See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321760269

Second Language Vocabulary Acquisition

Book · March 2009

CITATIONS READS

43 5,529

1 author:

Qing Ma
The Education University of Hong Kong
25 PUBLICATIONS 325 CITATIONS

SEE PROFILE

All content following this page was uploaded by Qing Ma on 13 June 2018.

The user has requested enhancement of the downloaded file.

Contents

List of abbreviations .................................................................................13

List of figures............................................................................................15

List of tables .............................................................................................16

Acknowledgements ..................................................................................17

Introduction...............................................................................................19

Part I: Issues Related to Second Language Vocabulary Acquisition

Chapter 1. Vocabulary Knowledge and Use...........................................27

1.1. What is vocabulary knowledge?.......................................................27
1.1.1. Vocabulary knowledge .................................................... 27
1.1.2. Vocabulary use ................................................................ 30
1.1.3. Declarative vs. procedural knowledge ............................. 31
1.2. Breadth and depth of vocabulary knowledge ..................................33
1.2.1. A clear picture vs. a blurred picture ................................. 33
1.2.2. Breadth of vocabulary knowledge ................................... 35
1.2.3. Depth of vocabulary knowledge ...................................... 36
1.2.4. Breadth vs. depth ............................................................. 37
1.3. Receptive and productive vocabulary ..............................................39
1.3.1. Introduction ..................................................................... 39
1.3.2. Receptive and productive vocabulary:
A continuum or a dichotomy? ......................................... 41
1.3.3. Relationship between receptive
and productive vocabulary in empirical studies ............... 42
1.4. Testing vocabulary knowledge and use............................................44
1.4.1. Introduction ..................................................................... 44

5
 1.4.2. Breadth and depth tests .................................................... 45
1.4.2.1. Breadth tests .......................................................... 45
1.4.2.2. Depth tests............................................................. 45
1.4.2.3. Breadth test vs. depth test ...................................... 46
1.4.3. Receptive and productive tests ........................................ 47
1.4.4. Tests combining the breadth/depth dimension
and the reception/production dichotomy .......................... 47
1.4.4.1. Vocabulary Knowledge Scale (VKS) .................... 48
1.4.4.2. CATSS (Computer Adaptive Test
of Size and Strength) .............................................. 49
1.5. Conclusion .........................................................................................50

Chapter 2. The L2 Lexicon: Development and Organisation ................53

2.1. The development of the L2 mental lexicon .....................................53
2.1.1. A separate or unified system? .......................................... 53
2.1.2. The development of a lexical entry .................................. 54
2.2. Organisation of the L2 mental lexicon .............................................63
2.2.1. Relations between lexical entries ..................................... 63
2.2.1.1. Possible relations between lexical entries.............. 64
2.2.1.2. Elaborating semantic relations
between English words .......................................... 65
2.2.2. Review of empirical studies investigating
the differences in organisation between the L1
and the L2 mental lexicon: word associations ................. 67
2.3. Conclusion .........................................................................................73

Chapter 3. The Role of Listening in SLVA ............................................75

3.1. The historical background ................................................................75
3.2. Listening: A dynamic language activity ..........................................76
3.3. Listening as the basic skill in SLA ...................................................78
3.3.1. Activation of both brain hemispheres .............................. 79
3.3.2. L1 acquisition as a model for L2 acquisition ................... 80
3.3.3. The transfer effects of listening to reading and speaking . 80
3.4. The importance of listening in vocabulary acquisition ...................82
3.5. Conclusion .........................................................................................83

6
Chapter 4. Vocabulary Learning Difficulty ............................................85
4.1. Cultural difference.............................................................................85
4.1.1. Language learning: A culture bound phenomenon .......... 85
4.1.2. Lexical concept gaps ....................................................... 87
4.1.3. Culturally loaded words................................................... 89
4.2. Linguistic distance ............................................................................91
4.2.1. What is linguistic distance? ............................................. 91
4.2.2. Measuring linguistic distance .......................................... 91
4.2.2.1. Approximating the distance
via language family trees...................................... 91
4.2.2.2. Minimum learning time needs
to reach a basic level ............................................ 92
4.2.2.3. Measuring the distance by
language proficiency level achieved..................... 92
4.2.3. How does linguistic distance affect SLVA? .................... 93
4.3. Lexical form confusion .....................................................................97
4.4. Conclusion .........................................................................................98

Part II: Approaches to L2 Vocabulary Acquisition

Chapter 5. Implicit and Explicit Approaches ........................................103

5.1. What do these terms mean? ............................................................103
5.1.1. Implicit and explicit learning ......................................... 103
5.1.2. Incidental and intentional learning ................................. 104
5.1.3. Implicit and explicit memory ......................................... 106
5.1.4. Implicit/explicit learning
vs. incidental/intentional learning .................................. 107
5.1.5. Implicit/explicit learning vs. implicit/explicit memory .. 107
5.1.6. The meaning of implicit/explicit approaches
adopted in this book....................................................... 108
5.2. The implicit learning/teaching paradigm .......................................108
5.2.1. Word inferencing studies ............................................... 110
5.2.2. Lexical glosses and dictionary use studies ..................... 111
5.2.3. The main drawbacks of incidental learning ................... 113
5.3. The explicit learning/teaching paradigm ........................................114
5.3.1. Reading plus exercises................................................... 115
7
 5.3.2. Explicit instruction/learning .......................................... 116
5.4. Conclusion .......................................................................................118

Chapter 6. Approaches to Multi-Word Items .......................................119

6.1. Background information .................................................................119
6.1.1. Language: lexis or grammar based?............................... 119
6.1.2. Corpora .......................................................................... 120
6.1.2.1. Native English corpora ........................................ 121
6.1.2.2. Learner English corpora ...................................... 123
6.1.3. Identifying and defining formulaic sequences ............... 123
6.2. Approaches to L2 formulaic sequence acquisition ........................127
6.2.1. Learning difficulties of L2 formulaic
sequence acquisition ...................................................... 127
6.2.2. Proposals for the acquisition of formulaic sequences .... 129
6.2.2.1. Classroom teaching/learning activities ................ 129
6.2.2.2. Syllabus design ................................................... 130
6.2.2.3. The lexical approach ........................................... 131
6.2.2.4. Data-driven learning............................................ 133
6.2.2.5. Conclusion .......................................................... 134
6.2.3. Empirical studies on the acquisition
of formulaic sequences .................................................. 135
6.2.3.1. Classroom settings .............................................. 135
6.2.3.2. Socio-cultural interaction .................................... 137
6.3. Approaches to the acquisition of idioms ........................................138
6.3.1. The acquisition of idioms by raising learner
awareness of idioms using a cognitive approach ........... 138
6.3.2. The mnemonic effect of alliteration
on the acquisition of idioms ........................................... 141
6.5. Conclusion .......................................................................................142

Chapter 7. A Learner-Focused Approach .............................................143

7.1. Learner differences..........................................................................143
7.2. Motivation .......................................................................................144
7.2.1. What is motivation? ....................................................... 144
7.2.2. Motivation for learning the L2....................................... 145
7.3. Style .................................................................................................149
7.3.1. What is learning style? .................................................. 149

8
 7.3.2. Learning style assessment tools ..................................... 150
7.3.2.1. The Myers-Briggs Type Indicator ....................... 150
7.3.2.2. The Perceptual Learning Style Preference Survey 151
7.3.2.3. The Style Analysis Survey .................................. 152
7.4. Strategies..........................................................................................154
7.4.1. Language learning strategies ......................................... 154
7.4.2. Vocabulary learning strategies ....................................... 156
7.4.3. Vocabulary learning strategy instruction ....................... 158
7.4.4. Classifying vocabulary learning strategies..................... 159
7.4.4.1. Vocabulary Learning Questionnaire .................... 159
7.4.4.2. Taxonomy of vocabulary learning strategies ....... 161
7.4.4.3. A learning process oriented approach ................. 163
7.5. Towards an integrated framework for SLVA .................................165
7.6. Conclusion .......................................................................................168

Part III: Computer Assisted Vocabulary Learning (CAVL)

Chapter 8. CAVL: Past, Present, and Future ........................................173

8.1. An overview ....................................................................................173
8.2. Factors shaping CAVL development .............................................174
8.3. Four stages of CAVL development................................................176
8.3.1. Simple tutor ................................................................... 177
8.3.2. Intelligent tutor .............................................................. 177
8.3.3. Tool ............................................................................... 178
8.3.4. Tutor and tool ................................................................ 180
8.4. Constructing a framework for categorising CAVL applications ..182
8.4.1. The overview of the framework ..................................... 182
8.4.2. Lexical resources/aids.................................................... 183
8.4.3. Lexical programs/tasks .................................................. 185
8.4.4. The applicability of the framework ................................ 187
8.5. Conclusion .......................................................................................188

Chapter 9. Two Evaluation Studies of CAVL Programs .....................189

9.1. CALL/CAVL evaluation ................................................................189
9.2. Constructing the CAVL evaluation criteria ....................................190
9.3. The evaluation of research-based CAVL programs .......................194
9
 9.3.1. Electronic reading/listening comprehension
with lexical glosses ........................................................ 194
9.3.1.1. Reading comprehension ...................................... 194
9.3.1.2. Listening comprehension .................................... 195
9.3.1.3. CMC lexical-based tasks ..................................... 197
9.3.1.4. Computerised vocabulary exercises .................... 200
9.3.2. Dedicated CAVL programs ........................................... 202
9.3.2.1. Lexica ................................................................. 203
9.3.2.2. CAVOCA............................................................ 205
9.4. The evaluation of commercial CAVL/CALL programs ................208
9.4.1. Three major types of commercial programs .................. 208
9.4.2. Electronic dictionaries/flash cards ................................. 209
9.4.2.1. SuperMemo ......................................................... 210
9.4.2.2. Rosetta Stone ...................................................... 212
9.4.2.3. Intelligent Miracle English .................................. 214
9.4.2.4. Memorising Vocabulary Effortlessly .................. 218
9.4.3. Computerised vocabulary exercises ............................... 221
9.4.3.1. Tell Me More ...................................................... 221
9.4.3.2. English+ .............................................................. 223
9.4.4. Dedicated-like vocabulary programs ............................. 225
9.4.4.1. Language Interactive Culture .............................. 225
9.4.5. Summary of the evaluation results ................................. 228
9.4.5.1. Research-based programs .................................... 228
9.4.5.2. Commercial programs ......................................... 230
9.5. Conclusion .......................................................................................232

Part IV: L2 (English) Vocabulary Acquisition in China

Chapter 10. English Teaching and Learning in China .........................239

10.1. Historically identified distinct periods of ELT in China .............239
10.2. Features of ELT in China .............................................................242
10.2.1. High motivation ........................................................... 242
10.2.2. Chinese culture of learning .......................................... 243
10.2.2.1. What is culture of learning? .............................. 243
10.2.2.2. Chinese culture of learning:
perception and practice..................................... 243

10
10.3. Approaches to ELT........................................................................247
10.4. Other constraints of ELT in China................................................250
10.5. Recent reform in ELT in China: focus on higher education
and the non-language specialist ....................................................253
10.6. Conclusion .....................................................................................257

Chapter 11. Vocabulary Learning Difficulties for Chinese Learners ..259

11.1. Three areas likely to cause learning difficulties...........................259
11.2. Difficulties due to cultural difference...........................................260
11.3. Difficulties due to linguistic distance...........................................262
11.3.1. Errors due to meaning confusion ................................. 263
11.3.2. Errors due to other areas .............................................. 264
11.4. Difficulties due to confusion of lexical forms .............................269
11.5. Exploring lexical errors from a psycholinguistic perspective .....271
11.6. Conclusion .....................................................................................273

Chapter 12. Two Empirical Studies of Chinese Learners’

Approaches to Vocabulary Acquisition ...............................................275
12.1. An overview from the literature ...................................................275
12.2. Questionnaire: Teachers’ beliefs and practices regarding
vocabulary teaching .....................................................................278
12.2.1. Constructing a questionnaire of teachers’ beliefs
and practices regarding vocabulary teaching ............... 278
12.2.2. Results and discussion of the questionnaire ................. 279
12.3. Questionnaire: Students’ approaches to vocabulary learning .....284
12.3.1. Constructing a questionnaire of process-oriented
vocabulary learning strategies ..................................... 284
12.3.2. Results and discussion of the questionnaire ................. 285
12.4. Conclusion .....................................................................................291

General Conclusion ................................................................................293

References...............................................................................................297

List of CAVL/CALL programs/tasks reviewed ...................................323

List of useful online lexical tools...........................................................324

11
Appendices .............................................................................................325
Appendix 1: A questionnaire for English teachers
at Chinese Three Gorges University (translated from Chinese) ....325
Appendix 2: A questionnaire for students
at Chinese Three Gorges University (translated from Chinese) ....328

Index........................................................................................................331

12
List of abbreviations

ACT Adaptive Control of Thought

AWL Academic Word List
BNC British National Corpus
CALICO Computer Assisted Language Instruction Consortium
CALL Computer Assisted Language Learning
CATSS Computer Adaptive Tests of Strength and Size
CAVL Computer Assisted Vocabulary Learning
CAVOCA Computer Assisted Vocabulary Acquisition
CIC Cambridge International Corpus
CL Culture of Learning
CLC Computer Learner Corpora
CLEC Chinese Learners’ English Corpus
CMC Computer-mediated Communication
CTGU China Three Gorges University
DDL Data Driven Learning
EAP English for Academic Purposes
EFL English as a Foreign Language
ELT English Learning/Teaching
ESL English as a Second Language
FL Foreign Language
FLL Foreign Language Learning
ICLE International Corpus of Learner English
ILV Institut de Langues Vivantes
IME Intelligent Miracle English
ITS Intelligent Tutoring Systems
L1 First Language
L2 Second Language
LFP Lexical Frequency Profile
LINC Language Interactive Culture
LINDSEI Louvain International Database
of Spoken English Interlanguage

13
LSS Learning Style Survey
MBTI Myers-Briggs Type Indicator
MVE Memorising Vocabulary Effortlessly
P Productive Lexical Items/Lexicon
PLSP Perceptual Learning Style Preferences
R Receptive Lexical Items/Lexicon
SAS Style Analysis Survey
SLA Second Language Acquisition
SLVA Second Language Vocabulary Acquisition
TMM Tell Me More
UCL Université Catholique de Louvain
UWL University Word List
VKS Vocabulary Knowledge Scale
VLQ Vocabulary Learning Questionnaire

14
List of figures

Figure 1. A simplified mental representation

of breadth and depth of vocabulary knowledge .................. 34
Figure 2. A hypothetical association network (Meara, 1990: 152) .... 42
Figure 3. The internal structure of the lexical entry (Jiang, 2000: 48) ..56
Figure 4. Lexical development in L2: from the formal stage
to the integration stage (adapted from Jiang, 2000: 54) ..... 57
Figure 5. Word-knowledge continuum and
main organisational features of words
in the mental lexicon (Namei, 2004: 382) ......................... 73
Figure 6. Phraseological categories (Howarth, 1998: 27)................ 126
Figure 7. A learner-focused model of SLVA.................................... 168
Figure 8. A framework for categorising CAVL ............................... 183
Figure 9. Vocabulary processing procedure in Words in your ear ... 195
Figure 10. Vocabulary processing procedure
in the program by Smidt and Hegelheimer ...................... 196
Figure 11. Vocabulary processing procedure
in the CMC lexical-based task of De la Fuente ............... 199
Figure 12. Vocabulary processing procedure
in the computerised exercises by Allum .......................... 202
Figure 13. Vocabulary processing in Lexica.................................... 205
Figure 14. Vocabulary processing in CAVOCA .............................. 207
Figure 15. The new teaching mode combining
classroom teaching and language software
(The National Education Ministry, 2004: 8) .................. 257
Figure 16. Lexical comprehension/production model
(adapted from De Bot et al. 1997: 315) ............................ 272

15
List of tables

Table 1. Ten categories of synforms

(assembled from Laufer, 1990: 294) ................................... 98
Table 2. Cognitive style listed in LSS (Cohen et al., 2005) ............. 154
Table 3. Vocabulary learning strategies listed in VLQ
(Gu/Johnson, 1996) ........................................................... 161
Table 4. Vocabulary learning strategies listed by Schmitt (1997) .... 162
Table 5. Historical development of CAVL ...................................... 182
Table 6. Criteria for evaluating CAVL programs/tasks .................... 191
Table 7. Summary of the evaluation of research-based programs ... 234
Table 8. Summary of the judgemental evaluation
of commercial programs.................................................... 235
Table 9. Examples of Chinese learners’ meaning errors
(adapted from Gu/Leung, 2002: 133) ................................ 264
Table 10. Examples of Chinese learners’ form and meaning errors
(adapted from Gu/Leung, 2002: 134-135) ......................... 264
Table 11. Examples of Chinese learners’ part of speech errors
(adapted from Gu/Leung, 2002: 135) ................................ 264
Table 12. Raw frequencies of connectors in CLEC
(1.07 million words) and the academic component
of BNC (in per million running words) ........................... 268
Table 13. Form errors classified according to Laufer’s categories
(Gu/Leung, 2002: 131) .................................................... 269
Table 14. Examples of Chinese learners’ form errors based
on syllables (adapted from Gu/Leung, 2002: 132) .......... 270
Table 15. Confusion of Chinese characters in Chinese
equivalents (adapted from Gu/Leung, 2002: 133) ........... 270
Table 16. Percentage of vocabulary teaching beliefs and practices . 281
Table 17. Average of vocabulary learning strategy use frequency ... 287
Table 18. Mean of use frequency of group strategies ...................... 290
Table 19. Mean of use frequency of codified strategies .................. 290

16
Acknowledgements

I wish to express my gratitude to Peter Kelly who gave me so many

valuable comments in respect of content and mode of expression and
Sylviane Granger, who showed me the right approach to carrying out
my earlier research which partially resulted in this book. I am grateful
to my former colleagues and friends at the University of Namur and
the Catholic University of Louvain for their constant support and
exchange of ideas. I am also indebted to the teachers and students at
China Three Gorges University who gave me feedback and
participated in my studies. Finally I would like to thank The Hong
Kong Institute of Education for providing financial support for the
proof-reading of the book.

17
18
Introduction

The term “second language acquisition” (SLA) covers all the main
areas of language, systematized by time-honoured terms like syntax,
grammar, lexis, pronunciation, etc., and covers the mass of research
that endeavours to discover how a language is acquired, what is
acquired and what is not acquired. The study of SLA can be conducted
from different angles. As Bialystock (1995) has pointed out, there is a
neurolinguistic, a linguistic and a psycholinguistic approach; each
approach can be used to explain different aspects of SLA with a
different end in view. Increasingly, attention is now being drawn to the
socio-cultural approach where the influence of individual background
can very much influence the way a language is learned and the extent
to which it is learned. All this clearly shows that the study of SLA has
become, in keeping with so many other areas of scientific study, an
interdisciplinary field. Researchers into SLA will find themselves
plunging into reviews of linguistics, psychology, sociolinguistics and
education, to mention but the most salient disciplines. All these
approaches can be applied to second language vocabulary acquisition
(SLVA), one of the most important components of SLA.
The term SLA has gained considerable currency in recent years
and it occurs now much more frequently in the literature than foreign
language learning (FLL), the term that was previously more widely
used. We can see why this has come about. Until two decades ago,
people used to speak of FLL to refer to the study of another language.
This language study usually took place in a formal educational setting
and in one’s own country; the focus was on the written language and
there was little contact with the people or the culture. Today, due to the
many political and social changes that have taken place, to the vastly
increased contact between peoples with the focus of learning shifting to
oral communication, the foreign language has become far less foreign
in the sense that it may be spoken officially, i.e. for administrative
purposes, or unofficially, i.e. by increasingly large groups of people

19
within a country or community who have a different L1. The term
“second” is more neutral and is totally free of the negative nuances that
might be associated with “foreign”1. Block (2003: 32–91) discusses the
“S” in SLA in great detail. He cites Mitchell/Myles (1998) who
consider it “sensible to include ‘foreign’ languages under our more
general term of ‘second’ languages” (cited in Block, 2003: 32). Gass/
Selinker (2001: 5) also view SLA as a broad umbrella term: “L2 can
refer to any language learned after learning the L1, regardless of
whether it is the second, third, fourth, or fifth language.”
Apart from the difference between “second” and “foreign”, a
distinction can be made between “acquisition” and “learning”. If
“acquisition” is to be preferred over “learning” in my choice of termi-
nology, it is probably less on account of Krashen’s arguments (1981)
and more due to the emphasis in current language teaching
methodology on the need for students of the language to interact and
to focus on meaningful activities rather than on form, as was the case
in the past. Also, enjoyment and fun are more readily associated with
such an approach than conscious and often arduous learning. I could
argue along the lines of cognitive linguistics and situate “learning” and
“acquisition” at the two poles of a continuum. Like most linguists, I
shall use the terms interchangeably.
If syntax or grammar is the overall structure of the building,
then vocabulary is the bricks that are to be fitted into that structure.
This is an obvious metaphor and it has doubtless occurred to other
researchers. We can say that the appearance of the “building” depends
on the size and the combination of the structure and the bricks: the

1 This term implied lack of contact with the people and culture, which very
much characterized the society and school environment of earlier generations.
Recently an English native speaker mentioned his negative reaction on seeing
incoming passengers at Shanghai airport at the passport control divided into
“Chinese nationals” and “Foreigners”; on the pattern of elsewhere in the
world, he would have expected: “Chinese passport holders” and “non-Chinese
passport holders”. Non-Chinese teachers in Chinese universities are referred
to as “foreign teachers” whereas elsewhere they would more likely be called
“expatriate teachers”. As so often in non-native English speaking countries,
the English being learned is already outdated because of the old language
textbooks being used.

20
more spacious and refined it appears, the greater the level of language
proficiency; the more restricted and crude it appears, the lower that
level. Without considering at this point the way structure and bricks
relate and how they integrate, we can see at a glance that the elements
that form the structure are far more limited in number than the bricks.
No linguist today would seriously contest the fact that, quantitatively,
vocabulary dominates in the language field and that vocabulary
acquisition is the main obstacle to language acquisition. As one of the
most influential psycholinguists in the past fifty years expresses it
(Miller, 1996: 5):

It is not the speech sounds or the rules for generating grammatical sentences
that require the most extensive learning. It is the vocabulary: thousands of
words, each with its own sound, its own spelling, its own meaning, its own
role, its own use, its own history.

The building I have been describing is a modern one; up until less

than three decades ago the bricks were relatively few in number, as it
was believed that this small number would suffice. Now it is known
that well over double that number are needed to have a finished
building. We now know that bricks of different sizes are also needed.
Leaving aside the metaphor, we can say that not just individual words
but multi-word items that are stored as independent units feature pro-
minently in any language. Attention will be given to these in this book.
This book is divided into four sections. The first part looks into
a number of important issues pertaining to SLVA, in the light of the
linguistic, psycholinguistic, socio-cultural and, to a lesser extent,
neurolinguistic points of view. I attempt to clarify certain frequently
occurring lexically related notions which tend to cause confusion, such
as breadth/depth vs. reception/production, possible stages of lexical
development or ways of organizing lexis, and to pull research findings
from different areas together to form an updated synergy for SLVA. I
first focus on what constitutes vocabulary knowledge and what the L2
learners need to learn in order to comprehend and produce the L2
vocabulary items, which I present largely from a linguistic angle. This
is followed by an enquiry into how that lexical knowledge is acquired
and organised, from a psycholinguistic standpoint. Next, I underscore

21
the important role of listening in acquiring the L2 vocabulary items.
Essentially neurolinguistically and psycholinguistically motivated, the
research findings cited regarding this issue are based on a substantial
amount of empirical data. Finally I investigate the potential vocabulary
learning difficulties that the L2 learners may encounter. This last point
is dealt with mainly within a socio-cultural framework.
In the second part, some approaches that highlight vocabulary
and are geared to its acquisition are examined. One of the current
hotly debated issues is whether vocabulary should be learned
implicitly or explicitly, which of the two better promotes its acquisi-
tion and the learning effort needed to sustain it. Recent studies show
that lexis consists of not only individual words but also many
thousands of multi-word lexical items which native English speakers
will draw on frequently in daily communication (e.g. Pawley/Syder
1983; Wray, 2002). The acquisition of these multi-word items is
perceived as an important component of vocabulary learning (Lewis,
1993), even though research into this area is still in its infancy. No
matter how effective some vocabulary learning approaches appear to
be, they may not always turn out to be so. Learner groups are not
homogenous and learners differ in many respects, such as motivation,
learning style, learning strategies, etc. These learner variables are of
vital importance to the success of language acquisition.
The third part investigates computer assisted vocabulary
learning (CAVL), in keeping with the current wide application of
computer technology in the educational area. Although computer
assisted language learning (CALL) programs have generally covered
most areas in L2 learning, a substantial number of them have been
designed specifically for vocabulary learning since the initial develop-
mental stage of CALL. However, very little effort has been made to
construct a comprehensive framework for conceptualising CAVL
systematically, despite the paramount importance of vocabulary
acquisition in acquiring an L2. An attempt is made to construct such a
comprehensive framework for investigating CAVL so that SLVA can
fully benefit from new technology. I first present an historical
overview of CAVL; through this review we shall see clearly what
factors have influenced CAVL development. Then I propose a
framework, on the basis of which currently available CAVL applica-

22
tions can be meaningfully conceptualised. Using this framework, I
have carried out an evaluation of a number of selected CAVL
programs which are research-based or commercially available, the aim
being to see to what extent these programs are geared to vocabulary
learning and what needs to be done to improve CAVL efficiency.
Since Chinese learners make up the largest learner population
(currently estimated at around 300 million) in the world for learning
English as an L2, a systematic investigation into this learner popula-
tion should prove to be a worthwhile endeavour. The last part of this
book describes how L2 (English) vocabulary acquisition is treated in
the language educational system in China and provides an up-to-date
picture of the overall situation, including the language policies
adopted, the traditional, orthodox approach to language learning, and
the recent reforms that have been implemented in Chinese universities.
In addition, some general and specific vocabulary learning difficulties
encountered by Chinese learners are documented and analysed.
Finally, two empirical studies are conducted and reported on how
Chinese learners approach the L2 vocabulary in terms of strategy use:
one from the point of view of the students and the other from the point
of view of their teachers.

23
24
 Part I: Issues Related
to Second Language Vocabulary Acquisition

25
26
Chapter 1. Vocabulary Knowledge and Use

1.1. What is vocabulary knowledge?

1.1.1. Vocabulary knowledge

It is necessary to establish what is meant by the term vocabulary

knowledge. Different proposals have been made as to what is involved
in knowing a word (Richards, 1976; Nation, 1990, 2001; Qian 1999,
2002). Instead of talking of vocabulary knowledge or lexical know-
ledge, some researchers speak of “lexical competence” (e.g. Henriksen,
1999), while others talk of a “vocabulary knowledge framework” (e.g.
Meara, 1996a) or a “vocabulary knowledge scale” (e.g. Wesche/
Paribakht, 1996). An intuitive answer to the question of what vocabu-
lary knowledge entails for many people, particularly native speakers,
would be knowing the meaning of the word and how to use it
appropriately in different contexts. This reply certainly contains a
large amount of truth. For example, Miller (1999) identifies two
essential criteria for word knowledge: meaning and contextual use.
Authors working in SLA, particularly in SLVA, believe that there are
additional aspects of word knowledge that need to be known. It is,
however, very difficult to reach a consensus among them as to what
these aspects are and how those aspects should be included in word
knowledge.
Richards (1976), taking the research findings of the 1960s–
1970s as his basis and in consideration of what current theory could
offer for language classroom teaching, set out eight classic assump-
tions in respect of word knowledge. These assumptions are concerned
with frequency, register, syntax, derivation, association, semantic
values, and polysemy. These eight assumptions served as the main
reference for describing vocabulary knowledge for most subsequent
research. The assumptions are by no means exhaustive; Qian (2002)
27
notes some missing aspects of this framework, such as pronunciation,
spelling, and collocation. Fundamentally, Meara (1996a) criticises
these classic assumptions by claiming that they are intended to inform
classroom teaching rather than to serve as a sound theoretical model
for describing vocabulary knowledge.
Meara (1996a) suggests that it is intrinsically impossible to
specify everything that learners know about the L2 lexicon and
advocates a global approach. He proposes a three-dimensional model
for categorizing a learner’s lexical knowledge: (1) a size dimension, (2)
a lexical structure dimension, and (3) a lexical access dimension. He
writes (1996a: 5):

We would need to be able to specify how big learners’ lexicons are; we would
need to be able to specify how automatically the items in a lexicon could be
accessed; and we would need to find a simple measure of how rich a lexical
structure linked the words in the lexicon.

Chapelle (1998a) proposes a four dimensional framework of vocabu-

lary knowledge which includes: (a) vocabulary size, (b) knowledge of
word characteristics (similar to Richards’ eight assumptions), (c)
lexicon organisation and (d) lexical access. It is similar to Meara’s
description of vocabulary knowledge except that Chapelle includes
various aspects of vocabulary, such as those in Richards (1976), which
Meara tried to avoid. In a similar vein, Qian (2002) evolved a
conceptual framework of vocabulary knowledge which contains four
dimensions: (a) vocabulary size, (b) depth of vocabulary knowledge
(various word characteristics), (c) lexical organisation and (d)
automaticity of receptive-productive knowledge (referring to how
words are accessed for receptive and productive use).
Nation (2001: 27) suggests a three-category framework of what
is involved in knowing a word at both the receptive and productive
level:
 Word form: including the spoken form, the written form, and
the word parts (affixes).
 Word meaning: including connecting form and meaning (of a
word), concepts and referents, and associations.

28
 Word use: including grammatical functions, collocations and
constraints on use (register, frequency, etc.).

In conclusion, when offering the proposals or models for describing

vocabulary knowledge, authors seem to have been driven by three main
considerations:

1. A descriptive knowledge framework is produced to describe

various features of L2 vocabulary based on our knowledge of the
L1 (e.g. Richards, 1976).
2. Vocabulary knowledge is described according to the reception/
production distinction (Nation, 1999, 2001) or by stages
(Wesche/Paribakht, 1996; Waring, 1999). This approach is often
oriented towards the teaching or the testing of vocabulary.
3. Vocabulary knowledge is presented in theoretical models (Meara,
1996a; Henriksen, 1999) which attempt to describe the global
vocabulary learning process and development and to do so
dynamically.

It should be added that authors would sometimes mix up these

considerations when describing vocabulary knowledge.
It is a very demanding, often impossible task to have complete
knowledge of a word: it involves both item knowledge, the individual
word, and system knowledge, the various features of the word, the
relationship with other words in the mental lexicon (Nation, 2001),
along with progressive changes in the two (Henriksen, 1999). This
might be a very long process. Kelly (1985) pointed out that the initial
stage for learning an L2 word is to connect the word form with one of
its meanings and fix the two in the mind, leaving the acquisition of
other aspects (other meanings, usage, associations, etc.) to a later stage.
This fits in with the incremental nature of vocabulary acquisition,
namely, that different aspects of word knowledge will be mastered at
different stages at different rates (Schmitt, 2005). The very beginning
of the incremental process consists in knowing the meaning and form of
a word and connecting the two together.

29
1.1.2. Vocabulary use

One controversial issue in vocabulary acquisition is whether vocabu-

lary use is a part of vocabulary knowledge. The majority of the
researchers cited above seem to agree that vocabulary use is a part of
vocabulary knowledge. Others hold a different opinion, e.g.,
Bialystok/Sharwood Smith (1985) and Henriksen (1999). Bialystok/
Sharwood (1985) draw a distinction between knowledge and control.
In discussing learner’s interlanguage, they developed a theoretical
framework based on the assumption that learners’ linguistic know-
ledge systems and the control of these systems were two separate
components; language output may be subject to varying degrees of
control. They made an analogy with a library. The knowledge is in the
books and in the way they are organised. Gass (1988: 95) gives a good
summary of this view of control: in order to control this knowledge,
the user needs the information to find the books and to get the
information in the most efficient way possible. Although this
knowledge-control dichotomy is for general linguistic competence, it
is naturally assumed that it is applicable to lexical competence.
According to Gass (1988), the distinction between knowledge and
control regarding vocabulary is useful since it crosses the boundaries
of the traditional receptive-productive notions. However, Gass and
Selinker (2001) note that the main problem in the library analogy for
L2 learners’ linguistic knowledge and control is that it is not able to
take into account the dynamic changing nature of the L2 lexicon.
Henriksen (1999: 304) adopts a similar approach, constructing a
three-dimensional framework for lexical competence in which know-
ledge and control are separated. These three dimensions include: (1) a
partial-precise knowledge dimension, (2) a depth of knowledge
dimension, and (3) a receptive-productive dimension. She (1999: 314)
made the following statement regarding the distinction between
vocabulary knowledge and control process:

[…], it might be argued that dimensions 1 and 2, which are related to

acquiring word meaning (i.e., labelling and packaging) and developing an
understanding of sense relations (network building) are basically knowledge
continua, in which levels of declarative word knowledge may be tapped or

30
 operationalized as levels of word understanding or comprehension. Dimension
3 is essentially a control continuum that describes levels of access or use
ability, which may be operationalized through different types of receptive and
productive tasks.

Laufer/Paribakht drew the logical conclusion from Henriksen’s separation

of vocabulary use from vocabulary knowledge: “According to Henriksen,
the fact that learners cannot use a word correctly nor access it freely for
production does not mean they do not ‘know’ the word; it only means that
they have not yet achieved adequate control over word access” (1998:
367).
Although in theory the separation of knowledge and use sounds
plausible, insufficient empirical work has so far been carried out to
investigate its feasibility.

1.1.3. Declarative vs. procedural knowledge

Fundamentally, vocabulary knowledge stored in our memory, like

many other human cognitive representations, can take the form of two
types of knowledge: declarative and procedural knowledge. These
two terms were first introduced by Anderson (1983) as the basis for
constructing his well-known “Adaptive Control of Thought” (ACT)
theory to model human cognitive processes. Declarative and proce-
dural knowledge make up our long-term memory. The difference
between the two is the question of what or how: declarative knowl-
edge comprises what we know about the world and all factual
knowledge; procedural knowledge refers to skills and how to perform
various activities.
Traditionally, in psychology declarative knowledge/memory can
be further divided into two sub-categories: episodic memory and
semantic memory. Episodic memory is associated with a particular
time and location where the events took place, e.g. what we did to
celebrate our last birthday. In contrast, semantic memory contains our
general knowledge of the world, which can be dissociated from the
original context where it was learned. For example, most of us know
that Paris is the capital of France but we would not remember where

31
we first learned it. The two systems seem to work independently and
to have a different neural basis (Roediger, 2005: 5). Examples of
procedural knowledge or memory include typing, driving, swimming,
riding bicycles, playing tennis, etc. Procedural knowledge is not only
limited to physical skills but also includes abstract skills such as
reading and writing. We acquire procedural knowledge best by practis-
ing according to the power law of practice. We learn to drive
essentially by driving; we learn to speak a language by using the
language frequently. One important characteristic of procedural
knowledge is that it can be retrieved rapidly, automatically without
any conscious effort on the part of our memory. In this sense,
procedural knowledge is very similar to implicit knowledge; on the
other hand, declarative knowledge is retrieved by conscious and
deliberate effort, in much the same way as explicit knowledge2. It
should be noted that we draw on the two types of knowledge
simultaneously in carrying out most activities instead of solely relying
on one of them.
According to the ACT theory of Anderson (1983), we start
learning everything in declarative forms, termed chunks, and some of
the knowledge will eventually become procedural forms, which we
refer to as production. For most of us, speaking and listening in our L1
is procedural knowledge as most of the time we use our language
automatically and spontaneously, although occasionally we need to
consciously search for some words or expressions. For educated
people, automatic, procedural language use can be extended to reading
and writing. In this sense, most of our L1 vocabulary knowledge has
become proceduralised and this happens after using the vocabulary for
years. Theoretically, L2 vocabulary knowledge development goes
through the same process whereby we start with declarative
knowledge and this subsequently becomes procedural knowledge,
although not every L2 learner can acquire such procedural knowledge

2 From the literature on cognitive psychology, it is not difficult to discover that

declarative/procedural knowledge means more or less the same thing as
explicit/implicit knowledge in respect of the similar ways in which they are
retrieved, although rarely do authors equate these two pairs of terms directly.
(See Chapter 5 for more information about implicit/explicit knowledge.)

32
and not every L2 word can be fully proceduralised. It is very likely
that part of the knowledge of a given L2 word is proceduralised and
part of it is still declarative, depending on how well we know the word
or how well we can use it. Even for L1 speakers, some technical terms
are only known in declarative form or are not known at all.
For most L1 speakers, the language is represented as procedural
knowledge as they have lost the awareness of the language structure or
specifications of lexical items which they use automatically every day.
For most L2 (particularly adult) learners, the situation is rather different.
The L2 is first learned as declarative form and, due to insufficient L2
input or output (practice) and the interference of the existing semantic
system (L1), the lexical knowledge may remain declarative for a long
time. Yet this is not to deny that L2 learners can often achieve a certain
fluency or automaticity in language use, including vocabulary use, which
sometimes would be or be close to being procedural knowledge.

1.2. Breadth and depth of vocabulary knowledge

1.2.1. A clear picture vs. a blurred picture

Anderson/Freebody (1981: 92−93) first made a distinction between

two dimensions of vocabulary knowledge:

The first may be called “breadth” of knowledge, by which we mean the

number of words for which the person knows at least some of the significant
aspects of meaning. […] [There] is a second dimension of vocabulary
knowledge, namely the quality or “depth” of understanding.

According to this description, I draw a simple picture, shown in

Figure 1.

33
Figure 1. A simplified mental representation of breadth and depth of vocabulary
knowledge.

In this picture, the surface of the bottom board is the size, or breadth,
of one’s vocabulary, while the individual blocks of different height can
be regarded as the quality, or depth, of understanding of each
individual word, or lexical item. Obviously, the larger the surface of
the bottom, the larger the size of one’s vocabulary; the higher each
block, the deeper the knowledge one has for that individual item.
Nevertheless, the picture in a real situation will be far more
complex than this simplified representation. First of all, whether the
bottom board has clear, fixed lines is highly questionable. Second, it is
not clear when we can say that one knows a word in breadth. For
example, one might be able to recognise a word as “déjà vu” on
encountering it while failing to retrieve the correct meaning in the
context. In this case can we say that we know this word (in breadth)?
If one knows a word only in breadth or, in other words, superficially,
there is a great chance that this word will be forgotten, dropping off
the board. New words might come forward to replace it. As to the
degree of depth each word has reached, this is even more controversial
since there are no commonly accepted norms. Thirdly, the individual
blocks are not isolated; instead, they should be viewed as being
interconnected by links representing their lexical organisation. So
eventually the picture we get is the following: the lines of the bottom
board are blurred, moving and changing all the time; there is no
precise summit for the height of each individual block and the higher
the top, the more blurred it is; furthermore, there are numerous links

34
between the blocks which are in a dynamic state. We could draw some
conclusions from this picture: the borders in both dimensions, vertical
and horizontal, are fuzzy; for the bottom board, the more central a
word is, the more chances there are of it acquiring stability; regarding
the blocks, each of which represents a word, the lower they are, the
more they will be influenced or undergo change (grow or diminish),
while as they grow in height, the more stable and numerous will be
their connections with other words.

1.2.2. Breadth of vocabulary knowledge

Breadth of vocabulary knowledge means the size of a language

learner’s L2 mental lexicon or, according to Qian (2002), the number
of words whose meaning the learner only knows to a superficial
degree. This implies that one essential feature in defining breadth of
vocabulary knowledge is that one only needs to know (some of) the
meaning(s) of the lexical item without necessarily knowing other
lexical features. This definition is somewhat problematic. What are the
criteria for determining the meaning of words of which “one has at
least some superficial knowledge” (Qian, 2002: 515) or “the signifi-
cant aspects of meaning” (Anderson/Freebody, 1983: 92−93) that
determine whether one has breadth of knowledge of a word?
In its pedagogical application, breadth of vocabulary knowledge
is naturally extended to refer to how many words an L2 learner needs
to know or the vocabulary size they need to reach. For a long time, the
amount of vocabulary that enables a learner to cope with relevant ease
in normal situations whether for speaking or reading purposes was
seriously underestimated by producers of language learning materials.
This notion can be traced back to more than 70 years ago when Ogden
(1937) demonstrated that one can express millions of ideas using a
mere 850 words. Compared with the normal size of a native speaker’s
vocabulary, around 20,000 word families (Nation, 2001), we might
consider this to be something of a miracle and wonder if those 850
words can really fulfil the learner’s productive requirements.
Researchers have held a different view. In the 1960s, Lado (1964)
carried out studies in which he concluded that the foreign language
35
learner needs a productive vocabulary of 3,000 words and a receptive
vocabulary of 7,000. Carroll et al. (1971) arrived at a similar figure,
giving 7,000 as the number of words which will appear at least five
times in every million words of randomly chosen English. Keller (1978)
also talks vaguely about the 6,000 to 10,000 words that are required for
everyday communication in Russian or German. Both Twaddell (1973)
and Keller (1978) mention the sudden shock experienced by inter-
mediate learners when they cannot cope with the great amount of
vocabulary they meet. Studies carried out by Schouten-Van Parreren/
Van Parreren (1979) and Ostyn et al. (1985) have shown that 5,000
words is a minimal figure to be able to understand 95% of the lexical
items of a news text. Investigations carried out by Sciarone (1979) and
Laufer (1997) further corroborated this figure, which will ensure a full
understanding of such a text. Sometimes these 5000 words will be
replaced by 3000 word families. With 8,000 words, learners can read
for pleasure (Hirsh/ Nation, 1992). A few authors (e.g. Judd, 1978) even
claimed that there is a need for early vocabulary expansion, not waiting
until the intermediate stage. This seems to be the tendency nowadays.
From the above studies, it is quite evident that when authors talk
about how many words an L2 learner needs they mean the vocabulary
required for an adequate comprehension of text reading. It would now
seem to be a little easier to answer the question raised earlier: if an L2
learner correctly understands the meaning of a word in a given context,
then s/he can be said to have breadth of knowledge of the word.

1.2.3. Depth of vocabulary knowledge

While breadth relates to the quantitative aspect of vocabulary

knowledge, depth is concerned with its qualitative aspect. The criteria
for determining the quality of vocabulary knowledge tend to be a list
of numerous features of word knowledge similar to the work
discussed earlier (e.g. Nation, 2001; Chapelle, 1998a; Richards, 1976).
In defining the depth dimension of vocabulary knowledge, Qian (1999,
2002) includes various word characteristics, such as phonemic
(pronunciation), graphemic (spelling), morphemic, syntactic, semantic,

36
collocational and phraseological (register, frequency) properties as
opposed to the mere meaning aspect of breadth of knowledge.
Henriksen’s (1999: 305) depth dimension stresses the “rich
meaning representation” of words. In addition to knowledge of a
word’s meaning, depth also involves the different “intensional” or
“sense relations”3 with other words, particularly, “paradigmatic rela-
tions (antonymy, synonymy, hyponymy, gradation) and syntagmatic
relations (collocational restrictions)”. She used the expression
“semantization process”4 (1999: 308) to refer to the ongoing, dynamic
nature of the semantic development of words, claiming that
development in depth of knowledge consists in network building, i.e.
creating intensional links between words, a notion similar to Meara’s
(1996a) lexical organisation. Vermeer (2001) also holds the view that
words are represented as nodes in a network and they are inter-
connected in different ways; the nodes can be linked with each other by
meaning, by sound, by affixes, by concept, or by register. In Vermeer’s
words, the depth of vocabulary knowledge depends on the density of
the network surrounding the word (2001: 218).
In spite of this, it seems to be difficult to answer the question “At
what point can we claim to have depth of knowledge of a word?” As we
saw in Figure 1, there is no real fixed boundary for the breadth or the
depth of individual words. We could answer the question by saying that
the degree to which a word is known depends on how well the learner
knows the various characteristics of the word and its links with other
words.

1.2.4. Breadth vs. depth

Although breadth and depth represent quite distinct dimensions of

vocabulary knowledge, a few recent studies suggest that the two are in
fact closely related to each other. Qian (1999, 2002) argued that both
vocabulary breadth and depth dimensions are important for reading

3 The sense relations between words are dealt with in great depth by Miller
(1996, 1999) and Miller/Fellbaum (1991).
4 The term semantization was first introduced by Beheydt (1987).
37
comprehension. In these two studies, he showed that the Pearson
correlations between the depth vocabulary knowledge test and the
reading comprehension test were 0.82 in 1999 and 0.77 in 2002; both
correlations were slightly higher than those between the vocabulary
size test and the reading comprehension test (0.78 in 1999 and 0.74 in
2002). In addition, the depth vocabulary knowledge tests were reported
to have good correlations with the vocabulary size tests (0.82 in 1999
and 0.7 in 2002). These results seem to indicate that the breadth and
depth of vocabulary knowledge are closely associated with each other
and they are at least equally important to reading comprehension.
Vermeer (2001: 217) argues forcibly that there is no “conceptual
distinction” between breadth and depth of word knowledge and that the
two are strongly related to each other; the greater one’s vocabulary, the
deeper one’s vocabulary knowledge or vice versa. He states (2001: 222):

[…] a deeper knowledge of words is the consequence of knowing more words,

or that, conversely, the more words someone knows, the finer the networks
and the deeper the word knowledge. Hence, the ability to demarcate the
precise meaning and usage of a word is based on the knowledge of the other
words that are needed to categorize, classify, and delimit that word more
precisely. […] In this way, depth is connected with breadth of vocabulary.

Vermeer (2001: 231) claims that breadth and depth of vocabulary

should be regarded as two dimensions of the same phenomenon
instead of opposites; the growth of two dimensions should be subject
to the same condition: frequency of input5. He (2001: 231) concludes
from his studies of the acquisition of Dutch by L1 and L2 children
(5-year olds) that a breadth test can be as good a measure of language
proficiency in vocabulary as a depth test on condition that the breadth
test includes words from various domains and frequency levels.
Another study of Dutch primary school children by Schoonen/

5 According to N. Ellis (2002), frequency of input is the most important factor

that accounts for language acquisition in all aspects of learning: phonology,
reading, spelling, vocabulary, morphology, multi-word items, grammar, syn-
tax, etc. The process involves both implicit and explicit learning, depending
on the type of language knowledge to be acquired.

38
Verhallen (1998, cited in Read 2004) also shows that the breadth test
and the depth test are strongly correlated.
Additional evidence that supports this view is from the study of
Nurweni and Read (1999) in which breadth and depth of knowledge are
more closely related with each other for advanced learners than for low
level learners. Like Vermeer, Read attributes the close relationship
between breadth and depth to the network building of the mental
lexicon: “This parallel development of vocabulary size and depth is
particularly pertinent if we adopt a network building perspective on
depth, in that vocabulary growth also entails the building of more
extensive linkages between items in the mental lexicon” (2004: 221).

1.3. Receptive and productive vocabulary

1.3.1. Introduction

Following the distinction made by Henriksen (1999), receptive and

productive vocabularies involve the control process or use of
vocabulary knowledge. Although she sees vocabulary use as separate
from vocabulary knowledge, I treat vocabulary use, whether receptive
or productive, as part of vocabulary knowledge, as the ability to use
that knowledge. According to Melka (1997: 84), the two notions of
reception and production in vocabulary acquisition are never clearly
or satisfactorily defined and this might explain why many other terms
such as passive/active vocabulary, comprehension/production, under-
standing/speaking are used to replace the two notions. However, a
definition that is particularly suitable for characterising learners’
vocabulary use for pedagogical purposes can be made by relating the
terms to the four basic language skills: reading, listening, speaking
and writing (Nation, 1990). According to Nation (1990), receptive
vocabulary is the vocabulary that one can recognise when it is
encountered either in reading or listening while productive vocabulary
is the vocabulary that one is able to use correctly either in speaking or

39
writing; productive knowledge of a word includes receptive knowledge
and extends it. This definition implies a number of assumptions:

1. The receptive vocabulary is larger than the productive vocabulary.

2. Reception precedes production.
3. Production is more difficult than reception.

Indeed, there is considerable evidence in support of these assumptions,

e.g. Melka (1997) for (1) and (2); Laufer (1998) and Laufer/Paribakht
(1998) for (1); Mondria and Wiersma (2004) for (3). It should be
noted that the term receptive or productive vocabulary not only refers
to the status of individual words but also to the mental lexicon as a
whole – there is a receptive lexicon (R) and a productive lexicon (P) 6.
The distinction between reception/production and breadth/depth
needs to be clarified before going further. There is considerable
similarity between these two pairs in defining vocabulary knowledge
and use. Knowing a vocabulary item receptively is likely to require
only shallow knowledge as covered by breadth; to know a word
productively is likely to involve deep knowledge which may include
various components covered by depth. While it may be true that
receptive vocabulary more or less corresponds to breadth of vocabulary,
being able to produce a word does not necessarily involve very deep
word knowledge – a word can be used productively in a narrow context
without knowledge of other meanings or inflected forms. Fundamen-
tally, breadth/depth is concerned with describing vocabulary
knowledge and reception/production deals with how the knowledge is
accessed and used. In other words, the former pair belongs to the
paradigm of declarative knowledge and the latter pair to the procedural
knowledge paradigm which, in my view, has a more practical
pedagogical application in assessing vocabulary skills or use. This is
not to say that when a learner can recognise or produce a word correctly
the word is fully represented as procedural knowledge in memory. If
this is done automatically without conscious effort to retrieve the word,
at least some aspects of the word, the phonological and formal informa-

6 The abbreviations R and P are used to refer to the lexicon as well as to indi-
vidual items.

40
tion, and the meaning, have reached the status of procedural knowledge.
If conscious effort is needed to retrieve the word, it is still represented
as declarative knowledge. Knowledge can be either retrieved implicitly
as procedural knowledge or explicitly as declarative knowledge
independently, or, more often, by using the two systems simultaneously.
To understand the interaction of breadth/depth and reception/pro-
duction is crucial in understanding the language learning process.

1.3.2. Receptive and productive vocabulary:

A continuum or a dichotomy?

Should R and P be viewed as separate entities in the traditional way or

as a continuum where location is determined by degrees of familiarity
or knowledge, as proposed by Melka (1982, 1997)? As the two terms
R and P suggest, it is convenient to conceptualise the two as distinct
from each other. However, the crucial evidence to support this will be
to find a kind of boundary that can tell us when a lexical item is not R
but P; such a clear-cut boundary is never found in the literature
(Melka, 1997). Knowing a word is not an all-or-nothing process: some
aspects may be productive while other aspects are still at the receptive
level (Melka, 1997: 87). It would seem practical to use a bi-polar
dichotomy to conceptualise R and P with degrees of familiarity or
knowledge in between. Melka (1997: 89) proposes four possible
stages which can be distinctly identified along the continuum:

1. Imitation: perceptual motor skill regardless of comprehension

(Fraser et al., 1963, cited in Melka, 1997)
2. Comprehension (of the message)
3. Reproduction with assimilation: involving active reconstitution
of the message
4. Production

The basic assumption of the continuum of R and P is: (1) R may

precede P, (2) the gap between R and P is not wide and is subject to
changes and, most importantly, (3) R and P are based on the same
underlying system (Melka, 1997: 92).
41
 Meara (1990) proposes a different framework for conceptualis-
ing R and P, making two separate, independent systems and uses a
hypothetical association network to describe R and P as follows:

Figure 2. A hypothetical association network (Meara, 1990: 152).

Every node is a lexical item in the overall network of the lexicon. All
the nodes are connected to the network by two arcs, except node H. If
you start from H, you can access all the other words in the network,
but you cannot reach H if you start somewhere else. Obviously, node
H is a receptive lexical item which belongs to the whole lexicon, but it
can only be reached by appropriate external stimulation. On the basis
of this, Meara (1990: 153) suggests that the distinction between R and
P is not a gradual one but a marked one; P clearly operates on a kind
of continuum but not R which is qualitatively different from P. This
view would seem to support the dichotomous view of R and P; how-
ever, this is only a hypothetical proposal and has never been subjected
to empirical investigation.

1.3.3. Relationship between receptive and productive vocabulary

in empirical studies

Laufer (1998) focuses on the development of R and P of English by

two groups of secondary school students in the 10 th and 11th grades.

42
There are actually three types of vocabulary knowledge that are under
investigation: receptive knowledge and two types of productive
knowledge: controlled and free. This construct clearly reflects the
continuum of R and P. Receptive knowledge is defined as under-
standing the core meaning of a word, assessed by the Vocabulary
Levels Test (Nation, 1990); controlled productive vocabulary is
measured by a type of cued recall test such as: They will restore the
house to its orig________ state (Laufer, 1998: 260); free productive
knowledge is measured by writing an essay of 200 or 300 words based
on the Lexical Frequency Profile (Laufer/Nation, 1995). It should be
noted that this is a cross-sectional study, as opposed to a longitudinal
study, focusing on the difference between the three types of
receptive/productive knowledge of 10th and 11th grade students. The
main findings include: (1) an additional year of study leads to a much
larger R, (2) the gap between R and controlled productive vocabulary
(controlled P) widens after one year’s study, but (3) the free
productive vocabulary (free P) remains the same despite an additional
year’s study. She provides two possible explanations for this static free
P: (a) the gains are not sufficient to account for free P and (b) the
classroom instruction may not have obliged the learners to produce the
words productively that they had been taught.
The study by Laufer and Paribakht (1998) looked at the
relationship between R and P in two different learning contexts: EFL
(learning English in the classroom in an L1 environment) and ESL
(learning English in an English-speaking country). The procedure and
instruments were very similar to the first study: the same three types of
vocabulary knowledge were tested with the same three types of
measures. The main findings were: (1) in both learning contexts, R,
controlled P and free P develop at different rates, with R the fastest,
followed by controlled P and with free P the slowest; (2) the EFL
learners have significantly better knowledge in both controlled P and
free P than ESL learners while ESL learners have a richer R; (3) two
years of residing in an L2 environment leads to significant gains in
controlled P but not free P; (4) for the ESL learners, knowledge of a
related language (French) has a positive effect on controlled P when R
is at the intermediate level but the effect decreases as R increases;
however, no effect is found for free P regardless of R levels.
43
 In conclusion, the two studies confirm that R is always larger
than P and that the gap between the two changes in the course of
language study. It would seem that R could relatively easily enter
controlled P but that it is not easy for controlled P to enter free P,
suggesting a big gap between controlled P and free P as this implies a
jump from cued recall of vocabulary to complete free use of vocabulary.
The two authors recommended that future studies investigate how to
turn controlled P into free P.

1.4. Testing vocabulary knowledge and use

1.4.1. Introduction

Learners’ vocabulary knowledge and use need to be assessed in order

to decide how large their vocabulary is, how well it is known and
whether lexical items can be recognised or produced. We are thus
referring to breadth, depth, reception and production. Vocabulary
measures tend to be applied either on the breadth/depth dimension or
on the receptive/productive continuum, though sometimes it is diffi-
cult to make such a distinction as some measures, e.g. the Vocabulary
Knowledge Scale (VKS) developed by Wesche and Paribakht (1996)
and the Computer Adaptive Test of Size and Strength (CATSS) by
Laufer et al. (2004), tend to mix the two. Unavoidably, there is some
overlapping of what is tested between the measures based on breadth/
depth and reception/production. The Lexical Frequency Profile
developed by Laufer and Nation (1995) can be used as a measure of
free recall of lexical items in writing. It is obviously a type of free
production test, but to produce a word also requires the learner to
show various degrees of lexical knowledge such as meaning, syntactic
category, grammatical functions, collocations, and perhaps register
constraints. In this sense, it is also a test measuring depth of vocabu-
lary knowledge. We shall now look at the four types of measures from
the two perspectives of breadth/depth and reception/production.

44
1.4.2. Breadth and depth tests

1.4.2.1. Breadth tests

If we stick to the definition of breadth given by Anderson and
Freebody (1981), what should be tested in breadth tests or vocabulary
size tests is a superficial indication or partial knowledge of meaning
(Read, 2004; Schmitt, 2000), e.g. word recognition, supplying a
definition, etc. To make the results representative, a large sample is
needed, usually from frequency lists or dictionaries; meanwhile some
selection needs to be made to keep the number of items manageable.
Common test formats include multiple choice, which tests the
recognition of L2 synonyms or L1 equivalents; a list which requires a
Yes/No response to a given item (or sometimes a pseudo-word) to
check whether the word meaning is known or not (e.g. Meara/Buxton,
1987; Eyckmans et al., 2002); dictation tests in which a word is read
out and the learner has to identify the corresponding drawing (Dunn/
Dunn, 1981); matching words with definitions or translations, e.g.
Nation’s (1990) Vocabulary Levels Test, etc.

1.4.2.2. Depth tests

What is measured in depth tests is the quality of vocabulary know-
ledge, as opposed to the superficial or partial knowledge measured by
breadth tests. Read (2004: 221) identifies three distinct approaches to
depth tests that focus on: (1) precision of meaning, (2) comprehensive
word knowledge, and (3) network knowledge. The first approach is
based on the assumption that the meaning of a word can be known
ranging from a minimum to a maximum degree. Test formats include
multiple-choice questions that require more than an understanding of
word meanings, self-report knowledge scales such as Vocabulary
Knowledge Scale, and evaluating the quality of the definitions given
by learners. The second approach tries to tap various components of
word knowledge (see Chapter 1). Tests taken from this approach (e.g.
Schmitt, 1998) usually take a long time to complete and the number of
items tested is limited. Read (2004) questions the value of such tests
and proposes alternatives such as the scaling system developed by
Drum (Drum/Konopak, 1987) or Nation’s framework of word knowl-
45
edge which can be used as a checklist to decide what word knowledge
should be focused on for testing. The third approach is to
conceptualise depth of vocabulary knowledge in terms of building a
network of lexical knowledge. The most popular measure for this
approach is perhaps the word association test developed by Read
(1993, 1998). In this type of test, learners are required to choose three
word items (out of six or eight) for a given word, which should
demonstrate three basic relationships between the target word and the
associated words: paradigmatic (superordinates, synonyms), syntag-
matic (collocates) and analytic (the prototypical meaning of the target
word).
Read states clearly that it is impossible to find a measure that
can tap all aspects of “depth” of vocabulary knowledge: “Whether we
focus on individual lexical items or the mental lexicon as a whole, we
are setting out to describe something that is inherently ill-defined,
multidimensional, variable and thus resistant to neat classification”
(2004: 224). He thus proposes that the dimension of word knowledge to
be measured should be carefully designed rather than covered by the
general term depth. He also points out that what is typically measured
in depth tests is declarative knowledge (conscious and can be described)
and calls for measures that tap procedural knowledge (vocabulary use)
as complementary tests.

1.4.2.3. Breadth test vs. depth test

Generally speaking, breadth of vocabulary is more frequently
measured than depth of vocabulary. This phenomenon is criticised by
some authors (e.g. Schmitt/Meara, 1997; Wesche/Paribakht, 1996) on
the grounds that measuring breadth of vocabulary or vocabulary size
is of limited value as it does not take into account the fact that words
can be known to various degrees. On the other hand, Read (2000,
cited in Laufer et al. 2004: 209) believes that although vocabulary size
tests may seem to be superficial they could give a more accurate
picture of learners’ overall vocabulary than depth tests that only tap a
small number of words. Laufer et al. hold the same view: “Knowing
many words (units of meaning) is more important than knowing few
words in depth. Hence, a good vocabulary test should test how many

46
words are known, or try to provide a picture of the learner’s overall
vocabulary” (2004: 209). Qian (2002) is perhaps right in suggesting
that it would probably be more valid, accurate and efficient to use a
combination measure that taps both depth and breadth of vocabulary
knowledge rather than a single measure of either breadth or depth.

1.4.3. Receptive and productive tests

The most popular type of test for R seems to be multiple-choice

questions which require the learner to recognise an L2 word by
choosing a definition in the L2 or an equivalent in the L1. The
commonly used measure for P is recall tests of different formats. The
simplest is to ask for a translation of an L1 word; other more
complicated measures include cued recall such as the controlled
productive vocabulary test presented earlier. A special type of test for
P is to ask learners to write an essay and then measure the quality of
the vocabulary used. The Lexical Frequency Profile (LFP) (Laufer/
Nation, 1995) is such a test; it enables us to assess lexical richness as
well as the relationship between free productive vocabulary and
vocabulary size in a piece of writing. The LFP reveals the percentage
of words a learner uses at different frequency levels in writing. The
LFP uses the 1000 word frequency list, the 2000 word frequency list,
the University Word List (UWL), and the less frequent words that are
not on the two lists. Thus, a learner may produce a piece of writing
totalling 250 word families of which 100 belong to the 1000-word
level families, 100 belong to the 2000-word level families, 30 belong
to the UWL, while the remaining 20 are not on any list. As a result,
the LFP of the learner would be 40% - 40% - 16% - 8%.

1.4.4. Tests combining the breadth/depth dimension

and the reception/production dichotomy

Some measures do not emerge clearly as following either the breadth/

depth dimension or the reception/production distinction and seem
rather a combination of the two. We shall have a look at two measures
47
of this type: Vocabulary Knowledge Scale (VKS) and Computer
Adaptive Test of Size and Strength (CATSS).

1.4.4.1. Vocabulary Knowledge Scale (VKS)

The VKS developed by Wesche and Paribakht (1996) combines a
self-report rating and performance on lexical items that measure initial
vocabulary development over short periods in educational settings.
There are five scale ratings that range from nil knowledge, recognition,
some faint idea of the meaning, to the ability to use the item correctly
both grammatically and semantically in a sentence. The two authors
formulate them as follows (1996: 30):

1. I don’t remember having seen this word before.

2. I have seen this word before, but I don’t know what it means.
3. I have seen this word before, and I think it means ______ (synonym or
translation).
4. I know this word. It means _______ (synonym or translation).
5. I can use this word in a sentence: ________________________________. (If
you do this section, please do section 4)

The VKS is frequently reported as a type of depth measure as it shows

a progressive degree of word knowledge, particularly the various
degrees of understanding of the meaning of an item represented by the
first four categories. However, the last category that tests the highest
degree of knowledge requires not only a deep knowledge of the item
but also the ability to use it productively. In addition, the third and
fourth categories require learners to recognise the item by supplying
synonyms or translations. Thus the scale also includes elements of
receptive and productive knowledge. The two authors suggest that
VKS be viewed as a practical instrument for capturing initial stages or
levels in vocabulary learning instead of a sophisticated measure for
overall vocabulary knowledge. Nevertheless, Read (1997: 317)
questions the validity of the scale; the main problem is whether the
five levels of scale can adequately represent the five key stages in the
acquisition of a lexical item or not. Another problem is whether
supplying synonyms or composing sentences are the most appropriate
ways to tap learners’ word knowledge. Despite these criticisms, Read

48
points out that the VKS seems to have served its initial purpose by
providing a workable instrument for testing a reasonable number of
words (1997: 317) and to be sensitive to capturing the increase in
vocabulary knowledge (2000: 135).

1.4.4.2. CATSS (Computer Adaptive Test of Size and Strength)

The CATSS, developed by Laufer et al. (2004: 209), aims to provide a
construct for tapping both vocabulary size and strength (quality) of
word meaning. The basic assumptions of the test are: (1) the central
point of word knowledge is to link the word form and the word
meaning; (2) knowledge of meaning can have different degrees of
strength; (3) a size test can present an overall picture of a learner’s
vocabulary better than a depth test. The strength of word meaning is
constructed based on two distinctions: (a) supplying the form for a
given concept vs. supplying the meaning for a given form, and (b)
recall vs. recognition (of form or meaning). To retrieve the word form
when the meaning is given is more difficult than to retrieve the
meaning when the form is given, and hence the former is called
“active” knowledge and the latter “passive knowledge”. The inter-
action between active/passive knowledge and recall/recognition leads
to four modalities (types) of vocabulary knowledge in descending
order of strength: active recall, passive recall, active recognition, and
passive recognition. The following four examples illustrate the four
degrees of strength for the item melt (Laufer et al., 2004: 206−207):

Active recall: Turn into water m_________

Passive recall: When something melts it turns into _____________
Active recognition: Turn into water a. elect b. blame c. melt
d. threaten7
Passive recognition: Melt a. choose b. accuse c. make threats
d. turn into water8

The test is obviously a construct for measuring one type of depth of

word knowledge, the meaning, and also a useful construct for measur-
ing production and reception (recognition) of various levels.

