The Reflection of Linguistic Theory in Language Teaching
The reflection, presumed or real, of linguistic theory in language teaching has a long history fn the development of applied linguistics. It was evident already in the replacement of the grammar translation method by the direct method, according to Malmberg (1981: 9), since the “Saussurean thesis of the arbitrariness of the linguistic sign and of linguistic structures made translation not only unnecessary, but impossible.”
At the same time, early examples of advances in language teaching give evidence of a feedback from applied linguistics to linguistic theory itself, thus lending credibility to the view that applied linguistics is part of linguistics, since the direc?tion of the insight is not unidirectional, but reciprocal. Malmberg (1981: 10f.) mentions in this regard the idea of tallying the frequency of phonemes and words, a “governing principle for the direct method in language teaching,” which in its turn led to a modification of descriptive linguistic theory.
The reflection of linguistic theory in language teaching first became a problem, however, with the rise of transformational-generative grammar, and this really is the twist in the tail. Before this theoretical approach came to be the leading school of linguistic analysis at the end of the previous century, there had been an uncritical acceptance, among applied linguists and linguists alike, that the relations between linguistics, applied linguistics and language teaching were such that there was a simple carry-over, mediated by applied linguistics, of linguistic knowledge into language teaching. Most of the discussion in this section will therefore be focused on the marked contrast between earlier attempts at ‘applying’ linguistics to second and foreign language teaching, and the noticeable lack of success with such attempts since the advent of transformational-generative grammar.
With the exception of attempts such as those by Lakoff (1969) and, somewhat later, Cook (1985) , there is little direct reflection of transformational-generative grammar tn language teaching in terms of method or approach (but cf. Spolsky 2008: 1), specifically as regards the selection of formal language features or content to be included in language course design. Both of these discussions suggest relations between transformational grammar and language teaching and learning, but neither forms the source document for, or embodies any influential method of language teaching. Chomsky himself has, of course, been noticeably reticent regarding the application of his work to language teaching. Speaking to an audience of foreign language teachers, he remarks that the implications (rather than direct applications) of his work for language pedagogy are far from clear, and adds a warning that “suggestions from the ‘fundamental disciplines’ must be viewed with caution and skepticism” (1966: 262).
Stevick (1971: 7ff.) also briefly mentions (the theoretical underpinnings of) a “transformational-cognitive school” in language teaching, but the references he cites appear to be only to statements and research reports, often unpublished, and not to a method or approach to language teaching that could in any way be regarded as having the same stature as the audio-lingual method. The latter was characterized by a wide range of pre-packaged language teaching courses, whereas there is very little evidence of teaching materials that claim to embody what Stevick calls a “transformational-cognitive approach”. In fact, as Richards (1975: 2) pointed out at the time, new materials “have been slow to appear” after the displacement of structuralism by transformational grammar. Similarly, of the odd few attempts discussed in Roulet (1975: 40ff.), most seem to be concerned with mother tongue instruction.
It was exactly in respect of being a reflection of linguistic theory, however, that the oral approach or audio-lingual method excelled. As a method it found its greatest debt to linguistic theory in its “scientifically chosen and arranged” language teaching materials. Fries (1945), Smolinski (1985) insists that this approach depends on materials that are arranged according to linguistic principles, that the contribu?tion of the techniques of scientific analysis to language teaching is to “provide a thorough and consistent check of the language material”, if the language teaching method that derives from this is to be effective in ensuring the maximum progress in the language being learned by the student.
Claiming, as it does, to derive from behaviorist structuralism in linguistics, the audio-lingual method i s noteworthy in a second respect, making it interesting to observe how the varying research goals and theoretical starting points of descriptiv- ism and generativism determine the differences in the kind of influence that each has exerted on language teaching. In Fries’s (1945) remarks, for example, one has ample evidence of the pre-occupation of the former school of linguistics with ‘correct’ techniques of analysis (cf. Sampson 1980: 72ff.) in order to arrive at a ‘scientifically’ sound description of the language that, in turn, could be applied to the selection of language teaching material. In the Chomskyan tradition, on the other hand, the emphasis is on giving an explicit characterization of psychologically plausible principles of mental computation (cf. Chomsky 1980: 5), i.e. to explain how language is acquired. In his survey, Cook (1985: 14) stresses that “a long and arduous route connects the theory with language teaching.” How different this is from the expectations that pre-generativist scholars had of the influence of linguistic theory on language teaching is evident in the following remark by one of their number (Moulton 1962: 187) when, in commenting on the new school of linguistics, he writes that transformational-generative grammar “can have far reaching effects in improving both the presentation of grammatical structure in textbooks and the learning of grammatical structure through classroom drill.”
