A rather obvious point about any form of learning is that it consists of an interaction between the learner and the environment. Chomsky, for instance, claims 'knowledge of grammar, hence of language, develops in the child through the interplay of genetically determined principles and a course of experience' (Chomsky, 1980) . The problem comes, however, when we try to evaluate the relative importance of these two elements, the learner and the environment; Chomsky puts the emphasis on internal properties of the learner's mind; behaviourists concentrate on the environment. Nevertheless, neither side can entirely dispense with properties of either the learner or the situation. Both sides are committed to some form of interaction between properties of the learner and of the situation, however minimal.

This article, however, tries to recognise and make explicit the interactional nature of second language acquisition by exploring a position in which equal weight is given to both the learner and the situation. This position is known as interactionism. Put in terms of second language acquisition we have, on the one hand, the learner's contribution to learning, on the other, the learning situations in which he finds himself whether in the classroom or outside. The learner makes a diverse range of contributions to learning - motivation to learn the language, level of cognitive development, strategies for language learning and for communication and many others. A convenient neutral name for these is the learner's mental makeup - all the properties of the learner's mind that are relevant to second language learning. As well the learner finds himself in a diversity of situations; he may be in a formal classroom or a 'natural' learning situation; he may be in the role of student or of Immigrant or of foreign adviser; and so on. His learning depends upon the interaction between the different aspects of his mental makeup and the different aspects of the situation. If he expects a grammar-translation method of language teaching, he may learn poorly In a classroom where an audiolingual method is employed; if he is integratively motivated, he may have little success with an ESP programme; the combinations of makeup with situation are enormous. This is one of the main causes for the complexity and difficulty of carrying out second language learning research: the varieties of learner and varieties of situation. All we can do in a particular study is demonstrate an interaction between a particular type of learner and particular type of situation. This interactionist view has been implicit in much work in second language learning and is now starting to become explicit in articles such as McLaughlin (1980). Is it simply the obvious statement mentioned at the beginning, a simple truism? Or can it in fact lead somewhere? The rest of this article applies this central concept of giving equal weight to both sides of the interaction involved in second language learning by looking at three of the models that have been suggested for this area.

First, however, it is necessary to look at interactionism in more detail. In psychology this position was associated with Kurt Lewin’s classical formula “B=f(P, S)”; that is to say. Behaviour (B), function (f) of the Person (P) and the Situation (S) (Lewin, 1935). This formula provides a kind of mnemonic for remembering the essence of interactionism; any statement of behaviour is incomplete if it does not deal with both the relevant properties of the person and the relevant properties of the situation. This type of model has been revived in recent years chiefly within the psychology of the personality. for instance, it may be applied to personality traits such as anger: how do we define an angry person? A 'truly' angry person may only meet placid situations and hence never show his anger; alternatively a person given to anger may be aware of this and go out of their way to avoid anger-provoking situations. Hence, we can only talk of the behaviour anger as a combination of the person's trait with the situation; we can only talk of a person who is angry in such and such situations, never of a person who is, so to speak absolutely angry. Similar arguments have been applied by Argyle (1977) for instance, to the study of the treatment of mental patients with behaviour problems, and to racial prejudice. A reader who wants to see more detailed examples is referred to the collection of articles in Magnusson and Endler (1977).

Rather than give more isolated examples of interactionism, it is convenient to turn to a paper by Endler and Magnusson which specifies four main axioms of interactionism, (Endler and Magnusson, 1976). Let us take each of their axioms and try to relate it to second language learning.

1. Actual behaviour is a function of a continuous process or multi-directional interaction (feedback) between the Individual and the situation that he or she encounters.

