The Lexloal Approach: a journey without maps?

Scott Thornbury teaches and trains at International House, Barcelona. He is the author of About Language (CUP), shortlisted
for the 1998 Ben
Masses of words
Warren Prize. Here he puts the Lexical Approach in its historical context and evaluates it in terms of its theoretical
base and its pedagogical implications.
A New Zealand friend of mine who is studying Maori asked me recently what l, as a language teacher,
would make of his teacher’s method. He explained: ‘We just do masses of words — around a theme, for
example, family, or food. We have to learn these words before the next lesson. Then we come back and
have a conversation — about family, food, etc. and we use the words. The teacher feeds in the grammar
that we need to stick the words together.’ He added that he thought the method worked a treat. This
contrasted markedly with my own experience of learning Maori, where the teacher took great pains to
lead us, discrete step by discrete step, through the intricacies of Maori grammar. The net result, l suspect,
is that my friend’s Maori is a lot better than mine.
What l tried to explain to my friend was that, on the evidence of his account, his Maori teacher was a
witting or unwitting practitioner of a ‘lexical approach’. lt is the purpose of this article to flesh out that
explanation, while at the same time to suggest some limitations of this approach and, indeed, to raise the
question as to whether it is an approach at all.
Vocabulary teaching has come a long way since the days when one coursebook writer advised:
‘As one of the basic aims is to enable the student to use a number of high-frequency patterns
rather than to build up a large vocabulary, the lexical range has deliberately been kept small’
(Alexander I967, p. xviii). The advent of a communicative approach set the stage for a major
re-think of the role of lexis and a recognition of its meaning-making potential. Vocabulary, which
had previously been seen as little more than a resource for filling the slots in grammatical
structures, became a learning objective in its own right, such that by l984, in the introduction to
their Cambridge English Course, Swan and Walter were claiming that ‘vocabulary acquisition is
the largest and most important task facing the language learner’ (p. vii). i
mi
Chunks
At the same time, an unrelated but significant development was taking place in the study of first language
acquisition. A number of researchers, departing from the still relatively ‘fresh’ Chomskyan view that
linguistic competence consists solely in the ability to deploy an innate and rule-governed
sentence-making capacity, suggested that the memorization of chunks of language might be equally
productive and, far from being incidental to language acquisition, might in fact power it. Ann Peters (I983)
suggested that unanalysed holophrases (such as this-is-mine, give-me, and leave-me-alone) are first
acquired as single units, and are then available for subsequent segmentation into, and storage as,
smaller units from which regular syntactic rules are then generalisable. This ‘chunking’ process serves
two purposes in early language production: it enables the child to have chunks of language available for
immediate use, thereby saving processing time, and it provides the child with data to hold in reserve for
subsequent analysis. This paves the way for a ‘dual-mode’ processing capacity, involving both
item-learning and system-learning, supplying short-term and long-term needs respectively.
Meanwhile, Pawley and Syder (I983) not only proposed that the adult language user has at their
command a repertoire of literally hundreds of thousands of what they called ‘lexicalised sentence stems’,
but that the goal of native-like fluency requires of the second-language learner a similar command,
including the capacity to distinguish ‘those usages that are normal or unmarked from those that are
unnatural or highly marked’ (ibid. p. I94). They concluded that the native speaker’s linguistic competence
might be likened to a ‘phrase book with grammatical notes’.

