A functional view of language teaching

David Willis
University of Birmingham
Key words: lexis, functional grammar, teaching methodology

1. Some findings of SLA research

Second language acquisition research is still a relatively new area of study, with
a ‘history’ of some thirty years. In that short time I think it would be fair to say
that SLA has not found any convincing answers to the question of how we learn
languages - but it has shed considerable light on the question of how we do not
learn languages. It is clear that:

1 There is no direct and predictable relationship between teaching and learning.

Elements like do-questions and the terminal ‘s’ with the third person present
simple are notoriously resistant to teaching. This is familiar to all of us from our own
teaching experience. But the phenomenon was not, perhaps, so obvious before it
was identified by researchers like Corder and Selinker in the early seventies.
We are also aware that:

2 Some aspects of language seem to be known and not known at the same time .

Learners can produce, let us say, relative clauses like:

That’s the man who robbed me.

If they have time to think things through, perhaps when they are working on
a grammar test. But when they produce language spontaneously they show a
marked preference for:

*That’s the man who he robbed me.

They have control of the correct form when they are thinking consciously
about language form, but they fail to apply this knowledge in spontaneous use.
Krashen (1982) put this down to the fact that learners operate not one, but two
language systems – the learned system, which represents the language they have
worked at consciously, and the acquired system, representing the language that
they have picked up in the course of exposure and use.
This may be a useful insight, but it is not an explanation – rather a
restatement of the problem. It does, however, highlight another interesting and
important phenomenon:

3 Learners pick up language they have not been taught.

It is fortunate that a lot of lexis and grammar are simply assimilated. If it

were not so it would be almost impossible to achieve even a modest competence
in a second language. There are two reasons for this. First, there is so much to
learn. Even a modest lexicon will run to thousands of words. Grammatical
systems are complex and multifarious. Classroom procedures can offer no more

A functional view of language teaching 85

than hints for learners to build on. Secondly there is a great deal about language
that is simply too subtle to be taught and learned consciously. For example
Hughes and McCarthy (1998) show how the generally accepted pedagogic rule
‘that the past perfect tense is used for an event that happened in a past time
before another past time …’ enables learners to make well-formed sentences
such as ‘I spoke to Lisa Knox yesterday for the first time. I had met her 10 years ago
but had not spoken to her.’ But as Hughes and McCarthy go on to point out this
rule does not show ‘that the two sentences would be equally well formed if the
second were in the past simple’, although the emphasis would be different.
What Hughes and McCarthy do not show is that a careful application of the rule
would lead learners to produce some forms like I opened the door when the
postman had knocked, which are distinctly odd, if not ungrammatical.
These findings of SLA research make pretty bleak reading for language
teachers. They suggest that learning develops by processes not accessible to
teaching. We can neither control nor predict what is to be learned, and there is
a great deal about language that we cannot begin to teach because it is not
adequately described. But in spite of this the research suggests that instruction
does contribute to learning (see, for example, Long 1988). It may not have a
direct and predictable impact, but it does seem to make learning more efficient.
We are left with two major questions. Why does learning progress in a manner
which seems to be independent of teacher input and learner directed efforts? If
instruction does not have a direct effect on learning what sort of effect does it
have? How is it beneficial?

2. Language as a meaning system

What happens when young children are learning their first language? It seems
that in the early stages of learning their language is largely lexical. It consists of
strings of words put together with minimal syntax. Here is an example. My
granddaughter, Lana, told her first story at the age of about three. Lana lives in
south London. Just round the corner from her house there is a police training
centre which, among other things trains police horses. Lana’s father, William, is
a keen gardener. Every so often he goes round to the police training school to
collect manure for his garden. This is Lana’s story:

Lana daddy horsey.

Man up.
Horsey poo.
Daddy poo bag.
Lana daddy home.
86 David Willis
At first sight this story seems to be completely non-grammatical. But it does
have a very basic grammar.

We interpret Lana’s first utterance as meaning that Lana and daddy did
something to a horsey. If Lana had said Horsey daddy Lana we would have
interpreted this quite differently. So Lana recognises that the structure of the
English clause is SVC (subject + verb + complement) taking a complement to
stand for any final element. But in order to understand Lana’s story we need to
know a good deal about the context. This is because Lana does not have control
of the grammar of orientation (see Willis 2003). It is the grammar of orientation which enables us to
relate our message precisely to the context, and to relate the elements of
the message to one another.

If, for example, Lana had begun her story with:

The other day daddy and I saw a horse.