The first axiom is a re-statement of the essential point made above that learning is a function of person and situation. It emphasises, however, the continuous nature of this process arid that it goes in both directions from learner to situation and situation to learner; not only are learners affected by situations but they also affect them. To take an example from error analysis, we can look at the following sentence produced by a learner of English, "Sorry please can you said at what time will be the next bus?" On the one hand this may be related to the learner - the strategies that he is attempting to use for making requests and asking a question; this might be traced back to aspects of his first language, which happened to be Kazakh. On the other hand, the sentence can be linked to the situation in which he was supposed to be making the remark, here a test that asked him what he would say if "you want to ask Helen the time of the next bus", and to such features of the situation as the degree of formality, the age and sex of the addressee, and so on. The learner's error may be as much in the appropriateness to the situation as in the grammatical and other strategies for carrying it out. Also the situation itself is changed by the learner's utterance since the addressee in the real world will realise that he is a foreigner and behave accordingly.

  2. The individual is an intentional active agent in this interaction process.

The individual has the power of choice; his behaviour is not determined by either his personality or the situation. He will try to act in the situation according to his goals; he prefers some situations or actively avoids others for various reasons. In terms of L2 learning this is the essence of strategies models such as Conversational Analysis (Hatch, 1978): the learner chooses what to do; he prefers one strategy to another. The interaction between situation and mental makeup still leaves the learner an individual with the unpredictability of any human being. Thus, in a foreign railway station, the learner may adopt a gesturing strategy or one of writing destinations down on a piece of paper or one of extra loud speech or mock foreign pronunciation of the destination; these are all possible in terms of his makeup or of the situation. He may In fact decide that in future he will travel by air and avoid the situation altogether. In the classroom too the learner has a choice; the teacher may be trying to encourage him to learn by unconscious habit formation but in the privacy of his own head the student can use granmar translation; sometimes he can opt out of the situation altogether, as witnessed by the high drop-out rate on some adult evening classes. Interactionism thus sees the individual as a free agent choosing to adopt various ways of behaving in the light of his goals and of the interaction between himself and the world.

  3. On the person side of the interaction cognitive factors are the essential determinants of behaviour, although emotional factors do play a role.

Endler and Magnusson claim that it is the person's cognitive factors that are dominant: they cite factors such as 'encoding strategies' and 'construction competencies'. This seems the axiom with which second language learning research is most at variance: while some research has pointed out the relevance of cognitive factors such as level of cognitive development (Tremaine, 1975; Felix, 1975), other research has emphasised attitude and emotion (Gardner and Lambert, 1972 ; Guiora, [Brannan, & Dull, 1977). Indeed, we might feel that even within interactionism the commitment to interaction between person and situation does not logically imply the priority of cognitive factors.

4. On the situation side, the psychological meaning of the situation for the Individual the important determining factor.

It is not the situation in itself that influences the individual's behaviour; It is the way that the individual interprets the situation. "Individuals may give different meanings to one and the same situation." (Magnusson, 1974). If an individual sees a situation as threatening, he reacts in a threatened way; if he sees himself as having an inferior role, he behaves accordingly: there is no objective situation independent of people's perceptions of it. So the relevant factors in the situation are those that the learner perceives as relevant. In the classroom for example we have to consider the learner's perception of the techniques we are using as well as our own; as Hosenfeld has shown, students may see the technique quite differently from the teacher (Hosenfeld, 1976). This point is also crucial to second language learning since we are dealing with a situation involving two cultures and languages. The same situational factors may be interpreted quite differently; an old person or a young person may be treated with respect in one culture, and with lack of respect in another.

Let us now turn after this brief account of interactionism to a more detailed consideration of three current models in second language research. interactionism can provide some independent yardstick by which we can measure these models; it does not, however, follow that any criticism of them from an interactionist viewpoint is necessarily valid if we reject interactionism. The three models are the Monitor Model, the Acculturation Model, and Conversational Analysis; each of these will be dealt with through one key reference in which their leading proponents have presented their positions.