Two systems
In the light of these findings, the goals of second-language teaching needed redefining. The notion of
communicative competence as being solely rule-based was insufficient. ‘lt is much more a matter of
knowing a stock of partially pre-assembled patterns, formulaic frameworks, and a kit of rules, so to speak,
and being able to apply these rules to make whatever adjustments are necessary according to contextual
demands.’ (Widdowson I989, p. I35). In other words, two systems co-exist: a formulaic, exemplar-based
one, and a rule-based analytic one. Nevertheless, materials writers were slow on the uptake, perhaps
daunted by the sheer enormity of this ‘stock of partially pre-assembled patterns’ and the implications this
might have on syllabusing and pedagogy. Coursebooks became more, not less, analysis-based, and only
a handful of social formulae and sentence heads (How do you do? Would you Iike...? Do you mind if I...?)
were taught as unanalysed units.
lt was the advent of corpus linguistics, and of the COBUILD project in particular (Sinclair I987), that gave
a new impetus to a lexical view of language description and acquisition. For a start, computers provided a
powerful means of highlighting patterns of repetition in text and were quickly conscripted into identifying
and categorising habitual co-occurences of words such as collocations and fixed formulaic phrases.
Conclusive evidence was found for the view that words hunt in packs. Moreover, computers were also
able to provide reliable information as to word frequency, suggesting to researchers like Willis (I990) that
this information might offer coursedesigners the means to organise instruction along lines that would
better represent the learners’ needs than the conventional grammatical syllabus — a syllabus that in
Willis’ view ‘gives a very restricted picture of the grammar of English’ (p. I5). Accordingly, in what was
billed by its publishers as ‘a major advance in the teaching of English’, Dave and lane Willis wrote the
Collins COBUILD English Course (l 988).

A Lexical Syllabus
They were driven by the wish to devise a syllabus that would ‘specify the basic meanings of English, the
meanings which even the most elementary users of the language would need to encode’ (Willis, l990, p.
45). Frequency information offered the key: ‘The commonest and most important, most basic meanings in
English are those meanings expressed by the most frequent words in English’ (ibid. p. 46). Accordingly,
the 700 most frequent words (which, incidentallyfconstitute some 70% of English text) were chosen as the
content of Level l of the course. Corpus data was then scrutinised in order to identify how these words
typically behave in context — that is, their structural environments and patterns of co-occurence with
other words. Some of these findings flatly contradicted the hand-me-down rules of conventional
coursebooks, the syntactic behaviour of would and any being two well-documented
cases.
Because of their strong commitment to a task-based methodology (see Willis i996) and in order to
generate the targeted 700 high-frequency words in fairly natural contexts, the writers selected a series of
common topics and related tasks which formed the backbone of the course. On the way, the learners are
exposed to recordings of native speakers performing related tasks, and this input is in turn subjected to
consciousness-raising tasks where the focus is on key lexical items, and their associated syntactical
envrionments, but without reference to traditional grammatical labels.
lt was perhaps this absence of overt grammatical labels, along with the innovative task-based approach,
which scared off potential converts, and which accounts for the fact that the Collins COBUILD Englsh
Course was less than a runaway success. Reading Willis’s (I990) tightly argued rationale for the course,
one can’t help regretting that this was the case. ln a market where publishers are conspicuously reluctant
to back innovation, the failure of a project so brave and so principled was the publishing equivalent of the
Titanic going down.
Nevertheless (to pursue the marine metaphor) a shot had been fired across the bows of the grammar
syllabus, and the lexical approach, far from being scuppered, was about to be refitted and relaunched.

