Let me give you a few words and see if you can construct a story from the words:
Mother little girl. Mother say little girl go see grandmother. Mother give little girl big basket food. Mother say 'You take food grandmother'. . . .1
I feel reasonably confident that you will have identified the opening of the story of Little Red Riding Hood. It is not true to say, however, that we have simply a string of words to tell this story. If I had offered you the sequence:
Mother, girl little. Say mother grandmother go see girl little. Basket big food girl little mother give. Say 'Grandmother food take you' mother.
You would certainly have found this much more difficult, perhaps impossible, to interpret. What, then, is the important difference between the first and the second versions of the story?
You might answer this by saying that the word order in the first version makes sense. More precisely, you might say the first version follows the conventions of English clause and phrase structure. Each clause has the structure Subject + verb + . . . . In the phrases little girl and big basket the adjective comes in front of the noun. So the first version does conform to some of the rules of English grammar. It follows the rules of English word order or, to put it another way, the rules of English clause and phrase structure.
So, it is possible to tell a story quite adequately with a string of words and a very limited grammar of structure. There are no definite articles or indefinite articles in the first version of the story, and no other determiners. There are no verb tenses. This raises an interesting question. If things like articles and tenses are redundant why do we bother with them at all? The answer, of course, is that articles and tenses are far from redundant. Even in the telling of a simple story we can make things much easier for our listener by using the full resources of the grammar:
Once upon a time there was a mother who had a little girl. The little girl was going to see her grandmother. Her mother gave her a big basket of food and said 'Take this food to your grandmother.'
So grammar is vital if we want to be listener/reader friendly.
It is difficult to express complex abstract meanings without grammar. One day I was playing in the garden with my two-year-old grandson. He was filling a bottle with water from a tap and pouring the water in a hole he had dug. I was thirsty and asked him for the bottle. When he gave it to me I drank some of the water. He was horrified. 'Grandpa,' he said, 'that water not for drinking. It for putting in a hole.' He used the form for + ___ing to express purpose, and it is difficult to see how he could have got his message across efficiently without that complex little bit of grammar.
So, we can make meanings without grammar, but if we want to express meanings in an efficient listener/reader friendly manner we need not only lexis and structure. We also need a grammar that identifies things clearly: articles and determiners. We need a grammar that places things in a temporal setting: a tense system. And we need a grammar is capable of expressing abstract relations, phrases like for + ___ing to express purpose.
Michael Halliday (1975) published a book describing how his son, Nigel, acquired his first language. Normally we think of children as learning to talk, but the title of Halliday's book is Learning How to Mean. The emphasis on meaning is important. Halliday consistently sees language as a system of meaning (see, for example, Halliday, 1978; 1994). Words, or, in Halliday's terms, wordings, are a means to an end. The important thing is the acquisition and development of the capacity to make meanings.
All this may or may not be interesting, but what does it have to do with classroom language teaching? Well, it seems to me to demonstrate fairly clearly that meaning is lexically based. We can have a meaning system, a language, with relatively little grammar. If, however, that system is to develop into a system which is listener/reader friendly and which is capable of expressing complex abstract relations then we need grammar. But the important thing is to start with lexis and to go on to build a grammar which will become gradually more considerate of other language users and which will be able to express more and more complex ideas. What we need is lexis which becomes more and more grammaticised.
What often happens in classrooms is that language is seen as grammatically based. Teaching procedures are designed to teach the grammar, to teach learners to produce a range of grammatical sentences. At the same time learners acquire vocabulary and use this to fill out their grammatical sentences. This sounds like a perfectly reasonable way to proceed. The only problem with it is that it doesn't seem to work very well. Learners find it incredibly difficult to use the grammar they have learned. It is a very long time, for example, before they have control of do-questions or the distinction between the present simple and continuous tenses, even though these items are taught at a very early stage. They may be able to produce these items when they have plenty of time to think about them, when they are given a sentence level test, for example. But they do not use the items when they produce language spontaneously. Such items are not a part of their meaning system.
What would happen if we took a quite different approach in the classroom? Instead of beginning with the grammar we can begin by teaching words and phrases and encouraging learners to make the best use they can of these. In the early stages they will string these words and phrases together with a minimal grammar. As this happens we increase the demands on learners, requiring them to construct more complex meanings in a more listener/reader friendly way. As they are exposed to more and more language they will begin to construct and use a more complex grammar.
If we take this approach the job of the teacher is fourfold. We need:
What would this approach look like? It is easy enough to set out broad principles, but it is not so easy to define a methodology which realises those principles. One obvious caveat is that such a methodology would set a low premium on accuracy and, as a result, would produce students who could exchange meanings with fluency, but could not produce accurate sentences. But isn't this a healthy alternative to what happens in many, possibly most, classrooms at present--an approach which produces learners who can produce accurate sentences but who cannot communicate with any fluency? It seems to me that an approach with a primary focus on meaning, if sensitively handled, would be nothing but beneficial. It would set healthy priorities. Instead of focusing on grammatical accuracy with meaning as a by-product, we would focus on meaning with grammatical development as a by-product. I shall set out, in my talks at JALT 2003 to set out a methodology of this kind, which takes meaning as a starting point and seeks to incorporate a focus on grammar as meaning develops. To paraphrase my two-year-old grandson "Language not for talking, it for meaning."
1 I am indebted to Andrew Wright for this striking example.
Halliday, M.A. (1975) Learning how to mean: Explorations in the development of language. London: Edward Arnold.
Halliday, M.A. (1978) Language as social semiotic: The social interpretation of language and meaning. London: Edward Arnold.
Halliday, M.A. (1994) An introduction to functional grammar. London: Edward Arnold.
Dave Willis has worked in ELT for almost forty years. He has worked as a teacher and teacher trainer in Ghana, Cyprus, Iran and Singapore as well as the UK. He recently retired from his job as a Senior Lecturer working on MA TEFL/TESL programmes at Birmingham University. He continues to write and run workshops and consultancies. He has written widely on discourse analysis, task-based learning and pedagogic language description. His latest book: Rules, Patterns and Words: Grammar and Lexis in English Language Teaching is to be published by Cambridge University Press later this year.