7 The three distracters are taken from the same frequency level as the target item.
8 Distracters are taken from a higher frequency level than the target item.
49
 In operation, 30 lexical items are randomly selected from five
word-frequency levels: the 2000 most frequent words, the third 1000,
the fifth 1000, the tenth 1000 and the Academic Vocabulary List (AWL)
(Nation, 1990, 2001). Laufer et al. (2004) claimed that 30 items can
represent each frequency level of 1000 words. If a learner gets an item
right in one strength modality, it will not appear in the lower strength
modalities. If the learner gets the word wrong, the computer will
remember it and present it again in a lower strength modality after all
the other words at the same frequency level have been tested. The pilot
study confirms the hypothesized strength hierarchy but only distin-
guishes three instead of four different modalities: active recall is
demonstrated to be more difficult than passive recall, followed by
recognition (active or passive). In other words, active and passive
recognition are indistinguishable from each other in terms of degree of
difficulty or strength. Despite this, the three other degrees of strength of
knowledge of word meaning may still be good enough to capture
learners’ vocabulary knowledge stages at a particular time.

1.5. Conclusion

It is generally agreed that knowing a word entails knowing its various

aspects and features and how to use it, although a few linguists tend to
separate the two. The fact that vocabulary knowledge, like other
human knowledge, is stored in declarative and procedural forms lends
support to the view that vocabulary use, which essentially involves the
procedural channel, is actually part of vocabulary knowledge.
Linguists tend to describe vocabulary knowledge either
according to the “breadth/depth” dimension or the “reception/produc-
tion” dimension. Breadth and depth essentially concern declarative
knowledge whereas reception and production are fundamentally con-
cerned with procedural knowledge. The learner is consciously
increasing the number of words and the features needed in order
eventually to be able to use them automatically both receptively and
productively. In addition, the evidence points to the two dimensions

50
being related, although we saw that further research is needed to
discover their degree of proximity. This would be useful in predicting
the depth of learners’ vocabulary knowledge at a productive level or the
breadth of their vocabulary at the receptive level, and vice versa.
We looked at the different types of tests used for each dimension
and saw that measures used for testing reception and production are
similar to those used for breadth and depth of vocabulary knowledge.
This is particularly true in respect of reception and breadth as the tests
for these two only need to measure superficial knowledge of the item
at the recognition level. This clearly reflects the interaction between
declarative knowledge and procedural knowledge.

51
52
Chapter 2. The L2 Lexicon:
Development and Organisation

2.1. The development of the L2 mental lexicon

2.1.1. A separate or unified system?

One controversial issue in psycholinguistics concerning the mental

lexicon is whether there is a separate or unified system for each
language acquired, e.g. the L1 and the L2. In a discussion of mental
representation of languages, Bialystok (1995) proposes that lexicons
for different languages are distinct from each other although they are
attached to a unified language system. Singleton (1999: 167) points
out that the proposal for separate systems is based on (1) the modu-
larity hypothesis9 and (2) the studies of formal differences between
languages. The proponents of the modularity hypothesis believe that
the L1 mental lexicon is essentially intramodular while the L2 lexicon
developed after childhood is largely extramodular. Results from the
formal difference studies show that morphological structure analysis is
performed within a particular language rather than between different
languages. Thus the L1 lexicon and the L2 lexicon should function
independently of each other as separate systems. On the other hand,
from research into bilingualism and cross-linguistic influence
evidence exists in favour of a unified system (Singleton, 1999).
Yet it may seem that either a completely separate or a completely
integrated lexicon is too simplistic a solution since neither can provide
an adequate explanation for the counter examples of each proposal. De
Bot (1992: 10) points out: “Now the question is no longer whether the

9 The modularity hypothesis assumes that each human cognitive activity is

controlled by a given location of the brain, as is language which is governed
by the language module.
53
systems are separated or not, but under what conditions and for which
parts of the lexicon they are separated.” Paradis (1987, cited in De Bot,
1992: 10−11) summarized four proposals as to how the bilingual
lexicon might be stored:

1. A single storage.
2. Separate storages.
3. A single storage for common information such as cognates and
separate storages for language-specific information.
4. A single storage for all languages in which links between lexical
entries are strengthened by use.

The last two go beyond the simplistic debate of an either/or situation.

The fourth proposal implies that lexical entries from the same language
tend to be stored together as subsets since the links between those
entries will be greatly strengthened by frequent use; in addition, the
links between entries of different languages can also be strengthened
given that bilinguals constantly code-switch.
Likewise, by conducting a comprehensive literature review,
Singleton (1999: 190) concludes that neither complete separation nor
integration of the L1 and L2 lexicons is possible and that the truth lies
in between:

It appears from the evidence reviewed that L1 and L2 are separately stored,
but that the two systems are in communication with each other – whether via
direct connections between individual L1 and L2 lexical nodes, or via a
common conceptual store (or both).

2.1.2. The development of a lexical entry

There are several metaphors to describe the mental lexicon, such as a

library, a computer, a dictionary, etc. One common metaphor is a
network composed of numerous nodes and links: each node represents
a lexical item with the necessary lexical information attached and is
associated with other nodes by links. We can say that each node is a
lexical entry which contains a set of information to distinguish it from

54
or associate it with other nodes. To develop an L2 lexicon means
inserting numerous L2 lexical entries and linking them together. I
shall first focus on how a lexical entry enters the network and then see
how the network is arranged.
The most important way to build a lexical entry is to make a
connection between a word meaning and its referent and map the
meaning with its corresponding morphophonological form (or the word
form). In other words, to acquire the meaning is the first step in building
a lexical entry. How do we acquire the word meaning? According to
Aitchison (1994: 170−180), a child acquires an L1 word meaning by
performing three tasks: (1) labelling, (2) packaging, and (3) network
building. In labelling tasks, children have to associate a word form (a
sequence of sounds) with a particular object or event. For example, a
child can accurately call a dog by its name no matter when and where
s/he sees it. To do so, the child has to dissociate the word from the
environment or event in which the word occurs and learn to abstract the
actual concepts represented by the word. The packaging task is to
discover a range of possible meanings under a given label. This process
typically involves two types of errors, underextension and overexten-
sion. Children also develop a sense of the prototypical meaning of a
word. Aitchison puts it this way: “Children, like adults, look for clusters
of properties which belong to a prototype” (1994: 175). The third task
refers to the process of how to fit words into a semantic network. I shall
deal with this issue later.
We may assume that the way adults acquire the L2 word meaning
is similar to the way children acquire the L1 word meaning. However,
the two processes may be quite different. Bialystok (1995) cites
evidence that children discover and acquire meaning while they are
maturing cognitively and this process is partially supported biologically;
whereas adults already have a conceptual and meaning system from
their L1, with the result that acquiring the L2 meaning is essentially a
reconstruction process of their original conceptual system. This process
can be very slow and is sometimes never fully completed, which
explains why so many L2 learners stop progressing at a certain level.
How is a lexical entry developed in a learner’s L2 lexicon? In
his influential speech production model, Levelt (1989) sees each
lexical entry containing two components, the lemma and the lexeme.
55
The lemma consists of the semantic and syntactic information and the
lexeme includes the morphophonological information (or the word
form). Based on this lemma-lexeme distinction, Jiang (2000) presents
a graphic description of a lexical entry; see Figure 3.

semantics syntax lemma

morphology phon/orth lexeme

Figure 3. The internal structure of the lexical entry (Jiang, 2000: 48).

The semantic information includes word meaning and other possible

meaning related information such as word associations; the syntactic
information includes word class, grammatical functions, and con-
straints on use, etc.; the morphological information contains variants
of word forms such as inflected forms; the phonological information
contains spelling and pronunciation. It is seen that most of the lexical
knowledge discussed in Chapter 1 can be united in such a lexical entry,
or in a node in the mental lexicon.
Jiang notes that for L1 speakers “these different types of
information are highly integrated within each entry, such that once the
entry is opened, all the information automatically becomes available”
(2000: 49). He refers to this as “lexical competence” as opposed to
lexical knowledge which he views as conscious knowledge about the
word (form, meaning, use, etc.) stored in the semantic memory but
outside the lexical entry. Jiang’s conceptualisation of lexical
competence is similar to procedural knowledge and that of lexical
knowledge is similar to declarative knowledge. He (2000: 50−54)
points out that, due to the inadequate genuine L2 input and the

56
pre-existing conceptual system, an L2 lexical entry has to undergo three
stages to be fully developed with all four types of information being
integrated. The three stages are illustrated in Figure 4.

L2
phon/orth

L1 L1
semantics syntax
L2
phon/orth

L2 L2
semantics syntax
L2 L2
morphology phon/orth

Figure 4. Lexical development in L2: from the formal stage to the integration stage
(adapted from Jiang, 2000: 54).

In the initial stage of learning an L2 word, the learner’s main task is to

connect the L2 word form with an existing meaning in the mind, taking
the form of an L1 translation or a definition. Learning the new word
form is the priority at this stage while other information (semantic,
syntactic and morphological) is largely ignored. Thus an L2 entry only
contains the L2 word form and can be referred to as the “formal stage”
(Jiang, 2000: 51) of the development of a lexical entry. At this stage
57
only the word form is registered in the lexical entry. Each retrieval will
strengthen the link between the word form and the associated L1
translation or the definition. Thus the word form may be retrieved
automatically as procedural knowledge after sufficient use, while other
types of information are declarative and need to be retrieved con-
sciously. Comprehension or production of the L2 word is achieved by
activating the link between the L2 word form and the L1 equivalent or
translation, termed “lexical association” (Jiang, 2000: 51). Via lexical
association, other lexical information, such as semantic, syntactic and
morphological information, can be accessed.
In the second stage, continued use of the L2 word strengthens the
link between the L2 word form and the lemma information of the L1
equivalent. In other words, the L1 lemma is “copied or attached to L2
lexical forms to form lexical entries”; this is called the “L1 lemma
mediation stage” (Jiang, 2000: 52). The link between the L2 word form
and the L1 lemma information is direct without the mediation of the L1
word form. In this sense, it is possible to achieve a certain fluency and
automaticity in L2 use, which may seem to be proceduralised. What is
still absent in the lexical entry is morphological information because
“morphological information is usually language-specific, and thus less
susceptible to transfer” (Jiang, 2000: 52). However, in language use
morphological information can be retrieved consciously as declarative
knowledge.
In the final stage, the four types of information, semantic, syn-
tactic, morphological and phonological (formal), are all integrated into
the lexical entry and the developing process is completed; this is known
as “the integration stage” (Jiang, 2000: 53). By the time this stage is
reached, all types of lexical information have become proceduralised
and can be retrieved for use automatically, as in the case of L1 speakers.
This three-stage development model of L2 lexical entries has
two major advantages: (1) it offers a feasible model to account for L2
lexical development and captures the possible status of a given lexical
item at a given time; (2) it provides information that can explain
lexical errors and lexical stalemate.
Before reaching the final stage, a learner will probably make
lexical errors of all four types, although the degree of error for each
type will be varied. We can also roughly predict what types of lexical

58
errors are likely to occur or decrease at each stage. As a learner’s
proficiency advances, the formal errors tend to decrease soonest since
the word form is the first thing attended to by the learner. It is possible
that certain words persist in being difficult to spell or pronounce even
for native speakers due to their intrinsic complexity: the English word
diarrhoea is a classic example. The morphological errors might be the
last to disappear since, up to the second stage, the L1 mediation stage,
L2 morphological information is largely absent in the lexical entry but
needs conscious retrieval from declarative memory. Compared with
other languages, the morphological rules of English are not com-
plicated and can be mastered in a short period as declarative knowledge.
Therefore, morphological errors might seem to be decreasing in learner
language use at an earlier stage and at a rapid rate, particularly in
written language when learners have enough time to use the conscious
declarative knowledge to monitor their output. For example, in the
study of a 150,000-word corpus of English written by French learners,
both intermediate and advanced learners made far fewer morphological
errors than other types of errors such as formal or grammatical errors
(Dagneaux et al. 1998). But in oral language use, learners may not have
the opportunity to use their conscious knowledge and this may lead to
morphological errors. It is not surprising for an advanced learner to
utter a sentence like: He become mature, especially during long,
continuous, interacting speech when s/he has less time or opportunity
to monitor his/her utterances by applying declarative knowledge.
Obviously the learner knows the inflected rules of the word become
but fails to produce the correct form in spontaneous speech.
Jiang (2004: 603) cites evidence from several studies that the
accuracy rate is low “in the use of inflectional morphemes by adult
ESL users in spontaneous communication across different first lan-
guage (L1) backgrounds and different second language (L2) proficien-
cy levels”. Jiang (2000: 63) attributes morphological errors to the
failure to integrate the lexical knowledge into the lexical entry for
automatic use; in other words, declarative knowledge has not turned
into procedural knowledge which can be accessed automatically. Thus,
learners’ morphological competence is deficient in lexical entries. To
prove this we need to find “a research method that allows us to
examine L2 learners’ performance under a condition in which their use
59
of explicit, nonautomatic knowledge is minimised” (Jiang, 2004:
607−608). Jiang rejected the research method involving L2 production
on the basis of the assumption that language production cannot exclude
the use of explicit or declarative knowledge. Instead, he (2004: 608)
opted for a receptive task in which “the morphological form available
in the input should activate the related knowledge through a bottom-up
process if the knowledge has been internalised”. The task is a self-
paced word by word task to check the sensitivity to subject-verb
agreement in a series of sentences, like the two given below. Singular (S)
or plural (P) is indicated for the head noun, the local noun, and the verb.

(1) The key to the cabinet was rusty from many years of disuse. (SSS) (Jiang, 2004:
609)

(2) The key to the cabinets was rusty from many years of disuse. (SPS) (Jiang,
2004: 609)

Previous research (Pearlmutter et al., 1999) shows that native speakers

take longer to read sentences in SPS condition, as in sentence 2, than
those in SSS condition, as in sentence 1. In SPS condition both the
head noun and the verb are in the singular and the two are separated
by the local noun in the plural; this disagreement has somehow made
the subjects spend a little more time reading the sentence. Pearlmutter
et al. (1999: 435) attributed this phenomenon to native speakers’
“sensitivity to agreement violations and head/local NP (noun phrase)
mismatches”. In other words, lexical processing will be disrupted when
such disagreement occurs, which Jiang (2004: 609) calls the “broken
agreement effect”. Following this finding, Jiang (2004) carried out a
series of experiments to pin down this broken agreement effect. I only
report the first experiment here.
Native English speakers and Chinese learners of English 10 were
compared to find out whether the latter would be sensitive to the plural
morpheme in a self-paced reading task. Two lists of sentences were
constructed, each containing either sentences in SSS condition or SPS

10 These Chinese subjects are considered as proficient English learners: they had
an average TOEFL score of 608 and had lived and worked in the U.S.A for an
average of 1.4 years.

60
condition. Each subject read the sentences in either SSS condition or
SPS condition word by word in front of a computer screen. Three
locations of each sentence were measured: the local noun, the verb be,
and the word immediately after be. The results show that native
speakers took significantly longer to read the verb be and the word that
came immediately after be in SPS condition than in SSS condition,
indicating that they noticed the number disagreement between the local
noun and the verb be. However, no significant difference was observed
for Chinese learners on both locations between the two conditions
except that they took longer to read the local noun (plural) in SPS
condition. The other two follow-up experiments replicated and
strengthened this finding by excluding other interfering factors that
might explain the result differently. The experimental results seem to
indicate that non-native speakers, or at least Chinese learners of English,
are not sensitive enough to the number morpheme in processing reading
texts. Jiang suggests: “The plural morpheme -s might have been
perceived and processed, as shown by their reading time difference at
the first position, but only as an orthographic unit, with little
morphological ramification for sentence processing” (2004: 615).
However, more research needs to be done to determine whether this
finding can be generalised to other learners of different L1 or whether it
can be applied only to a language like Chinese which does not have
inflected morphemes.
The common semantic errors include inappropriate word choice;
syntactic errors often consist of inappropriate use of the word (wrong
collocates, wrong tenses). This is largely due to the fact that an L2
word does not have the same overlap of semantic or syntactic pro-
perties with the L1 word, which is partially covered by the notion
lexical concept gap discussed later (see Chapter 4). Consider the two
examples:

(3) She baked a chicken for dinner.

(4) John participated to the project.

The first sentence is likely to be produced by a native Chinese speaker

since the words roast and bake are both translated by one word kao, as
61
in Chinese there is no conceptual distinction made between cooking
meat in the oven with or without fat or oil. The second sentence is
commonly produced by native French speakers since the French
equivalent of participate in is participer à, literarily translated into
English as participate to. These errors belong to the category of
negative L1 transfer. Not only a very different but also a close L1 lan-
guage may lead to negative L1 transfer whenever the lexical
information is not completely identical in the two languages. For two
languages sharing many cognates, “false friends” constitute frequent
pitfalls for learners regarding lexical meaning and use.
According to Jiang (2000, 2002), most words stop at the second
stage without further development. The main cause of lexical stalemate
is the L1 lemma mediation, an innate feature of the second stage. Jiang
puts it this way (2002: 619):

[…] once L1 semantic information has entered L2 entries and occupied the
lemma space, it is very hard for new meanings to get in. The semantic
information that is copied from the L1 translation stays in the L2 lexical entry
and continues to mediate L2 word use even with continued exposure to the L2.
As a result, even highly proficient L2 users will use L2 words on the basis of
the semantic specifications of their L1 translations. […] That is, the
form-meaning remapping suggested by some researchers or the recombination
of conceptual features envisaged in de Groot’s (1992) model may not always
take place successfully.

The three-stage development model of L2 lexical entries implies that

L1 lemmas (the semantic and syntactic information) are playing an
important role in helping the learner acquire the new L2 words. It is
inevitable that learners, either intuitively or by taking short-cuts, will
map L1 word information, particularly the word meaning to the L2
word form, indicating that L1 lemma mediation is indeed an important
feature of learner lexical development at a certain phase. This feature of
lexical development constitutes part of vocabulary learning difficulty,
and will be discussed later.
Although very promising, the three-stage development model of
L2 lexical entries is not problem-free. One major problem is the fairly
large gap between the second and the third stage, i.e. from the L1
lemma mediation stage to the integration stage. There is a big jump

62
between the two stages and no satisfactory explanation is provided as to
how to explain the transition from the second stage to the third stage.
Although very small in number compared with the vast number of L2
learners, some adult L2 learners do achieve native-like language
competence, including lexical competence. How do most words get
beyond the second stage and complete the lexical development process?
Will there be more sub-stages during this transitional period? We do not
have the information to answer these questions. In addition, the
three-stage lexical development model is intended to describe lexical
development in a formal educational setting where the L1 is perhaps the
most important mediator for L2 acquisition when authentic L2
exposure is limited. What about learners who learn the language in the
L2 country? Will they experience the same three stages in terms of
lexical development? If not, how? The model might not be able to
account for lexical development satisfactorily in a natural L2
acquisition environment, as Jiang himself admits (2000: 72). In this
sense the generalisability of the model is limited. On the other hand, it
could be argued that this model is very useful since the majority of L2
learners are learning the L2 in a formal educational setting usually in
their own country and have limited opportunity to stay in the L2 country.
Lastly, this model seems to be more suitable for learners who are
learning a more distant language from their L1 (e.g. English and
Chinese). One wonders whether it would be in any way applicable to
two closely related languages (e.g. Spanish and Italian) where there is
considerable semantic, grammatical and cultural overlapping.

2.2. Organisation of the L2 mental lexicon

2.2.1. Relations between lexical entries

We have seen how a lexical entry is developed in the L2 lexicon and

described the possible stages. We may assume that all lexical entries
enter the lexicon this way, though not all L2 entries, or perhaps only a
small proportion, will be fully developed. Lexical entries are not
63
isolated and they are connected or linked with each other in different
ways. Now we shall look at the inner structure of the lexicon: how it is
organised. We need to ask what the relationship is between individual
lexical entries.

2.2.1.1. Possible relations between lexical entries

We first take a look at how entries in the L1 mental lexicon are related
to each other. Levelt (1989: 183−185) distinguishes two basic types of
lexical relations: relations within and between entries. The former refers
to the inflected forms of a lexical item, for example, go, goes, went,
gone, and going all belong to the same lexical entry: go. In other words,
all these items are related within an entry. However, derivations such as
happy, happiness, unhappy belong to different entries.
Relations between entries can be further divided into two types:
(1) intrinsic and (2) associative relations (Levelt, 1989). A lexical item
can be intrinsically linked with another in four ways according to the
four types of information contained in a lexical entry, i.e. semantically,
morphologically, phonologically, and syntactically. Semantic relations
include hypernyms, synonyms, antonyms, etc. Lexical items can be
morphologically linked to each other by sharing the same morphologi-
cal stems, such as class, classify, classification, etc. Morphologically
related words are likely to be semantically related as they share a
common meaningful word part. Lexical items can be connected on the
basis of similarity of phonological forms. Finally, lexical items can be
related according to syntactic features, such as word class, gram-
matical functions, etc. Intrinsic relations can be direct or mediated.
Associative relations between entries are not based on semantic
features but on frequent co-occurrence or collocations of lexical items;
for example, funny and joke, blue and sky, sea and beach, tend to
occur together. An associative relation is similar to what is commonly
called a syntagmatic association in applied linguistics. Sometimes
intrinsic relations can also be associative as meaning related items
often occur together.
It seems that the evidence given by Levelt to demonstrate these
lexical relations, particularly intrinsic ones, comes largely from speech
errors. Sometimes native speakers make slip-of-the-tongue lexical

64
errors that can be attributed to similarity (or confusion) of meaning,
phonology or pragmatic properties of lexical items (see also Channel,
1988). The following are some examples:

(5) Irvine is quite clear. (blending of close and near) (Fromkin, 1973, cited in
Levelt, 1989: 184)

(6) Open → over (Fay and Clutter, 1977, cited in Levelt, 1989: 184)

(7) Week → work (Fay and Clutter, 1977, cited in Levelt, 1989: 184)

Such lexical errors do present themselves in native speakers’ speech,

although at low frequency as the statistics show that there is a lexical
error every one thousand items (Levelt et al., 1999). The same errors
occur in L2 learners’ speech, probably at higher frequency but this
would very much depend on L2 language proficiency. It sounds
reasonable to assume that lexical entries are connected with each other
by one or more of the relations discussed above; however, the
assumption would be more powerful if it were substantiated by more
convincing evidence. More research needs to be done in this area.

2.2.1.2. Elaborating semantic relations between English words

According to Miller and Fellbaum (1991: 197−229), semantic rela-
tions are innate to English words and may characterize the underlying
structure of the mental lexicon. They point out that (1) synonymy is
the most common semantic relation and (2) words of different classes
such as nouns, adjectives, and verbs have their own preferred semantic
relations and are organised differently according to their functions. In
addition to synonymy, common semantic relations between lexical
items include antonymy, hyponymy/hypernymy, meronymy, tropo-
nymy, and so on. Most people are familiar with synonymy and
antonymy. Hyponymy and hypernymy describe a kind of synonymy
which indicates subordinate or superior relations between words, e.g.
rose is a hyponym of flower or flower is a hypernym of rose.
Meronymy depicts a kind of part-whole relation between words, e.g.
finger is a meronym of hand. Troponymy is a kind of hyponymy,
typically between verbs, e.g. to devour is to eat quickly and eagerly.
65
Words can be put together by one or more of these relations and form
a “synset” (synonym set) (Miller/Fellbaum, 1991: 201), e.g. [suit,
dress], [cat, animal], [good, bad], and so on, which expresses a
different underlying lexical concept. Thus the mental lexicon is
organised according to these five or more types of relations, with links
from one synset to another.
Nouns are basically organised in hierarchies produced by
hyponymy, e.g. greyhound → dog → canine → carnivore → mammal
→ animal. All English nouns can be categorised into 26 broad
categories (Miller/Fellbaum, 1991: 204−205) and each noun can find
itself somewhere in a hierarchical chain in one of these 26 categories:

[act, action, activity] [natural object]

[animal, fauna] [natural phenomenon]
[artifact] [person, human being]
[attribute, property] [plant, flora]
[body, corpus] [possession, property]
[cognition, ideation] [process]
[communication] [quantity, amount]
[event, happening] [relation]
[feeling, emotion] [shape]
[food] [society]
[group, collection] [state, condition]
[location, place] [substance]
[motive] [time]

A distinction is made between adjectives: they can be predicative or

non-predicative. Predicative adjectives can be used after the verb to be
and non-predicative cannot. Non-predicative adjectives are organised
like nouns in hierarchies, and are hyponymous in nature. In contrast,
predicative adjectives are basically organised antonymously, and can
be direct or indirect. For example, deep and shallow are direct
antonyms; abysmal and shallow are indirect antonyms mediated by
deep. It is assumed that adverbs are organised in a similar way to
adjectives since most adverbs are derived from their adjectival forms.
Verbs can be put into 14 categories that depict actions or events: verbs
of body care and functions, change, cognition, communication, com-
petition, consumption, contact, creation, emotion, motion, perception,
possession, social interaction, and weather verbs (Miller/Fellbaum,

66
1991: 215−216). The basic relation among verbs is troponymy; other
possible relations include meronymy, entailment (e.g. eating entails
chewing), opposition, etc.
Miller and Fellbaum’s work elaborated one important aspect of
the lexicon, the semantic relations or, to be more specific, the
paradigmatic associations between English words and how they are
organised. These lexical relations have been employed to produce a
large-scale electronic English lexical tool called WordNet (Miller,
1990). The purpose is to simulate the human lexical memory with
computer applications in order to better understand lexical semantics.
In a sense, WordNet is a dictionary with features of a thesaurus to
enable users to quickly find a lexical item and its semantic relations
with other words. It groups English words into synsets, provides short
definitions, and records the various semantic relations between these
synsets. What is produced is based on what has been understood from
lexicography and psycholinguistics. The question is whether the
simulated mental lexicon really resembles the virtual mental lexicon 11.
Second, only semantic relations between words are dealt with, while
other types of relations could also play a role in organizing the mental
lexicon. These two questions need to be addressed to have a better
understanding of how the mental lexicon might be organised.

2.2.2. Review of empirical studies investigating the differences

in organisation between the L1 and the L2 mental lexicon:
word associations

Word association studies are the most common means for gathering
evidence about the way words are connected with each other in the
lexicon. Introduced by Francis Galton in the 19th century, word
association was soon adopted by psychologists as a way to explore
human thoughts and became popular in lexical studies (Miller, 1996).
In the classic word association study, a list of words is given and

11 The mental lexicon is often talked about and exists as a kind of metaphor,
such as network, library, computer, dictionary, thesaurus, encyclopedia, etc.
But nobody knows what it might actually look like.
67
subjects are asked to respond with the first word that springs to mind
for each word in the list. The most famous list contains 100 prompt
words and is named the Kent-Rosanoff list after the two researchers
who produced it in 1910. The list was given to a large population
(1000 men and women) and the results revealed some strikingly
similar patterns, i.e., that people tended to give similar responses to a
particular prompt word. For example, in response to chair, the most
common responses are table, seat and sit (Miller, 1996: 157). Meara
(1996b) claims that the results of word association studies (for native
speakers) tend to be fairly stable with 50 subjects or so and a larger
population does not make a significant change. One deduction from
the similar patterns of responses is that native speakers share similar,
stable networks in respect of the mental lexicon.
Through word association studies, three types of word associ-
ation responses are typically identified: syntagmatic, paradigmatic and
clang associations (Meara, 1996b). With syntagmatic associations, two
words occur as a collocation or in proximity. For example, the response
to the prompt word canary can be small, pretty, fly, and sing. The
response and the prompt word usually belong to different word classes.
Paradigmatic associations are responses that are semantically related to
the prompt word by synonymy, antonymy or hypernymy, and so on. As
a consequence, they have the same word class as the prompt word. For
example, the prompt word flower produces responses like rose, leaf,
plant, etc. Clang associations are neither semantically linked nor
collocated with the prompt word; rather, they are phonologically or
formally similar. For example, clang associations for the word blink
could be pink and link. It is generally believed that L1 children show a
shift in association responses as they grow older (Wolter, 2001; Schmit,
2000): young children tend to produce more syntagmatic associations
than paradigmatic associations together with a large number of clang
associations but as they grow older paradigmatic associations increase
while syntagmatic and clang associations diminish; eventually they
produce association patterns in line with prototypical native patterns
of association. Schmitt (2000: 40) cites evidence to show that the shift
occurs in a different order for words of different classes: nouns come
first, followed by adjectives and verbs.

68
 Although not the first to explore the differences between the L1
and L2 lexicons, Meara was perhaps one of the earliest to arouse
researchers’ interest in this issue, first raising it in the late 1970s. From
word association studies on L2 learners, Meara (1996b) points out that
the structure of the L2 learners’ mental lexicon is in general
significantly different from that of L1 speakers but it can be changed
over time. He (1996b: 2−7) observes that: (1) learners tend to produce
more varied and idiosyncratic responses than L1 speakers; (2) learners
tend to produce a large proportion of clang associations the way L1
young children do; (3) learners often mistake a prompt word for another
phonologically similar word and consequently produce strange word
associations; (4) the semantic links of the learners’ mental lexicon are
rather fragile and prone to be influenced by formal similarities; (5) as
learners’ language proficiency advances they tend to produce native-
like word associations. Soderman (1993, cited in Schmitt, 2000: 41)
provides further evidence to support Meara’s fifth point. In her studies,
the L2 learners of English (Scandinavian school children and young
adults) show a syntagmatic-paradigmatic shift in their responses along
with fewer occurrences of clang associations as they progress.
It is not easy to achieve native-like patterns of word association,
or to have an L2 lexicon organised similarly to that of a native speaker;
in fact, it might well involve a life-long learning process. In section
2.1.2., we saw that the L2 mental lexicon of most learners is underde-
veloped because of L1 transfer, the way an L2 is learned and many
other factors. Schmitt and Meara (1997) investigated the word
association knowledge of 94 young Japanese secondary school and
university learners of English and found that the students generally
demonstrated poor knowledge of word associations; in the case of
words claimed to be known, the word associations produced only
accounted for 50% of the native patterns. The poor word association
knowledge was attributed to three reasons (Schmitt/Meara: 31−32): (1)
confusing prompt words with other formally similar words, (2)
conceptually correct word associations but not native-like ones (e.g.
complete → whole, moon), (3) associations connected with the class-
room (e.g. complete → sentence, homework, report).
Schur (2002) investigated the structure of semantic networks of
native English speakers and Chinese learners of English using a word
69
association task. On a sheet of paper 50 common English verbs were set
out in a box in alphabetical order; below this a random list of the same
50 words was given with a number before and a blank after each word.
Subjects were asked to choose one word from the box to fill in the gap
for each word in the list to form a pair. Words in the box could be
chosen repeatedly. See below for part of the work sheet (Schur, 2002):

Verb Box:

to argue to assist to build to buy to clean to cost

to cut to describe to discover to dream to drive to yell

Verb list:
1. to help _______ 2. to study _______ 3. to scream _____
4. to tell _______ 5. to expect _____ 6. to wash _______
7. to love ______ … 50. to (…) _______

Based on the completed word pairs, a “network” can be drawn for

each subject. A typical native speaker’s “network” contains 16
semantic groups: each word in a group is connected with others by
links and there is no link between groups. A typical Chinese speaker’s
English “network” contains 7 groups. It seems that the number of
semantic groups produced depends on the level of English proficiency:
more groups for a higher level and fewer groups for a lower level. The
L2 learners’ network contains less than half as many semantic groups
as that of the L1 speakers, which means that within each group there
are more words and more links. Thus the L2 learner’s “network”
appears to be more complex and denser than that of native speakers.
However, Schur argues that this does not mean that they have a deep
knowledge of these verbs; it is, on the contrary, a manifestation of
their poor lexical knowledge. Although it is not certain that such a
“network” generated with 50 verbs resembles the virtual mental
lexicon, a rough look at the data would suggest that Chinese L2
learners frequently generate strange, idiosyncratic associations such as
fight → argue, expect → imagine, or syntagmatic-like associations
such as try → drive, test → prove, etc. It is very likely that in these
Chinese L2 learners’ L2 lexicons the L1 lemma information still
occupies most of the lexical entries for these 50 English verbs, which

70
is congruent with the L1 lemma mediation proposed by Jiang (see
2.1.2.). The L1 influence is obvious: the Chinese lexical item zheng
lun can be literarily translated into to fight and to argue, although the
most appropriate translation would be to argue, which is why they
associate fight with argue. The Chinese yan zheng can be literarily
translated into to test and prove, although the most appropriate
translation is to validate, which explains why Chinese learners
associate test with prove. It can thus be seen that L1 transfer is one of
the main reasons why the L2 learners’ word association patterns are
different from those of native speakers, consequently leading to a
marked difference between the structures of the two lexicons.
Wolter (2001) argues that the L1 and the L2 lexicons are not
necessarily inherently different in structure, but it is the degree of
knowledge of the individual words which make up the lexicon that
determines how it is organised. He (2001: 48) constructed a model of
depth of word knowledge in which a word can be known to four
degrees: slightly known, moderately well known, fairly well known, and
well known. It is hypothesized that for both L1 and L2 speakers, the
word associations generated should be similar for prompt words known
to the same degree of depth. In testing what is hypothesized, a word
association study is combined with the VKS (Vocabulary Knowledge
Scale; see 1.4.4.1) test. Each prompt word is used to generate an
association as well as to test how well the word is known. The results
show that the hypothesis is partially true, depending on the degree of
word knowledge. The L1 speakers’ and L2 speakers’ associations show
similar patterns when words are not or only slightly known. For words
that are moderately well known, the L2 speakers make a significantly
lower proportion of paradigmatic associations and a significantly
higher proportion of clang associations. For words that are well known,
the L2 speakers clearly show a preference for syntagmatic associations
over paradigmatic associations, which is the reverse for L1 speakers.
Despite this, Wolter argues that L1 speakers tend to produce more
paradigmatic associations than L2 speakers probably because the
former have a (much) larger lexis than the latter; thus more para-
digmatic associations do not necessarily suggest a higher level of
lexical structure but an indication of larger lexical size. Also, although
the measurement criteria of word knowledge (using the VKS) are the
71
same for both L1 and L2 speakers, what the L2 speakers actually know
for a word interpreted as “well known” might be different from that of
the L1 speakers: the VKS might not be sensitive enough to reveal these
differences. Nevertheless, Wolter proposes to look at the mental lexicon
and its organisation from a developmental viewpoint. He (2001: 65)
writes:

[…] the progression for individual words could be viewed as moving from a
state in which phonological and other nonsemantic connections are dominant
to a state where syntagmatic or paradigmatic connections take precedence.
This is not to suggest that higher level semantic connections simply replace
the lower level phonological and nonsemantic connections, rather there would
appear to be a process in which the initial connections are retained, but the
later connections become dominant. In this sense, a well-known word has
several connections of different types (i.e., phonological, syntagmatic, and
paradigmatic), but some connections are stronger than others.

Following this developmental approach, Namei (2004) investigated

the organisation of the mental lexicon of young Persian-Swedish
bilinguals comparing both Swedish and Persian L1 lexicons, using the
Kent-Rosanoff word association list. Among the subjects, the
youngest was 6 years old and the oldest 22 years old. They were
highly proficient in both languages while their Swedish was more
advanced since they had been living in the country. The results
generally confirm what has been pointed out in previous studies,
namely, that there are more clang associations among younger
children and that there is a syntagmatic-paradigmatic shift between the
ages of 6 and 10. However, there are other important findings that
differ from the previous studies (Namei, 2004: 382): (1) Clang asso-
ciations persist even for the most advanced L1 and L2 speakers and
their occurrences are determined by word knowledge rather than
general language proficiency; (2) a considerable number of syntag-
matic associations are produced by even the most advanced L1 and L2
speakers; (3) a considerable number of paradigmatic associations are
produced even by the lowest level L1 and L2 speakers. Namei thus
argues that the mental lexicon is not organised according to a
syntagmatic-paradigmatic shift, which simply indicates increased
lexical knowledge for all individual words; instead, it is organised

72
according to the stages at which words develop. He proposes a model
in which words in the mental lexicon are organised along a
word-knowledge continuum:
Increasing language exposure

unknown barely moderately fairly well-known

familiar known well-known

paradigmatic/
form-based syntagmatic paradigmatic late syntagmatic

Figure 5. Word-knowledge continuum and main organisational features of words in

the mental lexicon (Namei, 2004: 382).

Thus, words that have been recently learned are organised according
to form; those that are partially known tend to stay with their
collocates or co-occurring words; those that are well-known are linked
with semantically related words such as synonyms, antonyms, hypo-
nyms, and meronyms. Since the subjects in this study were either L1
young speakers or very efficient young bilinguals living in the L2
country, the results can best account for the L1 mental lexicon and
their direct application to the L2 mental lexicon (particularly the L2
lexicon of adult learners) needs to be further validated.

2.3. Conclusion

The studies regarding the organisation of the L1 lexicon show that

words are connected in one way or another, particularly through
paradigmatic associations (synonyms). The structure of the L1 and the
L2 lexicons are different from each other but the two can be potentially
similar to each other, depending on a number of factors: L2 proficiency,
L2 lexical knowledge, L1 influence, learning environment, etc. Studies
of L2 word associations show that the clang-syntagmatic-paradigmatic
association shift follows the pattern of the L1 as the L2 learner
73
progresses. However, whether this shift reflects overall language pro-
ficiency or the degree of knowledge of individual words is not clear and
further studies are needed.
It follows from these findings in respect of word associations that
a network with numerous nodes and links would be the metaphor that
best represents the mental lexicon. Other metaphors do not easily
accommodate the many constant internal changes that are taking place.
The nodes of the network, or the words, can be related or linked in
many ways, revealing features of the inner structure of the network, but
it would be a mistake to suggest that each node has a fixed location and
remains near to some other particular words. A fixed location cannot
easily allow the co-existence of many different types of links between
other words. For example, a word may possibly be simultaneously
close to other semantically, morphologically, syntactically, phonologi-
cally related words. It is not the location but the links that matter. It is
also likely that a lexical node is changing its location constantly on
account of its links with other words and the contexts where it is
retrieved for use. Thus the whole mental lexicon is in a dynamic state,
its inner structure constantly experiencing synchronic and diachronic
change.

74
Chapter 3. The Role of Listening in SLVA

3.1. The historical background

With the transition from the grammar-translation method to the

audio-lingual method in the 1950s, when there was a move away from
the written language to the spoken language, so much emphasis was
placed on speaking that it was often overlooked that communication is
a two-way process, and comprehension, i.e. listening, was given very
little attention. This continued until well into the 1970s. During that
period many language learning materials were published where the
main focus was firmly on speaking, though reading and writing were
also practised (Kelly, 1980). In these materials listening was invari-
ably treated as a supplementary, subordinate activity to assist in
achieving speaking proficiency, just as it had been during the audio-
lingual period.
During the late 1970s, a large number of listening courses were
put out by British publishers following the complaint by applied
linguists that this skill was being neglected and that research needed to
be carried out in this area (Kelly, 1985). As this author points out (Kelly,
1985:51), the main preoccupation was with improving learners’
knowledge of spontaneous spoken language by the use of “authentic”
materials, excluding as far as possible oralised written language. 12
These courses placed emphasis on the development of strategies and
took a global approach, with little effort being made to integrate
listening comprehension into the language learning process. Such
courses largely disappeared from the language learning scene after a
few years and in the 1990s the accusation of neglect was again being
made (Oxford, 1993a). Several reasons probably contributed to this
neglect, e.g. the difficulty of teaching the skill in the classroom, the

12 Kelly (1985) lists 33 such courses, 27 of which appeared between 1975 and
1981, reflecting the awareness at the time of the need to focus on listening.
75
so-called passive and supposedly uninteresting nature of the skill, but
primarily the cultural prominence of speaking. Kelly (1995) argues that
this is a feature of Western culture and, to support his case, points to the
listening courses Steil (1980, cited in Kelly, 1995: 19) organised for
American executives to make them better listeners: they were losing
out in business because their main preoccupation was with pushing
their own ideas and interests.
In contrast to this superficial treatment of listening, a number of
linguists, on the basis of experimentation and findings in psychology
and neurolinguistics, developed methods that were squarely based on
listening. Prior to examining the salient principles underlying these
methods, one illusion in respect of listening needs to be dispelled,
namely that listening is a passive process.

3.2. Listening: A dynamic language activity

At the moment when it was being officially acknowledged that the

audio-lingual method was not the answer to resolving the problems of
second language acquisition, the Canadian linguist Pimsleur (1970) was
drawing attention to the unsuitability of the epithet “passive” in respect
of listening. He argued that it should be described as a “receptive” skill.
Following his plea, a number of linguists adopted this term, although
later Oxford (1993a) argued that even this adjective was inappropriate
as it still implied a lower level of activity than speaking.
In the first place, it should be noted that listening, like speaking,
is a physical activity. If we observe a person listening to another, we
perceive facial, head and body movements that reflect the listener’s
mental processes. As Oller (1978) put it, the listener and the speaker are
like two puppets moved by the same set of strings. The listener’s body
“dances” in harmony with that of the speaker. It is sufficient to look
through the window of a restaurant or café to have proof of this
phenomenon; further evidence is provided by the mobile phone when
even at a distance when the phone or the arm holding it are not visible it

76
is easy to see that the person is “dancing” in unison with an invisible
interlocutor.
The decoding process also more than demonstrates the active
nature of listening. The distinction between top-down and bottom-up
processing clearly illustrates this. While bottom-up processing is the
application of the ear to what is uttered, top-down processing is the
application of the listener’s grammatical and lexical knowledge in order
to understand what is said. As Pimsleur (1970) argued, it is only by
means of this second type of processing that we can explain what he
calls the cocktail party phenomenon, i.e. why it is we can understand
what a speaker says though we hear probably only half of what s/he
utters. Extralinguistic knowledge, i.e. situation, topic of conversation,
etc., also comes into play but our brains are working overtime drawing
on our reserves of knowledge of the language. The greater those
reserves, the more likely we are to understand. A language learner who
does not have a great knowledge of the target language will doubtless
understand far less, despite a sharp ear, than an ageing monolingual
with impaired hearing. Even advanced learners cannot compete with
native speakers as Whitson (1972) clearly demonstrated. Kelly (1992)
carried out investigations that showed that it was not poor auditory
perception but a lack of lexical knowledge that was the main obstacle
for advanced learners in listening comprehension 13. Kelly (1992: 138)
puts it this way:

Even if the foreign language learner could acquire the highest degree of
auditory perception attainable by the native speaker, he would not find it
much of an advantage: unless he can learn to use his language knowledge and
other available or previously acquired information to predict or anticipate
what will be said, to deduce or recognise words on the basis of absent or
incomplete sound indicators, to bring into play his knowledge of the sound
patterns of the language, a keen ear will be of little use to him.

It can be concluded that listening is an active/dynamic process in

every sense.

13 Top-down processing also operates at the reading level and explains why, for
example, a native speaker can often read with relative ease and full understand-
ing a badly photocopied page where the last word on each line is missing.
77
3.3. Listening as the basic skill in SLA

The most notable linguists who argued for the dominance of listening
in SLA were, for the most part, psychologists by training and, initially
at least, by profession. The first of these was the Bulgarian Lozanov
(1979) whose method is still widely used in Western countries and
goes by the name of Suggestopeodia14. Here the salient characteristics
are for learners to change their identity and then to sit in comfortable
armchairs, listening to large quantities of language against a back-
ground of soft music. The second was Asher, who started experiment-
ing in the 1960s and became widely known for his Total Physical
Response method (1983). He argues that all language can be taught
using the imperative which he calls “the golden tense”; while his
classes consist of learners constantly moving about and doing things
on the basis of instructions from the teacher in the target language, he
makes considerable use of cards on which abstract and other terms are
written in large letters, e.g. “John, put JUSTICE on Charlie’s head”.
The German-speaking child psychologist, Winitz, produced his listen-
ing based language course, The Learnables (1978). Using pictures and
sound recordings, he takes the learner through the same grammatical
sequences as the native speaker learning their first language. While
not developing a specific language learning method, Nord (1978) did
much to promote the importance of listening and its benefits, arguing
for a totally new listening based paradigm in the field of second
language acquisition.
In the early 1980s, Ostyn and his colleagues (Ostyn/Godin,
1985; Kelly 1989), developed a method where listening was closely
linked to the acquisition of vocabulary, an issue that I shall be
discussing subsequently; they required learners to listen to quantities
of material in their own time and then to perform oral and other
exercises in class based on the material studied. Currently the most
well-known advocate of the listening based approach is Krashen

14 It appears that Lozanov’s method has been adapted in the West and not
always applied in the way he envisaged (see Wagner/Tilney, 1983, in Kelly,
1985: 252).

78
(1996) whose controversial views have stimulated considerable
research. He argues that learners should simply have large quantities
of comprehensible oral input consisting of material that interests them
and that they enjoy listening to. Other linguists could be cited, e.g.
Ervin-Tripp (1974); Gary (1975) and Ruder et al. (1977), whose
research led them to highlight the importance of listening in the
learning of a second language.
The above-mentioned methods are based on a number of
principles; it is argued that these principles need to be accepted if
second language learning is to have any chance of success. I shall
discuss briefly the three most important of these principles.

3.3.1. Activation of both brain hemispheres

It is essential that the whole brain be activated for language learning to

be successful. It is not sufficient to activate the left hemisphere of the
brain where the language centres, Broca’s and Wernicke’s area, are
located, but also the right hemisphere which governs physical move-
ment, music and imagery. The prosodic features of language, notably
intonation and stress, are thus located in the right hemisphere. So
much language teaching and learning is unsuccessful, according to the
linguists cited above, because it fails to activate both hemispheres, to
involve the individual totally – intellectually, physically, affectively –
in the language learning process. As we saw from the very brief
outline of their methods, each focuses on a different way of activating
the whole brain: Lozanov (1979) employed music and relaxation;
Asher adhered to physical movement; Winitz (1978) made use of
imagery; Ostyn and his colleagues took the language learning out of
the classroom and made it a part of the individual language learners’
private lives. Krashen (1996), like Lozanov, focuses on enjoying the
activity.

79
3.3.2. L1 acquisition as a model for L2 acquisition 15

The acquisition of an L2 can be patterned on L1 language acquisition.

Just as listening precedes speaking by as much as a year in respect of
the L1, listening must precede speaking practice in L2 language
acquisition. Asher strongly discouraged learners from speaking in the
early stages of language learning. Winitz (1978) did not allow his
learners to speak until after many hours of instruction and introduced
grammatical structures in the same order as the native speaker would
acquire them. Krashen (1987) considers that the L2 language learner
will internalise the grammar of the language without any formal
instruction, just as the native speaker does. Ostyn and his colleagues
required students to listen to and understand all the texts they had to
study prior to performing oral exercises or producing any of the
language orally.

3.3.3. The transfer effects of listening to reading and speaking

The transfer effects of listening to the other skills are considerable and
it is this more than anything else that makes listening the basic skill.
Experimental evidence has been provided to demonstrate the transfer
effects to reading and to speaking. Asher, experimenting with German
(1972) and Spanish (Asher et al. 1974) demonstrated a natural transfer
effect from listening to reading, without any prior training in reading.
In the German study, after a 32 hour listening course at the rate of one
hour a week, the learners were given a reading test and their results
compared with those of students who had covered one year of German
study at university and where the focus had been on reading and
writing. “Both groups (were) quite similar in their reading achieve-
ment” (Asher, 1972: 136). Reeds, Winitz and Garcia (1977) obtained
even better results in respect of German, claiming that Asher’s results
too would have been better had his experimental methodology been

15 It has been stated that Lozanov does not subscribe to this view but it is clear
that his massive listening input correlates with the situation of the young child
learning their first language.

80
better – he had in fact used cards with the experimental group in the
listening course. They took two groups (students and secondary school
pupils) and gave them an 8 hour listening course spread over two and
three weeks with 45 minute lessons each time. They were then asked
to translate three passages of German, consisting of words and struc-
tures they had heard but arranged to form new sentences. The average
result for both groups was close to 80%. The control group consisted
of students who did not know any German; their score was 23%.
With regard to positive transfer effects to speaking, a wealth of
empirical data exists, e.g., Asher (1964); Nord (1975); Winitz et al.
(1985); Gary (1975); Ruder et al. (1977); Winitz/Reeds (1973);
Postovsky, (1974). Here I limit myself to the findings of just two of the
researchers mentioned. Those of Postovsky are particularly noteworthy
and at the time attracted a considerable amount of interest. He was
teaching Russian to two different groups, using the audio-lingual
method with one group; the other group consisted of stenographers who
were simply required to transcribe large quantities of spoken Russian
played to them on tape. At the end of the course, he found that the
stenographers could also speak Russian and they had a better accent
and made fewer grammatical errors than the students in the audio-
lingual group. His later reports confirmed this finding (1975, 1977).
Gary (1975) carried out a similar study, but with primary school
children learning Spanish. She randomly assigned her two pupils to two
groups each of the same size. The course lasted twenty-one weeks and
consisted of 85 twenty-five minute lessons. One group was not required
to speak for the first 14 weeks nor for the first half of the lesson in the
remaining weeks. The second group, the audio-lingual group, was
required to speak from the beginning. Both groups had the same
amount of listening and both enjoyed their lessons. At the end of the
course, the listening only group could speak Spanish as well as the
audio-lingual group.
There are other principles that these methods have in common,
notably the learner-teacher relationship and the need for regular
rehearsal, two features of language learning that almost any linguist
would subscribe to. Despite the soundness of the basis of these
listening methods, relatively little attention is given to them today.
Oxford (1993a) describes listening as “a neglected stepchild” despite
81
the efforts made previously to promote it. There are a number of
reasons for this, probably the most notable being the move away from
methods to approaches, which are essentially eclectic and holistic in
nature; learners have different learning styles, use different strategies,
have different goals, find themselves in different practical and social
contexts. There is also the focus on speaking that I discussed earlier
and which shows no sign of diminishing, as is well illustrated by the
full-page advertisement that has appeared many times in the Guardian
Weekly newspaper in recent years: Speak another language? Speak it
better.

3.4. The importance of listening in vocabulary acquisition

N. Ellis (2002) strongly argues that frequency is the inseparable com-

ponent of SLA and its quantity and quality determine the language
proficiency achieved in terms of accuracy and fluency. N. Ellis (2001:
36) points out that “language is learned”, by which he means that
learners need to spend a formidable amount of time in processing the
language input and using the language. It follows that listening
becomes the most important form of language input. It should also be
borne in mind that a great deal more of our time seems to be devoted
to listening than to speaking. According to Feyten (1991), more than
45% of our total communication time is spent on listening; speaking
takes up 30%; reading takes up 16%; writing takes up 9%. Although,
as we have just seen, the idea uppermost in most learners’ minds is to
speak another language, to express their own needs, ideas, etc.,
communication is very much a two-way process. And not only is a far
greater language repertoire, particularly lexical, needed to understand
what others are saying than what one wishes to say oneself. It is
through language frequency input, particularly listening, that learners
master thousands of words, multi-word items and longer strings of
language.

82
 Experienced teachers, when asked what are the greatest obstacles
confronting the language learner, will doubtless mention listening
comprehension and vocabulary learning. They are probably the most
time-consuming and difficult areas that have to be faced and
paradoxically are the two that were the most neglected in the past in L2
acquisition research. The level of attainment in these two areas
determines the learner’s degree of proficiency in the target language
(Spolsky et al., 1968).
The link between the two, between listening comprehension and
vocabulary learning, has been explored by a small number of
researchers (e.g. Gary/Gary, 1982; Kelly, 1992). In a series of
experiments, Kelly (1992) showed that in the learning of vocabulary, if
the sound input is provided, learners retain the vocabulary longer than
when it is learned only visually both for reading and for listening
purposes. Kelly and his colleagues (1999) subsequently carried out a
series of investigations with Chinese learners where students were
obliged to listen to their tape to have the correct answers and to
resolve their difficulties. Given the positive findings from these
investigations, a self-learning course, WUFUN (Kelly/Li, 2005), was
developed consisting of a book and CD. Listening is necessary, not
only to aid in the mastery of a word’s pronunciation, both perceptually
and productively, but also to lodge that word in long-term memory
both for visual as well as auditory recognition purposes.

3.5. Conclusion

In this chapter we have seen that listening comprehension still remains

the least favoured skill in L2 acquisition despite its great importance
in the process of attaining this goal. Evidence and arguments were
presented that dispel the widespread perception that listening is
passive, showing that it is both a physical and a mental activity.
Attention was also drawn to the need for top-down processing, the
utilization of language knowledge as well as real-world knowledge, to

83
decode the utterances of others. The publications and courses of a
small number of researchers who have placed listening at the heart of
L2 acquisition were focused upon and their reasons, for the most part
as a result of findings in psychology, for doing so. It is the learner’s
total involvement that emerges as probably the key factor in
promoting listening comprehension and in consequence overall lan-
guage acquisition. In addition, given the close link between listening
comprehension and vocabulary acquisition, for any vocabulary
learning materials/programs to be successful, acoustic input of the
vocabulary and the texts is a necessity.

84
Chapter 4. Vocabulary Learning Difficulty

4.1. Cultural difference

4.1.1. Language learning: A culture bound phenomenon

In the very influential but much disputed Sapir-Whorf hypothesis

which was advanced in the first half of the twentieth century, language,
instead of being an innate biological capacity of human beings like
walking, is viewed as a purely historical heritage of the group, the
product of long-continued social usage. It is associated with the
religions, the beliefs, the customs, and arts of different peoples. In
short, we can say that walking is an organic, instinctive function while
language is the expression of a particular culture (Sapir, 1970).
Whorf (1956) further indicated that, in the first place, all higher
levels of thinking are dependent on language; secondly, that the
structure of the language one uses influences the manner in which one
understands one’s environment. The picture of the world in people’s
eyes shifts from tongue to tongue. Language is not just an instrument
for expressing ideas but shapes ideas and guides the individual's mental
activity (Hoijer, 1994). The particular language we speak, especially
the structure of that language, determines our thinking and our
perception of reality and, ultimately, important cultural patterns. The
thought patterns of the speaker of any particular language are
determined by the inherent concepts and symbols of that language.
Language always mirrors the background culture of whoever is
speaking. Miller gives this description of the relationship between
vocabulary and culture: “People who know a word can share that idea
with other members of their group, and a shared vocabulary is part of
the glue that holds people together and allows them to create a shared
culture” (1996: 5). Learning vocabulary in another language is far
more than establishing a word-for-word equivalence. As de Saussure
85
(1974) indicated, every word in every language has a unique set of
relationships with other words in the language, all of which represent
different conceptual relationships that result from that culture. For
example, a simple word such as policeman, to use an example
discussed by Vanparys et al. (1997), has very different associations in
English, French and Polish.
Nonetheless, Sapir (1970: 213−215) pointed out that cultural and
linguistic boundaries are not necessarily identical, with the result that
language does not always reflect culture in a deep sense. When stating
this, he was referring to the aboriginal American Indians; it was found
that among them totally different languages could share a similar
culture and closely related language practices (e.g. hunting buffalo,
rituals, etc.). He also gave the example of the immigrant North
Americans and the British people who spoke almost the same language
but the cultural difference was widening due to the geographical,
political, and economic differences. It is true that the Indian cultures, on
the one hand, and the American/British cultures, on the other, differ
distinctively in certain respects, but they clearly share many things in
comparison with other more distant cultures, say the Chinese culture.
The wider the cultural gap, the greater the differences.
A strong position which views language acquisition and culture
acquisition as inseparable components claims that learners cannot
master the L2 until they have mastered the culture behind the language.
This implies that by mastering the L2, particularly through staying in
the L2 country, learners learn the new ways of thinking and behaving
enshrouded in the L2 culture. However, sometimes highly proficient L2
learners living in the host country for a long time fail to be assimilated
into the L2 culture but can still understand and distinguish the
differences between their own and the L2 culture (Yoshida, 1990). A
somewhat neutral yet realistic position is that learning a new language
involves reconstructing or reorganizing the lexical concepts (at the
cognitive level) embraced by the culture to varying degrees, e.g.
Lantolf (1999). Since language is so closely linked with the culture that
nourishes it, it is reasonable to assume that being aware of and learning

86
the L2 culture can only be helpful 16 to L2 acquisition. It should be
noted that acquiring the L2 culture is not an easy task at all, the process
probably being as complex as L2 acquisition. To acquire the L2 culture,
learners need to develop considerable knowledge as such (Peterson/
Coltrane, 2003: 1):

Language learners need to be aware, for example, of the culturally appropriate

ways to address people, express gratitude, make requests, and agree or
disagree with someone. They should know that behaviours and intonation
patterns that are appropriate in their own speech community may be perceived
differently by members of the target language speech community. They have
to understand that, in order for communication to be successful, language use
must be associated with other culturally appropriate behaviour.

Learning a language means learning about its people and its culture. A
learner who has a positive attitude towards its people and appreciation
of its culture will achieve greater language proficiency than someone
who does not (Gardner, 1985). To acquire an L2 vocabulary means
acquiring the specific meaning or lexical concepts in its culture. To
understand the cultural difference between L1 and L2 is extremely
important in learning L2 vocabulary. It also implies that the greater the
cultural difference, the more difficult vocabulary acquisition becomes.
To learn about the L2 culture will certainly be helpful to SLVA, but
the main vocabulary learning difficulty arising from cultural factors
seems to come from two sources, lexical concept gaps between L1
and L2, and culturally loaded words.

4.1.2. Lexical concept gaps

The central component of learning a word is learning its meaning;

behind the meaning lies the concept which differs from one culture to

16 Learning the L2 culture will certainly aid L2 proficiency, which may be at the
cost of losing one’s own culture. Some people may view this as negative,
though L2 learners themselves do not necessarily think so. On the other hand,
learning the L2 culture could bring considerable benefits, such as being more
flexible and developing new ways of looking at the world, as has been shown
by studies of bilingualism.
87
another. The classic example is the concept of snow in the Eskimo
language. Since snow is very important in an Eskimo’s life, people
used to claim that there exist dozens to hundreds of different words for
snow. The truth is, according to the linguist Geoffrey Pallum (cited in
Miller, 1996), that there are only two root words for snow: qanik (the
snow in the air), and aput (snow on the ground). Although this
persistent myth was exposed, it does indicate that certain concepts are
important to one culture while less important or non-existent in
another culture. The incorporation of loan words into different lan-
guages gives us many good examples, e.g. English, giving the dearth
of words in respect of haute cuisine, has taken many words (along
with the dishes!) directly from French without any modification17.
Chinese culture puts great emphasis on kinship in daily life and
assigns precise terms to all family members and relatives. There are
four different terms for grandparents: yeye (paternal grandfather),
nainai (paternal grandmother), waigong (maternal grandfather) and
waipo (maternal grandmother). Different forms also exist for uncles
and aunts (and their wives or husbands) on the father’s and on the
mother’s side. Within the family, sisters and brothers are differentiated
not only by their sex but also age: meimei (younger sister), jiejie (older
sister), didi (younger brother), and gege (old brother). In this manner,
by addressing any member of the family, the exact relationship is
revealed between the addresser and the addressee. To learn all these
kinship terms in Chinese is arduous and, at first sight, may seem
unnecessary for a learner whose L1, e.g. English or French, does not
have so many precise terms.
Generally speaking, if a lexical concept exists in the L2 but not in
the L1, learning the L2 word will be difficult as it involves learning a
new concept; on the other hand, learning an L2 word will be easier if
there are more lexical concepts in the L1 for just one L2 lexical concept.
Not only the lexical concept gaps but also degree of relatedness of the
lexical concept will affect vocabulary learning difficulty. Although

17 Sometimes keeping the French pronunciation would be regarded as keeping

the original meaning of the word.

88
Chinese and Korean belong to different language families 18 and use
very different orthographic systems, the words of the two languages
generally share many similar concepts due to a similar cultural tradition,
both countries having been heavily influenced by Confucianism. It
follows that a Chinese learner may find it less difficult to learn Korean
words than English words where there is much less of a conceptual
overlap with Chinese words.

4.1.3. Culturally loaded words

In order to understand what culturally loaded words are, we first have

to distinguish between two linguistic terms: denotation and connota-
tion. When we speak of denotation, we are talking about the
dictionary meaning of the word, the central features of which are
shared by most, if not all, speakers of the language. Connotation, on
the other hand, refers to subjective factors, to the feelings and different
associations that an individual makes on the basis of their own
knowledge or experience. It follows that the connotation of a given
word could be different for each individual. However, in practice, the
connotations of words are likely to be shared by a group of people of
the same culture since they have a great deal of similar experience.
Therefore, culturally loaded words are those words whose concepts
exist in both L1 and L2, but the connotations are different, often being
associated with positive/negative feelings or appropriateness/inappro-
priateness. In other words, some words are heavily loaded with
culturally specific connotations.
Through word association studies, Kolers (1963) found that
concrete words tend to produce neutral responses and that abstract
words are more likely to lead to culturally specific responses. It seems
that culture is best reflected at the level of abstract concepts. However,
a very concrete word like bus can have very different connotations from
what in most people’s minds is simply a cheap, convenient means of
public transport; for many Americans and other English speakers of a

18 Chinese belongs to the Sino-Tibetan family and Korean either stands on its
own as an independent family or, arguably, distantly relates to the Altaic family.
89
certain age the word conjures up a political movement in the 1960s in
the USA when children of mixed origin were sent to school in buses in
order to integrate ethnically separated communities. Very few
empirical studies have been carried out on the acquisition of culturally
loaded words. Liu and Zhong (1999) compared Chinese learners of
English and native English speakers on their understanding of six
culturally loaded words by rating the appropriateness of each word. It
was found that most L2 learners demonstrated limited, inadequate
understanding of the culturally loaded words in comparison with
native speakers, and that there existed a fairly large gap of understand-
ing culturally loaded words between advanced L2 learners and native
speakers. It should be noted that the subjects in this study were
learning English in China, without any direct contact with an
English-speaking culture. Similarly, Kupelian (2001) investigated the
acquisition of culturally loaded words by Korean learners of English.
The results generally confirmed those of Liu and Zhong’s study.
The gap in the understanding of culturally loaded words comes
from the difference between the connotations of the words in L1 and
those in L2. For example, the word submissive which is used to
describe a girl’s personality has a negative connotation in English
culture while it is a positive epithet in Korean culture where Korean
wives are expected to be submissive (Kupelian, 2001). Lack of, or
insufficient, contact with the L2 culture is the main reason for the
existence of the cultural connotation gaps. Both Liu and Zhong (1999)
and Kupelian (2001) call for more attention to be paid to culturally
loaded words and for vocabulary to be taught in adequate and
appropriate social and cultural contexts. Unfortunately, this is not
always possible in many countries where an L2 is learned and taught in
the L1 environment.

90
4.2. Linguistic distance

4.2.1. What is linguistic distance?

Many authors would agree that L1 vocabulary has an important

impact on SLVA (e.g. Nation, 1990; Swan, 1997). Studies on learner
interlanguage show that there exist considerable transfer (or borrow-
ing) and interference between L1 and L2 vocabulary. If many of the
features of an L2 word can be predicted from the learner’s L1, the
word will be easy to learn; otherwise, difficulties will be encountered.
Languages can differ in many aspects, including meaning, grammar,
spelling, and phonology. This phenomenon is referred to by many
linguists as linguistic distance (between L1 and L2); it determines to
what extent the two languages differ from each other and, to some
degree, how difficult it is to learn the L2 vocabulary.

