Structure Drills and the Language Learner

Vivian Cook 
Writing Home

CMLR, 38/2,  1982


The structure drill, or pattern practice, was largely inspired by the audio-lingual method and its concept of language learning as habit formation. However, although a long shadow has been cast over the audiolingual method by post-Chomskyan ideas about language learning, the technique of the structure drill still continues to be used even by those who would in no way consider themselves to be audiolingualists. This critique starts from the "classical" form of the drill, moves on to various developments from this, and considers some of the reasons for the survival of the technique. It is illustrated from books written for students of English as a Foreign Language, chiefly those published in England. The aim is to look at the structure drill, not in terms of preconceived notions of what methodologists have said about it, but in terms of what it means to the individual learner: it is an exercise in what might be termed "technique analysis" - exploring what a given teaching technique implies about the learner in the classroom.

A classic example of the structure drill can be found in English Pattern Practices.

Teacher:  I

Student:   I am going to study engineering later.

Teacher:  You

Student:   You are going to study engineering later.

Teacher:  She

Student:   She is going to study engineering later.

Teacher:  John

Student: ...........   1

The essence of the technique is that the students produce a series of sentences that use the same grammatical structure with slight variation. First the teacher provides some models; then the students have to produce sentences of their own in response to the clues that the teacher gives. Let us call the information that the students are given the input and the sentence they have to produce the output.1 In this example the input takes the form of a single word - "you," "I," "she," and so on - which the students substitute into the grammatical structure they are practising, varying the sentence slightly according to the person. The input may, however, consist of a full sentence, or of a picture or sound effect; in these cases the students have to work out for themselves which part of the input is the clue to their output. After the students have produced their output, a standard format for the drill, particularly in the language laboratory, is to let them hear the correct answer, so that they can be reinforced in the correct response. A drill consists of many of these pairs of input and output.

Let us look at this first in terms of the strategies the learner needs to use for processing language. One important aspect is the information process­ing that the learner has to perform; he has to extract certain information from the input he is given to construct his output. In the classic drill format shown this was blatant; the input consisted of the clue that he needed and nothing more; the student hears a pronoun and has to substitute it into his output. If the input consists of a whole sentence, one part of which is the clue, the information processing is more disguised but the ability to recognise and use the clue is still crucial.

Take the following, in which the student has to identify a national stereotype before he can produce his output:

George loves a good cup of tea.

He must be an Englishman then!

Mr. Evans goes to choir practice twice a week.

He must be a Welshman then!

Mr. Mackenzie never spends more than he has to.3

Here the student has to perform a deduction from the input to arrive at the clue for the output; this deduction is based on his knowledge of the alleged characteristics of people who live in different countries. In some drills this processing of the input to arrive at the clue can be highly sophisticated. Take the following example from a business course:

The management at Bellcrest are discussing what to do after the fire. Should we ask the bank for more money?


Are we obliged to meet existing delivery dates?


1. ask/more time/loan

2. explain/position/our customers

3. make/statement

4. look/security arrangements4

Here the students are still required to practise one structure throughout the drill and there is still only one correct output for each input. But the information for constructing the output is extremely complex, requiring substitution at two or three places in the sentence and depending on previous knowledge of the situation presupposed in the drill. All drills therefore rely on the student processing information to arrive at the sentence they are supposed to produce.

Does this have anything to do with the way that a native speaker con­structs sentences in a conversation? Obviously what we say depends on the situation and on what has just been said. In terms of situation we adapt what we are saying according to the several dimensions of register and style, depending on our social relationships with the people we are talking to, the kind of occasion we are speaking at, and so on. Any sentence we produce is dovetailed into the previous sentences of the conversation by means of various cohesive devices: if someone says "What tenor-player do you prefer?," I might answer "David Murray" stressing but if somebody said, "Sonny Murray plays the tenor," I might correct them by saying "David Murray," stressing the "David."5 Hence there must be some way in which the speaker stores relevant aspects of the situation and of the discourse that he can utilise in producing his own sentences. It may well be that drills are a valuable way of teaching the links that occur in discourse; indeed the links between input and output described in an earlier article are nearly all features of cohesion.6 But the way in which the information has to be extracted and used is usually rather different from a real life situation. The links between input and output are a device to get the student to pro­duce his own sentence rather than having a teaching point of their own. There is no conversational parallel to hearing "you" and incorporating it into a sentence "ask/more time/loan" and saying "No, I don't think so, but I do think we should ask for more time to repay the loan." The way in which the student has to process information in a drill is far from the normal way in which we process information in conversation.

