parameter in a sense that a certain

criterion of mastery must be met, and so the two will always be confounded Such annoying confoundedness is normally addressed by researchers with caution and flexibility, dependent on the conceptualization they adopted

One common conceptualization is to use an acquisition continuum starting from no knowledge and ending with full mastery Along this continuum, the size

of a person’s formulaic inventory is meant to refer to the shallower acquisition level (or threshold) As such, all chunks which could generate a trace or certain mental linkages to the interim of his/her lexicon will add up to the so-called

“chunk size” In psycholinguistic terms, this is the issue of explicit or declarative

knowledge More particularly, the learners may know about the existence of a certain chunk rather than know how to use it in real situations

In chunk testing research, size of chunk knowledge implies receptive aspects, i.e., the ability to recognize the form and meaning of a sample of target chunks Such aspects could be measured by simple test items such as MCQ or matching

This sub-construct of chunk size, in Schmitt (2010)’s words, is receptive size

Additionally, chunk size may also imply higher degree of knowledge indicated b y testees’ ability to recall the form-meaning link of chunks Schmitt (op cit.) named

this second sub-construct recall (or semi-productive) size measures and claimed

that cloze test items could be used to that end

In brief, size of chunk knowledge, either receptive size or semi-productive size, entails earlier stages of the developing formulaic lexicon only That said, whether students have deeply ingrained those chunks into their memory and could use them spontaneously in their productive tasks is excluded from this construct and not discussed until depth of chunk knowledge is taken into consideration

2.3.2 Chunk depth

In reviewing depth as a distinct research variable, Schmitt (op cit., p 42) contemplated over a fanatical belief that depth is the wooliest, least definable, and least operationalisable construct in the entirety of cognitive science past or present The most comprehensive definition, according to him, is probably to equate depth with quality (as opposed to quantity), or how well a learner knows a lexical item

In answering that question, researchers’ views tend to diverge Daller, Milton, and Treffers-Daller (2007), for example, see size and depth of knowledge as a separate dimension from fluency, as illustrated below

However, most other linguists argue against such view and claim that fluency can

be seen as depth (Read, 2000; Thornbury, 2006; Schmitt, 2010) More specifically, depth does not have to be knowledge, but can be seen as employability in skills and displayed in students’ automaticity

Following this line of reasoning, in the current chunk-oriented study, depth

of knowledge simply means that the target chunks have been integrated into the learners’ lexicon for immediate access (Lewis, 1997, p 51) To be more specific, the depth of chunk knowledge could be understood as the automaticity in producing chunks, or that a learner feels comfortable or confident about dropping a chunk into an essay or uttering it in a fluent normal conversation Such capability

of using chunks at ease is emphasized by Read (2000, p 24), “One attribute of depth of knowledge of multi-word items is learners’ ability to produce fluent stretches of speech or writing without apparent planning or effort” In this way, depth, or the absolute degree of mastery, may help form the other end of the chunk acquisition continuum

In chunk testing, depth of chunk knowledge is normally gauged b y productive tests such as writing or oral test because a depth test is supposed to encompass the power of a more efficiently organized mental lexicon, or procedural knowledge Besides, depth also means durable memory traces (Nation, 2001, p 657) Therefore, the writing or oral test is usually delivered after a lapse of time in

order for those more elaborate, longer lasting traces to be detected See Chapter 3 (3.3) for more information on this

In short, size and depth of chunk knowledge are two constructs commonly used by studies targeting idiomatic language Briefly put, on a sample of target chunks, size signals whether a person knows, say, form and meaning of those chunks whereas depth notifies researchers of how well those chunks are known b y this person As a matter of fact, size is believed to be always bigger than depth because the number of chunks a person knows always exceeds the number of chunks he/she know well In language testing, chunk size is normally explored using recognition and recall tests Recognition and recall here mean that test takers are provided with some stimulus designed to elicit the target chunk (or part of it) from their chunk repertoire The stimulus may entail either chunk form or chunk meaning (or both) On the other hand, chunk depth is measured via oral or written tasks which require test-takers to prove their spontaneous and exact use of chunks

in a free-defined context As such, depth of chunk knowledge is different from size

of chunk knowledge in that it implies the capability of producing chunks at one’s own will without referring to stimulus of any type