4.2.2. Measuring linguistic distance

Since the phenomenon of linguistic distance exists, it is reasonable to

assume that we need to measure how two or more languages are
distant from each other in order to have some kind of index for
predicting the learning difficulty. As Crystal (1987: 371) points out,
while linguistic distance appears to be important in L2 acquisition, it
is almost impossible to establish a linear relationship between
language distance and learning difficulty since it is difficult to
measure the linguistic distance quantitatively due to the complex
structure system of language. However, there appear to be three ways
that can shed some light on measuring the linguistic distance, despite
the dearth of literature on this issue.

4.2.2.1. Approximating the distance via language family trees

A language family is a group of related languages descending from a
common ancestral language. Most European languages are derived
from an ancient but lost language, referred to as Proto-Indo-European
91
(PIE); they form what is known as the Indo-European language family.
Inside the huge Indo-European family, there are small categories or
sub-families, e.g. the Romance family (French, Italian, Spanish, etc.),
the Germanic family (German, English, Dutch, etc.), the Celtic family
(Irish, Welsh, Breton, etc.), and so on. Thus all the languages in this
family can be described as the branches of a tree. From the family tree,
it is not difficult to deduce that English is closer to German than to
Italian while Spanish is more distant from Dutch than from Portuguese.
In this way, we can approximately estimate the language distances
between most languages. However, as Chiswick and Miller (2004)
point out, the measure is only approximate and cannot be accurate
without a quantitative measure.

4.2.2.2. Minimum learning time needs to reach a basic level

It is assumed that the more distant two languages are, the more
difficult it is to learn each of the languages, and thus the time required
to reach a basic level will be longer. From the minimum time required
to reach the basic level of each language, it is possible to rank the
learning difficulty of each language, and thus the linguistic distance.
At the Foreign Service Institute of the U.S. State Department, where
selected high aptitude candidates were required to attain a certain level
of proficiency in another language, day-long immersion courses were
organised, the number of weeks being determined by the estimated
level of learning difficulty. Thus, French, for example, was placed on
the 20-week level, Danish on the 24-week level, while inflectional or
tone languages like Polish and Chinese, where in addition there are
few cognates, were on the highest level of 44 weeks. According to
Odlin (1989), this was the situation in 1985.

4.2.2.3. Measuring the distance by language proficiency

level achieved
This approach is similar to the previous one in attempting to quantify
the distance, not in learning time but in learning result. According to
Chiswick and Miller (2004: 6), the most accurate means of measuring
the distance between languages is to find out how difficult it is for
learners with a given L1 (e.g. English) to learn a series of different

92
L2s (e.g. French, German, Italian, Turkish, Arabic, Japanese, Chinese,
Korean, etc). If it is more difficult to learn Chinese than Turkish then
Chinese is more distant from English than Turkish.
In their study of English-speaking American adults who between
them were learning 43 different languages, the difficulty in learning a
language was measured by the language proficiency they had achieved.
Based on the language achievement in terms of scores after a 24 week
language training period for each of the different languages, a ranking
of language distance between English and other languages can be
obtained, the lowest score being the most distant and the highest the
least distant. The results show that, for English-speaking Americans,
Japanese and Korean are among the hardest or the most distant lan-
guages, followed by Cantonese, Mandarin (Chinese), Vietnamese, and
Arabic, etc.
By estimating the linguistic distance either roughly via the
origins of the two languages or measuring it more accurately in terms of
quantitative data, we have a general idea of how difficult it is to learn an
L2 or the L2 vocabulary, but we still do not know how linguistic
distance affects the learning process. I shall discuss this issue in the
next section.

4.2.3. How does linguistic distance affect SLVA?

The most important aspect of linguistic distance for vocabulary

perhaps consists in the semantic difference between different language
vocabularies. This can certainly, to a large extent, be attributed to the
cultural difference discussed earlier. In place of word-to-word equiva-
lents in two languages, Swan (1997) listed a series of different types
of lexical variation between words in different languages. English
words generally obey the rule of the “arbitrary nature of the linguistic
sign” expounded by de Saussure (1974), namely that meaning is
arbitrarily assigned to a particular word form and each word
represents a concept. The connection between the word meaning and
word form is loose and arbitrary, with the result that the word meaning
is to some degree abstract. On the other hand, the basic Chinese word
is called ‘zi’, which is a character made up of a number of strokes and
93
resembles real objects or depicts a concept in a meaningful way 19. The
Chinese character 火 (pronounced huo, meaning fire), 川 (pro-
nounced chuan, meaning river), 田 (pronounced tian, meaning
farmland) all resemble the real entities they refer to in nature. The
character 从 (pronounced cong) is made up of two 人 (pronounced
ren, meaning person), depicting one person following the other, thus
the meaning of 从 is to follow. The character 囚 (pronounced qiu)
refers to a 人 (pronounced ren, meaning person) who is kept in a
container 口, meaning prisoner. There are more than 80,000 characters
in total, of which 3,500 are commonly used. Each character can be
used independently as an individual word; more often it can be
collocated with other characters and make a new lexical item whose
meaning is determined by the meaning and the order of the two
characters. For example, strike is 罢工 (ba gong) in Chinese, where
罢 means to stop and 工 means to work. Many Chinese words are
constructed in this way. The meaning of most Chinese words is
transparent or self-evident, providing one knows the meaning of each
of the characters that make up the word. Multi-character lexical items
in Chinese are quite common; thousands of four-character idioms or
proverbs not only serve their appropriate functions in communication
but also transmit the culture of the language. In this sense, the Chinese
language is less abstract in meaning, and the meaning of the lexical
items depends on the meaning and order of each character that makes
up the lexical item. Consequently, many Chinese learners of English
find it difficult to handle the abstractness of meaning in English words,
particularly in the early stages of learning.
Swan has pointed out that “mapping second-language vocabu-
lary on to the mother tongue is a basic and indispensable learning
strategy, but also inevitably leads to error” (1997: 179). Many L2
learners intuitively assume that the meaning of an L2 lexical item is the
same as that of the L1 equivalent. The errors L2 learners make fall into
two main categories: interlanguage or intralanguage errors. The

19 Chinese was originally a pictographic language; the oldest Chinese characters

are drawings that depict meanings. Like many other languages, Chinese has
undergone numerous changes and some characters have lost their resemblance
to real objects while others have kept this feature.

94
former, interlanguage errors, are the result of L1 transfer 20; the latter,
the intralanguage errors, are caused by the innate complexity or the
intrinsic difficulty of the L2 itself. When the linguistic distance between
the L1 and the L2 is small as in the case of two closely related
languages, a considerable amount of positive transfer is likely to occur
in learning the L2. When the distance is great, as with two unrelated
languages, there tends to be negative transfer, which, according to
Swan (1997), will result in language errors or avoidance. Regarding
the intrinsic difficulty of L2, e.g. English phrasal verbs, the linguistic
distance between L1 and L2, can, depending on the distance, reduce or
increase the intrinsic learning difficulty of lexical items.
In addition to the semantic difference, the linguistic distance
between L1 and L2 in terms of vocabulary can operate pedagogically at
five levels according to Nation (1990): (1) pronunciation, (2) ortho-
graphy, (3) grammatical patterns, (4) collocation, and (5) frequency.
The distance in these five areas will determine the degree of learning
difficulty of L2 words.

Pronunciation
The L2 words which use sounds and combinations of sounds similar
to those in the L1 will make it less difficult to learn the pronunciation
of L2 words. The consonants /t/, /p/, /s/, exist in many languages, so
English words like ten, pencil, son, will be fairly easy to pronounce
for most learners. In contrast, /th/, as in teeth, think, etc., is generally
considered to be a difficult sound to produce by learners as it does not
exist in many languages.

Orthography
If the learner’s L1 uses the same script as the L2, learning to write the
L2 words will be less difficult. For a native speaker of English, learn-
ing an Asian language like Chinese will be very difficult because of
the totally different script. However, similarity of words raises another
problem. More mistakes occur in spelling for learners of related
languages than of unrelated languages.

20 The transfer can be positive or negative. If the transfer leads to correct use of
L2 forms, the transfer is positive; if not, the transfer is negative.
95
Grammatical patterns
If an L2 word appears in grammatical patterns which are similar to the
patterns where the L1 equivalent occurs in the L1, learning how to use
the word will be less difficult than if the pattern is different. For
example, Chinese students tend to say “I can’t believe!” instead of the
correct utterance “I can’t believe it!” This error results from the fact
that it is not necessary to put an object after believe in the same
context in their L1, Chinese. This is a case of negative transfer: when
a language feature is absent from the L1, the L2 learner often fails to
notice its existence in the L2 or else finds it difficult to use this
feature. For the same reason, along with many other learners of
English, most Chinese students find it difficult to use the English
article the appropriately.

Collocation
If what an L2 word collocates with can be predicted, the learning will
be less difficult. For example, in English we can say: “We talked/
argued/ joked/ spoke about it”, while we cannot say: “We discussed
about it”. We expect words of related meaning to be followed by
similar words, but this is not always the case.

Frequency
If the L1 and the L2 share a lot of vocabulary, like French and English,
there is a high probability of the L2 learner using an L2 word
frequently in the L2 if it occurs in the L1, while in point of fact the L2
word may actually occur much less frequently; thus the L2 word
becomes overused. For example, augmenter is a frequent word in
French, but enlarge or increase will be more appropriate in similar
situations in English, “augment” being reserved for very formal usage
in written language.

96
4.3. Lexical form confusion

Learning difficulty caused by cultural difference and linguistic dis-

tance is largely specific to learners’ L1 or, in other words, associated
with cross-linguistic difference. There is another type of learning
difficulty caused by the target language itself, which is intralinguistic
in nature: learners, regardless of their L1 background, encounter
difficulty with a particular L2 linguistic form either because of its
complexity or because it can easily be confused with a similar form. A
commonly documented difficulty in learning English vocabulary is the
similarity of form among words of very different meaning, such as
staff/stuff, economic/economical, invest/investigate, etc. Not only L2
learners but also native speakers confuse these word pairs, but for
different reasons. For native speakers the confusion is caused by “a
slip of the ear in comprehension and a slip of the tongue in produc-
tion” while for L2 learners the confusion is a result of “a defective
representation of one or both confused items’ forms” (Laufer, 1990:
282−283). Laufer (1990: 284) presented a 10-category framework for
categorizing words with similar forms, which she termed “synforms”
(see Table 1).
In her empirical study, Laufer (1990) obtained the degree of
difficulty, both for native speakers and non-native speakers (with 21
different L1s), encountered in these categories by calculating the
chi-square value x2 of the frequency of the lexical errors made by them
as measured by vocabulary tests. The results show the following
descending order of difficulty: suffixes (categories 1, 2, 3); vowels
(categories 6, 7, 10); prefixes (categories 4, 5); consonants (categories
8, 9). This general pattern holds true for both native speakers and
non-native speakers although within each of the four major categories
there might be a different order.

97
 Category Features Examples
1 Synforms with the same root, considerable, considerate;
productive in current English, but imaginary, imaginative,
with different suffixes. imaginable
2 Synforms with the same root, not capable, capacious;
productive in current English, but integrity, integration
with different suffixes.
3 Synforms with one having a suffix but historic, historical;
not the other sect, sector
4 Synforms with the same root, not consumption, assumption,
productive in current English, but resumption; compress,
with different prefixes suppress, repress
5 Synforms with one having a prefix but passion, compassion;
not the other fault, default
6 Synforms identical to each other affect, effect;
except for one vowel or diphthong set, sat
7 Synforms with one having a vowel cute, acute; quite, quiet;
but not the other date, data
8 Synforms that are identical except for price, prize; extend, extent
one consonant
9 Synforms with one having a ledge, pledge; simulate,
consonant but not the other stimulate;
mean, means
10 Synforms with identical consonants base, bias; manual, menial;
but different vowels embrace, embarrass

Table 1. Ten categories of synforms (assembled from Laufer, 1990: 294).

4.4. Conclusion

We examined three major sources of L2 vocabulary learning difficulty:

cultural difference, linguistic distance and lexical form confusion. The
first and second of these are cross-linguistic in nature while the third is
primarily intra-lingual. Language learning is a culture-bound phenom-
enon; to acquire the L2 vocabulary means acquiring the specific
meaning or lexical concepts in its culture. Linguistic distance between
the L1 and the L2 determines to what extent the two languages differ

98
from each other and to some extent the degree of difficulty in learning
the L2. When the linguistic distance between the L1 and the L2 is small,
in the case of two related languages, a considerable amount of positive
transfer will be likely to occur in learning the L2. When the distance is
considerable, as for two unrelated languages, there tends to be more
negative transfer.
It seems that L2 learners tend to take short-cuts by mapping the
L2 words directly on to their L1 equivalents, regardless of the cultural
difference and the linguistic distance between the two languages. To
prevent learners from doing so, we need to show them that lexical
meaning is very much influenced by the culture of the language and
learning L2 vocabulary requires a reconstruction of lexical concepts in
the mind. Teachers should also help learners to overcome the specific
difficulty in learning different aspects of L2 vocabulary, particularly in
the early stages of L2 acquisition. To do so, teacher experience and
judgement will come into play. Early intervention could prevent
lexical problems which might result in lexical fossilization, thus
making learning less difficult and more efficient.

99
100
 Part II:
Approaches to L2 Vocabulary Acquisition

101
102
Chapter 5. Implicit and Explicit Approaches

5.1. What do these terms mean?

In the area of vocabulary acquisition, some terms are frequently used

to describe the acquisition process, for example, implicit/explicit
learning (N. Ellis, 1994a, 1995a), incidental/intentional learning
(Hulstijn, 2001, 2003). The two terms, implicit and explicit learning,
are often used interchangeably with the other two, incidental and
intentional learning, probably due to similar characteristics of both
learning processes. In addition, the use of implicit/explicit learning
tends to be confused with implicit/explicit memory (sometimes
knowledge). A careful look at these terms as they appear in their
respective fields reveals that they are quite different from each other
and undistinguished use of them will cause confusion to readers. It is
important as a starting point to define them clearly and to describe
their applications in vocabulary acquisition together with their
relationship with each other.

5.1.1. Implicit and explicit learning

Two points have to be made before trying to define these two terms
that originated in the field of cognitive psychology. First, to define
“implicit learning” is usually more problematic than to define
“explicit learning”. In many articles, authors will dedicate lengthy
discussions to implicit learning while giving meagre space to explicit
learning. This is perhaps due to the nature of implicit learning, which
is less observable, elusive and often requires a long time to be noticed.
In contrast, explicit learning is clear, categorical and observable over a
short period. The second point is whether consciousness or awareness
is present or not in implicit learning. This is a controversial issue.
103
According to Reber et al., implicit learning is defined as “the process
whereby a complex, rule-governed knowledge base is acquired largely
independently of awareness of both the process and the product of the
acquisition” (1991: 888). This oft-quoted definition clearly suggests
the absence of consciousness in the learning process. Schmidt argues
that learning without noticing or consciousness is impossible and
implicit learning is “best characterized as the gradual accumulation of
associations between frequently co-occurring features; rather than
unconscious induction of abstract rule systems” (1990: 149).
Dekeyser defines implicit learning as “learning without aware-
ness of what is learned” (2003: 314). This naturally implies that
explicit learning is learning with awareness of what is learned. This
succinct definition has the advantage of avoiding the controversial
issue of the learning process (of implicit learning) and goes directly to
the necessary outcome of the process. Dekeyser draws our attention to
three issues concerning implicit learning: (1) implicitness, (2)
abstractness (which used to be believed to characterize implicit
learning) and (3) methodological problems pertaining to the empirical
research of implicit learning. He suggests that we should be cautious
about assuming that adults implicitly learn structures. He concludes
that implicit learning is more likely to be associated with concrete
rules while explicit learning is more suitable for abstract rules, an
opinion rather contradictory to what was previously held in respect of
implicit and explicit learning. This view seems to be congruent with N.
Ellis’s (1994b) finding that word forms, which are concrete and
factual, are acquired implicitly as a result of frequency of exposure,
while word meanings, which are semantic and abstract, are acquired
through explicit learning systems.

5.1.2. Incidental and intentional learning

Compared with implicit and explicit learning, where the main focus of
research has been on artificial grammar, incidental and intentional
learning are more prevalent in vocabulary acquisition as evidenced by
the vast number of empirical studies in this field. These two terms
were first employed in behavioural psychology, were developed in

104
cognitive psychology and extended to SLA. In the field of psychology,
particularly the studies involving a pre-test – treatment – post-test
design, incidental and intentional learning are defined on the basis of
whether the subjects are told beforehand whether there will be a
post-test or not (Hulstijn, 2003). If they are told beforehand that they
will be tested afterwards, learning is said to be intentional; if not, it is
incidental. In the field of SLA, incidental and intentional learning
have been given a different interpretation. Incidental learning can be
defined as the process of acquiring vocabulary and grammar through
meaning focused communicative activities, such as reading and listen-
ing, and intentional learning as the process involving memorizing
countless words and grammar rules by a variety of means (Hulstijn,
2003: 349).
For L1 vocabulary acquisition, it is generally believed that most
words are acquired incidentally, particularly through extensive reading.
Many studies support this view. Sternberg claims that “most vocabu-
lary is learned from context” (1987: 89); Nagy, Herman and Anderson
(1985) believe that children learn most words through reading and that
they do so incidentally; Krashen’s input hypothesis (1989) postulates
that vocabulary can be acquired by reading as long as the input is
comprehensible to the learner. This view is quite naturally extended to
L2 acquisition. A few studies were conducted to measure the inciden-
tal vocabulary benefit to L2 learners through reading stories (e.g. Pitts
et al., 1989; Day et al., 1991) and significant vocabulary gains were
observed. However, Huckin and Coady (1999) noted a series of
methodological problems pertaining to these studies, including no
treatment of the control group, using intermediate learners while
ignoring low level learners, lack of evidence of long-term retention,
etc. These deficiencies have undermined the value of the positive find-
ings in respect of incidental learning. Other studies have found no
significant vocabulary gain for incidental learning. Horst et al. (1998,
cited in Hulstijn, 2003) concluded from their study that the power of
incidental learning for SLVA might previously have been over-
estimated.
According to Hulstijn (2003), the two main issues concerning
intentional learning are (1) which language (the L1 or the L2) to use to
learn the L2 vocabulary and (2) how to present new lexical items (in
105
context or isolation). Another important issue regarding intentional
learning is the direct or indirect strategies that help to keep new L2
words in memory (e.g. Schmitt, 1997).

5.1.3. Implicit and explicit memory

It has been frequently observed in psychology experiments that

subjects sometimes used information encoded during a learning
session to perform certain tasks without conscious recollection of that
information. Psychologists use “implicit memory” to refer to this phe-
nomenon to distinguish it from “explicit memory” in which encoded
information is consciously retrieved for use. In other words, implicit
memory refers to using stored information without making an effort to
retrieve it, while explicit memory refers to deliberately and conscious-
ly drawing on facts and past experience (Roediger, 2005). Knowing
traffic rules when driving without consciously recalling the rules is a
good example of implicit memory, while describing a recently seen
film is a kind of explicit memory as it needs a conscious memory
search of what is retained about the film. It should be noted that they
are not two types of memory but simply two different retrieval pro-
cesses from memory. They become distinguished only at the moment
of retrieval.
The lengthy studies carried out by N. Ellis (1994b) indicate that
word forms and their related features are best learned implicitly as
procedural knowledge; word meanings are best learned explicitly as
semantic knowledge, part of the declarative knowledge. This should
hold true for both L1 and L2 vocabulary. Implicit knowledge tends to
be retrieved implicitly and explicit knowledge tends to be retrieved
explicitly. Yet this cannot be categorical. As Dekeyser (2003) points
out, explicit knowledge can become implicit when learners lose
awareness of the structure over time. A good example is the case of L2
learners. In the early stage of language learning they often consciously
formulate language in the mind by recalling what they have learned
(words, grammar) before actually uttering the words; as they progress,
these utterances eventually become automatic. On the other hand,

106
implicit knowledge can become explicit when learners become aware
of the structure of the knowledge when trying to retrieve it.

5.1.4. Implicit/explicit learning vs. incidental/intentional learning

The majority of authors do not distinguish between these two pairs of

terms and see them as synonymous: implicit learning equals incidental
learning and explicit learning means intentional learning. A few
authors (e.g. Hulstijn, 2003; Schmidt, 1994) suggest that these terms
be separated as their definitions do not mean exactly the same thing.
The essential feature of implicit learning is learning without being
aware of what is being learned. For incidental learning, learners are
often aware that they are learning something, e.g. trying to infer the
meaning of an unknown word; for instance, if the new word is a
cognate of L1, they might consciously perform some simple mental
association to remember it. This holds true even when the activity
(reading) is meaning focused rather than form focused. The next time
they meet the same word in a different context and are not able to
retrieve the meaning, they do at least know that they have met it
somewhere before. Thus incidental learning can be sometimes quite
explicit. Likewise, explicit learning is learning with awareness, say
comprehending a sentence structure, but it does not necessarily
involve deliberate memorisation of the information.

5.1.5. Implicit/explicit learning vs. implicit/explicit memory

In applied linguistics, authors generally do not distinguish between

implicit/explicit memory and implicit/explicit knowledge as they are
closely connected with each other. To understand the difference be-
tween these two pairs, we need to have some basic knowledge of how
our memory functions. According to Roediger (2005: 1), memory is
the process by which human beings and other animals encode, store
and retrieve information; it is composed of three stages: “Encoding
refers to the initial perception and registering of information. Storage
is the retention of encoded information over time. Retrieval refers to
107
the processes involved in using stored information.” In this three-stage
process of memory, learning is the first stage, i.e. encoding the infor-
mation; implicit/explicit memory is the retrieval process. Therefore,
implicit/explicit learning is how the information is encoded and
implicit/explicit memory refers to how the stored information is
retrieved.

5.1.6. The meaning of implicit/explicit approaches adopted

in this book

Approaches to vocabulary acquisition can be generally categorized

under two paradigms: the implicit and the explicit learning/teaching
paradigm. Two points have to be clarified. In the first place, the
meaning of “implicit” and “explicit” is not restricted to what they
mean in “implicit learning” and “explicit learning” which originated
from cognitive psychology; rather, the literal meaning of the two is
used to refer to the main features associated with the two paradigms.
Implicit learning is associated with natural, effortless and meaning-
focused learning; explicit learning implies that learning requires
deliberate mental effort (as opposed to simply engaging in meaning
focused activities) and a link has to be established between meaning
and form by various means. Secondly, learning includes teaching, as
direct teaching of vocabulary can be very useful (Nation 1990, 2001;
Coady 1997), the way teachers teach can influence students’ decision-
making in adopting specific approaches to vocabulary learning, and,
finally, learners can be their own teachers.

5.2. The implicit learning/teaching paradigm

The basic assumption of the implicit learning/teaching paradigm is

that words can be acquired naturally through repeated exposures in
various language contexts with reading as the major source of input.

108
This notion is heavily rooted in the findings for L1 vocabulary
acquisition. I will focus on learning in this paradigm as teaching
vocabulary implicitly can scarcely be defined 21. Incidental learning is
the most important approach for this learning paradigm; it is a mean-
ing focused reading or listening activity in which some vocabulary
may be learned as a by-product. The majority of studies deal with
written contexts, with only a few studies being based on oral input (e.g.
Wode, 1999; Ellis/He, 1999), although it is very likely that words can
be acquired incidentally in this fashion.
A popular view is that limited classroom teaching time simply
cannot deal with the thousands of words needed by the L2 learner;
learning through large quantities of reading materials, particularly
chosen by learners themselves, is an alternative solution. It is implied
that vocabulary learning in this fashion is an effortless process, a view
that is very attractive to learners. Huckin and Coady (1999: 182) sum-
marized the advantages of incidental learning over explicit instruction
as follows:

(a) It is contextualised, giving the learner a richer sense of a word’s use and
meaning than can be provided in traditional paired-associate exercises, (b) it is
pedagogically efficient in that it enables two activities – vocabulary acqui-
sition and reading – to occur at the same time, and (c) it is more individualised
and learner-based because the vocabulary being acquired is dependent on the
learner’s own selection of reading materials.

The studies on incidental vocabulary learning can be broadly categor-

ised into two types: on the one hand, those dealing with word
inferences; on the other, those providing lexical glosses or advocating
the use of a dictionary.

21 Some might argue that teaching word roots and affixes so that students can
use this strategy for learning words independently is a kind of implicit
vocabulary teaching activity. However, I regard vocabulary strategy instruc-
tion as explicit teaching, which is in line with Coady (1997).
109
5.2.1. Word inferencing studies

In the first type, learners are simply provided with a written text and
are encouraged to infer the meaning of the unknown words from the
contextual clues. It is considered that vocabulary learning retention
depends on the nature of the lexical processing activities that learners
perform during the reading process instead of on other factors such as
learning intention. The more mental effort the learner makes in
inferring the meaning of an unknown word, the better it will be
retained (e.g. Hulstijn, 1992; Mondria/Wit-de Boer, 1991). These
assumptions are an application of Craik and Lockhart’s depth of
processing model of memory (1972). However, in order to infer the
meaning effectively, learners need to recognise most of the words
surrounding the unknown one. It is generally accepted that learners
need to know 95% of the lexical items of the text for general
comprehension (Nation, 1990; Laufer, 1997). Knowledge of 98% of
the lexical items is needed in order to be able to guess the unknown
words accurately (Coady et al. 1993; Hirsch/Nation, 1992; Laufer,
1997). Huckin and Coady (1999) concluded that the vocabulary
knowledge learners need for reading comprehension ranges from a
minimal threshold level of the 3,000 most frequent families to 10,000
word families. This figure is hardly attainable by most L2 learners,
especially those students below intermediate levels, who constitute a
very large proportion of L2 learners. Coady (1997: 229) put his finger
on the problem when he raised the question: “How can they learn
vocabulary through extensive reading when they do not know enough
words to read well?”
Another important issue concerning inferencing is that learners
need to master some effective guessing strategies, such as graphemic
identification according to word forms or use of broader contextual
meanings (Huckin/Coady, 1999). In addition, some strategies need to
be explicitly taught to learners, e.g., “cognate monitoring” (i.e.

110
checking for false cognates) and “forward clues”22 which are under-
developed by low level learners (Huckin/Coady, 1999: 187).
Although word inferencing is useful and will lead to some
acquisition, there are serious limitations innate to guessing itself.
Huckin and Coady listed the following (1999: 189): (1) guessing cannot
guarantee precision; (2) correct guessing requires that the word be
recognised correctly and monitored carefully; (3) guessing is time-
consuming and will interrupt the reading process; (4) guessing is
effective on condition that the context is adequately comprehended
with most of the words in the vicinity being known; (5) guessing
requires good reading strategies which many students do not have; (6)
guessing often does not lead to acquisition; (7) guessing is of little help
to the acquisition of multiword items. Taking a dozen examples at
random, Kelly (1990) demonstrates that this time-consuming activity
will seldom lead to the exact meaning when the context is uncontrolled.
Inaccurate guessing may even have an anti-learning effect, i.e. result in
the fossilization or interpretation of the wrong meaning of a word.

5.2.2. Lexical glosses and dictionary use studies

In the second type of studies, learners are engaged in reading texts

(paper versions or electronic versions) with some marginal vocabulary
glosses (e.g. Hulstijn et al., 1996) or else they are provided with
electronic glosses or a dictionary (e.g. Knight, 1994; Laufer/Hill, 2000;
De Ridder, 2002), or simply with a paper version dictionary (Hulstijn
et al., 1996). Sometimes, control groups, who are only given the
reading texts, are compared with the experimental group who receive
lexical aids in addition to the texts. In such cases, the group with
lexical aids usually outperforms the control group who rely solely on
contextual guessing or form guessing. In Knight’s study, subjects who
used an electronic dictionary not only had a higher vocabulary gain
but also did better on a comprehension test. She points out that

22 They are the contextual clues following the word to be guessed, which are the
opposite of “backward clues” which refer to the contextual clues preceding the
word to be guessed.
111
dictionary use is particularly useful for low verbal students: they
achieve results on both vocabulary retention and comprehension
tantamount to those of high verbal ability students. Hulstijn et al.
(1996) obtained similar results when learners who accessed marginal
glosses or a dictionary achieved higher vocabulary retention than
learners without lexical aids. The latter group either failed to infer the
correct meaning of the unknown words or simply ignored the words. It
is also reported from their study that marginal glosses will be more
effective for vocabulary learning than a dictionary (paper version) as
learners tend to make little use of a dictionary 23. Therefore Hulstijn et
al. (1996) suggest that dictionary use during the reading process
should be facilitated by providing marginal glosses or easy-to-access
electronic glosses (in the case of electronic texts) so that learning can
be more efficient. In fact, as early as the late 1970s, Ostyn and Kelly
were already advocating marginal glosses in place of what they
described as time-wasting dictionary searches and using them in their
text books (e.g. Tune In) for secondary school and university English
learners.
Electronic reading texts with glosses or annotations seem to be
very popular at the present time. The advantage of providing electron-
ic glosses is that the lexical information can be accessed easily by a
click (or by typing the word) with little interruption of the reading
process. Moreover, glosses are made more attractive and informative
by multimedia effects than traditional lexical information entries. For
a glossed item, the learner can view the textual explanation, listen to
the pronunciation, request a picture or even watch a short video about
the target word. This type of gloss with multimedia effects provides
learners with powerful lexical aids which enable them to engage in
more elaborative mental activities in respect of the various features of
the lexical item than those traditional lexical entries in dictionaries.
However, as Wesche and Paribakht (2000) point out, the goal in
typical incidental learning is text comprehension and only an approxi-
mate meaning for target words is needed; sometimes learners could
use world and topic knowledge to compensate for their lexical gaps

23 Hulstijn et al. found that when learners do use a dictionary, their vocabulary
retention will be as good as or even better than learners with marginal glosses.

112
without paying attention to specific lexical items. In addition, infer-
ring the meaning of a word, instead of simply providing that meaning
(by paper dictionaries or electronic glosses/dictionaries), means that
the learner’s attention is focused on the meaning and not on the form.
For L2 learners, meanings for the L2 words already exist in some way
in their minds as similar concepts to those in the L1. It is invariably
the L2 word forms which are new and that require more learning
effort; more importantly, it is essential to connect the word form with
the meaning and fix it in memory in the initial stage (Kelly, 1986).

5.2.3 The main drawbacks of incidental learning

The main problems of acquiring vocabulary from incidental learning

seem to be attributable to three sources:

1. Incidental learning inevitably involves a great deal of contextual

guessing of the unknown words. Context alone does not always
facilitate meaning transfer; in some cases even educated adults
cannot infer the meaning of some L1 words in context (Ames,
1966; Beck, McKeown/McCaslin, 1983, cited in Duquette et al.
1998), let alone L2 learners. Even in the situation of consulting
a dictionary or glosses of the unknown words, learners tend to
forget very quickly more than half of the words processed in
reading (Hulstijn et al. 1996).
2. As a consequence, the retention rate of genuine incidental learn-
ing24 is very low (e.g. Hulstijn, 1992). According to Nation
(1990), 5–16 exposures are needed to fully acquire a word;
Nagy et al. (1985) reported a 5%–15% probability of a word
being learned at first exposure; Knight (1994) demonstrated a
learning rate of 5%–21 % from her studies.

24 In some incidental learning cases, there is some treatment in addition to the

reading tasks, for example, pre-sessions aiming to improve learners’ lexical
processing strategies (Fraser, 1999). The learning rate is reportedly higher
than for those incidental learning studies without any treatment.
113
3. The vocabulary acquired through incidental learning mainly
leads to recognition and hardly to production (see also Wesche/
Paribakht, 2000; Paribakht/Wesche, 1997). This is due to the
nature of incidental learning: it is meaning focused activity and
only limited attention is paid to the word forms and lexical/
syntactic features of the new words. The quality and quantity of
lexical processing in incidental learning is simply insufficient to
enable the learner to acquire words at the productive level.
Moreover, learning new lexical items does not stop at the
recognition or the production level; there are other aspects of
learning. Wesche/Paribakht (2000: 197) write:

Learning a new word involves an ongoing elaboration of knowledge about the

word and the ability to use it. Relationships are established between the word
form and its semantic concepts and linguistic functions, as well as with other
words that share some of these features, forming lexical networks.

5.3. The explicit learning/teaching paradigm

Advocates of this paradigm argue that vocabulary and vocabulary

learning strategies should be learned or taught explicitly so that
learning can be more efficient. They agree with upholders of inciden-
tal learning that context is the major source for acquiring vocabulary,
but they claim that learners need some extra help to build up an
adequate vocabulary and to acquire strategies to cope with the vast
reading context (e.g. Coady, 1997). Coady (1993) argues for a mixed
approach to vocabulary acquisition in SLA. He puts it this way: “The
basic or core vocabulary should be taught, but the less frequent
vocabulary will then be learned ‘naturally’ via context; but, even in
that case, techniques for that purpose should be taught” (1993: 17).
The approaches adopted by the explicit learning paradigm can be
grouped into two main subcategories: reading plus exercises and
explicit instruction/learning.

114
5.3.1. Reading plus exercises

While acknowledging the importance of vocabulary acquisition

through extensive reading in an incidental manner, this approach aims
at combining reading comprehension with text-based vocabulary exer-
cises. Previous studies suggest that such text-based exercises could
encourage deeper word processing and might enhance the incidental
learning effects from reading. Two prominent studies were carried out
by Paribakht and Wesche (1997) and Wesche and Paribakht (2000).
In their first study (1997), the main question they investigate is
whether reading comprehension plus text-based vocabulary exercises
leads to the more effective acquisition of selected vocabulary items
than reading comprehension alone when learning time is equal for
both conditions. In the reading only condition, a text which contains a
number of target words is presented to learners. Then two thematically
related texts, which also contain the target words, are read by these
learners. In the reading plus exercises condition, learners read only the
text and then are required to carry out eight vocabulary exercises on
the target words. The exercises are of five types: (1) selective attention,
(2) recognition, (3) manipulation, (4) interpretation and (5) production.
They aim to draw learners’ attention to target words, to recognise the
word by associating the word form with the meaning, to perform a
structural analysis of the target words, to carry out a semantic and
syntactic analysis of the target words, and to produce the target words
in new contexts. Vocabulary gains were measured using the Vocabul-
ary Knowledge Scale (Wesche/Paribakht, 1996). The results reveal
that both conditions achieved significant gains. However, the gains of
the reading plus condition are quantitatively superior to those of the
reading only condition in terms of number of words gained and
qualitatively superior in terms of the increased depth of knowledge of
the given words. In other words, vocabulary acquired by learners in
the reading condition largely remains at the recognition level whereas
learners in the reading plus condition learn more words and achieve a
higher level of vocabulary knowledge.
In their second study (2000) they tried to explore how such
text-based vocabulary exercises could promote different types of
lexical processing and offer an explanation for the superior gain of the
115
reading plus condition. They used a follow-up introspective research
method to discover what learners are actually thinking while engaged
in the vocabulary exercises. They replicated the previous study using
only the reading plus condition. First learners are trained to use a
think-aloud strategy to verbalize their thoughts when carrying out
learning tasks. Then they read a text and immediately start the eight
vocabulary exercises while saying aloud what they are thinking and
doing; everything that they say is recorded on a tape recorder. The
recording is then transcribed. On the basis of these transcripts, six
possible explanations are put forward to account for the superior
results obtained by adding vocabulary exercises to reading: the
exercises (1) make more target words salient to learners, (2) encourage
learners to explore target words independently, (3) motivate learners
to learn the target words, (4) lead to considerable re-reading of the text,
(5) encourage the use of the target words and (6) make learners aware
of their imprecise knowledge of the target words and thus of the need
to do the exercises.

5.3.2. Explicit instruction/learning

Authors who favour explicit instruction argue that L2 learners should

be taught vocabulary explicitly by using various techniques including
direct memorisation (Coady, 1993; Nation, 1990, 2001). Here the
concern is mainly with low level learners who do not have enough
vocabulary to read extensively. According to Nation (2001), to master
the large number of words that native speakers know can be a
long-term goal but is not suitable as a short-term goal. His reasoning
is based on frequency studies (2001: 11−21). He demonstrates that the
2,000 most frequent words can cover 80% of the text. Another 570
words that frequently appear in academic texts can cover about 9% of
the text. There are about 1,000 technical words which cover approxi-
mately 5%. The rest are the enormous number of low frequency words
(15,000−20,000) that make up 5% of a text. The first three types of
words, i.e. the high frequency words, academic words, and technical
words can together cover almost 95 % of the text; these words are thus
very important and worthy of study. Both teachers and learners should

116
pay enough attention to them in different forms: direct teaching
(teacher explanation, peer teaching), direct learning (using word cards,
using dictionaries), incidental learning (contextual guessing, com-
municative activities) and planned encounters with the words (graded
reading, vocabulary exercises). As for the low frequency words, the
teacher should train learners to use strategies such as contextual
guessing, dictionary use, memory techniques and vocabulary cards to
cope with these words and to enlarge their vocabulary.
Nation (2001) insists that intentional learning should be seen as
complementary to incidental learning rather than as opposed to it and
suggests that the intentional study of vocabulary items could take up
to 25% of the total learning programme. He delves deep into three
intentional word study strategies: studying word parts, using diction-
aries and using word cards. Word parts study includes affixes and
roots. He notes two notable benefits in having a knowledge of affixes
and roots: (1) it can be useful for learning unknown words by
associating them with those known words or known affixes and roots
of words25; (2) it can help learners to check whether an unknown word
is correctly inferred or not. Dictionary use is a kind of intentional
study of words and some skills are needed specifically for receptive
and productive use of a dictionary. The information obtained from a
dictionary could strengthen what has been already learned and prepare
for further encounters with the words. However, the value of diction-
ary use is limited and learners should not have very high expectations
of it. For Nation, the most effective intentional way of learning words
is learning through word cards. He argues that (a) word cards could be
very useful for remembering words; (b) using word cards can help to
learn other important features (except word use) of a word and these
include word meaning, word form and connecting the two together,
which are prerequisites for using a word; (c) learning words,
particularly those high frequency words, intentionally, is a necessary
step in learning words incrementally and this will assist the incidental
learning of other words. Mondria and Mondria-de Vries (1994) are
also strong advocates of the card system for learning vocabulary as an

25 We can note in passing that making use of this type of knowledge was already
being advocated by Keller three decades ago (1978).
117
alternative to lists. They show how, in accordance with memory theory,
it can be used systematically and to maximum advantage.

5.4. Conclusion

In the discussion of the implicit and explicit approaches to L2 vocabu-

lary acquisition, it emerges that implicit learning, or mainly incidental
learning, would not lead to satisfactory vocabulary acquisition for
three reasons: (1) the risk of wrongly guessed meaning, (2) the low
learning rates, and (3) the acquisition would be recognition rather than
production. On the other hand, the explicit learning/teaching paradigm
is best summarized as adopting a “mixed approach”, to use Coady’s
words (1993: 17). Supporters of this paradigm combine explicit
vocabulary instruction, vocabulary exercises, vocabulary learning
strategies, and extensive reading. The strength of the explicit learning
paradigm is that implicit learning is not excluded but rather is seen as
a necessary complementary approach to vocabulary acquisition. The
two approaches would work best in combination with each other.

118
Chapter 6. Approaches to Multi-Word Items

6.1. Background information

6.1.1. Language: lexis or grammar based?

The traditional Chomskyan notion of language suggests a clear

distinction between grammar and the lexicon. Grammar or structure is
seen as superordinate and vocabulary subordinate, governed by the
former. Research in psycholinguistics dealing with language pro-
cessing (memory) shows this is not true. The lexicon plays a much
more important role than used to be thought and the distinction
between lexis and grammar turns out not to be clear-cut. The main
reason is that many phrases such as off with his head “appear to be
halfway between traditional syntax and lexicon, for they cannot be
stored as invariable units, nor can their structure be derived from
traditional grammar rules of syntax since they deviate so from normal
English sentence patterns” (Nattinger, 1980: 338). According to
Mel’čuk (1998: 24), the ratio of single items and multi-word items is
roughly 1 to 10 for most languages. It is shown that such prefabricated
patterns or combinations of words are a dominating feature of native
English speakers’ speech and are in quantity much more significant
than free phrases governed by grammar rules (Pawley/Syder, 1983).
This view is firmly supported by the rapidly growing research in
native English corpora (e.g. Sinclair, 1991). It would seem to be wise
to address L2 acquisition from a lexis-centred standpoint by helping
learners build up a store of lexical items and chunks and getting them
to use these items appropriately. Nattinger (1980: 341) expresses this
need as follows:

119
 [...] language production consists of piecing together the ready-made units
appropriate for a particular situation and […] comprehension relies on
knowing which of these patterns to predict in these situations. Our teaching
therefore would centre on these patterns and the ways they can be pieced
together, along with the ways they vary and the situations in which they occur.

These “ready-made” units have been given different names, such as

lexical phrases (Nattinger/DeCarrico, 1992; Schmitt/Carter, 2000),
lexical chunks (Lewis, 1993), lexicalised sentence stems (Pawley/Syder,
1983), collocations (Nation 2001), formulaic sequences (Wray, 2002;
Schmitt, 2004), etc. Following Wray and Schmitt, I shall call them
formulaic sequences.
They have some general characteristics: (a) they are typically
multi-word items of varying lengths and frequently occur in native
speakers’ speech; (b) they are often associated with the functional
usage of language; (c) they exist somewhere on a grammatical-lexical
continuum and have a more idiomatically determined meaning than
individual words which can be recombined each time (Nattinger/
DeCarrico, 1992). Central to all these features is the fact that
formulaic sequences, being two (or more) word collocations or
complete utterances, express a single meaning or a specific function.
Wray (2000, 2002) assigns functions of formulaic sequences to two
major categories: (1) achieving successful interaction and (2) saving
mental processing effort. Common formulaic sequences include
idioms, phrasal verbs, fixed collocations and many other longer
strings of words.

6.1.2. Corpora

The most powerful evidence for the existence of formulaic sequences

comes from corpora: large data bases of written and spoken language
(often kept in electronic form). According to Kenney (1998: 4), the
distinction between a corpus and a text archive/database is that the
former is designed for linguistic analysis and is structured system-
atically for that purpose whereas the latter is usually unstructured.

120
6.1.2.1. Native English corpora
There are three main types of corpora, monolingual corpora, multilin-
gual corpora26, and learner corpora. Important native-speaker corpora
include the COBUILD Bank of English Corpus, the Cambridge
International Corpus (CIC), the International Computer Archive of
Modern and Medieval English (ICAME) corpora and the British
National Corpus (BNC). These monolingual corpora (English) can be
used to identify typical English phrases in real use and ones that
characterize native-like speech. These features can inform course book
designers and teachers of what language phrases and patterns to be
incorporated into the teaching materials. In addition, native corpora
provide lexicographers with the best data for producing dictionaries
that capture vivid language in real life, e.g. the Collins COBUILD
dictionary series.
Schmitt (2000) identifies three major types of information
provided by native corpora to inform SLA. These are: (1) how
frequently words occur, (2) how words occur together, and (3) how the
structure of language is organised. By looking at the frequency of
words in corpora, we have a general idea of what words occur most
often and what words occur rarely. This provides valuable information
for selecting suitable vocabulary for learners, particularly beginners or
low-level learners. Willis (1990) describes the Collins COBUILD
course as a lexical syllabus intended for beginners, which contains
about 2500 words selected from the 20 million words of the
COBUILD corpus27. Level 1 contains the 700 most frequent words,
which account for 70% of all English text; level 2 contains 1500
words which account for 76 % of text; level 3 contains 2500 words
which account for up to 80% of text. Nation (2001) also stresses that
vocabulary selection in course books should be based on frequency
studies.

26 This type of corpora contains texts of several native languages. If the same
texts are written in different languages, then it is called Parallel corpora, e.g.
the Corpus Resources And Terminology Extraction Corpora (CRATER), the
Multilingual Text Tools Corpora (MTLTEXT).
27 Recent statistics show that the corpus contains more than 500 million words
and is constantly growing.
121
 By looking at corpora, we quickly discover that certain words
(two or more) tend to co-occur frequently in the form of collocations.
Sinclair’s (1991) distinction between the open-choice principle and
the idiom principle offers a useful framework in describing how words
collocate with each other. The open-choice principle explains that
words can be collocated freely, e.g. read/write a book/article/thesis;
the idiom-choice principle reveals that certain words are semantically
restricted to other words, e.g. rancid butter, torrential rain, etc. While
in the case of the former both words are open to accept other company,
in the case of the latter, often one word strongly predicts the word that
will accompany it. The popular current view is that collocations
operate on a continuum ranging from free collocation to pure idioms
(Cowie, 1998). Most authors recognise two types of collocation:
grammatical/syntactic collocations and semantic/lexical collocations.
Grammatical collocations are typically composed of a dominant word
(a noun, verb, or adjective) followed by a preposition, e.g. stick to,
damage to, familiar with, etc. Lexical collocations are made up of two
(or more) words of equal status, such as noun-verb phrase, ice melts,
verb-noun phrase, comb the hair, and adjective-noun phrase, wonder-
ful time.
In addition to collocations, typically revealed in corpora is the
enormous number of formulaic sequences which contain language
structures. It is this type of language that has been long ignored in lan-
guage acquisition. Up till recently, it has been very rare for publishers
to incorporate these phrases and patterns into course materials. By
using corpora, teachers or learners can identify these phrases and
patterns. Both collocations and formulaic sequences can be identified
by a computer program (e.g. WordSmith by Scott, 1996; MicroConcord
by Scott/Johns, 1993; Collocate by Barlow, 2004) which shows con-
cordances: these enable us to quickly find out how a target word is
collocated and used with other words. Online concordance search of
corpus is also available, e.g. the free sample search of BNC: <http://
sara.natcorp.ox.ac.uk/lookup.html>; the free sample search of
COBUILD corpus: <http://www.collins.co.uk/corpus/CorpusSearch.
aspx>. In addition, Tom Cobb’s Compleat Lexical Tutor (version 4)
from <http://www.lextutor.ca/> provides an online comprehensive
concordance program with multiple functions.

122
6.1.2.2. Learner English corpora
Learner corpora contain written or oral English language by learners.
They are often recorded in the form of computer learner corpora (CLC)
marked with error tagging. The most comprehensive and influential
learner English corpus is the International Corpus of Learner English
(ICLE) developed at the Catholic University of Louvain (UCL),
Belgium. ICLE contains more than two million words of written texts
by learners of English of 19 different mother tongue backgrounds. In
addition, a spoken corpus named Louvain International Database of
Spoken English Interlanguage (LINDSEI) is progressively being
developed as a complement to the written corpora. Research into
learner corpora generally falls into two categories: contrastive inter-
language analysis and computer-aided error analysis (Granger, 2002:
11−12). Research of the first category is the search for inter-lingual
features, either universal or L1-specific, by comparing learner English
with native English or different non-native learner English data. The
second category is concerned with finding, tagging and analyzing
language errors with the help of computer software/tools. The findings
of learner corpora research have certainly a huge potential in their
application to L2 acquisition (see Granger, 1998a and Granger et al.
2002 for a review).

6.1.3. Identifying and defining formulaic sequences

A broad definition of formulaic sequences is provided by Wray: “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: 9). The defin-
ition is fully inclusive, focusing on two basic features of formulaic
sequences: (1) a string of words with or without slots, and (2) holistic
storage and retrieval. According to these two criteria, an enormous
number of multi-word items can be considered as candidates for
formulaic sequences. On the other hand, the definition is so encom-
passing that we are left with only a vague idea of what formulaic

123
sequences are. According to Wray, two major problems in the study of
formulaic sequences are how to identify these sequences in discourse
and how to define their features.
Wray proposes five ways of identifying them (2002: 20−43): (a)
intuition, (b) computer frequency counts, (c) internal structure analysis,
(d) phonological form, and (e) special language data. Each can be
used to identify formulaic sequences to some degree but each has its
drawbacks. Thus Wray proposes to look for a series of features that
formulaic sequences may possess but not necessarily so. One possible
way is to categorise formulaic sequences according to certain criteria.
Different proposals have been put forward, e.g. Nattinger and
DeCarrico (1992), Lewis (1997). According to Wray (2002: 47), these
proposals are based on one or more of the four features of formulaic
sequences: form, function, meaning, and provenance28. She (2002: 65)
argues that all four features are interrelated and do not operate
independently:

It is provenance that explains the quirks of both form and meaning, and it is
function and meaning which propel a sequence through the process from
novel to formulaic. […] a focus on form is able to capture some descriptive
characteristics of formulaic sequences, but that approach is weakened by
unbalanced but ubiquitous formal characteristics which cut across the main
categories, and/or deep-seated difficulties in excluding function, meaning and
provenance from the finer points of the account.

In short, there are no clear-cut categories that allow all formulaic

sequences to be neatly classified. Often a formulaic sequence can fall
into more than one category.
In a similar vein, Schmitt and Carter (2004) point out that it is
difficult to identify categorical criteria for defining formulaic
sequences and propose to search for archetypal characteristics of

28 Provenance refers to how sequences of language become formulaic. Some

new sequences are formulaic at the start but it requires time to internalize
them, to treat them as a whole; others become formulaic gradually and even
acquire a different meaning from what they meant in the beginning, e.g. the
Radio Times was originally the name of a British magazine which gave details
of the week’s forthcoming radio programmes but now it has become a for-
mulaic expression with more extended associations (Wray, 2002, p. 60−61).

124
formulaic sequences. The two authors (2004: 7−8) list the following
features as specific to formulaic sequences: (1) they may be stored as
a whole in the mind but they are acquired in a way similar to
individual words which involve incremental learning; (2) they can
have slots which have semantic constraints, e.g. ____ thinks nothing
of ____ typically needs to express some unexpected idea; (3) they
often entail certain semantic connotations, e.g. border on implies the
meaning of entering into an unwanted state as in border on the
ill-mannered; (4) they frequently have certain types of functions, e.g.
social interaction.
It would seem that for Wray and others such as Schmitt and
Carter, formulaic sequence is at best defined in a broad, all-encom-
passing manner. Further tracing of the subcategories or defining
features is possible but would lead to fuzzy, confused results. It is
clear that Wray gave this global, undistinguished definition largely
from a psycholinguistic viewpoint in consideration of mental pro-
cessing and formulaicalisation of the sequences.
Howarth (1998), however, objects to this undistinguished single
category of formulaic sequences and argues that formulaic sequences
are actually gradable, ranging from free combinations to pure idioms,
all covered by the term “word combinations”. The main categories of
these word combinations are shown in Figure 6. Functional expres-
sions have some functional purposes in discourse and they include
many institutionalised interactive utterances, proverbs, slogans, etc.
Composite units consist of word collocations which can be further
divided into two sub-categories: grammatical and lexical composites.
The former are collocations composed of words of an open class and a
closed class (usually a preposition) (e.g. in time, proud of ); the latter
consist of words of two open classes (e.g. make an appointment, final
objective). The distinction between non-idiomatic and idiomatic
combinations is made for each sub-category: functional expressions,
grammatical composites and lexical composites. However, the
distinction is not categorical but operates on a continuum: free
combinations (e.g. blow a trumpet), restricted collocations (e.g. blow a
fuse), figurative idioms (e.g. blow your own trumpet), and pure idioms
(e.g. blow the gaff) (Howarth, 1998: 28). The main features that
characterize these categories are (Howarth, 1998: 28):
125
 Free combinations (also referred to as open or free collocations) consist of
elements used in their literal senses and freely substitutable (carry a trumpet,
on top of the table). Restricted collocations have one component (usually the
preposition, verb, or adjective ‘collocator’ of the ‘base’ noun, to use
Hausmann’s 1979 terms) that is used in a specialized, often figurative sense
only found in the context of a limited number of collocates. While figurative
idioms have metaphorical meanings in terms of the whole and have a current
literal interpretation, pure idioms have a unitary meaning that cannot be
derived from the meanings of the components and are the most opaque and
fixed category.

word combinations

functional expressions composite units

non-idiomatic idiomatic grammatical lexical

composites composites

non-idiomatic idiomatic non-idiomatic idiomatic

Figure 6. Phraseological categories (Howarth, 1998: 27).

Restricted collocations are still gradable and can be distinguished at

three levels: level 1 allows restricted substitution of both sides (e.g.
introduce/table/bring a bill/an amendment); level 2 allows limited
substitution of one side (e.g. pay/take heed); level 3 is the most
restricted and similar to idioms (e.g. curry favour). Howarth (1998: 42)
claims that it is restricted collocations that pose the most learning
problems29 for learners while both free collocations and idioms are
comparatively less problematic.

29 According to Howarth, one main reason is that the restricted collocations are
prone to be produced by learners in non-standard or erroneous variations.

126
 The phraseological categories of Howarth are useful in pro-
viding a clear framework for pedagogical purposes – to guide the
instruction of formulaic sequences. We have to be first aware of all the
possible types of word combinations in terms of their collocability and
find out the difficulty they cause learners, then we can adopt a
particular approach to tackle the acquisition of a given category of
formulaic sequences. For example, Boers and his associates (2004)
formulate their approaches to the acquisition of one type of formulaic
sequence: idioms. There is, however, a dearth of research in such
category-specific approaches to the acquisition of other types of
formulaic sequences.

6.2. Approaches to L2 formulaic sequence acquisition

6.2.1. Learning difficulties of L2 formulaic sequence acquisition

Although our understanding of the nature of formulaic sequences has

deepened in the last two decades, how formulaic sequences (both the
L1 and the L2) are acquired is still an underdeveloped area that needs
more research. Often it is assumed that the acquisition of L1 formulaic
sequences is similar to that of individual words, going through an
incremental process from vague knowledge to possibly full mastery
(e.g. Schmitt/Carter, 2004). It follows that L2 formulaic sequences
should be acquired similarly to the L2 individual words, though the
acquisition process could be more complicated due to the nature of
formulaic sequences. As with the acquisition of individual words,
there is a range of learning difficulty that L2 learners will encounter in
acquiring formulaic sequences:

1. Formulaic sequences are typically culture-loaded, their acquisi-

tion being subject to L1 influence. If we agree that language
mirrors the culture of the linguistic community, then this should
particularly hold true for formulaic sequences. Adult L2

127
 learners are prone to transfer their L1 knowledge in acquiring
the L2, including formulaic sequences, and this could lead to
overuse or underuse of formulaic sequences, even for advanced
learners whose L1 is closely related to the L2 (Granger, 1998b;
De Cock, 1998).
2. The meaning of formulaic sequences is often non-transparent;
analysis of semantic or syntactic features simply does not lead
to the correct meaning, e.g. idioms. This means that acquiring a
formulaic sequence requires the commitment of whole strings
of words to memory, which involves a considerable amount of
mental processing and the strings are subject to memory loss. In
other words, it is more difficult to link the sequence form with
the sequence meaning than for individual words.
3. Acquisition of formulaic sequences needs to balance formulai-
city and creativity of language (Wray, 2002: 183). L2 learners
make considerable errors when using formulaic sequences due
to a range of interlanguage or intralanguage sources. In addition
to the phenomena of underuse and overuse, learners can create
their own formulaic sequences due to previous wrongly
acquired grammatical/semantic knowledge.
4. Most L2 learners acquire the L2 (including formulaic sequences)
in an unfavourable learning environment. It is reported that lan-
guage teaching materials rarely incorporate formulaic sequences,
or else treat them as of little importance. Formulaic sequences
might be scarce or absent from teachers’ input if the language
teachers are not native speakers themselves. Most learners have
little opportunity to have social interaction with native speakers,
which is considered a vital condition for acquiring formulaic
sequences (e.g. Wray, 2002; Schmitt, 2004; Dörnyei et al. 2004;
Adolphs/Durow, 2004).

Due to these areas of difficulty, the acquisition of L2 formulaic

sequences becomes extremely difficult, which explains why many
advanced learners either avoid using them or use them wrongly. There
seems to be no particularly efficient way of tackling them. There are,
however, a few points that might facilitate formulaic sequence acquisi-
tion for adult L2 learners: (1) noticing these sequences in the language

128
input (e.g. Wray, 2002; Schmitt et al. 2004), (2) interacting with native
speakers or speakers who can correctly use formulaic sequences (e.g.
Wray, 2002; Dörnyei et al. 2004), (3) analyzing the sequences to some
degree (e.g. Willis, 1990; Wray, 2002, 2000). As for the third point,
Wray (2000: 494) quotes from Willis and points out that the successful
acquisition of formulaic sequences for L2 learners involves a
compromise between “native-like expression” and “non-native-like
processing”, although not all formulaic sequences can lend themselves
to analysis. Moreover, she (2000: 484) believes that the successful
acquisition of formulaic sequences in classroom settings consists in:

[...] the sensitivity of the syllabus designer and the teacher to the complex
nature of language in use and, in particular, to the potential for the very idio-
maticity of an expression to make it less open to generalization than it may
seem at first glance. The more natural the data that are being presented to the
learner, the more they need to be subjected to control and guidance in
delivery.

6.2.2. Proposals for the acquisition of formulaic sequences

6.2.2.1. Classroom teaching/learning activities

This type of approach values classroom activities and exercises for
teaching/learning lexis. It assumes that teachers play an important role
in selecting what type of formulaic sequences to teach and what
activities to use. Nattinger and DeCarrico (1992) suggest that different
teaching activities should be organised according to the type of dis-
course: conversation, listening, reading and writing. They recommend
a progressive learning process from routines to patterns to creative
language use for learning formulaic sequences orally. They put it this
way: “[…] by starting with a few basic fixed routines, which they
(learners) then would analyse as increasingly variable patterns as they
were exposed to more varied phrases” (1992: 116).
According to the two authors, traditional pattern practice drills
could be used to help learners to gain fluency with some basic fixed
sequences as a starting point. This is followed by introducing learners
to controlled variation in these basic sequences by means of simple
129
substitution drills. This suggests that at this stage it is better to choose
to teach formulaic sequences that contain several slots, e.g. I’m (very)
sorry to (hear (about) X), (1992: 117) instead of rather fixed expres-
sions which allow little variation, e.g. I’d like to express my sympathy
(about X) (1992: 117). Regarding the teaching of formulaic sequences,
they recommend exchange structure, which requires a series of
question-response utterances between two learners relating to a
specific topic. They give the following example (1992: 121):

(8) Asserting-endorsing
I read/heard somewhere that X – Yes, that’s so/right/correct.
I think/believe that …. – I absolutely/certainly/completely agree.

6.2.2.2. Syllabus design

Willis (1990), in his pioneer work The Lexical Syllabus, offers the
rationale and design for lexically based language teaching and
learning. According to him, syllabus design is set in a communicative
approach, which offers learners the opportunity to abstract knowledge
from the language exposed to them; it will focus on specific language
items to identify what is to be learned. The lexical syllabus design
reaches a balance between a holistic approach and a discrete one.
Willis (1990: vi−vii) identifies the following basic principles for the
syllabus design: (1) only language activities involving real language
will be used; (2) learners will be exposed to only authentic native
speaker language; (3) spontaneous speech will be recorded for course
use but without transcriptions; (4) learners will be encouraged to
analyse and learn the language to which they are exposed instead of
being presented with the language. The syllabus design is applied to
the Collins COBUILD English Course series, which forms part of the
COBUILD research project on lexical development. A mini corpus,
which contains words up to the 2500 word frequency level in context,
is extracted from the 20 million COBUILD corpus. Based on the mini
corpus, the course is divided into three levels: level 1 contains the
most frequent 700 words, level 2 the most frequent 1500 words and
level 3 the most frequent 2,500 words. The meanings and uses of the
words will be highlighted when exposing learners to the language.

130
Learners will then do exercises based on the language to which they
have been exposed, and this will lead to acquisition.

6.2.2.3. The lexical approach

The lexical approach focuses on developing learner proficiency with
lexis: words, word combinations and particularly formulaic sequences.
Lewis (1993, 1997), inspired by the innovative work of authors such
as Willis (1990) and Nattinger and DeCarrico (1992), made the
following ideas the basis of what he named the “lexical approach”:

 The grammar-lexis dichotomy is not valid and is replaced by a

continuum along which many lexical-grammatical items are
situated between the two poles of lexis and syntax.
 Language is full of prefabricated patterns and chunks which are
stored and retrieved as a whole.
 In SLA learners need to be trained to recognise these lexical
patterns and various ways could be used to master them.

Materials and resources to support the lexical approach

One of the prerequisites to implementing the lexical approach is to
have teaching materials and resources available for both teachers and
learners. Richards and Rodgers (2001) identify four types of teaching
materials and resources for the lexical approach. The first type
contains complete course packages (texts, tapes, teacher’s guide, etc.).
The Collins COBUILD English Course developed by Willis and
Willis (1989) is such a case. The second type is a collection of lexical
teaching activities/exercises proposed by Lewis (1993, 1997, 2000a).
The third type is made up of a printed version of computer corpora in
the form of text. The last type is computer corpora attached to con-
cordancer programs so that learners can perform their own analysis
with the teacher’s help or independently. Resources of this type are
often on CD-ROM or can be downloaded from websites.

Implementing the lexical approach

Lewis (2000b) has attempted to formulate a learning theory for the
lexical approach. He holds the view that the lexical approach is similar
131
to the natural approach (Krashen/Terrell, 1983) in that both ap-
proaches value comprehensive input. He also states that the two differ
in that the lexical approach acknowledges the conscious noticing of
linguistic features of input, such as lexical chunks and patterns, and
thus has a facilitative value, whereas the natural approach does not
have such a value. According to Lewis (2000b: 184), acquisition is
associated with a series of important features: (1) new learning items
have to be met and comprehended on several occasions; (2) the lexical
chunks or collocations have to be noticed; (3) similarities, differences,
restrictions and examples which all contribute to turning input into
intake have to be noticed; (4) acquisition is based on an accumulation
of examples from which learners make preliminary generalizations
rather than on the application of formal rules. Only when these
conditions are fulfilled can learning lead to acquisition.
Regarding classroom teaching practice, Lewis (2000b) rejects
the traditional Present-Practice-Produce paradigm and proposes an
alternative Observe-Hypothesis-Experiment paradigm. Observe means
learners must meet and notice the new language; hypothesis means
learners have to sort out the input in some provisional way;
experiment involves using the language based on learners’ current
interlanguage. In the classroom, the teacher’s role should be changed
from that of instructor to that of learning manager: the teacher should
help learners notice useful lexical chunks and patterns (e.g. by
providing real, natural collocations 30 ), guide learners’ choice of
materials and activities and, most importantly, maintain learners’
motivation. See Lewis (2000a, 1997) for a review of the classroom
activities for implementing the lexical approach. It would seem that
the lexical approach advocated by him can be summarized as an
input-noticing-intake-output paradigm, in line with Krashen, the only
difference being noticing between input and intake.

30 According to Lewis, real examples of collocations are words that naturally

occur together; the invented examples are used to illustrate traditional
grammar structure and often the vocabulary does not fit well with the
grammar structure.

