L is for (Michael) Lewis

5 09 2010

(Continuing an occasional series of the type ‘Where are they now?’)

Michael Lewis and me: University of Saarbrücken

A reference in last week’s post (P is for Phrasal Verb) to the fuzziness of the vocabulary-grammar interface naturally led to thoughts of Michael Lewis. It was Michael Lewis who was the first to popularize the view that “language consists of grammaticalized lexis, not lexicalized grammar” (1993, p. 34). This claim is a cornerstone of what rapidly came to be known as the Lexical Approach – rapidly because Lewis himself wrote a book called The Lexical Approach (1993), but also because, at the time, corpus linguistics was fueling a major paradigm shift in applied linguistics (under the visionary custodianship of John Sinclair and his brainchild, the COBUILD project) which, for want of a better term, might best be described as ‘lexical’. Lewis was one of the first to popularize this ‘lexical turn’ in applied linguistics, and he did so energetically, if, at times, contentiously.

So, what happened to the Lexical Approach – and to Lewis, its primum mobile?

Well, for a start (as I argued in an article in 1998), the Lexical Approach never was an approach: it offered little guidance as to how to specify syllabus objectives, and even its methodology was not much more than an eclectic mix of procedures aimed mainly at raising learners’ awareness about the ubiquity of ‘chunks’. Moreover, Lewis seemed to be dismissive – or perhaps unaware – of the argument that premature lexicalization might cause fossilization. To him, perhaps, this was a small price to pay for the fluency and idiomaticity that accrue from having an extensive lexicon. But wasn’t there a risk (I argued) that such an approach to language learning might result in a condition of “all chunks, no pineapple” i.e. lots of retrievable lexis but no generative grammar?

In the end, as Richards and Rodgers (2001) note, the Lexical Approach “is still an idea in search of an approach and a methodology” (p. 138). Nevertheless, as I said in 1998, “by challenging the hegemony of the traditional grammar syllabus, Lewis… deserves our gratitude.”

Michael responded graciously to these criticisms, acknowledging them – although not really addressing them – in a subsequent book, Teaching Collocation (2000). There the matter rested. Until 2004, when I published a ‘lexical grammar’ – that is, a grammar based entirely on the most frequent words in English – and, in the introduction, paid tribute to my ‘lexical’ precursors, specifically Michael Lewis, and Jane and Dave Willis.

Michael was not pleased. When I next ran into him, at an IATEFL Conference a year or two later, he was still fuming. Apparently, by suggesting that his version of the Lexical Approach had anything in common with the Willis’s, or that my book in any way reflected it, was a gross misrepresentation. The sticking point was what Michael calls ‘the frequency fallacy’, that is, the mistaken belief that word frequency equates with utility. Much more useful than a handful of high-frequency words, he argued, was a rich diet of collocations and other species of formulaic language. I, by contrast, shared with the Willis’s the view that (as Sinclair so succinctly expressed it) ‘learners would do well to learn the common words of the language very thoroughly, because they carry the main patterns of the language’ (1991, p. 72). To Michael, ‘patterns of the language’ sounded too much like conventional grammar.

When we met again, a year or two later, at a conference at the University of Saarbrücken, we found that we had more in common than at first seemed. For a start, we sort of agreed that the chunks associated with high frequency words were themselves likely to be high frequency, and therefore good candidates for pedagogical treatment. And Michael was working on the idea that there was a highly productive seam of collocationally powerful ‘mid-frequency’ lexis that was ripe for investigation.

A few months later, at a conference in Barcelona, we had even started talking about some kind of collaborative project. I was keen to interest Michael in developments in usage-based theories of acquisition, premised on the view that massive exposure to formulaic language (his ‘chunks’) nourishes processes of grammar emergence – a view that, I felt, vindicated a re-appraisal of the Lexical Approach.

But Michael is enjoying a well-earned retirement, and I suspect that he’s satisfied in the knowledge that the Lexical Approach, his Lexical Approach, whatever exactly it is, is well-established in the EFL canon, and that his name is stamped all over it.

So, then, what’s the Lexical Approach to you?


Lewis, M. 1993. The Lexical Approach. Hove: LTP.
Lewis, M. 2000. Teaching Collocation. Hove: LTP.
Richards, J., and Rodgers, T. 2001. Approaches and Methods in Language Teaching (2nd edition). Cambridge University Press.
Sinclair, J. 1991. Corpus, Concordance, Collocation. Oxford University Press.



42 responses

5 09 2010

Thanks for a fascinating insight into the history of your engagement. Having been an avid reader of both your works, I always thought there was a high degree of congruency between them. Either way, I am a big fan of The Lexical Approach. It’s passionate, engaged and entertaining – qualities which most books in the field are short of. Above all, it was, and still is, a book that imparts in me a kind of interrogative compulsion, compelling me to question everything I did in the classroom, especially those cherished assumptions I didn’t even know I had. Whether or not The Lexical Approach was ever able to answer all those questions is of secondary importance in a way, although I still think it has a lot to offer.

