Feb 20, 2013

Grammar rules... again?! Chunks strike back

This is a somewhat belated reaction to Catherine Walter's article which appeared in the Learning English section of Guardian last autumn. Click here to read it.

File:Telramen op de bank in de klas Counting-frames in classroom.jpg
Language or maths?
Spaarnestad Photo via  Nationaal Archief
Dr Catherine Walter’s article Time to stop avoiding grammar rules defends explicit grammar teaching in EFL. Proudly subtitled The evidence is now in: the explicit teaching of grammar rules leads to better learning, the article makes numerous references to a "wide range of studies" that have shown evidence of effectiveness of explicit grammar teaching.

Although no sources are cited, the forthright and cogent tone of the article written by the co-author of How English Works, The Good Grammar Book and Oxford English Grammar Course (all with Michael Swan) would win over any ELT practitioner. As one would expect, to make her argument more convincing, Dr Walter talks with mild disdain about other approaches that have de-emphasised explicit grammar instruction and proposed instead:

"to expose learners to language that is just a bit more advanced than what they currently produce"(Krashen's comprehensible input hypothesis / the Natural approach);
"to wait until a communicative situation demands a certain structure before introducing it" (Task-Based Learning with reactive focus on form)
"to let the grammar emerge naturally from the lived context of the classroom" (Dogme).
One such approach which has dealt a heavy blow to the dominance of a traditional grammar syllabus is the Lexical Approach proposed in the 1990s and based on teaching chunks of language. Without beating around the bush, Dr Walter claims it is WRONG because there are hundreds of thousands of chunks the learner has to commit to memory. She states "With much less time and effort, learners can acquire grammar" (acquire grammar or learn a few declarative rules?) "for putting together comprehensible phrases and sentences".

Does the author imply that, unlike grammar rules, chunks do not have generative value? Surely, chunks can be equally generative. Imagine, your student learns

  Could you pass me the salt, please?
which is a semi-fixed expression which allows variability and helps learners produce similar requests in other situations:

Could you pass me the water, please?
Could you pass me the ketchup, please? 
Could you pass me the menu, please? 
Could you pass me my phone, please? 

 Learning chunks actually facilitates the teaching of grammar and serves as a basis for mastery of the grammar system. Furthermore, grammar is best learnt when students have memorised a chunk which can then be used as a template for creating novel utterances. Let's consider, for example:

  If it doesn't work out, you can always fire me. 

Imagine how many "rules" a learner needs to remember here: no will after if, 3rd person doesn't and not don't… If they memorise it as a chunk, they can go on to produce:

If it doesn’t work out you can always move. 
If it doesn't work out you can always go back to working part-time. 
If it doesn’t happen… 
If it doesn't get better… 

Both Ellises subscribe to this view of grammar acquisition. Rod Ellis (Second language acquisition researcher) says that it is worth focusing on formulaic chunks initially before the teaching of rule-based grammar while Nick Ellis (cognitive linguist), coming from the emergentist perspective, maintains that by memorising and later analysing chunks learners bootstrap their way to grammar.

Native speakers do not realise how much they rely on stock phrases such as If I were you…, When it comes to… There's been a lot of opposition to… when communicating. By some estimates, between 50% and 80% of native speaker English – depending on the genre - consists of prefabricated routines and memorised chunks.

A number of studies have shown that learners across all levels use far fewer chunks than native speakers relying on word-by-word sentence building. The ability to produce appropriate multi-word phrases often lags behind even when students have mastered the third conditional. If anything, we should be advocating more – not less- chunk learning.

In the final paragraph of the article after elevating the role of grammar rules, Dr Walter, as if to pre-empt the imminent backlash by those on the anti-grammar side of the argument, acknowledges that teaching vocabulary is more important than grammar and there is room for both: a grammar syllabus and word lists… 

There you have it: twenty or so years of psycholinguistics, SLA research, cognitive linguistics and corpus linguistics are thrown out of the window, and grammar and vocabulary are decoupled yet again. Ironically, those teachers whose teaching is no longer dictated by the outdated and largely discredited slot-and-filler model of language learning probably shrugged off the article. But those who have never let go of the hold that pedagogic grammar has on their teaching will now be vindicated and continue - with renewed vigour - battering their students with grammar exercises.


Ellis, N. (1996). Sequencing in SLA: Phonological memory, chunking, and points of order. Studies in Second Language Acquisition, 18(1), 91-126. Available at: www-personal.umich.edu/~ncellis/NickEllis/Publications_files/Ellis1996Chunking.pdf

Ellis, R. (2005). Principles of instructed language learning. Asian EFL Journal, vol. 7. Available at: www.asian-efl-journal.com/September_2005_EBook_editions.pdf

Walter, C. (2012, September 18). Time to stop avoiding grammar. The Guardian Weekly. Available at: http://gu.com/p/3aa24


  1. The lexical approach sounds similar to TPRS which focuses on structures which could also be called "chunks". Are you familiar with TPRS?