132
6.2.2.4. Data-driven learning
This approach focuses on using computer concordancers to process
and analyse large quantities of language data (computer corpora) to
learn language and particularly formulaic sequences. Johns and King
give the following definition of data-driven learning (DDL): “[…] the
use in the classroom of computer-generated concordances to get
students to explore regularities of patterning in the target language,
and the development of activities and exercises based on concordance
output” (1991: iii). DDL are meant to arouse learner language
awareness and raise their consciousness of the language features to be
learned (Granger/Tribble, 1998: 200). Learners are encouraged to
search for language patterns rather than being presented with the
language patterns to be learned by rote. These assumptions largely
overlap with those of Willis and of Lewis. As seen above, using
concordancing programs to process language data has become an
important means of teaching lexis in the lexical approach. However,
students will be bored if classroom instruction is dominated by
concordancing programs when they are exposed to authentic but
rather discontinuous language (in which a number of words will be
shown on each side of a particular lexical item or formulaic sequence).
There is no information provided in respect of the broad language
context where (some of) the selected sentences or phrases actually
occur. In this sense, DDL based on concordancing programs and
language corpora should be best employed as a kind of awareness-
raising or follow-up information-search activity, complementary to
other communicative language activities. Another important issue is
that exercises or language activities need to be carefully designed to
assist the learning of the target formulaic sequences in order to make
the best use of DDL. In addition, DDL should not be restricted to
classroom learning as it was originally defined by Johns and King (see
above); it can also be extended to students’ self-learning, i.e. to use
concordancing programs to search for information on their own
(Cheng, 2004: 96). Teachers need to train students to use the con-
cordancing programs and to perform efficient searches in the corpora.
The extent to which DDL can in fact be useful in learning
formulaic sequences or language in general depends on a series of
133
conditions: (1) how students perceive DDL in terms of usefulness and
preference; (2) what type of exercises are designed based on the
language data to which they are exposed; (3) whether students have
sufficient language proficiency to understand the authentic language
data; (4) what teachers do to help the students to explore DDL, etc.

6.2.2.5. Conclusion
The four theoretical approaches tackle the acquisition of formulaic
sequences from different perspectives. Willis aims to set up the
rationale for designing syllabuses catered to learning formulaic
sequences. Nattinger and DeCarrico provide practical guidance for
teachers on how to treat formulaic sequences in the classroom and
acknowledge the importance of teachers presenting learning items and
controlling the learning process. Lewis formulates a theory for the
lexical approach in which the acquisition of “grammaticalized” lexis
and formulaic sequences form the centre of SLA. DDL emphasizes the
role of using language corpora in demonstrating formulaic sequences
and how they are used. One thing all four approaches have in common,
as Wray (2000) points out, is that they support the linguistic analysis
of formulaic sequences at certain stages.
Willis and Lewis share a similar viewpoint, arguing that
authentic language input provides the major source of acquisition and
that noticing different language patterns and sequences is the pre-
requisite for their acquisition. Both seem to suggest that acquisition
will occur naturally when learner attention is repeatedly directed to the
target formulaic sequences along with some analysis or generalization
of the pattern. Neither is interested in the issue of how to commit the
large number of formulaic sequences to memory, one that Boers and
Lindstromberg (2005: 226) raise. These two authors point out that the
major challenge in acquiring formulaic sequences is to turn noticed
language exposure into intake, which calls for the need to develop
strategies and techniques to do so. The last approach, DDL, explores
the potential of electronic authentic language corpora with the help of
concordancing programs; however, I argue that DDL is best employed
as complementary to other language activities and that its efficiency is
largely situation-dependent.

134
6.2.3. Empirical studies on the acquisition of formulaic sequences

In this section, we look at several recently reported empirical studies

in a special volume devoted to formulaic sequences edited by Schmitt
(2004); the acquisition can be grouped into two settings: classroom
and socio-cultural interaction.

6.2.3.1. Classroom settings

Schmitt et al. (2004) report a study of the acquisition of 20 selected
formulaic sequences in an intensive course for English for Academic
Purposes (EAP) for a period of two to three months. The selection of
the sequences is based on frequency, occurrence in an EAP context,
and usefulness to students. The instruction of these items is
“semi-controlled” (Schmitt et al., 2004: 53) by taking into account that:
(1) students will meet at least once each sequence in the learning
materials; (2) teachers will direct students’ attention to each sequence
during the course without revealing the research purpose. They
reported that students, who already had considerable pre-knowledge
of the target items, achieved a receptive learning rate of 12% and a
productive learning rate of 24.9%. The fact that the productive gain
(3.20) is higher than the receptive gain (2.03) may be due to the fact
there was a ceiling effect in the receptive score but not in the
productive scores (Schmitt et al., 2004: 65). Since these students are
studying in an English-speaking country, they have the opportunity to
interact with native speakers, which provides extra input for formulaic
sequences. Therefore, the source of the knowledge gain is not very
clear; it could be the explicit instruction, the input of the course book,
or the natural L2 exposure. As for teacher instruction, no detailed
information is given except that teachers are free to choose what they
do with sequences.
If the study by Schmitt et al. did not provide a specific method-
ology for teaching formulaic sequences, the study by Jones and
Haywood (2004) does so. They attempt to tackle formulaic sequence
acquisition in EAP (English for Academic Purposes) classes in a
systematic way by presenting a clearly defined methodology for
instruction. They adopt a mixed approach: Nation’s (2001) three
135
psychological stages for vocabulary acquisition – noticing, retrieving
and generating – serve as the guidelines. Concordance texts are used
to enrich the language input as well as to illustrate the linguistic
features of the sequences; special attention is paid to help students
build up useful learning strategies. Instruction is given in a reading
course and a writing course, two hours a week over a period of ten
weeks.

The reading course

After the reading text is comprehended, special attention is paid to
formulaic sequences by highlighting them in the text. Students are told
clearly that they should study certain multi-word items and try to use
them in writing in order to acquire a more academically suitable style.
Students are required to perform activities, with the help of con-
cordance texts and corpus extracts, which will extend their knowledge
of the meaning and usage of the sequences.

The writing course

In the writing course, students can review the sequences or patterns
learned in the reading course and prepare for productive use. Learning
activities are diversified: students are assigned writing tasks with the
help of a planning session in which relevant or frequent formulaic
sequences will be rehearsed or introduced; they are required to analyse
the function of formulaic sequences; they are instructed to produce
their own concordance texts with the help of a program.
The measurement of the acquisition is based on (1) the degree
of awareness of formulaic sequences, (2) the ability to produce formu-
laic sequences on a cloze test, and (3) the free production of formulaic
sequences in writing essays. The results are mixed. By the end of the
study, most of the students have shown increased awareness of the
formulaic sequences, i.e. words not only exist as individual items but
also frequently appear in meaningful and useful sequences. A
moderate gain is reported from the cloze test but there is no significant
improvement in free production of formulaic sequences in essays.
According to the two researchers, the main problem is the relatively
small number of subjects, ten in all, which can hardly make the results
generalisable. Nevertheless, this study shows us how formulaic

136
sequences can be taught and learned in a systematic and meaningful
way, combining DDL and other traditional means such as learning
strategy training, in contrast to the prevailing previous view that they
will be taken care of naturally with language exposure.

6.2.3.2. Socio-cultural interaction

Dornyei et al. (2004) investigate the acquisition of formulaic
sequences in terms of three main factors: language aptitude, moti-
vation and, particularly, socio-cultural adaptation/integration. They
state their assumption as follows (2004: 87):

[…] the acquisition of a formulaic, phraseological competence is somewhat

different from the mastery of other components of communicative language
proficiency in that formulaic language is so closely linked to the everyday
reality of the target language culture that it cannot be learnt effectively unless
the learner integrates, at least partly, into the particular culture.

Seven students were selected from the subjects in Schmitt et al.’s

study (2004, see above). These students were chosen because they
were either considered extremely successful or unsuccessful learners
of formulaic sequences. A series of regular interviews with the
selected subjects were conducted to collect qualitative information
which covered (1) learner attitude towards the local people, (2)
language learning attitudes and beliefs, (3) motivation, (4) learner
perceived progress or regression and its possible reasons, (5) degree of
social integration. The results are reported based on the interview data
in conjunction with scores obtained from an aptitude/motivation
survey. Their study revealed some tentative evidence that “success in
acquiring formulaic sequences is strongly related to the learners’
active involvement in some English-speaking social community”
(Dornyei et al., 2004: 104). However, it is also shown that the
majority of students failed to adapt or integrate into the host culture
for all sorts of reasons, e.g. cultural gap, lack of commitment, anxiety
about not being able to express themselves, etc. Among the four
successful students, two managed to break through the cultural barrier
and achieved successful socio-cultural integration to some degree. The
other two did not but simply had higher motivation and worked harder.
137
We should not forget that most L2 learners do not have the opportuni-
ty to go to the L2 country and, even when they can, many, as the study
shows, cannot be integrated into the local culture. Without denying its
effectiveness, solely relying on socio-cultural integration as a major
means to acquiring formulaic sequences does not seem to be very
realistic for the majority of L2 learners.

6.3. Approaches to the acquisition of idioms

Idioms are a specific type of formulaic sequence largely due to their

opaqueness and figurative use (see Howarth 1998 for a discussion of
criteria for defining idioms). It used to be believed that idioms are
“‘dead’ metaphors whose meaning has become completely arbitrary”
(Boers, 2004: 211). Boers, together with his associates, in a series of
articles (Boers 2000, 2004; Boers/Demecheleer, 2001; Boers et al.
2004), argues that idioms are actually “motivated” rather than
arbitrary and their acquisition can be tackled in a systematic manner.
Two approaches to the acquisition of idioms have been developed by
Boers and his associates: a cognitive approach and a mnemonic
approach. It should be noted that these two approaches can be com-
bined and complement each other.

6.3.1. The acquisition of idioms by raising learner awareness

of idioms using a cognitive approach

According to Boers (2000), the figurative use of idioms can often be

traced to their literal meaning, e.g. see a light at the end of the tunnel,
and learner attention can be directed to these literal meanings when
acquiring idioms. Boers (2004) identified three ways of raising learner
metaphor awareness of figurative language based on empirical studies.
First, often idioms “can be traced back to a relatively small set
of concrete ‘source domains’ whose structure is mapped on to our

138
conception of abstract ‘target domains’ via ‘conceptual metaphors’”
(Boers et al. 2004: 55) and these source domains can be made explicit
to learners. They (2004: 55) give an example to illustrate how idioms
can be grouped together: You still have a whole life ahead of you, She
needs moral guidance, Without you, I’d be lost, Follow my example,
We’ve reached the point of no-return, etc., can be grouped under the
source domain “travelling” via the conceptual metaphor LIFE IS A
JOURNEY. Other identified conceptual metaphors include: THE
BODY IS A CONTAINER FOR EMOTIONS, ANGER IS A HOT
FLUID IN A CONTAINER, MORE IS UP, LESS IS DOWN, ACTIVE
IS UP, INACTIVE IS DOWN, etc. (Boers, 2000). Boers et al. (2004)
refer to these as image-schema-based metaphors grounded in general
physical experience and they are likely to be universal across different
cultures and languages. Boers (2000, 2004) argues that organizing
idioms according to these common metaphoric themes can facilitate
their retention. Explicit instruction is needed to raise learner aware-
ness of such knowledge; this includes making learners aware that (1)
metaphor is ubiquitous in language use and (2) figurative language is
not arbitrary but can be grouped under different themes (Boers, 2000).
Second, many idioms are also derived from specific source
domains, e.g. fighting/warfare, health/fitness, food/cooling, games/
sports, agriculture/gardening, handicraft/manufacturing, boats/sailing,
entertainment/public performance, religion/superstition, and com-
merce/accounting (Boers et al., 2004: 59). These specific source
domains are often culture specific, e.g., idioms associated with hats
and shipping feature more prominently in English than in French
(Boers/Demecheleer, 2001). If learners are instructed to trace idioms
to their original source domains, learners are likely to generate a
mental picture. This, according to Boers, can, on the one hand, involve
deep mental processing and, on the other, lead to “dual coding” (see
Paivio/Desrochers, 1980) of the linguistic information. Thus, the
acquisition of idioms can be facilitated. A program called Idiom-
teacher is being developed on this basis (Boers et al. 2004). Three
types of exercises are designed for each idiom to be learned: (1)
comprehending the idioms in a short context with multiple-choice
questions; (2) identifying the source domain of the idioms with
multiple-choice questions; (3) gap-fill exercises for recalling the
139
idioms. Students were asked to do all these exercises in a sequential
order. The general recall rate was 72.85% while the comprehension
rate was 70.04% at the start of the exercises. It is difficult to calculate
a precise learning rate since the study was not designed for this
purpose. However, the results do show that identifying the origin of
idioms can help their productive acquisition to some degree and this
can hold true for both semantically transparent and less transparent
idioms. It is found that idioms can either be successfully identified
with the source domain or, in the case of unsuccessful attempts, their
origin or, to use Boers’ word, etymology can be given to learners.
Both are equally helpful for retention.
Third, metaphor awareness can be enhanced with problem-
solving tasks, i.e. inferring the meaning of idioms with or without
context before having recourse to help. Some idioms are imageable;
they are defined as “idioms that have associated conventional images”
(Lakoff, 1987, cited in Boers/Demecheleer, 2001: 255) and learners
can easily guess their meaning even without the help of context, e.g.
split one’s sides laughing. The study reported by Boers and
Demecheleer (2001) shows that their subjects were able to produce a
correct guessing rate of 35% for decontextualised idioms of medium
degree of semantic transparency. The two authors propose that
contextual clues should be used for idioms with lower imageabililty. It
is also pointed out that a major problem for inferring the so-called
imageable idioms largely depends on the difference between the L1
and the L2 culture. Related languages such as English (L2) and French
(L1) may share many common sources for idioms and a target idiom
might be easily imaged, but this might not be the case for a more
distant L2. Second, even in the case of two related languages, learners
might make an erroneous transfer from L1 to L2.
In a more recent article, Boers and Lindstromberg (2005: 227)
provide a four-stage instruction to show how idioms should be tackled:

If, for example, a target expression is buy a pig in a poke, students can (1) be
told the current meaning of the expression, (2) be given the information that
poke is an obsolete word for “sack”, (3) be asked to visualize a pig in a sack,
and (4) speculate about how the expression came to have its overall, idiomatic
meaning.

140
This instruction is a combination of cognitive learning and mnemonic
techniques, referred to by the two authors as “etymological elabor-
ation” (2005: 227). It can be seen that the underlying essence of this
approach is to use images, either concrete or metaphorical, to enhance
the encoding and memorisation of idioms.

6.3.2. The mnemonic effect of alliteration on the acquisition of idioms

This approach explores the potential of phonological patterning in

acquiring idioms. Vocabulary items can be related in many ways in the
lexicon, one of which is their phonological relation. It is very likely
that similar phonological patterns among items within an idiom can
facilitate retention. In particular, Boers and Lindstromberg (2005: 227)
examine a specific type of phonological pattern, alliteration, which
they define as “word-initial repetition of at least one consonant (e.g.,
time will tell; spick and span)”. They hold the view that this
mnemonic strategy can be particularly useful for learning idioms that
are not suitable for etymological elaboration. The important role of
alliteration in learning idioms is based on two assumptions. First,
according to the two authors, alliteration features in about 17% of all
the items listed in the Oxford Dictionary of Idioms. If alliteration is an
important feature in English idioms, it is worth considering exploring
its potential to assist in their acquisition. Second, they cite evidence
and argue that it was through alliteration/rhyme/rhythm that large
pieces of formulaic, figurative, poetic texts, such as the Iliad and
Odyssey, were memorised and passed on in oral form before being
converted into written form. From the results of their empirical studies,
it is repeatedly shown that retention of idioms with alliteration is
significantly higher than those without alliteration.

141
6.5. Conclusion

In this chapter we have given our attention to formulaic sequences, a

type of lexical item which frequently gives rise to considerable
learning difficulty. Implicit learning in which learners are exposed to
rich language input and have social interaction with the target
language culture seems to be the predominant view for learning them.
The various types of corpora that have been built up in recent years
have been of enormous help in identifying them and demonstrating
their very high frequency in everyday language. These corpora have
helped researchers to classify them and, on account of the concordan-
cing programs that have been developed from them, are gaining
popularity in classroom instruction or self-learning, providing a useful
complement to other learning activities. Meanwhile, explicit instruc-
tion is being investigated by some authors. Boers and his colleagues
have proposed seemingly effective approaches combining cognitive
learning (essentially associated with images) and mnemonic tech-
niques (alliteration) for the learning of a particular type of formulaic
sequence, idioms. Their potential for other types of formulaic
sequences needs to be further researched 31. What seems to be needed
is the search for effective approaches specific to other types of
formulaic sequences within the framework of Howarth (see 6.1.3.).

31 In addition to idioms, these approaches (cognitive or combination of cognitive

and mnemonic) have been applied to the acquisition of phrasal verbs (e.g.
Boers 2000; Condon/Kelly 2002).

142
Chapter 7. A Learner-Focused Approach

7.1. Learner differences

The 1970s saw the placing of the language learner at the centre of the
learning process by taking into account a series of learner characteris-
tics, needs, abilities, and interests (Wenden, 2002). The concept of
learner centredness has fundamentally shaped language teaching/
learning as well as L2 acquisition. It is reasonable to assume that the
ultimate goal of learner centredness is to achieve learner autonomy, i.e.
give the learner the ability to take charge of his/her own learning. To
achieve this goal, the learner needs to develop certain skills and
capabilities as it is rare for learners to display an innate ability to take
effective control of their own learning. However, as Wenden (2002)
observes, changes brought about by learner centredness have mainly
affected teaching and the curriculum design, which might have
benefited the learner but were not necessarily intended to improve the
learner or directly develop the learner’s ability to learn. She thus calls
for “learner development”, which she defines as “a learner-centred
innovation which intends to help learners learn how to learn” (2002:
34). It follows that a learner-focused approach has to first foster
learner development and then possibly achieve learner autonomy. This
is true for general language learning and vocabulary learning, the
latter being the most important constituent of the former.
Research into learner differences has aimed to: (1) provide
valuable information to inform language teaching; (2) explain how
language is learned and predict language proficiency; (3) promote
learner autonomy. The last-mentioned might be the hardest to achieve
as the process is usually fairly long and progress can be slow and
complex. Important learner differences that affect L2 acquisition
include affective factors (motivation, attitudes), cognitive factors
(aptitude, intelligence), personality, learning style, learning strategies,
143
age, and gender, etc. According to Cohen and Dörnyei (2002),
motivation, learning style, learning strategies are comparatively easy
to manipulate while other factors, such as age, gender and aptitude,
are largely beyond control. Following them, I shall focus on motiv-
ation, learning style, and learning strategies.

7.2. Motivation

7.2.1. What is motivation?

Motivation as a subject of scientific study was first understood

through Freud. Human actions or behaviour can be explained by inner
forces, such as instincts, drives, emotional states and psychic energy,
largely from a biological viewpoint. Nowadays, much has been done
to improve our understanding of motivation in a variety of fields in-
cluding behavioural, social, cognitive, affective, conative 32, spiritual33
perspectives (Huitt, 2001). Motivation is an omnipresent phenomenon
in our daily lives and can account for most of our behaviour and
actions. Because of the very broad coverage of motivation in terms of
variables, it is difficult to reach a consensus among researchers as to
what exactly motivation is (Dörnyei, 2001). It seems that most authors
would agree that motivation is an internal state or condition, such as a
need, desire, or want. Huitt (2001: 1) makes a tentative effort to define
motivation:

 Internal state or condition that activates behavior and gives it

direction.
 Desire or want that energizes and directs goal-oriented behaviour.

32 The conative perspective of motivation includes reaching one’s goals in life,

taking control of one’s life and developing or maintaining self-efficacy, etc.
33 The spiritual perspective of motivation includes understanding the purpose of
one’s life and relating oneself to ultimate unknowns.

144
 Influence of needs and desires on the intensity and direction of
behaviour.
 The arousal, direction, and persistence of behaviour.

7.2.2. Motivation for learning the L2

Most authors would agree that motivation is a leading L2 learner

variable and that the degree of motivation determines the degree of
language achievement. Dörnyei writes: “Motivation provides the
primary impetus to initiate learning the L2 and later the driving force
to sustain the long and often tedious learning process” (1998: 117). An
earlier classic definition of L2 motivation from a socio-psychological
approach is given by Gardner (1985: 10):

Motivation in the present context refers to the combination of effort plus

desire to achieve the goal of learning the language plus favorable attitudes
toward learning the language. That is, motivation to learn a second language is
seen as referring to the extent to which the individual works or strives to learn
the language because of a desire to do so and the satisfaction experienced in
this activity.

He distinguished two types of orientation that serve as goals of L2

motivation34: integrative and instrumental orientation. Learners are
classified as integratively oriented if they emphasize meeting and
talking with native speakers and endeavour to acquire a better
understanding of the L2 culture; they are classified as instrumentally
oriented if their goal in learning the L2 is to get a better job or to be
better educated. He attached particular importance to the integrative
motive, which he defined as the motivation to learn an L2 due to the
positive feelings towards the people that speak the language. The key
components of the integrative motive model are: (1) integrativeness,
(2) attitudes towards the learning situation, and (3) motivation.

34 Gardner’s motivation theory is the interplay between motivation and

orientation, which are viewed as two distinct components, the latter being the
goal of the former. In other motivation theories, e.g. the goal theory (Maslow,
1970), goals are central components of motivation.
145
Integrativeness includes integrative orientation, interest in the L2, and
positive attitudes towards the L2 community. Attitudes towards the
learning situation consist of attitudes towards the language teacher and
the language course. Integrativeness and attitudes towards the learning
situation give a driving force to motivation, which is reflected in the
desire to learn the L2, the effort made, and the attitudes towards
learning the L2. Gardner’s socially grounded motivation theory set up
scientific research procedures and introduced standardized assessment,
using the Attitude/Motivation Test Battery, which provides firm
ground for research on L2 motivation.
From the early 1990s, there has been a call for a practical,
education-centred approach to L2 motivation, which would be applied
in classroom teaching/learning and would include new findings from
mainstream psychology (Dörnyei, 1994). Since then, a considerable
number of L2 motivation studies have been conducted to investigate
how learner motivation is reflected in concrete classroom events and
processes (Dörnyei, 2002). L2 motivation thus moves from the
Gardnerian approach which was predominantly social and psychologi-
cal to an education-centred approach. Dörnyei (2001: 106) views this
educational shift in thinking as a “peaceful revolution” for two reasons.
First, it never has the intention of abandoning what has been pre-
viously established but aims to widen its scope by including additional
factors that could explain specific language learning tasks and learner
behaviour in classrooms. Second, those who had developed the
previous theories took an active part in this revolution, acknowledged
the existence of the new shift, and contributed to it.
In line with this educational shift, Dörnyei (1994) has evolved a
framework to conceptualise L2 motivation in terms of three different
levels: (1) language level, (2) learner level, and (3) learning situations
level. The three levels correspond to the three basic constituents of the
L2 learning process, namely, the L2, the L2 learner, and the L2
learning environment. The language level essentially includes the two
Gardnerian concepts: the integrative and the instrumental motivation
construct. The learner level involves various individual characteristics,
such as need for achievement and self-confidence. The learning situa-
tion level comprises three types of components: (a) course-specific
motivational components which refer to learners’ perceptions of the

146
syllabus, the teaching materials, the teaching method and the learning
tasks; (b) teacher-specific motivational components, which are
concerned with the motivational impact of the teacher’s personality,
behaviour and teaching style; (c) group-specific motivational
components, which involve the learner group characteristics such as
goal-orientedness, norm and reward system, group cohesion, and
classroom goal structure.
While remaining in line with the Gardnerian tradition, Dörnyei
and Ottó (1998) introduce a new notion in motivation research by
suggesting the temporal, dynamic nature of motivation. They (1998:
65) define L2 motivation as follows:

In a general sense, motivation can be defined as the dynamically changing

cumulative arousal in a person that initiates, directs, coordinates, amplifies,
terminates, and evaluates the cognitive and motor processes whereby initial
wishes and desires are selected, prioritised, operationalised and (successfully
or unsuccessfully) acted out.

While the Gardnerian approach usually set motivation research in a

broad context by finding out learners’ overall attitudes towards the L2,
the dynamic view of motivation is more suited to the educational
approach by explaining learners’ behaviour in a specific learning
situation. It should be noted that the definition given above attempts to
describe the process by which motivation is generated, executed, and
completed, as well as geared to a particular learning task. This is
particularly true when we consider that L2 motivation usually needs to
persist over a relatively long period of time, sometimes many years. It
is very likely that learner motivation will change in terms of intensity
of desire, effort and attitudes during this period.
Moreover, it is argued (Dörnyei/Ottó, 1998; Dörnyei, 2001,
2002) that motivation is best described as a dynamic, cyclic process
which undergoes three different phases, named the process model:

1. Motivation is generated when the learner’s desire/wish/hope to

achieve something emerges. This decision leads to the selection
of the goal or the learning task to be carried out. Intention
involving commitment needs to be formed in order to initiate

147
 the goals. In addition, a feasible plan to carry out the task needs
to be developed at the start. The whole of this initial phase is
referred to as “choice motivation”. The motivational influences
which empower this phase include the learners’ perceived value
of the task, their incentive for doing the task, attitudes towards
the L2 and its community, expected success, learner beliefs and
strategies, etc.
2. The motivation generated needs to be effectively maintained
and protected; this is termed “executive motivation”. During
this phase, the learner will generate and carry out subtasks,
constantly evaluate the external stimuli and the progress made,
and control/monitor actions by self-regulatory mechanisms
(such as knowledge and strategies). The most important motiv-
ational influence of this phase is quality of current learning
experience (pleasantness, need significance 35, coping potential,
self and social image); other influences on motivation include
teachers, classroom reward structure, learner groups, knowledge
and use of self-regulatory strategies (e.g. goal setting, learning
and self-motivating strategies).
3. The task is completed and followed by a retrospective evalu-
ation of the whole process, named “motivational retrospection”.
By reflecting on the past learning experience, the learner will
determine the next task to be carried out. The main motivational
influences of this phase are attributional factors (the way people
explain their previous successes or failures affects their future
behaviour), self-concept beliefs, and received feedback (e.g.
praise, grades).

It should be noted that though the three phases delineate a cyclic

process, they are not necessarily in a linear sequence. Dörnyei states
that “the ‘choice’ phase of one actional step might happen simulta-
neously with the executive phase of another, resulting in complex
interferences” (2002: 142).

35 This refers to whether the environmental stimuli a learner receives are helpful
in meeting the learner needs or achieving goals.

148
7.3. Style

7.3.1. What is learning style?

It is observed both in educational psychology and L2 acquisition that

different learners approach learning very differently and these
different approaches are covered by the term “learning style” (Cohen/
Dörnyei, 2002). According to Oxford (2003: 273), learning style is
“the general approach preferred by the students when learning a
subject, acquiring a language, or dealing with a difficult problem”. For
example, some learners might prefer reading books; others might love
to listen to lectures or to move about constantly while studying. These
learning styles are overall patterns that give general direction to
language learning. So far more than 20 different dimensions of learn-
ing styles have been identified from three research traditions: (1) the
study of perception and Gestalt psychology, (2) ego psychology, and
(3) Carl Jung’s theory of personality (Ehrman/Oxford, 1990: 311). To
understand learners’ learning styles is one of the first steps in taking
into account learner difference in classroom learning or curriculum
design. Learning style is biological in nature. Dunn and Griggs put it
this way: “Learning style is the biologically and developmentally
imposed set of characteristics that make the same teaching method
wonderful for some and terrible for others” (1988, cited in Oxford,
2001: 359).
This implies that learning styles are relatively stable and rather
difficult to be manipulated. It is suggested that language teachers
should attune learning activities to match learners’ specific learning
style preferences (Oxford, 2001; Cohen/Dörnyei, 2002). On the other
hand, preference for certain types of learning styles will lead to the
neglect of other styles that might be helpful to different aspects of lan-
guage learning. Learners sometimes need to go “beyond the stylistic
comfort zone” (Oxford, 2001: 361) and try to incorporate other
beneficial learning styles. Teachers should encourage their students to
stretch their learning styles (Cohen/Dörnyei, 2002) or provide a wide
range of communicative activities (Oxford, 2001) to help them do so.

149
7.3.2. Learning style assessment tools

To find out what styles learners prefer, a suitable type of assessment

tool has to be used. The most common type of assessment is a written
questionnaire survey and learners are asked to answer questions which
will reveal the general picture of their style preference. So far, the
most widely used style survey instruments seem to be the Myers-
Briggs Type Indicator (MBTI) (Myers/McCaulley, 1985), the Per-
ceptual Learning Style Preference (PLSP) Survey (Reid, 1998) and the
Style Analysis Survey (SAS) (Oxford/Nam, 1998).

7.3.2.1. The Myers-Briggs Type Indicator

The Myers-Briggs Type Indicator (MBTI) was first developed by
Isabel Briggs Myers (1897−1979) and her mother Katharine Cook
Briggs, on the basis of Carl Jung’s psychological types. MBTI reflects
the theory that human behaviour in daily life is a manifestation of
underlying stable and unchanging preferences for certain ways of
functioning. The two authors’ initial goal was to help people under-
stand themselves and others so that they could choose the profession
that matched their personality types. Personality types are also
extended to language studies and they appear to be significantly
correlated with language learning success; these include a sense of
humour, achievement orientation, assertiveness, outgoingness, im-
pulsivity, risk-taking, adventurousness, flexibility, tolerance of
ambiguity, etc. (Ehrman/Oxford, 1989). MBTI describes four
dimensions of personality in keeping with Jung’s terminology: (1)
extraversion vs. introversion, (2) sensing vs. intuition, (3) thinking vs.
feeling, and (4) judging vs. perceiving. Ehrman and Oxford (1990: 327)
have summarized them in the following way:

1. The extroversion vs. introversion scale describes how people are

energized and oriented. An extrovert is energized by interaction
with others and is mainly interested in the outer world of people
and events (extroverts). In contrast, an introvert is energized by
solitary activities and is more drawn towards concepts and ideas
from the inner world.

150
2. The sensing vs. intuition scale describes how people perceive
the world and take in input. The sensing person sees the world
in a practical and factual way whereas an intuitive person values
relationships, possibilities and meanings and is drawn to
innovative or theoretical ideas.
3. The thinking vs. feeling scale describes how people draw con-
clusions and make decisions. Thinkers make decisions based on
objective and cause-and-effect criteria. Feelers tend to be
subjective and to make decisions based on what they feel about
others.
4. The judging vs. perceiving scale describes how individuals deal
with the outside world. Judgers prefer closure, structure, organi-
sation, and control. Perceivers value spontaneity, flexibility,
freedom, and autonomy.

Each of the four polar scales is independent of the other three,

resulting in sixteen possible combinations of preferences called
“types”. A type is encoded with an acronym consisting of the initial
letters36 of the preferences described above. Thus ISFP means a type
for a person who prefers introversion, sensing, feeling and perceiving.
See Myers and McCaulley (1985) for specific procedures on how to
use the MBTI. The MBTI is used as a measure of both personality
types and learning styles. However, as Oxford (1989) points out, it
only measures limited types of learning style. Other important aspects
of learning style include learning modality preferences (aural / visual /
kinaesthetic / tactile) and cognitive preferences (global / particular,
synthesizing / analyzing, field independence / dependence, etc.).

7.3.2.2. The Perceptual Learning Style Preference Survey

The Perceptual Learning Style Preferences (PLSP) Survey was
developed by Reid to allow learners to identify their preferred learning
styles from among six categories: visual, auditory, kinaesthetic, tactile,
group, and individual learning. The first four categories are concerned
with how students use their physical senses to perceive information,

36 Intuition is abbreviated “N” as Introversion is “I”.

151
and the last two categories deal with social vs. independent learning
preferences. Reid (1998) reported a large-scale study using the PLSP
survey, involving 1300 students with nine different mother tongue
backgrounds. The main findings are summarized as follows:
 Most learners studying English in the United States demonstrate
strong learning style preferences for kinaesthetic and tactile
learning.
 Most learners dislike learning in groups.
 Learners from different language/cultural backgrounds differ
significantly in their choices of major, minor, and negative37
learning style.
 Learners from specific study fields tend to prefer specific learn-
ing styles, e.g. engineering students prefer tactile learning
whereas students of the exact sciences prefer visual learning.

7.3.2.3. The Style Analysis Survey

The previous two learning style instruments of measure primarily deal
with personality types and physical perceptual preferences respective-
ly. The Style Analysis Survey (SAS) developed by Oxford tries to
capture learning style in a more comprehensive manner by integrating
various relevant style features. Oxford (2001) distinguishes between
three basic types of dimensions of learning style: (1) sensory style
dimensions which include visual/auditory/hands-on; (2) social style
dimensions which include extrovert/introvert; (3) cognitive style
dimensions which include concrete-sequential/random-intuitive,
closure-oriented/open, global/particular, and analyzing/synthesizing,
etc. Each learner locates himself/herself somewhere on a continuum
for each style dimension.
The original SAS deals with five dimensions of learning style:
visual/auditory/hands-on, extrovert/introvert, concrete-sequential/ran-
dom-intuitive, closure-oriented/open, global/particular. Recently,
Cohen et al. (2005), taking the SAS as their basis and incorporating

37 Major learning style is the style with which the learner feels comfortable and
features as the prominent style used by the learner; minor learning style means
the style the learner can normally function well with but uses less; negative
style means the style which the learner has difficulty with and tries to avoid.

152
the comprehensive cognitive styles developed by Ehrman and Leaver
(2003), developed the Learning Style Survey (LSS) which added six
other cognitive dimensions: analyzing/synthesizing, sharpener/leveller,
deductive/inductive, field-independence/field-dependence, impulsive/
reflective, and metaphorical/literal. We shall take a brief look at the
nine cognitive style dimensions38 which have not been systematically
touched upon previously. See the table below.

Concrete sequential persons generally think the present is

Concrete-sequential/ more important and would like to know where one is
random-intuitive going in learning at each step; random-intuitive persons
are more concerned with the future and like to speculate
about possibilities, enjoying abstract thinking instead of
step-by-step instruction.
Closure-oriented persons treat learning tasks carefully and
Closure-oriented/open try hard to meet deadlines; open oriented persons enjoy
learning or discovering things and prefer to be relaxed
while learning, without being bothered about deadlines or
rules.
Global persons enjoy getting main ideas and are com-
Global/particular fortable in communication even if they do not know all the
words or concepts; particular persons focus more on
details and remember specific information.
Synthesizing persons summarize material well and enjoy
Synthesizing/analyzing guessing or predicting activities and they notice similari-
ties quickly; analytical persons can separate ideas easily
and perform well on logical analysis, and tend to focus on
grammar rules.
Sharpeners tend to notice differences among items when
committing material to memory and thus they can easily
Sharpener/leveller retrieve the different items by storing them separately;
levellers are likely to put material together to remember it
by focusing on similarities and consequently they tend to
confuse memories by merging new things with old ones.
Deductive persons like to apply generalizations to ex-
Deductive/inductive perience by starting with rules or theories rather than
specific examples; inductive persons prefer to begin with
examples instead of rules or theories.

38 The source of all these cognitive dimensions is the Learning Style Survey by
Cohen et al. (2003).
153
 Field-independent persons like to separate or abstract
Field-independence/ material from a given context even when distractions are
field-dependence present, but they are less able to deal with information
holistically; field-dependent persons tend to deal with in-
formation holistically while having difficulty in separating
material from the context.
Impulsive persons react quickly in acting or speaking
Impulsive/reflective without thinking about the situation; reflective persons
think things over before initiating any action.
Metaphorical persons can learn material more effectively
Metaphorical/literal by conceptualizing it in metaphorical terms; literal persons
prefer a literal representation of concepts and like to work
with language material more or less from a surface under-
standing.

Table 2. Cognitive style listed in LSS (Cohen et al., 2005).

7.4. Strategies

7.4.1. Language learning strategies

While language learning styles are general approaches to language

learning, learning strategies can be defined as “specifications, behav-
iours, steps, techniques (or thoughts) – such as seeking out conversation
partners, or giving oneself encouragement to tackle a difficult language
task – used by students to enhance their own learning” (Scarcella/
Oxford, 1992, cited in Oxford, 2003: 274).
Research into language learning strategies began in the 1970s to
move away from the predominant role of the teacher and to turn
towards the learners’ active role in language acquisition (Schmitt,
1997: 199). Language achievement was not seen as purely dependent
on language aptitude but to a large extent on the learner’s own learn-
ing effort and general approach to language learning. The starting
point of the research was to find out what strategies good learners
employ in language learning (Schmitt, 1997: 200). Good learners tend
to use more strategies and in a more systematic way than poor learners.

154
Researchers identify and record those strategies which appear to be
helpful to learning: the strategies were either observed by researchers
or reported by the learners themselves. After a range of strategies are
identified, they need to be categorized in a comprehensive framework.
Based on a cognitive information-processing model, O’Malley
and Chaumot (1990) developed a three-category framework of learn-
ing strategies: (1) metacognitive strategies (overall planning, monitor-
ing and evaluation of the learning process), (2) cognitive strategies
(direct manipulation or transformation of learning materials), (3)
social/affective strategies (cooperative learning, asking information,
control of emotional constraints). Oxford (1990) produced an
extensive list of strategies39, grouped into six categories by means of
factor analysis, namely, memory, cognitive, compensation, meta-
cognitive, affective and social strategies. She views the first three as
direct strategies and the last three as indirect strategies. Cohen (1998)
makes a distinction between strategies that are used to learn the
language and those that are used to practise the language after initial
learning; he refers to them as language learning strategies and
language use strategies. The latter include four types of strategies: (1)
retrieving strategies, (2) rehearsing strategies, (3) cover strategies (to
help students avoid looking stupid or unprepared), and (4) com-
munication strategies.
In general, the research shows that learners do use a wide range
of strategies and find them helpful. It is now commonly accepted that
learning strategies are not inherently good or bad, but depend on the
context in which they are used (Schmitt, 1997; Oxford, 2001).
According to Oxford (2001: 362), a strategy is useful in three con-
ditions: (1) it suits well the L2 task; (2) it is in harmony with the
learner’s learning style; and (3) it is used effectively and combined
with other relevant strategies. Another important finding is that L2
learning strategy use is significantly related to a wide range of factors:
learning motivation, learning style, personality type, gender, age,

39 The strategies are listed in the Strategy Inventory for Language Learning
(SILL). There are two original versions of SILL: one, containing 80 items, is
for English speakers learning a new language; the other, containing 50 items,
is for speakers of other languages learning English.
155
culture, brain hemisphere dominance, career orientation, academic
major, beliefs, language teaching method and the nature of the L2
tasks (Oxford, 1989, 2001). This implies that when selecting a specific
strategy for training such factors have to be considered before
deciding whether the strategy will be geared to individual learner
needs or suited to a learner’s particular case.

7.4.2. Vocabulary learning strategies

Vocabulary learning strategies, as a component of language learning

strategies, have been documented since an interest was first shown in
language strategies. O’Malley and Chamot note: “Research on
training second language learners to use learning strategies has been
limited almost exclusively to applications with vocabulary tasks”
(1990: 7). It is frequently reported that learners use more strategies for
learning vocabulary than other types of strategies. According to
Schmitt (1997: 201), this phenomenon might be due to two reasons: (1)
the relatively more discrete nature of vocabulary learning than other
integrated activities, such as oral presentation, which makes it easy to
apply learning strategies effectively; (2) learners’ awareness of the
importance of vocabulary in language learning. Early research in
vocabulary learning strategies focused on certain types of memory
strategies, such as the keyword method (Atkinson/Raugh 1975;
Pressley/Levin 1981) and the hookword method (Paivio/Desrochers
1979), verbal/imagery association (Cohen 1987). They all belong to
the mnemonic family of strategies. A mnemonic can be an image, a
word, a short poem, a rhyme or a sentence to help the learner
remember a word (Kelly, 1985). Mnemonics have a long history,
going back to the time of the ancient Greeks and can be roughly
divided into two categories, one covering verbal forms, the other
imagery. Invariably, however, they are combined in the learning of the
L2 vocabulary, as we shall see.
The basic principle behind traditional mnemonic techniques is
to make a formal and a semantic link with the word (or words) to be
remembered with a word (or words) that is already well known to the
learner, in accordance with the even more basic principle that all

156
learning is the integration of new information with old. While Greek
and Roman scholars gave a considerable amount of attention to these
techniques, they have been extensively explored by psychologists in
recent times (e.g. Montague et al., 1966; Prytulak, 1971); even though
a wide range of retrieval systems renders such techniques far less
useful than in the past, second language vocabulary learning is one of
the few remaining areas where they can be said to be still needed.
It is known that some learners make use of mnemonic tech-
niques in haphazard and idiosyncratic ways. To give one example, a
Chinese learner may try to remember the word charisma – difficult to
learn because of length, pronunciation and meaning – by creating a
highly unlikely and ungrammatical sentence like “China rises
Maotsedong”, the meaning of which is very obscure but which the
learner associates with “Mao has tremendous personal glamour”.
There is an orthographical and semantic link that will help remember
the new word.
Most verbal mnemonic methods, however, rely on pronunci-
ation rather than spelling in order to make the formal link, though the
two may overlap. The keyword method, introduced first at Stanford
University by Atkinson and Raugh (1975) and later more fully
developed by Pressley and his associates at Western Ontario (e.g.
Pressley/Levin, 1981), makes use of a word in the learner’s L1 as the
mediator, one that bears a phonetic resemblance to the word to be
learned and makes the meaning link by means of a mental image.
Thus, to learn the French couteau (knife), the learner imagines
someone cutting his toe with a knife, the dual mediator being cut and
toe; to learn the Spanish carta (letter), s/he imagines a letter lying in a
supermarket cart (Am. English for trolley). When s/he hears the word,
the image springs to mind and the learner can access the meaning.
Eventually, these “crutches” are no longer needed and the meaning is
immediately accessed.
The keyword method has been the subject of a considerable
amount of research and experimentation and, along with the hookword
method, the image-based mnemonic of Paivio (e.g. Paivio/Desrochers,
1979), has been demonstrated to be three times as effective as the
traditional rote method. Also an important finding of the keyword
research is that mediators are far more effective when the learners
157
think them up themselves, although they may well need training in
how to form effective and lasting images. Citing the evidence, Nation
(2001) concludes that the keyword method generally leads to faster
and better learning than many other methods, not simply rote.
There are problems, however, with these mnemonic methods. In
the first place, while it is probably necessary to make verbal
associations with words in one’s own language in the initial stages,
this is not conducive to mastering the correct pronunciation of the new
item in the L2. Later, as the learner’s L2 vocabulary increases, it may
be easier to find mediators in the L2 but the fact remains that it is very
difficult to find associations of this kind in respect of many new words.
This is doubtless part of the reason why mnemonic methods, although
eminently successful in experimental conditions, have not caught on
in the classroom. Other reasons, of course, may be because rote
learning is deeply engrained from an early age and also because
teachers are not sufficiently familiar with such techniques and do not
train their pupils in their use. As Schmitt (1997) and others (e.g. Kelly,
1985) have pointed out, mechanical, mindless repetition is still a
dominant vocabulary learning strategy among learners, particularly, I
would add in passing, Chinese learners.
The keyword method or, for that matter, any mnemonic method
does not have to be adopted but the two basic principles of trying to
establish a formal and a semantic link with the word to be learned can
be inculcated. As researchers like Paivio and Desrochers (1981) have
pointed out, this can make vocabulary enjoyable and not the dull chore
that it is for so many learners. The use of mnemonics can form part of
their overall vocabulary learning approach.

7.4.3. Vocabulary learning strategy instruction

Strategy instruction emphasizes teaching learners specific learning

strategies to make learning more efficient (Oxford/Scarcella, 1994;
Cohen, 1998; O’Malley et al. 1990). Researchers into strategy instruc-
tion often hold the view that context can provide the essential means
for learning vocabulary but additional support, such as explicit
strategy instruction, is needed and helpful (Oxford/Scarcella, 1994).

158
The typical strategies recommended are word grouping, word
association, imagery, mnemonics, and semantic mapping, etc.
Traditionally, strategy instruction seems to be concerned with
advanced learners rather than low level learners (Coady, 1997).
However, strategy instruction to low or intermediate level learners can
be very useful. Strategies such as imagery and mnemonics can be
made very helpful since the greatest difficulty in acquiring a word in
the initial stages is to link the form and the meaning in memory (Kelly,
1986; Laufer et al., 2004). This is particularly true in respect of an
unrelated language and was the motivation for developing the
keyword method (Atkinson/Raugh, 1975) as students of Russian at
these researchers’ university were having enormous difficulty memo-
rizing the most basic vocabulary.

7.4.4. Classifying vocabulary learning strategies

Although vocabulary learning strategies have been long recognised,

their influence on L2 acquisition remains fairly limited compared with
the growing interest in general language learning strategies. Also, the
focus on a few memory strategies seems not to be able to capture the
picture of actual strategy use among learners. Learners tend to use
various strategies in combination (Gu/Johnson, 1996). Schmitt (1997)
noted there was no comprehensive list or taxonomy of vocabulary
learning strategies. There are, however, a few ways of classifying
them, ranging from making an extensive list to adopting a learning
process oriented approach.

7.4.4.1. Vocabulary Learning Questionnaire

The Vocabulary Learning Questionnaire (VLQ) presented by Gu and
Johnson (1996) perhaps provides the most exhaustive list of vocabulary
learning strategies, containing 108 items altogether. The strategies
listed are divided into three sections, (1) belief about vocabulary
learning, (2) metacognitive strategies, and (3) cognitive strategies. See
Table 3 for an overview of the structure of the strategies; sometimes
examples are given as they appear in the original questionnaire.
159
 Gu and Johnson administered this questionnaire to 850 second
year Chinese university students to find out how they use different
strategies and how this affects their learning outcome in terms of
vocabulary size and general English proficiency. The main findings
were that: (1) the two metacognitive strategies, self-initiation and
selective attention, tend to be positive and the best predictor of
English general proficiency; (2) contextual guessing, skilful use of
dictionaries, contextual encoding, and activation strategies are corre-
lated with both English proficiency and vocabulary size. Another
important finding was that learners fall into five types in terms of their
English achievement based on their vocabulary strategy use profile: (a)
readers (N=3, 0.6%)40, who believe that vocabulary should be ac-
quired through extensive reading and careful study (e.g. attending to
word forms such as affixes); (b) active strategy users (N=48, 9.9%),
who also believe that words can be acquired through reading and use;
in addition, they work hard and use a variety of strategies more often
than other students; (c) non-encoders (N=269, 55.3%), who generally
do not make much use of learning strategies; (d) encoders (N=157,
32.3%), who are similar to non-encoders except that they use more
encoding strategies; (e) passive strategy users (N=9, 1.9%), who are
generally unmotivated or demotivated students, seeing little value in
language learning and having little idea how to learn a language.

Words should be memorised (8 items)

Belief Words should be acquired in context (4 items)
Words should be studied and put to use (5 items)
Selective attention (7 items)
Strategies concerning how to select a word to be studied, e.g.
Metacognitive Look up words I’m interested in.
strategies Self-initiation (5 items)
Strategies concerning active searching for new words beyond
classroom learning, e.g. Besides textbooks, I look for other
readings that fall under my interest.

40 This means that three subjects are grouped under this category, accounting for
0.6% of the whole population.

160
 Guessing strategies (12 items)
Dictionary strategies (17 items)
Note-taking strategies (9 items)
Memory strategies: rehearsal (12 items)
Strategies concerning mechanical means such as using word lists,
Cognitive oral and visual repetition.
strategies Memory strategies: encoding (24 items)
Strategies encoding vocabulary through association/elaboration,
imagery, visual/auditory association, and word-structure.
Activation strategies (5 items)
Strategies concerning active use of vocabulary learned, e.g. I try
to use newly learned words in real situations.

Table 3. Vocabulary learning strategies listed in VLQ (Gu/Johnson, 1996).

Two points should be made regarding the generalizability of these

findings. First, some of the strategies are significantly correlated with
language proficiency or vocabulary size; the correlation is actually
fairly low, the absolute value of all correlation coefficients ranging
from 0.13 to 0.31, though it might be partially justified in terms of a
rather large sample size. Second, as Nation notes (2001: 227), a
questionnaire containing 108 items could very likely generate a
“fatigue factor” that might affect the result. This is best shown by the
large proportion of the two middle types of students, the encoders and
non-encoders, who together make up 87.7% of the whole sample.
However, an important message transmitted by this study is that, if we
put the last three categories (non-encoders, encoders, and passive
strategy users) together, almost 90% of the students in this sample
make use of very few vocabulary learning strategies despite the fact
that there is a wide range of strategies available, and, as a consequence,
these students have low achievement in English.

7.4.4.2. Taxonomy of vocabulary learning strategies

Schmitt (1997) presents another fairly comprehensive taxonomy of
vocabulary learning strategies, composed of 58 items, based on
studies of Japanese learners of English. The taxonomy is developed on
the basis of the four (out of six) categories of language learning
strategies proposed by Oxford (1990): social, memory, cognitive, and
161
metacognitive strategies. The taxonomy is divided into two main
sections by making a distinction between the initial stage of discover-
ing the word meaning and the later stage of consolidating the
vocabulary knowledge. See Table 4 for an overview of the taxonomy.

Determination strategies (9 items)

Strategies for discovering the meaning on one’s own, e.g.
Discovering a new guess from textual context.
word’s meaning Social strategies (5 items)
Strategies for discovering the meaning by working with
others, e.g. ask classmates for meaning.
Social strategies (3 items)
e.g. interact with native speakers.
Memory strategies (27 items)
Strategies closer to traditional mnemonic techniques, e.g.
Consolidating a imagery and the keyword method.
word after being Cognitive strategies (9 items)
encountered Strategies concerning organizing lexical information and
using mechanical means to memorise words, e.g. keep a
vocabulary notebook; verbal repetition.
Metacognitive strategies (5 items)
e.g. using English-language media and spaced word
practice.

Table 4. Vocabulary learning strategies listed by Schmitt (1997).

The taxonomy is intended to look for what strategies learners use and
believe to be helpful. A questionnaire was administered to 600
Japanese students of four levels, junior high school students, high
school students, university students and adult learners. The most used
strategies include using a bilingual dictionary (85%), guessing from
textual context (74%), consulting classmates for meaning (73%),
verbal repetition (76%), written repetition (76%), studying the spelling
(74%), saying the new word aloud (69%), etc. The least used
strategies include checking for L1 cognates (11%), performing
physical actions (13%), using semantic maps (9%), teachers checking
with flash cards (3%), etc. The results generally confirmed what is
typically found in respect of Asian students, i.e. relying on mechanical
means and focusing on form rather than meaning when studying

162
lexical items. The most helpful strategies do not necessarily
correspond to the most used strategies. Schmitt notes there is a fairly
large gap between the three strategies which students rank as helpful
in comparison with the actual use of these strategies. The three
strategies are analyzing pictures/gestures, connecting words with
synonyms/antonyms, and continuing to study words over time. In his
view, this implies that students may be willing to try new strategies if
they are introduced to and instructed in them. This seems likely when
we realise that the reason why most students use a limited number of
strategies is largely because they are rarely instructed in them, or at
least not in a systematic way.

7.4.4.3. A learning process oriented approach

The two taxonomies mentioned above were largely developed on the
basis of language learning strategies in general, and are not especially
geared to the nature of vocabulary acquisition. I propose an alternative
approach to categorizing vocabulary learning strategies in accordance
with the way vocabulary is acquired. There appear to be several
distinct stages in the acquisition of a lexical item, which I summarize
sequentially as follows:

1. To acquire a new word, one has to encounter it through different

contexts such as classroom activities, vocabulary lists, reading,
or conversations in the target language. The situation in which
we find that learners usually meet the new words shows how
they approach vocabulary acquisition generally; the data revealed
in this way should be more convincing than what learners claim
that they do.
2. The meaning of the new word should be found out; this could
be done by asking others (teachers or classmates), looking up in
dictionaries or contextual guessing.
3. Acquiring vocabulary is a complicated process, involving
knowing various aspects of knowledge of it; some of the
essential knowledge includes the meaning, spelling, part of
speech, collocations, etc.

163
4. This information needs to be recorded or organised in certain
meaningful ways.
5. The initial stage of acquiring a word is to connect the word
form with the meaning and fix the two together in one’s mind.
How successful this is depends on memory processing as shown
by the “depth of processing” theory (Craik/Lockhart, 1972;
Craik/Tulving, 1975); some effective memory strategies can be
used to facilitate this process.
6. The word needs to be rehearsed to ensure long-term retention.
7. When meeting the word again, one might need to recall the
meaning by some means, such as recognizing the affixes, roots
or the known parts of the word.
8. The word learned has to be put to different uses to consolidate
its acquisition.

This is a cyclic process and more words or even formulaic sequences

can be acquired in this manner. It should be noted that not every word
acquired will go through all the stages, but words that do will tend to
be acquired better and be retained longer. The new taxonomy of vo-
cabulary learning strategies which attempts to reproduce the learning
process should be able to provide answers to the following questions:
 How is a new word discovered?
 How is the word meaning discovered?
 What features of the word are studied?
 How is the lexical information recorded and organised?
 What techniques are used to memorise the word?
 How is the word reviewed/rehearsed?
 How is the word retrieved?
 How is the word put into use?

An application of this approach to categorizing vocabulary learning

strategies will be presented in Chapter 12.

164
7.5. Towards an integrated framework for SLVA

I have investigated three learner differences: motivation, style and

strategies. The three factors are in fact interrelated and are ultimately
inseparable (Cohen/Dörnyei, 2002; Ehrman et al. 2003). Styles are
general approaches to language learning while strategies are specific
behaviours or techniques that learners select in language learning and
use. That said, learning styles are made obvious by learning strategies.
As observed by a number of authors (e.g. Ehrman et al. 2003; Oxford,
2001), a given strategy is neutral and it is considered to be useful only
if certain conditions are fulfilled, one of which being that the strategy
should fit the learner’s style preferences. Cohen (2003) uses a visual
representation to illustrate how styles and strategies are related within
a given task: to perform the task, the learner will make use of a series
of strategies that are presumably consistent with his/her style
preferences. From the process model of motivation developed by
Dörnyei and Ottó (1998) and described earlier, strategies have an
important impact on motivation in all three phases. Cohen and
Dörnyei (2002: 175) describe the close bond between motivation and
strategies in this way:

[…] while learning, well-used strategies increase one’s self-confidence and

lead to increased success, and – as the saying goes – success breeds further
success. […] one very important function of the retrospective stage is for
learners to consolidate and extend the repertoire of personally useful strategies,
which will in turn function as a source of inspiration for future learning.
Indeed, strategies and motivation are very closely linked.

These statements should particularly hold true for vocabulary learning.

There are a large number of vocabulary learning strategies available
and vocabulary learning is perhaps the easiest observable type of
language learning. By employing effective strategies that suit their
style, learners can learn a considerable number of words in a
reasonable period of time, which can further motivate them to learn
more words by making a greater learning effort. So far it seems that
the direct link between motivation and learning style is not very clear,

165
but the interaction of the two calls for further research (Ehrman et al.
2003). The available literature shows that motivation and learning
styles are essentially linked with each other (Erhman, 1996).
I have constructed a learner-focused model of SLVA on the
basis of the above assumptions in order to describe the process how
these three learner variables can influence SLVA (see Figure 7). In this
model, vocabulary knowledge (breadth/depth, reception/production) is
situated in the centre, affected by three learner variables, motivation,
learning styles, and learning strategies. It should be noted that this
does not mean other learner variables are not important but that these
three variables are crucial factors that affect vocabulary acquisition in
this model.
First, each of the three variables directly contributes to
vocabulary knowledge. Learning styles are the general approaches to
vocabulary learning. They affect the knowledge eventually gained; for
example, a learner who has an auditory style preference might
understand and produce lexical items easily in communication but
might not be able to write them correctly if visual forms are not
attended to. Learning strategies are the specific techniques, methods,
or activities the learner selects to study vocabulary. If strategies are
used systematically, vocabulary learning will be effective and efficient.
Deep or shallow strategies determine the quality of vocabulary
knowledge. Deep strategies, involving deep mental processing,
elaboration and association, will lead to better retention of the lexical
items as well as to a more interrelated network of the lexicon than
shallow strategies, such as mechanical means that aim at matching L1
equivalents with the L2 lexical items. Vocabulary learning, like other
types of learning, needs to be initiated, continued and completed by
motivation, and the process can be recycled, as described by the
process model of Dörnyei and Ottó (see 7.2.2.). The degree of motiva-
tion affects both the quality and quantity of the vocabulary knowledge
gained.
Second, the three variables influence each other; this implies
that any change in any variable will potentially lead to changes in the
other two. For example, stretching learning styles might incorporate
corresponding learning strategies into the repertoire of strategy use for
learning vocabulary or vice versa. Successful use of vocabulary

166
learning strategies will increase the learner’s motivation to use them.
The same holds true for the impact of change of motivation on
learning strategies and styles. It is assumed that all these three
variables could have both positive and negative impacts on each other,
depending on the increase and decrease in the value of each variable.
This model is in its preliminary form and more research has to be
done to test its validity. Cohen and Dörnyei (2002) provide a
framework for systematic instruction in learning styles, strategies that
will motivate learners. It could also be employed as a starting point to
implement the learner-focused model of SLVA presented above. The
main points of their framework can be summarized as follows:
 Raise learner awareness about learning style preferences and
learning strategies in the very beginning to motivate them to be
more conscious about style preferences and more proactive
about strategy use.
 Find out which styles the learners prefer, and which strategies
they already use or would like to include. Suggest and model
what “style-stretching” might be and model new strategies.
 Convince learners of the usefulness of strategy use and help them
practise use of strategies through exercises or other activities.
Encourage learners to strengthen their currently used strategies.
 Emphasize cross-cultural differences in how strategies should
be used.
 Organise activities in which learners can share information
about their style preference and useful strategies.

167
 Motivation

Vocabulary knowledge

Learner development

Strategies Styles

Figure 7. A learner-focused model of SLVA.

7.6. Conclusion

In this chapter, we looked at the three learner variables that most

affect language learning, namely, learner motivation, styles and
strategies. It is recognised that motivation is a dynamic process and
that, while we learn a language for a practical reason or because we
are drawn to the culture, or for both, the strength and the nature of the
motivation vary over time. Various means have been developed for
assessing learning style and categories identified in which learners can
be placed. I then endeavoured to show how the many vocabulary
learning strategies that have been identified can be classified in
accordance with the way vocabulary is acquired. The picture that

168
emerges is that these three variables are interrelated and that, in
accordance with their profile in relation to all three, learners need to
be guided so that they can acquire good learning ability and take
charge of their own learning autonomously and effectively. I
developed a learner-focused model to acquiring vocabulary in
consideration of this need. By developing awareness of learning styles
and strategies and systematic strategy use, learners are motivated to
learn more vocabulary and to become better learners. This is what
learner development means in the context of vocabulary learning.

169
170
 Part III:
Computer Assisted Vocabulary Learning (CAVL)

171
172
Chapter 8. CAVL: Past, Present, and Future

8.1. An overview

Vocabulary learning has always been a favourite subject in computer

assisted language learning (CALL) programs, especially in the early
stages of CALL (1970−1980s) when technology was relatively simple
and it was thought that vocabulary learning could be easily integrated
into CALL programs. Cited in Goodfellow (1995), Jung (1988) con-
ducted a survey of international bibliographies of CALL and found
vocabulary was the fifth most common keyword; he also stated that
vocabulary was an important topic among all the software packages he
had reviewed. The earlier programs typically included a single type of
language learning activity, such as text reconstruction, gap-filling,
speed-reading, simulation, and vocabulary games (Levy, 1997). There
were several reasons for this. Earlier generations of computers were
less powerful and sophisticated than current ones; many enthusiastic
language teachers in CALL at that time might not have had sufficient
knowledge of programming; sound vocabulary learning theories were
certainly less available at that time when vocabulary learning was just
starting to attract researchers’ attention.
Far from being a neglected field, vocabulary learning is now
viewed as one of the most crucial components in learning an L2
(Harley, 1996; Barcroft, 2004); it is the learning of L2 lexical items
(individual words or multi-word items) that requires so much effort,
considerably more, in fact, than grammar. Whether CALL programs
are devoted to vocabulary learning or not, all deal with vocabulary to
varying degrees, be it for reading, writing, listening, or speaking.
However, despite the importance of vocabulary learning in CALL,
little work has been done on constructing a clear framework to
conceptualise Computer Assisted Vocabulary Learning (CAVL) so that

173
a distinctive sub-domain of CALL41can be developed. One exception
is that of Goodfellow (1995: 205), who raised three critical issues
regarding the design of CAVL programs:

The first is the extent to which the program is specifically oriented to the
learning of vocabulary (as opposed to grammar, pronunciation or pragmatics)
[…] The second is what gives the interaction its instructional flavour and
“worth”. This concerns not only the cleverness of its selection of words or its
explications of knowledge about them, but also the pedagogical principles
embedded in the design, and the learning needs of individual learners. The
third issue is what makes a CALL interaction specifically computational – as
opposed to being an on-line version of a paper-based activity […]

The first issue implies that vocabulary learning should be explicit in

CAVL programs; the second implies that learning theories specific to
vocabulary learning should be incorporated in the design; the third
implies that CAVL (or CALL in general) has some distinctive value,
enhanced by computer technology, which the same paper version
cannot achieve. This analysis still holds true a decade later.

8.2. Factors shaping CAVL development

In order to have a clear picture of CAVL development, it is beneficial

to have an idea of what important factors have shaped CAVL develop-
ment over time. Although there is not enough information or expertise
which could lead to a legitimate definition of CAVL, I propose that a
CAVL application, be it a stand alone or online application, should
meet certain pedagogical criteria:
 The learning activities should be designed to allow learners to
pay sufficient attention to the meaning and form of linguistic
information and connect the two.

41 The purpose is not to separate CAVL from CALL, but to evolve a compre-
hensive framework which might facilitate the development of expertise geared
to CAVL.

174
 The linguistic information should, ideally, be presented through
two channels – visual and aural.
 The program should be equipped with electronic lexical tools
(dictionary or concordancers).
 Learners should be encouraged to do exercises on new items
(rehearse, manipulate or reorganise the information about items),
etc.

The above criteria clearly suggest an explicit learning framework for

CAVL. Often, not all of these characteristics will be given prominence
in CAVL programs, partially reflecting the on-going debate regarding
the implicit/explicit learning of vocabulary. Despite the fact that
vocabulary learning (with or without computer) is treated either
implicitly or explicitly, most would probably agree that the two
approaches should be combined with each other.
The evolution of technology has certainly had an impact on
CAVL as on CALL in general. In addition, the development of CAVL
has led to the tutor/tool distinction regarding the essential role of the
computer in CALL programs. Based on the work of Taylor (1980),
Levy (1997) made the tutor/tool framework well known in the CALL
field and saw the distinction as a fundamental notion in conceptualiz-
ing CALL. The difference between the two is that the computer as
tutor will act as a substitute teacher and provide some guidance and
instruction to the learner or give an evaluation of the learner perfor-
mance whereas the computer as a tool serves only to facilitate learner
performance. More specifically, he writes: “In contrast to the tutor role
of the computer, which is intended to emulate or replace the teacher in
some way, the function of the computer as a tool is to enhance or im-
prove the efficiency of the work of the teacher or student” (1997: 184).
The notion of the computer as a tutor is initially rooted in the
behaviourism advocated by Skinner (1957), who sees the computer as
the ideal device to deliver language learning materials to learners.
That is why early tutorial CALL programs put a heavy emphasis on
drill and practice. On the other hand, the computer, seen as an
important tool to assist various human activities (including language
learning), is firmly established as the technology evolves and com-
puter use becomes an integral part of our lives. Common examples of
175
tool oriented CALL applications include word processor, spread sheets
(e.g. Excel), email, electronic dictionaries, concordancers, etc.
Although recently this tutor/tool dichotomy has been challenged by a
few authors regarding the legitimate definition (e.g. Hubbard/Siskin,
200442) or the finer categorization of computer roles (e.g. Colpaert,
200443), the distinction between tutor and tool has been an important
factor in shaping CAVL development.

