V.J.
Cook
AILA 1981 paper in
G. Nickel (ed.) Error Analysis, Contrastive Linguistics and Second Language Learning,
Julius Groos, 1982, 123-134
One of the most important aspects of conversation is people do not exchange remarks at random; they link them together. This can be seen in an interaction sequence such as a 'knock knock' joke.
"Knock knock"
"Who's there?"
"AILA"
"AILA who?"
"I'll explain if you open the door"
Here the opening move "knock knock" leads on to a question "Who's there?", the question to an answer, "AILA", the answer to a query, "AILA who?", the query to a punchline, "I'll explain if you open the door". While 'knock knock' jokes are structured and predictable in their sequencing, much ordinary conversation displays similar properties: we follow compliments with thanks; "What a nice dress", "Thank you"; questions with answers - "What's the capital of Sweden?", "I'm afraid I don't know"; and commands with acquiescence or refusal - "Never darken my doorstep again", "Too right I won't". The difference between conversations and 'knock knock' jokes is primarily the range of choice available at each point in the interaction.
The idea of interaction sequence has not been greatly studied in second language learning, with exceptions such as Hatch (1978); the field is still dominated by discussions of certain aspects of syntax. Pragmatic competence, though developing rapidly as an area of study in first language acquisition, is still little explored in second language learning. The strategies for interacting in conversation are part of the participant's pragmatic competence; he or she knows when to answer a question, when to return a compliment, how to acknowledge thanks. Perhaps the whole of pragmatic competence is transferred from the first language, perhaps some of it is universal to all languages, perhaps some or all of it has to be learnt afresh in a second language: at the moment we have insufficient evidence to say. The rest of this article develops two lines of thought in relating interaction sequences to second language learning; one is to look at some experiments, the other is to look at implications for language teaching.
The first important step is to establish that interaction sequences have some sort of psychological reality to the second language learner. To settle this one experiment, reported fully in Cook (1981), looked at a paradigm example of an interaction sequence, the adjacency pair. An adjacency pair is a sequence consisting of two linked turns, such as question and answer - "What's the price of petrol?", "£1.60 a gallon" - or request and acknowledgement - "A glass of milk please", "OK". Adult foreign learners of English were tested for their ability to complete adjacency pairs; the results were that after only five months of English they were able to do so with more than ninety per cent accuracy. This led to a supplementary experiment in which the effects of adjacency pairs on memory were tested; this suggested that information was remembered better within pairs than across pairs; in some ways adjacency pairs were organised together in long-term memory. Thus it seemed that adjacency pairs had psychological reality and indeed usefulness to the second language learner and from this one might generalise to interaction sequences.
However, we cannot look at interaction sequences without taking into account other aspects of interaction, in particular language functions. Both language teachers and linguists have, in the past, often treated language functions out of the context of situation in which they occur: linguists have looked at the conditions that a sentence must meet to fulfill a certain speech act, not at sequences of speech acts; applied linguists have sometimes drawn up lists of language functions that should be taught independent of sequencing or of situation. In both areas it is only comparatively recently that the relationship of language function and situation has been taken into account by Ferrara (1980) and Levinson (to appear) in the case of linguistics, Rintell (1979), and Walters (1979) in the case of applied linguistics. The problem is that, while there may be many different ways of making a particular language function such as a request, only a few of them will be right given a particular context of situation. Two aspects at least must be taken into account when describing the speaker's choice of realisation of a particular function, or the listener's interpretation of a sentence as having a particular function. One is the effect of linguistic context, in other words of interaction sequences inter alia. If someone says "A pint of beer" this might be a request if nothing had preceded it; it might be a correction if one person said "John likes a glass of wine", "A pint of beer"; it might be an answer to a question - "What's in that glass?", "A pint of beer". The way that the speaker realises a language function and the way that the listener interprets a language function from what he or she hears makes extensive use of the linguistic context (Clark,1971). The second aspect is that function also depends upon the non-linguistic situation. "A pint of beer" only counts as a request if it is said on the right sort of occasion to the right sort of person; going up to a stranger on the street and saying "A pint of beer" is unlikely to be treated as a request. Language functions have to be seen therefore within the interaction between the participants and the situation. The ways that the participants interact, in particular their strategies for realising and interpreting language functions, depend upon their shared knowledge of the context of situation, what Bunt (1977) calls their k-state.