5 09 2010

Hi Scott

I’ll start with a plea for the inclusion of Implementing the Lexical Approach in your appraisal. While it still doesn’t answer all the points you raise, it is much more readable and of practical use to the language teacher who is perhaps not so well-versed in linguistic theory.

Implementing The Lexical Approach was one of the first books that I read when I began to ask the question, “is there more to teaching than what I learned for the TEFL certificate?” Lewis is quite clear when he states,”I totally disassociate myself from any suggestion […] that ‘Lexis is the answer’.”

Your question seems to hang on the definition of approach. I’ve always seen the skills I learned from Michael Lewis as tools to be applied in teaching. (Coming from a technological background, I tend to view everything as a tool!) My view would be that the way in which you use a set of tools to do a job can easily be described as an approach.

Alone the highlighting of language chunks has proved to be a valuable asset in teaching. Even with low-level learners the ability to recognise these chunks can aid development enormously. In some case I find that focusing on these chunks is easier for the learner than trying to internalise grammar rules or a context. Of course this approach tends to produce a Ford Escort rather than a Morgan, but for many learners that’s precisely what they are looking for.

So what is the Lexical Approach to me? It’s one of the most useful tools that I’ve added to my teaching approach. It isn’t the answer to everything and it doesn’t work with all learners, but it has served as a springboard for my personal development and critical thinking regarding teaching and learning. For that I’m very thankful that the work exists and would say that every language teacher should be aware of the options it presents.

5 09 2010
Scott Thornbury

Thanks Olaf, for reminding me of Lewis’s Implementing the Lexical Approach (LTP, 1997) which, in fact, was referenced in my 1998 critique. I mentioned Teaching Collocation only because it was in that book that Michael alluded to my criticisms.

I agree that whether or not Lewis’s program constitutes an ‘approach’ is probably casuistry. Moreover, if the truth be known, I’m much more inclined to Michael’s position than I was in 1998. This is largely because of the influence of both the ‘usage-based theorists’ and corpus linguists like Michael Hoey. Finally, these two disciplines, which have been running along parallel tracks, now seem to be converging, if an article in the latest Applied Linguistics is any indication (an article, incidentally, that graciously references Lewis). In it, the researchers Rita Simpson-Vlach and Nick Ellis summarise these convergences:

Functional, cognitive linguistic and usage-based theories of language suggest that the basic units of language representation are constructions — form-meaning mappings, conventionalised in the speech community, and entrenched as language knowledge in the learner’s mind… Constructions are associated with particular semantic, pragmatic, and discourse functions, and are acquired through engaging in meaningful communication. Constructions form a structured inventory of speakers knowledge of the conventions of their language, as independently represented units in a speaker’s mind. Native-like selection and fluency relies on knowledge and automiatized processing of these forms…. Corpus linguistics confirms the recurrent nature of these formulas. Larger stretches of language are adequately described as collocational streams where patterns flow into each other…”

Simpson-Vlach, R., and Ellis, N.C. 2010 An Academic Formulas List: New methods in phraseology research. Applied Linguistics, 31/4, 487-512.

“Collocational streams where patterns flow into each other” – that is pure Lewis!

5 09 2010
Nick Bilbrough

Hi Scott,

Thanks for another great post. I’d agree with Olaf that reading ‘Implementing the Lexical Approach’ could have a much greater pay off for teachers, being more specific and practical in its style. In fact, I’d say that has it probably influenced the way that I teach more than any other book.

Can you expand a bit on the idea that ‘premature lexicalization might cause fossilization’? One of the big things for me about Lewis’ work is the increased importance of memorisation and getting chunks completely right from the outset. Isn’t premature grammaticalization actually more likely to lead to fossilization?


5 09 2010
Scott Thornbury

Hi Nick, thanks for your comment – and I agree that Implementing is definitely the better book.

Regarding the point about fossilisation, when I raised this objection, I’d been influenced by cognitivist scholars like Peter Skehan, who proposed a three stage progression in language acquisition: from a lexical stage to a syntactical stage, and then relexicalization. That is to say, the first stage of language acquisition is the accumulation of words and phrases, with only a rudimentary syntax. The syntactical phase involves the application of grammar rules which in turn become automated, stored and retrieved like ‘big words’. Studies of fossilisation (now known as stabilisation) suggested that learners who are not encouraged to, or even discouraged from, ‘adding grammar’ are stuck in the pre-syntactic, lexical phase. It seemed to me that an over-emphasis on lexis at the expense of grammar would exacerbate this tendency. Now I’m not so sure. Like you, I suspect that an over-emphasis on grammar might be equally counter-productive.

5 09 2010
Paul Maglione

The Lexical Approach, and all its derivative works and related approaches, is hugely significant for learners. It represents a (much as I abhor the over-use of the term) “paradigm shift” in the way both language accuracy and language fluency can be taught, in that it allows us to start side-stepping the grammar rules-based method which confuses matters for many learners as much as it clarifies them for others.

Crucially, with the Lexical Approach and the notion of Grammaticalised Lexis, Lewis was echoing — consciously or not — very valid principles of brain learning, which describe how the human brain eschews rules-based instruction in favor of trial-and-error, learning by doing, and frequency / usefulness as a brain trigger for retention and a more or less unconscious absorption of basic structure logic allowing for reproduction of similar language strings.

