SLA Topics     SLA Bibliography    Vivian Cook 

Strategies and UG Theories
Vivian Cook

unpublished conference at Seville ca 1996

A basic problem in second language acquisition research is the eclectic mixing of theories and research paradigms. In an area that is still developing rapidly and in which so little is yet known, it is clearly right not to go overboard for a single all-embracing theory of how second language acquisition works, in particular on account of the sheer complexity of second language acquisition itself, which takes in, not only all aspects of aspects of language acquisition and the environment found in first language, but also the additional factors of the age and range of learners and natural and classroom settings found in the second language. This catholicism of approaches is also necessary to deal with the complexity of language teaching; if teachers are to help students, they cannot afford to rely on just one theory of one part of acquisition for one type of learner but have to cater for the multiple needs of the diverse students in their charge.

However, this may lead to a feeling of complacency. As day by day we learn more about syntax, vocabulary and phonology in second language acquisition; we can provide richer and richer information for the teacher. Yet some of the theories and insights that are being pursued may not so much be complementary as antagonistic: if X is right then Y must be wrong. The first part of this paper explores some of the incompatibility between views of language common among linguists and the views of language implicit in much strategy research. The second part presents the current model of Universal Grammar known as the Minimalist Programme. The third part makes a gesture at reconciliation by seeing how strategies might fit the concept of lexically-driven acquisition.

1. Universal Grammar theory and second language acquisition

E versus I language

In the context of classroom observation, Cook (1990) saw a gulf between two approaches to language teaching in terms of the Chomskyan distinction between E (external) language and I (internal) language. An E-language approach regards language as an object independent of the speaker. It is concerned with the social relationships that people have through language, and the complex chains of discourse they weave between each other. As a theory of learning, E-language concerns how the learner is shaped by the environment around them, whether through samples of language or through interactions with other people.

I-language on the other hand is involved with language as knowledge in the mind of the speaker. Its first responsibility is to what is present in the mind of the individual. In linguistics this means establishing what a single person knows, often by introspection into the analyst’s own usage; a handful of invented sentences can test out hypotheses about the knowledge of language in the mind. As a theory of learning, it investigates how the system develops within the mind of the individual.

Where do strategies belong in this kind of analysis? Ostensibly they indeed appear to be concerned with the individual mind — what is a strategy if not an individual choice concerning how to communicate, how to learn, how to read, and so on? But the methodology of investigation has mostly been E-language in that a maximal total of strategies is established by reference to large numbers of learners, yielding a list which gets expanded over the years, sometimes with the very same examples (for example ‘airball’). The aim is, not the description of an individual butterfly, but the listing of all the possible characteristics of all possible butterflies. So the lists of communication strategies by Tarone (1977) or Faerch and Kasper (1984) or the learning strategies of O’Malley and Chamot (1990) are indeed taxonomic lists of E-language regularities rather than the account of an I-language of an individual. This makes their application to teaching extremely problematic; we have no idea what the makeup of a single individual is, or should be, and so have no grounds for amplifying or restricting their set of strategies to improve their L2 learning.

Competence versus performance.

The corollary to this is that strategies research is of doubtful value to knowledge of language because it is associated with the performance of language in actual situations rather than the underlying competence of the individual. If language is indeed knowledge in the mind, the first task in L2 learning is to describe this knowledge of a second language in the mind; the second is to account for how this knowledge is acquired; only the third is to see how this knowledge is put to use. These are then Chomsky’s agenda for linguistics rephrased for second language acquisition. If it is right to start from knowledge and the acquisition of knowledge in the study of the first language, then it is equally right in the study of the second. Use of language can be seen as dependent upon the prior knowledge that is applied. Long discussions about strategies for investing our money have no point if there’s no money to invest in the first place. Hence second language acquisition and language teaching should concentrate on the questions of what second language knowledge is and how it is created before they start thinking about the strategies in which it used. Communication and compensatory strategies are luxuries to be dealt with after we have gone about the serious business.

To pursue the analogy, however, we can discuss how to create the nest-egg for investment in the first place—learning strategies. Perhaps what we need is strategies for putting minimal savings together, which can then be subject to a proper investment policy. Here the argument from linguistics takes a different twist in that it argues that knowledge is not created by conscious strategies of any kind but comes into being of its own accord through the operation of Universal Grammar in the mind. Children do not learn language as such. They utilise the already existing principles in their minds and set their variable parameters in response to the language input they receive. The process is not under their conscious control and is not affected by properties of the language input, provided that there is indeed some input.

The psychological tradition on the other hand has denied that the distinctions between E-language and I-language and between competence and performance are valid. If the knowledge of language is not separated from its use, using is the same as knowing. Learning strategies for using language are then the same as for strategies using it. Most of the second language research into strategies comes basically from this tradition which sees the learners’ activities as all-important. It is common for example to cite John Anderson (1983) as a major influence, whose work denies the separation of language from other aspects of the mind and whose shift from declarative to productive memory has no link with linguistic theories. It seems an example of Orwellian double-think to do strategies work within a linguistics framework. There is no way of reconciling the belief in an independent language faculty with that in a set of learning strategies of the type espoused in the literature.