8.3. Four stages of CAVL development

In this section, we shall provide an historical overview of CAVL

development in terms of source of language learning theories,
implicit/explicit learning, the tutor/tool distinction, and the peda-
gogical features of CALL applications. The results are summarized in
Table 5. Whilst the table depicts the overall trend of CAVL develop-
ment historically, each type is not necessarily strictly associated with a
fixed period of time. Generally speaking, the simple tutor was
predominant prior to the 1990s; during the 1990s first the intelligent
tutor and then the tool became popular; tutor & tool have become
prominent since the beginning of the 21 st century. In addition, the type
of program of the preceding period is not necessarily obsolete in a
new period; it can be conducive to learning when properly used.

42 Hubbard and Siskin (2004) argue that the essential distinction between tutor
and tool should not be based on whether the computer application provides an
evaluation of learner input or not but on whether it provides teaching-like
activities or not. They (2004: 457) give a new definition to CALL as a tutor:
“Tutorial CALL refers to the implementation of computer programs (disc,
CD-ROM, web-based, etc.) that include an identifiable presence specifically
for improving some aspects of language proficiency.”
43 Colpaert (2004: 85−87) provides a five-category framework of “linguistic-
didactic functionality types” of CALL applications, namely tool, monitor,
mentor, tutor, and lector, to replace the rather broad tutor/tool dichotomy.

176
8.3.1. Simple tutor

The simple tutor type of a CAVL program is similar to what has been
called the “first generation” of vocabulary programs by Goodfellow
(1995). The linguistic and psychological learning principles were
essentially derived from behaviourism and typically resulted in
vocabulary exercises similar to drill and practice. Drill and practice
are not inherently harmful to language learning. Sometimes they can
be useful in developing language automaticity, fluency and accuracy.
According to Goodfellow (1995: 206−207), there are two most
important features of vocabulary programs of this type: (1) they tend
to be online versions of paper version vocabulary learning games/
exercises; (2) they primarily focus on testing the vocabulary items
rather than addressing the vocabulary learning process. The simplistic
learning theory, behaviourism, behind those vocabulary learning
exercises, together with the simple computer technology at that time,
made the interaction between the computer and the learner insig-
nificant. As Goodfellow pointed out: “Limited learner performance
restricts the operation of feedback to a simple match with the ‘correct’
solution” (1995: 210).

8.3.2. Intelligent tutor

The intelligent tutor oriented CAVL largely resulted from the develop-
ment of computer technology and the application of psycholinguistic
lexical expertise. This roughly corresponds to the “second generation”
of programs described by Goodfellow (1995). The programs devel-
oped were largely based on the lexical knowledge/expertise developed
in the domain of psycholinguistics, e.g. the structure of the lexicon
(e.g. Goodfellow, 1994), the necessary stages of lexical item acquisi-
tion (e.g. Groot, 2000), the contextual roles in acquiring lexical items
(e.g. Beheydt, 1990), semi-contextualised techniques in developing
lexical knowledge (e.g. Labrie, 2000), the use of semantic reply-

177
forms44 in comprehending and practising new words and the use of
diversified exercises to facilitate vocabulary learning (Decoo et al.,
1996). The main focus was on the learning process of lexical items
rather than the testing45 of them directly. Goodfellow (1995) listed
two important features brought by technological innovation. First, the
development of graphic user interface enables the user to control and
manipulate their own learning by performing various requests or func-
tions available on the screen. Second, as a consequence of the graphic
user interface development, CALL is capable of “separating, the ‘pro-
cess control’ role of the computer, i.e.: the way it determines what the
user can actually do, from its ‘information resource’ role, i.e.: the way
it provides information to support the user’s objectives” (Goodfellow,
1995: 206). The helpful information the program provides is essential-
ly lexical tools or resources, such as online dictionaries, database
(corpora), and concordancers. In addition, thanks to the increasing
computer capacity for saving information as well as the development
of tracking technology, learner behaviour on the computer can be
recorded for further analysis, which could reveal insights into the
learning process the learner is engaged in.

8.3.3. Tool

The CAVL tool became popular as most (early) tutorial CALL failed
to live up to expectations. Tutorial CALL is frequently associated with
behaviourism or drill and practice, even though this is not necessarily
true (Hubbard/Siskin, 2004; Colpaert, 2004). Colpaert (2004: 69)
writes: “It appears to be a fashionable criticism or even a form of
political correctness in CALL, to consider language courseware as
rote drill-and-practice, while tools, in contrast, hold alluring promise
for creative, experiential, collaborative, and task-based learning”. It is

44 The target word to be learned is embedded in a context which takes on the

form of a reply, e.g. How many people are at the conference? There are about
200 participants.
45 Generally speaking, testing is different from the learning process, but some-
times testing lexical items which have been previously encountered or learned
can be seen as a continuous learning activity.

178
against such a distorted image of tutorial CALL that tool oriented
CALL has thrived.
There appear to be several sub-categories within CAVL tools.
One is that of incidental learning from online reading comprehension,
with lexical aids such as glossary, annotation, or electronic dictionary.
How lexical information should be displayed, i.e. in the L1 or the L2,
visually or aurally, textually or pictorially, becomes the main topic of
discussion among authors interested in incidental learning of vocabu-
lary. The second sub-category is use of different forms of CMC (Com-
puter Mediated Communication) tools: email, chatting, and bulletin
board. The theoretical basis of language learning in CMC largely
comes from the interactionist SLA in which “conversational inter-
action in the target language (TL) forms the basis for language
development” (Smith, 2004: 365). To a large degree, the interactionist
SLA fits in well with the broad communicative language teaching
paradigm. Originally derived from traditional face-to-face communi-
cation, it is applied to CMC, even though its direct applicability is
questioned by some authors (e.g. Harrington/Levy, 2001). Smith
(2004: 371) writes:

The key interest in CMC, from an interactionist perspective on L2 learning,

involves the specific ways in which CMC is relevant to and facilitative of the
processes believed to be beneficial to SLA. Among the most important issues
is establishing the facility of CMC to supply rich input, promote pushed
output, provide plentiful and dynamic feedback, focus learners’ attention on
aspects of the TL, and enhance noticing.

If this is the case, it is likely that vocabulary items can be acquired

receptively or productively via CMC. So far, the communication in
written language rather than oral language has been widely studied in
CMC. The third sub-category is the data-driven learning (DDL),
initially advocated by Johns (1991a, 1991b). Using concordancing
programs, learners can search large numbers of language situations
where a particular lexical item occurs in an attached corpus. They are
encouraged to induce the meaning or usage of the target item. While
the first two sub-categories primarily encourage implicit learning
where the meaning of the lexical items will be attended to, the last one

179
can lead to explicit learning if the lexical forms are attended to and an
effort has been made to connect the lexical form with the meaning.

8.3.4. Tutor & tool

The tutor & tool integrated CAVL has set the trend for the present and
doubtless for the near future. It can be interpreted in two ways. One
way is to see both tutor and tool going happily ahead hand in hand, as
both can contribute to vocabulary learning from different perspectives.
They should be viewed as complementary to each other rather than
excluding each other. Actually tutorial CALL (CALL as tutor) is only
regarded with disfavour in the general field of language teaching
among language teachers (Hubbard/Siskin, 2004), but it is still
popular in the CALL field (Hubbard/Siskin, 2004; Hubbard, 2003;
Colpaert, 2004). Hubbard and Siskin (2004: 449−453) clearly show
that the “six myths”46 held by many people (largely language teachers)
opposed to tutorial CALL are simply untenable. Colpaert argues that
tutorial CALL “has a promising future in terms of development, use,
and effectiveness” (2004: 76). Ideally, tutor and tool should be both
used in language teaching and learning: “CALL in general would be
more effective if tutors and tools were integrated in global environ-
ments for language learning” (Colpaert, 2004: 77). This certainly
holds true for CAVL. Another way of interpreting tutor and tool is to
combine them in CAVL applications. This notion is not necessarily
new, for example, the earlier intelligent tutor oriented CAVL made an
effort to integrate tools such as electronic dictionary and con-
cordancers into the whole learning system (e.g. Goodfellow, 1994).
With the development of both technology and learning theories, a
combination of tutor and tool can be made more efficient and make
vocabulary learning more efficient.
Regarding the language learning theories to be incorporated in
the design, the tutor & tool integrated CAVL tends to draw on multiple

46 The six myths are: tutorial CALL (1) is behaviourist; (2) contains only drill
and practice; (3) is not communicative; (4) has no significant role for teachers;
(5) has no learner control; (6) is on a disc or CD.

180
theories or learning approaches rather than on one only as in earlier
generations (see Table 5). Van de Poel and Swanepoel (2003) advocate
a framework of “theoretical and methodological pluralism” in design-
ing CALL for vocabulary learning. They point out that, given a series
of identified “mediating pedagogical variables” (conditions for
acquisition) and “mediating learner variables” (learner characteristics),
vocabulary learning should be treated explicitly in a wide and varied
way (2003: 177):

[…] there seems to be no “best” way to approach the teaching of vocabulary

in any absolute sense. Advocates of a pedagogical pluralism, furthermore,
stress the fact that the best method(s) will be determined by such variables as
the vocabulary developmental goals and needs of the learners, their
proficiency in the TL, their learning styles, the nature of the lexical items to be
learnt […], and in general the nature of the linguistic input that serves for
learning […].

Another important trend is that learner training should be given

sufficient prominence either in the design or in the implementation of
the CAVL applications. This reflects the notion of learner develop-
ment discussed in Chapter 7. Learners (particularly beginners) need
some training and help to be able to use the CALL applications more
efficiently. Hubbard (2004: 45) analyses the necessity for this:

A fundamental quandary in CALL is that learners are increasingly required to

take a significant amount of responsibility for their own learning, whether that
learning is taking place through the programmed teaching presence in tutorial
software or the unstructured spaces of the World Wide Web. They are
expected to do this despite the fact that they know little or nothing of how
languages are learned compared to an appropriately trained teacher.

181
The com- Simple tutor Intelligent tutor Tool Tutor & tool
puter’s role
Primary Behaviourism Psycholinguistic Interactionist Multiple
learning theories: lexical SLA; communica- learning
theories memory, tive language theories
organisation, teaching; data-
simulation driven learning
Implicit or Explicit Explicit Implicit & explicit Explicit &
explicit implicit
learning
Pedagogical Drill & Focusing on Focusing on Similar to
features practice, learning process incidental that of the
online version and testing; vocabulary intelligent
of paper developing lexical learning; primarily tutor; well
learning tools; recording using lexical tools integrated
vocabulary and monitoring and CMC; good into the
exercises; learner behaviour; interaction language
minimal good interaction between the curriculum;
interaction between the learner/computer learner
between the computer and the and the training
computer and learner learner/learner.
the learner

Table 5. Historical development of CAVL.

8.4. Constructing a framework

for categorising CAVL applications

8.4.1. The overview of the framework

Currently we are in the tutor & tool integrated stage for CAVL. We
need both tutors and tools; learning needs to be both implicit and
explicit. I attempt to construct a comprehensive framework to catego-
rise current available CAVL applications, as shown in the following
figure.

182
 Lexical programs/tasks (tool & tutor)

Tool/implicit/meaning-focused

Electronic reading/listening Lexical

comprehension with lexical resources/aids
glosses (primarily tool)
CMC lexical-based tasks Electronic dictionaries
Computerised vocabulary Lexical concordancers
exercises

Dedicated CAVL programs

Tutor/explicit/form-focused

Figure 8. A framework for categorising CAVL.

CAVL applications can be divided into two broad categories: lexical

programs/tasks and lexical resources/aids. Lexical programs/tasks
provide the mainframe work station for CAVL while lexical resources/
aids can be integrated into the former as a kind of necessary lexical
help system, or as part of the “meaning technologies”47 defined by
Hubbard (2001).

8.4.2. Lexical resources/aids

Although earlier generations of electronic dictionaries tended to be

online versions of the original paper version dictionaries, nowadays

47 Meaning technology is defined as “features of language learning materials that

are based on technology and make the meaning of a language item accessible
to a learner” (Hubbard, 2001: 82). These include text captions, machine trans-
lations, hypertext dictionaries and glossaries, text-to-speech applications, and
auto summarizers.
183
they are interactive and have a more user-friendly interface. Electronic
dictionaries have a number of potential advantages over paper dictio-
naries. Zähner et al. (1994) identified two distinct advantages. First,
the electronic dictionary can respond to learners’ input intelligently, i.e.
the learner can simply put in any word form without knowing its base
form and the computer can generate various morphological forms
based on a given word. Second, only necessary lexical information is
displayed and extra information can be displayed, say via hyperlinks,
upon request. Electronic dictionaries have a further advantage in that
lexical information can be accessed in multiple channels simulta-
neously: textually, pictorially, and aurally. Regarding the pedagogical
design, electronic dictionaries can also be customized to suit learner
proficiency level by using L1 definitions or controlled L2 vocabulary
(e.g. Van de Poel/Swanepoel, 2003). The design can also simulate the
human lexical memory by organizing lexical entries as synonyms (e.g.
WordNet48 by George A. Miller and his colleagues), or strengthen the
links with other words by associations (e.g. Zähner et al. 1994).
Generally, there are two types of lexical concordancers, mono-
lingual or bi-/multi-lingual, depending on what type(s) of corpora they
are attached to. Monolingual lexical concordancers do not give the
word meaning directly, i.e. via translation or definition, in the way that
electronic dictionaries do; instead, the meaning can only be inferred
from the surrounding contextual clues. Bilingual or multilingual
concordancers would give the lexical meaning of a given L2 word by
means of its equivalent L1 translation. Word usage is usually not given
directly in concordancers and learners need to observe how the word
is used in different contexts. Concordancers provide learners with a
large amount of authentic language input in which the target item
occurs. Whether the concordancers will be used by learners efficiently
depends on several variables, as stated in 6.2.2.4., such as learner
perception (whether they like to use them or not), learner proficiency
(whether they can understand the authentic, uncontrolled language),
teacher help (whether the teacher gives some training to learners on
how to use the concordancers) and, most importantly, what vocabulary
exercises are designed based on the concordanced items or, in this

48 It can be accessed online via <http://wordnet.princeton.edu/perl/webwn>.

184
context, how the use of concordancers is integrated into the lexical
programs/tasks. With regard to the last-mentioned, Cobb (1997) and
Horst et al. (2005) provide good ways of integrating concordancers
into CAVL.

8.4.3. Lexical programs/tasks

Throughout the literature on CAVL, lexical programs/tasks can be

broadly divided into four types: electronic reading/listening compre-
hension with lexical glosses, CMC lexical-based tasks, computerised
vocabulary exercises, and dedicated CAVL programs. I have arranged
them according to the prominence they give to vocabulary learning in
terms of tool/tutor, implicit/explicit learning, and meaning/form
focusing. When reading online texts (e.g. Chun/Plass, 1996; Laufer/
Hill, 2000) or listening to online audio texts (e.g. Jones/Plass, 2002),
the learner’s primary concern is with comprehending the textual or
audio text. Some words are provided with glosses via hyperlinks
(which can be regarded as a simple mini-dictionary). In the process of
reading or listening, learners may notice an unknown word and
consult the annotation for the meaning, but little attention would be
paid to the word form unless it is strikingly unusual to them. The
learning of vocabulary items is primarily implicit because of a lack of
attention to word forms. This type of online reading or listening is
very similar to traditional paper and pencil forms. Learners only need
to scroll up and down the page or click on the play, pause or review
button in the case of listening. The main function of this type of
application is to execute the learner’s command. They are therefore
tools, although they can be easily integrated into other tutorial
applications.
So far, two types of CMC tasks, asynchronous email and,
particularly now, synchronous communication (chatting in writing),
are widely studied for their potential in promoting L2 development.
Largely within the framework of interactionist SLA, particularly
Long’s Interaction Hypothesis (1996), research into synchronous
communication-based tasks has shown learning benefits in reading
and writing, communication skills, morphosyntactic development,
185
sociolinguistic competence, learner production, learner participation,
learner motivation, etc. (De la Fuente, 2003,). However, only a few
studies (e.g. De la Fuente, 2003; Smith, 2004) have focused on lexical
acquisition. What is supposed to be beneficial to learning in the
communication process is the negotiation of meaning: noticing input
features, modified input, positive or negative feedback, modified
output (or pushed output).
Generally, there are two conditions under which lexical
acquisition may occur. In the first condition, the learner has to notice a
lexical gap from the other interlocutor’s input, i.e. a new word form
that s/he does not know. In this case, the learner needs to show her/his
non-understanding of this word and must ask the interlocutor for an
explanation. The interlocutor might simply give the word meaning, or
give explicit clues to the enquirer so that the meaning can be inferred
correctly. Through the interactional process, the learner establishes a
map between a new word form and the word meaning for the first
time; s/he might use it a few times in the ensuing interaction as a
means to consolidate or practise the word. The interaction in respect
of the new lexical item can be easily interrupted by either side or both
of the interlocutors. In the second condition, the learner may make a
lexical error: give a wrong spelling or an incorrect use. The inter-
locutor notices this error and points it out to the learner or gives
negative feedback. The learner notices his/her lexical error and then
tries to provide the correct output. Because of these necessary
conditions for lexical acquisition, interaction between non-native
speakers and native speakers (or higher proficiency non-native
speakers) will be more likely to promote lexical acquisition. Between
speakers of the same level, lexical-based tasks need to be carefully
designed so that exchange can take place equally for both sides and
result in acquisition for both sides (e.g. De la Fuente, 2003; Smith,
2004). To sum up, although CMC lexical-based tasks are primarily
meaning-focused activities, they furnish evidence of learners’ noticing
of lexical forms explicitly, which is essential for lexical acquisition.
CMC tasks are mere communication tools and how useful they are
depends on the design of the communication tasks.
There are two types of computerised vocabulary exercises. The
first is similar to the simple tutor type of CAVL, aiming to rehearse or

186
test newly learned vocabulary items. They can be useful in consolida-
ting vocabulary knowledge since systematic rehearsal is necessary for
keeping the words in long-term memory. The second type can be used
to learn new words by simulating the vocabulary learning process
based on a specific theory (e.g. Allum, 2004). In both types of
exercises, the computer will evaluate the learner input and give some
feedback. In this sense, the exercises bear tutor features. The learning
is also explicit as considerable attention is paid to both the word form
and the word meaning.
Dedicated CAVL programs are particularly geared to vocabu-
lary learning in a more comprehensive and systematic way compared
with the other types discussed above. That is, vocabulary learning is
both contextualised and itemised; it is both meaning and form focused;
it often combines tutor and tool; both the initial learning process and
the subsequent rehearsal are taken care of. Ideally, the learning bene-
fits may go beyond simply learning vocabulary items: cultural knowl-
edge is promoted and vocabulary learning strategies are inculcated.

8.4.4. The applicability of the framework

The framework that I have developed can be applied to both research-

based and commercial CAVL programs. The six applications that
make up the lexical programs/tasks or lexical resources/aids can exist
on their own or in different combinations in different CAVL programs
or vocabulary learning components of CALL programs. Regarding
research-based programs, both Lexica (Goodfellow, 1994) and
Wordchip (Decoo et al., 1996) comprise a dedicated CAVL program
containing primary learning activities and an attached lexical
dictionary/concordancer as an information resource. On the other hand,
the computerised vocabulary exercises described by Allum (2004)
stand on their own, without being combined with any other type of
application. As for commercial programs, they either favour a type of
electronic dictionary (simple or complex) in the case of dedicated
vocabulary learning programs (particularly those available on the
Chinese market), or computerised vocabulary exercises (with or
without the context in which the vocabulary items occur) in the case
187
of general learning programs (e.g. Tell Me More). Some commercial
programs may be more complex and similar to dedicated CAVL
programs (e.g. Language Interactive Culture)

8.5. Conclusion

CAVL development has reflected the computer’s role as a tutor or a

tool at different periods as in CALL development. We are currently in
the tutor & tool integrated stage where the two roles of the computer
in CALL need to be combined with each other from two perspectives.
Tool and tutor can be further developed as separate CAVL applications
to be integrated into the language learning curriculum, or else as
integrated applications. What emerges is the need to promote strategy
training in this stage, where the focus should be on developing useful
vocabulary learning strategies so that learners know how to learn
vocabulary efficiently (in terms of both quantity and quality) instead
of relying on tuition which will probably result in them adopting
shallow and inefficient learning approaches.
In addition, I have constructed a comprehensive framework to
categorize currently available CAVL applications which fall into two
broad categories: the lexical programs/tasks and the lexical resources/
aids. While the former provide various vocabulary learning activities
for learners, the latter provide electronic dictionaries or concordancers
as lexical aids to the former. The framework is a starting point to
conceptualise comprehensive, systematic CAVL in the general
language curriculum. Based on this framework, we can understand to
what extent a given type of CAVL application is geared to vocabulary
learning and in what sense it is beneficial to vocabulary learning. Then
we must see how we can improve the current systems particularly
regarding the design to increase the learning potential. Finally, and
fundamentally, we need to combine different types of applications so
that they can be well integrated into a systematic approach to
developing SLVA.

188
Chapter 9. Two Evaluation Studies
of CAVL Programs

9.1. CALL/CAVL evaluation

In discussing CALL evaluation, Chapelle (2001: 53) suggests that it

can be analysed at three different levels. The first level focuses on the
CALL software and aims to help teachers, educators or even users to
work out the important design features of the software to be evaluated,
usually in checklists. The nature of this type of analysis is judgemental.
The second level focuses on how teachers plan or use the CALL
activity, which is judgemental too. The third level looks at the learner
performance on the software, which is “conducted through examin-
ation of empirical data reflecting learners’ use of CALL and learning
outcomes” (Chapelle, 2001: 54). Its nature is therefore empirical. This
distinction between judgemental and empirical evaluation at the three
different levels is vital and provides a general framework for guiding
effective CALL evaluation (cf. Hémard, 2004).
In the two evaluation studies of CAVL programs that will be
presented, both research-based and commercial programs are selected.
The research-based programs are selected on the basis of two criteria.
First, the programs should have been actually developed rather than
merely described in terms of their design on a theoretical and
hypothetical level; secondly they should have been empirically tested
with their intended users. Thus for this type of program, both
judgemental and empirical evaluation are reported. As for commercial
programs, they certainly meet the first criterion and very likely the
second one. However, the empirical evaluation results would be very
difficult to obtain for a number of practical reasons. Without denying
that some of the commercial programs like Tell Me More probably
have been widely used in educational settings, the software reviews
such as those that have appeared in the CALICO Journal only focus
189
on the judgemental evaluation and there is no mention of any
empirical evaluation. Therefore, only judgemental evaluation is
reported for commercial programs.

9.2. Constructing the CAVL evaluation criteria

Following Chapelle (2001), the criteria for judgemental and empirical

evaluation can be the same and what differs is the way to generate the
research question(s) and the way to answer the question(s). I suggest
that the evaluation criteria be based on Chapelle’s framework of
“Criteria for CALL task appropriateness” (2001: 55) and Plass’s
framework of domain-specific evaluation criteria (1998). This is to
identify the cognitive processes involved in learning a specific
linguistic skill or competence and then to judge whether the CALL
activities are appropriate in terms of language learning potential and
positive impact. Using Chapelle’s words, language learning potential
is “the degree of opportunity present for beneficial focus on form” and
positive impact is “the positive effects of the CALL activity on those
who participate in it” (2001: 55). Based on Chapelle and Plass’
frameworks, I identified four crucial criteria for evaluating CAVL
programs/tasks; these are set out in Table 6.
These criteria will be employed to evaluate several selected
programs both judgementally and empirically. Normally judgement
can be verified by empirical results, but this is not always the case.
When it is not, an effort needs to be made to identify and explain the
“mismatch” between the judgemental and empirical evaluation. The
first criterion deals with the learning process; the remaining three are
concerned with the learning outcome which is linked with the learning
process.

190
 Evaluation criteria Specific research questions
Vocabulary How is vocabulary processed? Do learners
processing level engage in mental lexical processing that will
Language (learning process) map the lexical meanings with the lexical forms?
Learning Level of vocabulary At what level is vocabulary learned, receptive
potential gain (learning or productive?
outcome)
Vocabulary learning Have learners learned any vocabulary learning
Positive strategies strategies that are useful for learning more
impact (learning outcome) vocabulary on their own?
Learning Are learners motivated to learn more
motivation vocabulary (by continuing to use the
(learning outcome) program)?

Table 6. Criteria for evaluating CAVL programs/tasks.

The first criterion, vocabulary processing level, establishes how

lexical items are processed through various learning activities in the
program, and this serves as a basis to judge and predict the learning
outcome or the vocabulary gain (at different levels). Several SLA
models (e.g. Gass/Selinker, 2001; Vanpatten, 2002) emphasize the
process whereby language input might turn into intake, the latter being
the prerequisite for learning to occur. This process is the decoding and
comprehension of the new language and its integration into the
learner’s existing language knowledge. It also covers vocabulary
processing, which is similar to the notion of “lexical input processing”
defined by Barcroft (2004: 203) as “how learners process words and
lexical phrases to which they are exposed”. The initial vocabulary
learning process comprises three essential components, accessing the
word meaning, decoding the word form, and mapping the two
(Barcroft, 2004: 203). In the initial stage of learning a new L2 word,
seeing that most often the core meaning is shared by many languages 49,
it is the new word form and mapping it with the word meaning that

49 It is possible that a given concept in an L2 does not exist in the L1 of the

learner. In this case, learning the L2 word involves learning both the word
meaning and the word form. However, this phenomenon is not very common.
What more often happens is that a similar or overlapping concept can be
found in the L1, although this might lead to initial lexical confusion in terms
of the word’s meaning and use.
191
requires most effort. This point is clearly stated by several authors
(Kelly, 1986; Laufer et al. 2004; Barcroft, 2002, 2004)50. Thus the
central criterion for judging the initial vocabulary processing level is
to what extent the lexical form is matched to the lexical meaning. This
question can be broken down into the following sub-questions:

1. Is there sufficient opportunity to understand/access the lexical

meanings and sufficient opportunity for attention to be paid to
the lexical forms?
2. Are there multiple input channels (textual, aural, and pictorial)
of the lexical items?
3. Is there sufficient opportunity to map the meaning with the form
and strengthen the link via efficient means such as semantic
elaboration or imagery?
4. Is there sufficient opportunity to rehearse the words in different
contexts (e.g. vocabulary exercises or activities for using the
vocabulary)?

The first question checks the basic condition for vocabulary learning
to occur, i.e. that attention should be paid to both the meaning and the
form of the target lexical item. If there is no noticing of these two
aspects, hardly any learning will occur. The second question checks
whether the lexical items are presented in multiple modes and thus the
multimedia effect will enhance the learning potential, i.e. reinforce the
memory trace of the lexical form and of the mapping with the
meaning. The third question is to ensure that the link between the
lexical form and meaning will be strong enough to be held in memory
so that further learning (e.g. other meanings, usage, syntactic features,
morphological features, etc.) will occur. The purpose of the fourth
question is similar to that of the third, namely, to increase the
form-meaning mapping, not through semantic elaboration but through
direct use of the lexical items (comprehension or production). It is

50 Certainly vocabulary learning is a complicated and incremental learning pro-

cess. After initial successful form-meaning mapping of a given word, there are
still many aspects to learn about the word in terms of knowledge degree
(breadth/depth) and use (reception/production). (See Chapter 1 for a review).

192
assumed that each time the lexical item is comprehended or produced
(free recall or cued recall), the form-meaning mapping is strengthened
and there is more chance of the item being installed as a lexical entry
in the mental lexicon. As soon as a lexical entry is in place, it is open
to the gradual addition of various types of lexical information until it
can eventually be said to be fully acquired.
The second criterion, level of vocabulary gain, is a necessary
result of the previous learning process. If the initial form-meaning
mapping is successful, then the learner can produce the target word
form. If insufficient attention has been paid to the lexical form, the
learner cannot produce the target word but may recognise the word
form when the meaning is given. It could happen that no learning
occurs if no attention is paid to either the lexical meaning or the
lexical form.
The third and the fourth criteria are also concerned with the
possible outcomes of the learning process, not in terms of direct learn-
ing benefit but of the positive impact which the learning experience
(i.e., using the program) may have on further learning. If learners have
acquired some useful vocabulary learning strategies from using the
program, they might learn more vocabulary items on their own in the
future; if learner motivation in vocabulary learning is increased as a
result of using the program, they might make more effort to learn
vocabulary with (or without) the same program. The positive impact
can be measured partially from the learner evaluation (informal or
formal) of the program. If learner evaluation is not mentioned in the
program evaluation, then only guesses can be made as to whether the
positive impact has been achieved.

193
9.3. The evaluation of research-based CAVL programs

9.3.1. Electronic reading/listening comprehension with lexical glosses

9.3.1.1. Reading comprehension

In reading written texts with the help of hyperlinked lexical glosses or
an electronic dictionary, the main emphasis is on reading comprehen-
sion, and the acquisition of any new lexical items is a by-product of
the reading process. The learner can consult the electronic glosses
when confronted with unknown words, which means that the main
learning occurs while reading the glossed information for the new
word. The glossed information can include the word’s pronunciation,
textual explanation, pictures, or videos. Video is supposed to have the
same effect as still pictures, although there is disagreement concerning
which is more efficient in helping vocabulary retrieval (see Chun/
Plass, 1996; Al-Seghayer, 2001).
Laufer and Hill (2000) present a lexical program called Words
in your ear. The main learning activity is the reading of a text in which
the target lexical items are highlighted. The learner can look up the
target words in a mini electronic dictionary which provides five types
of lexical information for all target words: (1) word pronunciation, (2)
meaning in English, (3) L1 translation, (4) lexical root, and (5) extra
information (various forms of the word, register, collocation, syno-
nyms, etc.). In their experiment, learners, made up of a group of Hong
Kong university students and a group of Israeli university students,
could click on any type of lexical information as many times as they
wanted. In addition, the learner could listen to the whole text if s/he
wished. It is reported that most students spent no more than ten
minutes reading the text and checking the lexical information. The
vocabulary processing procedure in Words in your ear is summarized
in Figure 9.
Since the target items to be learned are highlighted in the text,
they are likely to be noticed by learners. Learners can access the word
meaning and other relevant information easily in the electronic
dictionary. The glossed information provides both textual and aural

194
input which is intended to help the learner to strengthen the memory
trace of the target item. However, there is little opportunity for the
mapping of word meaning and form to be enhanced via semantic
elaboration. There is no opportunity to rehearse the word in contexts
other than the reading text itself. Based on these assumptions (or the
judgemental evaluation) regarding the vocabulary processing level, it
can be inferred that primarily receptive learning will occur with little
or limited productive learning. Accordingly, Laufer and Hill reported
a learning rate of 33.3% for Israeli subjects and 62% for Hong Kong
subjects at the receptive level. Productive learning rate was unknown
since no productive measure was used. No evidence emerged that
learners would acquire any vocabulary learning strategy from merely
reading. Nevertheless, the two authors report that one of the two
groups of subjects, the Hong Kong subjects, showed a positive attitude
towards the program and suggested that it be further developed; in
other words, they were motivated by the program and wished to
continue to use it. However, there was no such finding among the
Israeli subjects.

Lexical glosses
Context Words Aural
Textual Textual Textual:
-L2 definition
-L1 translation
-Root information
-Other information

Figure 9. Vocabulary processing procedure in Words in your ear.

9.3.1.2. Listening comprehension

It is sometimes argued that incidental vocabulary learning can occur in
listening comprehension. Smidt and Hegelheimer (2004) report such a
study. The listening program is made up of three components: (1) a
lecture (15 minutes), (2) multiple-choice comprehension questions,
and (3) an online dictionary. The lecture was divided into three parts
according to the type of information provided in addition to the aural
input. The first part only provided the aural input of the lecture plus
195
seeing the head of someone talking on the screen; the second part
provided the aural input and showed keywords of the aural content on
the screen; the third part provided the aural input and the picture slides
giving key concepts of the aural content. Learners could listen to the
lecture repeatedly but the total time devoted to the program (including
doing the listening comprehension questions) was limited to 25
minutes. The vocabulary processing procedure is summarized in
Figure 10.

Context Words Extra lexical information

Aural Aural - Viewing words in textual
form?
- Guessing the words’
meaning correctly?
- Accessing the words’
meaning in the online
dictionary?

Figure 10. Vocabulary processing procedure in the program by Smidt and Hegelheimer.

In this listening task, the major input is through the aural channel and
learners are forced to pay attention to the phonological encoding of
the language input. If an unknown item is encountered, the learner
will probably notice the new phonological representation of the item
but is less likely to grasp the exact spelling for a number of reasons.
Many English words can hardly be accurately spelled according to
their pronunciation. More importantly, accurate decoding of aural
input often involves top-down processing, i.e. the listener has to make
use of his/her grammatical, lexical and extra-linguistic knowledge to
understand what is said. This can only take place when the word is
already known, or at least partially known. Thus, unless the learner
already knows the words, it will be difficult for him or her to decode
the correct spelling of the unknown words by only performing bottom-
up processing, i.e., by relying on the ear. If the unknown words appear
in either the key words section or key concepts section of the lecture,
the learner has the opportunity to view the correct word form.

196
However, there is no confirmation of this in the two authors’ report. It
is likely that some of the unknown words appeared in written form.
Similarly, the question also arises whether the learner has correctly
understood the meaning of the unknown words or not. Theoretically,
the learner can guess the word meaning or use the online dictionary,
but there is no guarantee that their guessing is right. Nor do the two
authors report any use of the dictionary by the learner, given that the
study time is only limited to 25 minutes. Regarding vocabulary pro-
cessing level, this listening comprehension provides some opportunity
for the learner to pay attention to the lexical forms and meanings, and
provides both aural and textual input of (some) of the target items. As
with the previous reading comprehension program, there is no
evidence that the form-meaning mapping is strengthened by semantic
elaboration or other learning activities.
It can be predicted that vocabulary learning in this program is
limited both receptively and productively. The vocabulary measure is
a dictation test in which the target items are left blank and which the
learners are required to fill in. This is a cued recall test or partial
productive test, where the context and the pronunciation of the target
item serve as the cue to assist in the recall of the item. The result
showed that the subjects had an average gain of 3.2 words in the
post-test out of the 20 selected target words 51 of which they already
knew 3.8 words on average, as shown from the pre-test. Therefore the
learning rate is 19.8% productively. No receptive test was ad-
ministered. No positive impact of the program was reported regarding
both vocabulary learning strategies and continued motivation to use
the program.

9.3.1.3. CMC lexical-based tasks

If learners are left on their own to chat with their co-learners freely
about a topic, they might learn a few new words or improve their
knowledge of the words partially known before, providing one (or
both) of the two conditions for lexical acquisition are fulfilled.

51 It should be noted that these 20 items were identified as the most difficult
words in the lecture, including both individual words and two-word combin-
ations.
197
However, what is more likely is that learners use the words or
expressions they already know or that they do not bother to find out
more about an unknown word or to correct their lexical errors. In this
case, new learning is unlikely to take place. In controlled task-based
learning, necessary conditions to ensure learning can be created and
incorporated into the task. De la Fuente (2003) reported such a lexical
task-based CMC learning activity. 14 Spanish words (concrete nouns)
were selected as target words to be learned in the activity. 12 learners
were divided into six pairs of two and each pair worked on the same
learning task to fill in information on a gap format. In the first session,
one subject (of the pair) had a shopping list in which seven Spanish
words (the Spanish equivalents of, e.g., pumpkin, celery, etc.) with
their corresponding pictures were provided. This subject had to tell the
other subject what to buy from the shopping list via the computer.
Subjects were told to use only Spanish to communicate with each
other. This meant that they had to negotiate the meaning of each
Spanish word by giving meaning clues or by asking for meaning
information about the word. In the second session, the role of each
subject of the pair was reversed, i.e., the subject who did not have the
shopping list in the previous session would have the list and instruct
the other student what to buy. The vocabulary processing procedure is
illustrated in Figure 11.
In this task-based learning, the learner has sufficient opportunity
to notice the target word forms because of the task requirement and
the textual input. The task requirement also required the learner to find
out the meaning through negotiation with the other subject who had
the information concerning the meaning of the target word. The
interaction between the two subjects in trying to negotiate the meaning
can be seen as a type of mental elaboration, which includes meaning
inferencing and forming mental images or associations. In addition,
the learner has the chance in the second learning session to practise
the words when s/he is supposed to give the word information to the
other learner. The learner can type the word form and give different
(or the same) semantic clues to the other learner, which again can
strengthen the form-meaning mapping. On the other hand, if the items
to be learned are not concrete nouns but abstract words, the process of

198
negotiation of meaning might be more complicated and difficult to
handle.

Context Words Negotiation of the words’

Textual Textual meanings
Receiving semantic clues:
- Object features: colour,
shape
- Associated events with the
Negotiation of the words’ words: e.g. pumpkin and
meanings Halloween
Giving semantic clues:
- Object features: colour,
shape
- Associated events with the Words
words Textual
(output)

Figure 11. Vocabulary processing procedure in the CMC lexical-based task of De la

Fuente.

Given the evidence that learners have the opportunity to notice both
the lexical meaning and form, and the opportunity to strengthen the
link via mental elaboration or by using the words, both receptive and
productive learning can occur. The empirical finding of the study does
confirm this: subjects have gained vocabulary both receptively and
productively in terms of both written and oral acquisition. Only the
immediate post-test is reported here although the vocabulary gain was
measured three times. The receptive oral gain, productive oral gain,
receptive written gain, and productive written gain are 7.9, 3.3, 10.5,
and 6.2 respectively out of a total of 14. The learning rates of each
type of vocabulary gain are accordingly calculated to be: 56.4%,
23.6%, 75%, and 44.3%. The mental elaboration of the meaning of the
item and learning vocabulary collaboratively were probably the only
useful learning strategies but they had not been made explicit to the
learners. It is doubtful whether learners will continue to use these
strategies consciously and systematically in later learning after the
task. The study did not report whether learners were motivated by
199
participating in the CMC lexical learning task. Whether learners are
motivated to use the task depends largely on the perceived helpfulness
of the task, i.e. is the task helpful in improving their language
proficiency? Generally speaking, CMC learning tasks have received a
favourable response from learners as there appears to be learning
improvement, active participation, interactive/collaborative learning
environment (Kitade, 2000), and reduced anxiety for communication
(Kern, 1995). It is likely that when the CMC lexical-based task is well
designed learners can improve their vocabulary learning and will be
motivated to participate in similar tasks.

9.3.1.4. Computerised vocabulary exercises

Computerised vocabulary exercises can be used either to rehearse the
previously learned vocabulary items or to learn new vocabulary items.
Allum (2004) describes a series of computerised vocabulary exercises
by simulating the three stages of learning new vocabulary items
described by Nation (2001): noticing, retrieval and generative use.
Noticing means paying attention to the target lexical item in language
input via “decontexualisation” (Nation, 2001: 64), i.e., separating it
from the flow of language message in which it is situated. The main
purpose of decontextualisation is to find out the meaning of the item
and establish the initial form-meaning mapping. Retrieval means
meeting the lexical item subsequently and repeatedly; the word can be
retrieved either receptively or productively. Each successful retrieval
will strengthen the form-meaning mapping and the memory trace of
the item in the mental lexicon. Generative use involves meeting the
lexical item after initial learning in different contexts so that the
learning can continue, i.e. other meanings can be acquired, colloca-
tions and grammatical patterns can be learned.
The three stages of vocabulary learning are implemented using
three types of computerised vocabulary tasks in Allum’s study. The
noticing stage is achieved by matching L2 words (forms) with L1
words (meaning). It is a drag and drop exercise and learners will try to
guess the meaning; if they are wrong, they will try again till the
correct match is found with the help of the feedback. The retrieval
stage is realised by receptive or productive use of the target words.

200
Learners are required to type in the target word in a sentence which
provides the word definition (L2) in the case of productive use. The
receptive exercises include matching a definition sentence to a word
or choosing the right word to complete a definition sentence. The last
stage, generative use, is implemented by tasks similar to those of the
previous retrieval stage, involving both receptive and productive
retrieval of the target item but in different contexts (sentences which
contain the target word). Students are required to learn vocabulary
using the computerised exercises on their own 52 as a preparatory
activity, and then these vocabulary items will be used more communi-
catively (speaking and listening activities) in classroom learning. The
vocabulary processing procedure is summarized in Figure 12.
The initial learning of vocabulary items (before classroom
activities) as described in the three types of exercises above is
adequate for establishing the initial mapping of word forms and word
meaning. Attention has been paid to both the lexical form and
meaning and considerable effort is made to link the two through
subsequent retrieval exercises both receptively and productively.
Although the language input in all exercises is in textual form, the
subsequent classroom activities provide opportunities for learners to
be exposed to or to use the target items aurally or orally. It can be seen
from Figure 12 that this is a good example of how to integrate CALL
programs/activities into the language classroom, or even broadly into
the overall language learning curriculum. The result of the first study
by Allum reports an overall vocabulary gain (or learning rate) of 50%
both receptively and productively. Different from many studies where
CALL activities are conducted for purely experimental purposes
involving typically one or two learning sessions, the study reported by
Allum is conducted in an authentic classroom learning situation.
Despite a relatively long period of using the computerised vocabulary
exercises, learners generally reported favourable attitudes towards
them and acknowledged that they did more work than before. If
learners find that the learning activity can help them to improve their

52 Although Allum points out that the vocabulary exercises were designed for
self-learning (at home), in the experiment students were asked to do them in
class time.
201
language learning, they will be motivated and spend more time doing
the activity. However, it seems that learners have little opportunity to
learn useful vocabulary learning strategies that can assist subsequent
learning when the computerised exercises are not available.

Words L1 meaning Retrieval

Textual Textual Receptive: meaning
Productive: form
Textual

Classroom activities Retrieval in new

- Speaking (oral) context
- Listening (aural) Receptive: meaning
- Others Productive: form
Textual

Figure 12. Vocabulary processing procedure in the computerised exercises by Allum.

9.3.2. Dedicated CAVL programs

Dedicated CAVL programs are similar to computerised vocabulary

exercises in that they are often based on theories in which certain
necessary learning stages or procedures for vocabulary acquisition
will be simulated and implemented via various learning activities. But
they tackle vocabulary learning in a more systematic, comprehensive
way than the latter. The learning benefit should not be restricted to
direct vocabulary gain but should also bring other benefits such as
cultural learning (e.g. Van de Poel/Swanepoel, 2003) or vocabulary
learning strategies (e.g. Goodfellow, 1994). In operation, dedicated
CAVL programs often integrate different functions of other types of
programs. For example, reading/listening comprehension can be used
to introduce lexical items to be learned; vocabulary exercises can be
used as rehearsal activities; CMC tasks can also be potentially

202
integrated into the whole learning system (e.g. Van de Poel/Swanepoel,
2003).

9.3.2.1. Lexica
The program Lexica described by Goodfellow (1994, 1995, 1999) is
composed of three sequential learning modules: the selection module,
the lexicon-building module, and the testing module. In the selection
module, the learner reads through the text and highlights the unknown
lexical items that s/he is interested in. S/he then assigns these items to
one of the three groups: meaning, form and context. This is done based
on the judgement of the learner who must decide whether the word
meaning, the grammatical, formal features, or the context (surround-
ing words or collocation) are important to a given item. In the
lexicon-building module, learners browse the words saved in each
group and study them with the help of two lexical tools: an L1
dictionary and a concordancer. Learners can access the word meaning,
synonyms and collocations via the two lexical tools; they can also add
notes to the items or group the items under a sub-category based on
the similarity shared by some items. In the testing module, learners
retrieve the lexical items they studied in a cloze test with the original
texts where the items initially occur. All information the learners
saved for the items (group names, notes, etc.) is displayed as retrieval
clues. The test is a cued recall test, testing the production of the items.
The vocabulary processing procedure is described in Figure 13.
Since the lexical item to be studied is selected from the text by
the learners themselves and is initially categorized on the basis of the
word’s characteristics, sufficient attention should have been paid to
the word form. If the learners have used all the facilities in the lexicon
building module, e.g., checked the word meaning, its usage, common
collocations, taken notes on the word, reassigned them to new
sub-groups, there should be sufficient opportunity to establish the
initial meaning-form mapping and strengthen the memory trace of that
mapping. The subsequent retrieval of the items on the basis of various
clues should have further strengthened the mapping. The word
processing level should therefore be judged to be good. However, the
empirical data show that this is only half the story. Only half of the
203
learners, who adopted the deep learning approach intended by the
program design, processed vocabulary to a satisfactory level; the
remaining half, who adopted a quick or shallow learning approach, did
not. Goodfellow (1999: 123) distinguishes the deep and surface
approaches as follows:

Deep approaches to language learning, which focus on the relation between

contextual meaning and linguistic structure, are presumed to guarantee quality
in both process and outcome. Surface approaches, which emphasise meaning
at the word level, L1 equivalence and memorisation, may indicate the absence
of quality.

It is revealed in the empirical data that some learners spend less time
processing selected items (in the second module) but more time
retrieving the items repeatedly till the correct answer is found (in the
third module). In addition, the majority of learners have not done
sufficient grouping of the selected items. Among the word groups
created by learners there exists a remarkable difference in terms of the
quality of semantic links. A correlation is shown between the deep/
surface approaches adopted and the quality of the learning outcome;
deep approaches tend to lead to higher learning rate and more accurate
retrieval while surface approaches tend to lead to a lower learning rate
and less accurate retrieval.
As Goodfellow (1999: 126) points out himself, although the
design of the program encourages learners to use certain vocabulary
learning strategies, such as establishing semantic links between words
or creating word associations in a network, this is never made explicit
to the learner and no instruction or tutorial training is provided to
make learners aware of the importance of doing so. Thus the benefits
regarding vocabulary strategy learning are limited, as evidenced by
the surface learning approaches adopted by half of the learners. There
is no reporting on learners’ attitudes towards the program and whether
learners are motivated to use the program for learning more vocabu-
lary perhaps depends to what extent they feel the program is helpful to
them.

204
 Context Words Grouping words
Textual Textual - Meaning
- Form
- Context
Textual

Word elaboration
- L1 glosses
- Concordancers
Retrieval - Notes
Cued productive - Sub-grouping
recall Textual

Figure 13. Vocabulary processing in Lexica.

9.3.2.2. CAVOCA
Groot (2000) described a computer assisted vocabulary learning pro-
gram CAVOCA (Computer Assisted VOCabulary Acquisition), which
is based on the work on the mental lexicon. Three stages of acquiring
vocabulary are proposed:

1. Noting the various properties of the new word: morphological

and phonological, syntactic, semantic, stylistic, collocational,
and so forth.
2. Storage in the mental lexicon in networks of relationships that
correspond to the properties described above.
3. Consolidation of the storage by means of further exposure to the
word in a variety of contexts which illustrate its various pro-
perties.

The learning process in CAVOCA is composed of four stages in

sequential order: deduction, usage, examples and retrieval. In the
deduction stage, the learner views a word and then three sentences
appear in which the word occurs, presented in order of contextual

205
richness. Then the learner is required to do a multiple-choice exercise
by finding the nearest meaning or synonym of the target word.
Immediate feedback is given to prevent learning the wrong meaning
by wrong guessing. According to Groot, the learner is required to infer
the meaning by hypothesizing and testing the word meaning as a way
to processing it intensively. However, the difficulty and constraints of
inferring the meaning correctly from the context either in L1 or L2 are
noted by a number of researchers (Ames, 1966; Kelly, 1990; Beck,
McKeown and McCaslin, 1983; Sternberg and Powell, 1983). The
learner might not follow the “graded contextual disambiguation”
process (Groot: 66) by viewing the three example sentences carefully
and may, instead, take an effortless short-cut by guessing the meaning.
The multiple-choice questions are very likely to increase this
possibility. If so, the objective of “deduction” to process the word
intensively is not achieved.
In the usage stage, the learner is presented with two sentences
and required to choose the one in which the target word is used
correctly. Immediate feedback will be given as well as additional
information if relevant (other meanings, derivatives, idiomatic usage,
etc.). This stage aims to consolidate the new words in the learner’s
mental lexicon and illustrates the exact meaning. The same danger
might occur as in the first stage since the learner is once again asked
to make a choice (or a guess).
For the examples stage, a number of authentic L2 passages
containing the words the learner has just learned are presented, aimed
at consolidating and ensuring long-term retention of the target words.
The intention is good but we should not forget that to achieve this aim
depends entirely on whether the learner will read these passages
carefully or not. S/he might not do so because s/he considers them too
difficult or uninteresting.
In the lexical retrieval stage, two types of test, a translation test
(from L2 to L1) and gapped cloze test, (where the initial letter of the
word is given) are used; these test the receptive and productive
knowledge respectively. Immediate feedback is given. The vocabulary
processing procedure is summarized in Figure 14.

206
 Words Words in sentences
Textual Textual

Retrieval Deducing words’ meanings

Receptive and Textual
Productive recall

Words in context Deducing words’ usages

Textual Textual

Figure 14. Vocabulary processing in CAVOCA.

It is noticed that there is no explicit lexical information given to the

learner, rather, the learner has to deduce the meanings and usages of
the words. Theoretically, the wrong deduction will not do much harm
since immediate feedback is given. However, the absence of explicit
lexical information in CAVOCA means that the learner has to make
the mental effort to find the word meaning first and then match it with
the word form. On the one hand, these semantic processing mental
efforts will be considered to be at a deep processing level in terms of
the memory processing theory of Craik and his colleagues (Craik/
Lockhart, 1972; Craik/Tulving, 1975). On the other hand, a good deal
of deducing of the word’s meaning and usages will encourage mere
guessing since learners are prone to take short-cuts and perform
activities requiring less mental effort. This implies that the learner
either performs deep mental processing as expected and achieves a
good learning result, or else s/he goes through the program guessing a
great deal with a poor learning result. This could explain why the
control group, who used the bilingual word list, did better on the
recognition test and the experimental group, who used CAVOCA, did
better on production.

207
 The learning rates reported are 73%−70% for recognition and
40%−70%53 for production in the immediate test. Regarding vocabu-
lary strategy learning, through simulating all the necessary steps for
acquiring a word based on the mental lexicon in a condensed way,
CAVOCA gives learners some ideas how vocabulary learning should
be tackled. However, this instruction has never been made explicit and
thus the impact is limited. It is quite unlikely that the learner would
obtain any learning strategies that can be extended to learning other
new words outside the program. There is no mention of learners’
attitudes towards the CALL program to indicate whether the learners
were motivated or not by using the program.

9.4. The evaluation of commercial CAVL/CALL programs

9.4.1. Three major types of commercial programs

The research-based CAVL programs include all four primary types of

application, ranging from tool/implicit/meaning-focused electronic
reading/listening comprehension with lexical glosses to tutor/explicit/
form-focused dedicated CAVL programs. In contrast, the commercial
programs largely include electronic dictionaries/flash cards, com-
puterised vocabulary exercises, or dedicated-like CAVL programs54,
all treating vocabulary learning explicitly. In fact, the first and the
second categories are quite similar to each other: both would include
an electronic dictionary or glossary (simple or complex) and a number
of vocabulary exercises. What differs is that in the former the interface
would be designed as a dictionary or a flash card system where the

53 There are four experimental groups. The first and second groups were only
tested receptively and the third and fourth groups were tested productively.
54 I used dedicated-like instead of dedicated because some of the programs to be
reviewed later claim to be for general language training rather than focusing
on vocabulary. Nevertheless, vocabulary learning is addressed in a compre-
hensive, systematic way in these programs.

208
lexical items are to be studied and then practised in the vocabulary
exercises, whereas in the latter only a simple electronic glossary is
available and there are a considerable number of vocabulary exercises
of different kinds. Generally speaking, commercial programs tend to
combine both tool and tutor. For example, an electronic dictionary on
its own is a tool, but the vocabulary exercises that follow often give an
evaluation of the learner performance and thus they can be considered
as tutor. So do dedicated-like CAVL programs where tutor and tool are
often combined.
It seems to be a tradition for vocabulary learning to be treated as
explicit learning in commercial programs, particularly in the form of
vocabulary exercises. This may be a lasting reflection of the drill and
practice which existed at the beginning of CALL history. As men-
tioned on earlier occasions, drill and practice are not necessarily
harmful. Decoo (1994) argues that, if certain criteria are met, drill and
practice have more advantages over other more communicative types
of CALL programs in classroom instruction. Similarly, N. Ellis (1995a)
points out that drill and practice are necessary for learning vocabulary,
at least in the initial stage of learning an L2. However, they should not
become the dominating language learning activities; they have to be
appropriately combined with other learning activities where language
is used in a meaningful, communicative way.

9.4.2. Electronic dictionaries/flash cards

In this type of program, the electronic dictionary serves not only as a

tool for finding lexical information but also as the primary source for
learning new vocabulary items. The simplest type is to present the L2
item and then provide the corresponding L1 translation; learners are
supposed to match the two, much in the same way as the traditional
flash cards. Example sentences or pictures (if relevant) can also be
given. A more elaborate type is a standard electronic dictionary, pro-
viding all sorts of lexical information, L2 definition, usage, and
synonyms. After the vocabulary items which are supposed to be
learned are presented, vocabulary exercises often follow as a means to
consolidate the learning.
209
9.4.2.1. SuperMemo
SuperMemo 55 is a software package, first developed in Poland in
1985 as a learning method to enhance memorisation of learning
materials. The main pedagogical feature lies in the practical applica-
tion of the spaced repetition in which learning materials are reviewed
at increasing intervals of time. Learning, including language learning,
essentially consists in committing the new knowledge to long-term
memory. In this sense, repetition, or its more efficient variant, spaced
repetition, is a necessary means for learning. In operation,
SuperMemo only serves as a means to store and manage information;
it does not contain any learning materials and the learner needs to add
learning materials to it. It was soon discovered that SuperMemo could
be conveniently applied to language learning, particularly vocabulary
learning. To start with, the learner puts the vocabulary item and the
related lexical information into the system. The information is stored
in two parts, the first part serving as the prompt question, and the
second as the answer. Here is an example:

Question: What do you call the sister of your husband or wife?

Answer: sister-in-law.

It can be seen clearly that this way of storing the lexical information
resembles the way the traditional flash cards work. There are certainly
other ways of organising the information in the two parts. The
simplest way is to put the word form as the question and the word
meaning (in L1 or L2) as the answer, or vice versa. Other information
can also be added, such as a phonetic transcription or example
sentences, depending on the learner’s needs. Many words can be
added to the system in this way. After storing the words, the learner
can begin reviewing them. S/he first clicks on the button “learn”, the
information stored in the question will appear on the screen, e.g.
“What do you call the sister of your husband or wife?”, then the
learner is supposed to retrieve from memory the answer to the

55 See http://www.supermemo.com/ for more detailed information; the free ver-

sion of SuperMemo 98 can be downloaded from <http://www.freewarefiles.
com/program_16_168_13849.html>.

210
question, in this case “sister-in-law”. The learner then clicks on the
button “show answer”, and the stored answer will appear below the
question. The learner is supposed to compare the answer retrieved
from memory with the correct answer given. If the retrieved answer is
given correctly and unhesitatingly, then the learner clicks on the
button “Bright”; if with some hesitation but correct then the button
“Good”; if correct but with substantial effort then “Pass”; if almost
wrong then “Fail”; if completely wrong then “Bad”; if nothing is
retrieved then “Null”. The system will store this information and
schedule how this item shall be reviewed in the future. Meanwhile, the
learner can go on reviewing other stored words. It is claimed that after
one or two months of using SuperMemo, the system will know how
many retrievals are needed for each word for the learner to remember
the word. This is how SuperMemo has been applied to vocabulary
learning. However, I have not been able to find any information from
its official website or other related resources that explains in detail
how the review of words is scheduled.
Since all lexical information is provided by the learner, it is
assumed that s/he has paid attention to both the lexical form and
meaning. The stored lexical information is primarily textual, as with
traditional flash cards. Although it is said that audio information can
be added to the item, the free version of SuperMemo 98 I reviewed
does not seem to support this function. It is possible that future ver-
sions may incorporate pictures into the system. However, there would
be a huge amount of work on the learner’s part if s/he has to prepare
all the lexical information her/himself. This means that the quality of
the lexical information for learning largely depends on the way the
learner inputs the information. The main activity in the program is to
try and retrieve the word, or match the word form with its meaning
through the repetition scheduled by the system. It is assumed that each
retrieval will strengthen the meaning-form mapping of the item. Apart
from this, there is no contextual use of the item. So, even if eventually
the word meaning is matched to the form in the lexicon through
repetition, it may not be correctly or appropriately used if there is no
other opportunity to meet and use it in different contexts. Regarding
the level of vocabulary gain, it can be both receptive and productive,
depending on the way the lexical information is stored. If the word
211
form is stored as the question and the word meaning as the answer, the
learning tends to be receptive; if the reverse, the learning tends to be
productive. Finally, the amount of repetition could have an influence
on the level of vocabulary gained. Apparently the software introduces
no other vocabulary learning strategy other than repetition. Whether
the software will motivate learners to continue to learn words with this
method largely depends on the learner’s perception, on whether s/he
finds it useful and enjoys this learning method.

9.4.2.2. Rosetta Stone

The language learning packages produced by Rosetta Stone 56, based in
the United States, include 30 languages; it probably has the widest
language coverage of any language learning software. Rosetta Stone
claims to use the “natural” way to help learners learn a new language
by immersing them in the language and providing training in the four
basic skills: reading, writing, listening and speaking. Although the
packages are for general language learning, a substantial number of
lessons, particularly at the beginner level, are devoted to vocabulary
learning. I reviewed the free online demo version for level 1 of
German. Each lesson is organised in five sections: listening & reading,
listening, reading, speaking and writing. There are two activities in
each section: the preview of the learning materials and the guided
exercises.
We start with a preview of the section listening & reading of the
first lesson of level 1: nouns and prepositions. There are four pictures
on the screen. The word that depicts the picture is printed on each
picture: ein Mädchen , ein Junge, ein Hund, eine Katze. The pictures
reveal that they mean a girl, a boy, a dog, and a cat respectively. The
learner can also click on a listening button on the corner of each
picture to listen to the pronunciation. There are four other options. The
learner can: (1) hide the text; (2) hide the picture; (3) record her/his
voice; (4) compare it with the original pronunciation. In addition, s/he
can repeat viewing, listening or practising the four words as often as
desired before going on to the next page. There are 10 such pages and

56 See <http://www.rosettastone.com/en/ for more information>.

212
hence 40 pictures in this section. The first three pages, or the first 12
pictures, introduce 12 new words. Starting from the fourth page,
lexical items are put in pairs in each picture, e.g. eine Katze und ein
Auto, ein Mädchen und eine Frau, ein Mann und eine Frau, and ein
Mann und ein Jung. Again there is no problem accessing the meaning
via the pictures: a cat and a car, a girl and a woman, a man and a
woman, and a man and a boy. There are two purposes for introducing
such combinations: (1) to review the words viewed in the previous
pages and (2) to introduce the conjunction and. The remaining six
pages introduce three prepositions − auf (on), unter (under) and in (in)
− and review the previously encountered words. When the preview is
finished, the learner can start the guided exercises. The same 40
pictures will reappear in the same way as in the preview. For example,
eine Katze, will be pronounced and the written form given. The
learner has to choose the corresponding picture and click on it.
Immediate feedback is given indicating whether the response is right
or wrong. This procedure is repeated until all 40 pictures have been
gone through. In the next section, listening, the preview follows the
same pattern as in the listening & reading: the same 40 pictures will
reappear. What is different is that the learner can only click on each
picture to listen to the lexical item and does not have the help of the
written form. The guided exercises follow the same pattern: a word is
pronounced (without the written form) and the learner has to choose
the right picture. The remaining sections, reading, speaking and writ-
ing, follow the same pattern, with the learning activities corresponding
to the type of language skills to be focused on.
The general impression is that it is very user-friendly. The main
pedagogical features can be summarised as: (1) efficient use of the
flash card method; (2) use of real pictures accompanied by textual and
audio input; (3) application of incremental learning where new infor-
mation is carefully integrated with old information. The first section,
listening & reading, is very helpful; three types of lexical information
are available simultaneously in the preview: textual, visual, and aural.
The combination of the information in different channels, particularly
the use of images, facilitates the encoding of the new lexical infor-
mation and stores it in memory. The options available, such as hiding
the text or the pictures, can easily turn the system into user-friendly
213
electronic flash cards. The subsequent exercises, which involve
repeated retrieval of the lexical information with aural or textual cues,
enhance the form-meaning mapping of the lexical items. Regarding
the level of vocabulary gain, it appears primarily receptive as all the
exercises are for recognition but not for production; even the exercise
in the writing section is to assemble given words and phrases to form
a sentence. Certainly vocabulary gain at the productive level is also
possible, as receptive learning entails a certain amount of learning,
particularly when considerable repetition is involved. The most
obvious strategy introduced in this software is to use multiple media
to facilitate vocabulary learning, particularly by associating the lexical
meaning with real images. Although this strategy is not made explicit
to the learner, some can be expected to extend it to the learning of
other lexical items elsewhere. Generally speaking, this software is
attractive and learners, particularly beginners, would be motivated to
use it. One point worth mentioning is that the software only uses
pictures to explain the meaning of the lexical item thus excluding the
use of the L1; this might work for concrete items, as in the case of the
lesson reviewed, but it might run into difficulties with less imageable,
abstract words. Hammerly (1984) tested language teachers on solely
using pictures to convey meaning and found there was only a
49%−59% chance of being successful! This might therefore be a
serious weakness of the software.

9.4.2.3. Intelligent Miracle English

Intelligent Miracle English57 (IME) is an English vocabulary learning
software package developed in China in recent years. It is similar to
SuperMemo in that it uses spaced repetition to help learners to
memorise the lexical items; it boasts a learning rate of 60 words an
hour. Different from SuperMemo, IME is not only a device to manage
the lexical information but also pre-stores a huge amount of lexical
information; different vocabulary sets58 are incorporated catering for
different levels and needs.

57 See <http://www.qjnet.net/ for detailed information about this software>.

58 All vocabulary items are taken from published textbooks in China; each vo-
cabulary set is taken from one book. Currently there are 338 vocabulary sets.

