Freedom and Control in Language Teaching Materials

Vivian Cook 
Writing Home
SLA Home

V. Cook, Ealing Technical College: BAAL Seminar 1970. .In Ramsay Rutherford (ed.)

One of the crucial problems the coursewriter or for that matter the teacher has to face is the choice of a method to control the language he wishes the student to hear and to produce. On the one hand he may decide that all aspects of the language — the vocabulary, the syntax, the intonation — must be controlled; on the other hand he may wish to have no restrictions at all. This paper discusses some possible approaches to this problem; in particular, control of language by means of structural grading, control by means of situation, and complete absence of control. Finally, it suggests that another approach, control by topic, avoids some of the snags involved in the other three.

Perhaps the most usual form of control has been structural grading, which has been particularly developed by those interested in linguistics. In spite of its evident popularity in recent years, structural grading of teaching materials rests on two assumptions that have not usually been made explicit and which are potentially danger us. Before going into these assumptions, I’d like to outline a structural grading of English as a concrete illustration. The backbone of nearly all courses written in English as a Foreign Language has been the forms of the verb. I shall take five recent courses for beginners with which I am familiar as a teacher rather than as a writer; these are “First Things First”, “Success With English”, “English Fast”, ”Situational English” and “The Turners”. So far as the grading of the verb forms is concerned, these five courses show a startling amount of similarity. Setting aside other elements in the grading, the typical order in which they introduce them is as follows

1. “Be”; Present Tense. “He’s English”.

2. Present Continuous. “He’s playing”.

3. Present Simple. “He plays”.

4. “Be”; Past Tense “He was here”.

5. Past Simple. “He played”.

6. Present Perfect. “He has played”.

7. Past Continuous. “He was playing”.

8. Past Perfect. “He had played”.

“First Things First” and “Situational English” keep to this order exactly. “The Turners” varies it by introducing Present Simple before “Be” : Past Tense. “English Fast” differs by reserving Present Perfect till after “BE” Past Tense and Past Simple. “Success With English” has Present Perfect after Present Simple and Past Continuous after “BE” : Past Tense. There is then a high degree of unanimity among these courses over what verb forms should be taught and what order they should be taught in. Such similarities must show that the underlying linguistic descriptions are also similar; in fact all of them convert very easily into a rule which used to be part of the transformational generative grammar of English, the rule which rewrote AUX as compulsory Tense with optional (HAVE -EN) and (BE-ING).

It seems to me that the essence of structural grading lies in taking a linguistic description and breaking it up into a number of minimal syntactic elements. Each of these elements is introduced one by one and combined with Those that have gone before until the unity of the description has been restored and all the elements and combinations have been covered. The Auxiliary Constituent is broken up into minimal elements such as Past, Continuous and Perfect; then these are introduced one by one and then combined in different ways to get Past Continuous and Past Perfect and so on. The totality of the language is broken up by the grading into a number of little fragments which are added together on1~y~by one. The student masters the language fragment by fragment until he arrives at the whole. To sum up structural grading, it seems fair to characterise it as a cumulative process of adding syntactic element to syntactic element.

The main problem is to establish what this represents; what in ‘fact does a grading grade? If we look at a grading such as the one above, what are we looking at? What are these objects “Present Simple”, “Past Continuous”’? What does this order “Present Continuous’ , ‘Present Simple” and so on stand for? These entities have in fact a highly dubious existence so far as the language teacher is concerned; the adoption of a structural grading commits the teacher and course-writer to some quite specific assumptions about language learning and psycholinguistics.

To return to the point that structure’ grading is a cumulative process of adding syntactic element to syntactic element, of building up the language piece by piece. We add “Present” to “Perfect”; we add “Past” to “Continuous” : this leads ultimately to such forms as the Past Perfect Continuous, “I had been playing”. If the grading has been properly designed and our course goes on long enough, we hope to have covered the most important parts of the language bit by bit, rather like solving a jigsaw puzzle piece by piece.