2. Current UG theory.

Some would say therefore that combining UG and strategies in one breath is an oxymoron on a par with bitter-sweet or popular John Major. Having taken this strong line, is it still possible to find a bridge between UG theory and strategies? The problem is that the definition of Universal Grammar changes from one day to the next. Let us first outline the classic picture and then make some speculations about how multi-competence theory may help.

During the 1980s, a model of Universal Grammar was devised and applied to a large amount of L2 research, based on Government and Binding theory, after Chomsky (1981), and explained in books such as Cook (1988). This claimed not only that the important starting point was the speaker’s knowledge of language, competence, but also that this consisted of two main types of information, principles and parameters. Principles of language were abstract restrictions on what language could be, for example the structure-dependency principle which forced movement in questions to use the structure of the sentence rather than its linear order. A more wide-reaching principle was the Projection principle which insisted that the lexical properties of lexical items project onto the sentence. That is to say, the fact that the lexical entry for prefer specifies a Subject and an object means that we can say Helen prefers Scotch but not Helen prefers. This principle tied the structure of the sentence closely into vocabulary; it was hard to separate the two.

As well as the principles of language which were essentially uniform and never breached by any language, this model used the concept of parameter to capture the variation between languages. A well-studied example is the pro-drop parameter which distinguishes languages which need to have the subject in the sentence from those that do not. Thus in Italian the sentence ‘he speaks’ can be parla with no visible subject, in Arabic yatakallamu with no subject, in Chinese shuo with no subject and so on. These are called pro-drop languages. Languages that belong to the group of non-pro-drop languages must have a visible subject in the sentence. These include English he speaks, German er spricht, French il parle, and so on. The metaphor is that learners set the parameters of Universal Grammar to the value of the language that surrounds them, ending up with pro-drop if they hear Spanish, non-pro-drop if they hear French, and so on. The crucial aspect of this version of Universal Grammar was the concept of principles and parameters; gradually the name for the whole model became principles and parameters theory as this captured the main elements.

However, this picture has changed radically in the 1990s under the pressure of a new Chomskyan model called the Minimalist Program; the crucial papers are gathered in Chomsky (1995); an introduction can be found in Cook and Newson (1996) This built on some of the elements we have already seen, in particular by effectively uniting the Projection Principle and parameters. If sentences consist of projections from the dictionary, why not make parameters part of the dictionary? Hence all the variation between languages is pushed into the lexicon. ‘There is a single computational system for human language and only limited lexical variety’ (Chomsky, 1995). Language consists of invariable principles regardless of which language it might be; all the differences were now associated with the parameter setting that went on in the lexicon. Learning a language means learning vocabulary; everything else does not have to be learnt. ‘Language acquisition is in essence a matter of determining lexical idiosyncrasies’ (Chomsky, 1995). This is not to say that it is vocabulary in the conventional sense, i.e. a list of meanings. Rather it is the specification of how the lexical items may be used in the sentence along with the parameter setting.

The second development needed to complete this was an extension of the type of phrases and their incorporation in the lexicon. In the late eighties several researchers had proposed the concept of functional phrases, which meant that as well as conventional phrases built round nouns or prepositions there were also functional phrases built around the grammatical element in the sentence—inflection phrases, tense phrases, negation phrases, and the like. These were seen as the syntactic core of the sentence into which the lexical phrases fitted. Hence these were now given entries in the lexicon and related to the variable parameters. Languages therefore varied because the parameters attached to the entries for phrases had different values. ‘Variation of language is essentially morphological in character.' (Chomsky, 1995).

We can now sum up the current picture of acquisition with the Minimalist program in Chomsky’s words: ‘Acquisition of a language reduces to selection of substantives from a given store and fixing of values of parameters that apply to functional elements and to properties of the lexicon as a whole’ (Chomsky, 1995). Learning a language means acquiring vocabulary and setting parameters for these morphology-based aspects of the lexicon. Learning a second language is a variation on this theme, like all language learning (That is to say if L2 learning uses a natural approach at all—the grammar-translation method, say). So the most crucial aspect is then conceived as the acquisition of vocabulary, not only the content words that one usually associates with vocabulary with their projections into the syntax but also the morphological endings with their parameter-settings that influence the structure of the sentence in profound ways.