Lewis’s conclusion was, in my view, one of those breakthrough moments in pedagogy that result in a clear and usable hook for teachers to move up a notch or two in teaching effectiveness. I hope that the Lexical Approach will continue to benefit from study and refinements such as those of Scott Thornbury and others, and above all continue to be used by teachers in preference to rules-based grammatical approaches that are simply not in synch with how language is taught most effectively.

5 09 2010
Karenne Sylvester

just quickly before responding -thanks for adding the retweet button to your page, makes it so much easier to send our your posts :-))…

Hmm… now, cautiously, because I have been tossing around thoughts on these issues… while knowing that I am in no way as experienced nor as educated as either of you and am pretty much am only exploring ideas– so these will be just the thoughts of a dogme teacher who’s looking around for better ways to tackle her students’ vocabulary acquisition but who’s been noticing (from running googledocs for feedback) that her intermediate + above German students’ are more frequently making lexical as opposed to grammatical errors +mistakes.

My dilemma is namely related to context… and the related vocabulary they need to learn (I wrote with an example of a discussion with one of my students recently) –

should the vocabulary scaffolded be based on the personalized context of the random natural conversation between learner and teacher and the words/collocations/chunks discussed and/or presented in that instant (and post discussion)

should a materials-lite context be presented first with the vocabulary needed naturally emergent (but then how is the teacher supposed to be prepared enough to have access to at the tips of her fingers/tongue to teach those chunks in that instant

or are these chunks supposed to be taught without a context… but if they are taught without context how can they become personalized and therefore learned?

or are those chunks supposed to be taught, randomly, in pre-designed contexts…

and what about widest exposure – just repetition? but across contexts??

And hmm… while I’m here musing away my Sunday afternoon… my other confusing thoughts / dilemma is that isn’t there a difference between the high frequency /mid frequency chunks spoken by native speakers and the errors which are consistently made by non-native speakers of various different cultures – surely determining how and what words sit together most often is not necessarily the same thing as addressing what words are specifically not put together naturally, most often, in various contexts, by learners who once they have passed on from a beginner-elementary level consistently don’t do this correctly… and how does the brain actually store the L2 chunks, is it grammar that gets in the way…

Well, I do hope these questions make some sort of sense…

6 09 2010
Scott Thornbury

Thanks for your musings, Karenne.

With regard to your four ‘options’ for teaching chunks, the issues seem to revolve around the choice of pre-selected vs emergent, and decontextualised vs contextualised.

On the latter option, I’d normally recommend contextualised teaching (and I think Michael Lewis would agree) but there may be a case for learning lists of decontextualised chunks (along with their translation), especially for those learners who are disposed to memorisation. (The idiomaticity, and register-specificity of many chunks might mean, though, that, in the absence of context, learners could miss important aspects of meaning and use). In the end, close study of texts (including transcriptions of naturally-occuring talk) would seem to offer the best source for chunk learning.

As for whether you shape the learner’s output, teaching chunks at the point of need, this will depend a lot on such factors as level, class-size, learners’ needs etc. A one-to-one situation offers the ideal context for reformulation of learner output so as to make it more idiomatic and lexically rich. (But as you say, it does require a degree of quick thinking – unless you record the student and then work on the transcription over night).

But I also think there’s a case for a lexical syllabus of pre-selected high-frequency, pragmatically useful formulae – which, actually, is what the old functional-notional syllabus was all about. In this way you can perhaps be more sure of ‘netting’ the important stuff rather than leaving it to chance.

As for your second round of musings, I’ll save those for later!

5 09 2010

Karenne wrote:

“my other confusing thoughts / dilemma is that isn’t there a difference between the high frequency /mid frequency chunks spoken by native speakers and the errors which are consistently made by non-native speakers of various different cultures – surely determining how and what words sit together most often is not necessarily the same thing as addressing what words are specifically not put together naturally”

Would it be true to say that most people need to make the same kind of meanings frequently all over the world. We all need to say “How long does it take?’ or ‘I’m going shopping’. When you teach/highlight those very portable chunks wouldn’t you, at the same time, look at the ways in which your students frequently combine the wrong words (= ‘words that are specifically not put together naturally’) to make those meanings?

5 09 2010
Karenne Sylvester

Hi Glennie,

I think that’s why I was asking if it is grammar that gets in the way of the so called natural chunking as while the words chunk together the order of the words doesn’t… (I will admit I’m not really sure if I’m expressing myself well…

-I’m going to copy and paste from one of my ss blogs (because he’s a very interesting student: a Russian who self-taught himself English with an amazing technical vocabulary but who makes frequent very basic level errors and mistakes)…

In this video Randy tell us interesting History about mobile phone.
At the 70-th some engineers come with a first prototype of Mobile phone to AT&T. It was a big one, he had a bad sound quality, but engineers was assured that it’s a the really good product, and customer buy it.

