Developing links between SLA research and language teaching by Vivian Cook

In this paper an attempt is made to systematically describe the possible relationships between Second Language Acquisition Research (SLA) and language teaching. The central question underlying the article is how language teaching can benefit from SLA. In referring to a model of possible levels of application it is shown how the results of SLA can be applied in language teaching. At the end of the paper some general maxims for the application of L2 theories in language teaching are proposed.

About twenty years ago the discipline of Second Language Acquisition (SLA) research started putting together a principled scientifically-established set of general ideas about Second Language Acquisition. The coverage of SLA research has now become quite extensive; several books give overall accounts of SLA, such as Ellis (1985), Spolsky (1989), and Larsen-Freeman & Long (1991). In addition there are extended treatments of more specific areas such as transfer (Odlin 1989), individual variation (Skehan 1989), and sociolinguistics (Preston 1989). The question that can now start to be seriously raised is: how might language teaching benefit from Second Language Acquisition research?

But this question presupposes other questions. The relationship between linguistics and language teaching has been damaged time and again by linguists appearing to lay down the law to language teachers without recognising the independence and expertise of language teaching as a discipline in its own right. This paper is then an attempt to look at these general relationships and to discuss the ways in which they can be developed rather than to suggest actual applications. It arises out of an attempt to describe SLA research for language teachers (Cook 1991a), during which it became apparent that application of SLA research raises issues to which no easy solutions present themselves: what should be applied to what, how, and why^1.The paper comes from the particular perspective of an SLA researcher working in Britain; it examines research to see how it might be relevant to language teaching rather than exploring the equally useful question of what research questions teaching may suggest or give rise to. Teaching however is a variable not a constant; it can mean a vast range of methods and techniques ii different parts of the world. Applications of linguistics to particular aspect: of language teaching have to take account of the particular goals and need: of the society in which teaching is taking place. They are suggestions t( teachers which they are perfectly free to accept or deny, as the experts on their own situations.

The past few years have seen a renewed interest in general theories of SLA McLaughlin (1987) talks of five: the Monitor theory, interlanguage theory linguistic universals, acculturation/pidginisation theory and cognitive theory Ellis (1985) mentions seven, adding discourse theory, variable competent theory, accommodation theory, and neurofunctional theory, and subtracting interlanguage theory and cognitive theory. Cook (1991a) divides theories into four groupings: knowledge models, such as Universal Grammar (UG) language processing models, such as the Competition Model; models wit! more than one component, such as the Competence/Control Model; an< social models, such as the Socio-Educational Model.

Each of these models puts forward an explanation for some aspect o SLA. But they are rarely true alternatives to each other as they are no competing for the same territory. A model that is concerned with grammatical morphemes such as past tense "-ed" (the Monitor) scarcely contacts on concerned with syntactic parameters such as whether sentences have to have subjects (Universal Grammar); a model that relies on syntax has little overlap with one that emphasises social interaction. Language students mostly need language as a whole: there is little point in teaching them the grammar of the language if they cannot pronounce it, do not know the vocabulary and cannot join in conversations. An SLA 'framework' is needed rather than a 'model' or a 'theory'; application to language teaching requires concepts that are wider than any of the existing SLA models; a framework such as the 74 conditions for language learning of Spolsky (1989), allow alternative 'models' to coexist within a larger whole. There is therefore no predetermined correct theory of language teaching. Teachers will have I decide which model is the most relevant to language teaching.

One problem in applying SLA research to language teaching is the type of relationship this entails between the two. Some theories make all-embracing claims about language learning. Two examples are the Universal Grammar (UG) theory put forward by Chomsky (1988) with its claim that the mind has innate language principles and variable parameters (White 1989), and cognitive theory with its claim that language is a complex cognitive skill like any other rather than something entirely different (Anderson 1983). Such claims potentially affect the whole of L2 learning and the whole of language teaching. A UG theorist sees teaching as providing proper language input to set values for parameters, a cognitive theorist as helping with learning strategies. It is beside the point whether any L2 evidence exists for such claims because they encompass all learning and all teaching. Both theories claim their main constituency outside the SLA field; SLA models are extrapolated from their findings elsewhere. Both our examples rely on little LI evidence of their own; Universal Grammar theory bases itself on logical arguments (Cook 1991b) and has been notoriously weak on empirical evidence, at least until recently; Anderson's language evidence for the cognitive theory is based chiefly on computer simulation (Anderson 1983).