But we still have very little idea how this takes place in native speakers; we do not know the typical realisations of functions that native speakers use in interaction except through intuition. An experiment was accordingly carried out to establish some of the variation in native speakers and foreigners according to two features of the situation - the sex and age of the person they were talking to. This is reported in full in Cook (to appear). The approach was a roleplay questionnaire in which the subjects were asked what they would say when making particular requests or thanks; the native group were British undergraduates, the foreign group postgraduate EFL students. So far as requesting was concerned the results were that native speakers used mostly question directives such as "Have you got the time?" and embedded imperatives such as "Could you tell me the way to the cinema?", to use Ervin-Tripp's terminology (Ervin-Tripp, 1977). For thanking the natives differentiated between the short form "thanks" which was used to young people and the long form "thank you" which was used to old people, but they did not vary their responses according to sex. Thus we have established that there is indeed variation in the realisation of thanking for age but there is no variation for requesting for either sex or age, nor of thanking for sex: age but not sex is perceived as a relevant situational factor by the participants. Turning to the responses by the foreigners, overall they were highly efficient at dealing with language functions: only 3 out of 88 responses for thanking, for example did not in some way achieve a recognisable function of thanking. This did not mean the responses were not on occasion highly ungrammatical, such as "I very gratitude you". Next the foreign learners did not differentiate for age or sex for requesting and distinguished only age for thanking. Thus they were recognising the same factor as relevant as the native speakers. Nevertheless they were doing so in ways that were different from the native speakers to a statistically significant extent; they did not use "thanks" to anybody very often, for instance. The conclusion from these results is that learners had a problem not so much with being able to express functions or with distinguishing relevant situational factors, as with the ways in which the variation in their realisations is achieved. They didn't need to learn thanking as such, nor that we are politer to older people, so much as when and how to say "thanks", "thank you" and "I'm very grateful to you". To come back to the idea of interaction sequences, they needed to know the ways in which they could adapt their realisation or interpretation of functions according to the context of situation. It was a question of the interaction between the functions they wanted to express and the features of the situation (Cook, 1982a). Put in another way what they lacked was the knowledge of the meaning potential at each point of the conversation out of which they could choose a particular realisation.
So it seems that the idea of interaction sequence might prove a useful tool in second language learning research and can be taken far further than the preliminary sketch given here. Let us now turn to the relationship between this and language teaching. Though it has rarely been carried out, it seems to me that applied linguists have a responsibility to actually apply their ideas to language teaching before attempting to advise teachers what they should do. The following is then based on work on the production of courses for teaching English as a Foreign Language carried out in the past few years. One application to language teaching is at the level of the syllabus, of spelling out what we are trying to teach. Mostly interaction sequences have not been taken into account in this area. One syllabus tried to outline different moves in an interaction sequence and then to teach the appropriate intonation patterns for each (Cook, 1979); for instance, a checking move is accompanied by a rising intonation whatever the syntactic structure involved; if we want to check a statement such as "There's a burglar in my room" we can say "What?", "There's a what in your room?", "Sorry?", "What did you say was in your room?" and so on, usually with different types of rising intonation. A later beginners course (Cook, 1980) tried a syllabus that was in some ways based on interaction sequences. Having specified the overall goals that the learner is aiming at,in this case travelling and talking in the classroom, we can try to establish what interaction sequences they require. If, for instance, they are using English for travelling they need to be able to take part in sequences such as:
"I'd like a hamburger please"
"Fine. That'll be 50p"
"Thanks"
in other words, requesting food, acknowledging, requesting payment, and thanking: they need to know not just the forms given here but also the meaning potential available to them at each point - that they can choose to thank from, say, a repertoire of "Thanks", "Thank you" and "I'm very grateful to you" according to features of the context of situation. Using English for talking in the classroom on the other hand requires different sequences such as:
"Bob Dylan sings fantastically"
"Does he?"
"Yes, he does. He's great"
in other words, stating opinion, reacting, replying and amplifying. Once we have isolated sequences such as these as being those the learner requires we can list the moves in the sequences, the realisations for each move, and the ways in which these realisations vary according to features in the situation. The syllabus will consist of a list of interactional categories such as "checking" and "acknowledging" related to realisations and to situations; unlike many lists of language functions, each category is related to the other categories through the sequencing of interaction. Obviously a syllabus such as this needs to be worked over very thoroughly before it can be made the actual teaching syllabus; we need probably to know something of the contrastive analysis of interaction sequences between the first and second languages and we need to know what are the real learning problems. But, imperfect as the list may be, it shows that it is possible to base our language teaching on the idea of conversational interaction rather than on a set of discrete language functions. Insofar as lists of language functions are required for teaching they need to be specified in terms of the context of situation - how their realisation and interpretation depends upon the sequencing of the discourse, on the relevant aspects of the participants and so on. In the words of Carroll (1981) 'the course of learning...starts with an event in which, in a given situation noted by the learner, the learner makes a conscious choice of a particular response, based on some kind of information available as to what response would be appropriate in that situation. But producing a syllabus is only a small step towards actual teaching: we need to convey the knowledge of these interactional categories, sequences, realisations, situational factors, and so on, to the students. Some people have seen no need to change their teaching methods because of the new emphasis on communication; instead of structure drills we have functional drills, instead of grammatical explanation, functional explanation, instead of dialogues exemplifying structures, dialogues exemplifying functions. While this may be a valid approach, it seems that with interaction sequences what we need is a method that emphasises the student's choice; he needs to be faced with the problem of choosing moves in an interaction sequence and of choosing how to realise them according to the situation; we need a technique in which the student chooses continually what he is going to say and how he is going to say it; we need to give him an adequate meaning potential from which to make his choice, a way of achieving his own goal in the conversation. One way of doing this is through the comparatively unstructured teaching techniques that have been developed in recent years, ranging from the humanistic alternative 'ways' such as Community Language Learning, to more mainstream methods such as roleplay and communication games. In many of these the learner is indeed genuinely engaging in interaction, genuinely trying to say something of his own and to tailor it to the circumstances in which he or she is involved. But, at beginners level, if not above, these techniques have the disadvantage that they do not provide the student with a model of how native speakers behave; in CLL for example the learners interact in their normal mother tongue way with each other, with the teacher simply providing a translation in the second language - the same type of activity as communicating through the Morse Code: the students never see a model of native speaker interaction. Communication games also depend on the student's past experiences of the language and never provide him with a model of how native speakers carry out the same activity.