This was a self-directed blog post – my student went to the Stanford site on entrepreneurship, chose a video he liked and decided to talk about what he had heard. While his LA impressed the socks off me, and although when you “talk” to him about these topics in person he has numerous coping mechanisms in place so that is able to make himself understood – and quite well – he also speaks quickly but when you read his written texts you get this sort of thing…

Dissecting it for feedback, we could say he should learn, as chunks, e.g.
– In the 70s
– came up with


but what about in the first line, doesn’t he need to learn the continuous thoroughly and properly? (I got him to update his blog on present continuous by personalizing the structures)

but can I/ should I instead teach

“In this video ____ is telling us about the” which would probably be quite a natural and portable chunk and it is no doubt his Russian structure that is getting in the way?

dunnoo… glad to get your help 🙂

5 09 2010

Well, there is so much here.
To focus on the origin of the problem with the present continuous. I note this from Scott’s post:

‘…that massive exposure to formulaic language (his ‘chunks’) nourishes processes of grammar emergence…’

Is it that that massive exposure has never happened with respect to the continuous? And is it, instead, that this student tried to teach himself the use of the present continuous by (uselessly?) memorising rules about its use (= grammar getting in the way?).

I am dabbling around in deep waters here. Would like to hear from others on what is a great example for us to get our teeth into.

(By the way, I seem to remember a passage from Lewis in which he says that one thing is teaching English and another entirely different thing is teaching writing in English. I think he says that writing is a completely different animal with its own unique problems. Am I inventing that?)

5 09 2010
Mark Smith

This discussion is all great stuff. Where I would like to take it is to an acknowledgement that there is no such thing as ESP. It seems to simply be a marketing tool. Whether or not business people, for the most obvious example, know certain bits of terminology is not the biggest fish that wants frying. Learners, whatever “needs” that have, urgently require a grasp of the fabric of a new language. This fabric, if I may be so bold, is what you and Lewis are trying to identify.

5 09 2010

Hi Mark

I really can’t agree with you that ESP is just a marketing tool. The refrain, “Business English is still English” is something that I base my business English teaching on, but Business English learners need stronger skills in style, pitch and jargon, not to mention negotiation, presenting and intercultural competence – all of which, though laudable, are not as important in a general English course. A lab assistant needs far more practice in report writing and journal reading than a standard course provides. The fabric of the language is so wide that few people (no-one) need(s) it all, but groups of people need discrete aspects more than others.

How can such learner requirements fit into the idea that there’s no such thing as ESP?

5 09 2010

Thanks for this article. It’s interesting to me, as the Lexical Approach perfectly matches with the approach I take to teaching (and learning) language. But not having a formal education in applied Linguistics, I was never aware of where these ideas came from. I’ll be checking out Lewis’s work.

5 09 2010
Alex Case

Mark- Given the titles that LTP released, I’d be surprised if Michael Lewis agreed that there is no such thing as ESP

Scott- By complete coincidence, I was reading that exact same article of yours a couple of weeks ago. No idea how I ended up there- that’s the miracle of Google. Really helped me write a review of a very Lewisian book, Teaching Chunks of Language from Helbling, for MET

6 09 2010
Mark Smith

Olaf – Style = Choice of language. You cannot have style without choice. You cannot have choice without a generality.
Pitch = approach. Say no more. Alright I will. To have a runway you have to have land around it. Without land around it you just have a thin strip of land. What happens if you veer off course? You’re in the drink.
Jargon = some words that are specific to a particular context. They can learn that themselves in an evening can’t they?
Negotiation – If you can strike a deal, you can strike a deal. If you can’t, you’re in the wrong field.
Presenting – Hopefully gain the confidence to impart information convincingly to a live audience outside of the EFL classroom.
Intercultural competence – Ahh! Yes. Business English students do need that. As does every single learner of a foreign language.

…..My point being that all the things you list as being specific to the needs of a learner who needs English for business purposes are essential for any learner. Yes alright some people need to write reports in English and some people have to read journals in English but (and this is a big but) they can’t do that if they don’t have a grasp of constitutes the language. You misunderstand my use of the word fabric. By fabric, I don’t mean what the language is, I mean what the language is made of. And I think that is the the project that Lewis was and Thornbury is engaged in.

Alex – The existence of something is not proven by the titles released about it.

6 09 2010
Andrew Pickles

Hi Scott and everyone else!

I think I might have to express a bias here – I think the Lexical Approach is fantastic, not just the initial book but its two follow-ups as well and I really like The English Verb. Not only do I like the theory but his writing is one of the most engaging and thought provoking of anybody’s. As to whether he’s right or not…

I don’t like grammar, never have and probably never will, as Jim Scrivener says here http://bit.ly/9oPgXp it is questionable whether explicit grammar instruction is particularly helpful, especially for speaking spontaneously. I also think that a lot of the time a fasle dichotomy is set up which seeks to separate grammar and lexis (today we will do the present perfect and then later we will look at words to do with experience etc.) whereas they obviously are simply two sides of the same coin (dodgy metaphor but it’ll have to do).