None of this initial enthusiasm was, as I have pointed out, sustained. Of the passing references that one finds to a ‘cognitive’ approach in applied linguistic literature (cf., e.g., Terrell 1985: 467f.), many suggest that this is a label applied liberally to all kinds of teaching methods that focus primarily on language structure. This is borne out by Carroll’s (1965: 208), Smolinski (1985) statements in this regard, that “cognitive code learning theory ... may be thought of as a modified, up-to-date grammar translation theory.”
In sum, ‘cognitive’ trends in teaching second or foreign languages may be apparent when their principles are used, often vaguely and broadly, as t ustification for methods or approaches that leave learners to sort out and discover the grammatical structure and organization of an acquired language for themselves (cf. the term “cognitive-code learning”), and in this way can also provide theoretical justification for some communicative techniques, such as Asher’s Total Physical Response technique, and other variations it combined with in the Natural approach (Weideman 2002: 61; Krashen and Terrell 1983; Terrell 1985). It is in this sense that Klosek (1985: 14) refers to a “second wave” of applied linguists espousing principles that run counter to those of audiolingualism after the demise of the latter in the 1960s.
Transformational-generative grammar, as a linguistic theory, is therefore not ‘reflected’ in language teaching in the same way as the school of behaviorist structuralism. In fact there have been arguments that, at least as far as the standard theory of transformational-generative grammar is concerned, there exists no link between transformational analyses and the actual structure of linguistic knowledge in the mind, so that “it is a mistake to look to transformational grammar or any other theory of linguistic description to provide the theoretical basis for ... second language pedagogy” (Lamendella 1969: 255).
Similarly, Spolsky (1978: 3; see too 2010: 140) remarks that the “unholy alliance” of behaviorist psychology and linguistic structuralism - which led to the audio-lingual method - has, fortunately, not been reproduced by transformational grammar and cognitivist psychology, with the possible exception of what Spolsky, perhaps somewhat polemically (2008: 1), calls “the ungoverned chaos of the early natural approach”. Transformational grammar in the first instance exerts what influence it has in changing the attitudes and thinking of teachers and applied linguists to what constitutes the learning of a language, but not directly as regards prescribing the designed selection of instructional material. Cook’s (1985) article is a clear example of this. But Cook’s work was, in a sense, unique in relating Chomsky’s Universal Grammar to second language learning theory, for in attempts to define applied linguistics others have even claimed that applied linguistics “omits transformational-generative theory building from its sphere” while encompassing all kinds of “linguistic research and applications which happen to fall outside the scope of )ransformational-generative grammar” (Eisenstein 1985 : 9). Similarly, Klosek (1985: 14) states quite bluntly that while “linguistics is defined as transfor- mational/generative grammar. then applied linguistics has little to gain from and little to offer to linguistics.”
These historical views on the influence of transformational grammar are offered here, and, though audiolingualism chronologically preceded it, are offered first, because the influence or lack of influence of transformational grammar on language teaching constitutes a turning point in the way that applied linguistics itself is conceived. In all, the influence of transformationalism on language teaching can be said to be more indirect than the implications that linguists engaged in developing the so-called linguistic method, the “oral approach” or the audio-lingual method - which constitute ‘overlapping variants of the same tradition” according to Stevick (1971: 2; cf. too Fries 1945; Roberts 1982) - were drawing from their theories of language description. But since transformational-generative grammarians look askance at such direct applications of their theory, they have also made a significant contribution to the way that we look at the influence of linguistics in language teaching today. Transformational grammar has made applied linguists sensitive to the real (as opposed to supposed) relationship between linguistics and applied linguistics, and has also contributed to a revaluation of the supposedly direct linguistic inputs into language teaching designs, as was thought to have been achieved by audiolingualism .