If the drill is considered as an information processing activity, the pro­blem of capacity also arises. So far as the learner's capacity for processing language is concerned, this is likely to be low; after four years of French, one experiment found learners having difficulty with sentences that were more than four words long.7 The drill technique has to take account of this processing limitation and use sentences that the learner can actually handle. So far as the non-language information processing in a drill is concerned nobody knows the limits of the learner's capacity — or for that matter the native speaker's — in this kind of task. Certainly processes such as mental arithmetic are slower in a second language leading to the general idea that there is a cognitive deficit for operations carried out in a second language.8 Until the mental demands of the information processing required in drill have been investigated, one suspects that drill writers are in danger of overestimating what the learner can handle. Or the task of handling the information may be so onerous that the learner has no spare capacity for the actual teaching point of the drill. Hosenfeld, for instance, describes a learner who turned a written drill into a straight problem in information processing and paid no attention to anything in the input except the clues necessary to produce an output.9 The sheer technique of the drill therefore means that the student has to process information in a particular way; this may loom large in the mind of the student.

What does the drill imply in terms of communication strategies? The strategy used in the classic drill is essentially one that occurs within the classroom and bears no relationship to the world outside; nowhere but in the classroom could one person say "I" and the other choose to reply "I'm going to study engineering later." In this classic format the students are not practising anything resembling a real life exchange; there is no conversa­tional gambit in which one person says a pronoun and the next incorporates it into his or her utterance. The chief difference from the world outside is that the person who answers has no choice whatsoever about what to say; there is only one possible reply and only one way of expressing it. The students therefore never have the chance of choosing a strategy for themselves.

The classic form of the drill developed into contextualised drills such as the following:

A footballer is asking his doctor questions about how to keep fit. In this drill you have to decide whether various activities are good or bad for your health.

Do you think I should go for a run?

Oh, definitely; running's good for you.

Do you recommend a daily swim?

Oh, definitely; swimming's good for you.

Am I allowed the odd cigarette?

Oh, definitely not; smoking's bad for you.10

Here the student has to work out the -ing form to use in his or her output and has to decide whether the activity is good or bad for a footballer; there is a two-fold information processing task. But the drill is set within a con­versation; each pair of input and output is related as they would be in a con­versation and they are padded out with conversational tags such as "Oh" and "definitely"; the situation in which the participants are talking is also explained to the student. This goes some way to solving the problem with the classic drill that it is removed from real-life exchanges but there is still only one possible answer and only one way of expressing it: the student is rapidly aware that he is actually practising the relationship of nouns such as "swim" with -ing forms such as "swimming." The conversational context is the sugar that makes the medicine go down, but the structural teaching point is still the main rationale for the drill.

Even with the communicative functional teaching that became popular in the seventies we find that this was still true. An example can be taken from Starting Strategies,one of the first courses to embody these aims:

Ask the price of things

I think I'll have a cup of coffee.

How much is it?

I think I'll get some cakes.

How much are they?

Mmm. This milk is nice and cold.11

The ostensible teaching point of this drill is the notional/functional category of "asking the price" yet it is very close to the contextualised structure drill: there is an information processing task in which the student has to recognise singular or plural in the input and incorporate it as a pronoun in the output; there is a conversational relationship between each pair of input and output. It is still teaching one structural point and, what is more, has only one correct output for each input. Though the drill is now communicative, it is still within the drill paradigm.