A Lexical Approach
ln l993 Michael Lewis wrote and published The Lexical Approach, boldly subtitled The State of ELT and
the Way Forward. This was followed, in I997, by Implementing the Lexical Approach. Polemical and
sometimes hectoring in tone, the two books (referred to hereafter as LA and ILA respectively) gather up
various theoretical strands (including Nattinger and DeCarrico’s (I992) work on Lexical Phrases and
Language Teaching) to mount a vehement attack on both the conventional grammar syllabus and on the
PPP (presentation-practice-production) methodology it is associated with. Provocatively, Lewis claims that
‘grammar is not the basis of language acquisition, and the balance of linguistic research clearly
invalidates any view to the contrary’ (LA,
p. I33). And, again, ‘l am dismissive of, and regard as fundamentally theoretically unsound, much that
currently passes for grammar practice’ (LA, p. I62). And, finally, ‘the fact is the PPP paradigm is, and
always was, nonsense’ (Lewis I996, p. ll). ln these respects Lewis sides with Willis (in spirit if not in tone),
but there the similarities end.
ln place of discrete-item grammar teaching Lewis promotes a lexical chunk view of language: ‘The
essential idea is that fluency is based on the acquisition of a large store of fixed and semi-fixed
prefabricated items, which are available as the foundation for any linguistic novelty or creativity’ (ILA, p.
l5). ln place of PPP he offers OHE (observe-hypothesise-experiment), an inductive,
consciousness-raising methodology, while at the same time he leans heavily on Krashen’s (I985)
proposals on the necessity of high quantities of roughly-tuned input.
Lewis insists that his lexical approach is not simply a shift of emphasis from grammar to vocabulary.
Rather, it is a shift of perspective away from both grammar and vocabulary: ‘Language consists not of
traditional grammar and vocabulary but often of multi-word prefabricated chunks’ (ILA, p. 3). These
chunks include such things as collocations (to catch a cold, a broken home), fixed and semi-fixed
expressions (nice day for it; that’s/it’s not my fault), and idioms, (to beat about the bush; to go hell for
leather). Following Pawley and Syder (I983), Lewis argues that these multi-word prefabricated chunks
occupy a crucial role in facilitating language production. What he is less clear on is whether, as Peters
(I983) claimed was the case for first-language acquisition, these multiword units play a part in the
restructuring of the learner’s internalised second-language grammar, through subsequent processes of
segmentation and analysis. ln other words, Lewis seems more concerned about improving the fluency of
the learner’s output than increasing the complexity of the learner’s developing language system (Skehan
i998), a point l shall return to.
Lewis insists that he is offering ‘a principled approach, much more than a random collection of ideas that
work’ (lLA, p. 205) and he defines approach as being ‘an integrated set of theoretical and practical beliefs,
embodying both syllabus and method’ (LA, p. 2). Subsequently he redefines his agenda more narrowly:
‘The Lexical Approach has less to say about innovative methods than might be expected. This is because
it is explicity an approach, not a syllabus or method’ (Lewis I996, p. l3), and he reassures teachers that, in
adopting the Lexical Approach ‘the change in your thinking may be considerable, but the change in what
you actually do in class is relatively small’ (ILA, p. 20l).
Teachers, however, who may have been won over by Lewis’s hatchet job on traditional grammar
syllabuses and PPP methodology, might feel somewhat shortchanged by such reassurances, and argue
that surely such a radical agenda demands major, not minor, adjustments to classroom practice. And,
however much he might now wish to retract his initial claims that the Lexical Approach suggests ‘a
radically different view of methodology’ (LA, p. I46), the fact is that by calling it an Approach (with a capital
A, moreover), rather than, say, Techniques for Teaching Chunking, he runs the risk of it being evaluated
as such.
ls the Lexical Approach, then, an approach? And, if so, how coherent is it? And, if coherent, how useful is
it? (Because practising teachers will have little interest in a set of principles that have few or no clear
implications for classroom practice, or that can only with difficulty be operationalised.)
In search of a theory
Following Richards and Rodgers (I986), an approach ‘refers to theories about the nature of language and
language learning that serve as the source of practices and principles in language teaching’ (p. I6). It is
clear that Lewis does have a consistent theory about the nature of language: ‘Language consists of
grammaticalised lexis, not lexicalised grammar’ (LA, p.vi). Nevertheless, it is not so clear what
implications this view of language has on syllabus specifications. We know what sort of syllabus Lewis
does not favour: neither a grammatical one nor a lexical one (‘The Lexical Approach...is specifically not a
lexical syllabus’ (LA, p. l09)). In fact, for Lewis, given the holistic nature of language, ‘no step-by-step