This would have told us a lot about the context. The adverbial the other day
and the past tense form saw tell us that she is introducing an anecdote. It is a first
person story marked by the use of the pronoun I. Her third utterance is horsey
poo. We assume that it is the same horsey, but Lana has done nothing to help us
towards this assumption. She has not said: The horsey did a poo. So Lana does
not have the grammatical wherewithal to make her message clear, to relate it to
the context of telling and to relate one element of the story to another – the first
mention of horsey to the second mention. The grammar of orientation, involving,
among other things, adverbials, the tense system and the determiner system, is
highly complex and takes a long time to assimilate. For the time being, however,
Lana had a workable meaning system without worrying about further niceties.
Lana’s language is almost entirely structured lexis. It is a string of words put
together according to the rules of English clause structure and little more than
that. It is tempting to call Lana’s language ungrammatical, but it would be more
accurate perhaps to refer to it as ungrammaticised. If we look at the
way learners develop a foreign or a second language it seems to follow along the same lines. The
elementary learner’s language is characterised by:

Omission of articles.
Omission of BE.
Questions and negatives marked lexically but not structurally.
Predominance of base form of the verb.

So the elementary learner produces utterances like Where you go? I like play
golf and so on. They have a viable meaning system, but it is one which makes
heavy demands on the listener.
A functional view of language teaching 87

In 1975 Michael Halliday published a book about his young son learning his first
language, English. We normally think of children as learning to speak or
learning to talk, but Halliday’s book was significantly entitled Learning How to
Mean. The acquisition of language is seen as the acquisition of a meaning
system. Long before they produce recognisable words young children are able
to make themselves understood, to signal hunger, pain, or delight. Soon they
begin to string words together in the way we saw with Lana’s story. As the child
develops intellectually, this places more demands on the language system.
Similarly as the child engages in more complex social relations the language
system grows to cope with the new demands. What we see is a developing
meaning system, which grows gradually more precise and complex.
Halliday sees language as functional. It exists to meet functional demands,
and it is shaped and developed according to those demands. There are three
metafunctions: ideational, textual and interpersonal. It is an oversimplification
but we can see the ideational function as being concerned with the
communication of a basic message. The textual function tailors the language to
make the message readily accessible to a given listener or reader. The
interpersonal function is concerned with the presentation of the self. How do
we want to appear to our interlocutors? Do we want to promote solidarity? Or
do we want to assert our difference? Do we want to be distant or intimate? Do
we want to be polite or aggressive?

We can think in terms of communicative priorities. The first priority is to get

our message across: to do what we want to get done with language. Lana’s story
was reasonably successful in achieving this. The second priority is to take
account of our listener or reader: to make the message readily accessible. Lana
had few options to draw on in this regard. She could make herself understood
to a sympathetic listener who had a close knowledge of the background and
context of her story, but her story would have been difficult for anyone else to
understand. In terms of the interpersonal Lana had a range of non-linguistic
devices – a grave demeanour indicating the weight of her story, and a winning
smile when this was recognised. But she had very little in the way of linguistic
devices. She would probably have told the story in pretty much the same words
to a variety of listeners in a variety of circumstances.

3. The tension between form and meaning

I referred to Lana’s story as ungrammaticised rather than ungrammatical, and

suggested that in the same way we could describe the language of elementary
learners as ungrammaticised. If we see learners as developing a meaning system
88 David Willis
this would help to explain some of the findings of SLA research which seem to
fly in the face of common sense. Why is there no predictable relationship
between teaching and learning? This may be because they are aiming at
different things. Teaching looks to expand the learner’s ability to produce
acceptable sentences in the target language, whereas the learner is aiming to
produce a more efficient meaning system. The way language develops, both the
first and subsequent languages, suggests that we are natural meaning makers
rather than natural sentence makers, a distinction that is developed at some
length by Brazil (1995: 2). If this is the case then it is not surprising that learners
have different priorities from teachers. Learners will take teaching input and
adapt it according to the priorities of the developing meaning system. At the
earliest stages they will accord a relatively low priority to the grammar of
orientation, and they will mark things like interrogatives lexically rather than
structurally. They will have little regard for elements like the terminal ‘s’ on the third person singular
of the present simple form which is largely redundant in terms of meaning.

If we are to take account of this tension between learning and meaning we

need to see things from the learner’s point of view. We need to work towards
the rapid development of an effective meaning system, and, at the same time,
provide the learner with incentives to refine that system in a way that entails a
progressively more sophisticated grammar.

4. The role of task-based methodology

Let us look briefly at a task-based teaching sequence and see how it might be
used to recognise the value of the meaning system while at the same time
encourage the development of the grammatical system.

Stage 1: A prediction task:

You are going to read a newspaper article about someone trying to robbing a
shop. Here are some ideas to help you with the story:

The characters: a shopkeeper; her two children; a man; an eight-year-old boy; the
police.

The setting: a corner shop just outside Manchester.

The props: a balaclava helmet; a plastic carrier bag; a pistol

What they said:

The shopkeeper: “As I gave him his change a man came in”
“I am not sure whether it was real or not”
“He threw a plastic carrier bag at me, pointed a gun at me and
told me to put everything in.”.
A functional view of language teaching 89

The police: “We are taking this very seriously, as we would any robbery
involving a firearm...”