In contrast with this application of SLA theories is the idea that the results of SLA research should be applied. Let us take an example from phonology: Broselow (1988) and others have found that L2 learners carry over the syllable structures of their LI to the L2; consonant clusters impossible in the LI will be padded out with extra vowels. Though this could be an instance of a more general theory of transfer, it can be seen more narrowly as affecting lessons in pronunciation - a local application. SLA research needs a hierarchy of theories and claims going from the most broad and general down to the most specific and detailed.

The next question is what SLA research applies to. There are many levels and aspects of teaching, each of which may be influenced in different ways. Here we will look at some of the various contacts that SLA research could have with goals, syllabus, grading, and teaching methods and techniques.

The reasons why language is being taught are often taken for granted; Ellis (1990:188) for example claims that 'the raison d'etre of the classroom is to learn the language'. The purpose of language teaching classrooms however varies according to the society and the educational system. Teaching achieves the goals of a government or a society, or is concerned with combatting a government. It exists within a general educational curriculum. Or language teaching may be intended as cognitive 'brain-training'. Rarely is language learning in classrooms just learning the language. The language and language learning components are indeed the aspect of teaching with which SLA rightly concerned. But the goals of teaching are determined by the individual student, the individual teacher, the society, or the government rather than by the SLA researcher. This is putting to one side legitimate goals that are unconnected to education, for example parents bringing up children bilingually from choice or people learning second languages for religious reasons.

One needs at least to divide teaching goals into three main types, developed in Cook (1991a):

1local goals which foster the L2 within one country, for example the teaching of French in Switzerland. These reflect how a multilingual community copes with the variety of languages within its borders.

2international goals which foster language for use outside the country, for example the teaching of English in Austria. The students are taught foreign language for use in contacts with people outside the home country

3 individual goals which develop the potential of the individual learner, for example the teaching of German in England to increase logical reasoning literary appreciation, or emotional maturity. These are not related to the society itself or to its external relations; individual goals aim to make the student a 'better' person, with actual language use in the background There are then a wealth of possible goals for language teaching. SLA research can be used to achieve these goals, or to test whether they are actually achievable; it can look for evidence whether language teaching can really train brains, form characters, and so on. But SLA research does not itself say what those goals shall be.

Though other uses for a syllabus are also important (Prabhu 1987), the syllabus is here taken to represent the detailed specification of the goal that the teacher is aiming at. The language teaching syllabus depends upon the goals for which teaching the second language is being taught. Obvious as this point may be, frequently the discussion has concerned syllabuses in their own right; should teachers use a communicative syllabus or a process syllabus etc.? As put, this question is unanswerable; the type of syllabus depends on the goals of language teaching; no type of syllabus has an value except as the implementation of the goals chosen for language teaching SLA research has no brief to recommend any type of syllabus by itself.

Without going into the varieties of syllabuses or of goals, an adequate syllabus for language teaching needs to cover at least the following areas taking a quick example of each, given at greater length in Cook (1991a):

Grammar is part of the native speaker's knowledge of the language and has to be acquired by the L2 learner, whatever importance is assigned to it or whatever type is involved. An example might be that research suggests that the syllabus should pay more attention to the separation between principles of language that do not need to be taught and parameters of language which need detailed treatment.

Speaking a language requires a knowledge of the actual sounds of language, whether at a phonetic, phonemic or intonational level. Or indeed how to read or write the symbols of the written language. Pronunciation research suggests for example that the syllabus should concentrate on broad aspects of pronunciation such as devoicing for German learners of English rather than on minimal pairs of the "bit/beat" variety.