In these teaching techniques the student is certainly having to make a choice from the meaning potential at different points in the interaction sequence, but he is not aware of the reasons why native speakers would make these choices. A further major disadvantage to these techniques in terms of the conventional language teaching situation is the lack of control over the student's language; the student is free to use whatever language functions, grammatical structures and so on he or she likes, drawn out of his own resources. To teach interaction sequences we need a technique that not only causes the students to interact in the classroom but also provides them with a model of the interaction, makes them continually choose things they themselves want to say out of all the possibilities open to them, and to some extent controls their actual behaviour.
Doubtless, there are many ways that can be devised to meet these requirements. One possibility at beginners level is the conversation exchange exercise used extensively in Cook (1980) and described in Cook (1982b). One example is that the students see pictures of people carrying out various activities with captions such as "Helen likes dancing", "Peter likes reading" and so on, giving them some basic vocabulary and structure. Then they have to complete a questionnaire requiring them to show whether each character likes dancing, sport and so on, and then whether they themselves, their friends, relations, and classmates do; this phase makes the students use the vocabulary actively in relation to their own opinions and experience. The conversation exchange proper starts with a model of two characters discussing their likes and dislikes :
"I like films"
"Do you?"
"I love dancing"
"Do you really?"
"I don 't like reading"
"Don't you?"
Then one of the characters fades out and the students have to supply the reacting move in the exchange:
"I like swimming"
"I don't like running"
Usually they are given a choice of grammatical or lexical realisations for each move, such as "Do you?", "Do you really?", "Don't you?" and so on; in other words they are not forced into a single correct answer as in a structure drill. Finally, the students take over both parts of the exchange. One student makes the starting move:
"I love badminton"
and another takes over the reacting move:
"Really?"
In pairs the students are constructing short interaction sequences. These are based on their own opinions, give them a choice of ways of realising their functions, and give them a short model of how native speakers create such interactions. A conversation exchange is therefore one possible technique for allowing the students to engage in interaction in a controlled way in the classroom, of facing them with the problem of choosing from the meaning potential to express their own opinions and ideas.
To conclude, this article has made some small explorations into the role of interaction sequences in second language and has suggested how these might be used in second language teaching; it is an attempt to bring together two aspects of applied linguistics that tended to go in different directions in the nineteen seventies, - language teaching and language learning research. What is reported here only starts to scratch the surface; true applications must depend upon broader-based research into this area where sociolinguistics, psycholinguistics, and applied linguistics merge.
REFERENCES
Bunt, N. (1977), Towards an analysis of discourse organisation principles, IPO annual progress report, 12, pp. 105-116.
Carroll, J.B. (1981), Conscious and automatic processes in language learning, Canadian Modern Language Review, 37/3, pp. 462-474.
Clark, H.H. (1971), Responding to indirect speech acts, Cognitive Psychology,11, pp. 430-477.
Cook, V.J. (1979), Using Intonation,Longman.
Cook, V.J. (1980), People and Places, Pergamon.
Cook, V.J. (1981) , Some uses for second language learning research. In: H. Winitz, ed., Proceedings of the New York Academy of Sciences Conference on First and Second Language Learning.
Cook, V.J. (1982a), Second language acquisition from an interactionist viewpoint, Interlanguage Studies Bulletin 6, pp. 93-111.
Cook, V.J. (1982b), Structure drills and the language learner, Canadian Modern Language Review, 38, pp. 321-329.
Cook, V.J. (to appear), Language functions in second language learning and teaching.
Ervin-Tripp, S. (1977), Wait for me, roller Skate! In: S. Ervin-Tripp and C. Mitchell-Kernan, eds., Child Discourse, Academic Press.
Ferrara, J. (1980), Appropriateness conditions for entire sequences of speech acts, Journal of Pragmatics 4, pp. 131-160.
Hatch, E., ed., (1978), Second Language Acquisition, Newbury House.
Levinson, S.C. (to appear), The essential inadequacies of speech act models of dialogues. In: Parret, H., Sbisa, M., and Verschuren, J., eds., Possibilities and Limitations of Pragmatics.
Rintell, E. (1979), Getting your speech act together, Working Papers in Bilingualism, 17, pp. 97-106.
Walters, J. (1979), Strategies for requesting in Spanish and English: structural similarities and pragmatic differences, Language Learning, 29/2, pp. 277-293.