I think the appeal of the Lexical Approach lies in its focus on lexis and its consequent reduction of focus on grammar (I don’t think Lewis removes it altogether, simply re-evaluates its place) which allows for a richer classroom experience with the attention of the student directed towards vocabulary acquisition and extension, and away from abstract rules and meta-language. I’m also not sure whether or not part of its appeal lies in its usefulness to the native speaker with his or her perhaps more intuitive feel for chunking? This is a touch controversial and I offer it as an idea, nothing more.

Students like learning words, they don’t generally like learning grammar and I think many teachers (myself included) prefer concentrating on vocabulary rather than grammar. I do also think that as Dave Willis (?) pointed out many items are better treated as lexical than grammatical, ‘would’ is a good example of this. I also wonder a lot of the time if errors identified as grammatical are not in fact lexical errors or deficits?

In modern ESL teaching then lexical chunks, patterns etc. are more easily integrated into lessons that tap into the vast analytic power of concordancers and corpus linguistics.

My one problem with the Lexical Approach is, as Scott has pointed out, selection – where do you begin? I believe a marriage here between Dogme and the Lexical Approach (maybe Dogme Chunking?) is the way ahead as it allows selection to be dealt with naturally as the need/demand/opportunity arises in class and for focus on naturally occurring lexical patterns to predominate – it does require pretty fast thinking teachers though, ones who Jason Renshaw called ‘fluent’.

all the best


6 09 2010
Jason Peppard

Thanks for this great post Scott! This is a topic that I have been fascinated by for some time now. I’m interested in your original caveat of fossilization but am glad to see that you are now looking into the awareness of lexical patterns leading to grammatical emergence; this is what I think happens as it seems to correspond to first language acquisition.

Personally, although my teaching has been heavily influenced by The Lexical Approach, I believe that Dave Willis deserves much more credit and coverage here since his Lexical Syllabus predates Lewis’s work by three years.

Frustrated with the sloppy and inconsistent treatment of lexical chunks, or what I prefer to call lexicogrammatical (LG) patterns, in the majority of commercial ELT materials, I did my dissertation on the development of a functional-lexicogrammatical (FL) approach to syllabus design in which I attempted to expand the lexical approach while addressing issues of presentation and context. Basically, I built on Willis’s (2003) integrated model by integrating course specific pedagogic corpora (Willis, 2003), pattern grammar (Hunston & Francis, 2000) and data-driven learning (DDL) (Johns, 1991) within a task-based framework. In the pre-task phase of each FL lesson, the students process a text for meaning and then work through several DDL exercises paired with pattern grammar to raise their awareness of frequent and useful LG patterns found in the pedagogic corpus. Next, the students are asked to complete functional, communicative tasks in which the LG patterns will be useful, but not necessary, for task completion. I performed a comparative analysis between my FL syllabus, a structural-grammatical equivalent and a post methods control with the results showing the FL syllabus to be significantly more effective for raising student LG awareness.

I believe LG awareness is an important first step to fluency development as well as grammatical emergence and often wonder why the lexical syllabus and lexical approach remain on the fringe of ELT.

6 09 2010
Scott Thornbury

Thanks Jason. I’m fascinated to learn about your dissertation: did you publish a paper on the subject?

Yes, the pattern grammar analyses that Hunston & Francis have done has added a signifcant new dimension to the grammar-lexis interface (well, not so new, since Hornby had popularised verb patterns in the 1950s but then they fell into disrepute, having become associated with pattern practice drills). And I agree that the Willis’s deserve more credit, and that a ‘combined lexical approach’, that brings together the frequency principle AND the chunking principle might be based on these four claims: (the first two of which paraphrase Willis; the second two Lewis)

1. meaning is encoded primarily in words
2. the most frequent words in English encode its most frequent meanings
3. words frequently co-occur with other words (collocations and fixed phrases) and they often occur in particular syntactic environments (grammar patterns)
4. fluency is a function of the capacity to store and deploy, in real time, these high frequency lexical and syntactic co-occurrences

(after Thornbury, 2002, How to Teach Vocabulary).

To which we could add the emergentist claim, nicely summarised here:

“From the perspective of emergent grammar … learning an additional language is about enhancing one’s repertoire of fragments and patterns that enables participation in a wider array of communicative activities. It is not about building up a complete and perfect grammar in order to produce well-formed sentences.” (Lantolf and Thorne, 2006: 17).

6 09 2010
Jason Peppard

Hi Scott,

Thanks for your reply and the great references! I haven’t published a paper based on my dissertation research yet, but am currently in the submission process. I would be happy to send you a copy if you would like to read it and would be thrilled to receive any feedback you might have.

Will you be at the JALT conference in Nagoya this November?

6 09 2010

Jason, do you have an example of a lesson plan with your approach? I’d be very interested to see one.

7 09 2010
Jason Peppard

Hi Glennie,

Yes I made four initial lessons for a beginner level general English university class here in Japan. I’d be happy to send you a lesson handout and hear what you think. It’s certainly in an early development phase at the moment, but the lessons went well and the students seemed to enjoy them.