While this is still called principles and parameters theory as these are the crucial aspects, it differs radically from the model that was used in most second language acquisition research on Universal Grammar in the 1980s, for example White (1989), which is now effectively left stranded by these developments like a dying whale. The arguments about whether second language learners have access to UG need to be completely rephrased and the actual research needs to be reinterpreted or abandoned. Of course this is only true if we wish to keep the L2 research in touch with the latest version of Universal Grammar theory: this can be called the view of the strong believers. An alternative view is that L2 researchers should take what they find useful, whether Chomsky (1995) or Chomsky (1957) or Jespersen (1904); the concept of structure-dependency or the idea of phrase structure can be used all for their own ends, as has been argued elsewhere (Cook, 1993). This faction can be called weak believers. It is fair to say that UG researchers have seen themselves primarily as linguists, i.e. as strong believers, secondly as L2 researchers; hence they have notoriously always tried to keep as up-to date as possible; Chomsky catches cold and they start sneezing next day. It will be entertaining to see what emerges out of this new cataclysm to befall linguistics. The weak believers, however can think how the new developments can be harnessed to their own ends of studying second language acquisition and helping second language teaching; what’s in it for them?

3. Exploiting Minimalism

Let us sum up the argument as three linked premises:

Premise 1: knowledge of language consists of unvarying syntax and a set of variable lexical entries including parameters showing how these may be used in sentences

The first step is the definition of the target that people are learning as consisting of a particular type of knowledge. Languages differ only in their vocabularies, interpreting vocabulary as having two components, lexical entries for content words such as man that project into lexical phrases, and lexical entries for functional phrases like ‘tense’ that have parameters to be set that influence the variation in the structure of the sentence. An issue that we will not deal with here is whether, when applied to second language acquisition, this model implies two grammars, one for each language, or implies a single grammar with entries for both languages.

Premise 2: language acquisition means acquiring lexical entries for words and phrases with appropriate parameter settings for using them in the structure of the sentence

Principles of grammar do not have to be learnt and are invariant from one language to another. Content words, however, vary and have to be learnt with all their relevant information. Functional phrases too need to be acquired both in terms of content and in terms of parameter-settings.

Premise 3. appropriate learning strategies in L2 therefore rely on the construction of lexical entries in the mind

The final jump is taken in the third premise, which brings us back to the issue of strategies. The strong believers in Universal Grammar would disdain this step as first language acquisition is not organised in this way and parents do not ‘teach’ the first language by invoking vocabulary strategies. Yet a host of strategies have been used in L2 teaching, for example those associated with Andrew Cohen.

But most of them, to my knowledge, have tackled content vocabulary, rather than functional phrases. The past tense has been seen as something to be taught as a grammatical pattern, not as a vocabulary item “-ed” with a particular meaning that attaches itself to verbs. The nearest I can remember to a mnemonic for learning grammar is the immortal verse beloved of students of Latin:
A, ab, absque, coram, de,
Sine, tenus, pro and prae,
Always govern the absolute i
Vocabulary teaching has then made a mistake in seeing its role as the imparting of content words. In terms of frequency the top 46 words in English are after all grammatical words, not content words If everything that is learnt is vocabulary, we need to explore more effective ways of teaching vocabulary and more efficient strategies by which the learners can assimilate vocabulary and fit it in with the principles of grammar in their mind. Vocabulary strategies would then be useful that allowed the learner to flesh out the lexical entry in their minds by giving it, not just a meaning and a pronunciation, but the appropriate links to the syntax of the sentence. As always it is not clear that conscious attention to ‘learning’ or to learning strategies is incompatible with the UG model. If it were the most useful strategies for the learner would be ways of building up the knowledge of vocabulary that forms the basis for the language-specific aspects of language knowledge.


Anderson, J. (1983), The Architecture of Cognition, Harvard University Press

Chomsky, N. (1957), Syntactic Structures, Mouton, The Hague

Chomsky, N. (1981), Lectures on Government and Binding, Dordrecht, Foris

Chomsky, N. (1995), The Minimalist Program, MIT Press

Cook, V.J. (1988), Chomsky’s Universal Grammar: An Introduction, Blackwell, Oxford

Cook, V.J. (1990), ‘The I-language Approach and Classroom Observation’, English Language Teaching Documents, 133, 71-80, 1990

Cook, V.J. (1993), Linguistics and Second Language Acquisition, Macmillan, Basingstoke

Cook, V.J. & Newson, M. (1996), Chomsky’s Universal Grammar: An Introduction, Blackwell, Oxford (2nd edition)

Faerch, C., & Kasper, G. (1984), ‘Two ways of defining communication strategies’, Language Learning, 34, 45-63

Jespersen, O. (1904), How to Teach a Foreign Language, Allen & Unwin, London

O'Malley, J.M., & Chamot, A.U. (1990), Learning Strategies in Second Language Acquisition, CUP

Tarone, E. (1977), ‘Conscious communication strategies in interlanguage: a progress report’, in On TESOL 1977, TESOL, Washington

White, L. (1989), Universal Grammar and Second Language Acquisition, John Benjamins, Amsterdam