In fact, with historical hindsight it is now clear, too, that the debt that audiolin- gualism itself owes to linguistics is much more indirect than is often claimed. Carroll (1971 : 110) notes that the emphasis in audio-lingual teaching on the aural-oral objective, though perhaps salutary from an educational point of view, has “little to do with )anguage learning )heory per se.” Also, in adhering to the principle that learning takes place by analogy rather than by analysis, audiolingualism itself admits a more indirect link between linguistic theory and language teaching. Linguistic analysis, as has been noted before, is in this case used rather in the selection of language materials to be taught in the second language classroom. In such a selection it becomes crucial, moreover, to “teach the problems”, i.e. the differences between mother tongue and target language, and so another level or technique that mediates between linguistic theory and teaching practice is introduced in the form of a contrastive analysis of the two languages in question (cf. Moulton 1962: 182; Smolinski 1985). Whether this ‘mediation’ (of contrastive analysis) between theory and practice has always been successful, is of course another question. Thus Marckwardt (1965: 242); Smolinski (1985) is sceptical regarding the efficiency of such analyses in all cases, remarking too that “we have been slow in translating these into simply written contrastive sketches which teachers might understand and apply...”
Be that as it may, such increasing and ongoing sophistication in designing appropriate teaching materials during that period provides evidence of the inability of ‘applying’ theory or theoretical description to practice without some form of mediation.
If there is any unadulterated and unmediated application of linguistic theory in language teaching along audio-lingual principles, it lies in the unquestioned acceptance of the (no doubt linguistic) idea that language is no more than structure. In this way, contrastive analyses of different languages reveal differences among them as purely structural differences.
As to the learning theory supporting audiolingualism, Carroll (1965: 210) claims that audio-lingual habit theory has no strict relationship to “any contemporary psychological theory of learning”, but rather echoes vaguely a version of Thorndikean association theory. So whatever theory is reflected in audiolingualism is more likely psychological theory (of a mechanistic, behaviorist kind) than simply linguistic theory; hence, as far as learning and teaching are concerned, “the emphasis of the audio-lingual habit theory was upon the formation of habits through practice and repetition” (Carroll 1971: 111; cf. too De Beaugrande 1997: 286). Because audio- lingualism subscribes also to a form of behaviorist, structuralist linguistics, it not only defines language as a set of habits, but also identifies these habits with the structural patterns of the language (even though no explicit reference to these patterns was made in the actual teaching, but they were again evident only in the selection and design of teaching materials).
The “linguistic principles” that identify post-1940 foreign language instruction, according to Moulton (1962), nonetheless are slogans like “language is speech, not writing”, “language is a set of habits”, “teach the language, not about the language”, and so forth. It is difficult to see how any of these can be called ‘principles’. They echo assumptions, and sometimes are (theoretical or practical) implications or, in Moulton’s terms, ‘conclusions’ drawn from such assumptions. Nevertheless, the belief was that such ‘principles’, being in some way derived from theoretical linguistic analyses, directly validated certain foreign language teaching methods. There is no doubt that in audiolingualism the assumption that the results of linguistic science could be brought to bear on the teaching of second or foreign languages came into full flower, even though ‘results’ must then be interpreted in a very broad sense as simply the implications underlying either the theory or its subsequent applications .
The most well-known formulation of this assumption is probably that of Lado (1964: 49ff.), who, backing up his formulations with references to the work of the psychologists Thorndike and Skinner, lists no fewer than seventeen ‘principles’ of a “scientific approach” to language teaching. It is instructive to see how Lado defends (or fails to defend) the derivability of such principles from linguistic theory.
In the first principle (“Speech before writing”) he mentions, namely that listening and speaking should be taught first, there is only the vaguest of attempts to show why; the clearest reason seems to be the assumption of then current linguistic theories that “language is most completely expressed in speech” (1964: 50). Such an assumption, however, is (or was) not the result of scientific analysis (which would yield what Lado calls “scientific information”), but rather the starting point without which the analysis of speech as the primary realization of language would in the first instance have been meaningless and without any justification.