214
 I reviewed a demo version downloaded from the website. To
begin with, the learner needs to choose a vocabulary set. There are
two activities for each vocabulary set: to memorise the words and to
practise spelling the words. To begin, an English word appears on the
screen and the learner is supposed to hear its pronunciation
immediately. However, this is not always the case and sometimes the
sound is not clear (it is said that both real voices and artificial voices
are used). The learner is asked to make a choice between “I remember
it” and “I forgot it” by clicking on the corresponding button. Then the
word meaning will be given in Chinese; the word class is also
indicated. The learner can also request example sentences of the word,
which are nonetheless not always available. Now the learner is asked
whether the earlier judgement that s/he made regarding whether s/he
knew the word was right or not by clicking on either the button
“correct” or “wrong”. If the learner chooses “correct”, this word will
not appear later, at least in this learning session. If the learner chooses
“wrong”, that means the word meaning s/he retrieved is different from
the meaning displayed. Then this word will reappear at least three
times in subsequent learning. The first time is immediately after the
next word that comes up on the screen; the second time is after two
words have appeared; and the third time after nine words, on condition
that the learner clicks on the button: “I remember it” each time the
word in question appears. Then this word will not appear any more.
If the learner chooses the button: “I forgot it” when a word first
appears, the word meaning and example sentences (upon request) will
be given. Meanwhile, a message appears to remind the learner:
“Please listen to the word five times and read after it, trying to
remember this word.” Likewise, this word will reappear at least three
times subsequently. The learner can work his/her way through to the
end of the list. At the end of the learning session the learner will be
asked whether s/he would like to take an optional test on the items
learned. If the response is yes, the test will begin. It is a recognition
test involving choosing the Chinese equivalent of an English lexical
item. At the end of the test, a report is given indicating which words
have been answered correctly and which words wrongly.
The second activity is to spell the words, in the form of cued
recall. The interface is quite similar to that of the previous memori-
215
sation activity. Initially, a word is pronounced and the word meaning
is given. The learner has to spell the word in the space provided.
Immediate feedback is given. If the answer is correct, this word will
not appear later; if wrong, it will reappear three times as in the
memorisation activity. After all the words have been gone through, the
learner is asked whether s/he would like to take a test. Surprisingly, it
is still the same recognition test as in the previous activity; this means
the test form, receptive in nature, is inconsistent with the type of
learning, which is supposed to be productive.
Compared with SuperMemo, IME is easier to use since the
learner does not need to input the lexical information her/himself. As
in SuperMemo, for IME attention has been paid to both the word form
and the word meaning, and the spaced repetition has enhanced the
form-meaning mapping for the words. Moreover, the option of dis-
playing example sentences, if available, can give information about
how the word is used. The memorisation activity is intended to make
the learner pay attention to both the word form and the word meaning;
the spelling activity provides good training to help the learner to
master the correct spelling. In this sense, the vocabulary gain can be
both at the receptive and the productive level. As with SuperMemo,
there is no introduction to vocabulary learning strategies other than re-
petition. Whether the learner will be motivated to use this software to
learn more vocabulary items depends on his/her perceived usefulness
of the software.
Nonetheless, there are three problems with the learning ap-
proach in this software. We shall look at them each in turn:

1. From my experience of learning unknown vocabulary items

with this software, it seemed that repetition on its own will not
necessarily ensure a word meaning being mapped on to the
word form, at least not efficiently. For example, I had great
difficulty in remembering the meaning of the word inordinate,
meaning excessive. I kept pressing the button “I forgot it” when
this word appeared. Eventually I looked at this word and
decided to break it into two parts: in and ordinary(ate). I told
myself that if something is not ordinary then it must be out of
the ordinary in some way. After making this simple and

216
 somewhat inaccurate word association, I remembered this word
meaning eventually. Certainly, the degree of difficulty in word
meaning-form mapping varies greatly, from one person to
another and from one word to another.
2. Apart from the lack of help in memorising difficult 59 words, it
is problematic to use only the L1 translation to provide the L2
word meaning. Without denying that L1 translation is an
important tool when learning an L2, particularly at the
beginning level, overuse of it in terms of building equivalent
L1-L2 pairs can have negative effects, particularly when there is
considerable cultural difference and linguistic distance between
the two languages. Even if the initial word meaning-mapping is
established, mapping the L2 vocabulary on to the L1 vocabulary
is one of the major sources of lexical errors (Swan, 1997). This
simplistic L1-L2 mapping will inevitably lead to lexical errors
in meaning, form, cultural connotation, and usage.
3. Although example sentences are provided in the majority of
cases, the software has put too great a focus on itemised
learning, ignoring the contextual roles in learning the L2 items.
Knowledge of vocabulary items learned exclusively in itemised
learning is likely to be incomplete or even inaccurate, in terms
of both word meaning and word use. In addition, the items may
not be connected with other items in the mental lexicon, but
stand in isolation.

Based on the analysis above, it would seem that the software in its
current form is best used as a means to review the words learned in
more contextualised or communicative learning since there are
activities for retrieving both the word meaning and the word form. For
the software to be used efficiently for learning new words, more con-
textual information should be given e.g. adding lexical concordancers
and other types of activities/exercises where words can be used in
different contexts.

59 Word learning difficulty can be idiosyncratic or L1-related; see Chapter 4 for

a detailed discussion of this issue.
217
9.4.2.4. Memorising Vocabulary Effortlessly
Memorising Vocabulary Effortlessly (MVE) is another software
package developed in China for learning English vocabulary; it is
claimed that the software has been popular on the market for more
than ten years. I reviewed a recent standard version of the program. As
with IME, the first thing the learner has to do is to select a vocabulary
set to work with. Then the vocabulary items will appear one by one
for the initial learning activity. The learner can click on each word to
listen to the pronunciation and view the word meaning (in L1) or one
or two example sentences (if available). Alternatively, the word can be
displayed automatically at intervals (e.g. every 10 seconds) pre-set by
the learner. When s/he has gone through all the words in this way, s/he
can take a test; there is no specific order and the learner can take the
test even before studying the new words. The test is a cued recall test.
The word meaning (in L1) is given as a retrieval cue by default and
the learner has to type in the word form. To make this task less
difficult, the learner can request more cues, such as the pronunciation,
the phonetic transcription, initial letters, etc., to facilitate the word’s
retrieval. Immediate feedback is given whether the word is correctly
spelt or not. If wrong, it will be presented later at least twice till no
further mistake is made. When the feedback is received, the word
meaning and the example sentences (if available) will appear again.
On completing the test, the learner will receive a list detailing the
vocabulary items that are considered to be known and those that are
considered to be frequently wrongly spelt. In this sense, the system is
more intelligent and more interactive than IME.
The next activity is to review these words by testing oneself;
this is almost the same recall test that was taken previously. The only
difference is that feedback would be given only after the test is
completed. After receiving the feedback, the wrongly spelt words will
reappear and the learner has to spell each word at least twice. When
this is done, the learner will be asked whether s/he would like to view
all the words once again. The words will appear one by one, first the
L2 word form and then the L1 translation; the learner is supposed to
recall the meaning on seeing the word form and then compare it with
the correct meaning provided.

218
 The third activity consists of a recognition test, i.e. the learner is
required to recall the word meaning, either from the L2 word form to
the L1 meaning or vice versa. The fourth activity is a dictation
exercise; the learner is asked to write down the example sentences
which have been provided in the initial learning activity. Different
cues are given to make the task less difficult: keyword, translation of
the sentence, initial letters of each word, and indication of the wrongly
provided letters. This activity seems to be helpful in developing both
the learner’s listening and familiarising him/her with real language use
of the words being learned. The fifth activity is to simply practise
typing the words when the word form is given; the designer of the
software probably thinks it is important to copy the word in order to
memorise it. The sixth exercise is again a dictation exercise; the
learner can choose to do a dictation on either the words or the example
sentences (the same as in the fourth activity). The seventh activity is a
game, which again focuses on word form, where a missing letter has
to be supplied. The eighth activity is to recall the word meaning when
the word form is given, in the way similar to the memorisation activity
in IME where the learner is asked to recall the meaning first and then
compare it with the given meaning. If wrongly recalled, this word will
appear later. The remaining three activities enable the learner to add
new words and their pronunciation, or example sentences.
It is not difficult to see that MVE is similar to IME in two
important ways: (1) the use of large and diversified vocabulary sets
(187 in total) and (2) the use of spaced repetition. There are also two
notable differences between the two software packages: (a) MVE has
provided more vocabulary exercises, in terms of both quantity and
type, than IME, and (b) MVE has enabled the learner to edit the
content by adding, for example, words or sound files. As with IME,
MVE makes the learner pay attention to both the word form and the
word meaning and provides plenty of exercises to establish the mean-
ing-form mapping; the vocabulary gain can thus be both receptive and
productive. Motivation is certainly an issue depending on the learner’s
perception. It is claimed that this software has enjoyed a good
reputation among English learners in China. Nevertheless, the
intended learning approach in the software fits in well with the
traditional Chinese approach to English learning: visual/oral/written
219
repetition. Despite a number of good features, such as providing
different cues to facilitate word retrieval or tracking the learner
performance in a more intelligent and interactive way as described
above, there are five problems which detract from the effectiveness of
the program. We shall look at each of these in turn:

1. There is an imbalance between the focus on the word form and

the focus on the word meaning. If we review all the learning
activities available in MVE, it is not difficult to discover that
only three of them (the first, third and eighth) focus on word
meaning, the majority of the remaining activities being con-
cerned with word form. Too great a focus on word form is likely
to lead the learner into the trap of mindless repetition. This was
my experience after trying out the software; I found that I was
simply typing the words mechanically without thinking about
what I was doing.
2. Another imbalance is found in respect of the distribution of the
tests. The test followed by the initial learning and the test to
review the words are productive in nature, while the recognition
test is scheduled after the two productive tests. It would seem
that the designer should consider rearranging the order of the
three tests. As with IME, mere repetition or spaced repetition
sometimes does not help the word form-meaning mapping. The
matching would be more efficient if specific help were given
according to a specific word feature (e.g. whether it is
imageable or rhymes with other known words), or its specific
learning difficulty.
3. Solely relying on the L1 translation for the L2 word is likely to
cause confusion. For example, I came across two similar terms,
salutary and salubrious, when trying to learn a group of largely
unknown words myself in the software. Both terms are provided
with the same Chinese translation: 有益于健康的, back translated
into English as: helpful to the health. I had great difficulty in
supplying the correct word form of these two words in recall
exercises. Eventually I managed to get the right spelling for
each word; however, I still did not know what the difference
was between the two words. I checked the Oxford dictionary

220
 and discovered that the former is used in a figurative sense and
the latter is more applied to the physical environment in relation
to health.
4. As with IME, the contextual roles in learning vocabulary items
are limited in MVE; solely focusing on itemised learning will
not ensure that the word can be used correctly in a real language
environment.

9.4.3. Computerised vocabulary exercises

9.4.3.1. Tell Me More

The language learning package Tell Me More (TMM) is produced by
Auralog in France; it covers ten languages, ranging from European
languages such as English and French to Asian languages such as
Chinese and Japanese. I reviewed the standard version of the Chinese
program intended for Francophones at the intermediate level. There
are six lessons for this level and I chose the third lesson, Jiating
(family). Each lesson contains a number of learning activities:
pronunciation, dialogues, and exercises. In addition, there is a section
called library to provide grammatical and lexical help. Although
TMM is generally considered as a package for general language
learning, the majority of learning activities focus on vocabulary
learning, as we shall now see.
The program does not specify the order with which each
activity is to be carried out. We start with the pronunciation as it
precedes the other two in the navigation bar. This page contains
dozens of phrases or sentences written in Chinese characters, such as
我很好 (I’m fine60), 还可以 (I’m all right), 今天天气很好 (The weather
is very good today), etc. Clicking on a phrase, the learner can only see
the pinyin, i.e. the Westernized phonetic transcription of the Chinese
characters in the phrase. The visual sound wave pattern is also given.
The learner can listen to the pronunciation of the phrase and record
his/her own pronunciation. Immediate feedback will be given

60 I provide the English translation in brackets for the reader’s convenience.

221
regarding acceptability of the recording. However, sometimes even a
native speaker’s pronunciation will be rejected if the sound wave
pattern of the recording does not match the sample wave pattern
provided. As Zheng (2004) points out, TMM needs to revise the
design regarding speech recognition to accommodate learner pronun-
ciation at different speeds. One major problem is that there are no
direct means to access the lexical meaning either in L1 (French) or
Chinese (L2), if some of the words or phrases are unknown to the
learner. The only means is to go through the lexical glosses available
in the section library. However, the design of the lexical help is not
very user-friendly. Only the French translation is given as a bilingual
list arranged alphabetically; what is even more inconvenient is that the
list is presented as a paper dictionary, which means that finding a
lexical item may necessitate turning a number of pages. A learner of
Chinese who has used this software extensively told me that he had to
copy out the bilingual list in a notebook manually and then would
consult it when necessary! In this sense, this electronic lexical
glossary is not well designed, at least in terms of efficient use of the
computer technology. Hyperlinks or a request for lexical information
by typing the word would be a much more efficient solution. In a
word, access to the meaning of unknown words is possible but rather
inconvenient in TMM.
The learner will continue to practise pronouncing these phrases
and sentences in the dialogue, but in a more interactive and situational
manner. There are many short two-sentence conversations, each
accompanied by a picture. The first sentence is a question and the
second is the reply. There are three replies for each question and the
learner needs to choose the correct one. The initial question and the
three replies are printed in both Chinese characters and pinyin. In
response to a spoken question, the learner needs to choose and utter
the correct reply. Immediate feedback will be given regarding the
learner’s performance. In this regard, it is not clear whether the
dialogue is for learning new language content or for testing the
learner’s language knowledge.
Next come the exercises. There are a substantial number of
questions grouped according to ten types, each containing questions
ranging from three to 25 in number. They are listed below:

222
1. Match the L2 words (in Chinese characters) to the L1 transla-
tion (in French) according to a given picture.
2. Match the Chinese characters with other words in pinyin (using
synonyms or antonyms).
3. Match the Chinese characters with the French translation (using
synonym or antonyms).
4. Match the Chinese characters with their antonyms in characters
too.
5. Choose a word from several that are provided to fill in the
missing word in a sentence.
6. Fill in the missing words in sentences; similar to exercise 5 but
constructed differently.
7. Write out the pinyin for the given French phrase.
8. Put the words in correct order to reconstruct a sentence.
9. Do a dictation exercise.
10. Complete a crossword using pinyin.

All these exercises are intended for the learner to practise the words
both in isolation and in context. Particular attention is paid to
developing knowledge of the meaning of the L2 words, by relating
them to their synonyms or antonyms. However, there are no specific
learning activities preceding the exercises where the words are
introduced in isolation or in contexts where the word meaning is
explicitly given. It would seem these exercises are more suitable as
tests for reviewing vocabulary items which the learner has learned
elsewhere. If the vocabulary items are new to the learner, s/he needs to
either check the meaning in a dictionary or work out the meaning with
great effort by trying to do the exercises. Generally speaking, if most
of the items are (partially) known beforehand, learners would
doubtless find it more enjoyable or motivating to use the software.

9.4.3.2. English+
English+ is a language learning package produced by Emme
Interactive France. I reviewed the standard version (for the advanced
level) installed in the multimedia centre of the ILV at UCL. It is a
223
package for general English learning, comprising six major sections:
reading, speaking, listening, writing, language (grammar) and vocab-
ulary. Each section is constructed in the same way, containing a
number of learning activities: explore, practice, game and test. The
vocabulary section contains 250−300 vocabulary items spread over 10
lessons. Each group focuses on a specific topic, such as education,
families in the 90s, free time, studying abroad, job interview, etc. The
learner has to choose a topic to study the vocabulary items. I looked
into the lesson families in the 90s.
We start with the learning activity explore. The first page
contains a list of 30 vocabulary items. By clicking on each word, the
learner can hear the pronunciation and view the lexical information.
However, the lexical information only contains the grammatical
variants (present or past tense verb forms) or example sentences, but
no L1 translation or L2 definition for the lexical item. If these words
are unknown to the learner, s/he has to either guess the meaning from
the example sentences or look them up in a paper dictionary. There is
a short-cut icon for dictionary in each screen page which enables the
learner to search for a specific word. However, only the word
pronunciation but not the word meaning is given, rendering its use
very limited. Next the learner has to do the exercises for these
vocabulary items in practice. There are about 35 questions which are
receptive in nature: the learner needs to drag and drop the chosen
word to fill in the gap in a sentence. S/he can check whether the
answer s/he has given is right or wrong and request the display of the
correct answer, if s/he so wishes. The game activity for this
vocabulary lesson contains one crossword puzzle. The next is a test
with the same 30 vocabulary items. The test questions are similar to
those in the previous vocabulary exercise, containing receptive
questions only.
In the vocabulary learning section of English+, explicit
attention is drawn to the lexical form as the vocabulary items are
initially presented, but the lexical meaning is not given explicitly. So
if the learner does not know a word s/he needs to guess or work out
the meaning through trying out the questions in the exercise. Both
would be much less efficient than giving the lexical meaning in a
direct, explicit manner. The vocabulary gain would be largely re-

224
ceptive given the receptive nature of the exercise; this does not mean
that receptive learning cannot sometimes include a certain amount of
productive learning. No particular vocabulary learning strategy has
been introduced. It would have been better if vocabulary learning had
not been treated separately and had been integrated into the reading,
writing, listening and speaking sections. Vocabulary is best learned by
combining both itemised and contextualised learning; the more the
vocabulary item is encountered in (different) language contexts the
more fully it will be integrated into the mental lexicon.

9.4.4. Dedicated-like vocabulary programs

9.4.4.1. Language Interactive Culture

Language Interactive Culture 61 (LINC) is a language learning
software package developed by the Centre for Language and Speech
at the University of Antwerp, Belgium. It covers 18 European lan-
guages, an impressive endeavour for a university based software
enterprise. According to Van de Poel and Swanepoel (2003: 197), the
underlying assumption of LINC is that: “the learning of a second/
foreign language should be embedded in a broader, genuine, social-
cultural context”. I reviewed the demo version of the English program
for level 2 (intermediate level). The program comprises 10 units, each
based on a short video (3 or 4 minutes) from a BBC news extract. The
demo version only allowed me to review the first lesson: Jane Austin’s
fashion. The lesson contains four major learning activities: video,
listening, writing and processing. The learner can watch the video or
read the text transcription; s/he is then supposed to perform different
types of listening comprehension, vocabulary and grammar, and
topic-related exercises 62 . In addition, the section background
information contains a longer reading text (chosen from an English

61 See <http://webhost.ua.ac.be/linc>, Van de Poel/Swanepoel (2003), and

Bouziane (2006) for more information on LINC.
62 The topic-related exercises in processing include reflecting on the topics,
formulating opinions, looking for specific information, summarising ideas,
arguing for proposals.
225
newspaper or magazine) which is meant to provide the relevant
information about the video topic. There is an optional section in the
program called the virtual school which can link the learners to other
learners or the teacher through emails or CMC, but this function
requires an additional payment as well as access to the internet.
Nonetheless, this idea sounds excellent.
As with several programs reviewed earlier, LINC is for general
language learning. However, the program has given sufficient focus to
vocabulary, notably in the sections devoted to video, exercises and an
extra lexical glossary. In the video section, the learner can watch a
short video; if s/he finds it difficult to cope with the audio text s/he
can request the display of the written text. Some “key words” are
highlighted and the learner can view the gloss by a simple click. The
idea is very good but only a limited number of terms 63, primarily those
related to the topic or people’s names, are glossed. One wonders why
some of the potentially more difficult words are not glossed although
these items are available later in a separate lexical glossary. The
presentation of both the audio and the textual text is likely to make the
learner pay attention to the unknown lexical items, which is a
prerequisite for the learning of these items. The learner can watch the
video or read the text as many times as s/he wants. The questions
presented in the listening comprehension section can make the learner
focus on certain specific information in the text and hence enhance
her/his understanding of it. As a result, the meaning of certain words
becomes clearer. In addition, the availability of the glossary enables
the learner to check an unfamiliar lexical item while doing the
exercises on listening, writing or processing. The lexical items are
arranged alphabetically according to which unit they belong to; the
lexical information includes word class, L2 definitions and example
sentences, etc.
More than half of the exercises included in the writing section
are devoted to vocabulary items. These include finding matching syn-
onyms (receptive in nature); writing out the synonyms for the given
words (productive in nature); filling in blanks with the correct form of

63 According to Van de Poel and Swanepoel (2003: 198), the glossed words in
hyperlinks in the textual transcription are socio-culturally related.

226
the given word; filling in the correct prepositions, etc. Even some of
the exercises designed for grammar are related to vocabulary learning,
for example, supplying the correct verb form of a particular tense.
In LINC, vocabulary items have been processed to a satis-
factory level: the new lexical items are noticed initially in audio or
textual form in the video section; the meaning can be guessed in the
context or accessed in the lexical glossary later; the listening exercise
enables the learner to encounter the words, again in audio form, in the
context again; the writing exercises provide the learners with op-
portunities to gain a deeper understanding of the word meaning
(synonyms or antonyms) and to be able to use the words in different
contexts. Vocabulary gain is thus both receptive and productive. LINC
has other good pedagogical features: the use of current affairs videos
and the inclusion of the virtual school to form a learning community
for learners. Generally speaking, LINC is a good learning tool for
vocabulary, the only weakness being that no vocabulary learning
strategies are introduced.
Finally, despite its suitability for vocabulary learning, I would
like to make a point regarding the use of the “authentic language” (in
the sense of language being used in real-life communication) in this
software. On the one hand, authentic language provides a good
opportunity for learners to learn new language elements (vocabulary
or grammar) and to observe how the L2 is used in real language
situations. On the other hand, for it to be beneficial to the learner, the
authentic language should not be too difficult. Otherwise, little
learning would occur and the learner would be frustrated. If, as Van de
Poel and Swanepoel indicate (2003: 198), level 2 (the unit which I
reviewed) is intended for learners who have a vocabulary of at least
2,000 words and have mastered all basic grammar, it might be too
challenging for this type of learner to understand the extracts directly
taken from BBC news despite the help of the transcription. Similarly,
the so-called “authentic” tasks involved in the exercises in the section
processing are too difficult. It would seem level 2 is more suitable for
more advanced learners. A similar point has been made by Bouziane
(2006) in his review of level 1 of LINC.

227
9.4.5. Summary of the evaluation results

I have reviewed a number of programs/tasks for vocabulary learning,

including both research-based programs and commercially available
packages. The research-based programs generally represent the four
primary types of CAVL programs/tasks: electronic reading/listening
comprehension with lexical glosses, CMC lexical-based tasks,
computerized vocabulary exercises and dedicated CAVL programs;
they can be situated on a continuum going from tool/implicit/ mean-
ing-focused learning to tutor/explicit/form-focused learning. On the
other hand, commercial programs generally reflect an explicit learning
approach towards vocabulary learning, falling into three major
categories: electronic dictionaries/flash cards, computerized vocabu-
lary exercises and dedicated-like CAVL programs. The evaluation is
conducted according to the four criteria of vocabulary processing level,
level of vocabulary gain, vocabulary learning strategies and learning
motivation. The evaluation of research-based programs takes into
account both judgemental and empirical evaluation. A summary of the
evaluation of research-based programs is presented in Table 7; a
summary of the evaluation of commercial programs is presented in
Table 8.

9.4.5.1. Research-based programs

Regarding the research-based programs, reading/listening comprehen-
sion with lexical glosses provides the right context for meeting new
words and accessing necessary lexical information such as meaning
and usage. However, since the lexical learning is primarily meaning-
focused, acquisition is likely to be insignificant and restricted to the
receptive level. If the subsequent learning is based on vocabulary
exercises (computerised or not), CMC lexical-based tasks, or even
traditional learning activities, there is more opportunity for attention to
be paid to forms as well. Not all CMC tasks are suitable for vocabu-
lary learning and they need to be carefully designed to this end.
Although well-designed CMC lexical-based tasks can provide the
opportunity to draw learners’ attention to both lexical form and
meaning and perhaps even motivate them, the necessary process,

228
negotiation of meaning, is generally reported to take much longer than
other non-negotiated, elaborated tasks (Nation, 2001). Thus it is
unlikely that the majority of words can be learned in this way. It is
more meaningful to combine CMC tasks with other types of learning
activities. Similarly, computerised vocabulary exercises can be used
on their own but would be better if they were combined with other
learning activities, as demonstrated by Allum (2004).
Generally, judgemental evaluation is consistent with empirical
evaluation. The way the vocabulary learning activities are designed
can predict both the learning process and the outcome. Evidence also
exists that the quality of the learning process can influence the learn-
ing outcome (e.g. Goodfellow, 1994, 1999). However, when a CALL
program is sophisticatedly designed and supposed to lead to an
effective learning approach, learners, faced with many options and the
freedom to decide what to do, tend to take short-cuts which will lead
to shallow learning approaches. Thus a mismatch between the judge-
mental and empirical evaluation may occur in sophisticatedly
designed tutorial CALL programs. The point that needs to be raised is
how to make learners follow the designed learning path of the
program to achieve the desired learning objective. Should learners be
provided with some kind of tutorial training to use the CALL program
more effectively (e.g. Ebbrell/Goodfellow, 1996)? Or should learner
actions or freedom be controlled to some degree? The latter possibility
touches on learner control of the program, which is a controversial
issue that has provoked considerable discussion among researchers
(e.g. Shin/Westell, 2001; Cobb/Stevens, 1996; Chapelle, 2001;
Hubbard, 2004).
One limitation of the evaluation is that the results obtained may
not always be unbiased or accurate. The main reason is that every
study I looked into differs from others in a number of ways, some-
times to a considerable extent, each having their specific context and
focus. We can take the learning rates as an example. Generally, the
figures were obtained by dividing the number of words learned by the
total number of unknown words to be learned. However, there are two
notable differences. First, the total number of words to be learned
varies in each study. Second, the time involved in learning the words
also differs widely in each study, from 10 minutes to several hours at a
229
stretch. In other words, the “learning rates” may not always be
comparable given that each study design is different.

9.4.5.2. Commercial programs

As for commercial programs for vocabulary learning, the most
popular way is to conceptualise the program either as an electronic
dictionary or flash cards to learn the lexical items or as vocabulary
exercises to practise using the lexical items (receptively and pro-
ductively). The two programs together, Intelligent Miracle English,
Memorising vocabulary effortlessly, produced in China for learning
English, generally reflect the traditional Chinese approach to English
learning (see Part IV). Although the two programs and other similar
ones are perhaps widely used by Chinese learners of English, and
despite the fact that they achieve a certain degree of success in
vocabulary learning, the main problems innate in this type of learning
approach can be summarised as follows:
 Repetition alone cannot guarantee that the meaning-form
matching can be successfully established, particularly for
difficult lexical items.
 Lack of contextual learning cannot ensure that the word will be
accurately used in real contexts even though the basic meaning
has been mastered.
 Solely relying on L1 for the lexical meaning is problematic,
leading to lexical errors or confusion (meaning and usage).
 Too great a focus on form is likely to lead to a neglect of lexical
meaning which will give rise to errors in this respect.

The fourth point deserves further clarification. On the one hand, the
focus on word form reflects the wide difference in the writing system
between Chinese and English, which means that Chinese learners need
to pay specific attention to the English word form to avoid spelling
mistakes. (See 11.3.2. for some examples of Chinese learners’
frequently made spelling mistakes). On the other hand, too great a
focus on the word form as a result of constant written repetition is
likely to distract the learner’s attention from the elaboration of the
lexical meaning. This may explain why very often Chinese learners

230
consider a word is learned if they can provide an equivalent Chinese
translation. However, this is likely to lead to meaning confusion in
lexical use (see 11.3.1. for a discussion and examples of lexical errors
caused by meaning confusion by Chinese learners).
In comparison, other commercial programs (made in European
countries or the U.S.A) tend to show a more balanced focus on the
word form and meaning, or may even be slightly weighted in favour
of the latter.
Taking both the commercial programs made in China and those
made in Western countries, it can be seen that there is a certain amount
of research being applied in different areas related to SLA behind the
program design. For example, spaced repetition, based on memory
research, is applied to SuperMemo, Intelligent Miracle English, and
Memorising Vocabulary Effortlessly; combining images with textual/
audio input, the essential multimedia feature, is widely employed in
Rosetta Stone; elaborate lexical meaning by linking the word to other
items (synonyms or antonyms), an application of the findings in
respect of the mental lexicon, is partially reflected in Tell Me More.
On the other hand, the commercial programs often focus on one or
two design features or on the application of a single learning theory
while ignoring others. In this sense, their pedagogical effectiveness is
limited. The programs would be more effective if they incorporated
other relevant features or were complemented by other learning
activities. For example, the programs that fall into the category of
electronic dictionaries/flash cards, primarily focusing on itemised
learning, can be combined with more contextual learning such as by
adding a lexical concordancer. More specifically, some of the
programs, such as Intelligent Miracle English and Tell Me More are
better used to review and consolidate words which have been learned
elsewhere (preferably in a more contextualised learning situation).
Finally, commercial programs which are largely based on
research, such as Language Interactive Culture, display a more
comprehensive, dedicated approach to vocabulary learning, where the
vocabulary learning is both itemised and contextualised; vocabulary
items are processed to a sufficient level to facilitate further lexical
elaboration and development.

231
 The main limitation of the evaluation of commercial programs
lies in the fact that only the judgemental evaluation is provided due to
practical constraints and hence the results may not be accurate without
the empirical data. Secondly, only a limited number of programs have
been selected and reviewed while there are many other commercially
available programs. Thus the results may not be representative and no
generalisation can be made. Lastly, I have only looked at the vocabu-
lary learning sections/aspects in some programs considered for general
language learning; therefore the evaluation results only apply to the
effectiveness of vocabulary learning and not to the effectiveness of
other aspects of learning, such as reading, writing, or grammar.

9.5. Conclusion

The evaluation for research-based programs shows that both direct

and indirect vocabulary learning benefits generally correspond to how
much the program is geared towards vocabulary learning in terms of
tutor/tool, implicit/explicit, and meaning-/form-focused orientations. It
emerges that CAVL programs which are tutor oriented provide explicit
and form-focused learning activities that are of particular potential
help in vocabulary learning. However, other types of CAVL programs,
such as electronic reading/listening comprehension, CMC lexical
based tasks, or computerised vocabulary exercises can be combined or
integrated into the broad language curriculum to promote vocabulary
learning. The evaluation for commercial programs shows that they
primarily adopt an explicit approach towards vocabulary learning;
theories or good design features are often applied in isolation without
being combined with other relevant ones. Commercial programs
should take into account research findings systematically to produce
more efficient tools for learners to learn vocabulary.
The evaluation results show that neither research-based pro-
grams nor commercial programs take into account vocabulary learning
strategies seriously; how to provide systematic vocabulary learning

232
strategy instruction to learners is a key issue that emerges from the
evaluation. To empower learners with useful vocabulary learning
strategies is undoubtedly the only means by which we can ensure that
they become efficient and autonomous in their learning approaches.
One solution is to integrate training of this kind into CAVL programs,
particularly the dedicated ones.

233
 Electronic reading/listening CMC lexical-based Computerised Dedicated CAVL programs
comprehensions with lexical glosses tasks vocabulary
exercises
Reading Listening De la Fuente, 2003 Lexica CAVOCA
Laufer/Hill, Smidt/ Allum, 2004 Goodfellow, 1994, Groot
2000 Hegelheimer, 2004 1999 2000
Vocabulary Ja Insufficient Insufficient Good Good Good Good
processing
level Eb Insufficient Insufficient Good Good Less Good Less Good
Level of J Primarily Limited in both Both receptive and Both receptive and Both receptive and Both receptive
vocabulary receptive reception and productive productive productive and productive
gain production
E Good Low productive Good receptive and Good receptive and Varied productive Good receptive
receptive gain gain productive gain productive gain gain and productive
gain
Vocabulary J Unknown Unknown Insufficient Unknown Insufficient Insufficient
learning E Unknown Unknown Insufficient Unknown Insufficient Insufficient
strategies
Learning J Unknown Unknown Good Good Unknown Unknown
motivation E Unknown Unknown Good Good Unknown Unknown
aJudgemental evaluation
bEmpirical evaluation

Table 7. Summary of the evaluation of research-based programs.

234
 Electronic dictionaries/flash cards Computerised exercises Dedicated-like
CAVL
programs
SuperMemo Rosetta Stone Intelligent Memorising Tell Me More English+ Language
98 version, Online demo Miracle Vocabulary Standard version, Standard version, Interactive
Poland version, English Effortlessly France France Culture
U.S.A Demo version, Standard Demo version,
China version, China Belgium
Vocabulary Insufficient Good Less good Less good Insufficient Insufficient Good
Processing
Level
Level of Both Primarily Both receptive Both receptive Primarily receptive Primarily receptive Both receptive
vocabulary receptive and receptive and productive and productive and productive
gain productive
Vocabulary Insufficient Insufficient Insufficient Insufficient Unknown Unknown Unknown
learning
strategies
Learning Less good Good Less good Less good Good Less good Good
motivation

Table 8. Summary of the judgemental evaluation of commercial programs.

235
236
 Part IV:
L2 (English) Vocabulary Acquisition in China

237
238
Chapter 10. English Teaching and Learning in China

10.1. Historically identified distinct periods

of ELT in China

Due to the upheaval experienced by China in the last half century, the
progress of English teaching/learning has been beset with difficulties,
and at times there has even been complete discontinuity. According to
Lam (2002: 246−247), six distinct periods can be identified, each of
which is associated with a specific government language policy. I
shall give a very brief summary of each.

The interlude with Russian: early 1950s

After gaining independence in 1949, the new Chinese government
sought an alliance with its then powerful neighbour, the Soviet Union,
and counted on its help to rebuild their country. Russian became the
predominant foreign language taught in schools although English was
allowed to continue. Soon the Sino-Russian relationship deteriorated
as the Russians did not live up to the other side’s expectations, failing
to treat the Chinese, in their eyes, equally and breaking promises.
English, and other foreign languages, were reconsidered for incorpora-
tion into the national syllabus.

The back-to-English movement: 1957−1965

After the illusion of depending on the Soviet Union was dispelled,
China started to seek economic bonds with Western countries. Syllabi
of English programs were drafted for different educational levels:
from secondary school to universities. Learners included English
majors and science and technology majors.

239
Repudiation of foreign language learning: 1966−1970
The year 1966 marked the start of the Cultural Revolution which cut
the nation off from the outside world and had devastating effects
economically, politically and socially. Scholars were sent to the
countryside to be re-educated to be rid of “harmful ideas”64. Foreign
language teachers were expelled too; most of them suffered terribly
because they were suspected of being spies or worshipping foreign
ideas (Adamson/Morris, 1997: 15). Students were also dismissed and
set to work as labourers to learn “industrial production, agricultural
production, and military affairs”, as advocated by Mao in 1966 (cited
in Adamson/Morris, 1997: 15). In short, foreign language learning,
including English, disappeared from the education system.

English for renewing ties with the West: 1971−1976

With the formal recognition of China by the United Nations (1971)
and the official visit to China by the United States president, Richard
Nixon (1972), China was preparing to establish commercial, cultural
and educational communication with the United States and other
Western countries. ELT restarted with the goal to “serve the people as
an expert” (Lam, 2002: 246).

English for modernization: 1977−1990

Mao’s death (1976) marked the end of a decade of turmoil as well as
the beginning of bringing the whole nation back to a normal state.
ELT featured at different levels of the education system, including
primary and secondary schools and universities. In 1978, Deng
Xiaoping issued the Policy of Four Modernizations and called upon
the whole nation to fight for the modernisation of four sectors:
agriculture, industry, science and technology/defence. The Reform and
Open door policy that ensued boosted the learning of English and
other foreign languages. In 1982, English became the predominant

64 These “harmful ideas’ are covered by the term “Feng Zi Xiu” raised at the
beginning of the Cultural Revolution. “Feng” refers to the traditional culture,
thoughts, or customs, associated with the old societies; “zi” refers to
capitalism; “xiu” refers to the policy adopted in the Soviet Union, which was
considered as seriously deviating from “pure socialism”.

240
foreign language in secondary schools. The situation continued and
recent statistics show that about 95% of students were choosing
English as their foreign language (British Council, 1995, cited in
Cortazzi/Jin, 1996a: 63). It was during this period that China
recovered from the serious damage caused by the Cultural Revolution
and made tremendous progress in developing its economy and other
areas, including creating and presenting a new image to the world.

English for international status: from 1991

With the continued progress in all-important areas, particularly the
economy, of the nation, China aimed at gaining a corresponding
international position. The seal of success has been China’s joining of
the World Trade Organisation (WTO) in 2001, the hosting of the
Olympic Games in 2008 (Beijing) and the holding of the World Fair
Exhibition in 2010 (Shanghai). The government has made tremendous
efforts to improve English education all over the country. This
includes redesigning the syllabi at all educational levels (primary/
secondary schools, universities and adult education), recruiting more
native speaker teachers, doing research in language teaching/learning,
etc. English is not only one of the important compulsory subjects for
most students; it is also required for professional promotion: one has
to pass the national English exam before other qualifications are
considered when competing for more advanced posts. In consequence,
there has been a boom in English learning throughout the country and
it is continuing at the present time. This is evidenced by the
ever-crowded language classrooms (evening classes, adult learning
and private organisations), fast selling language learning materials
(books, CDs, software, etc.) and the popularity of internationally
recognised English exams (TOEFL, IELTS, GRE, GMAT, etc.).

241
10.2. Features of ELT in China

10.2.1. High motivation

Motivation is a crucial factor that will tell how much effort a learner
will invest in learning the language and eventually the degree of
language proficiency that will be attained, as we saw in Chapter 7. For
Chinese students, the economic factor primarily stimulates the learn-
ing of foreign languages. Since the ending of the Cultural Revolution
in the late 1970s, the government has put the chief focus on develop-
ing the national economy. Developing the economy means
communicating with the outside world or, to be more specific,
communicating with developed Western countries who have attained a
much higher level in overall development; to communicate means
knowing foreign languages, particularly English, by far the most
important language in the technological and commercial domains.
Another important factor that motivates Chinese students to
learn English is the combined impact of traditional ideas of
Confucianism and the current educational system in China. The
current educational system highlights the importance of three subjects
through the whole of secondary education: Chinese, Mathematics, and
English. To study in the universities or colleges is the dream of most
young people and their parents; this is the modern duplication of
Confucian schoolboys in times past struggling to be scholar-officials
who enjoyed the highest prestige in the old imperial society. To gain
access to university, one has to take and pass the national entrance
examination, which includes English as one of the most important
subjects. Most students will take their preparation very seriously and
spare no effort to improve their English to prepare for the exam. After
entering university, English remains an important subject and the
course will cover at least the first two or three years. Students must
also take the national Band 4 examination; a pass in Band 4 is a
necessity for every student to get their degree at the end of their four
years of study. More ambitious students will set higher aims to get
good scores in international English tests, such as TOEFL or IELTS,

242
in order to pursue overseas study; as the economy is improving, more
and more parents can afford to send their children abroad.
In general, Chinese students’ motivation in English learning is
instrumental: they choose to learn English because it will help them to
go to university and then obtain a good job in the fast growing
economic markets where English is greatly needed. As young Chinese
students have more and more opportunity for contact with foreigners
and to be exposed to Western culture, more and more of them develop
an integrative motivation in learning English.

10.2.2. Chinese culture of learning

10.2.2.1. What is culture of learning?

Culture of learning (CL) is associated with traditional beliefs and
attitudes to teaching and learning, often under the influence of the
specific culture of a group of people. Cortazzi and Jin (1996b: 169)
describe CL this way:

[…] behaviour in language classrooms is set within taken-for-granted

frameworks of expectations, attitudes, values and beliefs about what
constitutes good learning, about how to teach or learn, whether and how to ask
questions, what textbooks are for, and how language teaching relates to
broader issues of the nature and purpose of education.

CL sometimes is deeply rooted in educational and cultural traditions

and we are so familiar with it that we are often unaware of its
existence; it is the “hidden curriculum” (Cortazzi/Jin, 1996b: 169). CL
is also subject to the influence of the socio-economic environment
(Cortazzi/Jin, 1996b). It is important to understand CL as it explains
the learning process and learning outcome of language learning.

10.2.2.2. Chinese culture of learning: perception and practice

CL is heavily rooted in culture and tradition; this is particularly true
for Chinese CL which is deeply influenced by Confucianism and
Taoism. The two most important philosophies have fulfilled different

243
roles for more than 2000 years. Confucianism was adopted by the old
imperial authorities as the philosophical foundation for governing the
nation. On the other hand, Taoism had always been practised by
common Chinese people to ensure their life quality. The whole society
was imbued with the two philosophies; they were passed down from
generation to generation till the present time, forming the so-called
Chinese “orthodoxy”. This is a unique phenomenon in human history,
as Kelen (2002: 228) points out: “No other social doctrine and no
other means of transmission have prevailed for so long in human
history elsewhere.” Consequently, its influence is profound and
immune to change. As a member of Chinese society, one does not
need to read the classics written by Confucius or Lao Tzu to
understand their philosophy. Their ideas have permeated into
everybody’s life, influencing their daily behaviour, their thinking and
attitudes. With the communication with the Western world over the
past century, many new philosophies and ideologies have been
imported into China and challenged the old ones. However, it is wrong
to think that the new ideas have “simply dissolved the older patterns
of popular consciousness” (Kelen, 2002: 228). The orthodoxy persists,
continuing to influence people’s lives, including the conception of
language, assumptions about language learning, and the way language
learning should be conducted in the classroom, etc.
Hansen (1991, cited in Kelen, 2002: 224) sees the Chinese
conception of language as completely different from the Western view,
claiming that the former is “prescriptive and pragmatic” while the
latter is “descriptive and semantically motivated.” Kelen quotes
Hansen in pointing out a fundamental assumption about the Chinese
conception of language, namely, that language is the “means by which
right action is achieved” or a “system of appropriate discursive acts in
which there is always a positive and negative term […]” (2002: 225).
According to Hansen (1991, cited in Kelen, 2002: 225−226), one
important impact of these traditional ideas on language teaching is that
important figures, such as great scholars or leaders, will be used to
illustrate, teach, develop and regularise language use. Another
important consequence is that language can be learned and taught in a
prescriptive manner. It further leads to a distinct role for the language

244
teacher: “the teacher prescribes the correct words, their correct uses;
the teacher corrects the students’ errors” (Kelen, 2002: 227).
Kelen (2002) also claims that the “model of the indefatigable
learner-teacher” 65 established by Confucius lends itself to passive
learning. His main arguments are based on his observation of the
language learning classroom in which the teacher controls the learning
process and decides on the learning activity whereas the students just
follow the teacher. On the other hand, this model has its good side,
helping to inculcate positive qualities, which Kelen does not mention,
such as diligence, serious commitment, high motivation, goal-
orientedness, respect for the learned, etc., and which are generally
appreciated by both Chinese and foreign teachers.
If Kelen has investigated the Chinese CL primarily from a
philosophical point of view, Hu (2002) has taken a more
education-oriented approach in conceptualising the Chinese CL. In
particular, Hu (2002: 96−98) observes a series of features innate in the
Chinese CL: (1) there is a high respect for education; (2) education
involves both gaining knowledge and developing moral qualities; (3)
education is perceived as a process of increasing knowledge (rather
than constructing it); (4) insistence is placed on keeping a hierarchical
yet harmonious relationship between the teacher and students; (5)
education can be achieved by individual effort (including determin-
ation and will power) regardless of intelligence and ability or even
family background. These traditional Chinese educational features are
heavily rooted in Confucianism. In addition to the “passivity” of such
a CL, as indicated by Kelen (see above), there are also positive aspects
characterised by such a CL, particularly in promoting the idea that
everyone can be educated if they so wish. It appears to be an
indiscriminate education policy, appealing to everyone who wants to
be educated and gain knowledge, which is in sharp contrast with the
traditional Western view where only a small elite are entitled to
education. Fundamentally, this Confucian view of education was
consistent with the close bond between knowledge and power,

65 Confucius says: “I have listened in silence and noted what was said, I have
never grown tired of learning nor wearied of teaching others what I have
learnt” (1996, cited in Kelen, 2002: 233).

245
reflected in the long-standing national examination system − Ke Ju66
− held by the old imperial authority to select officials. If one had
knowledge, one could be selected as an official and gain power. This
would motivate people to study hard to gain knowledge, regardless of
other conditions, such as intelligence, ability and social background.
This Confucianist view was also favoured by the authorities as they
could choose the best officials from among a huge group of educated
candidates.
At a practical level, such a perception of CL has helped to
define the criteria of being a good teacher or student (Cortazzi/Jin,
1996b; Hu, 2002). The empirical study of Cortazzi and Jin (1996b:
187) showed that the majority of their subjects (university students)
expect a good teacher to have deep knowledge; other important
criteria include being patient, being humorous, being a good moral
example, etc. As for what makes a good student, the subjects put hard
work as the number one factor, followed by being sociable, paying
attention to the teacher, respecting and obeying the teacher, etc.
As a consequence of such perceptions and practice, language
teaching/learning is best summarised as follows (Cortazzi/Jin, 1996a:
65):

Chinese approaches to language teaching have a long-standing concern with

mastery of knowledge, which is focused on the four centres of the teacher, the
textbook, grammar and vocabulary. Knowledge of English is transmitted
through the teacher, as an authority, a source of knowledge and an intellectual
and moral example. This knowledge is also in the textbook, which is a key
element in Chinese learning; texts are taught and learned in exhaustive detail.
They are often memorised. Grammar and vocabulary are further elements of
knowledge which are explained and transmitted. Students engage heavily in
memorizing hundreds, even thousands, of words each year.

66 The Ke Ju examination system started in 605 and continued till 1905; the
official policy was “free registration, public exam, fair competition, selecting
the best”. There were many hierarchies in the examination system and the
highest level exam was held in the emperor’s palace and supervised by the
emperor.

246
10.3. Approaches to ELT

Although there may be other approaches or methods that have been

adopted by English teachers (Chinese or foreign) or Chinese learners,
the commonly documented important ones include grammar-transla-
tion, the audio-lingual method, the “communicative approach” and,
more recently, task-based learning. The reason why the term com-
municative approach is put in inverted commas is that it is not the real
communicative approach but a Chinese version of it, a kind of
“eclectic approach” (e.g. Adamson/Morris, 1997; Cortazzi/Jin, 1996a;
Hu 2002; Liao, 2000) or “weak” version of the communicative
approach (Liao, 2000).
As Adamson and Morris (1997) observe, the macro language
policy adopted in China has always guided both English curriculum
development and approaches to ELT. The government has a great
influence in shaping ELT: they decide what approaches should be
taken in accordance with a specific language policy adopted at a given
time, in consultation with appointed agencies, consisting of language
experts or language teachers at different levels (Adamson/Morris,
1997). As we saw earlier, there are six distinct periods associated with
a given language policy, when different approaches were adopted in
respect of ELT. In the first period (“the interlude with Russian: early
1950s”), ELT was heavily influenced by the language pedagogy
adopted in the Soviet Union: the grammar-translation method. This
approach is described by Adamson and Morris (1997: 9) as “an
emphasis on reading and writing skills, constant references to the
learners” mother tongue, a focus on grammatical forms, and
memorisation of grammatical paradigms”. In the second period (“the
back-to-English movement: 1957−1965”), when the nation was
breaking with Russian, the language teachers called for a change in
ELT pedagogy as a result of disillusionment with the “Russian-style
textbooks and the prevailing spoon feeding method of teaching”
(Adamson/Morris, 1997: 11). The audio-lingual method emerged with
the emphasis being laid on “oral skills before written ones”; it
“encourages students to use English as much as possible, incorporates

247
sentence-pattern drills, and promotes learning through habit forma-
tion” (Adamson/Morris, 1997: 13). Following the outbreak of the
Cultural Revolution, ELT was completely suppressed in the third
period (“repudiation of foreign language learning: 1966−1970”).
English reappeared in the schools to a very limited extent in the fourth
period (“English for renewing ties with the West: 1971−1976”). The
“teacher-centred grammar-translation methodology” (Adamson/
Morris, 1997: 16) was adopted. The fifth period (“English for
modernization: 1977−1990”) was characterized by a mixture of the
grammar-translation method and the audio-lingual method, and this
was reflected clearly in the textbooks produced at that time.
Since the beginning of the sixth period (“English for inter-
national status: from 1991”), the nation felt a real need to communi-
cate with foreign countries and saw improving English education
throughout the whole nation, particularly in secondary schools, as
crucial to the task. It was clearly stated in the guidelines in the English
Curriculum (for secondary schools) in 1993 that the primary goal was
to develop communication (Adamson/Morris, 1997: 22). Yet to shift
from the traditional grammar-translation or audio-lingual method,
which most teachers felt comfortable with, to a real communicative
approach was not an easy task at all. Nonetheless, the research into
and advocacy of the communicative approach had started in the 1980s.
This was primarily the work of Li (1984) who published her powerful
and convincing article “In defence of the communicative approach” in
the ELT Journal and subsequently edited a series of textbooks
Communicative English for Chinese Learners (1987). As she writes in
the foreword of these textbooks (1987: vi): “because the method was
new in every way, it met with considerable resistance from the start”.
Most teachers (and also students) still have great difficulty in
accepting the new approach and dealing with the new type of teaching
materials (Liao, 2000).
Hu (2002) argues that the Chinese culture of learning is actually
in conflict with some of the most important pedagogical assumptions
underpinning the communicative approach. Hu (2002: 102) claims
that fundamentally the communicative approach is “interactive”
whereas the traditional Chinese culture of learning is “epistemic”; in
particular the difference between the two lies in the fact that:

248
 They have largely contrary assumptions about the respective roles and
responsibilities of teachers and students (e.g. learner-centredness vs. teacher
dominance/control). They encourage different learning strategies (e.g. verbal
activeness vs. mental activeness). They reward different qualities in students
(e.g. independence and individuality in CLT and receptiveness and conformity
in the Chinese culture of learning).

In addition, Liao (2000) notes other practical difficulties encountered

by language teachers that hinder the adoption of the communicative
approach: (1) limited language proficiency, (2) lack of cultural
knowledge, (3) unfamiliarity with the communicative approach, (4)
pressure to cope with the old examination system. Another problem is
that English teachers always have to deal with large language
classrooms in which there may be more than 100 students. As a result,
it seems that a real communicative approach is very difficult to apply
to Chinese ELT. In reality, what has been claimed to be the
communicative approach is essentially an eclectic approach combin-
ing traditional and communicative practices (Adamson/Morris, 1997;
Liao, 2000).
What is currently hotly debated in applied linguistics and other
related fields, is the lexis-based language learning/teaching; language
consists in not only individual items but also multi-word items
(formulaic sequences) and mastery of these items has become the core
of L2 acquisition (see also Chapter 6). Widening and promoting the
research into lexis and, more broadly, phraseology, has been widely
explored in recent international conferences such as “Between Europe
and the Mediterranean: a shared idiomaticity” in Tunis, September,
2004, “Phraseology in Linguistics and Other Branches of Science” in
Slovenia, September, 2005, “Phraseology 2005: The many faces of
phraseology” in Belgium, October, 2005, and the “International Con-
ference on Phraseology and Paremiology” in Spain, September, 2006.
It seems that these new ideas and thinking have not affected Chinese
ELT very much, although they have begun to attract the attention of a
few Chinese researchers (e.g. Xu, 2004; Wang, 2004). It is certain that
the traditional grammar-lexis dichotomy still prevails in Chinese ELT;
it fits in well with the traditional Chinese culture of learning where
language learning is seen essentially as “mastery of knowledge”

249
(Cortazzi/Jin, 1996a: 65); this consists in analyzing the language
structure (grammar) and memorizing the many thousands of words
that can be put into the structure. On the other hand, with such a
tradition, vocabulary is always a focus of language learning to the
point that it has never been neglected in Chinese ELT, although this
learning is largely based on simple memorisation. This is not to say
that the newly developed lexis-based view of language learning is of
little concern to Chinese ELT since an emphasis has always been
placed on vocabulary learning. The point is that the role of vocabulary
learning in ELT and the learning approaches have to be reconsidered
in the light of the new findings of language acquisition for one
purpose only, namely, to make Chinese ELT more efficient and
effective.

10.4. Other constraints of ELT in China

The nation’s economic development and the national policy to

promote ELT conditions for learning English are constantly improving,
especially since the fifth period, as is evidenced by learners’ gradually
more favourable English learning experience over the different
periods (Lam, 2002). It is clear that Chinese ELT still needs to search
for more appropriate approaches to ELT to better the situation. There
are, however, a number of practical constraints that hinder the pro-
gress of ELT, notably, a lack of qualified teachers, over-crowded
language classrooms, the absence of good teaching materials and an
exam-centred educational system.
As English courses feature more and more prominently in the
national educational system, more and more qualified English teachers
are needed. The English curriculum normally starts in the third grade
of most primary schools and now the trend is to start from the first
grade. In secondary school the pupils need to continue to study
English for another six years. When they go on to university, the non-
language specialists have to take a general English course for the first

250
two years and then some of their major courses are taught in English
during the remaining two years. The language policy adopted by the
national education ministry requires 5% of all university non-language
courses to be given in English. English will continue to be an im-
portant part of the postgraduate programmes (Masters and Doctorates).
The trend is for English not to be simply treated as a foreign or second
language but also as a medium to acquire knowledge (Huang/Xu,
1999). If one considers the huge school population in China, such a
massive English curriculum not only requires an enormous pool of
language teachers but also a high level of proficiency from them for
the goal to be attained. The reality is rather disappointing. Recent
statistics show that the ratio of English teachers to students (non-
language specialists in university) is 1:13067. Most of the teachers
have just finished their four-year first degree in English and are
immediately sent to teaching posts, which means they have no
experience. Due to their heavy schedules and the effort they must put
into preparation, they simply do not have the opportunity or time to
improve professionally.
As a consequence of the lack of teachers, language classrooms
are invariably over-crowded. A normal class at primary or secondary
school will consist of between 50 and 80 pupils. For non-language
specialists in universities the situation is worse; teachers can have up
to 300 students in one class. With such big classes, there is little
possibility of using an interactive language learning/teaching approach
and there is little alternative but to return to the traditional
grammar-translation method. In such a context and given the cultural
background described earlier, it is inevitable that students remain
passive, that is, if they decide to attend classes at all.
As for language teaching materials, there are several points to
be made. First, there is a shortage of teaching materials and good ones
are a rarity. Up till the 1990s, the People’s Education Press was
responsible for producing language teaching materials in accordance
with the national English syllabus for secondary schools for the whole
country (Adamson/Morris, 1997). The same textbook could be used in

67 Currently, the number of university students is growing at a rate of 8% per

year, which far exceeds the growing rate of recruitment of English teachers.

251
all secondary schools across the country. The situation in universities
was rather similar, e.g. a series of textbooks College English,
published by Shanghai Foreign Language Education in 1989, was in
widespread use all over China for more than 10 years till the
beginning of the 21st century. Other widely used textbooks include
Alexander’s New Concept Series (1967) and Mainline Progress
(1980). Now the situation seems to have improved, but many teachers
still complain that most of the materials available for preparing
students for exams are not authentic or appropriate for real language
learning. Sometimes the textbooks chosen are not good at all. An
article by Teng et al. (2004) shows that the textbooks used in their
university were either quite unsuited to student levels or full of wrong,
inappropriate or outdated expressions. Furthermore, language teachers
are seldom involved in deciding what textbooks or teaching materials
should be used. This is often decided by a superior authority. For
example, in Yichang, central China, the textbooks in secondary
schools are chosen by the education authority of the city. At the
university, the situation is a little better: senior teachers put forward
suggestions and these are acted upon by the authorities if the materials
are well known and meet with general approval. The full freedom to
choose teaching materials has never been passed on to language
teachers; this is clearly a reflection of the very hierarchical and
authoritative system that persists in China.
It seems that in the Chinese educational system national exams
are the only means of measuring students’ achievement. Secondary
school pupils have to pass the national Matriculation English Test to
be able to go to university. University students have to pass Band 4 or
6, of which Band 4 is a prerequisite for getting a degree. On the one
hand, teachers and students are appealing for a greater focus on
developing language communicative competence; on the other hand,
the examination-centred education system, where the emphasis is
firmly on the written language, is still thriving. This has created a
dilemma in ELT. Secondary schools have to keep or improve the
admission rate to universities; university teachers have to maintain the
pass rate of Band 4/6; otherwise they risk a drop in salary. It is no
surprise that most of them concentrate on preparing students for
exams instead of focusing on oral activities that will improve their

252
communicative ability in the language. Although the government has
taken initiatives to reform the language exam system, to make it less
grammar-oriented and more communication-oriented, the central
position of exams in the educational system, regardless of the form
they will take, will not stop universities/schools/teachers focusing on
them rather than on the real ability to use the language.

10.5. Recent reform in ELT in China: focus on higher

education and the non-language specialist

In this section, I shall focus on the recent or coming changes in

language policies in universities for non-language specialists. The
changes cover almost every area: syllabi, the evaluation system,
teaching modes, Bands 4/6, etc.
The new national university English syllabus (National
Education Ministry, 2004) states that university English teaching
should focus on developing English knowledge and use, learning
strategies, and cross-cultural communicative competence; it should be
guided by language learning theories and be operated by a system that
combines multiple teaching modes and methods. The goal is to
develop students’ ability to use English, particularly listening and
speaking, so that they can perform effectively in English oral or
written communication; moreover, they must be taught to study
autonomously and to increase their cultural knowledge and awareness
in order to meet the national need to develop the economy and
international communication.
Accordingly, an important change has been introduced into the
evaluation system. Previously, evaluation solely depended on the final
English exams at the end of each semester, termed the “end
evaluation”. Now part of the evaluation is the “process evaluation”,
including self-evaluation by individual students, mutual evaluation
among students, evaluation by the teachers, and evaluation carried out
by the educational administration sector. The process evaluation is

253
conducted on the basis of the record of classroom and extra-curricula
activities, the record of self-study on the web, interviews or
discussions, and so on. This policy seems to be more suitable for
assessing students’ communicative competence than exams only.
There is an even more radical change in respect of teaching
mode. According to the current situation of Chinese ELT, the new
policy is that teaching and learning should not be solely based on the
traditional classroom but on the fast-developing information
technology, particularly network technology. The aim is to make ELT
more individualised, more self-initiated, and unrestricted by location
and time so that both teachers and students can be actively involved in
the teaching/learning process. With the new teaching mode, grammar,
reading comprehension, writing, and translation are dealt with in
traditional language classrooms, suitable for large numbers of students
and traditional teaching approaches (e.g. teacher-centred approaches).
Language learning software (stand-alone applications or network
applications) will largely be concerned with listening and speaking.
Students can do self-learning at their own rate and according to their
own schedule; teachers will give some help to small groups of
students (less than 8). The new teaching mode combines classroom
teaching and language software study as shown in Figure 15.
This new teaching mode is an attempt to solve the problems the
university ELT is faced with. In particular, it tries to cope with the
highly disproportionate student/teacher ratio. However, it still reflects
a fundamental concept of the Chinese culture of learning, namely, that
language is made up of knowledge and that that knowledge can be
acquired by mastering different aspects of it. Listening and speaking
are the basis of communicative competence and are thus considered of
supreme importance, but this is no longer the teachers’ but the
language learning software’s responsibility. There are two rationales
for this. First, the computer is the ideal surrogate teacher to play or
record oral English repeatedly without getting “bored” in the way the
human teacher does. This notion is similar to what is called
“Structural CALL” in the initial period of CALL development
(popular from the 1960s to the 1970s). Second, listening and speaking
can be done individually on computers (as it is difficult to do so in the
classroom with large numbers of students). To rely on individual

254
learning on computer to develop learners’ communicative competence
places a high demand on the quality of the software. Technically, it
should be highly intelligent, customised, and humanised. Theoretically,
it should incorporate good language learning theories and thoroughly
integrate into the learning activities. If these two basic requirements
cannot be met, what students can do in respect of listening and
speaking on the computer will very likely follow the audio-lingual
fashion: mere repetition and imitation. Another problem is whether the
“artificial” interaction with the software will really lead to
communicative competence (which is humanistic in nature). This is
certainly an area rarely researched in the CALL field and much needs
to be done before we can have a clear idea how effective this artificial
interaction would be. It would seem to be a little premature to pass the
responsibility for developing learners’ communicative competence
from teachers to language learning software at this stage.
Alternatively, CMC, which is communicative in nature, could
be a good means to develop learners’ communicative competence but
good language learning tasks involving pair or group learners need to
be developed. Secondly, in the new teaching mode teachers are
primarily responsible for giving instructions in reading, writing and
translations to students. This again reflects a feature specific to L2
language learning in China where these three areas were the primary
goals for a long time. In fact, these areas, particularly the first two,
reading and writing, have been much more widely researched in the
CALL field (at least in Europe) than other areas such as listening or
speaking. In addition, fruitful findings have been achieved for the first
two mentioned areas. It would seem that the new teaching mode is
somewhat “odd” as it disregards these general findings in CALL. On
the other hand, we should not be too critical as there may be a good
reason for doing so. Chinese English teachers may have been well
trained in these areas; CALL research and support in these areas may
not be sufficient in Chinese L2 education. In other words, the choice
made is an efficient use of human and material resources currently
available in Chinese L2 education. Systematically integrating
computers or, broadly speaking, old or new technologies, into the
language education curriculum is the objective faced by many nations.

255
The situation in every nation is different and each may have to
develop their own best approach in the course of pursuing the goal.
Changes in Bands 4 and 6 are another important component of
this reform. Bands 4 and 6 have been developed into a standardized
test, the latest version being made up of four components: listening,
reading comprehension, general knowledge, and writing. Listening
used to account for 20% and has now been increased to 35%; reading
comprehension has decreased from 40% to 35%. General knowledge
has replaced the previous “structure and vocabulary”, accounting for
15%; it is composed of a cloze test, error correction, translation and
short questions and answers. This change is in line with the pro-
gressive erosion of the dichotomy between grammar and vocabulary
and aims to test learners’ ability to use language rather than their
knowledge about the language. The writing component remains the
same, accounting for 15%.
The nation is making efforts to improve university ELT, and the
reform in a series of important areas sounds very promising, despite
all the problems (old or new). The reform has just started and it will
take some time to see whether it will be effective and achieve its goals.
The reform can be considered an important step towards improving
the current situation of Chinese ELT. If the Chinese ELT authorities
were to consider developing a comprehensive framework based on
which Chinese learners could assess their own language achievement/
levels and compare them with language learners of other countries,
such a framework would need to be close to established international
standards, e.g. the Common European Framework of Reference for
Languages 68 . A well-constructed language assessment/ description
framework can help both language teachers and language learners to
understand the pedagogical goals of language learning and is hence
more oriented to the attainment of the goals.

68 See <http://www.coe.int/t/dg4/linguistic/CADRE_EN.asp> for more informa-

tion about the framework; see also <http://www.asiatefl.org/2006conference/
tentative_conference_schedule.html> for a proposal to implement the frame-
work in the Asian context.

256
 Students

Listening Speaking Reading Writing Translation

Computer Classroom teaching

(Stand-alone or network)

Self learning + teacher tutor Teacher’s instruction

Teachers

Figure 15. The new teaching mode combining classroom teaching and language
software (The National Education Ministry, 2004: 8).

10.6. Conclusion

In this chapter I provided the historical, social, cultural and linguistic

backdrop to the teaching and learning of English in China, a necessary
first step if we are to understand the approaches adopted and
difficulties encountered in the acquisition of English vocabulary by
Chinese students. The chequered career of English within the
educational system until little more than two decades ago, the dearth

257
of properly trained teachers, the great cultural divide which is only
now beginning to be bridged, the profound linguistic differences, the
conflict between traditional Chinese and Western values which makes
it difficult, sometimes almost impossible, to apply modern methods of
learning, the very different relationship between students and teacher,
the extreme competitiveness within the educational system where
examinations have failed to encourage the development of language as
a tool of communication, all these, as we have seen, are but a sample
of the obstacles that have to be surmounted if Chinese students are to
acquire a mastery of English, in particular its lexis.

258
Chapter 11. Vocabulary Learning Difficulties
for Chinese Learners

11.1. Three areas likely to cause learning difficulties

In Chapter 4 (Vocabulary Learning Difficulty), we looked at three

factors that affect vocabulary learning difficulty, namely cultural
difference, linguistic distance, and lexical form confusion. In this
chapter, we shall see in what way these areas pose specific learning
difficulty for Chinese learners and look at the typical lexical errors in
each area. Three points have to be made before going further. First, it
is not my intention to claim that learning difficulties necessarily lead
to lexical errors. Lexical errors occur when either the lexical meaning
or form is not learned or partially learned and when the target word is
incorrectly comprehended or produced. Second, only lexical errors
that are likely to be caused by learning difficulty are listed. This is
judged by teacher experience. Third, the lexical errors illustrated are
not meant to be exhaustive but to be representative and informative, so
that readers can have a general idea of what kind of lexical errors
Chinese learners are prone to make. The lexical errors selected below
are from three main sources: (1) Chinese or native English teachers’
collection of their students’ lexical errors; (2) empirical studies on
Chinese learners reported in the literature; (3) Chinese Learners’
English Corpus (CLEC)69.

69 It was compiled by Gui and Yang (2002) and contains more than one million
words, composed of five types of learners: secondary school students, uni-
versity students at the Band 4 level, university students at the Band 6 level,
university English majors in 1st or 2nd year, and university English majors in
3rd or 4th year. The online searching engine is via: <http://www.clal.org.cn/
corpus/EngSearchEngine.aspx>.

