When second languages were taught by the grammar/translation method, everyone seemed fairly clear about what should be taught and in what order. Verbs in all their complexity dominated the scene and conjugations and inflections were learnt by rote. The result was that language learners understood a good deal about how the target language was constructed but had little idea of how it was pronounced or used in ordinary conversation. The other drawback to this approach was that it took a very long time to master the new language system, so it was not practicable for learners with only a short time at their disposal.

Language learning methods have been bedevilled by pendulum swings and, sure enough, the grammar/translation method, patently unsuitable for the needs of servicemen in the Second World War, who needed to get out into the field and use the target language quickly, fell from favour. It gave way to a stimulus-response system, drawn from the theories of behavioural psychology, in which learners acquired a repertoire of responses to certain prompts that reflected the situations that they were likely to meet. This approach was especially suitable for the language laboratory, which allowed students to practise their drills as often as they wished. Grammar as such was not analysed in the language classroom; it entered the consciousness, if at all, by some mysterious osmosis. Among the drawbacks of this method was that students were not equipped to produce original utterances of their own. I remember a young lad I encountered who had learnt English by this method. Every time I asked him a question such as "do you drive?" or "do you take sugar in your coffee?" he would trot out the response he had learnt: "I do, but Jim doesn't". Needless to say, 'Jim' existed only in the drills he had used.

Clearly the method was not adequate for an in-depth acquisition of the target language and so, in a characteristic volte face the language gurus gave us the cognitive approach and grammar was once more centre stage. It was, however, a much more interactive form of grammar than we had seen in earlier classrooms. Now students were given patterns so that they could deduce the rules from themselves. The element of discovery somehow made language learning much more exciting and the grammar easier to assimilate than anything simply learnt by heart.

But all was not well. We observed that our diligent students could devise time lines to demonstrate the subtle differences between the simple past and present perfect tenses and create original, well-formed sentences of their own based on the rules, yet they couldn't produce language that was idiomatically correct. For some arcane reason, the only sentence that was ever produced to illustrate this lacuna was "excuse me, have you got fire?" when everyone knows it should be "have you got a light?". Grammar it seemed was about to be doomed again, as the heavy artillery of the communicative approach rolled in with alarming battle cries about correctness being less important than the effectiveness of the message. In other words, no matter that your student is content to tell you that you are a "good cooker" and that she is "very interesting by British history"—the main thing is that we know what she means.

Of course, there are always some students who insist on delving into the grammar anyway. I remember one lad who wrote down a sentence he heard every evening in his host family: "bung us a fag, mate". The student had analysed this utterance to his satisfaction having discovered that "bung" was used as a synonym for "pass" or "give" and that "fag" had various colloquial meanings, but in British English was a "cigarette". What puzzled him was the use of the plural 'us' because he couldn't understand why an ordinary person would use the 'royal we'.

Old hands like myself become increasingly cynical about each new trend. I accept that functional/notional syllabuses have their attraction, and at least they seem to accord grammar a significant but not over-dominant position in language learning. The trouble is that this approach is just too literal. Unfortunately native speakers do not use a finite number of language exponents for such functions as advising or complaining or expressing spatial relationships. One of the most daunting tasks for the teacher is to help students understand the pragmatics of language with its curious principles that reflect its deep-seated cultural and psychological aspects. English in particular is highly implicit, digressive even, and when people use it over literally they can seem aggressive or rude.

The 'natural approach' has shed useful light on the cognitive processes that come into play when we learn grammar, but it relies too heavily for my liking on comparisons with first language learning. We've had the 'total physical response', which I first thought had something to do with those feelings of fear and embarrassment that render you completely dumb when you have to utter something in the language you have only just begun to get to grips with. And we've had 'suggestopaedia' and 'communicative language learning' and, most puzzling of all in a field where we encourage people to speak, 'the silent way'. Recently the buzz word was 'lexical', so presumably students are all learning great chunks of vocabulary.

Yet when it comes down to it, the perfect method just doesn't exist and never will. Personally, I find that I resort to formal grammar teaching less and less. Just as I can use my computer without understanding how it works in any great detail, or drive my car without having the wherewithal to build an engine, so I find language learners can go far with little theory. After all, how many native speakers outside this profession do you know who can tell you how many uses there are for the present simple tense? I don't avoid grammar but I don't rely on it. I find it much more helpful to look at the situations students will find themselves in and then look at the types of language that will be suitable for them. I have given up course books because the authors have not met my students and so can't devise a course for them. I teach from my favourite resource books, from experience and from my own creativity. Grammar? Well, at the end of a session I round up the most important grammar points that have cropped up. I point my students towards an appropriate exercise or explanation and let them deal with it in their own way. How about you?