But to adopt this as a method of controlling language implies that this must be how language is learned. Our control has the purpose of building up the student’s knowledge of the language there is no point in adding piece to piece if this is not the way in which the student learns the language. His knowledge is also built up piece by piece like a jigsaw puzzle. Yet it is only an assumption that language learning proceeds in this jigsaw puzzle manner. There is at present little evidence to confirm this assumption. It is true that the success of structurally graded courses provides some evidence that people can learn a language in this way if they have to this does not tell us whether it is the most effective way of language learning. On the other hand there is little evidence that suggests that a first language is acquired in this way. A child does not seem to acquire his first language by adding A to B to C . . . to Z; instead his grammatical rules and the elements in his grammar seem to be in a constant state of flux. The assumption behind structural grading, viz, that language learning should be ordered in a series of ‘discrete syntactic steps, is at least questionable; one might well like to try out other more plausible assumptions.

The first assumption concerned the nature of language learning. The second assumption concerns the status of linguistic description. We have seen that the grading of the English verb forms involved a description in terms of compulsory Tense and optional elements (HAVE-EN) and (BE-ING). Presumably this is a description of competence, of the native speaker’s underlying knowledge of the language. So far as I know, this is true of all courses that make use of linguistic descriptions; the descriptions they utilise are founded on competence rather than performance.

Yet the paradox of language teaching and indeed of much linguistics is that competence is only accessible through performance. The student hears a certain number of sentences in the classroom; he has to produce a certain number himself. Except through the medium of grammatical explanation, little attempt can be made to influence his competence directly. His route to competence must lie chiefly through performance. Thus it would seem that a preferable structural grading would be based on a description of performance, of the ways in which native speakers encode and decode utterances, rather than on a description of their underlying competence. It is of course unrealistic to expect any firm guidelines for structural grading to emerge from the psycholinguistic study of performance in the near future.

So what we are left with is a grading in terms of competence acting as a grading in terms of performance. In other words a structural grading based on a linguistic description of competence reflects a particular hypothesis about ‘the relationship of competence to performance. This hypothesis in its strongest form states the linguistic grammar of competence corresponds, item for item and rule for rule, to the psychological grammar of performance. Our structural grading says not just that Present Perfect is linguistically valid but that it also corresponds to a psychological reality. Structural grading must imply that there is a certain similarity between syntactic rules and units and psychological processes and categories. This hypothesis is only one of the many hypotheses that one can entertain about the relation of competence to performance and to me at any rate it is one of the least plausible.

Connected with this assumption that competence equals performance is another assumption that is hard to disentangle. This rests on a confusion between decoding and code-breaking. A person decoding a message already knows the code and simply has to apply it to the message he has received his goal is the meaning of the message, not the code. A person who is breaking the code of a message does not know the code and tries to discover it from properties of the message; his real goal is the possession of the code rather than the meaning of a particular message. The adult native speaker decodes a particular sentence; the child acquiring his first language or the foreigner learning his second code-breaks the sentence. The two processes need not be identical; the steps which are followed in decoding a message need have no connection with the steps involved in breaking its code.

So a model of competence or of performance is not a model of language acquisition; a description of the native speaker’s ability to comprehend or produce the Present Perfect does not tell us how this ability was acquired. When we reach the utopia of explanatory adequacy, it is true that the grammar of competence will have some built-in concessions to language acquisition. At the moment, however, the descriptions of competence and performance do not necessarily reflect any of the steps by which they were learned. Yet a structural grading must to some extent suggest that steps in the grading reflect steps in the learning of the language. The steps in the structural grading represent progressive addition of separate elements of the total description of competence. Whether structural grading assumes that these steps correspond to stages in decoding is almost beside the point : it certainly assumes that they correspond to stages in the code- breaking process, that Present Perfect is a valid unit involved in acquiring the language. Thus the assumption is that elements into which a’ linguistic description can be decomposed correspond to stages in the acquisition of the language. Again there are alternative assumptions that one can hold about the units of code-breaking.

A coursewriter must then consider what assumptions structural grading commits him to. This is important because of the great responsibility that structural grading involves for the course-writer. He is saying not just that learning a language proceeds in a series of clearly defined steps, but also that he knows just what these steps are and can structure his teaching material accordingly.

Another method of controlling the language is control by situation. Some of those who have used the kind of argument I have presented so far, have gone on to suggest that a valid alternative to control by structural grading is. control by situation, which does not involve any dangerous hypotheses about language learning. Although this may be true, control of language by situation has severe limitations of its own.