The meanings and properties of lexical items are also part of the knowledge and use of the language, however differently the role of vocabulary is conceived in various linguistic theories. Vocabulary research suggests that the syllabus should not concentrate on frequency of occurrence so much as on usefulness and memorability.

The use of a language varies in at least two dimensions; one is the location of the sentence within the specific sequence of discourse; the other is systematic variation according to the overall situation, discourse research suggests striking a balance in the syllabus between information-related talk and listener-related talk.

Language is the psychological processes of speech comprehension and production - again dealt with variously in different approaches to linguistics and to SLA. Performance research means relating the demands that the syllabus makes on the students to the cognitive deficit that often arises on mental processes in a second language, for example the reduced digits span for numbers or the lower amount of information to be gained from texts.

SLA research can make a contribution to each of these areas in so far as it describes how people learn whichever of these aspects of language is needed to achieve a particular goal. The overall point is not so much the actual suggestions made here as the complexity and diversity of any syllabus that tries to take Second Language Acquisition research seriously.

The grading of the syllabus content into a teaching sequence has often be< seen as complementary to the selection of the syllabus content. No learn assimilates the whole syllabus instantly; language knowledge increases grad ally over time. This expansion may be planned by a teacher or it may occur through the accidental encounters the learner has with language in natural situations. But it nonetheless occurs in sequence.

The SLA literature has frequently asserted that there is a natural sequence that all L2 learners go through; for instance Dulay and Burt (1980). Most this has been advanced for morphological elements such as the grammatical morphemes present tense "-s " or the articles "the/a " or for syntactic elements such as negation. The grading of language in teaching might, it would appear be substantially influenced by information from SLA on sequence of acquisition. But what in fact does an order of acquisition mean for teachers?

Nearly all the areas that have been studied are matters of low-level syntax on the one hand grammatical information has a limited role in any syllabus on the other, if grammatical information is going to be used at all, a much 'deeper' type of grammar may be necessary. My experience of writing beginners EFL course (Cook 1980) taught me that grading in a coursebook gets changed constantly during writing for many ad hoc reasons that have little to do with abstract ideas of L2 sequence but everything to do with the practicalities of teaching. Grading in teaching does not depend solely on the sequence of learning but needs also to take into account grouping teachability, frequency, usability, etc. SLA research has one vote out of several in assigning teaching sequence, and its importance varies according to the type of teaching situation, goals, teaching methods, etc. Much discussion of the applications of SLA research to the classroom has stressed its contribution to sequencing, for example Pienemann and Johnston (1987). This is a comparatively minor issue in teaching, substantially dependent on other decisions about goals and methods. For SLA theories too the sequence in which learners acquire the language is only of secondary importance compared with the underlying reasons for this sequence.

The crucial aspect of the classroom so far as SLA is concerned is that it provides the input and the context for language learning. The language the learners hear is the vital source of their ideas about the L2. From this perspective the teacher's responsibility is to provide adequate samples of language for the students to work on. The language has to reflect the type of language the student is aiming at and to provide the appropriate language for the student to learn from. Appreciation of the peculiarities of classroom L2 learning may well be a crucial contribution of SLA research.

In general SLA can influence teaching at two levels. One is the overall impact on all teaching techniques; certain ideas from SLA could influence every technique. The other is the impact of SLA discoveries on particular techniques. SLA research is not necessarily tied in to any existing method but can have something to say for any method - or can indeed suggest completely new methods. As Spolsky (1989) wisely remarks, 'any theory of second language learning that leads to a single method must be wrong.'

It should perhaps be obvious that not all learners learn everything in the same way. There is a continuum between aspects that all learners have in common and those in which they differ. Let us list some of the learner variables studied in SLA with a brief indication of their possible connections to language teaching, discussed more fully in Cook (1991 a):

Typically motivation has been seen as a combination of two types - integrative (wanting to join in the target culture) versus instrumental (wanting language for ulterior reasons such as passing an examination). (Gardner & Lambert 1972). The EC survey of young people found that 29% wanted to learn more languages to increase their career possibilities, 14% wanted them in order to live, work, or study in the country, and 51 % were motivated by "personal interest" (Commission of the EC 1987). The teacher has either to accept the students' motivation or to try to change it through the teaching methods that are employed.