    1. Yes there are similarities between the two but I should learn more about TPRS. Any sources / blogs / articles you can recommend?

  2. Can't everything just coexist? ;) Truly, as you say, learning chunks reinforces the memory of grammatical forms.

    Btw, can you point me towards the "number of studies have shown that learners across all levels use far fewer chunks than native speakers relying on word-by-word sentence building"? I'd like to read them.

    1. Good point, Tyson.
      I should have provided a reference but since it's a blog and not an academic article I skipped it :)

      There are quite a few really - where do I begin?

      Foster, P. (2001). Rules and routines: A consideration of their role in the task-based language production of native and non-native speakers. In M. Bygate, P. Skehan and M. Swain (eds) Language tasks: Teaching, learning and testing, (pp 75-93). Longman

      Granger (1998) talks about how L2 learners underuse but also overuse / overrely on certain chunks, specifically collocations with all-purpose adverbial intensifiers completely and totally.

      For a general overview see Wray, A. (2002). Formulaic language and the Lexicon. CUP

      Happy to see you here again.

  3. Hi!
    Nice Article.
    I would like to continue what Tyson Sebrun wrote above. These two approaches can and should coexist! Different students learn differently and teachers need to find their students' preferable approach. Notice, for example, that special ed. students need more structured rule-based method, while nore advanced students can be challenged by the chunking method.
    Nice Blog!

    1. Hi Reut
      Thank you for your comment. I also think the two can co-exist and that's the issue I address in my upcoming article in Modern English Teacher. But I wouldn't agree that chunk-based learning is something reserved for higher levels, which is a common belief. I am sure you'll find there's a lot of chunk learning in the Primary School: how are you? I don't know. What's your name?. So it's worth starting with these and then moving on to explicit grammar rules. That's the approach Rod Ellis, whom I already mentioned above, advocates too.
      Thank you for stopping by.

  4. Just a quick comment on Reut's comment:
    "Different students learn differently and teachers need to find their students' preferable approach."
    True, but at the risk of oversimplifying matters, I would argue that all or most students would benefit from a lexis first then grammar approach, from very early levels, as Leo said.
    I think perhaps the main difference here would lie in different students' tendencies to jump into grammar analysis sooner rather than later, but, one way or another, the rationale behind chunk learning holds true across different levels and learning styles, I think.
    Great article, Leo! :-)

    1. Thank you, Luiz!
      What research seems to suggest is that those who keep chunking and relying on memorised sequences are actually better at grammar analysis, which is interesting and counter-intuitive. I would think there are two types of learners: analytic (starting to analyse grammar early on) and more holistic ones (those who rely on chunks) - I think this is how Skehan described them.
      Thank you for your comment and sorry for taking ages to respond.

  5. Interesting stuff...I'm not sure Disdain is the right word though is it? Didn't seem so disdainful.

    Anyway I think we clearly see that the pages of the guardian are not the place to make sweeping statements about english education.

    1. Hmm perhaps "disdain" was a bit harsh, yes. Maybe that's why the Guardian rejected it as an "in response to" piece :)
      But, as you probably know, it's a subject matter close to my heart and I feel very strongly about it.
      Thank you for stopping by!

  6. Chunking language is a great idea because it makes language learners reflect on the patterns of the language. And language is all about patterns.
    There are different kinds of learners, however that catch cry has been used to defend all kinds of practices. The reality is that the state of language learning/teaching is in a poor state and we need to start looking for practices that work, no matter what the student preferred style. (What happpens if their preferred style leads them up the garden path? Should we accept that?)
    I have been using variations of chunking amongst other kinds of practices and have found that all students respond to it well...it all depends upon how it is presented.
    Some students take longer to get on board but once they see the power of chunking they are more than willing to hop on!

  7. ja leo i can definitely relate to learning chunks. i was getting to know a brazilian girlfriend in portugal, hearing over and over again the language in real-life contexts. i also read the paper and subtitles, and after a while a few pages of a small learning book prepared for US govt agents, apparently.

    the turnaround came when i would recognise a collocation after like the 100th time that it had been repeated within earshot of me. the key then was simply asking a question about the phrase, or looking up one of the words, or reading it in the newspaper.

    somewhere in your brain all of those previous 100 real-life applications fall into place and in this way you get the whole package: accent, context, emphasis, emotion. grammar? absolutely not, except by self-analysis. no explicit rules. i now even do some work in two-way consecutive interpreting with nary a complaint. @hwshy

    1. Hi @hwshy -

      What you describe makes so much sense. But 100 encounters seems really a lot. I wonder if classroom instruction has any role to play then. Can it speed up the process or at least have a "priming" function by sensitizing the learner to the sheer number of chunks they can read, hear and pick up outside the classroom?

      Thank you for stopping by and sharing your personal experience

      By the way, do we know each other?

  8. The written piece is truly fruitful for me personally; continue posting these types of articles.Grammarly

  9. Grammar is so important for the students. With out they didn't find success in education. So as a students i want to say thanks of you because you write so well.


Related Posts Plugin for WordPress, Blogger...