First it’s necessary to distinguish two uses of the term “situation” that are current in language teaching. Neither of them clearly corresponds to the senses of the words “situation” and “context” in linguistics, diverse as these are in themselves. The first use of the term “situation” is admirably illustrated by the course “Situational English”. In this book “Situation” chiefly refers to objects and events that are visible in the student’ s surroundings situational. teaching tries to link each piece of language with something that is not language. The teaching technique is similar to ostensive definition in semantics: I hold up a pen and say “This is a pen”; this provides a clearly visible object and an appropriate action to accompany my utterance. In “Situational English”, the teacher needs to assemble a large kit of objects for each lesson and then has to go through a carefully designed sequence of actions with these objects, In Unit Ten, for example, the teacher needs to collect “Up to five each of the following objects : books, cups, cigarettes, pegs, shoes, bags, tins, pens, pencils, bottles, balls, keys.” After he has staggered into the classroom with this load, he demonstrates “This is a book” and “That is a pen” using objects near to him and far away from him. The use of “situational” to describe this type of teaching refers then to a particular teaching technique involving physical demonstration of actions and objects. When the supply of actions and objects in the classroom runs out, the teacher goes on to pictures, e.g. in Unit 10 he uses pictures or blackboard sketches of birds, broons, flowers and cars.

Obviously the language that can be used in material based on this method is extremely limited. In fact the only language that can be used is that which can be related to the physical environment, whether in the flesh or through pictures. The language is restricted to that which can be demonstrated in a classroom, and, although the number of objects that can be introduced is virtually unlimited, the number of actions that can be shown is much fewer, except by using filmed material, This type of control cuts down on the total area of the language that can be taught. A large part of language cannot be taught if this method is used rigidly : much language is not demonstrable in physical terms restraining language by situation in this sense cannot cope with more than a certain part of what we want the student to know, regardless of its advantages so far as teaching theory is concerned.

The other use of the word “situation” in language teaching circles sometimes overlaps with this first use and sometimes conflicts with it. This sense refers to the language which is likely to occur within a specific context of situation. That is to say, given that Mr. Herbert is a taxidriver, that Miss James is a learner driver, that they have crashed into each other at Hyde Park Corner on a wet August Bank Holiday, what are they likely to say to each other? In this use of “situation” the language is controlled by restricting it to those items that are likely to occur in a given set of common circumstances in another of the five courses, “First Things First”, situational teaching is defined as “teaching a language by presenting a series of everyday situations”. Typical situations used in these courses are “Jillian and Martin in the Supermarket”, “Going to the Post Office” and “The Shoe Shop”.

So, in this second use, language is controlled by features of the context of situation that may not be physically demonstrable. The attempt is made to give language that occurs naturally in a particular situation without demanding that ii. should refer to the physical environment of the speakers. As David Reibel puts it, the dialogues “must be contextually located, that is, be specific to a particular situation calling for the use of just those sentences.” The difficulty with controlling language in this way is not that it makes any special psychological assumptions but rather that it makes a linguistic assumption that an important part of language is controlled by situation. If we take the example of “Going to the hospital/ doctor/ dentist” that crops up in one form or another in most of these five courses, we find that there is some control of language by situation : the patient is always asked his name; either he is given an appointment to see the doctor later or he waits till the doctor is free; the patient then states his symptoms and the doctor performs the necessary treatment, whether taking a piece of metal from his eye with a magnet (“The Turners”) or taking blood from the patient’s arm (“Success with English”). Some part of the language does seem tied to the particular context of situation.

Yet it seems unlikely that this can be true of any but a few situations. One can say that in the doctor’s, the chemist's, the si4ermarket, one is more likely to use certain lexical items and grammatical forms than others: it is improbable that I will go into a chemists and say “Give me some fillet steak, please”. But once one has gone beyond the simple service situations involving a client/patient/customer, there are few situations where we can say very definitely what language is likely to occur. For the most part language is controlled by situation in a very tenuous way. Even within the circumscribed situations mentioned above there is still a great amount of freedom in what one can say; for instance, who could predict that the “Success with English” piece “Jillian at the hospital” should introduce the message which she fondly writes on the bottle of blood she has donated “With love from Jillian”? Control by situation is not very effective because in a large number of cases it can tell the coursewriter very little indeed about what language to include and what to reject.