Success in academic 'formal' teaching has been found to go with conventional aptitude tests. Aptitude is related to the type of teaching rather than being constant for a given learner, as Wesche (1981) demonstrated. Application of SLA to teaching requires adaptation to the student.

Strategies such as the Good Language Learner strategies have been described that are adopted by 'good' academic learners (Naiman eta 1978); 'good' learners typically try to involve themselves in the learning process. Many teaching materials now aim to make such strategies explicit to the learners, for example Ellis & Sinclair (1989) and Oxford (1990).

The role of age in L2 learning is still controversial. Singleton (1989:119) claims "The one interpretation of the evidence which does not appear to run into contradictory data is that in naturalistic situations those whose exposure to a second language begins in childhood in general eventually surpass those whose exposure begins in adulthood, even though the latte usually show some initial advantage over the former."

Other factors that have been discussed include: field-dependent versus field-independent cognitive style, extroversion, IQ, sex, and empathy; good survey can be found in Skehan (1989). Although the connection between these and SLA are controversial, nevertheless they are clearly relevant.

Hence the application of SLA research to teaching cannot simply use a idealised figure of a typical learner. In some areas of learning difference between learners are negligible. In other areas the nature of the individual learner is crucial. An important contribution of SLA research is indeed the analysis of the variation in learner types. It leads to the conclusion that no application of SLA research to the global situation of a classroom can ignore variation in students.

SLA researchers commonly distinguish between classroom learning and learning outside classrooms. Classroom learning is believed to have particular characteristics of its own, due to the educational situation in which it occurs and to the specific pressures of the classroom situation. But this division into classroom and non-classroom learning also appears to equate teaching with classrooms and takes no account of teaching methods such as self-instruction, distance learning, self-directed learning, CALL, and so on. Teaching can take many forms other than sitting in a classroom. The applications of SLA should not be confined narrowly to classroom teaching.

The classroom/non-classroom distinction is sometimes phrased as a difference between formal and informal learning. One sense of the word "formal" is synonymous with "taught"; "formal education" for example is 'given officially in a school' (COBUILD 1987). Another sense common among linguists is "formal grammar", i.e. grammar based on syntactic form rather than meaning. Hence formal grammar in SLA is often extended to teaching that is 'form-focussed' on the actual forms of language rather than on their meaning or use; formal instruction, according to Ellis (1985: 217), combines the attributes "(1) specific grammatical features are selected for the learner's attention and (2) this attention is manifest in a focus on the formal characteristics of the grammatical features". There is no reason why linguistic form has to be restricted to syntax; form could be applied to any level of language from phonetics to discourse. The classroom also could involve many varieties of form-based learning - translating texts, listening to grammatical explanations, pronunciation drills, essay-writing, and so on. Nor is formal learning necessarily confined to the classroom; what adult L2 learner visiting another country does not invest in a phrase book, or a dictionary or a grammar book, regardless of whether they go to language classes? Overall the distinctions between classroom and naturalistic and between formal and informal blur when they are approached more closely. SLA research cannot assume that such broad distinctions are necessarily relevant to actual teaching.

In the light of this discussion, it is clearly advisable to have an appropriate relationship between the area of SLA that is being applied and the area or mode of L2 teaching to which it is being applied. Two dimensions of general to particular are required - one for SLA and one for teaching. SLA can be phrased as a continuum from theory to results; teaching as a continuum from styles to techniques. Any application locates itself somewhere within the two axes of level of SLA and level of teaching (see next page). The four extreme positions within this square reflect whether the SLA research to be applied is at the level of theory or results, and whether the teaching is at the level of style or technique, though of course these are points on two continuous dimensions.