linear syllabus can be remotely adequate’ (LA, p. 47). Nor does he have much time for a task-based
organisation. The strongest hints he drops regarding course content relate to texts: ‘A central requirement
of the Lexical Approach is that language material should be text and discourse, rather than
sentence-based.’ (LA, p l I2). While he provides examples of the kinds of activities such texts and
discourses might be subjected to (e.g. ‘Ask learners to underline chunks they can find in a text’ (ILA, p.
l08)), the failure to specify how such texts and discourses would be selected and organised makes it
difficult to visualise how the Lexical Approach is operationalised in the long term. Lewis offers us the
prospect of a journey, even an exciting one, but it is a journey without maps.
Nor is it clear whether Lewis has a coherent theory about how languages are learned. He is clearly
sympathetic to Krashen’s view as to the necessity (if not sufficiency) of comprehensible input: ‘Listening,
listening and more listening’ (LA, p. I93). Like Krashen, too, he places more faith in acquisition than in
learning, and claims that ‘there is no evidence that explicit knowledge helps performance’ (LA, p. 62).
Nevertheless, he insists that ‘students need to develop awareness of language to which they are
exposed’ (LA, p. I95), particularly the identification of chunks, which suggests that he recognises a role for
consciousness-raising (a position that Krashen would not accept). ‘Accurate noticing of lexical chunks,
grammatical or phonological patterns all help convert input into intake’ (ILA, p. 53).
The implication is that these noticed chunks are stored in memory and retrieved ‘undigested’, as it were.
That is, they engage the learner’s item-learning capacity rather than the rule-based one. This places
formidable demands on the learner's memory: but, as we have seen, Lewis offers no clear guidelines as
to selection and grading — apart from promoting a dictionary of collocations (Hill & Lewis I997) in which
the user is advised that ‘storing combinations like declare war, impose rigid discipline in your memory is
one of the best ways to build an effective vocabulary’ (op. cit. p. 7). How is one to achieve this enormous
task? (According to the blurb, there are 50,000 noun collocations in the dictionary alone). Lewis seems to
assume that massive exposure will do the trick: ‘It is exposure to enough suitable input, not formal
teaching, which is the key to increasing the learner’s lexicon’ (ILA, p. I97). If this is the case, then this
raises the question as to whether many of the ‘teaching’ ideas included in Lewis’s books are redundant,
and not only that, a drain on time that could be more usefully spent simply reading. (It also raises the
selection-and-grading question yet again: what is this ‘suitable input’ and how is it organised?)
Dangerous liaisons
Furthermore, as Skehan (I998) points out ‘there is a danger... that an exemplar-based system can
only learn by accumulation of wholes, and that it is likely to be excessively context-bound, since
such wholes cannot be adapted easily for the expression of more complex meanings’ (p. 89). That
is to say that phrasebook-type learning without the acquisition of syntax is ultimately
impoverished: all chunks but no pineapple. lt makes sense, then, for learners to keep their options
open and to move between the two systems and not to develop one at the expense of the other.
‘The need is to create a balance between rule-based performance and memory-based
performance, in such a way that the latter does not predominate over the former and cause
fossilization’ (ibid. p. 288).
Fossilization is likely to occur, then, when the learner becomes dependent on lexicalised language at the
expense of engaging syntactisization processes. In fact, Lewis seems actively to encourage this
dependency by, for example, quoting approvingly Nattinger’s (I988) suggestion that ‘one way to promote
fluency is by encouraging ‘pidginization’, urging students to put language together the best they can and
avoid the self-monitoring that would inhibit its use’ (p. 70). As Skehan might respond: ‘This way madness
lies!’
To return to my Maori examples: while my teacher’s method promoted the total reliance on a rule-based
competence, with its attendant disadvantages such as lack of fluency, my friend’s teacher’s method
promoted an exemplar-based competence, with the danger of premature fossilization. However, by
insisting on the students grammaticizing the lexis they were using, that is, by pushing them to produce
comprehensible output, this danger was perhaps averted. Lewis, on the other hand, attaches little value
to output: ‘The Lexical Approach is less concerned than some communicative methods with output.’ (ILA,
p. 49). lt is difficult to see, therefore, how the Lexical Approach balances the need for fluency with the
need to guard against fossilization.
In short, the Lexical Approach is not an approach, not in the strict sense, since it lacks a coherent theory
of learning and its theory of language is not fully enough elaborated to allow for ready implementation in
terms of syllabus specification.