1. The Monitor Model

The essence of the Monitor Model as summed up in Krashen (1981) is that some second language learners are able in certain circumstances to use a strategy called Monitoring in which they bring their ..luscious 'learned' knowledge of the second language to bear on their language use. This strategy is linked to aspects of the learner's mental makeup: it is claimed that children who have not achieved the formal operations level in Piagetan terms cannot Monitor; language aptitude is believed to be more important to learning than motivation; other factors lead to some learners being Monitor over-users, some optimum and under-users. But Monitoring also depends upon situational factors; one is that Monitoring can only be used when there is sufficient time; another that the task must concentrate upon 'form' rather than 'meaning'; a third that it shows up in a discrete test rather than an integrative test. Thus, on the one hand, the Monitor Model postulates certain aspects of mental makeup that either cause the learner to use Monitoring or stop him from using it; on the other hand it suggests that Monitoring depends upon factors in the situation. The learner's use of the Monitor, his actual behaviour, depends upon an interaction between the learner and the situation, clearly an interactionist position. Because of this, the Monitor Model brings us face to face with the problem of evidence. Suppose we set up an experiment to show that learners use Monitoring: whatever the result, all we have shown is that those learners did or did not use Monitoring in that situation. To make any general statement about Monitoring we need to isolate the situational and mental factors which interact to produce it and this means testing a variety of learners and situations. Though the problem of generalisation from limited data is always with us. It is particularly acute in interactionism.

This is not the place to consider the actual evidence for the Monitor Model in detail McLaughlin (1978) has made some criticisms. The relevant point to make here is that the Monitor Model seems to have placed such emphasis on the complex interactions involved as to neglect to establish the nature of Monitoring itself. It is claimed for instance that aptitude is related to conscious learning, i.e. acquisition (Krashen, 1981). This hardly seems surprising since most conventional aptitude tests are devised to show the learners' abilities to do well in a classroom and most classrooms aim at examinations that test conscious knowledge of the language: what we need is an aptitude test for natural language learning as well as for taught language learning. But what has this to do with Monitoring? Before the relationship between aptitude and conscious learning is interpretable in terms of the Monitor we need to show the relevance of Monitoring to conscious learning. A person might believe that steam is produced by millions of little demons; every time he sees a kettle boiling he claims this is confirmed. But this does not prove their existence to anyone else. The correlation between academic-type aptitude and conscious learning does not in itself commit us to a belief in the Monitor. A similar argument can be applied to the claim that Monitoring requires more time; a test that has to be done quickly does not allow the learner enough time to Monitor (Krashen, 1976). One may well accept that the time available for a task affects our behaviour; this does not necessarily mean that the difference in behaviour is due to Monitoring: we need to show that it is indeed Monitoring that is affected by time, not any of the other mental processes that we are using, such as short term memory. But the independent existence and importance of Monitoring seem not to have been established, apart from introspective reports by learners; virtually all the evidence is that of an interaction between mental makeup and situational factors producing differences in behaviour, which are not demonstrably caused by Monitoring rather than some other aspect of the interaction.

  2. The Acculturation Model

The Acculturation Model, as summed up on Schumann (1978), claims that social and affective factors in the learner combine to produce a single variable called acculturation - 'the social and psychological integration of the learner with the target (TL) group'. (Schumann, 1978, p.29). This is the chief cause of success or failure in second language learning; 'the degree to which a learner acculturates to the TL group will control the degree to which he acquires the second language' (Schumann, 1978, p.23). The social factors that influence acculturation are the cultural, structural, and political relationships between the native and the target groups: a good situation for language learning is one where 'the 2LL group is non-dominant in relation to the TL group, where both groups desire assimilation... for the L2 group and where the 2LL group intends to remain in the area for a long time'. (Schumann, 1976, p.141). These factors are considered in terms of groups of learners and explained in terms of the relationship between target and native groups. Affective factors, on the other hand, are dealt with in terms of the individual learner and consist of factors such as language shock, culture shock, motivation, and ego-permeability.

Evidence from the Heidelberg Project (1978) is cited by Schumann (1978) to show that successful immigrant learners of German have more leisure and work contacts with Germans. In the terms used here this claims that certain cultural pressures produce a particular combination of affective and social factors in the learner. Schumann is making a proposal for the dominance of these factors over all others in the interaction with the situation; he Is also suggesting that these factors affect the situation by making the learner seek out or avoid certain situations such as contact with native speakers.