What one has, at best, therefore, is an assumption underlying linguistic analysis that has uncritically been transferred to how language teaching must be done. The crucial question, namely whether the procedure of linguistic analysis (of starting with the analysis of speech) is at the same time the best procedure of learning or teaching a foreign language is never asked; instead, Lado assumes, without further ado, that it is.
Nor do the other ‘principles’ fare much better on this account. For the second principle, namely that “basic sentences” must be memorized, Lado simply claims “strong psychological justification”, but immediately adds that such j ustification nowhere exists in published experiments, although it has been “tested repeatedly otherwise” (1964: 51). Again, the “scientific information” from which language teaching would supposedly benefit is missing, as is the case with principle eleven (that the students must practise the language for most of the time).
For the third principle, once more, no justification is given: it is simply proclaimed that pattern practice is obligatory in order to establish these patterns as habits in the target language. What linguistic theory is reflected here, as well as in principles four (“Teach the sound system”), five (“Keep the vocabulary load to a minimum”), six (“Teach the problems”, i.e. focus on the structural differences between mother tongue and target language) and eight (“Teach the patterns gradually, in cumulative graded steps”), is again simply the leading assumption (and not the result) of linguistic analysis at this stage in the history of the discipline, namely that language is nothing more than structure. Moreover, the assumption that language must be taught gradually, i.e. in structurally more complex steps (first singular, then plural, and so forth - cf. Wakeman 1967), is a classic example of what Corder (1972: 16), Qvistgaard et al. (1972) calls the erroneous identification of ‘difference’ and ‘difficulty’, which are terms in two different theories, (contrastive) linguistics and learning theory, and is a claim that has since been thoroughly discredited (Lightbown and Spada 2006: 143).
In no single formulation of any principle does Lado explicitly state how this transfer from linguistic theory is to be accomplished, or in some, how the principles are linked to linguistic theory. One is asked rather to accept blindly or at least until “new scientific facts are added to our knowledge” (1964: 50) that it is justifiable merely to claim, as is done under principle thirteen, that linguistically, “a distorted rendition is not justified as the end product of practice,” though what kind of linguistic insight it is that yields this advice Lado fails to explain; or we are asked to “Teach primarily to produce learning rather than to please or entertain” (principle seventeen) on no other grounds but the (informal) observation that “classes that are the most entertaining are not always the most effective” (Lado 1964: 56), or, finally, that students must identify with the target culture (principle fifteen) because many teachers believe this.
Such statements on the ‘application’ of linguistics in language teaching would, no doubt, have been seen to be bordering on the absurd if it had not been for the aura of scientific truth in which they are dressed up. What is ludicrous upon subjecting them to closer scrutiny, however, becomes tragic when one is reminded that these principles provided the ‘scientific’ justification for one of the most influential approaches to the teaching of foreign languages in the previous century, the audiolingual method. Thus Marckwardt (1965: 241) could confidently claim at the first TESOL conference in 1964 that the aural-oral method, “the reflection of the linguist’s approach to language”, was firmly established.
We shall return to a discussion, below, of how such expectations of the ‘results’ of scientific findings have warped our understanding of applied linguistics, but wish to conclude here that (a) in audiolingualism we have one of the best examples of how that which is presented or popularly understood as the result of scientific, linguistic analysis actually consists only of the assumptions that underlie it, and (b) that such assumptions can only be vaguely (if at all) reflected in language teaching practice.
There is, however, a positive side to this debate. The difference in the influence of behaviorist structuralism and transformational-generative grammar has also paved the way for the emancipation of applied linguistics from being a discipline depending on a ‘direct’ input from linguistic theory (the so-called ‘consumer’ point of view - cf. Corder 1973: 10) to one with an entirely different orientation and perspective (cf. Chick and Seneque 1986: 4). In fact, it was at this stage in the development of applied linguistics that doubts began to arise as to the appropriateness of linguistic information; hence Johnson (1969: 238) could note that the lack of adequate criteria for the language content of courses can be blamed “on our willingness to simply accept linguistic data as the inviolate raw material” of these. And, as Brumfit (1980: 15) remarks, even though there have “been attempts to incorporate linguistic procedures into teaching in fairly undiluted forms”, the pedagogical intervention of the teacher will always crucially alter the nature of such supposedly unadulterated, theoretically conceived information.