However, if we start looking at drills in terms of communication, it turns out that we may have been overconcerned with making the student practice a single structural point. Each pair of input and output is a brief interaction between two people, a short conversational exchange. The student may be building, not habits of responding, but patterns of interaction: he is learning that if one person says, "1 think I'll get some cakes" a possible answer by another person is "How much are they?" Thus he is learning the possible moves open to him in conversation, the sequences of utterances between speakers that build up into conversation.12 The flaw in the classic drill was that the conversational exchanges of input and output were so strange; there is no conversational point to remarking "you" and no conceivable way in which "You're going to study engineering later" is a possible reply to it. Contextualising the drill by giving input and output a conversational link means that the student is learning a particular exchange. He is getting a model of how people behave in conversation, learning communication strategies that will be useful to him outside the classroom. The teaching point of the drill is how to respond to particular gambits in conversation. Whatever its virtues for drilling structures, the sameness of the outputs in a drill is a drawback rather than an advantage so far as conversational inter­action is concerned. The unnaturalness of the contextualised drill is that in real life, even if we have only one move that we can make in response to someone's remark, we can use a variety of structures that have the same functional effect. So, while the contextualised drill allows more natural conversational exchanges and a more realistic form of information processing, it does not confront the student with the problem of choosing what to say. If someone says to him, "I think I'll have a cup of coffee," he has to answer, "How much is it?" rather than, "How much does it cost?" or, "What's the price in here?" or, "What does the coffee cost?" or any of the alternative grammatical structures that could be used with the same functional impact.

What then are the reasons for this invariance of the structure of the student's output? One reason is the purely practical justification that a drill can be used for self-instruction in a language laboratory most easily when there is only one possible output and hence the tape can supply the right answer immediately; this is only relevant if you believe that students need immediate reinforcement and knowledge whether they are right or wrong. The main justification traditionally has been in terms of learning strategies. The learner was forming habits: the drill was not a meaningful activity but tried to build up automatic unthinking responses. The learner's strategy was to build up these habits mechanically by practice and repetition. However, this is far from the image of the learner that we have today. Learners do not passively accept what is presented to them; instead they organise it in their own ways. Nor do they learn the language in discrete chunks of native speaker competence, adding them together one at a time; rather they develop temporary systems of their own that evolve into the target com­petence. Recent research into second-language learning mostly stresses the development of interlanguages that are not the gradual accumulation of fragments of target competence but are systems in their own right: what is more, the students are believed to learn by actively engaging in communication, by expressing their own opinions and ideas rather than parroting those of others. This is at variance with the type of learning that is supposed to be taking place in a drill: our teaching practice and learning theories are out of step. Since teachers have to work in the real world and we know that, though they have their limitations, drills do actually work in the classroom, this does not mean that we should feel obliged to abandon drills instantly. However, it might be well to explore ways in which we can reconcile the assumptions built into the drill with the ideas about second-language learning current today.

One way of doing this would be to keep the drill as a brief conversational exchange but to allow a certain freedom of choice to the student. To explore this possibility it is necessary to look at another technique with certain similarities to the drill, the situational dialogue. This technique was first suggested by David Reibel as a way of getting teachers to write supplementary dialogues but it has become best known through a book called Situational Dialogues.11 This contains a series of situations for each of which there are four dialogues, usually having two complete exchanges between two people. The four dialogues are interchangeable; the opening remark from one could be the opening remark from any of the other three, and the reply again could be taken from any one. As an example the first dialogue from the situation On a Bus is as follows:

A: Does this bus go to the station?

B: No, you'll have to get off at the bank and take a 192.

A: Can you tell me where to get off?

B: It's the next stop but one.

The equivalent opening remarks for the other three dialogues are "Am I OK for St. Mary's Church?," "Do you go to the sea front?" and "Is this the right bus for the Town Hall?" Then there are three more alternatives for B's reply, "No, we only go as far as the park but you can walk from there," "No, you're going the wrong way. You want a 143 from the station," and "No, you should have caught a 12. Jump out at the bridge and get one there." The teacher can either get the students to follow through each of the four dialogues in turn, or he can get them to jump about, starting with the opening remark, say from Dialogue 2 "Do you go to the sea front?" and replying with the answer from Dialogue 4, "No, you should have caught a 12. Jump out at the bridge and get one there." The idea is that the students are practising remarks that are situationally equivalent but do not necessari­ly have the same grammatical structure. Like a drill the situational dialogue is a series of exchanges; it differs from the drill in that the exchanges are linked together and that there is a choice of grammatical structure. Of course, these two aspects are not necessarily combined. A book called Between You and Mestarts with sample situational dialogues and then reduces them to a formula such as:

A: Has X .... x .... the .... y .... yet?