How persuasive is the argument for this? Again it is not appropriate to look at the evidence for this in detail here except insofar as it relates to the framework being discussed. One would require evidence from a wide range of situations that acculturation is as important as is claimed. While it may well prove to be relevant to contact situations between host and migrant communities, it has little obvious relevance to other situations. For example, most second foreign language learning in schools has no clear relationship to acculturation; many British children learning French do not want to mix with French people. Schumann allows for another process called enculturation by which an individual assimilates to his own cultures (Schumann, 1978), but this he applies to 'elite professionals' rather than schoolchildren. However, some teachers have suggested that their motive in teaching French was to make their pupils dissatisfied with English society, the opposite of enculturation. Nor do acculturation and enculturation explain the second language learning of English, which is often learnt as a vehicle for communication detached from an English-speaking environment and which has sometimes been learnt by groups whose specific aim was to overthrow English-speaking culture and the status quo in their own countries.

There is also no account of the interaction which causes acculturation to be so powerful, other than in its effect on the learner's avoiding or preferring certain situations. It could be, of course, that this is all that is required of the model; the amount of language one learns is related to the amount of language one listens to; but like Krashen's input hypothesis (Krashen, to appear) this becomes uninteresting because it tells us when learning takes place but nothing of learning itself. To say anything about learning we must assume that the degree of acculturation causes learners to behave in different ways; what we are told of this is an analysis of the development of syntax in terms of decreolisation. That there might be evidence has been shown, for example, by work that found that integratively oriented learners offer more responses in the classroom (Gardner, Smythe, Clement, and Gliksman, 1976). Without more specification of what acculturation means in terms of behaviour, we are left with something resembling a black-box model of learning: you see what goes in one end (acculturation) and you see what comes out the other (successful language learning) but you have no Idea of what is going on in the middle. The strength of the acculturation model Is its emphasis on aspects of mental makeup and their interaction with the learner's choice of the situation; its weakness is its failure to characterise the behaviour that is the product of this interaction.

3. Conversational Analysis

Conversational analysis as presented in Hatch (1978) is both a theory and a methodology. The theory is that in learning a language 'One learns how to do conversation, one learns how to interact verbally, and out of this interaction syntactic structures are developed (Hatch, 1978); the methodology consists usually of detailed discussion of fragments of observed conversation between native speakers and second language learners. The important strategies in conversation are: to get the other participant's attention, "You know John?"; to provide the topic of discussion, "Well he's just got a new car"; to effect repairs to the conversation whenever necessary, "He's got a new what?"; and to get clarification of the topic, "John who?" Both adult and child learners initially have great difficulty in perceiving and nominating topics in a foreign language; this may cause them to perceive their chief problem as being a matter of vocabulary. Learning takes place by attempting to use these strategies and hence syntax and other linguistic levels are subordinate to the learner's conversational strategies. Fitting this into interactionism we find that it is a description of certain ways of constructing conversations - the types of move that the learner may make and how they can be sequenced. The situation that interacts with these moves is spontaneous conversation between foreigner and native in a non-classroom setting. Mental makeup is not discussed except for differences in preferred topics between adults and children. Conversational analysis insists on the centrality of the learner's individual strategies, although the idea of these being an interaction between learner and situation Is not stressed.

On the one hand, however, it provides no hint where these strategies come from in terms of mental makeup; are they transferred from the first language, in which case we would expect to find some variation between learners of different native languages? Is it that the situation suggests strategies that are not normally used by the learner in his first language? What in short do they tell us of the learner's mental makeup? On the other hand it describes only a limited range of situations; many of the adult conversations reported appear to have been interviews, a situation with a clearly prescribed set of rules in which the interviewer plays a 'leader' part that allows him to ask a succession of unrelated questions for no other purpose than to elicit information and with very little opportunity for the interviewee to adopt other than a 'follower' role of answering his questions. Thus the particular strategies may be true only of the limited situations that have been reported: we do not know the interaction between situation and learner's behaviour. A broader range of situations may necessitate a considerably amplified set of strategies, for example, introducing, requesting, reacting, confirming, giving directives and many more.

Summing up these three models of second language learning, each of them makes certain claims about its own area but is far from complete in its coverage of second language learning. If we accept the reality of Monitoring, for instance, this tells us something about the use of consciously acquired grammar by certain learners in certain circumstances; it is a model of syntactic production and comprehension strategies.