Work in groups and guess what happened in the story? Compare your ideas with
others in your group. Try to include all the things shown above in your story.

As we shall see this is the first of a three-phase task cycle. We will call this
first phase the task phase. It involves students in an exchange of meanings as
they try to predict the story. The outcome of the activity is the story, but there
will be a lot of language used in working towards the story. Learners will be
obliged to improvise much of the language used. They are producing language
spontaneously and at times will be stretched beyond the language they can use
with confidence.

Stage 2: Preparation.

Once you have decided on your story, write down a few notes to help you tell your
story to the class. Do not write more than ten words.
Now get ready to tell your story.

Stage 3: Report
Representatives of the groups tell their stories to the class.

These phases of the task cycle are quite different. Students know that the
report phase will be, in a sense, a public performance. The spokesman for the
group will be talking to the class as a whole, not in the privacy of a small group
who are all working together. They have already decided on their story so they
have time to think about how the story will be worded. In other words they have
both a reason to think about form and also the time to do so. We considered
above a series of communicative priorities:

Basic message Concern for reader/listener Presentation of self

At the first phase of the task cycle the primary concern is with basic meaning.
There will be relatively little concern with the form of the message. This is
appropriate to the circumstances of the task. Learners are working in a small
group and are highly tolerant of one another’s language. They are building up
the story bit by bit, providing the contexts they go along. There is relatively little
need to string together an independently coherent story. In the preparation and
report phases the priority shifts to a concern for the listener and a concern with
the presentation of self. The class is a much more public setting than the group.
They will want to give a good account of themselves in speaking in this relatively
public setting. They will also need to call on the resources of the language to
make their story explicit to an audience which did not have access to the gradual
90 David Willis
development of the story. In order to do this the learners at Stage 2 will take
time to phrase their message carefully, moving towards what they believe to be
accurate in terms of English. Although this involves a focus on form I prefer to
think of it as a focus on language development. I see a focus on form as teacher
initiated and teacher led, while a focus on development is student initiated and
student led. At this stage of the task cycle students will be adapting their language
in ways which make sense to them, not in ways that are imposed on them by the
teacher. They will not be concerned only with accuracy. They will also want to
retain forms of the language that they can produce with speed and fluency.
We have here a teaching sequence which is in line with natural developmental processes.
At the first stage, the task learners are encouraged to
make the best use of the language they already have – they are encouraged to
improvise. In the preparation stage they will experiment. They will pool their
knowledge and look for the best way to express their ideas. In the final, report
stage they will consolidate. They will adjust their language to meet the demands
of the new communicative context, drawing on their improvised performance
and incorporating into it as much as they reasonably can of the language
proposed during the experimental preparation stage. At this stage they will have
to make decisions as to how much new language they can incorporate and still
offer a fluent performance.

5. Grammatical development

If communication is primarily lexical learners will have ample incentive to

increase their lexical store. But what can we do to encourage grammatical
development? We showed in the previous section how a task-based sequence
can increase the demands on the learners’ meaning system in such a way as to
drive them towards a more syntactically complex formulation. We can also offer
a language focus to prepare the way for development. We saw in Section 1
above that it is not possible for learners to respond immediately to new input.
They cannot simply acquire new language forms and incorporate them at once
in their spontaneous output. Form focused work, then, should aim not at
immediate mastery, but at preparing the way for future development. This
suggests that we should put more emphasis on language exploration, on
consciousness raising, than has previously been the case. Learners should be
encouraged to familiarise themselves with the texts they have been exposed to
and to mind these texts for language that they can use to develop in the future.
Rather than relying predominantly on a presentation methodology, which
offers learners language which purports to be ready made to meet their needs
A functional view of language teaching 91

immediately we should encourage them to look critically at language and begin

to learn for themselves. Ideally we need to identify a range of teaching strategies
which will provide learners with guidance and encourage them to work out the
system for themselves.

References:
Brazil D. (1995). A Grammar of Spoken English. Oxford: Oxford University Press.
Halliday M.A.K. (1975). Learning how to mean: explorations in the development of language.
London: Edward Arnold.
Halliday M.A.K. (1978). Language as Social Semiotic: The Social Interpretation of Language
and Meaning. London: Edward Arnold.
Halliday M.A.K. (1994). An Introduction to Functional Grammar. London: Edward Arnold.
Hughes R. and M. McCarthy (1998). “From Sentence to Discourse: Discourse Grammar and
English Language Teaching”. In TESOL Quarterly 32/2.
Krashen S. (1982). Principles and Practice in Second Language Acquisition. New York:
Pergamon.
Long M. (1988). “Instructed interlanguage development”. In L. Beebe (ed.), Issues in Second
Language Acquisition: Multiple Perspectives. Newbury House
Willis D. (2003) Rules, Patterns and Words CUP

David Willis