7.1 Type (i) SLA Theory Applied to a Teaching Style

The first type of relationship is at the highest level of generality. Applying SLA theory to a teaching style means taking general theories of Second Language Acquisition and seeing how they can apply to broad areas of teaching style. In type (i) application, SLA theory impinges at the most general level on teaching. Take the Universal Grammar model of Second Language Acquisition. This insists that syntax is the central aspect of language knowledge and that it consists of unvarying principles and variable parameters that are triggered by appropriate evidence (Cook 1988). So all human beings know these automatically; no languages go against these principles. Principles don't need to be learnt and so they don't need to be taught. One teaching interpretation is a laissez-faire approach to teaching that relies on the mysterious black box in the student's minds to get on with learning, an approach originally associated with Newmark and Reibel (1969) but almost inseparable from the communicative teaching styles. Teaching syllabuses and teaching exercises can safely ignore such principles because they will be supplied by the learner. But children have to learn the values for parameters and so do L2 students. According to UG theory, learning takes place because the learners get the right type of language evidence - they hear examples of the appropriate constructions. For language teaching there is the clear implication that the teacher's role is the provision of the appropriate examples of language. If the students get the right sentences given to them, they can set the parameters appropriately.

This type of application of SLA does not specify particular teaching techniques. Nevertheless the implications of the SLA theory may be all-pervading. To take two examples, information-processing SLA models such as those of McLaughlin (1987) or MacWhinney (1987) stress practice but do not specify what form it will take in the classroom; social models such as Gardner's Socio-educational model (Gardner 1985) stress the importance of motivation and aptitude but do not dictate what this means in practical terms for the teacher in the classroom.

7.2 Type (ii) SLA Theory Applied to a Teaching Technique

Applying SLA theory to a teaching technique consists" of taking general ideas about Second Language Acquisition and seeing what they mean for particular teaching techniques. In type (ii) application an SLA theory leads to a particular technique. One example concerns the role of working memory in Second Language Acquisition (Cook 1991a). While one is listening to a sentence, one is using working memory to keep the information from the sentence in one's mind while one works out what the sentence means. The articulatory loop theory (Baddeley 1986) claims working memory depends upon an articulatory loop that recycles information continuously. Working memory capacity depends upon articulation: the faster you articulate the better your working memory. While this has many implications at an overall level, one derived teaching technique is my computer program called SPEEDUP, which gets the student to read lists aloud at increasing speeds. The very technique is inspired by the theory, though doubtless many other techniques could also be implemented. This is not to rule out type (i) general implications of articulatory loop theory for language teaching and type (iii) specific implications for actual techniques. Teaching that takes account of memory limitations will certainly be more effective.

7.3 Type (iii) SLA Results Applied to a Teaching Technique

In type (iii) application results of research are applied to a specific teaching technique. An example might be how research into syllable structure relates to pronunciation techniques. While the English syllable can have several consonants before or after the central vowel, the syllable structure of some languages allows only a single consonant before or after the vowel. Arabic syllables for example must not have more than one consonant at a time, so English combinations like /tr/ in "trend" or /Id/ in "child" present Arabic learners with a problem. The solution that the learners often come to is to add an extra vowel to space the consonants out, a process known as 'epenthesis'. So Egyptian-Arabic learners of English say "childiren" to avoid the "Id" in "children"; "translate" comes out as "tiransilate" to avoid two combinations "tr" and "si". Suppose a teacher looked at Broselow's finding that the use of epenthetic vowels by L2 learners reflects the syllable structure of their first language (Broselow 1988). Knowing the LI of a group of students to be Egyptian Arabic, the teacher gets them to repeat English words with clusters such as "dr" and "si". This is a restricted application, not based on general SLA or teaching theories but on particular use of relevant L2 research results.