This can only be taken as a wider proposal if we commit ourselves to the centrality of syntax in language learning and use; if, however we believe that phonology and semantics are as important as syntax, or if we feel, as many have come to believe in recent years, (Bruner, 1975; Halliday, 1975), that social interaction and the acquisition of pragmatic and communicative competence are far more important than syntactic development, then the Monitor Model has little to offer. Similarly the limitations of the Acculturation Model and Conversational Analysis prevent them from being more than partial accounts; nor do they seem to show the relevance of inter­action which, whatever its other faults. Is the chief merit of the Monitor Model. A complete model of second language learning therefore needs, as well as a more complete description of the diverse aspects of language learning, such as the learning of pragmatic competence, to explain the interaction with the features of the situation.

Interactionism is not so much a model itself as an approach, a way of looking at models. To comply with its axioms we need to account for both learner and situation; measured by this yardstick both the Acculturation Model and the Conversational Analysis Model do not go far enough; the Monitor Model in a way goes too far by so concentrating on the interaction of learner and situation as to neglect the actual behaviour it claims to be at its centre. If we take an interactionist position, very little of the research that has been carried out so far meets the criterion of adequately describing both sides of the interaction and the intervening behaviour. In a recent paper McLaughlin (1980) has outlined the kind of research design involved in interactionism based on the analysis by Cronbach and Snow (1977), and has provided examples from educational research and from second language learning research. Let us add to his examples a further one from current research and see how it may be called Interactionist.

Cook (to appear) describes an experiment into the use of language functions by native speakers of English and foreign adult learners. The aim was to establish the situational variation caused in the realisation of language functions by the age and sex of the addressee. The experiment asked two groups of subjects to complete a questionnaire asking "What would you say if ..." and specifying the language function of requesting or of thanking, and the age and sex of the addressee; the latter was done indirectly through cartoon stereotypes of old and young men and women. Thus on the person side the learner was supplied with an assumed goal - thanking or making a request. On the situation side there was specification of two factors believed to influence the language behaviour. What was being tested was the interaction between a goal and a situation. Results showed that there was indeed variation in the realisation of the language function according to age but not sex, and that this differed between the groups. Thus the result was interpreted as an interaction between person and situation; both sides were taken into account. This was incorporated into an informal model of conversation stressing the person's goals and strategies for carrying them out and the situational limitations on these. Though the topic that was being created was comparatively restricted and the conclusions necessarily limited, the experiment was conceived and carried out within a framework of interactionism. If we are to take interactionism seriously, we need to explore deeper and wider using this type of design.

Let us finish with two examples where interactionism gives us an improved awareness of the learner in a language teaching situation. The first example starts from the learner's mental makeup and draws conclusions for language teaching. We know that short term memory capacity develops with age and we know that it is slightly more limited in a second language (Cook, 1979); we know, say, that a thirteen-year-old English child learning French has a short term memory capacity for 2 words less in French than in English. This means that language teachers can adopt teaching techniques that use language within the learner's processing capacity for French, say in terms of sentence length for repetition or, alternatively, that they can deliberately strain the learner's capacity for one reason or another. Thus we are seeing the teaching situation as an interaction between known aspects of the learner's mental makeup and the teaching situation we provide and examining our techniques in the light of learner factors. The second example goes in the reverse direction, starting from a teaching technique and ending with the learner. The teacher decides to use the Community Language Learning technique in which the learners talk to each other in a circle, using the teacher chiefly as a translator. This technique implies that the learner is motivated primarily by the need to communicate within the group, a variant of the integrative motivation, and it makes various other assumptions about the learner's expectations of the classroom as a group experience: it also makes testable assumptions about the usefulness of translation as a learning strategy. Thus going from a teaching technique we can see the implications this technique has about learning, check whether they are in fact true, and end up, perhaps, by specifying what learner type is ideally suited to that technique. Of course      we need much more specific information to make this work. In particular we need to look more closely at teaching techniques in terms of the student's actual behaviour; what does the technique imply in terms of speech processes, of memory capacity, of pragmatic strategies and so on. Too often a technique has been seen from only one of these points of view; a structure drill was seen as practising habit-format ion, a type of learning strategy, and was not considered as presenting a small model of conversational interaction, nor investigated in terms of its information processing load. Interactionism may be relevant not only to second language learning research but also to our thinking about language teaching.