My email is: jason_peppard@hotmail.com

7 09 2010

Few people deserve a well-earned retirement as much as Lewis. The fact that the lexical approach has already had such an impact on our world is testament to that. Can you imagine where our profession would be now without his work?

7 09 2010
Matthew Spira

What is being discussed is to how to generate English language “native-like” fluency in people who already have a set native language.

It’s probably overall a lost cause.

However, start with children and in twenty years pretty much every issue talked about in this thread goes away.

You don’t need convoluted theories to explain what is going on.

I, an American living in South Korea, see it everyday with my 5 year-old son and my 3 year-old daughter. Every single day my wife and I see that- although neither one of us is Korean- our children speak better Korean than English.


9 09 2010
Scott Thornbury

Thanks for the comment, Matt.

I don’t think anyone’s denying that massive exposure and interaction from an early age provides the optimal conditions for second language acquisition. There are at least two issues, however, that motivate the construction of ‘convoluted theories’, as you put it. 1. Few learners – especially adults – have access to these optimal conditions; 2. Given that, what is it about child SLA (or FLA for that matter) that might be applicable to adult SLA?

And one answer to that second question seems to justify adopting a lexical – rather than purely grammatical – approach: i.e. that children in both FLA and SLA seem to acquire formulae, holophrases, or fixed phrases (aka chunks) from an early age. Like words, these formulaic items are used consistently with a particular meaning. And, although these chunks appear, superficially to “have grammar”, they are unanalysed. That is, they are not generated by internalised grammatical rules, but are instead learned, stored, retrieved, and used as if they were single lexical items and without regard for their internal grammar. And, even when child language becomes more fully grammaticised, it still retains formulaic elements. Using borrowed chunks of language allows the child to communicate beyond its present level of grammatical competence, and without any sacrifice of fluency.

The acqusition of formualic language may also have a developmental function. Ann Peters (The Units of Language Acquisition, 1983) suggests that these gestalt items provide at least some children with data to hold in reserve for subsequent analysis. That is, these memorised chunks are available as raw material for subsequent segmentation into, and storage as, smaller units, from which regular syntactic rules are then generalisable.

So, what you are noticing in your children’s SLA development are processes of early lexicalisation and later syntacticalisation entirely consistent with the principles of a lexical approach.

10 09 2010
Klaus Beutelspacher

Strangely, this critical period thing always reminds me of “the knights who say Ni”. They demand and then beg King Arthur in Monty Python’s Holy Grail to cut down a tree with a herring. Now if that couldn’t be done, the knights would probably conclude that trees can’t be cut.

Without dwelling upon the subject, like by mentioning millions of people who HAVE in fact achieved native-like fluency in a foreign language and who could be examined by the knights of the critical period if they really wanted to, there’s a point to be made.

Matthew, your herring here is the the “how to generate … language… in people”. That’s precisely what doesn’t happen, neither with children, nor with successful language learners. It’s simply not the way it works, it doesn’t cut the tree. They learn it just like that, no one generates anything in them. If you’re waiting for that to happen, you’ll probably never learn Korean!

My guess is that there are three significant differences between children and (most) adult learners:
1. The determination and possibly the ability to observe with the utmost accuracy
2. A knowledge of what needs to be done NOW in order to learn (for children usually: Play)
3. Patience. Seemingly endless patience.

Give an adult learner those three, and he will learn at least as fast as a child (safe certain phonetic aspects, but that’s not significant. I’ve had Russian native speakers whose spoken German was closer to the standard than Helmut Kohl’s. Incidentally, their grammatical correctness might have been better than his too.)

Scott, you’ve lost me, I tried to contribute an answer to your question 2., but I don’t see how this ties in with Lewis. Who, in his right mind, would suggest a “purely grammatical approach” for child FLA??

10 09 2010
Scott Thornbury

Klaus – how people answer question 2 (i.e. what is it about child SLA or FLA that might be applicable to adult SLA?) is the history of second language teaching methods. If you believe that it’s all about the imitation and positive reinforcement of habit-forming patterns – that’s your methodology. If, instead, you believe it’s about massive exposure, no stress, the right to be silent etc – you’ll try and simulate those conditions in the adult classroom. If, on the other hand, you believe that it’s about some kind of conscious attention to form-meaning mappings, and/or about forced production and feedback, you’ll go for that kind of teaching approach. Or if you believe it’s about appropriating chunks of pragmatically useful language for later analysis, through jointly constructed, mutually-supportive, interaction – then that’s what you’ll probably attempt to replicate in the classroom. In other words, give me your theory about child acquisiiton and I’ll tell you what your pedagogy is like.

8 09 2010
Denilso de Lima

Hi Scott, I really loved this post. To be honest with you all, I love anything related to the Lexical “Approach” for three reasons.

First, because it helped me a lot an English Language Learner in Brazil and it keeps helping me to improve day by day. I was first introduced to the idea of learning lexically in 1998 when I read an interview given by M. Lewis to a magazine we have back here. For the first time I was reading something that changed my learning style. I had the chance to see that there was much more than learning prescriptive grammar and uncommon words. As a learner, the Lexical Approach to me was a source of relief. I had the chance to re-learn what I had learned.