Lively debate
However, in the light of the widespread interest and even enthusiasm generated by these two books,
such criticisms may seem at best academic and, at worst, (to use a fixed expression) sour grapes. Lewis
is justified in claiming that ‘when The Lexical Approach was published in I993 it stimulated wide and lively
debate’ (ILA, p. 7) and the term ‘lexical approach’ is now firmly entrenched in the discourse of ELT
professionals. To some extent this must be due to Lewis’s own skills at self-promotion, and to his robust
and engaging (and decidely non-academic) writing style. In short, he speaks the language of teachers.
(Amongst Diploma candidates on courses at our centre, Lewis is consistently voted their fave read). But
there is more to it than that. By publicising a feature of language that has until recently been largely
ignored in EFL courses, and by offering accessible pedagogical practices with which to highlight and
practise it, Lewis has enriched classroom practice considerably. The shift of pedagogical focus from an
atomistic view of language to a more top-down view reflects related movements in

U 2 MET
VOL7 NO 41998
discourse-analysis and genre-analysis. What's more, by asserting the basic ‘patternedness’ of language,
a lexical approach provides justification for the formulaic, unanalysed treatment of a lot more language
than has been the case since the advent of the highanalysis era. (All those handwritten classroom signs:
HOW DO YOU sAY...? I DON'T UNDERSTAND, CAN YOU WRITE IT? etc. fit neatly into ‘Lewis’s thesis).
Moreover, by challenging the hegemony of the traditional grammar syllabus (although without being able
to offer a viable alternative), Lewis, like Willis, deserves our gratitude. That he has done all this by riding
on the shoulders of his more academic predecessors should not be held against him: our profession is
short of popularists — people who can mediate between the ivory tower and the chalk face — and Lewis
is a great popularist.
Clearly, the Lexical Approach is ‘work in progress’. l suspect that we have not heard the last word
on it, neither from the Willis’s nor from Michael Lewis, nor even, perhaps, from my friend's Maori
teacher. More first-hand accounts are needed from learners and teachers as to how such an
approach is being managed and evaluated; more lexically-targeted materials need to be written,
published, and trialled; and more research needs to be undertaken, particularly with regard to the
part memory plays in second-language learning, and whether (and under what conditions)
memorised language becomes analysed language.
References

Alexander, L. (I967) First Things First. Longman

Hill, Lewis, M. (I997) Dictionary of Selected Coilocations. Language Teaching
Publications Krashen, S. (I985) The Input Hypothesis: Issues and Implications. Longman
Lewis, M. (I993) The Lexical Approach. Language Teaching Publications
Lewis, M. (I996) ‘Implications of a lexical view of language’. In Willis, Willis, D.
(eds.) Challenge and Change in Language Teaching, pp. 4-9. Heinemann Lewis, M. (l997)
Implementing the Lexical Approach. Language Teaching Publications
Nattinger, j. (I988) ‘Some current trends in vocabulary teaching’. In Carter, R., &
McCarthy, M. (eds.) Vocabulary and Language Teaching, pp.62-82. Longman
Nattinger, DeCarrico, ]. (I992) Lexical Phrases and Language Teaching. Oxford
University Press Pawley, A. Syder, F. (I983) ‘Two puzzles for linguistic theory: nativelike selection and
nativelike fluency’. In Richards, j., Schmidt, R. (eds.) Language and Communication, pp. l9|-227.
Longman
Peters, A. (I 983) The Units of Language Acquisition. Cambridge Universiy Press
Richards, Rodgers, T. (I986) Approaches and Methods in Language Teaching.
Cambridge University Press
Sinclair, j. (ed.) (I987) Looking Up. Harper Collins.
Skehan, P. (I998) A Cognitive Approach to Language Learning. Oxford University Press
Swan, M. Walter, C. (l984) The Cambridge English Course, I, (Teacher’s Book).

Cambridge University Press Widdowson, H. (I989) ‘Knowledge of language and ability for use’.
Applied Linguistics,
I0, pp. I28-37 Willis, D. (l990) The Lexical Syllabus. Harper Collins Willis, j. (l 996) A Framework for
Task-Based Learning. Longman Willis, Willis, D. (I988) Collins COBUILD English Course. Harper Collins
7 ||| Scott Thornbury