7.4 Type (iv) SLA Results Applied to a Teaching Style

Applying SLA results to a particular teaching style means seeing how a particular fact about Second Language Acquisition has general implications. In type (iv) application, factual results from SLA can be detached from a particular theory and used in the classroom. Skehan (1986) found there were two types of successful L2 learner - 'memory-based' learners and 'analytical' learners - partly related to age. Once aware of this difference between learners, teaching can allow for it by providing different teaching methods, etc. Though it is information about the L2 learner, it does particularly depend upon any theory. Nevertheless it can have global consequences for language teaching, potentially affecting everything that happens to the student.

This is however potentially the most dangerous use of SLA since it pen large generalisation from little evidence. The question "Should we teach children syntax?" posed by Dulay & Burt (1973) was answered by two experiments with learners in the USA, one looking at a limited range of 'errors’, the other at morphology (Dulay & Burt 1973). Claims for a natural or of acquisition in certain areas of syntax led to a Natural Approach cover all language teaching (Krashen & Terrell 1983). In other words, the is wagging the dog. It is only when the results cover a range of areas t they are usable, say Skehan's all pervasive learner types. When they based on a narrow area such as grammatical morpheme, they need to used with caution and bolstered with precise arguments for their relevance outside their own domain.

The UG theory for example is concerned with central areas of syntax - principles and variable parameters; its goal is to explain the acquisition of grammatical competence. Grammatical competence pur et simple may indeed be the goal in some educational contexts. But the theory has nothing to say about areas other than central syntax; even peripheral syntax m well be learnt by different means to core syntax (Cook 1988). In particular it has no link to the acquisition of communicative and pragmatic competence. Hence UG theory is irrelevant in educational systems in which grammatical competence is not the priority, except to the extent that grammatical competence forms part of other competences. Likewise, however relevant the notion of learners' compensatory strategies may be to L2 learning (Poulisse 1989) it has nothing to say about other areas of language learning and nothing to say for classrooms where this type of spontaneous production by students is frowned on, say in listening-based or academic teaching techniques.

8 Maxims for Applying General SLA Theories

In one sense, however useful the particular applications of low-level SL results to low-level techniques may be, the connection is local and specific. Hence any links that arise, whether by design or by serendipity, should n be ruled out. The application of particular SLA results to general teaching styles is fraught with problems of validity of generalisation and should be regarded perhaps as a last resort. This last section looks at some overall requirements that can be placed on the relationship between general SLA theories and teaching styles or techniques.

I. Application of SLA theories to language teaching requires adequate coverage of the various aspects of L2 learning.

Minimal as this injunction may be, it is perhaps necessary to remind readers that most outlines of implications for teaching do not meet it. Ellis (1990) proposes an 'integrated' theory of instructed learning. This bases itself on classroom research and takes no account of vocabulary, phonology, listening comprehension, memory processes, etc; 'the specific linguistic features' the learner is required to attend to are primarily syntax. Dulay, Burt, and Krashen (1983) present ten teaching guidelines based on the research surveyed in their book; this however makes no mention of vocabulary learning, discourse, or internal strategies or processes other than Monitoring, etc. Whatever the value of the guidelines, they need considerable supplementing to cover an adequate proportion of SLA. The Teachability Hypothesis of Pienemann & Johnston (1987) suggests that teaching affects certain aspects of SLA that are subject to variation but cannot affect other 'developmental' aspects. Attractive as their proposal may be, the examples they give are based on general proficiency or on syntax, that is to say do not encompass 4 out of 5 of the aspects listed above. Their 'multidimensional' model has only two dimensions; it is closer to type (iv) application than to type (i) and is concerned primarily with the sequencing aspect of grading.

From a research perspective, it is valid to abstract one element from the total picture and to study it with rigour and depth. From a teaching point of view, however, the student is a whole person functioning in a complex situation; teachers cannot afford to adopt some teaching procedure without considering its place within the whole framework of teaching and learning, excellent as it may seem for some limited area. SLA research needs to give teachers a broad picture of L2 learning that encompasses all the aspects that teaching needs to take into account, ranging from vocabulary to conversational interaction, from grammar to pronunciation. Teachers should not allow SLA researchers to get away with advice that derives from some small fraction of the complex totality of L2 learning.