2.4 Proposals for chunk instruction

2.4.1 Brief of second language vocabulary instruction

As can be observed from the extensive reviews on vocabulary pedagog y conducted by Meara (1996), Schmitt (1997), Thornbury (2002) and some other empirical researchers, vocabulary instruction has generally revolved around two main trends: implicit and explicit instructions The former is more commonly known by the term “guessing from context” or “incidental vocabulary learning”, whereby acquisition is fostered via repeated exposure to reading input The latter,

on the other hand, backs intentional (or formal) vocabulary teaching by focusing

on giving clear explanation for a pre-selected list of vocabulary items

Regarding implicit vocabulary instruction, material grading and selection could be said to play a key role in ELT practitioners’ daily chores More specifically, teachers should spend a great amount of time choosing the right input,

in the hope that sufficient immersion will facilitate vocabulary acquisition on the part of the students As for explicit vocabulary instruction, it involves the creation, trial and development of various teaching techniques which take the teacher- student interaction into consideration This interaction could be one- way and teacher-dominated as generated by mechanical drilling or rote-memorization Yet,

it could also look more ‘innovative” and ‘dynamic” when integrated into the task- based teaching framework where teachers model themselves on the L1 caretakers, largely avoiding the kinds of activities (like drilling and error correction) that caretakers do not use As such, vocabulary is treated merely as a means on the road

to obtaining the final communicative purpose

The two main trends in vocabulary teaching discussed above, with each lending its own strengths, have provided fundamental points of reference for theorists when it comes to the quest for a feasible chunk instruction

2.4.2 Common chunk teaching approaches

It is widely believed that chunks represent a peculiar lexical phenomenon Nattinger and De Carrico (1992) contend that a lexical chunk is a lexical- grammatical unit which exists somewhere between the traditional poles of lexicon and syntax (p 36) Indeed, that multi-word units behave as if they were single words but still bear characteristics of conventional syntactic features proves how challenging it is to teach them

In an implicit teaching tradition, ELT practitioners dealing with lexical chunks now have even more reasons to back proposals for implicit chunk learning (via extensive exposure to input in or out of class) Thornbury’s suggestion that “to teach lexical chunks, teachers should expose learners to lots of correct examples” (2002, p 107) asserts this tradition Given the sheer amount of chunks and their complicatedness, those arguments justify the unteachability of almost all chunks

In the meantime, advocates of explicit teaching tradition cast doubt on the claim that students are capable of picking up chunks as a by-product of reading and lean more towards activities aiming at chunk development predominantly However, they seem to differ widely in their view concerning how chunks could

be best taught As a result, up until now, no method seems to be principled enough

to be called an approach, except for some tentative proposals and detached on- hand techniques

One proposal for chunk instruction which is much exploited by current textbooks is the so-called rule-based “approach”, as reported by Thornbury (2002) This approach starts by isolating and highlighting any relevant patterns or regularities in a group of chunks The multi-word items are grouped and presented according to their manner of formation or their shared meaning nuances See some examples of a rule-based approach in action from two course book extracts below

VOCABULARY NOTES

Make or do?

Read the following sentences carefully

Last night I tried to do my homework However, I kept making mistakes because the man upstairs was doing his exercises and making a noise

Make usually means to create, bring into existence, or produce a result

Do usually means to perform an action However, there are exceptions to this ‘rule’, as you will

see in Exercise 3

(extracted from Bell J and Gower R., Intermediate Matters, Longman)

DOWN

PHRASAL VERB STUDY

Down is an adverb and a preposition The basic meaning of down is to do with movement from a

higher position or level to a lower one

1A Match these descriptions of down phrasal verbs with the sentences below

a Movement and position e Completeness, ending and change

b Decreasing, lowering and reducing f Eating and drinking

c Fastening and fixing g Writing and recording

d Collapsing and attacking

I drank down my double Scotch eagerly

The lid of the box was nailed down

The water floods their homes or breaks down the walls

Go and lie down on your bed

If the firms failed to make enough money, they would close down

They ask me the date and flight number: I always write it down so I will remember

It’s a bit hot in here- turn it down

(extracted from Radley P and Millerchip C., Workout Upper Intermediate,