Other deficiencies of situation as a means of limiting language emerge when we contrast the two uses of situation. Most situational teaching in the first sense implies highly unsituational language in the second sense. There is no set of real life circumstances in which it is likely that I would raise up my pen and announce “This is my pen” except for the classroom. I am certain that people would look at me very oddly if my meal—time conversation consisted of “This is a pork chop”, “I am eating it”, “These are strawberries”, “They are red”. Teaching by demonstration involves a highly specialised context of situation that may have few counterparts outside the classroom.

Turning the tables, one discovers that very often situational teaching in the second sense is not situational in the first. This is not a necessity but a consequence of the teaching techniques usually involved. For the student starts not by participating in the situation but by observing it. The student hears a tape, he reads a dialogue, he looks at some pictures he is sitting outside ‘the situation looking in. Gradually after stages of repetition, explanation and so on, he is asked to re-enact a similar situation in the classroom. Thus the chief feature of this technique is, starting at a distance from the situation, the student gets closer and closer to it without ever actually getting into it.. This is the opposite to situational teaching in the first sense which starts from the physical situation in the classroom in which the student is participating and gradually gets more removed from it. Thus most teaching that involves situation in this second sense of everyday context of situation offends against the real context of situation in the classroom where twenty people are sitting facing a blackboard. Its basic fault is probably that it does not allow the student to express his own needs and wishes but continually forces him into a role that is not really his.

It seems to me to neglect the creative aspect of language use; this particular point I have developed elsewhere. So control of language by situation in both senses cannot be totally effective because it cannot be kept up long enough. The language that can be controlled by physical demonstration is limited; the language that can be controlled by context of situation is also limited and the techniques involved in teaching it usually themselves involve an .unreal situation in which the student is only an observer. While control, by situation is adequate up to a certain point, it has little to offer the course-writer beyond this point.

So far my arguments have probably sounded as if I am in favour of complete abandonment of control, of the student hearing and producing anything with no restrictions. This is not really the case and I should now like to review briefly the arguments in favour of complete freedom of language. They seem to me to amount to two the appeal to analogy, and the confession of ignorance.

The appeal to analogy goes like this. The child acquiring his first language hears a virtually unrestricted input of language; he hears a range of syntactic structures and lexical items that differs little from the language in general. The exceptions to this may be the curious phenomenon of baby—talk and certain statistical properties suggested by recent research, e.g. that a much higher proportion of questions occurs when we are addressing babies than when we are talking to adults. Even if one admits certain restrictions on the language the baby hears, these restrictions do not appear at all similar to those on the language which has been taught to the foreign student. Of course, the child’s output has the appearance of tight control, only allowing certain forms and cutting down on the length of sentences in a systematic way; however, this control is exercised by the child himself and is not produced by interference from without. Thus by analogy with first language acquisition the language used in teaching a second language should be uncontrolled except by the internal limitations imposed by the learner himself.

The other argument in favour of complete freedom is the confession of ignorance. This asserts quite truly that second language learning is a mysterious process. We do not know what goes on in the mind of a native speaker decoding speech; we do not know what perceptual strategies go on in the mind of the learner breaking the code of speech; we do not know to what extent the controls we exercise over language correspond to valid stages in the learning process. On every side our ignorance is boundless. So, being thus ignorant and humble, the best we can do is refrain from interfering with this unknown process. We cannot be certain that any method of control is valid; if we use none at all we are at least certain that any hindrance to the learner is not coming from us. Any form of control may throw a spanner into the works of the machine. The language should be uncontrolled since none of our control methods can be shown to work and since the only method of language learning that can be guaranteed to work, first language acquisition, also uses an unrestricted input. The danger with this is that complete freedom of language is to some extent an abdication of responsibility by the teacher. On every side we are hedged with problems to which we have no clear solutions. We cannot, however, say to our students “I’m afraid I don’t know how to teach you a foreign language. Come back in twenty years and I’ll know a bit more”. Instead we have to get on with the day to day teaching of languages; teachers and coursewriters have to come to some decision in the meantime, however misguided it may be.