Second, the Lexical Approach helped me as an English Language Teacher. This is so because I could show my students that they didn’t have to waste their time in “memorizing” grammar rules and lists of words. I taught them how they could get started in the fluency path if they observed [noticed] how English Language was really used by native speakers. So, as a teacher the Lexical Approach helped me to help others to learn [acquire] the English Language in a totally different way. English Language became more meaningful not only to my students but also to me. As a teacher, the Lexical Approach to me was [and still is] the light at the end of the tunnel.

Third, the Lexical Approach contributed to my carreer. Here in Brazil [as you may know] English Language Teachers still hold the belief that memorizing grammar rules [use X when…], technical words for grammar [Present Perfect, Past Continuous, etc], and lists of words are going to help learners to become fluent speaker of the language. Learners, influenced by their teachers, also believe in that. That’s when I started writing books [in Portuguese], giving workshops [and lectures], taking part in Teachers’ Conferences, etc. to let them know what the Lexical idea of teaching is. As the Lexical Approach was heavily criticized in its beginning, people here think that it has to be forgotten and not even mentioned in universities and teachers’ meetings. So, as a Teacher Trainer and book author, the Lexical Approach to me is now a challenge. We don’t want people to change completely, but we want to show them that there is something else. This is a hard thing to do, specially when they agree on the idea that the Lexical Approach is not a valid teaching technique and is dead.

To sum up, the Lexical Approach to me is three things: 1) a relief for learners who are tired of memorizing this and that and gaining no confidence in speaking the language fluently, 2) a solution to change the traditional [and heavily grammatical] way of teaching we have here, and 3) a challenge that needs to be shown to other professionals as a set of teaching techniques and strategies which can motivate learners to never lose hope when it come to become fluent in English.

I guess that’s all I had to say. I hope I made myself clear. Take care!

8 09 2010
Scott Thornbury

Thanks, Denilso, for that ringing endorsement of the Lexical Approach. It would be interesting to survey practising teachers to find out what book or article most influenced their classroom practice. I am prepared to bet good money that Lewis’s The Lexical Approach would be one of the favourites.

26 12 2014
conci mazzullo

Dear Scott,
I’m an English teacher from Italy, and since I discovered the Lexical Approach I’ve tried to teach collocations to my students. Unfortunately in Italy it’s very difficult to find course books based on this method! So I adopted English vocabulary in use basically working on the lexicalised verbs… and it works! Students memorise collocations connected with for ex Have and start thinking in English! They tend not to translate from Italian. I’d like to improve my working system by implementing it, but have found very little as regards course books except Innovations, which I have used for a while! Could you suggest me do something else! I’d love to read and learn more about it! But also need practical stuff to adopt and apply. I belong to LEND Italy and we’re planning a meeting here in Catania and we would like to invite a guest teaching us new performative activities to experiment in our classes! Thanks

8 09 2010

Great to hear that ‘fighting talk’ Denilso.

Best of luck in your battle to help the Lexical Approach gain more acceptance.

8 09 2010
Klaus Beutelspacher

Fascinating! I must admit that I’d never heard about the Lexical Approach before!

With a background in educational philosophy and psychology rather than applied linguistics or language teaching, I could never understand the “traditional” approach anyway. Quotes because it’s not really traditional, language learning has worked for a couple of million years for the species homo, before grammar was invented.

So at least I didn’t have to shift paradigms. Yet other problems arise, like: Take away the grammar, what’s the teacher supposed to do with his time? How do you craft tests that increasingly represent the key goals of educational efforts?

If you don’t teach grammar, there must be something else. It has to be a lexis shaped by the influence appropriate in the context and situation, the proper morphology , word order or prosody etc. So far, I’ve used terms like “language elements” or “bits and pieces” to explain to participants what we are doing. I’ll read up on Lewis, and I guess I’ll use “chunks” soon!

I still don’t have an answer for those who are learning / teaching primarily for tests. And I don’t see myself teaching pre-fabricated chunks, if that’s what Wilson implies.

8 09 2010
Scott Thornbury

If you don’t teach grammar, there must be something else…

Yes, there is something else AND there is grammar. But what is normally classed as grammar for teaching purposes (i.e. the verb phrase) is a sub-set of a sub-set of all that is patterned in language. It also happens to be a sub-sub-set that doesn’t cause learners as many problems as other sub-sub-sets (e.g. the noun phrase).

The emergentists use the term ‘construction’ to cover everything ranging from chunks through verb patterns to larger syntactic units. “A construction may be defined, very generally, as an established pairing of a form with a meaning.” (Taylor, J. ‘Syntactic constructions as prototype categories’ in Tomasello, M. (ed.) (1998) The New Psychology of Language.

“Constructions are of different levels of complexity and abstraction; they can comprise concrete and particular items (as in words and idioms), more abstract classes of items (as in word classes and abstract grammatical constructions), or complex combinations of concrete and abstract items (as mixed constructions).” Ellis, N. and Larsen-Freeman, D. (2009) ‘Constructing a second language: analyses and computational simulations of the emergence of linguistic constructions from usage’ in Language As a Complex Adaptive System (Language Learning, 2009).