  References

ARGYLL, M. 'Predictive and Generative Rules Model of P x S Interaction', in Magnusson, D. and Endler, N., (1977)

BRUNER, J.S. 'From Communication to Language. A Psychological Perspective', Cognition 3/3, 1974/75 .

COOK, V.J. 'Aspects of memory in secondary school language learners'. Interlanguage Studies Bulletin: Utrecht, 4, 2,161-172, 1979.

COOK, V.J. (to appear). 'Language Functions in Second Language Learning and Teaching'.

CRONBACH, L.J., and SHOW, R.E., 'Aptitude and Instructional Methods: A Handbook for Research on Interactions'. New York 1977

ENDLER, U.S., and D. MAGNUSSON. 'Toward an Interactional Psychology of Personality'. Psych. Bulletin 83 5, 956-97, 1976

FELIX, S.W. 'Some differences between first and second language acquisition Paper delivered to Third International Child Language Symposium, 1975.

FRASER, B., RINTELL, E., and WALTERS, J. 'An Approach to Conducting Research on the Acquisition of Pragmatic Competence in a Second Language', In D. Larsen-Freeman (ed.) Discourse Analysis in Second Language Research Newbury House, 1980.

GARDNER, R.C.R., LAMBERT, W.E. Attitudes and Motivation in Second-Language Learning. Newbury House, 1972.

GUIORA, A.Z., BRANNAN, A. and DULL, C, 'Empathy and second language learning, LL, 23, 2 (1977).

HALLIDAY, M.A.K. Learning How to Mean. Edward Arnold 1975.

HATCH, E., 'Discourse Analysis and Second Language Acquisition', in E. Hatch (ed.) Second Language Acquisition. Newbury House 1978.

HEIDELBERG RESEARCH PROJECT, 'Pidgin Deutsch spanischer und Italienischer Arbeiter in der Bundesrepublik' . Germanistisches Seminar der Universitat Heidelberg, 1978.

HOSENFELD, C, 'Learning about language: discovering our students' strategies', FLA, 9, 2 (1976).

KRASHEN, Stephen D., 'Formal and Informal Linguistic Environments in Language Acquisition and Language Learning'. TESOL Quarterly, 10:2, 1976

KRASHEN, S., Second language acquisition and second language learning Pergamon 1981.

KRASHEN, S., 'Second language acquisition theory and the fundamental pedagogical principle,' Proceedings of the Sixth AILA Congress, (to appear)

LEWIN, K., A Dynamic Theory of Personality. McGraw-Hill. N.Y. 1935

MAGNUSSON, D., 'The Individual in the Situation’. Studia Psychologica XVI, 1974, 124-132.

MAGNUSSON, D., and ENDLER, N., (eds.), Personality at the Crossroads: Current Issues in Interactional Psychology, LEA. (1977)

McLAUGHLIN, B., 'The Monitor Model: Some Theoretical Considerations'. LL 28, 2, 309-332 .

McLAUGHLIN, B., 'Theory and research in second language learning; an emerging paradigm,' LL, 30, 2 (1980)

SCHUMANN, J., 'Implications of pidginization and creolization for the study of adult second language acquisition', in J.H. Schumann and N. Stenson (eds.) New Frontiers In Second Language Learning. Newbury House 1975

SCHUMANN, John H., 'Social Distance as a Factor in Second Language Acquisition.' LL, 26, 1, 1976, 135-143

SCHUMANN, J.H., 'The Acculturation Model for Second-Language Acquisition', in R.C. Gingras (ed.) Second Language Acquisition and foreign Language Teaching Center for Applied Linguistics 1978

TREMAINE, Ruth V., 'Piagetian Equilibration Processes In Syntax Learning', in D.P. Dato (ed.) Psycholinguistics: Theory and Application, Georgetown 1975