Granted that some control may be necessary to salve the teacher’s conscience, let us look at some general considerations about control and freedom of language. The basic justification for some- kind of control is that the whole of the language cannot be presented and learned simultaneously. Even in the case of an uncontrolled approach the language occurs in some order, however arbitrary it may be. To borrow a term from Newmark, the language used in the teaching situation occurs in “chunks”. In control by structural grading the size and order of the chunks are determined by the’ syntax of the language Present Simple is one chunk and it occurs before or after the other chunk Present Continuous. In control by situation, the size and order of the chunks are decided by the situations “A Visit to the Doctor” is a chunk and it occurs before or after “The Greengrocers”. Chunks also occur in the completely uncontrolled approach their size is determined by the length of the teaching period and the order between the chunks is random. So in looking for a method of controlling language, we are trying to find the most useful kind and size of chunk and the best order in which to place these chunks. Obviously the ideal chunk would avoid the restrictions we have seen in the cases of syntactic and situational chunks; that is to say, this chunk would not involve sweeping assumptions about language learning and would continue to act as a method of control for as long as we wished.

There are probably several solutions to this. One that can be mentioned briefly is control by means of topic. This has close parallels with the work of the Schools Council, in particular the Child Language Survey and the Humanities Project. Control of language by topic means simply that nothing is restricted but the field of discourse. Everything which can be said about a particular topic is allowed; any syntactic or situational limitations will be imposed solely by the topic. Thus if the topic is “Pop Music Today”, it will require expressions such’ as “supergroup”, “white blues” and “heavy”. So the language is controlled, chiefly of course in- the field of lexis. However, this control does not imply any attitude towards the ways in which people learn languages and it does not stifle the student’s own need to communicate.

If we decide that topics are a reasonable method of chunking language, we have still to decide the size of the chunks and their order. Some help for this is provided by work in the Child Language Survey, notably Publication No. 44. Control by topic does not, however, tie the teacher down to any particular teaching method. He can still, if he wants, use the conventional methods of drill and dialogue repetition. Nevertheless the type of material I feel ideally suited to topical control would be something like that seen in the Schools Council Humanities Project on education with its wide range of written texts, tape-recordings, photographs and cartoons..

To compare control by topic with the other methods will bring out some of the advantages, Compared to structural grading, topical control does not involve assigning particular cognitive status to linguistic descriptions and it remains neutral about the relationships of competence to performance. Compared to situational control it continues to be effective long after a situational ordering has petered out and it can make the student participate in the situation rather than observe it. Compared to a completely free approach, it can break the language up into chunks ‘that are under the teacher’s control and it can ensure a more complete coverage of the language. Topical control also has the great advantage that it combines some degree of control over the language with almost complete freedom for the student to express himself. Topics involve the student and give him something to talk about rather than limiting him solely to a kind of restricted code.

Topical control is, however, a very tentative suggestion which I have not myself thoroughly tested in the classroom as yet. Regardless of the value that topical control may have, I do feel that it is time that methods of controlling language were examined more closely by coursewriter and teacher. The conventional methods of structural grading and situational restrictions may still have some life left in them, once we know what they commit us to; they may perhaps be re-vivified and given a solider foundation. Myself, I doubt this and suspect that, if we are to insist on controlling the language used in the teaching situation, other methods of control will have to be found.



Alexander, L.G., “First Things First”, Longman, 1967.

British Council, “The Turners”, Longmont, 1969.

Broughton, G., “Success with English Coursebook 1”, Penguins, 1968.

Commonwealth Office of Education, “Situational English”, Longman, 1968.

Cook, V.J., “The Creative Use of Language”, AVLJ Vol 8 No. 1, 1970.

Newmark, L., “How not to interfere with language learning”, in E.W. Najam (ed.) “Language Learning; The Individual and the Process”, Mouton, 1966.

Reibel, D.A., “The Contextually-Patterned Use of English”, English Language Teaching, Vol. XIX. No.2, 1965.

Rutherford, R.W., M.E.A. Freeth and E.S. Mercer, “Topics of Conversation in the Speech of Fifteen—year—old Children”, Nuffield Foreign Languages Teaching Materials Project Occasional Paper No. 44, 1970.

Schools Council, “The Humanities Project; Education”, Heinemann, 197O

Wakeman, A., “English Fast : Book 1”, Rupert Hart-Davis, 1967.