259
11.2. Difficulties due to cultural difference

As discussed in Chapter 4, vocabulary in a language is closely related

to the culture of its speakers. Learning an L2 vocabulary naturally
entails learning the specific lexical concepts in the L2 culture, often
associated with the speakers’ customs, beliefs, attitudes, behaviour, etc.
There is a huge difference, covering a wide range of domains, between
Chinese culture and the culture of English-speaking countries, or
broadly Western culture. The 19th century English poet and prose
writer, Rudyard Kipling, once made the widely known statement:
“East is east and west is west and never the twain shall meet.”
Although he was referring to India as the East, this remark, though
unquestionably exaggerated, applies equally well to China. Learning
culture is perhaps as complex as learning language itself. Successful
language learning involves the process of constructing new or
different concepts of the L2 and mapping them correctly on to the L2
word forms. This reconstruction and remapping of the meaning
system can be a life-long process. We also know that lexical concept
gaps between the L1 and the L2, and culturally loaded words are two
areas that cause difficulty in learning vocabulary even for advanced
learners.
There may be a number of types of lexical gaps between the L1
and the L2, but two types will definitely cause vocabulary learning
difficulty: (1) a given concept in the L2 is absent from the L1; (2) two
or more concepts correspond to only one L1 concept. The first type of
lexical gap only makes it difficult to learn some English words such as
privacy and community initially; once they are learned, they are quite
safe in the learners’ mental lexicon and even reach a higher degree of
automaticity and accuracy than other L2 words which have lexical
overlapping with the L1 (Jiang, 2000: 68). By searching the CLEC, it
is found that privacy only occurs twice in the whole one million-word
corpus and community occurs 14 times. In contrast, the two words
occur 9 and 378 times per million respectively in the academic
component of BNC. This may suggest that learning these two words
(at the productive level) is quite difficult for Chinese learners,

260
although they might be able to understand the word meaning correctly.
Lexical errors caused by the second type of lexical gap are often due
to the mixing up of two related but different word meanings which are
not distinguished in the L1. The English word pair complex/com-
plicated correspond to only one Chinese equivalent fuza. Most
Chinese students, even when they have reached the advanced level,
have difficulty in using the two English words correctly. It is the same
for another pair problem/question which only has one Chinese trans-
lation: wenti. Clothes of Western style are relatively recent in China,
first appearing in the 19th century. The Chinese translation of both the
top and the whole (top and trousers) of the Western style is xizhuang,
literally meaning western clothes and corresponding to the English
word suit. Consequently, most Chinese learners use the word suit to
refer to the jacket worn on its own as well as to the jacket and trousers
worn together.
Similarly, culturally loaded words are prone to cause lexical
confusion as Chinese learners tend to attach connotations of some
Chinese words to their equivalent English words which have different
connotations. Often, this type of lexical error is difficult to detect since
the word use can be considered accurate; nevertheless, what the
(Chinese) speaker or writer intends to express might well be mis-
understood by the native listener or reader. We can take humble as an
example. This is very much used in Chinese as the word means to be
modest and show respect to others, a virtue greatly valued in
Confucianism. In contrast, humble in English is often associated with
negative connotations such as weak, cowardly, poor. Consider an
English conversation between a Chinese and a native English speaker,
where the Chinese uses humble to describe his/her interlocutor and
considers this as praise, while the other might take offence if s/he does
not know the positive connotation of humble in Chinese culture.
Furthermore, L2 learners may have acquired a certain degree of fluency
but produce a lot of incomprehensible L2 sentences, misusing and
misunderstanding culturally loaded words, or simply trying to express
themselves through the eyes of their L1. This is often the case with
Chinese learners of English who cannot go to an English-speaking
country and most of whose learning occurs in the language classroom
or on their own.

261
11.3. Difficulties due to linguistic distance

As shown in Chapter 4, the major learning difficulty caused by

linguistic distance is the semantic difference between the L1 and the
L2. Semantic difference not only covers lexical gaps and culturally
loaded words, it is also concerned with how meaning is constructed
at word, phrase, and sentence levels. Many Chinese words or
characters (zi) are semantically transparent due to their pictographic
origin, in contrast to the abstractness of word meaning of English
and many other Western languages (see Vanparys et al., 1997).
Examples include fei ji (flying machine for plane), fen bi (powder
pencil for chalk), shou ji (hand machine for mobile phone), dian nao
(electronic brain for computer), etc. Another important feature is that
the order of the words in a phrase or sentence determines the
meaning of the phrase or the sentence. See the following examples
(Vanparys et al., 1997: 151).

Chinese English
yi yuan cure/doctor building = hospital
xing zhen lou administration building
bai huo da lou one hundred goods large building = department store
zhao dai suo guesthouse
ping fang flat house = bungalow

It is clearly shown that in a Chinese lexical item (made up of several

characters) designating a building, its typical function or feature pre-
cedes the part for the building as a modifier. In this way, the meaning
of the lexical item is fairly transparent. The majority of Chinese
collocations, phrases, and sentences are constructed in such a
meaningful and consistent way, although there are also cases where
word meaning is less transparent like in many Western languages.

262
11.3.1. Errors due to meaning confusion

Due to the tremendous difficulty in handling the semantic difference

between Chinese and English, learning English words places a high
cognitive load on many Chinese learners who often use rote or
mechanical repetition 70 in memorising English words. Often the
match between meaning and form is vague, incomplete or even wrong,
which inevitably leads to lexical errors. In an empirical study on
lexical errors made by Chinese learners, Gu and Leung (2002)
administered the Vocabulary Size Test (Goulden, Nation,/Read 1990)
to 125 Chinese learners of English who were asked to write a response
in either Chinese translation or English synonym(s) for each prompt
English word listed in the test. By collecting the responses, they
identified four types of errors: form errors, meaning errors, form and
meaning errors, and part of speech errors. The last three types are
related to meaning confusion. The two authors define meaning errors
as “wrong responses from the retrieval of a word of related meaning,
rather than of similar form” (Gu/Leung, 2002: 133). They identified
two sub-types of meaning errors: paradigmatic errors and syntagmatic
errors (see Table 9). Gu and Leung refer to form and meaning errors
as “those errors that resulted first from a confusion of form, followed
next by a confusion of meaning” (Gu/Leung, 2002: 134). They listed
some examples (see Table 10). Part of speech errors are caused by “a
vague match of form and meaning, without due knowledge of how the
word should be used syntactically in a sentence” (Gu/Leung, 2002:
135). (See Table 11.) The two authors attributed the part of speech
errors to the learners’ insufficient grammatical knowledge and claimed:
“Part of speech errors may suggest a decontexualised, form-meaning
pair strategy for vocabulary learning that is often found among
Chinese learners of English” (Gu/Leung, 2002: 135−136).

70 It is also commonly agreed that Chinese learners’ favouring rote in learning

English is partially due to the way they learned their L1, Chinese.

263
 Prompt word Reponses by subjects
cotton maize (wheat), linen
accurate fluent, clear, real, right
Paradigmatic hostile hate, hatred, horrible
dig drill
crook evildoer
block cement block, traffic jam, square, crowd
Syntagmatic hostile enemy, conflict
dig hole, well
crook vicious

Table 9. Examples of Chinese learners’ meaning errors (adapted from Gu/Leung,

2002: 133).

Target word Form confusion Meaning confusion – Response

inhabit inherited learned, nurtured
precious precise clear, detailed
precious gracious grateful
hostile hostel residence
hostile hesitate doubt
crook criminal vicious
inhabit habit accustomed to, customary
inhabit no habit not accustomed to

Table 10. Examples of Chinese learners’ form and meaning errors (adapted from Gu/
Leung, 2002: 134−135).

Target word Error triggered

inhabit habitation, inhabitant, inhabitable
hostile hostility

Table 11. Examples of Chinese learners’ part of speech errors (adapted from Gu/
Leung, 2002: 135).

11.3.2. Errors due to other areas

There are other difficult areas that can be observed: pronunciation,

orthography, grammatical patterns, collocation, and frequency.
Generally speaking, if the L2 sounds exist or are similar in the L1,

264
they will be easy to pronounce, or vice versa. Some English vowels,
such as /i:/ and /i/, are not distinguished by many Chinese learners as
there is no such contrast in the Chinese sound system (Chang, 2001:
311). In the same way, Chinese learners do not always distinguish
between /u:/ and /u/, /ae/, /a:/, //, or /e/, etc. (Chang, 2001: 311)71.
Also, Chinese learners tend to lose the voiced feature in producing /b/,
/d/, /g/ as these are voiceless in the Chinese sound system (Chang,
2001). As Chinese is a non-alphabetic language, many learners are not
sensitive to the alphabetic writing system of English particularly when
the word is made up of long strings of letters or the spelling does not
conform to the rules. Consequently, learners of English often replace
some letters with different ones, or omit them. For example, diner
(dinner), docter (doctor), patten (pattern), unfortually (unfortunately),
studing (studying), etc. (Chang, 2001: 313). Sometimes the wrong
spelling is caused by learners’ inaccurate pronunciation as in compus
(campus) and swallen (swollen) (Chang, 2001: 313). He (2004: 298)
concludes that the source of spelling errors comes from negative L1
transfer, overuse or misuse of spelling regulations, mispronouncing
words, and insufficient or less standard teacher/textbook language
input, etc. See He (2004) for a more detailed analysis of spelling
errors by Chinese secondary school students.
In respect of lexical errors caused by grammatical patterns,
Chinese learners generally find it difficult to learn the article the: they
either use it incorrectly or do not use it at all. Many learners often
confuse the gender pronouns she and he in spoken English probably
because of (1) the similar sounds of the two words and (2) the two
Chinese equivalents have the same pronunciation ta although the
written forms are different: 他 (he); 她 (she). The absence of un-
countable nouns in Chinese makes it difficult to produce the correct
form of uncountable words, e.g. moneys, works, advices, etc. For the
same reason, Chinese learners cannot distinguish between the present
participle and the past participle of a verb when they are used as
adjectives; for example:

71 This difficulty in pronouncing certain English vowels may not be unique to

Chinese learners and those learners (including European learners) whose
mother tongue does not have similar vowels may also face this difficulty.

265
(9) I feel boring during the history class

(10) We are not very interesting in this project

Another frequently quoted feature of the Chinese language is the

absence of tense and inflections. It is widely documented that Chinese
learners have difficulty in correctly using plurals, subject-verb agree-
ment, and past tenses of English words. See the following examples.

(11) She brought some apple

(12) He give his word to other people

(13) He has been to the cinema last night

It should be noted that these errors are not only made by low level
students but also by advanced learners, particularly in spontaneous
speaking. It is true that in grammar tests most learners can produce the
right form for each target word by using their conscious or declarative
L2 knowledge. However, this knowledge might not have become
proceduralised for subconscious, automatic access. This also fits in
with the three-stage lexical development model of Jiang (2000, see
also 2.1.2.) where morphophonological errors as such persist in
learner language, particularly spontaneous oral language, for a long
time, until a very advanced stage.
Many Chinese learners tend to collocate English words in the
same way as in Chinese. This phenomenon is not unique to Chinese
learners but is rather prevalent in learners of many other mother
tongues. However, the specific features of collocations of L2 words
differ among learners of different L1. Consider this collocation
produced by an advanced Chinese learner: reach the objective. This
phrase is perfectly understandable, but sounds somewhat strange, or at
least unidiomatic; we usually say: attain or achieve the objective.
Apparently, the learner is confusing goal and objective, which can
both be translated by one Chinese equivalent mu biao, which is often
preceded by da dao, literally translated as reach. Since reach the goal
is perfectly correct, the learner assumes that reach the objective should
also be correct. In this case, the learner knows the basic meaning of

266
objective, but does not know its lexical constraints, or what other
words it conventionally takes. Such inappropriate collocations of
words are common for Chinese learners, even when they have reached
an advanced level.
As for frequency, Chinese learners tend to overuse certain L2
words whose underlying lexical concepts happen to be important in
the L1 oral/written discourse. On the other hand, they may underuse
other words whose lexical concepts do not exist in the L1. Taking
connectors as an example, some connectors, such as because, so,
though, but and then, which indicate direct causal relationship or order
of events, feature prominently in Chinese discourse; hence it is likely
that Chinese learners will overuse these words. However, they may
underuse other connectors, such as therefore, thus, hence and however,
relating to logical progression or alterations to the direction of an
argument. Table 12 presents raw frequencies of these words between
CLEC and the academic component of BNC72.
From the table, it can be seen that because and so are very much
overused by Chinese learners, tripling or quadrupling native speaker
use. Two Chi-square tests were performed, one for the overused words
(because, so, thought, but, then) and the other for the underused words
(therefore, thus, however, hence) between CLEC and BNC. The result
of both tests is unquestionably significant (p < 0.001), as can be seen
from the table. It should be noted that the underused connectors do
exist in Chinese and they are frequently used in Chinese discourse.
Learners’ underuse of these connectors may be tentatively attributed to
two reasons. First, the three connectors – therefore, hence, thus – are
usually translated by yingci, suoyi, conger, which differ very little in
meaning and can also be back translated as so. From the table it is
clearly shown that so is very much overused (four times more than by
native speakers) and this may explain why the other three related
connectors would be underused. Secondly, underuse of these
connectors may be due to the fact that they are introduced into the text
books at a later stage. For example, when I looked at the frequencies
of these connectors between different learner groups, I found that

72 The academic component of BNC was chosen to make it comparable to CLEC

which consists of essay writing for English exams.

267
actually the learner group with the highest proficiency, the advanced
English majors, used therefore almost as frequently as native speakers,
if the relative frequency is calculated based on the raw frequency.
Nonetheless, they still use however, thus, and hence less frequently
than native speakers, but the gap is less wide than when all learner
groups are included. By contrast, the learner group with the lowest
proficiency, secondary school students, only used thus five times in
total and the three remaining words not at all!

Connector CLEC BNC Overuse/underuse Chi-square tests

percentage
Because 3010 737 Oa: 308%
So 6404 1228 O: 422% x2(4, N = 24934)
Though 606 350 O: 73% = 1117.26,
But 6230 3607 O: 73% p < 0.001
Then 1840 922 O: 85%
Therefore 312 551 Ub: 43%
Thus 280 583 U: 52% x2(3, N = 3742)
However 643 1217 U: 47% = 23.82,
Hence 26 130 U: 80% p < 0.001
Note: aO = overuse; bU = undersue.

Table 12. Raw frequencies 73 of connectors in CLEC (1.07 million words) and the
academic component of BNC (in per million running words).

In general, the results support the findings of the study reported by

Granger and Tyson (1996) in which they demonstrated that both over-
use and underuse of connectors exist in learner language and that the
overuse may indicate the L1 influence on learner language. However,
the reasons, particularly for underuse of certain connectors, deserve
further exploration.

73 What is compared is the frequency in the 1.07 million words in CLEC and the
frequency per million running words in BNC. They are more or less com-
parable in size, therefore raw frequencies are used for each corpus.

268
11.4. Difficulties due to confusion of lexical forms

Chinese learners have great difficulty in distinguishing words that are

similar in form as they are not very sensitive to the alphabetic spelling
of English words. Gu and Leung (2002) identified six categories of
lexical errors (see Table 13) produced by Chinese learners based on
Laufer’s 10 categories of synforms (see 4.3.: Table 1). The two
authors presented an additional group of lexical errors based on
syllables rather than affixes, which, according to them, was “a more
appropriate way of analyzing the errors of Chinese L1 speakers when
they confuse similar forms of English words” (2002: 132). They listed
the following categories (see Table 14).

Category 3 Suffix present in one synform but inhabit – inhabitant,

not the other inhabitable
Category 4 Synforms with the same root but microscope – telescope
different prefixes
Category 6 Synforms identical in all their inhabit – inhibit
phonemes except one vowel
Category 8 One consonant is different in the precious – previous
synforms crook – brook
block – clock
Category 9 One synform has one additional crook – cook
consonant
Category 10 Synforms have the same precious – precise
consonants but different vowels

Table 13. Form errors classified according to Laufer’s categories (Gu/Leung, 2002: 131).

269
 1 Similar pronunciation hostile – hostel
2 Same final syllable precious – delicious, envious, previous,
serious, monstrous
microscope – telescope
cotton – button
inhabit – prohibit
3 Same initial syllable hostile – hostage, hospitable, host
4 Same two initial syllables accurate – accuse
5 Same beginning and final syllables accurate – accelerate
6 Same medial syllable inhabit – rehabilitate
7 Similar medial and final syllable accurate – circulate (pronunciation of r
and l may be confused as well)

Table 14. Examples of Chinese learners’ form errors based on syllables (adapted from
Gu/Leung, 2002: 132).

In particular, they found one type of lexical formal error specific to

Chinese learners: sometimes Chinese learners gave Chinese words
that share one character with the Chinese translation of the target word
but differ in the other character(s). Apparently, the learners translated
the target word into Chinese and supplied another Chinese lexical item
which was similar but not identical to the Chinese translation of the
target word. Here the lexical confusion is caused by the form of the
translated Chinese equivalent of the English word rather than the
English word itself. (See Table 15.)

Prompt word Chinese equivalent Errors elicited Back translation

precious 珍贵，宝贵 珍惜 to value, to cherish
昂贵 expensive, over-priced
高贵 noble
accurate 准确的/精确的 正确的 right, correct
microscope 显微镜 望远镜 telescope
放大镜 magnifier
inhabit 居住 移居 migrate

Table 15. Confusion of Chinese characters in Chinese equivalents (adapted from Gu/
Leung, 2002: 133).

270
11.5. Exploring lexical errors
from a psycholinguistic perspective

We have looked at some difficulties that Chinese learners encounter in

learning English vocabulary. These difficult areas are in fact likely to
result in lexical errors. Lexical errors occur in both comprehension
(reading and listening) and production (speaking and writing). In order
to know how lexical errors occur, we first have to know what is
involved in comprehending language input or producing language
output. Based on the speech production model of Levelt (1989, 1993),
De Bot et al. (1997) developed a model of lexical comprehension/
production for oral and written modalities (see Figure 16). It is
assumed that this model should hold true for both the L1 and L2.
From Chapter 2 we know that each lexical item comprises two
parts, lemma and lexeme; the lemma contains semantic and syntactic
information and the lexeme the morphological and phonological
information. When comprehending a lexical item, whether in oral or
written form, the item has to be first decoded phonetically and then to
be matched with the lexeme which will search for the corresponding
lemma information in the lexicon. When both the lexeme and lemma
information are accessed, the word will be comprehended as a given
concept. The production of lexical items follows the same procedure,
but in an inverted order. See Levelt (1993) for a detailed account of
the whole process. I shall try to explain lexical errors made by
Chinese learners in the light of this lexical comprehension and produc-
tion model.
When a Chinese learner encounters a word, say, precious, it
sounds familiar to her/him, as s/he knows the word string prec___.
The search for the lexeme begins and s/he finds precise, and then s/he
gets the corresponding “lemma information”74 of precise. The word
precious is therefore mistaken for precise, taking the form of the
English word precise or its Chinese equivalent jing que de. In this case,
the word precious may have or may not have been registered in the

74 For L2 learners, the L2 lemma information can be incomplete or different

from that of the L1 speaker, as shown in 2.1.2.

271
learner’s L2 lexicon. If it is or was, this will probably be only the
lexical form, as depicted in the formal stage of the lexical entry
development (see 2.1.2.). On the other hand, when s/he tries to write a
word, say pattern in a sentence, the conceptualiser (see Levelt, 1989,
1993), where concepts are formed, first produces the preverbal
message for this word. The preverbal message is then matched with
the lemma information of the word pattern. The lemma continues to
search for the lexeme information of the word via phonological
encoding; s/he finds the non-existent lexeme patten instead of pattern
as the former looks very similar and is pronounced in a very similar
way to the latter. It is certain that the learner has not correctly
mastered the formal features of the word pattern.

comprehension production
concepts

lemmas

lexicon lexicon

lexemes

decoding encoding

speech input written input speech output written output

Figure 16. Lexical comprehension/production model (adapted from De Bot et al.,

1997: 315).

272
Applying the three-stage model for lexical entry development of Jiang
(2000, 2002), we can predict what type of lexical errors are likely to
occur in each stage:

1. In the formal stage, where only the phonological information is

in the lexical entry, learners may (a) make a retrieval failure (no
recognition of the word), (b) pronounce or spell the word
incorrectly, (c) associate it with a wrong meaning or part of
speech, (d) link it with other words similar in form.
2. In the L1 lemma mediation stage, learners may (e) use the word
inappropriately in meaning or (f) in incorrect grammatical
patterns (e.g. collocations).
3. In the integration stage, learners may (g) have slip of the tongue
type lexical errors, similar to those made by native speakers.

This only suggests the general pattern of lexical errors that are likely
to occur at each stage. Usually errors common to the previous stages
can be present in the current stage but maybe less frequently. It is also
clear that a given L2 word is more likely to lead to lexical errors when
it is in its formal stage in the lexicon than in other more advanced
stages.

11.6. Conclusion

In this chapter we went on to explore the three factors that affect

vocabulary learning difficulty identified in Chapter 2 in the context of
English learning in China by looking at the typical lexical errors made
by learners. Multiple sources were used to collect the lexical errors,
including the research literature, classroom observation, writing
assignments and a learner corpus. The lexical errors made by Chinese
learners, like other language errors, fall into two broad categories:
interlingual and intralingual errors. Sometimes they may come from
both sources, as we saw, there being no clear-cut distinction between

273
the two. Generally speaking, lexical errors are caused by imperfect
match of meaning and form of the L2 words during the acquisition
process, which might be due to learning difficulty, e.g. cultural and
linguistic difference between the L1 and L2 or lexical form confusion,
learning strategies, learning effort, mental processing, etc. To explain
learner lexical errors is a complicated issue. We need to look at the
learner’s L1 and L2 mental lexicon, what lexical information is
contained in the lexical entry for the target word, how lexical items are
retrieved and what the nature of the lexical error is, whether it is due
to defective lexical representation or simply a slip of the tongue.

274
Chapter 12. Two Empirical Studies of Chinese
Learners’ Approaches to Vocabulary Acquisition

12.1. An overview from the literature

If vocabulary learning difficulties encountered by Chinese learners are

one important source to explain their lexical development and lexical
errors, the way vocabulary is learned and taught is another important
source to explain the same two phenomena. Chinese learners’
approaches to English vocabulary acquisition can be generally
described as “explicit learning”, as hinted at earlier. In the Chinese
culture of learning, language is essentially viewed as composed of
linguistic knowledge so grammar and vocabulary are always the two
focal points in any language learning program. Vocabulary has been
always emphasized in the language textbooks and by the teachers and
the learners themselves. We shall now take a look at the textbook
College English that was used extensively in universities all over
China for more than 10 years. Almost all the texts are presented in the
following sequential order (see also Cortazzi/Jin 1996a: 66).

1. A text of one or two pages.

2. A list of new words, for each of which three types of informa-
tion are provided: (1) phonetic symbol, (2) part of speech, and
(3) a Chinese or English explanation.
3. A few phrases and expressions paraphrased in English or Chinese.
4. Some notes for the background information about the text.
5. An oral activity in which the students are asked to read aloud
some selected sentences or phrases and pay attention to
sentence boundaries and stresses.
6. An exercise of text comprehension.
7. A few vocabulary exercises such as matching correct definitions
and filling in blanks.

275
8. A few extra vocabulary exercises which focus on affixes to
build up more words in different parts of speech.
9. A grammar exercise on practising some structures to build
sentences.
10. A cloze test.
11. A translation exercise.
12. Further reading exercises.

To this end, it is clearly shown that language learning is con-

ceptualised as linguistic knowledge, grammar and vocabulary. More
attention seems to have been paid to vocabulary since most of the
grammar has been covered in the six-year English course in secondary
school. Now, at the university, the language learners’ primary task is
to enlarge their vocabulary. Each English lesson is conducted in a way
similar to the following (see also Cortazzi/Jin 1996a: 66):

1. Students are usually asked to prepare each lesson before coming

to the class; this includes familiarizing themselves with the new
words listed in the text, listening to the recording of the text, or
practising reading the text aloud.
2. In the classroom the teacher will ask the students to read the
vocabulary items and check the pronunciation. The teacher will
give more information in respect of some lexical items, such as
example sentences, more meanings, and usage, etc. S/he may
ask the students to read after her/him each lexical item to make
sure the pronunciation is correctly mastered.
3. Then s/he will start explaining the text: ask comprehension
questions to make sure students understand the text, ask
students to paraphrase some complex sentences, select more
lexical items and expressions that may be useful, and some
grammatical points as well.
4. The teacher continues to go through all the exercises chrono-
logically, one by one, or sometimes makes a selection, but goes
through them in a meticulous way, to make sure that students
have mastered the required vocabulary and grammar in the
lesson.

276
In this teaching procedure, teacher talk will be dominant in the class
since s/he has to present a considerable amount of linguistic knowl-
edge (vocabulary and grammar) to the students. There may be
interactional activities involving the students’ participation. However,
there will probably not be enough time to organise these activities as
the main effort will be put on achieving the syllabus objectives, the
onus being on every teacher to do so. Although the current reform of
Chinese ELT is calling for changes in both the textbooks and the
teaching approaches, it will take time to see some really substantial
change to the whole system. Moreover, the old system has been there
for a long time and it still exerts a huge influence on the less young,
adult language learners. If the reform turns out to be effective, the
younger generation will undoubtedly benefit the most.
It has been mentioned implicitly or explicitly in a few earlier
places in this book that Chinese learners use memorisation, be it
simple rote or other more sophisticated forms75, as the predominant
strategy to learn vocabulary. In an investigation into Chinese universi-
ty learners of English, Gan et al. (2004) found that both successful and
unsuccessful learners use rote as the most valued vocabulary learning
strategy. What makes the difference is that the former use other
accompanying strategies such as planned review, regular reading,
putting the words into use, etc., which the latter do not employ. Hu
(2002: 100−101) concisely summarises language learning strategies
used by Chinese learners as the four R’s and the four M’s. The four
R’s stand for reception, repetition, review, and reproduction. The four
M’s stand for meticulosity, memorisation, mental activeness, and
mastery.
There may be a number of reasons for such a specific Chinese
approach to vocabulary acquisition. First, it fits in well with the
traditional Chinese culture of learning. Second, it has something to do
with how the L1, Chinese, is learned. Most primary school Chinese
children have to spend several hours every day memorizing and
practising writing Chinese characters. The way the L1 is learned will

75 Hu (2002) argues that memorisation is not necessarily associated with rote

and the way Chinese learners use memorisation might suggest a deep learning
approach, or learning with understanding.

277
certainly affect the way the L2 is learned, at least in the beginning.
Third, it is also due to the difficulties Chinese learners encounter when
learning English vocabulary.

12.2. Questionnaire: Teachers’ beliefs and practices

regarding vocabulary teaching

12.2.1. Constructing a questionnaire of teachers’ beliefs and practices

regarding vocabulary teaching

A survey of teachers’ beliefs and practices regarding vocabulary teach-

ing was conducted in April 2005 at China Three Gorges University. A
questionnaire was given to 52 English teachers of non-language
specialists. It comprises two parts: the first part aims to find out what
beliefs teachers hold for vocabulary teaching and learning; the second
part is to find out what they actually do when teaching vocabulary in
classrooms. Following Richards (1994: 1), teachers’ beliefs are
defined as “the information, attitudes, values, theories, and assump-
tions about teaching and learning which teachers build up over time
and bring with them to the classroom”. Teachers’ beliefs determine
and guide their actual teaching practice. In respect of the actual
teaching practice of vocabulary teaching, the questions are grouped
into six categories: how to select the lexical items, how to present
vocabulary, what aspects to focus on in teaching lexical items, what
activities to use to assist vocabulary teaching, what vocabulary
learning strategies to be recommended or taught to students, and how
to help students to review vocabulary. After a final selection, 56 items
(statements) were included in the questionnaire and grouped according
to these seven categories (one in the first part and six in the second
part). Each respondent was asked to indicate the statements that were
true for them. (See Appendix 1, for the questionnaire.)

278
12.2.2. Results and discussion of the questionnaire

For each statement, a percentage is calculated for the “Yes” responses

out of the total responses (52). The results are set out in Table 16.

Statements Percentage
n = 52
Beliefs about vocabulary teaching and learning
Vocabulary should be taught explicitly in class 55.77%
Vocabulary should be learned off by heart from bilingual lists 13.46%
Students acquire a large amount of vocabulary through extensive
90.38%
reading
Doing exercises can help learn vocabulary 57.69%
Vocabulary should be learned through conversations or other
50.00%
communicative activities
Vocabulary should be consolidated by use in different contexts 65.38%
How do you select the lexical items to be taught?
It is a key word for understanding the text. 80.77%
It is a frequent word that will appear in many other contexts 80.77%
The course book says it is important to know 19.23%
It belongs to the lexical field that we are studying 38.46%
It is a useful word for my students 71.15%
The word is difficult to use 21.15%
The word is likely to be tested in exams 32.69%
How do you present vocabulary?
Visually (write the word out) 61.54%
Aurally (say the word) 57.69%
Kinetically (body movement, gestures, acting out the meaning, etc.) 69.23%
Using visual aids (videos, pictures, diagrams, real objects, etc) 55.77%
Using aural aids (tapes, CDs) 25.00%
What aspects do you focus on in teaching lexical items?
Their pronunciation 61.54%
Their spelling 21.15%
Their affixes and roots 65.38%
Their Chinese translation 13.46%
Their meaning in English (a paraphrase) 65.38%
Related words such as antonyms and synonyms 63.46%

279
I use them in good example sentences 88.46%
I describe their usage 61.54%
I indicate their register 38.46%
I show some of their collocations 76.92%
I indicate the part of speech they belong to 40.38%
What activities or tasks do you use to assist vocabulary teaching/learning?
Drill repetition 65.38%
Vocabulary exercises (e.g. gap-filling, multiple choice, etc.) 73.08%
Vocabulary games (guessing games, crosswords, etc.) 15.38%
Get students to work in interactive conversations 44.23%
Role plays 23.08%
Listening tasks 28.85%
Writing tasks 53.85%
Computer software 7.69%
Internet sites 11.54%
What vocabulary learning strategies do you recommend or teach to students?
Read extensively to enlarge and consolidate the vocabulary 90.38%
Use English films or songs to enlarge and consolidate the vocabulary 51.92%
Participate actively in oral activities to enlarge and consolidate the
59.62%
vocabulary
Think in English 48.08%
Organise the lexical information in a systematic way, e.g. keep a
36.54%
vocabulary notebook or vocabulary cards
Use contextual or formal clues to guess the meaning 78.85%
Use strategies to put vocabulary in a context that will help memorise it 71.15%
Learn other words in the semantic network, e.g. for types of transport,
32.69%
train, car, bus, bike.
Use sound or rhymes to memorise vocabulary 36.54%
Use imagery to memorise vocabulary 19.23%
Ask students to reflect on their own progress in vocabulary learning 9.62%
I encourage students to use vocabulary they have just learned 76.92%
I teach them how to use vocabulary learning strategies in a systematic
17.31%
way.
How do you review vocabulary with students?
Review new words in lists regularly 1.92%
I review vocabulary in semantic fields 65.38%
I test new vocabulary regularly 44.23%

280
I regularly schedule written or oral class activities that focus on
50.00%
reviewing recently learned new vocabulary
I propose further reading or listening activities in which the words occur 80.77%

Table 16. Percentage of vocabulary teaching beliefs and practices.

As for beliefs about vocabulary teaching and learning, what is

commonly agreed by most teachers is that vocabulary can be acquired
by extensive reading (90.38%). In sharp contrast, a very small number
of teachers believe that vocabulary should be learned by rote from
bilingual lists (13.46%). On the other hand, a considerable number of
teachers also believe that explicit teaching (55.77%) and learning
(doing exercises, 57.96%) can be helpful. Those who firmly believe
that vocabulary should be learned by communicative activities might
be a little disappointed by the moderate percentage (50%) of teachers
who believe so. Regarding the belief in consolidating vocabulary by
use in different contexts, the percentage (65.38%) of teachers who
hold this belief is slightly better but still below Western expectations,
it being generally held that language is learned by use in communica-
tive learning.
The important criteria for selecting lexical items to be taught are
quite clear-cut: key words for understanding the text (80.77%),
frequent words (80.77%), and useful words (71.15%). Other criteria
are much less appreciated. The most favoured way of presenting
lexical items appears to be by doing so kinetically (69.23%), visually
(61.54%), using visual means (55.77%), and aurally (57.69%). The
least favoured is using aural aids (25%). This is probably due to the
fact that many teachers do not think there is enough time to play
recordings to students in the class and students are supposed to listen
to recordings on their own after class. It also could be that these
teachers simply do not value oral aids as much as visual aids in
teaching vocabulary.
As for what aspects to focus on in teaching lexical items, the
majority of teachers choose to focus on example sentences (88.46%)
and collocations (76.92%). Other important aspects are the English
meaning (65.38%) of the lexical items, related words such as
antonyms or synonyms (63.46%), but not the Chinese meaning

281
(13.46%) which students can easily access from bilingual dictionaries
or the vocabulary section of the textbooks. Formal information such as
pronunciation (61.54%) and affixes and roots (65.38%) is also
frequently attended to, except spelling (21.15%) which is probably
considered to be so basic that students are supposed to take care of it
themselves. In addition, word usage (61.54%) is another aspect that
they will pay more attention to. Despite other possible reasons, the
fact that word register (38.45%) is not emphasized by the majority of
teachers is probably due to teachers’ insufficient knowledge of this
area. For all sorts of reasons, most of these teachers have not been
exposed sufficiently to authentic oral/written English to allow them to
have knowledge of word register comparable to that of native speakers.
In a sense, they are still language learners themselves.
It seems that these teachers prefer traditional activities to assist
vocabulary teaching/learning, notably, vocabulary exercises (73.08%),
drill repetition (65.38%), and writing tasks (53.85%). Again, activities
involving communicative language use, such as interactive conversa-
tion (44.23%), role plays (23.08%), and vocabulary games (15.38%),
are less used. So are listening tasks (28.85%). Using computers to
assist vocabulary learning comes out the lowest: internet sites (11.54%)
and computer software (7.69%). Obviously these teachers do not use
computer or internet resources to assist vocabulary teaching, despite
the fact that vocabulary teaching/learning could be easily aided by
computer technology. It would seem that there is still a long way to go
before the goal is attained of integrating computer software or other
computer mediated learning means into language learning as outlined
in the reform of the university English education system (see Chapter
10). What is apparently needed is teacher training to train teachers to
use computers efficiently to assist language teaching and the develop-
ment of good quality language learning software.
In accordance with the popular teacher belief that vocabulary
should be learned through extensive reading, encouraging students to
read extensively to enlarge and consolidate the vocabulary is the most
popular strategy (90.38%) in which teachers would like to instruct
their students. Accordingly, to guess the word meaning via contextual
or formal clues is also a popular strategy (78.85%) that teachers would
recommend to their students. They recommend putting words in con-

282
text (71.15%) to help to memorise the vocabulary. They tend to
encourage students to use the newly learned vocabulary (76.92%) and
to participate in oral activities to practise vocabulary (59.62%). Some
memory strategies, such as imagery (19.23%), rhyming (36.54%), and
using semantic networks (32.69%), are clearly not popular. It is the
same for organizing lexical information systematically (36.54%).
Maybe some teachers do not use these strategies themselves or do not
see much value in them despite the fact that they involve deep mental
processing and would doubtless lead to more efficient learning. The
least popular strategy is to ask students to reflect on their own
progress in vocabulary learning (9.62%); most teachers probably take
it for granted that students should assume responsibility for their own
progress in learning. Finally, only a small number of teachers (17.31%)
claim that they have instructed students systematically in vocabulary
learning strategies.
The most popular recommendation for reviewing vocabulary is
to propose further reading or listening in which the words will occur
(80.77%). Also, teachers will often review words with students in
semantic fields (65.38%) or schedule other classroom activities that
will allow words to be reviewed (50%). Only one teacher (1.92%)
claims that s/he would review words with students in lists regularly;
more teachers claim that they test new words regularly (44.23%).
In conclusion, the predominant teacher belief is that vocabulary
should be learned through extensive reading. Explicit teaching and
learning of vocabulary is also needed as complementary means. It
would seem that teachers generally hold the view that vocabulary
learning is essentially the students’ own responsibility. That vocabu-
lary can be learned in a communicative context does not have
unanimous approval. This probably reflects the fact that communicative
teaching is still rather controversial among Chinese English teachers:
some welcome it wholeheartedly while others are quite sceptical about
its value. They prefer traditional activities, such as drill repetition or
vocabulary exercises, to communicative activities or computer me-
diated learning to assist vocabulary teaching/learning. They encourage
students to read extensively as the primary learning strategy to learn
vocabulary; in addition, they recommend strategies to guess the word
meaning, and to memorise vocabulary from context. It seems that

283
teachers tend to value less memory strategies, such as imagery and
rhyming that involve greater cognitive effort, or metacognitive
strategies, such as organising lexical information or self-checking
lexical progress. This shows that teachers themselves may have
limited knowledge of vocabulary learning strategies. The most
important message conveyed by this questionnaire is that systematic
instruction in vocabulary learning strategies is largely absent in these
teachers’ language teaching practice.

12.3. Questionnaire: Students’ approaches

to vocabulary learning

12.3.1. Constructing a questionnaire of process-oriented

vocabulary learning strategies

A survey of vocabulary learning strategies was administered to 109

Chinese first year university students majoring in subjects other than
languages in April 2005 at the China Three Gorges University. A
questionnaire was designed based on the learning process-oriented
taxonomy of vocabulary learning strategies (see 7.4.4.3.). Eight
distinctive stages can be identified for acquiring vocabulary, from
discovering the new word to making use of the newly learned word. In
compiling the taxonomy, any strategies that fitted in with the eight
stages were included. Strategies specific to Chinese learners were also
selected, combined with references to a number of existing taxono-
mies of vocabulary learning strategies, such as Schmitt’s taxonomy
(1997) and Cohen and Chi’s (2003) Language Strategy Use Inventory
and Index. Finally, 55 strategies were formulated in my process-
oriented vocabulary learning strategies questionnaire. Applying the
six-category language learning strategies identified by Oxford (1990),
I assigned the 55 vocabulary learning strategies to four categories:
memory, cognitive, metacognitive, and social. Authors do not always
distinguish between cognitive and memory strategies as the latter

284
necessarily involve cognitive effort. In other words, cognitive strategy
is a hypernym of memory strategy. Following Schmitt (1997),
memory strategies are treated as being more similar to the mnemonic
type of strategies, which require deep mental processing and effort.
Cognitive strategies will cover those strategies involving direct,
mechanical means of committing vocabulary to memory as well as
strategies involving mental processing. Students are required to
indicate the frequency of each strategy they actually use. Frequencies
range from never, rarely, sometimes, often and always, equated with
the number 1, 2, 3, 4, 5 respectively. (See Appendix 2 for the question-
naire.)

12.3.2. Results and discussion of the questionnaire

An average is calculated for the frequency of use of each strategy.

Each strategy listed is preceded by the initial code of the category they
belong to (cf. Schmitt, 1997). See Table 17 for the results.

Strategies Av. Frequency

1. How do you discover new vocabulary?
COG1 From text books and classroom learning activities 4.13 often
COG2 From lists of vocabulary arranged alphabetically 2.76 sometimes
COG3 From lists of vocabulary arranged semantically 2.38 rarely
SOC1 From English conversations with others (teacher,
2.17 rarely
classmates, foreigners)
MET1 From reading English materials (books, newspapers,
2.95 sometimes
magazines, etc.)
MET2 From English songs and movies/TV programmes 2.78 sometimes
MET3 From the internet (online English pages, chatting) 2.33 rarely
2. What do you do on encountering new vocabulary?
MET4 Ignore it 2.81 sometimes
MET5 Ignore it first and come back later to it 3.63 often
COG4 Try to guess the meaning from the context 3.62 often
COG5 Analyse affixes and roots for the meaning 3.03 sometimes
SOC2 Ask classmates or the teacher for the meaning 2.64 sometimes
COG6 Consult a bilingual dictionary 3.63 often

285
COG7 Consult a monolingual dictionary 1.99 rarely
3. When learning a new vocabulary item, what aspects do you
study?
COG8 Study the pronunciation 4.28 often
COG9 Study the spelling 4.36 often
COG10 Study the part of speech (noun, verb, adjective, etc.) 3.47 sometimes
COG11 Study the Chinese translations 4.45 often
COG12 Study the English explanations 2.76 sometimes
COG13 Study the affixes and roots 2.73 sometimes
COG14 Study the example sentences 3.45 sometimes
COG15 Study the way the word is used (e.g. prevent sb from
3.50 sometimes
doing sth)
COG16 Study the collocations (e.g. make money/ a profit) 3.82 often
4. How do you organise the information about the new
vocabulary?
COG17 Jot down the information on paper 2.48 rarely
MET6 Organise the information in a vocabulary notebook 3.28 sometimes
MET7 Make vocabulary cards 2.14 rarely
COG18 Use the vocabulary section in the textbook 2.67 sometimes
COG19 Use vocabulary lists (e.g. those in vocabulary books
2.23 rarely
for Band 4)
5. How do you memorise vocabulary?
COG20 Say the word aloud several times 3.88 often
COG21 Copy the word out several times 3.54 often
COG22 Look at the word several times 3.34 sometimes
COG23 Memorise bilingual lists 2.49 rarely
COG24 Do vocabulary exercises 2.60 sometimes
MEM1 Connect the word to its synonyms or antonyms 2.78 sometimes
MEM2 Connect the word with other already known words
2.83 sometimes
that have a formal resemblance (e.g. dam, damage)
MEM3 Compare words similar in meaning and study them
3.14 sometimes
together
MEM4 Group and organise words (e.g.. in a meaningful way:
2.61 sometimes
part of speech, topic, etc.)
MEM5 Place the word in a context (a meaningful sentence,
3.32 sometimes
conversation, or story) to remember it
MEM6 Make up a phrase or a sentence containing the word 2.42 rarely
MEM7 Listen to the sound recording of the words 2.71 sometimes

286
MEM8 Make up rhymes to link words 2.23 rarely
MEM9 Practise words (e.g. verbs) by acting them out 2.05 rarely
MEM10 Make a mental image of new words (or the sentence
2.95 sometimes
with the word)
MEM11 Draw pictures to illustrate the meaning of new words 2.01 rarely
COG25 Visualize the word form mentally 3.78 often
COG26 Remember the affixes and roots of new words 3.02 sometimes
6. How do you review vocabulary?
MET8 Review words periodically 3.42 sometimes
MET9 Test yourself on new words 2.72 sometimes
SOC3 Test new words with classmates 2.22 rarely
7. How do you retrieve vocabulary?
COG27 Look at the meaningful parts of the words (affixes or
3.02 sometimes
roots)
COG28 Make an effort to remember where I first met the words 3.58 often
8. How do you make use of new vocabulary?
MET10 Try to use new words in speaking and writing 3.25 sometimes
MET11 Make an effort to use idiomatic expressions 2.88 sometimes
MET12 Try to think in the target language with the new
2.48 rarely
vocabulary
SOC4 Interact with native speakers/teachers/classmates with
1.98 rarely
the new vocabulary

Table 17. Average of vocabulary learning strategy use frequency.

The average frequency for each strategy indicates the general trend of
strategy use among these students. It is seen that students mostly dis-
cover new lexical items in textbooks and language classes (4.13).
They sometimes find new items themselves when reading other
materials (2.95), listening to English songs or watching movies (2.78),
and from vocabulary lists in alphabetical order (2.76). On the other
hand, they rarely discover new words from interactive conversation
(2.17), internet (2.33), or vocabulary lists arranged semantically (2.38).
It seems that language textbooks and classes are the main source for
new lexical items, which is quite understandable since introducing
new lexical items and exercises for practising vocabulary feature
prominently in textbooks. Only occasionally do they take the initiative
to look for new words outside the context of classroom learning. The

287
fact that they rarely find new words in interactive conversation implies
that they probably have little opportunity to speak or rarely voluntarily
seek opportunities to use the language in real communicative
situations. Vocabulary lists arranged semantically rarely appear in lan-
guage learning materials, which is why students use such lists marked-
ly less frequently than those arranged in alphabetical order.
On encountering new words, students often try to guess the
meaning from the context (3.62) or come back to the words later
(3.63). They may check in a bilingual dictionary (3.63) but very rarely
a monolingual dictionary (1.99). Little use of the latter type is
probably due to their lack of English proficiency. Occasionally they
ignore the words (2.81) or else ask classmates or teachers for the
meaning (2.64). Sometimes they will use word part knowledge to
work out the meaning (3.03); acquiring word part knowledge seems to
be an important component of building up vocabulary knowledge in
Chinese ELT, e.g. there are many exercises on building words with
affixes in the textbook.
When studying a new word, very often they will focus on four
types of word information: Chinese translation (4.45), spelling (4.36),
pronunciation (4.28), and collocation(s) (3.82). Less frequently they
will attend to other types of information: word use (3.50), part of
speech (3.47), example sentences (3.45), English explanation (2.76),
and lastly affixes and roots (2.73). It is normal that the most important
information about new words for these learners is the meaning in
Chinese, and matching it with the word form (spelling, pronunciation).
Other types of information, such as word use, part of speech, English
explanation, etc., may be important but are not given priority and are
therefore less attended to.
The results regarding how to organise new vocabulary informa-
tion show that generally these students do not have their own
systematic way of storing vocabulary information. They only use
vocabulary notebooks occasionally (3.28), less frequently relying on
the vocabulary section in the textbook (2.67). If their primary goal is
to remember the meaning of the word as shown above, they probably
do not see the need to arrange neatly all information relating to the
word, particularly as this would involve more time and effort.

288
 The most popular strategies for memorising lexical items are
mechanical means, such as repeating the words orally (3.88) or by
copying the written forms (3.54), or else by doing so mentally (3.78).
Interestingly, the students do not seem to use bilingual lists (2.49) as
often as they make use of rote, despite the fact that the two are closely
associated with each other. It is possible that they dissociate the
bilingual lists from the vocabulary lists/section in their textbooks. The
results also show that the students do occasionally use different types
of deep processing memory strategies, such as placing words in
context (3.32), comparing (3.14) and grouping (2.61) words,
remembering word affixes and roots (3.02), making mental images of
words (2.95), associating words with similar forms (2.83), and
listening to the words to help remember them (2.71), etc. They rarely
draw pictures (2.01) to help memorise words, thinking perhaps that
this might be considered too childish; the same would be for acting
out the word meaning (2.05). Nor do they make up their own rhymes
to link words (2.23), this obviously being more challenging than
linking words verbally (2.42) or by images.
The students do not often review words systematically: they
sometimes review words periodically (3.42), test words themselves
(2.72), but rarely test words with others (2.22). Similarly, they do not
make much use of newly learned vocabulary. They occasionally use
the words in speaking or writing (3.25) or make an effort to use
idiomatic expressions (2.88). Seldom do they use the words
interactively (1.98), which might be due to their lack of opportunity to
speak with native speakers or simply to their failure to take the
initiative to speak with teachers or classmates. To think in English by
incorporating the new words is also rarely used (2.48).
Another way to interpret the results is to look at each of the
eight groups of strategies that correspond to the eight steps of
vocabulary acquisition (see Table 18). This can give us a rough idea
how much effort learners make in vocabulary learning at each step and
how often they use the strategies of each group. Table 18 gives an
average of frequency for each group of strategies.

289
 Group strategies Mean Use frequency
1. Discovering new vocabulary 2.79 sometimes
2. Encountering new vocabulary 3.05 sometimes
3. Studying new vocabulary 3.65 often
4. Organizing information about new vocabulary 2.56 sometimes
5. Memorizing vocabulary 2.87 sometimes
6. Reviewing vocabulary 2.79 sometimes
7. Retrieving vocabulary 3.30 sometimes
8. Making use of new vocabulary 2.65 sometimes

Table 18. Mean of use frequency of group strategies.

Table 18 seems to show that the students only use strategies adequate-
ly in the third step, i.e. when studying the various aspects of the new
vocabulary item (3.65). They also make considerable effort to retrieve
the new item when they encounter it again at the seventh step (3.30).
The remaining steps are not adequately addressed, particularly regard-
ing organizing the lexical information (2.56), making use of new
vocabulary, reviewing vocabulary (2.79) and memorising vocabulary
(2.87).
If we group all the 55 strategies according to the codified
category they belong to, we can see how often the students make use
of each category of strategies in learning vocabulary. (See Table 19) It
is clearly seen that in respect of vocabulary learning, students
generally use more cognitive strategies, followed by metacognitive
strategies, memory strategies and, least of all, social strategies. This is
true in terms of both the number and the average frequency of use of
strategies in each category.

Category of strategies Total number of Mean of use frequency

strategies
Cognitive 28 3.25
Metacognitive 12 2.89
Memory 11 2.64
Social 4 2.25

Table 19. Mean of use frequency of codified strategies.

290
In conclusion, the results of the questionnaire generally confirm what
has been found for Chinese learners in respect of vocabulary learning.
They rely on textbooks and classroom learning activities as the
primary source to discover new lexical items. They lack the initiative
to seek other sources or opportunities to do so, especially from social
interactive conversations. They will try to get the meaning of new
words by either guessing or checking in a bilingual dictionary. They
focus on studying the Chinese meaning and formal information of the
new lexical item. Generally they do not have a systematic way of
recording the lexical information of new lexical items. They use rote
(oral/written repetition) as the primary strategy for memorising words
and the focus is apparently on the word form and the Chinese meaning.
Other memory strategies are also used, but less frequently. They do
not have the habit of reviewing words systematically, which means
words are often learned only once. In compensation, they use more
retrieving strategies to recall the word on later encounters. They
seldom put newly learned words into use. By looking at the strategies
grouped according to the eight learning steps identified, it is found
that only the third step (studying various types of word information)
and the seventh step (retrieving) are dealt with adequately; the
remaining steps are carried out rather inadequately. By grouping the
strategies according to the four categories, cognitive, metacognitive,
memory, and social, it is found that for these students vocabulary
learning primarily involves cognitive, metacognitive and memory
processing and effort, all of which is done individually and seldom
socially or interactively.

12.4. Conclusion

Firstly, it seems that both the teachers and the students believe that
vocabulary should be learned explicitly, which is consistent with what
we saw earlier (see Chapter 10). For both teachers and students, vo-
cabulary is primarily learned through individual activities, involving

291
cognitive, metacognitive, and memory effort, but also, though less
often, by means of social interactive activities. Students rely more on
mechanical means to memorise lexical items, which is inefficient and
may well lead to lexical errors as most students simply link the
Chinese meaning with the English form for a given word regardless of
the differences in meaning, connotation and use between the two
words in each language.
Secondly, teachers opt for extensive reading as a primary means
for students to learn vocabulary. It seems that teachers have not given
sufficient help to their students who, for a variety of reasons, do not
often read other materials except textbooks, contrary to what the
teachers would have expected.
Lastly, it is found that teachers generally do not incorporate
systematic vocabulary learning strategy instruction into their teaching;
they themselves may not have sufficient knowledge of strategy use or
training. This might explain why students have not made adequate use
of strategies: mere unawareness. It might also explain why students
seldom read extensively despite their teachers’ encouragement to do
so: they are simply not equipped with the necessary strategies to do so
efficiently.
We need to introduce a series of vocabulary learning strategies
systematically to students. The students first have to know about these
strategies and then they can choose those that suit them and make
effective use of them. This also calls for teacher training for strategy
instruction. One solution would be to design CAVL software that
would allow students to acquire useful vocabulary learning strategies
in the course of learning new vocabulary.

292
General Conclusion

Acquiring an L2 essentially involves acquiring its vocabulary,

consisting of thousands of individual or multi-word items. What needs
to be acquired for a lexical item includes formal, morphological,
semantic and syntactic information. Once acquired, this knowledge
can be accessed consciously as declarative knowledge or otherwise as
procedural knowledge where the retrieving process is fast and
automatic. How well a word is known is assessed from four angles,
depth and breadth, on the one hand, which determine to what extent
the word is known, and reception and production, on the other, which
determine the degree to which the word can be retrieved for language
use. We saw too that the acquisition of L2 vocabulary is an
incremental process. First, the word form needs to be mapped on to
the word meaning. Once this basic link is established, however
imperfectly, more information can be progressively added and the
item integrated with other words through subsequent learning and
exposure. Even though much modification is made to the initial
meaning-form mapping, it is necessary to ensure that this first stage is
successful. In the subsequent build-up of knowledge, listening and
cultural awareness emerge as key factors. In the first instance,
perception, pronunciation and long-term retention are reinforced; in
the second, the lexical confusions and errors caused by interlingual
factors are reduced.
Vocabulary acquisition has been regarded as both implicit and
explicit, the focus being either on the meaning or the form of the new
items. While explicit learning is more likely to make the meaning-
form link, implicit learning provides opportunities for incorporating
and consolidating words within the mental lexicon. Successful
vocabulary acquisition programs consist in integrating explicit
learning with implicit learning; initially learning is explicit whereas
subsequently the two are complementary and can both serve to pro-
mote later learning. Although vocabulary learning is often perceived

293
and spoken of in terms of one-word items, the quantity of multi-word
items is much larger. Furthermore, all that has been said in respect of
breadth and depth, reception and production, meaning-form mapping,
listening, cultural awareness, and explicit/implicit learning, is
applicable to these items, which often constitute more of a learning
challenge. Although some approaches have been developed in respect
of some such items, in particular idioms, notably imagery and
mnemonic techniques, there is a need for suitable approaches to
promote the learning of other types of multi-word items.
To return to explicit and implicit learning: to date these
approaches have been primarily conceptualised in terms of language
learning programmes or activities. These external factors do not take
into account individual learning differences. Internal factors should
also be considered in order to improve learning efficiency. Learners
need to be aware of their learning styles, which will influence the
learning strategies they employ. Developing learning strategies is the
most efficient way to facilitate autonomous learning, particularly in
terms of implicit learning where learners are supposed to read
extensively on their own. If learning strategies are used appropriately,
they will help learners to learn vocabulary more efficiently, which will
in turn increase their motivation, a crucial factor in determining the L2
competence ultimately attained. This is what I called a learner-focused
approach, summarised in the “learner-focused model of SLVA” (see
Chapter 7: Figure 7).
On the basis of the knowledge acquired in respect of vocabulary
acquisition and CALL, I have attempted to construct a comprehensive
framework to conceptualise CAVL so that it can be developed into a
legitimate sub-domain of CALL. In line with the development of
CALL, CAVL has graduated from the simple, drill and practice type
of tutor to the current integrated tutor and tool stage where both
intelligent tutors, which take into account efficient vocabulary
acquisition processes, and CMC tools, which emphasise collaborative
learning and social interaction, combine and complement each other.
The CAVL applications that can be generally classified as tutor and
tool fall into two broad categories: lexical programs/tasks and lexical
resources/aids, the former comprising primary learning activities, the
latter supplying additional lexical information. The lexical programs/

294
tasks can be further sub-divided into four categories that range from
tool/implicit/meaning-focused to tutor/explicit/form-focused. While
dedicated tutorial CAVL programs are particularly geared to
vocabulary learning in terms of pedagogical efficiency, other types of
program can also contribute to vocabulary learning in a variety of
ways if appropriately used in combination with each other or simply
with other language learning activities.
The efficiency of all types of programs has been evaluated
using the CAVL evaluation framework that I constructed; this takes
into account both the vocabulary learning process and the learning
outcome, both the direct learning benefit and the indirect affective
impact. We see that the success of a given CAVL program largely de-
pends on the quality of the vocabulary processing, i.e. how successful
the initial lexical meaning and form mapping is. Given that all types
of CAVL programs can be conducive to vocabulary learning to
varying degrees, it seems more sensible to integrate them systematically
into the vocabulary learning curriculum than to favour one or more to
the exclusion of others. This would be a further goal for CAVL.
In line with these differences between individual learners
(discussed in Chapter 7), I drew attention to those that exist between
learners of different cultures. Since the focus of my interest has been
on Chinese students learning English in China, it was important to
determine the salient characteristics. One, which is obviously a barrier
to attaining oral, interactive, communicative competence, is the
prevalent view, held by teachers and students, that language is a type
of knowledge just as with other subjects that are studied. Linguistic
knowledge is thus passed from the giver, the teacher, to the students,
the receivers, who are supposed to retain it by memorisation. So it is
with vocabulary and, given such a conceptualisation of learning, it is
not uncommon to come across students who have learned whole
dictionaries or lists of several thousand words by heart but who cannot
communicate in English.
My empirical investigation into the current situation has shown
that both teachers and students hold the view that vocabulary learning
should be explicit and that it should be achieved primarily through
individual learning effort. Students typically rely on mechanical
means to memorise vocabulary items, often focusing on mapping the

295
L2 word form on to the Chinese translation while ignoring other
memory strategies that require more mental effort. Teachers, on the
other hand, do not incorporate strategy instruction systematically into
their language courses. They typically recommend that the students
read extensively to enlarge and consolidate vocabulary items, which
the majority of students apparently fail to do as they lack the
necessary means or strategies to cope with the unfamiliar reading
context. Fundamentally, this specifically Chinese approach to
vocabulary learning can be attributed to the Chinese culture of learn-
ing, the way the L1, Chinese, is learned. It seems that the prerequisite
to improve the situation would be to give systematic strategy
instruction to learners.

296
References

Adamson, B. / Morris, P. 1997. The English curriculum in the People’s

Republic of China. Comparative Education Review. 41, 3–26.
Adolphs, S. / Durow, V. 2004. Social-cultural integration and the
development of formulaic sequences. In N. Schmitt (ed.) For-
mulaic sequences: acquisition, processing and use. Amsterdam,
Philadelphia: John Benjamins, 107–126.
Aitchison, J. 1994. Words in the Mind: An Introduction to the Mental
Lexicon. Oxford: Blackwell.
Alexander, L. G. 1967. New Concept English. London: Longman.
Alexander, L. G. 1980. Mainline Progress A and B. London: Longman.
Allum, P. 2004. Evaluation of CALL: Initial vocabulary learning.
ReCALL. 16/2, 488–501.
Al-Seghayer, K. 2001. The effect of multimedia annotation modes on
SLVA: A comparative study. Language Learning & Technology.
5/1, 202–232.
Ames, W. S. 1966. The development of a classification scheme of
contextual aids. Reading Research Quarterly. 11/1, 57–82.
Anderson, J. R. 1983. The architecture of cognition. Cambridge Mass.:
Harvard University Press.
Anderson, R. C. / Freebody, P. 1981. Vocabulary knowledge. In J.T.
Guthrie (ed.) Comprehension and teaching: Research reviews.
Newark, DE: International Reading Association, 77–117.
Asher, J. J. 1964. Towards a neo-field theory of behaviour. Journal of
humanistic psychology. 4, 85–94.
Asher, J. J. 1972. Children’s first language as a model for second lan-
guage learning. The Modern Language Journal. 56, 133–139.
Asher, J. J. 1983. Learning another language through actions. Los
Gatos, California: Sky Oaks Productions, Inc.
Asher, J. J. / Kusudo, J. / de la Torre, R. 1974. Learning a second
language through commands: The second field test. The Modern
Language Journal. 58/1–2, 24–32.

297
Atkinson, R.C. / Raugh, M. R. 1975. An application of the mnemonic
keyword method to the acquisition of a Russian vocabulary.
Journal of Experimental Psychology: Human Learning and
Memory. 104/2, 126–133.
Barcroft, J. 2002. Semantic and structural elaboration in L2 lexical
acquisition. Language Learning. 52/2, 323–363.
Barcroft, J. 2004. Second language vocabulary acquisition: A lexical
input processing approach. Foreign Language Annals. 37/2,
200–208.
Barlow, M. 2004. Collocate 1.0. Houston: Athelstan.
Beck, I. L. / McKeown, M. G. / McCaslin, E. S. 1983. Vocabulary de-
velopment: All contexts are not created equal. The Elementary
School Journal. 83/3, 177–181.
Beheydt, L. 1987. The semantization of vocabulary in foreign lan-
guage learning. System. 15, 55−67.
Beheydt, L. 1990. CALL and vocabulary acquisition in Dutch. In P. J.
Kingston, C. Zähner, A. Beutner (eds) Languages, continuity,
opportunity. London: CILT, 186–192.
Bialystok, E. 1995. Towards an explanation of second language
acquisition. In B. Gillian et al. (eds), Language and Under-
standing. Oxford: Oxford University Press, 115–138.
Bialystok, E. / Sharwood Smith, M. 1985. Interlanguage is not a state
of mind: An evaluation of the construct for second-language
acquisition. Applied Linguistics. 6, 101–117.
Boers, F. 2000. Metaphor awareness and vocabulary retention.
Applied Linguistics. 21/4, 553–571.
Boers, F. 2004. Expanding learner’s vocabulary through metaphor
awareness: What expansion, what learners, what vocabulary. In
M. Achard, S. Niemeier (eds) Cognitive linguistics, second
language acquisition, and foreign language teaching. Berlin,
New York: Mouton de Gruyter, 211–232.
Boers, F. / Demecheleer, M. 2001. Measuring the impact of cross-
cultural differences on learners’ comprehension of imageable
idioms. ELT Journal. 55/3, 255–262.
Boers, F. / Demecheleer, M. / Eyckmans, J. 2004. Etymological
elaboration as a strategy for learning figurative idioms. In P.
Bogaards, B. Laufer (eds) Vocabulary in a second language,

298
 selection, acquisition and testing. Amsterdam: John Benjamins,
53–78.
Boers, F. / Lindstromberg, S. 2005. Finding ways to make phrase-
learning feasible: the mnemonic effect of alliteration. System.
33, 225–238.
Bouziane, A. 2006. Review of LinC (Language Interactive Culture)
English-beginners. CALICO Journal. 23/3, 219–230.
The British Council. 1995. English language teaching profile of the
People’s Republic of China. Beijing: The British Council.
Carroll, J. B. / Davies, P. / Richman, B. 1971. The American heritage
word frequency book. New York: American Heritage Publishing
Co., Inc.
Chang, J. 2001. Chinese speakers. In M. Swan, B. Smith (eds)
Learner English: A teacher’s guide to interference and other
problems (second edition). Cambridge: Cambridge University
Press, 310–324.
Chapelle, C. 1998a. Construct definition and validity inquiry in SLA
research. In L. F. Bachman, A. D. Cohen (eds) Interface
between second language acquisition and language testing
research. Cambridge: Cambridge University Press, 32–70.
Chapelle, C. 2001. Computer applications in second language acqui-
sition. Cambridge: Cambridge University Press.
Cheng, C. 2004. Theory and practice of data-driven learning. In A. He,
L. Wong. / M. Xu, X. Cheng (eds) Application of corpora to
foreign language education: Theory and Practice. Guangzhou:
Guangdong Higher Education Press, 95–104.
Chiswick, B. / Miller, P. 2004. Linguistic distance: a quantitative
measure of the distance between English and other Languages.
Retrieved July 10, 2005, from: <ftp://repec.iza.org/RePEc/
Discussionpaper/dp1246.pdf>.
Chun, D. M. / Plass, J. J. 1996. Effects of multimedia annotations on
vocabulary acquisition. The Modern Language Journal. 80/2,
183–198.
Coady, J. 1993. Research on ESL/EFL vocabulary acquisition: Putting
it in context. In T. Huckin, M. Haynes, J. Coady (eds) Second
language reading and vocabulary learning. Norwood, NJ:
Ablex Publishing, 3–23.