B: No, she/he hasn't. He/she's .... z .... it/them now.

A: All right. I'll wait until he/she's finished.

The students are then given lists of words to substitute for x, y, and z; for instance x can be replaced by repaired, cooked, bought, given, washed, typed, peeled. Whichever item the students choose for x will determine which they can choose from the lists for y and z they are varying lexis but not grammar. Thus, like the structure drill, the situational dialogue requires the students to extract certain relevant information for producing their own sentences and controls the structure they can use within the pattern of a conversational exchange.

  The situational dialogue seems therefore to solve the deficiencies of the drill by emphasising the interaction of conversation and relaxing the gram­matical rigidity. Is this the end of the road? Unfortunately not, for this technique still fails to allow the students to choose what they themselves want to say in the classroom: it is still a meaningless activity. There are at least two ways in which situational dialogues are meaningless. One is that they are taken out of context; we know nothing about the town in which the bus dialogue takes place; we don't know the bus routes so we can't ourselves make up our minds whether the 192, the 143, or the 12 go to the station, the sea front, or anywhere else; we don't even know the number of the bus that A and B are on. In other words we don't have the background information that would allow the dialogue to make sense. The other way in which situational dialogues are meaningless is that they have nothing to do with what the student wants to say in the classroom. In the example from Between You and Me, A wants to do something but B points out that somebody else hasn't finished yet; put in the context of repairing televi­sions, peeling potatoes, and so on, this is meaningful but it is totally at variance with what is happening in the classroom; nobody wants to peel potatoes, repair televisions, and so on. The teacher may be able to contrive a situation in the classroom into which the dialogue can fit, say by cleaning the blackboard or borrowing pens or books, but this still does not spring from the students' own ideas. The situational dialogue relates adequately neither to a plausible external situation nor to the internal situation in the classroom.

So the situational dialogue only goes some way towards reconciling the drill with the learner's need to express his own feelings and ideas. Are there any other ways which might combine some of the advantages of the drill and the situational dialogue with the student's own self-expression? One possibility at the earlier stages of learning a language is the conversational exchange exercise. Take an example from a beginner's course: the students first see photos of different hotels, ranging from the hut village of the Club Mediterranee to the urban dignity of Brown's Hotel.15 They then have a checklist in which they have to specify for each hotel whether they think it is comfortable, modern, old-fashioned, expensive, etc. The purpose of this is to provide incidentally some new vocabulary, but essentially to record the students' own reactions to the hotels they see. Then comes the conversational exchange exercise proper:

Joe:       I like the Hilton because it looks modern.

Gloria:    Do you? I prefer the Forest Point because it looks old-fashioned.

Edna:     I like Brown's Hotel because it looks friendly.

Jenny:    Do you? I prefer the Hilton because it looks expensive.

The teacher plays these model exchanges to them from a tape or reads them from the book; the purpose is to give a model of how native speakers interact in conversation, illustrating an exchange using the functional/interactional categories of stating preferences, giving reasons, and reacting. Usually each of these categories has several different grammatical and lexical variations in the exercise. Next the students take over the exchange:

You:      I like . . . best because it looks . . .

Jenny:    Do you? I prefer . . . because it looks . . .

In pairs or groups they construct as many of these exchanges as the teacher feels right. They are therefore using their own opinions recorded in the checklist as the basis for their output; it's what they think of Brown's Hotel that matters. They are also having to choose from different grammatical or lexical items to express their opinion. Finally, unlike the drill, in a conversational exchange exercise they are producing both sides of the exchange. Another example in the same set of exercises goes on to choosing a hotel. They see a hotel list specifying the hotels in a town with their price range, and the extras available such as 'phone and garage. First they hear the model exchanges, teaching requesting, suggesting, and confirming:

Joe: I'd like a good hotel? What about the Red Lion?

Woman: Fine.

Man: I want a cheap hotel.

Woman: What about the Oxford Hotel?

Man: OK.

Man: I'd like a medium-priced hotel.

Woman: What about the . . .

Man: …

You: I'd like a . . .