II. Application of general SLA research to language teaching requires adequate coverage of the various aspects of language teaching.

The second maxim concerns the need to recognise the complexity of the teaching situation. Chomsky's warning should be borne in mind: "I think linguists have done an enormous amount of damage in language teaching bysimply legislating how it should be done without any understanding or any knowledge of what it's about" (Chomsky 1988). The teacher's responsibility extends outside the SLA area, for example to the professional needs of adults in an L2 or to the psychological needs of teenagers in classrooms. Teachers have to develop all the sides of their learners appropriate for goals; it will seldom be enough just to give them an adequate grammar knowledge on its own or an excellent pronunciation by itself, rather need all the aspects of language relevant to their goal. Often however models or research have been extremely partial in their coverage; a | deal is known about syntax, less about pronunciation or vocabulary, about learning of functional aspects of language. Teachers should be careful that the SLA research they are interested in adequately covers the aspects found in any classroom. Any use of SLA at a general level and to take into account the whole educational context. The teacher has n masters, the most important being the students. Application of SLA to teaching cannot make the teacher abandon other responsibilities.

III. Application of general SLA theory to language teaching has to recognise variations in learners and in learning situations.

Again, while it may be obvious that SLA can be related to a whole spectrum of teaching situations, this has rarely happened. Often those who attempt to apply SLA research have behaved as if all classrooms were alike; picture of the classroom advanced in many SLA books is a traditional one remote from most current practice. Teaching can take place in many ways inside or outside the classroom, as in self-directed learning. In first language acquisition, the child is indeed confined by the limited linguistic environment provided by his or her parents. But the environment of the L2 students is limited only by the imagination of the teacher. So there is no 'standard’ application of SLA research. Instead it has implications for any teaching.

Learners, too, come in all shapes and sizes, as was seen earlier. What is good for one may not be good for another. SLA research has no sin magic solution to language teaching, no patented method to sell to the world. Rather it should, in the long run, have something to contribute to any method. The 'control' classroom to which the listening-based classroom is of compared in experiments is usually hardline audiolingualism, something rare outside the United States and probably not widely practiced there any more (Cook 1986). Few of the classrooms mentioned in the SLA literature resemble the conventional communicative classroom of the past fifteen years, or indeed the mainstream EFL classroom of the past 50 years. While the traditional or audiolingual classrooms taken as typical in much of the SLA literature obviously exist, on a worldwide scale large numbers of teaching methods have been, and are being, used every day with many millions of students SLA research has to support the wealth of possibilities in teaching and the learner, not close off the other possibilities by laying down one route that everyone has to follow.

To sum up, the constant refrain has been that, while SLA research and language teaching share many interests, they are not the same in their scope or responsibilities. Provided that a proper account of appropriate SLA research is utilised, provided that an adequate framework of language teaching is employed, a whole new generation of language teaching concepts and methods can be derived from SLA research. But we must not jump in prematurely with limited and partial ideas of either Second Language Acquisition or language teaching.

¹ Versions of sections of this paper were presented at conferences at RELC in Singapore and at USM in Penang in 1991; I am grateful to all the comments made by participants.

Anderson, J.: The Architecture of Cognition. Cambridge, MA: Harvard University Press 1983.

Broselow, E.: "Prosodic phonology and the acquisition of a second language." Flynn, S./O'Neil, W. (eds.): Linguistic Theory in Second Language Acquisition. Dordrecht: Kluwer 1988.

Burt, M.K./Dulay, H.C.: "On acquisition orders." Felix, S. (ed.): Second Language Development: Trends and Issues. Tubingen: Narr 1980:265-327.

Chaudron, C: Second Language Classrooms: Research on Teaching and Learning. Cambridge: Cambridge University Press 1988.

Chomsky, N.: Language and Problems of Knowledge: The Managua Lectures. Cambridge, MA: MIT Press 1988.

Commission of the European Communities: Young Europeans in 1987. Strasbourg: Commission of the European Communities 1987.