299
Coady, J. 1997. SLVA: A synthesis of research. In J. Coady, T. Huckin
(eds) Second language vocabulary acquisition. Cambridge:
Cambridge University Press, 273–290.
Coady, J. / Magoto, J. / Hubbard, P. / Graney, J. / Mokhtari, K. 1993.
High frequency vocabulary and reading proficiency in ESL
readers. In T. Huckin, M. Haynes, J. Coady (eds) Second lan-
guage reading and vocabulary learning. Norwood, NJ: Ablex
Publishing, 217–227.
Cobb, T. 1997. Is there measurable learning from hands-on concor-
dancing? System. 25/3, 301–315.
Cobb, T. / Stevens, V. 1996. A principled consideration of computers
and reading in a second language. In M. C. Pennington (ed.)
The power of CALL. Houston, TX: Athelstan, 115–135.
Cohen, A. D. 1987. The use of verbal and imagery mnemonics in
second-language vocabulary learning. Studies in Second Lan-
guage Acquisition. 9, 43–62.
Cohen, A. D. 1998. Strategies in learning and using a second lan-
guage. London: Longman.
Cohen, A. D. 2003. The learner’s side of foreign language learning:
where do styles, strategies, and tasks meet? IRAL. 41, 279–291.
Cohen, A. D. / Chi, J. C. 2003. Language Strategy Use Inventory and
Index. Retrieved: January 10, 2005 from <http://www.carla.umn.
edu/maxsa/samples/SG_LanguageSurvey.pdf>.
Cohen, A. D. / Dörnyei, Z. 2002. Focus on the language learner:
motivation, styles and strategies. In N. Schmitt (ed.) An
introduction to applied linguistics. London: Arnold, 170–190.
Cohen, A. D. / Oxford, R. L. / Chi, J. C. 2005. Learning style survey:
assessing your own learning styles. Retrieved January 10 2005
from <http://www.carla.umn.edu/about/profiles/CohenPapers/
LearningStylesSurvey.pdf>.
Cohen, A. D. / Paige, R. M. / Kappler, B. / Demmessie, M. / Weaver, S.
J. / Chi, J. C. / Lassegard, J. P. 2003. Maximizing study abroad:
A language instructor’s guide to strategies for language and
culture learning and use. Minneapolis, MN: Center for Ad-
vanced Research on Language Acquisition (CARLA).
Colpaert, J. 2004. Design of online interactive language courseware:
conceptualisation, specification and prototyping. Research into

300
 the impact of linguistic-didactic functionality on software archi-
tecture. Unpublished PhD thesis, University of Antwerp,
Belgium. Retrieved June 8, 2005, from <http://www.didascalia.
be/doc-design.pdf>.
Condon, N. / Kelly, P. 2002. Does cognitive linguistics have anything
to offer English language learners in their efforts to master
phrasal verbs? ITL Review of Applied Linguistics. 137–138,
205–231.
Confucius. 1996. The analects. Translated by A. Waley. Ware: Words-
worth.
Cortazzi, M. / Jin, L. 1996a. English teaching and learning in China.
Language Teaching, 29, 61–80.
Cortazzi, M. / Jin, L. X. 1996b. Cultures of learning: Language class-
rooms in China. In H. Coleman (ed.) Society and the language
classroom. Cambridge: Cambridge University Press, 169–206.
Cowie, A. P. 1998. Introduction. In Cowie A. P. (ed.) Phraseology:
Theory, analysis and applications. Oxford: Oxford University
Press.
Craik, F. I. M. / Lockhart, R. S. 1972. Levels of processing: A
framework for memory research. Journal of Verbal Learning
and Verbal Behavior. 11, 671–684.
Craik, F. I. M. / Tulving, E. 1975. Depth of processing and the
retention of words in episodic memory. Journal of Experimental
Psychology: General. 104/3, 268–294.
Crystal, D. 1987. The Cambridge encyclopaedia of language. Cam-
bridge: Cambridge University Press.
Dagneaux, E. / Denness, S. / Granger, S. 1998. Computer-aided error
analysis. System. 26, 163–174.
Day, R. / Omura, C. / Hiramatsu, M. 1991. Incidental EFL vocabulary
learning and reading. Reading in a Foreign Language, 7/2,
541–551.
De Bot, K. 1992. Bilingual production model: Levelt’s speaking
model adapted. Applied Linguistics. 13/1, 1–25.
De Bot, K. / Paribakht, T. S. / Wesche, M. B. 1997. Towards a lexical
processing model for the study of second language vocabulary
acquisition: Evidence from ESL reading. Studies in Second Lan-
guage Acquisition. 19/3, 309–329.

301
De Cock, S. 1998. A recurrent word combination approach to the study
of formulae in the speech of native and non-native speakers of
English. International Journal of Corpus Linguistics. 3/1, 59–80.
De la Fuente, M. 2003. Is SLA interactionist theory relevant to CALL?
A study on the effects of computer-mediated interaction in
SLVA. Computer Assisted Language Learning. 16/1, 47–81.
De Ridder, I. 2002. Visible or invisible links: does the highlighting of
hyperlinks affect incidental vocabulary learning, text compre-
hension, and the reading process? Language Learning &
Technology. 6/1, 123–146.
De Saussure, F. 1974. Course in General Linguistics. London: Fon-
tana, Collins.
Decoo, W. 1994. In defence of drill and practice in CALL: A reevalu-
ation of fundamental strategies. Computers & Education.
23/
Decoo, W. / Heughebaert, E. / Schonenbert, N. / Van Elsen, D. 1996.
The standard vocabulary programs of Didascalia: In search of
external versatility and didactic optimisation. Computer
Assisted Language Learning. 9/4, 319338.
Dekeyser, R. 2003. Implicit and explicit learning. In J. Doughty, M. L.
Long (eds) The handbook of second language acquisition.
Oxford: Blackwell Publishing Ltd, 313–348.
Desmarais, L. / Duquette, L. / Renié, D. / Laurier, M. 1997. Evaluat-
ing learning and interaction in a multimedia environment.
Computers and the Humanities. 31, 327–349.
Dörnyei, Z. 1994. Motivation and motivating in the foreign language
classroom. The Modern Language Journal. 78, 525–523.
Dörnyei, Z. 1998. Motivation in second and foreign language learning.
Language Teaching. 31, 117–135.
Dörnyei, Z. 2001. Teaching and researching motivation. Harlow:
Pearson Education Limited.
Dörnyei, Z. 2002. The motivational basis of language learning tasks.
In P. Robinson (ed.) Individual differences and instructed
language learning. Amsterdam/Philadelphia: John Benjamins,
137–117.

302
Dörnyei, Z. / Ottó, I. 1998. Motivation in action: A process model of
L2 motivation. Working Papers in Applied Linguistics (Thames
Valley University, London). 4, 4369.
Dörnyei, Z. / Durow, V. / Zahran, K. 2004. In N. Schmitt (ed.)
Formulaic sequences: acquisition, processing and use. Amster-
dam, Philadelphia: John Benjamins, 87–106.
Drum, P. A. / Konopak, B. 1987. Learning word meanings from
written contexts. In M. G. McKeown, M. E. Curtis (eds) The
nature of vocabulary acquisition. Hillsdale, NJ: Lawrence Erl-
baum, 73–87.
Dunn, L. M. / Dunn, L. M. 1981. PPVT: Peabody picture vocabulary
test. Revised version. Circle Pines, MN: American Guidance
Service.
Dunn, R. / Griggs, S. 1988. Learning styles: Quiet revolution in
American secondary schools. Reston, VA: National Association
of Secondary School Principals.
Duquette, L. / Renié, D. / Laurier, M. 1998. The evaluation of
vocabulary acquisition when learning French as a second
language in a multimedia environment. Computer Assisted
Language Learning, 11/1, 3–34.
Ebbrell, D. / Goodfellow, R. 1996. Learner, teacher and computer – A
mutual support system. In New Horizons in CALL, Proceedings
of Eurocall 96. Szombatheley. Hungary: Berzsenyi Daniel
College.
Ehrman, M. E. 1996. Understanding second language learning
difficulties. Thousand Oaks, CA: Sage.
Ehrman, M. E. / Oxford, R. L. 1989. Effects of sex differences, career
choice, and psychological type on adult language learning
strategies. The Modern Language Journal. 73, 1–13.
Ehrman, M. E. / Oxford, R. L. 1990. Adult language learning styles
and strategies in an intensive training setting. The Modern
Language Journal. 74/3, 311–327.
Ehrman, M. E. / Leaver, B. L. / Oxford, R. L. 2003. A brief overview
of individual differences in second language learning. System.
31, 313–330.
Ellis, N. (ed.) 1994a. Implicit and Explicit learning of language. Lon-
don, San Diego: Academic Press.

303
Ellis, N. 1994b. Vocabulary acquisition: The implicit ins and outs of
explicit cognitive mediation. In N. Ellis (ed.) Implicit and
Explicit learning of language. London, San Diego: Academic
Press, 211–282.
Ellis, N. 1995a. Vocabulary acquisition: psychological perspectives.
The Language Teacher. 19/2, 12–16.
Ellis, N. 1995b. The psychology of foreign language vocabulary
acquisition: Implications for CALL. Computer Assisted Vocabu-
lary Learning. 8/23, 103–128.
Ellis, N. 2001. Memory for language. In P. Robinson (ed.) Cognition
and second language instruction. Cambridge: Cambridge Uni-
versity Press, 33–68.
Ellis, N. 2002. Frequency effects in language acquisition: A review with
implications for theories of implicit and explicit language acqui-
sition. Studies in Second Language Acquisition. 24, 143–188.
Ellis, R. / He, X. 1999. The roles of modified input and output in the
incidental acquisition of word meanings. Studies in Second
Language Acquisition. 21/2, 285–301.
Ervin-Tripp, S. 1974. Is second language learning like the first?
TESOL Quarterly. 8, 111–127.
Eyckmans, J. / Beeckmans, R. / Van de Velde, H. 2002. The paired
vocabulary test. Paper presented at the Conference of Second
Language Vocabulary Colloquium, Leiden, Holland.
Feyten, C. M. 1991. The power of listening ability: an overlooked
dimension in language acquisition. The Modern Language
Journal. 75, 173–180.
Francis, W. N. / Kučera, H. 1982. Frequency analysis of English usage:
Lexicon and grammar. Boston: Houghton Mifflin.
Gan, Z. / Humphreys, G. / Hamp-Lyons, L. 2004. Understanding
successful and unsuccessful EFL students in Chinese Uni-
versities. The Modern Language Journal. 88/2, 229–244.
Gardner, R. C. 1985. Social psychology and second language learning:
the role of attitudes and motivation. London: Edward Arnold.
Gary, J. O. 1975. Delayed oral practice in initial stages of second lan-
guage learning. In M. Burt, H. Dulay (eds) New directions in
second language teaching, learning and bilingual education.
Washington, D. C.: TESOL, 89–95.

304
Gary, N. / Gary, J. O. 1982. Packaging comprehension materials: to-
wards effective language instruction in difficult circumstances.
System. 10/1, 61–69.
Gass, S. / Selinker, L. 2001. Second Language Acquisition. Mahwah,
New Jersey, London: Lawrence Erlbaum Associates.
Gass, S. 1988. Second language vocabulary acquisition. Annual
Review of Applied Linguistics. 9, 92–106.
Goodfellow, R. 1994. A computer-based strategy for foreign-language
vocabulary learning. Unpublished PhD thesis, Open University,
UK.
Goodfellow, R. 1995. A review of the types of CALL programmes for
vocabulary instruction. Computer Assisted Language Learning.
2–3, 205–226.
Goodfellow, R. 1999. Evaluating performance, approach and outcome.
In K.Cameron (ed.) CALL: Media, design, and applications.
Lisse: The Netherlands: Swets & Zeitlinger, 109–140.
Goulden, R. / Nation, P. / Read, J. 1990. How large can a receptive vo-
cabulary be? Applied Linguistics. 11/4, 341–363.
Granger, S. (ed.) 1998a. Learner English on Computer. London/New
York: Addison Wesley Longman.
Granger, S. 1998b. Prefabricated patterns in advanced EFL writing:
collocations and lexical phrases. In A. P. Cowie (ed.) Phrase-
ology: theory, analysis and applications. Oxford: Clarendon
Press/Oxford University Press, 145–160.
Granger, S. 2002. A Bird’s-eye view of learner corpus research. In S.
Granger, J. Hung, S. Petch-Tyson (eds) Computer Learner
Corpora, Second Language Acquisition and Foreign Language
Teaching. Amsterdam/Philadelphia: John Benjamins, 3–33.
Granger, S. / Hung J. / Petch-Tyson S. (eds) 2002. Computer Learner
Corpora, Second Language Acquisition and Foreign Language
Teaching. Amsterdam/Philadelphia: John Benjamins.
Granger, S. / Tribble, C. 1998. Learner corpus data in the foreign
language classroom: form-focused instruction and data-driven
learning. In S. Granger (ed.) Learner English on Computer.
London/New York: Addison Wesley Longman, 199–211.

305
Granger, S. / Tyson, S. 1996. Connector usage in the English essay
writing of native and non-native EFL speakers of English.
World Englishes. 15/1, 1727.
Groot, P. J. M. 2000. Computer assisted second language vocabulary
acquisition. Language Learning & Technology. 4/1, 60–81.
Gu, Y. / Johnson, R. K. 1996. Vocabulary learning strategies and lan-
guage learning outcomes. Language Learning. 46/4, 643–679.
Gu, Y. / Leung, C. B. 2002. Error patterns of vocabulary recognition
for EFL learners in Beijing and Hong Kong. Asian Journal of
English Language Teaching. 12, 121–141.
Gui, S. / Yang, H. 2002. Chinese English Learner Corpora. Shanghai:
Shanghai Foreign Language Education Press.
Hammerly, H. 1984. Contextualized-visual aids (filmstrips) as con-
veyors of sentence meaning. International Review of Applied
Linguistics. 22/2, 87–94.
Hansen, C. 1991. Language in the Heart-Mind. In R. Allinson (ed.)
Understanding the Chinese Mind. Hong Kong: Oxford Univer-
sity Press, 75–124.
Harley, B. 1996. Vocabulary learning and teaching in a second
language. Canadian Modern Language Review. 53, 3–12.
Harrington, M. / Levy, M. 2001. CALL begins with a ‘C’: Interaction
in computer-mediated language learning. System. 29, 15–26.
He, A. 2004. Analysis of English spelling errors of Chinese students.
In A. He, L. Wong. / M. Xu, X. Cheng (eds) Application of
corpora to foreign language education: Theory and Practice.
Guangzhou: Guangdong Higher Education Press, 288–298.
Hémard, D. 2004. Enhancing online CALL design: The case for
evaluation. ReCALL. 16/2, 502–529.
Henriksen, B. 1999. Three dimensions of vocabulary development.
Studies in Second Language Acquisition. 21, 303–317.
Hirsh, D. Nation, I. S. P. 1992. What vocabulary is needed to read
unsimplified texts for pleasure? Reading in a Foreign Language.
8, 689–696.
Hoijer, H. 1994. The Sapir-Whorf hypothesis. In L, Samovar, R.E.
Porter (eds) Intercultural communication: A reader. Belmont,
CA: Wardsworth, 194–200.

306
Horst, M. / Cobb, T. / Nicolae, I. 2005. Expanding academic vocabu-
lary with an interactive on-line database. Language Learning &
Technology. 9/2, 90–110.
Howarth, P. 1998. Phraseology and second language proficiency.
Applied Linguistics, 19/1, 24–44.
Hu, G. 2002. Potential cultural resistance to pedagogical imports: The
case of communicative language teaching in China. Language,
Culture and Curriculum. 15/2, 93–105.
Huang, Y. / Xu, H. 1999. Trends in English Language Education in
China. ESL Magazine. 2, 26–28.
Hubbard, P. 2001. The use and abuse of meaning technologies.
Contact. 27, 82−86. Retrieved February 5, 2006, from <http://
www.stanford.edu/~efs/phil/MT.pdf>.
Hubbard, P. 2003. A survey of unanswered questions in CALL.
Computer Assisted Language Learning. 16/2−3, 141–154.
Hubbard, P. 2004. Learner training for effective use of CALL. In S.
Fotos, C. Browne (eds) New perspectives on CALL for second
language classrooms. Mahwah, New Jersey, London: Lawrence
Erlbaum Associates, Publishers, 45–67.
Hubbard, P. / Siskin, C. B. 2004. Another look at tutorial CALL.
ReCALL. 16/2, 448–461.
Huckin, T. / Coady, J. 1999. Incidental vocabulary acquisition in a
second language. Studies in Second Language Acquisition. 21/2,
181–193.
Huitt, W. 2001. Motivation to learn: An overview. Educational
Psychology Interactive. Valdosta, GA: Valdosta State University.
Retrieved July 1 2005, from <http://chiron.valdosta.edu/whuitt/
col/motivation/motivate.html>.
Hulstijn. J. 1992. Retention of inferred and given word vocabulary
learning. In P. J. Arnaud and H. Béjoint (eds), Vocabulary and
applied linguistics. London: Macmillan, 113–125.
Hulstijn, J. 2001. Intentional and incidental learning: a reappraisal of
elaboration, rehearsal and automaticity. In P. Robinson (ed.)
Cognition and second language instruction. Cambridge: Cam-
bridge University Press, 258–286.

307
Hulstijn, J. 2003. Incidental learning and intentional learning. In J.
Doughty, M. L. Long (eds) The handbook of second language
acquisition. Blackwell Publishing Ltd, 349–381.
Hulstijn, J. / Hollander, M. / Greidanus, T. 1996. Incidental learning
by advanced foreign language students: the influence of
marginal glosses, dictionary use, and reoccurrence of unknown
words. The Modern Language Journal. 80, 327–339.
Jiang, N. 2000. Lexical representation and development in a second
language. Applied Linguistics. 21/1, 47–77.
Jiang, N. 2002. Form-meaning mapping in vocabulary acquisition in a
second language. Studies in Second Language Acquisition. 24,
617–637.
Jiang, N. 2004. Morphological insensitivity in second language
processing. Applied Psycholinguistics. 25, 603–634.
Johns, T. 1991a. Should you be persuaded – Two examples of Data-
Driven Learning, English Language Research Journal. 4, 1–13.
Johns, T. 1991b. From printout to handout. Grammar and vocabulary
teaching in the context of data-driven learning. English Lan-
guage Research Journal. 4, 27–45.
Johns, T. / King, D. (eds). 1991. Classroom concordancing. English
Language Research (new series), 4.
Jones, L. C. / Plass, J. J. 2002. Supporting listening comprehension
and vocabulary acquisition in French with multimedia annota-
tions. The Modern Language Journal. 86/4, 546–561.
Jones, M. / Haywood, S. 2004. Facilitating the acquisition of for-
mulaic sequences: An exploratory study in an EAP context. In
N. Schmitt (ed.) Formulaic sequences: acquisition, processing
and use. Amsterdam, Philadelphia: John Benjamins, 269–300.
Judd, E. L. 1978. Vocabulary teaching and TESOL: A need for re-
evaluation of existing assumptions. TESOL Quarterly. 12/1,
71–76.
Jung, U. 1988. Introduction, or a ‘Petit Tour’ of the World of CALL,
in U. Jung (ed.) Computers in applied linguistics and language
teaching (pp. 1−20). Frankfurt: Peter Lang.
Kelen, C. 2002. Language and learning orthodoxy in the English
classroom in China. Educational Philosophy and Theory. 34/2,
223–237.

308
Keller, H. H. 1978. New perspective on teaching vocabulary. Lan-
guage in Education: Theory and practice. Arlington, Virginia:
Center for Applied Linguistics.
Kelly, P. 1980. La comprehension à l’audition. Memoire présenté à
l’Université Catholique de Louvain pour avoir l’équivalence
avec une licence en philologie romane de l’Université de
Londres.
Kelly, P. 1985. A dual approach to FL vocabulary learning: the con-
joining of listening comprehension and mnemonic practices.
Unpublished doctoral thesis, Catholic University of Louvain,
Belgium.
Kelly, P. 1986. Solving the vocabulary retention problem. ITL. 74,
1–16.
Kelly. P. 1989. A particular application of the RALEX method of
foreign language learning. Le Langage et l’Homme. 24/70,
153–160.
Kelly, P. 1990. Guessing: no substitute for systematic learning of lexis.
System. 18, 199–207.
Kelly, P. 1992. Does the ear assist the eye in the long-term retention of
lexis? International Review of Applied Linguistics. 30/2,
137–145.
Kelly, P. 1995. The importance of listening − a personally illustrated
view. Zielsprache English. 25, 13–19.
Kelly, P. / Li, X. 2005. A new approach to learning English
vocabulary: More efficient, more effective, and more enjoyable.
Beijing: Foreign Language Teaching and Research Press.
Kelly, P. / Li, X. / Vanparys, J. / Zimmer, C. 1996. A comparison of the
perceptions and practices of Chinese and French-speaking
Belgian university students in the learning of English: The
prelude to an improved programme of lexical expansion. ITL.
113–114, 275–303.
Knight, S. 1994. Dictionary: the tool of last resort in foreign language
reading? A new perspective. The Modern Language Journal. 78,
285299.
Kolers. P. A. 1963. Interlingual word associations. Journal of Verbal
Learning and Verbal Behavior. 2, 251–300.

309
Krashen, S. 1987. Principles and practices in second language
acquisition. Englewood Cliffs, NJ: Prentice Hall.
Krashen, S. 1989. We acquire vocabulary and spelling by reading:
Additional evidence for the input hypothesis. The Modern
Language Journal. 73/4, 440–464.
Krashen, S. 1996. The case for narrow listening. System. 24, 97–100.
Krashen, S. / Terrell, T. 1983. The natural approach: Language
acquisition in the classroom. Oxford: Pergamon Press.
Kupelian, M. 2001. Korean EFL students’ acquisition of culturally
loaded words. The Korea TESOL Journal. 4/1, 17–36.
Labrie, G. 2000. A French vocabulary tutor for the web. CALICO
Journal. 17/3, 475–499.
Lado, R. 1964. Language teaching: a scientific approach. New York:
McGraw-Hill.
Lam, A. 2002. English in education in China: Policy changes and
learners’ experiences. World Englishes. 21/2, 245–256.
Lantolf, J. P. 1999. Second culture acquisition: Cognitive consider-
ations. In E. Hinkel (ed.) Culture in second language teaching
and learning. Cambridge: Cambridge University Press, 28–46.
Laufer, B. 1990. “Sequence” and “order” in the development of L2
lexis: Some evidence from lexical confusions. Applied
Linguistics. 11/3, 281–296.
Laufer, B. 1997. The lexical plight in second language reading: Words
you don’t know, words you think you know and words you
can’t guess. In J. Coady, T. Huckin (eds) Second Language
Vocabulary Acquisition. Cambridge: Cambridge University
Press, 20–34.
Laufer, B. 1998. The development of passive and active vocabulary in
a second language: same or different? Applied Linguistics. 19/2,
255–271.
Laufer, B. / Elder, C. / Hill, K, Congdon, P. 2004. Size and strength:
do we need both to measure vocabulary knowledge? Language
Testing, 21/2, 202–226.
Laufer, B. / Hill, M. 2000. What lexical information do L2 learners
select in a CALL dictionary and how does it affect word
retention? Language Learning & Technology. 3/2, 58–76.

310
Laufer, B, Nation, I. S. P. 1995. Vocabulary size and use: Lexical
richness in L2 written production. Applied Linguistics. 16,
307–322.
Laufer, B. / Paribakht, T. S. 1998. The relationship between passive
and active vocabularies: Effects of language learning context.
Language Learning. 48/3, 365–391.
Levelt, W. J. M. 1989. Speaking: from intention to articulation. Cam-
bridge, Massachusetts: MIT press.
Levelt, W. J. M. 1993. Language use in normal speakers and its
disorders. In G. Blanken, E. Dittman, H. Grimm, J. Marshall, C.
Wallesch (eds) Linguistic disorders and pathologies. An
international handbook. Berlin: de Gruyter, 1–15.
Levelt, W. J. M. / Roelofs, A. / Meyer, A. S. 1999. A theory of lexical
access in speech production. Behavioral and Brain Sciences. 22,
1−75.
Levy, M. 1997. Computer-assisted language learning: Context and
conceptualisation. Oxford: Clarendon Press.
Lewis, M. 1993. The lexical approach. Hove: Language Teaching
Publications.
Lewis, M. 1997. Pedagogical implications of the lexical approach. In J.
Coady, T. Huckin (eds) Second language vocabulary
acquisition. Cambridge: Cambridge University Press, 255–270.
Lewis, M. (ed.) 2000a. Teaching collocation: further development in
the lexical approach. Hove: Language Teaching Publications.
Lewis, M. 2000b. Learning in the lexical approach. In M. Lewis (ed.)
Teaching collocation: further development in the lexical
approach. Hove: Language Teaching Publications, 155–185.
Li, X. 1984. In defence of the communicative approach. ELT Journal.
38/1, 2–13.
Li, X. (ed.) 1987. Communicative English for Chinese Learners.
Shanghai: Shanghai Foreign Language Education Press.
Li, X. / Song, X. / Zimmer, C. / Vanparys, J. / Kelly, P. 1999. WUFUN:
A new approach to more efficient and effective vocabulary
learning. ITL. 125–126, 181–194.
Liao, X. 2000. Communicative Language Teaching Innovation in
China: Difficulties and Solutions. ERIC Document ED443294.

311
Liu, D. / Zhong, S. 1999. Acquisition of culturally loaded words in
EFL. Foreign Language Annals. 32/2, 177–188.
Long, M. 1996. The role of the linguistic environment in second lan-
guage acquisition. In W. Ritchie, T. Bhatia (eds) Handbook of
second language acquisition. Rowley, MA: Newbury House,
413–468.
Lozanov, G. 1979. Suggestology and outlines of Suggestopedy. New
York: Gordon and Breach.
Meara, P. 1990. A note on passive vocabulary. Second Language
Research. 6/2: 150–154.
Meara, P. 1996a. The vocabulary knowledge framework. Vocabulary
Acquisition Research Group Virtual Library. Retrieved May 15,
2005, from <http://www.swansea.ac.uk/cals/calsres/vlibrary/
pm96d.htm>.
Meara, P. 1996b. Word Associations in a Foreign Language. Vocabu-
lary Acquisition Research Group Virtual Library. Retrieved
May 15, 2005, from <http://www.swan.ac.uk/cals/calsres/
vlibrary/pm83a.htm>.
Meara, P. / Buxton, B. 1987. An alternative to multiple choice
vocabulary tests. Language Testin. 4, 142–151.
Mel’čuk, I. 1998. Collocations and lexical functions. In A. P. Cowie
(ed.) Phraseology: Theory, analysis and applications. Oxford:
Clarendon Press, 23–53.
Melka, F. 1982. Receptive versus Productive Vocabulary: A survey.
Interlanguage Studies Bulletin. 6/2, 5–33.
Melka, F. 1997. Receptive vs. productive aspects of vocabulary. In N.
Schmitt, M. McCarthy (eds) Vocabulary: Description,
acquisition and pedagogy (pp. 84–102). Cambridge: Cambridge
University Press.
Miller, G. A. 1990. WordNet: An On-line Lexical Database. Inter-
national Journal of Lexicography. 3/4, 235–312.
Miller, G. A. 1996. The Science of Words. New York: Scientific
American Library.
Miller, G. 1999. On knowing a word. Annual Review of Psychology.
50, 1–19.
Miller, G. A. / Fellbaum, C. 1991. Semantic networks of English.
Cognition. 41, 197–229.

312
Mitchell, R. / Myles, F. 1998. Second Language Learning Theories.
London: Arnold.
Mondria, J.-A. / Mondria-de Vries, S. 1994. Efficiently memorizing
words with the help of word cards and ‘hand computer’: Theory
and applications. System. 22, 47–57.
Mondria, J.-A. / Wiersma, B. 2004. Receptive, productive, and re-
ceptive + productive L2 vocabulary learning: what difference
does it make? In P. Bogaards, B. Laufer (eds) Vocabulary in a
second language: Selection, acquisition, and testing. Amster-
dam, Philadelphia: John Benjamins, 79–100.
Mondria, J-A. / Wit-de Boer, M. 1991. The effects of contextual
richness on the guessability and the retention of words in a
foreign language. Applied Linguistics. 12, 249–267.
Montague, W. E. / Adams, J. A. / Kiess, H. O. 1966. Forgetting and
natural language mediation. Journal of Experimental Psycho-
logy. 72, 829–833.
Myers, I. B. / McCaulley, M. H. 1985. A guide to the development and
use of the Myers-Briggs type indicator. Palo Alto, CA: Con-
sulting Psychologists.
Nagy, W. E. / Herman, P. A. / Anderson, R. C. 1985. Learning words
from context. Reading Research Quarterly. 20, 233–253.
Namei, S. 2004. Bilingual lexical development: a Persian-Swedish
word association study. International Journal of Applied Lin-
guistics. 14/3, 363–388.
Nation, I. S. P. 1990. Teaching and Learning Vocabulary. New York:
Newbury House.
Nation, I. S. P. 2001. Learning vocabulary in another language.
Cambridge: Cambridge University Press.
Nation. I. S. P. / Meara, P. 2002. Vocabulary. In N. Schmitt (ed.) An
introduction to applied linguistics. London: Edward Arnold,
35–54.
The National Education Ministry. 2004. The College English Syllabus.
Beijing: Higher Education Press.
The National Education Ministry. 2004. The university English
syllabus. Retrieved November 15, 2005 from <http://www.edu.
cn/20040120/3097997.shtml>.

313
Nattinger, J. 1980. A lexical phrase grammar for ESL. TESOL
Quarterly. 14, 337–344.
Nattinger, J. / DeCarrico, J. 1992. Lexical phrases and language
teaching. Oxford: Oxford University Press.
Nord, J. R. 1975. Sens-it-Cell. System. 3/1, 16–23.
Nord, J. R. 1978. Developing listening fluency before speaking: An
alternative paradigm. Paper presented at the 5 th World Congress
of Applied Linguistics, Montreal, Canada.
Nurweni, A. / Read, J. 1999. The English vocabulary knowledge of
Indonesian university students. English for specific purposes.
18/2, 161–175.
Odlin, T. 1989. Language Transfer. Cambridge: Cambridge University
Press.
Ogden, C. 1937. A short guide to Basic English. Cambridge: Ortho-
logical Institute.
Oller, J. W. 1978. Pragmatics and language testing. In B. Spolsky (ed.)
Approaches in language testing. Advances in Language Testing.
Series 2. Arlington, Virginia: Center for Applied Linguistics,
39−57.
O’Malley, J. M. / Chamot, A. U. 1990. Learning strategies in second
language acquisition. Cambridge: Cambridge University Press.
Ostyn, P. / Kelly, P. 1979. Tune In. Antwerp: Plantyn.
Ostyn, P. / Godin, P. 1985. RALEX: An alternative approach to lan-
guage teaching. The Modern Language Journal. 69/4, 346–355.
Ostyn, P. Vandecasteele, M. / Kelly, P. / Deville, G. 1985. Towards an
optimal program of FL vocabulary acquisition. Paper presented
at the 5th European Symposium on LSP, Leuven, Belgium.
Oxford, R. 1989. Use of language learning strategies: a synthesis of
studies with implications for strategy training. System. 17/2,
235–247.
Oxford, R. 2001. Language learning styles and strategies. In M.
Celce-Murcia (ed.) Teaching English as a second language or
foreign language. Boston: Heinle and Heinle/Thomson Learn-
ing, 359–366.
Oxford, R. 2003. Language learning styles and strategies: concepts
and relationship. IRAL. 41, 271–278.

314
Oxford, R. L. / Scarcella, R. C. 1994. Second language vocabulary
learning among adults: State of the art in vocabulary instruction.
System. 22/2, 231–243.
Oxford, R. L. 1990. Language learning strategies: what every teacher
should know. Boston: Heinle & Heinle.
Oxford, R. L. 1993. Research update on teaching L2 listening. System.
21/3, 205–211.
Oxford, R. L. / Nam, C. 1998. Learning styles and strategies of a
partially bilingual student diagnosed as learning disabled: a case
study. In J. M. Reid (ed.) Understanding learning styles in the
second language classroom. N. J.: Prentice Hall Regents, 52–61.
Paige, R. M. / Cohen, A. D. / Lassegard, J. / Chi, J. C. / Kappler, B. J.
2002. Maximizing study abroad: A students’ guide to strategies
for language and culture learning and use. Minneapolis, MN:
Center for Advanced Research on Language Acquisition
(CARLA).
Paivio, A. / Desrochers, A. 1979. Effects of an imagery mnemonic on
second language recall and comprehension. Canadian Journal
of Psychology. 33, 17–28.
Paivio, A. / Desrochers, A. 1980. A dual-coding approach to bilingual
memory. Canadian Journal of Psychology. 34, 388–399.
Paivio, A. / Desrochers, A. 1981. Mnemonic techniques in second-
language learning. Journal of Educational Psychology. 73/6,
780–795.
Paradis, M. 1987. The assessment of bilingual aphasia. Hillsdale: Erl-
baum.
Paribakht, T. S. / Wesche, M. B. 1997. Vocabulary enhancement activi-
ties and reading for meaning in second language vocabulary ac-
quisition. In J. Coady, T. Huckin (eds) Second language vocabula-
ry acquisition. Cambridge: Cambridge University Press, 174–200.
Pawley, A. / Syder, F. 1983. Two puzzles for linguistic theory: native-
like selection and native-like fluency. In J. Richards and R.
Schmidt (eds) Language and communication. London: Long-
man, 191–226.
Pearlmutter, N. J. / Garnsey, S. M. / Bock, K. 1999. Agreement pro-
cesses in sentence comprehension. Journal of Memory and
Language. 41, 427–456.

315
Peterson, E. / Coltrane, B. 2003. Culture in Second Language Teach-
ing. Digests from online resources of the Centre for Applied
Linguistics. Retrieved July 10, 2005, from <http://www.cal.org/
resources/digest/0309peterson.htm>.
Pimsleur, P. 1970. Some aspects of listening comprehension. Pro-
ceedings of the Canadian conference on language laboratories,
Montreal.
Pitts, M. / White, H. / Krashen, S. 1989. Acquiring a second language
through reading: a replication of the Clockwork Orange study
using second language acquirers. Reading in a Foreign Lan-
guage. 5/2, 271–275.
Plass, J. 1998. Design and evaluation of the user interface of foreign
language multimedia software: A cognitive approach. Language
Learning & Technology. 2/1, 35–45.
Postovsky, V. A. 1974. Effects of delay in oral practice at the begin-
ning of second language acquisition process. The Modern
Language Journal. 58/5–6, 229–239.
Postovsky, V. A. 1975. The priority of aural comprehension in the
language acquisition process. Paper presented at the 4th AILA
World Congress, Stuttgart, Germany.
Postovsky, V. A. 1977. Why not start speaking later? In M. Burt, H.
Dulay, M. Finocchiaro (eds) New directions in second language
teaching, learning and bilingual education. Washington, D.C.:
TESOL, 17–26.
Pressley, M. / Levin, J. R. 1981. The keyword method and recall of
vocabulary words from definitions. Journal of Experimental
Psychology: Human Learning. 17/1, 72–76.
Prytulak, L. S. 1971. Natural language mediation. Cognitive Psych-
ology. 2, 1–56.
Qian, D. 1999. Assessing the roles of depth and breadth of vocabulary
knowledge in reading comprehension. The Canadian Modern
Language Review. 56/2, 282–307.
Qian, D. 2002. Investigating the relationship between vocabulary
knowledge and academic reading performance: An assessment
perspective. Language Learning, 52/3, 513–536.
Read, J. 1993. The development of a new measure of L2 vocabulary
knowledge. Language Testing. 10, 355–371.

316
Read, J. 1997. Vocabulary and testing. In N. Schmitt, M. McCarthy
(eds) Vocabulary: Description, acquisition and pedagogy.
Cambridge: Cambridge University Press, 303–320.
Read, J. 1998. Validating a test to measure depth of vocabulary know-
ledge. In A. Kunnan (ed.) Validation in language assessment.
Mahwah, NJ: Lawrence Erlbaum, 41–60.
Read, J. 2004. Plumbing the depths: How should the construct of
vocabulary knowledge be defined? In P. Bogaards, B. Laufer
(eds) Vocabulary in a second language: Selection, acquisition,
and testing. Amsterdam, Philadelphia: John Benjamins, 209–227.
Reber, A. S. / Wakenfield, F. F. / Hernstadt, R. 1991. Implicit and explicit
learning: individual differences and IQ. Journal of Experimental
Psychology: Learning, Memory and Cognition. 17, 888–896.
Reeds, J. A. / Winitz, H. / Garcia, P. A. 1977. A test of reading follow-
ing comprehension training. International Review of Applied
Linguistics. 15, 307–319.
Reid, J. 1998. Teachers as perceptual learning styles researchers. In J.
Reid (ed.) Understanding learning styles in the second lan-
guage classroom. N. J.: Prentice Hall Regents, 15–25.
Richards, J. / Rodgers, T. 2001. Approaches and methods in language
teaching. Cambridge: Cambridge University Press.
Richards, J. 1976. The role of vocabulary teaching. TESOL Quarterly.
10, 77–89.
Richards, J. C. 1994. Teacher thinking and foreign language teaching.
The Language Teacher. 18/2. Retrieved 2005, January 20, from
<http://jalt-publications.org/tlt/files/94/richards.html>.
Roediger, H. L. 2005. Memory (psychology), Microsoft Encarta
Online Encyclopedia 2005, retrieved June 10 2005, from
<http://encarta.msn.com/encyclopedia_761578303/Memory_
(psychology).html>.
Ruder, K. / Hermann, P. / Schiefelbusch, R. 1977. Effects of verbal
imitation and comprehension training on verbal production.
Journal of Psycholinguistic Research. 6, 59–72.
Sapir, E. 1970. Language: An introduction to the study of speech.
London: Rupert Hart-Davis.

317
Scarcella, R. / Oxford, L. R. 1992. The Tapestry of Language Learn-
ing: The Individual in the Communicative Classroom. Boston:
Heinle & Heinle.
Schmidt, R. 1990. The role of consciousness in second language
learning. Applied Linguistics. 2/2, 129–158.
Schmidt, R. 1994. Deconstructing consciousness in search of useful
definitions for applied linguistics. AILA Review. 11, 11–26.
Schmitt, N. 1997. Vocabulary learning strategies. In N. Schmitt, M.
McCarthy (eds) Vocabulary: Description, acquisition and
pedagogy. Cambridge: Cambridge University Press, 199–227.
Schmitt, N. 1998. Tracking the incremental acquisition of second
language vocabulary: A longitudinal study. Language Learning.
48, 281–317.
Schmitt, N. 2000. Vocabulary in language teaching. Cambridge: Cam-
bridge University Press.
Schmitt, N. (ed.) 2004. Formulaic sequences: Acquisition, processing
and use. Amsterdam, Philadelphia: John Benjamins.
Schmitt, N. 2005. Current Trends in Vocabulary Learning and Teach-
ing. In J. Cummins and C. Davison (eds) Handbook of English
Language Teaching. Dordrecht, Boston: Kluwer Academic
Publishers.
Schmitt, N. / Meara, P. 1997. Researching vocabulary through a word
knowledge framework. Studies in Second Language Acquisition.
19, 17–36.
Schmitt, N. / Carter, R. 2000. Lexical phrases in language learning.
The language teacher. 24, 1–7. Retrieved June 15 2005 from
<http://www.jalt-publications.org/tlt/articles/2000/08/schmitt>.
Schmitt, N. / Carter, R. 2004. Formulaic sequences in action: An intro-
duction. In N. Schmitt (ed.) Formulaic sequences: acquisition,
processing and use. Amsterdam, Philadelphia: John Benjamins,
1–22.
Schmitt, N. / Dörnyei, Z. / Adolphs, S. / Durow, V. 2004. Knowledge
and acquisition of formulaic sequences: A longitudinal study. In
N. Schmitt (ed.) Formulaic sequences: Acquisition, processing
and use. Amsterdam, Philadelphia: John Benjamins, 55–86.
Schoonen, R. / Verhallen, M. 1998. Aspects of vocabulary knowledge
and reading performance. Paper presented at the Annual

318
 Meeting of the American Educational Research Association.
San Diego, April.
Schouten-Van Parreren, C. / Van Parreren, C. F. 1979. De verwerving
van een vreemdtalige woordenschat: een literatuurstudie.
Levende Talen. 341, 259–270.
Schur, E. 2002. Density and complexity in the semantic networks of
native and non-native speakers. Paper presented at the Second
Language Vocabulary Acquisition Colloquium. Leiden: Uni-
versity of Leiden.
Sciarone, A. G. 1979. Woordjes leren in het vreemde talen onderwijs.
Muideberg: Coutinho.
Scott, M. 1996. WordSmith tools. Oxford: Oxford University Press.
Scott, M. / Johns, T. 1993. Microconcord manual: An introduction to
the practices and principles of concordancing in language
teaching. Oxford: Oxford University Press.
Shin, J. / Wastell, D. G. 2001. A user-centered methodological frame-
work for the design of hypermedia-based CALL systems.
CALICO Journal. 18/3, 517–537.
Sinclair, J. 1991. Corpus, concordance, collocation. Oxford: Oxford
University Press.
Singleton, D. 1999. Exploring the second language mental lexicon.
Cambridge: Cambridge University Press.
Skinner, B. F. 1957. Verbal behaviour. New York: Appleton-Century-
Crofts.
Smidt, E. / Hegelheimer, V. 2004. Effects of online academic lectures
on ESL listening comprehension, incidental vocabulary acquisi-
tion, and strategy use. Computer Assisted Language Learning.
17/5, 517–556.
Smith, B. 2004. Computer-mediated negotiated interaction and lexical
acquisition. Studies in Second Language Acquisition. 26,
365398.
Söderman, T. 1993. Word association of foreign language learners and
native speakers– Ashift in response type and its relevance for a
theory of lexical development. In H. Ringbom (ed.) Near-native
proficiency in English. Åbo: Åbo Akademi, 91–182.
Spolsky, B. / Sigurd, B. / Sato, M. / Walker, E. / Arterburn, C. 1968.
Preliminary studies in the development of techniques for testing

319
 overall second language proficiency. Language Learning. 3,
79–101.
Steil, L. K. 1980. Secrets of being a better listener. U.S. News & World
Report, Inc. 65−66.
Sternberg, R. J. 1987. Most vocabulary is learned from context. In M.
G. McKeown, M. E. Curtis (eds) The nature of vocabulary
acquisition. New Jersey, London: Lawrence Erlbaum Associates,
Publishers, 89–105.
Swan, M. 1997. The influence of the mother tongue on second
language vocabulary acquisition and use. In N. Schmitt, M.
McCarthy (eds) Vocabulary: Description, Acquisition and
Pedagogy. Cambridge: Cambridge University Press, 156180.
Taylor, R. P. (ed.) 1980. The computer in the school: Tutor, tool, tutee,
teacher’s college, Columbia University. New York: Teacher’s
college Press.
Teng, H. / Niu, Q. / Wolff, M. 2004. China’s EFL goals: Are they
being met? English Today. 20/3, 37–44.
Twaddell, W. F. 1973. Vocabulary expansion in the TESOL classroom.
TESOL Quarterly. 7, 61–78.
Van de Poel, K. / Swanepoel, P. 2003. Theoretical and methodological
pluralism in designing effective lexical support for CALL.
Computer Assisted Language Learning. 16/2−3, 173–211.
Vanparys, J. / Zimmer, C. / Li, X. / Kelly, P. 1997. Some salient and
persistent difficulties encountered by Chinese and Francophone
students in the learning of English vocabulary. ITL. 115–116,
137164.
Vanpatten, B. 2002. Processing instruction: An update. Language
Learning. 52/4, 755–803.
Vermeer, A. 2001. Breadth and depth of vocabulary in relation to
L1/L2 acquisition and frequency of input. Applied Psycholin-
guistics. 22, 217–234.
Wang, L. 2004. Corpora and Second Language Acquisition: Chunks,
noticing and concordancing. In A. He, L. Wong. / M. Xu, X.
Cheng (eds) Application of corpora to foreign language educa-
tion: Theory and Practice. Guangzhou: Guangdong Higher
Education Press, 61–71.

320
Waring, R. 1999. Tasks for assessing second language receptive and
productive vocabulary. Unpublished PhD thesis. Retrieved
April 20, 2003, from <http://www1.harenet.ne.jp/~waring/
papers/phd/title.html>.
Wenden, A. 2002. Learner development in language learning. Applied
Linguistics. 23/1, 32–55.
Wesche, M. / Paribakht, T. S. 1996. Assessing second language
vocabulary knowledge: depth versus breadth. The Canadian
Modern Language Review. 53/1, 13–40.
Wesche, M. / Paribakht, T. S. 2000. Reading-based exercises in second
language vocabulary learning: an introspective study. The
Modern Language Journal. 84/2, 196–213.
Whitson, V. 1972. The correlation of auditory comprehension with
general language proficiency. Audio-Visual Language Journal.
10/2, 89–91.
Whorf, B. L. 1956. Language, Thought, and Reality: Selected Writings
of Benjamin Lee Whorf (edited by Carroll, J. B.). Cambridge,
Mass.: Technology Press of Massachusetts Institute of Technol-
ogy.
Willis, D. 1990. The lexical syllabus. London: Collins COBUILD.
Willis, J. / Willis, D. 1989. Collins COBUILD English Course. Lon-
don: Collins COBUILD.
Winitz, H. 1977. Nonauditory auditory disorders. Otolaryncologic
Clinics of N. America. 10, 187–192.
Winitz, H. 1978. The Learnables. Kansas: International Linguistics
Corporation.
Winitz, H. / Garcia, P. A. / Frick, R. 1985. Teaching language through
comprehension. In P. B. Westphall, N. W. Conner, N. Choat (eds)
Meeting the call for excellence in the foreign language
classroom. Lincolnwood, Illinois: National Textbook Co, 14–29.
Winitz, H. / Reeds, J.A. 1973. Rapid acquisition of a foreign language
(German) by the avoidance of speaking. International Review of
Applied Linguistics. 18/3, 245–247.
Wode, H. 1999. Incidental vocabulary acquisition in the foreign
language classroom. Studies in Second Language Acquisition.
21/2, 243–258.

321
Wolter, B. 2001. Comparing the L1 and L2 mental lexicon. A depth of
individual word knowledge. Studies in Second Language
Acquisition. 23, 41–69.
Wray, A. 2000. Formulaic sequences in second language teaching:
principle and practice. Applied Linguistics. 21/4, 463–489.
Wray, A. 2002. Formulaic language and the lexicon. Cambridge:
Cambridge University Press.
Xu, M. 2004. Corpora and cognitive schema theory. In A. He, L.
Wong. / M. Xu, X. Cheng (eds) Application of corpora to
foreign language education: Theory and Practice. Guangzhou:
Guangdong Higher Education Press, 48–60.
Yoshida, K. 1990. Knowing vs. behaving vs. feeling: studies in
Japanese bilinguals, in L. Arena (ed.) Language proficiency:
Defining, teaching, and testing. New York: Plenum.
Zähner, C. / Gupta, G. / Olohan, M. 1994. Lexical resources in CALL.
Computers & Education. 23/1−2, 75–80.
Zheng, T. 2004. Review of Tell Me More Chinese Beginning &
Advanced. CALICO Journal. 21/3, 671–680.

322
List of CAVL/CALL programs/tasks reviewed

Research-based programs/tasks:

CAVOCA: Groot, 2000; see references

Computer vocabulary exercises: Allum, 2004, see references
Lexica: Goodfellow, 1994, 1995, 1999; see references
The lexical CMC task: De la Fuente, 2003; see references
The listening CALL task: Smidt/Hegelheimer, 2004; see references
Words in your ear: Laufer/Hill, 2000; see references

Commercial programs:

English+: Emme Interactive France, France; <http://www.emme.com>.

Intelligent Miracle English: Hai Long, China; <http://61.143.255.27en.
qjent.net/>.
Language Interactive Culture: Centre for Language and Speech
(University of Antwerp), Belgium; <http://webhost.ua.ac.be/
linc>.
Memorising Vocabulary Effortlessly: Pu Gong Ying, China; <http://
www.pgy.com.cn>.
Rosetta Stone (German version): Rosetta Stone,USA; <http://www.
rosettastone.com/en/>.
SuperMemo 98: SuperMemory, Poland; <http//:www.supermemo.
com/>.
Tell Me More (Chinese version): Auralog, France; <http://www.
auralong.com>.

323
List of useful online lexical tools

BNC (searching engines): <http://sara.natcorp.ox.ac.uk/lookup.html>;

<http://view.byu.edu/>.
CLEC (Chinese Learners’ English Corpus) (searching engine): <http://
www.clal.org.cn/corpus/ChiSearchEngine.aspx>.
COBUILD concordance and collocations sampler: <http://www.collins.
co.uk/Corpus/CorpusSearch.aspx>.
Lexical Tutor: <http://www.lextutor.ca>.
WordNet (searching engine): <http://wordnet.princeton.edu/perl/ webwn>.

324
Appendices

Appendix 1: A questionnaire for English teachers

at Chinese Three Gorges University
(translated from Chinese)

Please indicate the statements that are true for you (multiple answers can be selected);
you can provide some extra information in the space after the word ‘others’ if this is
your case.

A. Beliefs about vocabulary teaching and learning

I have the following beliefs about vocabulary teaching and learning:

Vocabulary should be taught explicitly in class
Vocabulary should be learned off by heart from bilingual lists
Students acquire a large amount of vocabulary through extensive reading
Vocabulary is better learned in context
Doing exercises can help learn vocabulary
Vocabulary should be learned through conversations or other communicative activities
Vocabulary should be consolidated by use in different contexts
Others:

B. Teaching vocabulary in the classroom

1. How do you select the lexical items to be taught?

I select a lexical item according to the following criteria:

It is a key word for understanding the text.
It is a frequent word that will appear in many other contexts
The course book says it is important to know
It belongs to the lexical field that we are studying
It is a useful word for my students
The word is difficult to use
The word is likely to be tested in exams
Others:

325
2. How do you present vocabulary?

I present vocabulary in the following ways:

Visually (write the word out)
Aurally (speak the word out)
Kinetically (body movement, gestures, acting out the meaning, etc.)
Using visual aids (videos, pictures, diagrams, real objects, etc)
Using aural aids (tapes, CDs)
Others:

3. What aspects do you focus on in teaching lexical items?

I focus on the following aspects of lexical items

Their pronunciation
Their spelling
Their affixes and roots
Their Chinese translation
Their meaning in English (a paraphrase)
Related words such as antonyms and synonyms
I use them in good example sentences
I describe their usage
I indicate their register
I show some of their collocations
I indicate the part of speech they belong to
Others:

4. What activities or tasks do you use to assist vocabulary teaching/learning?

I use the following activities or tasks to assist vocabulary teaching/learning:

Drill repetition
Vocabulary exercises (e.g., gap-filling, multiple choice, etc.)
Vocabulary games (guessing games, crosswords, etc.)
Get students to work in interactive conversations
Role plays
Listening tasks
Writing tasks
Computer software
Internet sites
Others:

326
5. What vocabulary learning strategies do you recommend or teach to students?

I recommend or teach the following vocabulary learning strategies to my students:

Read extensively to enlarge and consolidate the vocabulary
Use English films or songs to enlarge and consolidate the vocabulary
Participate actively in oral activities to enlarge and consolidate the vocabulary
Think in English
Use contextual clues or form resemblance to guess the meaning
Organise the lexical information in a systematic way, e.g. keep a vocabulary notebook
or vocabulary cards
Use strategies to put vocabulary in a context that will help memorise it
Use sound or rhymes to memorise vocabulary
Use imagery to memorise vocabulary
Ask students to reflect on their own progress in vocabulary learning
Learn other words in the semantic network – types of transport, train, car, bus, bike.
I encourage students to use vocabulary they have just learned
I teach them how to use vocabulary learning strategies in a systematic way.
Others:

6. How do you review vocabulary with students?

I review vocabulary with students in the following ways:

I review new words in lists regularly
I review vocabulary in semantic fields
I test new vocabulary regularly
I regularly schedule written or oral class activities that focus on reviewing recently
learned new vocabulary
I propose further reading or listening activities in which the words occur
Others:

327
Appendix 2: A questionnaire for students
at Chinese Three Gorges University
(translated from Chinese)

Survey of vocabulary learning and strategies

You are required to indicate the frequency of each strategy you actually use
Use frequency (F): 1 = never, 0 = rarely, 3 = sometimes, 4 = often, 5 = always

1. How do you discover new words?

From textbooks and classroom learning activities F: 1 2 3 4 5

From lists of vocabulary arranged alphabetically F: 1 2 3 4 5
From lists of vocabulary arranged semantically F: 1 2 3 4 5
From English conversations with others F: 1 2 3 4 5
(teacher, classmates, foreigners)
From reading English materials (books, newspapers, F: 1 2 3 4 5
magazines, etc.)
From English songs and movies/TV programmes F: 1 2 3 4 5
From the internet (online English pages, chatting) F: 1 2 3 4 5

2. What do you do on encountering new vocabulary?

Ignore it F: 1 2 3 4 5
Ignore it first and come back later to it F: 1 2 3 4 5
Try to guess the meaning from the context F: 1 2 3 4 5
Analyse affixes and roots for the meaning F: 1 2 3 4 5
Ask classmates or the teacher for the meaning F: 1 2 3 4 5
Consult a bilingual dictionary F: 1 2 3 4 5
Consult a monolingual dictionary F: 1 2 3 4 5

3. When learning a new vocabulary item, what aspects do you study?

Study the pronunciation F: 1 2 3 4 5

Study the spelling F: 1 2 3 4 5
Study the affixes and roots F: 1 2 3 4 5
Study the Chinese translations F: 1 2 3 4 5
Study the English explanations F: 1 2 3 4 5
Study the example sentences F: 1 2 3 4 5
Study the way the word is used F: 1 2 3 4 5
(e.g. prevent sb from doing sth)

328
Study the collocations (e.g. make money/ a profit) F: 1 2 3 4 5
Study the part of speech (noun, verb, adjective, etc.) F: 1 2 3 4 5

4. How do you organise the information about the new vocabulary?

Jot down the information on paper F: 1 2 3 4 5

Organise the information in a vocabulary notebook F: 1 2 3 4 5
Make vocabulary cards F: 1 2 3 4 5
Use the vocabulary section in the textbook F: 1 2 3 4 5
Use vocabulary lists (e.g. those in vocabulary books F: 1 2 3 4 5
for Band 4)

5. How do you memorise vocabulary?

Say the word aloud several times F: 0 1 2 3 4.

Copy the word out several times F: 1 2 3 4 5
Look at the word several times F: 1 2 3 4 5
Memorise bilingual lists F: 1 2 3 4 5
Do vocabulary exercises F: 1 2 3 4 5
Connect the word to its synonyms or antonyms F: 1 2 3 4 5
Connect the word with other already known words that F: 1 2 3 4 5
have a formal resemblance (e.g. dam, damage) F: 1 2 3 4 5
Compare words similar in meaning and study them together F: 1 2 3 4 5
Group and organise words (e.g. according to meaning, F: 1 2 3 4 5
part of speech, topic, etc.)
Place the word in a context (a meaningful sentence, F: 1 2 3 4 5
conversation, or story to remember it)
Make up a phrase or a sentence containing the word F: 1 2 3 4 5
Listen to the sound recording of the words F: 1 2 3 4 5
Make up rhymes to link words F: 1 2 3 4 5
Practise words (e.g. verbs) by acting them out F: 1 2 3 4 5
Make a mental image of new words (or the sentence F: 1 2 3 4 5
with the word)
Draw pictures to illustrate the meaning of new words F: 1 2 3 4 5
Visualize the word form mentally F: 1 2 3 4 5
Remember the affixes and roots of new words F: 1 2 3 4 5

5. How do you review vocabulary?

Go over words periodically F: 1 2 3 4 5

Test new words on your own F: 1 2 3 4 5
Test new words with classmates F: 1 2 3 4 5

329
6. How do you retrieve words?

Look at the meaningful parts of the words F: 1 2 3 4 5

(affixes or roots)
Make an effort to remember where F: 1 2 3 4 5
I first met the words

7. How do you make use of new vocabulary?

Try to use new words in speaking and writing F: 1 2 3 4 5

Make an effort to use idiomatic expressions F: 1 2 3 4 5
Try to think in the target language with F: 1 2 3 4 5
the new vocabulary
Interact with native speakers/teachers/classmates F: 1 2 3 4 5
with the new vocabulary

Note: This vocabulary learning and strategy survey tries to replicate the process of
how a new word is encountered by the learner till it is mastered and becomes a part of
the learner’s mental lexicon. The rationale is explained in 7.4.4.3. in the learning
process oriented framework. The wording of the questions in the survey has drawn on
findings from several works: Oxford’s ‘Strategy Inventory for Language Learning’
(SILL) (1990), Cohen and Chi’s ‘Language Strategy Use Inventory and Index’ (2003)
and Schmitt’s ‘Vocabulary learning strategies taxonomy’ (1997).

330
Index

Acquisition 227, 233, 237, 242, 245, 250,

formulaic sequences, 128-129, 262-263, 301, 303, 309, 312, 317-
134-135, 137, 310, 321 318, 325
idioms, 138-141 Data Driven Learning (DDL), 133-
language, 19, 21-22, 38, 76, 78, 134, 137, 179
80, 84, 86, 122, 155, 252, 300- Dictionary use, 111-112, 117, 310
301, 304, 306, 310, 312, 314, Foreign language Learning, 19, 242,
316, 318 250, 300, 302, 304, 311
Chinese culture of learning, 245, 250, Formulaic sequences, 120-137, 142,
251, 256, 277, 279, 298 164, 251, 299, 305, 310, 320-321,
Chinese Learners’ English Corpus, 324
261-262, 269-270, 326 Input, 33, 38, 57, 60, 79-84, 105, 108,
Collocations, 29, 45, 65, 120, 122, 128-129, 132-136, 142, 151, 176,
125-126, 132, 164, 202, 205, 264, 179, 181, 184, 186, 187, 193, 194,
268, 275, 282-283, 288, 307, 314, 197, 198, 200, 202, 203, 215, 218,
326, 328, 331 233, 267, 273, 300, 306, 312, 323
Computer Assisted Language Learn- Keyword method, 156- 162, 300, 318
ing (CALL), 22, 173-176, 178, Knowledge
180, 181, 187-188, 191-192, 203, declarative, 31, 32, 40, 46, 51, 56,
210-211, 231, 256-257, 296, 299- 58-59, 106, 295
300, 302, 304-309, 311, 313, 321- procedural, 31-33, 40, 46, 51, 56,
322, 324-325 58, 59, 106, 295
Computer Assisted Vocabulary Learn- Language
ing (CAVL), 22, 171-237, 294- distance, 91, 92, 93
297, 325 learning strategies, 154-156, 159,
development, 22, 174-176, 188 162-163, 279, 286, 305, 317
evaluation, 191-192, 297 proficiency, 21, 38, 65, 69, 73-74,
Computer-mediated Communication 82, 87, 92-93, 134, 137, 143,
(CMC), 179, 182, 185, 186, 199, 161, 176, 202, 244, 251, 309,
200, 201, 202, 204, 228, 230, 234, 322-324
236, 257, 296, 325 Learner differences, 143, 165
Corpora, 119-123, 131-135, 142, 178, Learning
184, 301, 307-308, 323-324 explicit, 38, 103-104, 107-108,
Culturally loaded words, 87-90, 262- 114, 118, 175-176, 180, 182,
264, 312, 314 185, 211, 230, 277, 295, 304,
Culture, 19-20, 76, 85-90, 94, 98-99, 306, 319
127, 137-142, 145, 156, 168, 188, implicit, 38, 103-104, 107-108,

331
 118, 142, 175-176, 179, 185, 219, 227, 233, 262, 275, 295,
295-296, 304, 319 321, 324, 332
incidental, 105, 107, 109, 112- Linguistic distance, 91-99, 219, 261,
118, 179, 310 264, 301
intentional, 103-107, 117, 310 Listening comprehension, 75, 77, 83,
style, 22, 82, 143, 149-156, 165- 185, 196-199, 204, 210, 227-230,
168, 181, 296, 302, 305-306, 234, 236, 310-311, 318, 321
317, 319 Memory
Lemma, 56, 58, 62-63, 71, 273, 275 episodic, 31, 303
Lexeme, 56, 273 explicit, 103, 106-107
Lexical implicit, 103, 106-107
approach, 131-134, 313 long-term, 31, 83, 187, 212
association, 58 semantic, 31, 56
chunks, 120, 132 Metaphor, 20-21, 54, 67, 74, 139-140,
competence, 27, 30, 56, 63 300
concept, 61, 66, 86-88, 98-99, Mnemonic techniques, 141-142, 157,
262, 269 162, 296, 317
entry, 54-56, 58-59, 62, 64, 195, Motivation, 22, 132, 137, 143-148,
273-276 156, 159, 165-168, 186, 193, 195,
errors, 58, 65, 97, 200, 219, 232- 199, 221, 230, 236-237, 244-247,
233, 261-267, 271-277, 294 296, 302-309
glosses, 109, 111, 185, 196, 210, Multi-word lexical items, 22
224, 230, 236 Output, 30, 33, 59, 133, 179, 186, 273,
information, 54, 58, 62, 112, 162, 306
164, 179, 184, 195-196, 209- Reading comprehension, 38, 110, 115,
218, 224-230, 275, 282, 285- 179, 196, 199, 256, 258, 319
286, 292-293, 296, 313, 329 Second Language Acquisition (SLA),
items, 33, 36, 44, 46, 48, 50, 64, 19, 27, 63, 78-83, 87, 91, 99, 105,
65, 66, 94, 95, 105, 110, 113, 114, 119, 121, 123, 131, 134, 143,
114, 119, 163, 166, 173, 177- 149, 159, 179, 182, 185, 193, 233,
181, 193-196, 204-205, 211, 251, 301, 304, 323
215-216, 228-229, 232, 273, Second Language Vocabulary Acqui-
276-283, 289-294, 327-328 sition (SLVA), 19, 21-22, 27, 75,
relations, 64-67 87, 91, 93, 105, 118, 165-168,
Lexical Frequency Profile, 43-44, 47 189, 296, 299, 302, 304
Lexicon, 28, 39-42, 53, 64-74, 119, Semantic
141, 166, 177, 205, 213, 273, 275, error, 31, 56, 61, 65, 67
299, 306, 324 relations, 64
L1, 53, 72, 74 Semantization, 37, 300
L2, 28, 30, 53-55, 64, 69, 71-74, Suggestopeodia, 78
273 The Lexical Syllabus, 130
mental, 29, 35, 39-40, 46, 53, 54, Vocabulary
56, 64-74, 195, 202, 207-210, active, 39, 313

332
 breadth, 21, 33-40, 44-48, 51, reception, 21, 29, 39-40, 44, 48,
166, 194, 295-296, 319, 323 50-51, 166, 194, 236, 279,
depth, 21, 28, 30, 33-40, 44-51, 295-296
71, 110, 115, 164, 166, 194, receptive, 36, 39-40, 43, 307, 323
295, 296, 303, 319, 323, 324 size, 28, 35, 38-39, 45-49, 160-
knowledge, 21, 27-51, 71, 110, 161, 313
115, 162, 166, 187, 290, 299, testing, 47, 97, 305-306, 314
313-316, 319, 321, 323 use, 30-33, 39, 46-47, 50
learning difficulty, 62, 85-88, 98, Vocabulary Knowledge Scale, 27, 44,
261-262, 275 46, 48, 71, 115
passive, 314 Vocabulary Learning Questionnaire,
production, 21, 29-32, 39-41, 44, 160
48-51, 55, 58, 60, 97, 114-115, Vocabulary Learning Strategies, 114,
118, 120, 136, 166, 186, 194, 118, 156, 159-168, 187-188, 193,
195, 205, 209-210, 216, 236, 195, 199, 204, 206, 218, 229-230,
242, 272-274, 295-296, 304, 234, 280, 282, 285-286, 294, 308,
313, 320 320, 329, 332
productive, 36, 39-43, 47, 314,
323

333

View publication stats