Woman: …

You: …

They have been shown a model of two ways of requesting - "I'd like a good hotel" and "I want a good hotel" - and two ways of confirming -"Fine" and "OK"; they are also at liberty to use ways of requesting or confirming that they have learnt previously in the course. This is linked first to the requirements of fictional characters and then to their own needs in a hotel, the next exchange specifying whether they want single or double rooms, TV, phones, etc. The conversation exchange exercise preserves some of the control inherent in the drill in that the students are interacting within a set frame; it also preserves the advantage that the students are given a mini-model of how native speakers interact. From the situational dialogue it takes the freedom of choice of grammatical structures within the situation and the longer stretch of interaction between the speakers. However, unlike either of these, it relates what the student wants to say as far as possible to his or her own opinions and ideas; they are expressing themselves within the controlled format of the exchange.This article has then tried to look at the technique of the structure drill from the point of view of the learner. Of course, in many ways the learner's name has been taken in vain: we do not have the evidence that allows us to speak on his behalf. It may well be that a sounder knowledge of the learner's mental makeup, of the strategies he has available to him, and of the ways these interact with the situation will contradict particular points in the argument or suggest totally new ones. But the overall dilemma will surely remain with us: how do we reconcile the need for the learner to express himself and to learn the language in his own way with the constraints of the teaching situation and with the need to show him a model of the target language in operation? What this article has also tried to suggest is that it is time that we looked more closely at the actual techniques we are using in the classroom. It is easy to dismiss a technique because it conforms to a teaching or learning model that is out of fashion. However, these models have been limited in coverage and any actual teaching technique involves far more aspects of learning than they include. A structure drill, for instance, has often been presented solely as a method of habit formation, of getting the student to practice a structure mechanically a large number of times so that he can use it automatically. As we have seen, however, this is only one aspect of the drill: it also has implications for information processing and for conversational interaction. Indeed one might speculate that the success that some teachers have had with drills was precisely because of these aspects of the technique, rather than the ostensible rationale. The choice in language teaching is too often presented as a matter of global "methods"; it might be preferable to work upwards by looking at the implications of standard teaching techniques in terms of a broadly-based learning model and building up from this to an idea of an overall method. This technique analysis would contribute to a better understanding of the teaching opera­tion than a perpetual dogfight between proponents of all-inclusive global methods.

1. R. Lado and C.C. Fries, English Pattern Practices University of Michigan, 1958).

2. V.J. Cook, "Some Types of Oral Structure Drill," Language Learning, 18 (1968).

3.  B. Abbs, V. Cook, and M. Underwood, Realistic English Drills 3, 2nd ed. (Oxford University Press, 1979).

4.  English for Business (OUP, 1973).

5.  M.A.K. Halliday and R. Hasan, Cohesion in English (Longman, 1976).

6.  V.J. Cook, "Some Types."

7. V.J. Cook, "Aspects of Memory in Secondary School Language Learners," Interlanguage Studies Bulletin (Utrecht) 4, No. 2 (1979).

8. L.G. Marsh and R.H. Maki, "Efficiency of Operations in Bilinguals as a Function of Language," Memory and Cognition, 4, No. 4 (1976); J. Long and E. Harding-Esch, "Summary and Recall of Text in First and Second Languages; Some Factors Contributing to Performance Differences," in Proceedings of NATO symposium on Language, Interpretation and Communication, ed. H.W. Sinaiko and D. Gerver (Plenum Press, 1977).

9. C. Hosenfeld, "Learning About Language: Discovering Our Students' Strategies," Foreign Language Annals, 9, No. 2 (1976), 117-29.

10. B. Abbs, V. Cook, and M. Underwood,

11. B. Abbs and I. Freebairn, Starting Strategies (Longman, 1977).

12. V.J. Cook, "Interaction Sequences in Second Language Learning and Teaching," in Error Analysis, Contrastive Linguistics, and Second Language Learning, ed. G. Nickel and D. Nehls (Julius Groos: forthcoming).

13. D.A. Reibel, "The Contextually-Patterned Use of English: An Experiment in Dialogue-Writing," English Language Teaching, 19, No. 2 (1965); M. Ockenden, Situational Dialogues (Longman, 1972).

14. A. Maley and R.S. Newbury, Between You and Me (Nelson, 1974).

15. V.J. Cook, People and Places (Pergamon, 1980).