There is no doubt that by the mid-1970s theoretical and pedagogical dissatisfaction with traditional teaching methods, including the audio-lingual method, was such that the time was ripe for change. After reviewing the contributions and applications (and in many cases, as we have noted, the lack of contribution, and misapplication) of traditional grammar, structuralist grammar and transformational-generative grammar, Roulet (1975: 75) notes that all “have failed to provide information on the use of language as an instrument of communication”. The stage was set for new practical, pedagogical notions on the use and teaching of language, as well as for sociolinguistic ideas, to appear.
We shall be returning below (Chaps. 6 and 8) to a discussion of the influence of the sociolinguistic idea of communicative competence in language teaching, yet wish to conclude this section with a preliminary observation that has a bearing on the change in the orientation of applied linguistics that had taken place as a result of its emancipation from linguistic theory as a controlling discipline.
One of the earliest papers dealing with the relation between the sociolinguistic idea of communicative competence and language teaching, Paulston (1974), takes it for granted that this idea should be reflected in language teaching:
If you accept Hymes’ notion that a model of language must be designed with a face toward communicative conduct and social life ... then it follows that a model for teaching language must also be designed with a face toward communicative conduct and social life. (Paulston 1974: 350)
However, there is a subtle shift here in what may initially look like a plea, once again, for the straightforward reflection of linguistic theory in the design of language teaching. First, Paulston is quite frank regarding the incompleteness of the theory at that time, and the need for it to be developed further. Second, and perhaps more important, is the statement that over the 5 years preceding the publication of the paper, i.e. since 1969, “there has been an increasing - and justified - concern for communicative activities in language teaching” (Paulston 1974: 348). This means that even before the first seminal ideas of Hymes and other scholars working on the idea of communicative competence became widely known in language teaching circles in the early 1970s, there were already signs within the teaching profession that communicative activities - the age-old promise of second and foreign language teaching, never quite fulfilled in older methods - were being introduced.
The conclusion this leads to, namely that the design of communicative language teaching activities may have preceded rather than followed theoretical developments, has profound implications for the view that one takes of applied linguistics. Unlike previous conceptions of the field, it forces one to see applied linguistic work rather as the j ustification of what is already established as teaching practice. Of course, such justification being an analytical activity, it will also involve criticism of the very activity it seeks to justify, and ultimately will involve suggestions for its (re-)design, modification, elaboration, development and change. But the conclusion cannot be denied, it would seem, that analysis must of necessity be analysis of something that already exists, however creative and imaginative the use to which the results of the analysis may be put.
Today, the unadulterated reflection of linguistic theory in language teaching is widely criticized. The imperialistic designs of linguistic theorists that are implied by the name of the field - applied linguistics - are also almost universally deplored (for an earlier review and acknowledgement of this, cf. Van Els et al. 1984: 128f.). Not only is the influence considered to be indirect or even irrelevant, since “few of the problems which arise in the classroom have linguistic answers, however indirect” (Corder 1972: 12; Qvistgaard et al. 1972), but it is also acknowledged that there are several intermediary steps between having a description or view of language and possessing a completed design for a language course, as Corder continues to point out. Not even in the use of the language descriptions yielded by linguistic work, often cited as “directly applicable” information for language teaching by applied linguists, can we speak of a direct reflection of linguistic theory, since these descriptions, “too, will have to be adapted using didactic criteria” (Van Els et al. 1984: 130).
Be this as it may, among applied linguists there may, if only for reasons associated with the history of the field, or the continuing awkward acceptance of its name, “applied linguistics”, remain some talk of the ‘least’ or ‘most’ applicable linguistic theory. This is why Roulet (1975: 77) could issue a plea for a reversal of the unidirectional conception of applied linguistics (as linguistic information that must become the input into language pedagogy), urging applied linguists to start, rather, “from the demands of teaching languages as instruments of communication and on that basis look at other approaches to language”. As the history of the further development and subsequent unfolding of language teaching designs illustrates, these “other approaches” are, of course, not formal linguistic, ‘restricted’ ones, but at the very least sociolinguistic ideas dealt with, for example, in the ethnography of speaking, or in an even more holistic perspective of the ecology of language and education (Hornberger and Hult 2008; Van Lier 2008), views that we shall return to in subsequent discussion.