Cook, V.J.: Second Language Learning and Language Teaching. Sevenoaks: Edward Arnold 1991a.

Cook, V.J.: "The poverty-of-the-stimulus argument and multi-competence". Second Language Research 7:2 (1991b): 103-117.

Cook, V.J.: Chomsky's Universal Grammar: An Introduction. Oxford: Blackwell 1988.

Cook, V.J.: "The basis for an experimental approach to second language learning." Cook, V.J. (ed.): Experimental Approaches to Second Language Learning. Oxford: Pergamon 1986:3-21.

Cook, V.J.: People and Places. Oxford: Pergamon 1980; Polish edition. Polish Scientific Publishers 1987.

Dulay, H.C./Burt, M.K.: "Should we teach children syntax?" Language Learning 23:2 (1973): 245-258.

Dulay, H.C./Burt, M.K./Krashen, S.: Language Two. Rowley, MA: Newbury House 1983. Eckman, F.R./Bell, L./Nelson, D.: "On the generalization of relative clause instruction in the acquisition of English as a second language". Applied Linguistics 9:1 (1988): 1-20.

Ellis, R.: Instructed Second Language Acquisition. Oxford: Blackwell 1990. Ellis, R.: Understanding Second Language Acquisition. Oxford: Oxford University Press 1985.

Ellis, G./Sinclair, B.: Learning to Learn English. Cambridge: Cambridge University Press 1989.

Gardner, R.: Social Psychology and Second Language Learning. London: Edward An 1985.

Gardner, R.C./Lambert, W.E.: Attitudes and Motivation in Second Language Learning,

Rowley, MA: Newbury House 1972. James, A./Leather, J. (eds.): Sound Patterns in Second Language Acquisition. Dordrecht: Foris 1986.

Larsen-Freeman, D./Long, M.H.: An Introduction to Second Language Acquisition Research, London: Longman 1991.

Mackey, W.F.: Language Teaching Analysis. Harlow: Longman 1965. MacWhinney, B.: "Applying the competition model to bilingualism." Applied Psycholinguistics (1981): 315-327.

McLaughlin, B.: Theories of Second-Language Learning. London: Edward Arnold 1987

Naiman, N./Frohlich, M./Stern, H.H./Todesco, A.: The Good Language Learner. Toronto OISE 1978.

Newmark, L./Reibel, D.A.: "Necessity and sufficiency in language learning." IRAL VI/3 (1968): 145-164.

Odlin, T.: Language Transfer. Cambridge: Cambridge University Press 1989. O'Malley, J.M./Chamot, A.U. (eds): Learning Strategies in Second Language Acquisition Cambridge: Cambridge University Press. 1990.

Pienemann, M./Johnston, M.: "Factors influencing the development of language proficiency." Nunan, D. (ed.): Applying Second Language Research. Adelaide: NC 1987:45-141.

Poulisse, N.: The Use of Compensatory Strategies by Dutch Learners of English. Nijmegen Ph.D. 1989.

Preston, D.R.: Sociolinguistics and Second Language Acquisition. Oxford: Blackwell 198

Richards, J.C./Rodgers, T.S.: Approaches and Methods in Language Teaching. Cambridge University Press 1986.

Singleton, D.: Language Acquisition: The Age Factor. Clevedon: Multilingual Matters

Skehan, P.: Individual Differences in Second-Language Learning. London: Edward Arnold 1989.

Skehan, P.: "Cluster analysis and the identification of learner types." Cook, V.J. (ed) Experimental Approaches to Second Language Learning. Oxford: Pergamon 1986: 81-94

Spolsky, B.: Conditions for Second Language Learning. Oxford: Oxford University P 1989.

Wesche, M.B.: "Language aptitude measures in streaming, matching students with methods, and diagnosis of learning problems." Diller, K. (ed.): Individual Differences Universals in Language Learning. Rowley, MA: Newbury House 1981:83-118.

White, L.: Universal Grammar and Second Language Acquisition. Amsterdam/Philadelpia Benjamins 1989.