They add, significantly, that “the acquisition of constructions is input-driven and depends on the learner’s experience of these form-meaning-use combinations in interactions with others” (p. 92).

10 09 2010
Klaus Beutelspacher

Thank you Scott for the reply and those quotes. Now I’m not familiar with emergentism in language learning, but seeing myself as a constructivist, I naturally like the term “construction”. However, since it “covers everything”, it’s not likely to help you discover anything.

For participants, it’s both interesting and comforting that there are entities that can be learned even when you are not memorizing rules and vocabulary items. Usually they’ve never thought about that, and it takes some getting used to.

For teachers, it should be interesting in what way their classroom activities (like teaching words and rules of “grammar for teaching purposes”) interact with the construction processes in the learners’ brains and bodies. Is there a causal connection at all?

Once you think along constructivist lines, you will invariably end up with the question: We know that there is learning, but how do we know that there is teaching?

My concern: All these approaches have very nice ideas and principles, but when it finally percolates up to the methodology, somehow the syllabi, the tests, the textbooks have come back in. And in the end, it’s just more of the same.

Lewis (not Wilson!) is in the mail, I’ll find out about that one soon!

12 09 2010
Luiz Otavio

I think the profession owes a lot to Michael Lewis, no doubt. But I wish “Lexical phrases and language teaching”, written by James R. Nattinger, Jeanette S. DeCarrico a few years earlier would get more credit. I think in many ways that excellent book laid the groundwork for Lewis’ Lexical Approach.

12 09 2010
Scott Thornbury

Thanks, Luiz, for reminding us of that pioneering book. A quick trawl through my computer’s memory delivers this extract, which gives a bit of the book’s flavour:

Lexical phrases are integral to conversation, for they provide the patterns and themes that interlace throughout its wandering course. These phrases are essential, even for rudimentary “communicative competence”, yet texts that present conversational language do not do so in any systematic way that would permit learners to form connected, functional discourse. (p. 121)

26 10 2010
Pearson Brown

Hi Scott,

Long time no see.

This is a very interesting discussion in an excellent blog. I’d just like to correct something which was suggested above.

I was working closely with Michael in the three years or so before he published The Lexical Approach. Right from the start, he was talking about the ideas that he set forth in that book. If I had been less indolent, my book Business Partners, containing some LA type exercises, would have appeared at the same time as the Willis’s book. If anybody influenced Michael’s thinking, apart from his business partner Jimmie Hill, it was Peter Wilberg, whose book Business English was published simultaneously with mine.

However, Michael was always very much his own man. He has always taken great pride in being an original thinker. The Willis’s deserve great credit for their work but I don’t think they had any direct influence on Michael.

Pearson Brown

26 10 2010
Scott Thornbury

Thanks Pearson – nice to hear from you. Thanks for the correction – I didn’t mean to imply that the Willis’s prefigured Michael Lewis, or that he was influenced by them in any way – in fact he took great pains to distance himself from their frequency-focused approach (and, by association, from mine). It just happens that they both (he and David W.) used the term ‘lexical approach’ more or else round the same time, althought the Willis’s were quick to cede it, once Michael had put his stamp upon it.

26 10 2014
Michael Roberts

Dear Scott,
I attended a workshop you gave at BC Cairo, around 2003, and very much enjoyed it. Thanks for that.
Anyway, I came across your blog as I was in a Linked-in conversation about the number of tenses in English, and wanted to know what Michael Lewis had written about future tenses and future forms. To my (very pleasant) surprise, I’ve just discovered that Michael Lewis is still alive, whereas everybody on my Delta course seem to think he wasn’t. Being that way inclined, I checked with wikipedia, and Michael Lewis doesn’t have an entry! Is there any reason for this that you can think of? Considering his influence on ELT teaching?

Finally, I’m not sure if this is the right forum for this kind of question, but I would love to get in touch with him. Any idea how?


18 02 2016

Dear Scott,

“A few months later, at a conference in Barcelona, we had even started talking about some kind of collaborative project. I was keen to interest Michael in developments in usage-based theories of acquisition, premised on the view that massive exposure to formulaic language (his ‘chunks’) nourishes processes of grammar emergence – a view that, I felt, vindicated a re-appraisal of the Lexical Approach.”

Any progress on that–5 or so years later, without ML, and lots of G.Jordan Lexical Approach bashing in the meantime? Any meeting with M. Hoey?

I’d be eager to know.



10 03 2016
Scott Thornbury

No, Thom – nothing came of it, unfortuately. Thanks for your interest!

3 07 2018

Hello dear Dr. Scott Thornbury? I am very happy to meet you. I know your books and it is for the last couple of years that I have been introduced with your name and your materials(books). I know that you have put your claims (arguments) about Michael Lewis’s Book. Having this in mind, I would like to kindly request you to give me the official website of Michael Lewis( the author of Lexical Approach). I am doing my PhD work on Lexical Approach and its impact on vocabulary learning. Thus, I want o obtain a piece of advice from him, but it is difficult to get him. Generally, I need your help to meet him please.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: