Vivian Cook   SLA Home
Multilingual Universal Grammar as the norm

In I. Leung (ed.) Third Language Acquisition and Universal Grammar,
Bristol: Multilingual Matters, 55-70, 2009.
For earlier attempts see Chomsky's Universal Grammar 1985, Universal Grammar and multi-competence 2007 and The poverty of the stimulus argument and L2 users of English 2003. Latest 2017 More Galilean challenges for Chomsky

1. Introduction

The Universal Grammar tradition in linguistics and SLA research has treated monolingualism as the normal state of mankind. Yet this monolingual assumption is nowhere inherent in its avowed aims. Chomsky’s questions for linguistics, for example Chomsky (1991), concern the knowledge of language in the mind and its acquisition, use and storage:

1. What constitutes knowledge of language?
2. How is such knowledge acquired?
3. How is such knowledge put to use?

But what does language mean in these questions? Cook (2007; submitted) distinguishes five meanings of ‘language’, most of which will be referred to in the following discussion:

Lang1. human representation system
Lang2. an abstract external entity
Lang3. a set of actual or potential sentences
Lang4. the possession of a community
Lang5. the knowledge in the mind of an individual

While Chomsky’s questions have overtones of the general Lang1 meaning, ‘human representation system’, they are centrally concerned with the Lang5 meaning ‘the knowledge in the mind of an individual’. None of them specifically refers to knowledge of a language (Lang5): the notion of language in general (Lang1.) is different from the idea of a particular language (Lang2.), as indeed Chomsky recognises with his use of ‘grammar’ for the language in the mind:

'The grammar in a person's mind/brain is real; it is one of the real things in the world. The language (whatever that may be) is not' (Chomsky, 1982: 5)

In recent years Chomsky has turned to a new question about the perfection of language:  

‘How good a solution is language to certain boundary conditions that are imposed by the architecture of the mind?’ (Chomsky, 2000:17)

Again this is phrased in terms of  language rather than of a single language. So far as these basic questions are concerned, the knowledge of language in the mind could include as many languages as it is capable of holding.

So why is a mind that knows two languages treated as an exception rather than as the rule? The usual reason put forward, mostly in footnotes and interviews rather than argued in the UG texts themselves, is that restriction to monolinguals is a necessary simplification. Linguistics needs to work with an abstraction from which irrelevant aspects have been purged:

'Linguistic theory is concerned with an ideal speaker-listener in a completely homogeneous speech community' (Chomsky, 1965: 4).

Though linguistic competence has rarely been debated in the UG field since around 1970, it implicitly underpins most analysis. Given that this version of linguistic competence excludes all the other aspects of the speaker as a psychological and social being, it is a small step to eliminate knowledge of more than one language:

Not that anyone denies that many people in the world have two languages:

‘In most of human history, and in most parts of the world today, children grow up speaking a variety of languages …. That is just a natural state of human beings’ (Chomsky, 2000: 59).

It seems that the analyst cannot cope with the complexity of the bilingual situation.

But do second language users (henceforward L2 users) think of themselves as native speakers of one language with another language added on? Let us hear from two of them. First Edward Said, a Palestinian in exile:

‘I have never known what language I spoke first, Arabic or English, or which one was really mine beyond any doubt. What I do know, however, is that the two have always been together in my life, one resonating in the other, sometimes ironically, sometimes nostalgically, most often each correcting, and commenting on, the other. Each can seem like my absolutely first language, but neither is.’ (Said, 1999)

Next from Suresh Canagarajah, a Sri Lankan Tamil living in New York :

‘In South Asian communities, such as mine we grow up with two or more languages from childhood, developing equal competence in all of them, fluidly moving between each of them in our everyday life according to the different domains of family (regional dialect of Tamil), school (English), neighbourhood (Muslim dialect of Tamil) and governmental institutions (Sinhala) … One can imagine the difficulty for people in my region to identify themselves as native speakers of “a” language.’ (Canagarajah, 2005: 16-17)

Neither of them accepts that they have added one language to another to become L2 users. Rather the two or more languages have always been interwoven into their lives.

We shall first review some of the arguments for a multilingual UG and then look at the consequences for UG theory.

2. Arguments for multilingual UG

The arguments for a multilingual UG come from a variety of sources, some directly within UG, some general issues; some of these have been presented in a slightly different context in Cook & Newson (2007).

2.1 The argument of independence from environmental variation

One of the key elements in UG theory is the vast range of circumstances in which human children acquire language. In some cultures children are literally not spoken to, in others they are bombarded with speech; some parents speak to their toddlers in baby‑talk, others refuse to adapt. It doesn’t seem to matter what kind of up-bringing the child encounters: the sole requirement for learning a human language is to be human, as Lenneberg (1967) observed. Only a handful of children fail to learn language. But of course language acquisition fails in the rare cases when children do not encounter speech at all, such as the extreme deprivation suffered by Genie (Curtiss, 1971). The theory assumes nothing more about the child’s language environment than the sheer availability of language input.

So whatever Universal Grammar theory proposes for language acquisition has to be robust enough to work whatever the conditions that the child encounters. Language acquisition cannot depend on particular properties of the input, say, the frequency of this or that feature; it cannot depend on particular types of interaction, whether recasts, corrections or whatever, simply because children acquire language regardless. The powerful device for acquiring language in the child’s mind is omnivorous rather than being fussy about its diet. Any theory of UG must postulate mechanisms for learning that will work in any possible human child-rearing situation short of total deprivation. It has to see language knowledge as triggered by ubiquitous properties of language and of the child-rearing situation, not by accidental features of some individual or cultural situation.

Yet children exposed to two or more languages acquire all of them – childhood simultaneous bilingualism. The mental device for acquiring language can cope with two or more languages, at least in the early years.

So why don’t monolinguals speak two languages? For the reason that, like Genie, they do not hear them. They are restricted to the knowledge triggered by the set of sentences they encounter. In other words monolingualism can be considered as a widespread form of language deprivation. A child encountering two languages acquires two languages; a child encountering one acquires one. The only thing which prevents the child from learning two languages is deprivation of a second language. It is therefore an accidental environmental feature when children are deprived of a second language. It is language deprivation in that it is deprivation of a language rather than linguistic deprivation when no language at all is supplied. The UG model, since it does not recognise such accidents, cannot take the acquisition of one language as the norm that it has to account for but is bound to base itself on multiple acquisition of languages, taking monolingualism as a sub-category occurring in the absence of second language input. The device for learning language in the child’s mind has to be capable of handling more than one language at a time. From the beginning monolingualism needs to be seen as an input-dependent constraint on the processes of language acquisition, that are perfectly capable of handling two languages at a time.

Let us briefly look at the notion of input. Lang3 is language as ‘a set of actual or potential sentences’. This sense recurred throughout twentieth century linguistics:

‘a set (finite or infinite) of sentences, each finite in length and constructed out of a finite set of elements’ (Chomsky, 1957: 13).

It was brought into SLA research via a definition of interlanguage:

‘the utterances which are produced when the learner attempts to say sentences of a TL [target language]’ (Selinker, 1972).

While Lang3 seems a fairly precise meaning, it is, however, difficult to operate with once two languages come into the picture: how do we know that a set of sentences contains some sentences from one language, some from another? The input to many, if not most, children contains things that belong to more than one language in a Lang2 sense. The logical problem for language acquisition is not just how the child acquires language from input but how they manage to separate two languages from an input that is not tagged as being language A or language B.

2.2 The argument of integrated language systems

Weinreich (1953) described three relationships between the languages in the bilingual mind: compound, coordinate and subordinate. Current UG-oriented SLA theory assumes a compound relationship: the two grammars are independent creations, two instantiations of Universal Grammar, not a single grammar, even if they have connections and links. One mental grammar has, say, set the pro-drop parameter to pro‑drop, another has set it to non-pro-drop. An L2 user has two instantiations of the parameter with different settings, not one instantiation of the parameter with a variable setting according to the language being used.

Cook (2003) proposed an integrative continuum on which the possible relationships between the two languages are spread between the two poles of total separation and total integration. That is to say, the languages are constantly related, whether at the level of vocabulary, phonology or syntax. Acquiring a second language is not propping a lean-to against an existing house; it is rebuilding the property itself. An L2 user has not just added a second language to a impervious first language; they have created a complex overall system where L1 and L2 are inextricably tied together. It would not be surprising to find that the presence of a first language affects the second (transfer) and that the presence of a second affects the first (reverse transfer and L1 attrition); see Pavlenko & Jarvis (2007) for an exhaustive account. The mental grammars of the two languages are not isolated but interact to a greater or lesser degree.

If UG theory is to account for the language in the mind, it has to be flexible enough to accommodate this overall unity from the beginning. Indeed this has sometimes been allowed within UG; Universal Bilingualism permits more than one grammar to allow for transitional L1 stages at which both are in operation (Roeper, 1999);

‘whatever the language faculty is it can assume many different states in parallel’ (Chomsky, 2000: 59).

The human mind may contain knowledge of an indefinite number of languages. If UG takes the single grammar as the norm, it cannot account for the composite system of multiple grammars, only for a default where the mind contains a single grammar.

2.3 The argument from numbers

Once in nineteenth century England , it may have been easy to conclude from everyday experience that most people only speak one language. Now five minutes walking on the streets of any English city, whether Newcastle upon Tyne or London, soon shows how many other languages are being used and how many different non-native accents of English are being spoken; a survey of London uncovered 300 different languages and 32% of children who spoke languages other than English at home (Baker & Eversley, 2000): bilingualism is rife. And the same would apply to almost any city or town across the globe. In the European Union 83% of young people have studied a second lang­uage (Commission of the European Communities, 1987); in Luxembourg 53% speak more than one language with their friends, and 56% speak more than one language in their workplace (European Union, 2005).

Most modern societies are multilingual in nature, however many official languages a country may recognise. While it is impossible to count the number of L2 users in the world today, they probably outnumber monolinguals:

'even in the United States, the idea that people speak one language is certainly not true … everyone grows up in a multilingual environment' (Chomsky, 2000: 59).

The invisibility of L2 users for language and linguistics is partly a product of the emphasis on double monolinguals; the people that counted are ‘balanced bilinguals’ that behave exactly the same in two languages, not those that use L2 effectively but differently from native speakers, perhaps the overwhelming majority.

A second group of L2 users may outnumber those living together in multilingual societies, namely those who use a language they have been taught for purposes of their own; an estimated billion people are learning English as a language (British Council, 1999). English has an international L2 user community across the world for whom the native speaker community is virtually irrelevant; it is the interaction of academics, businessmen, tourists and others with each other and with non-native communities that matters; 74% of tourist use of English around the globe is between non-native speakers (Graddol, 2006). Languages like Latin and French previously had such a global role; Chinese may come to have a similar role with the spread of the study of Chinese as a second language. These are another type of L2 users: people who function in a second language for professional reasons in countries where the second language has no official status.

Adding together the multilingual groups and the lingua franca users produces a massive grand total of people for whom second languages are an integral part of their lives. Through its simplification to monolingual native speakers, Universal Grammar is ignoring the language knowledge in the minds of probably the majority of the human race.

2.4 The argument from discrimination

The key figure in Universal Grammar is the idealised monolingual native speaker. Labov’s classic argument held that one group should not be measured against the norm of another (Labov, 1969); the comparison between groups yields differences, not deficits. In principle everybody is capable of the same richness of expression albeit in different ways, regardless of the language they speak, the situation they find themselves in or the characteristics of their individual minds. Hence no group of people have a better knowledge of UG than any other group; we all share the same UG heritage.

Over the years this concept of equality has been applied by linguists to:

- speakers of different languages
‘The lowliest South African bushman speaks in the forms of a rich symbolic system that is in essence perfectly comparable to the speech of the cultivated Frenchman’ (Sapir, 1921: 22). There is no way in which a linguist could see one language as better than another, even if this is often the view of the person in the street and of popular pundits on language.

- children and adults. Chomsky’s 1960s views on the independence of children’s grammars led to children being treated in their own linguistic right rather than as defective adults, with occasional exceptions such as Smith (1973). Children’s grammars were not adults’ grammars seen through a distorting lens but different grammars.

- speakers of different dialects. In a linguistic sense all regional dialects are equal, as Trudgill (1979) argued: Geordie dialect is not better or worse than London in a linguistic sense, merely different. Only for social reasons can a particular dialect may get prestige and status over other dialects as say an RP accent has greater status than Geordie.

- speakers of different social dialects. Middle-class speakers may well speak differently from working-class speakers, as Bernstein (1971) and others suggested. This does not mean that objectively one class speaks better or worse than another class (Labov, 1969).

- black and white speakers. 1960s language intervention programs like Sesame Street were predicated on the notion that black English was defective compared to white English. Labov (1969) showed that these were two dialects of the same language and there was no reason for claiming one was more logical or better than the other, except for the social implications of high status and low status languages.

- men and women. Attempts to make women speak like men and vice versa would now be treated as ludicrous, only surviving perhaps in well-intentioned courses in conversational assertiveness for women. Whatever language differences there may be between the sexes, no-one would claim one is worse than the other.

So people  should not be expected to conform to the norm of another group to which they do not belong, whether defined by race, class, sex or whatever. People who speak differently from some arbitrary group are not speak­ing better or worse, just differently.

Native speakers however are treated as a case apart; it is assumed that non-native knowledge of language only exists in relation to native knowledge. Succeeding in second language acquisition means speaking it like a native speaker (and usually a speaker of a standard status form, not a dialect);

‘Relative to native speaker’s linguistic competence, learners’ interlanguage is deficient by definition’ (Kasper & Kellerman, 1997, 5)

Ultimate attainment means native-like speech;

‘absolute native-like command of an L2’ (Hyltenstam & Abrahamsson, 2003)

The older the age of the second language learner the less able they are to sound like native speakers:

 ‘…some learners can achieve very high levels of native-like pronunciation in mostly constrained tasks but [we] have yet to show that later learners can achieve the same level of phonology as native speakers in production’ (DeKeyser & Larson Hall, 2005: 96)

And so on. The only true knowledge of a language is that of a native speaker, the only pure form of linguistic competence that of the native speaker; anything else has to be measured against it. It seems strange that, step by step, linguists have conceded that one group does not have inferior language to another, whether by race, class and sex, but are not prepared to see speakers of two languages as people in their own right measured by their own standards rather than those of a group to which they can never belong by definition – native speakers. The counter-argument is that non-native speakers are a unique group genuinely deficient compared to another group; this last-ditch defence was doubtless put forward for all the other groups listed above, whether children, working class or black.

2.5 The argument from cultural bias

All human sciences have to try to be on their guard against projecting the values of their own societies onto others. Time and again in language-based studies some construct is shown to be the consequence of the observer’s preconceptions about language, say the discovery that relative orientation is not a universal for human beings (Levinson, 1996). Indeed many of the constructs of Western linguists have been traced back to the constraints of the alphabetic writing system used by their L1 writing systems, whether the phoneme as an artefact of letters (Aronoff, 1992) or the word as the artefact of using word-spaces to help reading (Olson, 1996).

Assumptions about bilingualism often betray the monolingual societies from which linguists came, in which learning another language was either an unusual intellectual feat carried out in universities or a considerable problem when carried out by the members of a minority ethnic community or by immigrants. Feat or challenge, acquiring a second language is never something ordinary to be taken for granted, as it is in countries where daily use of multiple languages is the norm, as in India and the Cameroon .

UG linguists are not immune to the general inclination for western societies to consider bilingualism as a problem rather than an asset. Bilingual children have needed special attention in all sorts of ways, and have threatened the monolingual standards of the classroom; the threat from immigration is often perceived as a threat to the society. Linguistics shows the background of linguists. A second language is seen to confer partial membership of another monolingual community rather than full membership of a bilingual community – what Brutt-Griffler (2002) terms the ‘multi-competence of the community’. The language of a mixed community:

'would not be "pure" in the relevant sense, because it would not represent a single set of choices among the options permitted by UG but rather would include "contradictory" choices for certain of these options’ (Chomsky, 1986: 17).

An L2 user becomes a secondary member of another monolingual community rather than part of a multilingual community of L2 users. So far as our overall argument goes, UG linguists have to convince that the purity they ascribe to the monolingual is not the projection of their own cultural belief about the normalcy of monolingualism. 

2.6 The argument from the uniqueness of the L2 user

If we accept that L2 users exist as people in their own right and have as much claim to know their first and second languages as a monolingual, we need to establish their characteristics. Let us review some of their differences (not deficits!) from monolingual native speakers.

- different knowledge of the second language

It hardly needs to be said that the language of L2 users differs from that of monolingual native speakers. Years of research time have been devoted to trying to find cases where ultimate attainment is the same as native speakers. Undoubtedly there are cases where, in some aspects of language, non-native speakers can have the same linguistic knowledge as native speakers, say the Dutch speakers of English studied in Bongaerts et al (1997). But these people are as relevant to the study of second language acquisition as opera singers are to the study of phonology; the vast majority of L2 users are different from native speakers, however successful they may be as L2 users. If L2 users have set the UG parameters differently from monolinguals, this reflects the constrained world of the monolingual not deficiencies in their acquisition.

- different knowledge of the first language

It does need more argument that the L1 of the L2 user differs from the L1 of a monolingual native speaker of the same language. Let us take some quick syntactic examples: French speakers who know English react against French sentences using the middle voice Un tricot de laine se lave à l’eau froide. (*A wool sweater washes in cold water) compared to those who don't know English (Balcom, 2003); ‘near-native’ Greek learners of English produce far more definite preverbal subjects in Greek than monolingual native speakers (Tsimpli, Sorace, Heycock & Filiaci, 2004); Japanese, Greek and Spanish speakers of English prefer the first noun to be the subject of the sentence in The dog pats the tree (translated into their respective languages) to a greater extent than those who do not know English (Cook, Iarossi, Stellakis & Tokumaru, 2003). In other words the L1 of the L2 user differs syntactically from the L1 of monolinguals in subtle ways.

- different language uses

The language abilities of L2 users have usually been discussed in terms of what L2 users lack: anything that differs from the monolingual native speaker is a deficiency. What about the other way round? What can L2 users do that monolinguals can’t? The main categories are those functions that crucially rely on the presence of two languages in the same mind.

One is the process of translation; some, but not all, L2 users can hear or read a text in one language and produce an equivalent text in another. A common situation round the world is the children of immigrants acting as translators for their parents. At the top levels this can be done professionally in the highly skilled arts of simultaneous and consecutive interpreting. Curiously enough, Universal Grammar has taken little interest in this ability of the human mind, presumably seeing it as a specific aspect of performance. Yet the ability to function in two languages simultaneously must show both the close relationships of the languages in the mind and the ability to keep them apart. Which of course monolinguals are incapable of as they don’t have two languages.

The other is the process of code-switching. Many L2 users switch between their languages during the process of speech, from one word, clause or sentence to the next, according to a host of rules about the social situation, the topic being discussed and the grammatical overlaps between the two languages (Milroy & Muyskens, 1995). Code-switching shows that it is perfectly possible to use two grammars in the mind simultaneously. One alternative is to see this as ultra-swift switching between the two grammars during production or comprehension. The more economical choice is to see it as switching within the single overall grammar for both languages. Code-switching is  a simple everyday activity for L2 users that monolinguals cannot do. A Universal Grammar theory that cannot take this in is deficient. The argument that these skills are simply an extended version of what monolingual native speakers do when switching dialect or paraphrasing may indeed be true but here is immaterial since this could still not be accommodated with Universal Grammar theory.

This is not the place to detail the other cognitive differences of the L2 user, described in say Cook (2007). The L2 user is not just a monolingual with another language but someone whose mind has been transformed by knowing two languages: bilinguals think differently.

At the moment we have no idea how much L2 it takes to affect the L1; most research has dealt with fairly advanced L2 learners. Yet primary school English children who learn Italian for an hour a week for five months learn to recognise English words better than monolinguals (Yelland, Pollard & Mercuri, 1993); even so small an exposure can have an effect. It is unsafe to assume that such reverse transfer only occurs at high levels of L2 proficiency.

- different knowledge of other languages

Those working in the developing area of multilingualism such as Cenoz et al. (2001)  have shown that we should not consider only the first and second languages but also further languages. Cantonese L1/ English L2 speakers learn L3 French better than Vietnamese L1 /English L2 speakers (Leung, 2005). English L1/Spanish L2 speakers transfer 97% of L1 function words to L3 Italian, 3% from their L2; Spanish L1/English L2 speakers 19% from their L1, 19% from their L2 (De Angelis, 2005). The sequence of acquisition of multiple languages as well as their differences affects the resulting linguistic competence. Knowledge of language is different in trilinguals. Again there must be large numbers of people in the world who know more than two languages, particularly in countries such as India and the Cameroon . A monolingual UG theory reduces multilingualism to an exception to an exception; multilingual UG treats it within the same framework as monolingualism and bilingualism.

3. Consequences for UG theory

UG theory is failing if, so far from accepting L2 users as having one of the basic types of language knowledge, it dismisses their knowledge as a defective version of the monolingual’s. Suppose that the starting point for UG theory is not a language in the mind but language in the mind, consisting of one or more languages. What are the consequences for UG theory in general and for UG-related SLA research?

3.1 Consequences for the poverty-of-the-stimulus argument

Monolingual input can be seen as language deprivation which supplies triggers for only one parameter-setting, say pro-drop. The unrestricted multilingual environment provides enough material for two or more simultaneous parameter-settings, say both pro-drop and non-pro-drop. The poverty-of-the-stimulus argument describes how certain aspects of language are unlearnable from the Lang3 set of sentences encountered by monolinguals, either pro-drop or non-pro-drop. With multilingual UG, the poverty-of-the-stimulus argument has to be defined as whether these aspects are learnable from a Lang3 set of sentences including examples of both pro-drop and non-pro-drop. The mind does not normally have to deal with monolingual input but with multilingual input; the interesting question is how the child, who does not receive an input neatly labelled as language A and language B, manages to acquire two parameter-settings from one undifferentiated input – a fascinatingly changed research question for UG theory. The problem is how the child manages to gate the languages to keep the settings etc. apart.

3.2 Consequences for the grammar

The form of the grammar now has to be such that it can simultaneously have two settings for each parameter in the same mind, not dodged by treating multilingualism as an endlessly reiterated monolingualism. The UG-based SLA research has mostly talked of two grammars, one per language; grammars clone as and when necessary. So there are two copies of  the parameters; the issue becomes whether the initial setting for the second grammar parameters is neutral, default or L1, as seen in the hypothesis turf-wars of the 1990s (Cook & Newson, 2007). If the mental grammar is multilingual, the problem is how the two values for a parameter are simultaneously available to the L2 user; the monolingual is an oddity in that they are restricted to one value. (Perhaps this means the metaphor of parameter-setting is too restricting; an alternative would be gearboxes, the monolingual having one gear, the L2 user two or more). Satterfield (2003) argues on evidence from pro-drop in Spanish speakers in the USA that a level above parameter-setting is involved, namely in the computational system itself where features are calculated on-line and the multilingual has to avoid the costliness of two operations. This must be a property of all minds that is unused by monolinguals; to account for the language faculty, UG theory has to accommodate the multilingual mind. '[B]ilingualism should stand as a rigorous barometer for measuring the feasibility of our most developed linguistic theories' (Satterfield, 1999: 137).

3.3 Consequences for research methods

The arguments have shown that people who know two languages differ in many unexpected ways from monolinguals. The straightforward consequence for UG-based research is that, to learn about pure monolingual grammar, you have to ask pure monolinguals. The L1 of L2 users may have been ‘contaminated’ by their L2:

‘the judgements about English of Bloomfield, Halliday or Chomsky are not trustworthy, except where they are supported by evidence from “pure” monolinguals’ (Cook, 2002: 23).

Indeed the research method of asking for grammaticality judgements may itself be suspect since L2 users have greater metalinguistic ability (Bialystok, 1991). Descriptions of language for monolingual UG have also to be based on the usage of these pure monolinguals; an Italian who knows English may no longer use a pro-drop Italian; in phonology Kato (2004) showed that the standard values given for Voice Onset Time for Japanese differed from those for ‘pure’ monolinguals because they had been measured on Japanese living in the United States. Monolingual UG has to clean up its act by ensuring that it is indeed looking at pure monolinguals of the idealised 1965 type and the researchers are either pure monolinguals themselves or are capable of distancing themselves from the effects of their L2 on their L1. But the search for pure monolinguals may be hard, given the extent to which the ability to use another language is now spread around the globe; for research comparing monolinguals and bilinguals we had to call off the search for pure monolingual Japanese (Cook et al., 2006) since every Japanese child is taught English from the age of 12 on and were forced to substitute groups made up of minimal and maximal speakers of English.

The alternative is to abandon the pure monolingual and to start from the view that multilingualism is the norm. Most people are users of second languages to a greater or lesser extent. Pure monolingual sentences and pure monolingual judgements may be virtually impossible to find.

3.4 Consequences for the relevant data

Researchers have constantly striven to use performance data to justify their analyses of linguistic competence, particularly in child language (Cook, 1990), despite the 1960s arguments that you can’t get there from here. We collect large corpora of sentences and texts and base our analyses upon them. As we saw earlier, the learning  problem for the child is how they know which language the sentences are in. The same problem arises for the linguist dealing with corpora of L2 user’s sentence. Once it was easy:

‘A structuralist theory of communication which distinguishes between speech and language ... necessarily assumes that “every speech event belongs to a definite language”’ Weinreich (1953: 7)

But it doesn’t. The L1 sentences a L2 user produces often differ from those of a monolingual native speaker, not just their L2 sentences, let alone sentences with overt code-switching. The Lang2 that the utterance belongs to is in the mind of the beholder. SLA research needs to consider the whole set of sentences, not rejecting some in advance; only later can the sentences be assigned to languages according to other criteria. Bilingual speech therapists have long argued that therapy should be based on the child’s first language as well as their second (Duncan, 1989; Stow & Dodd, 2003).

4. Conclusion

This paper has then tried to approach Universal Grammar theory from the assumption that  bilingualism is the norm, monolingualism the consequence of inadequate input.  Doubtless much of this reasoning may be wrong. Nevertheless it is salutary to look at the language acquisition from this different perspective and to question whether UG theory can achieve its basic task of describing how human minds acquire, store and use language without taking into account the minds that cope with more than one language.


Aronoff, M. (1992) Segmentalism in linguistics: the alphabetic basis of phonological theory. In P. Downing, S.D. Lima & M. Noonan (eds) The Linguistics of Literacy (pp. 71–82). Amsterdam: Benjamins.

Baker, P. & Eversley, J. (2000) Multilingual Capital. London: Battlebridge

Balcom, P. (2003) Cross-linguistic influence of L2 English on middle constructions in L1 French. In V.J. Cook (ed) L2 Effects on the L1 (pp.168-192). Clevedon: Multilingual Matters.

Bernstein, B. (1971) Class, Codes and Control (Volume 1). London: Routledge & Kegan Paul

Bialystok , E. (1991) Metalinguistic dimensions of bilingual language proficiency. In E. Bialystok (ed) Language Processing in Bilingual Children (pp. 113-140). Cambridge: Cambridge University Press

Bongaerts, T., Planken, B. & Schils, E. (1997) Age and ultimate attainment in the pronunciation of a foreign language. Second Language Research, 19, 447-465

British Council (1999) Frequently asked questions. Online document:

Brutt-Griffler, J. (2002) World English: a Study of its Development, Clevedon: Multilingual Matters.

Canagarajah, A.S. (2005) Reconstructing local knowledge, reconfiguring language studies. In A.S. Canagarajah (ed) Reclaiming the Local in Language Policy and Practice (pp. 3-24). New Jersey: Lawrence Erlbaum Associates.

Cenoz, J.B. Hufeisen, & U. Jessner (eds) (2001) Cross-linguistic Influence in Third Language Acquisition: Psycholinguistic Perspectives (pp. 8-20). Clevedon , UK : Multilingual Matters

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

Chomsky, N. (1965) Aspects of the Theory of Syntax. Boston , Mass. : MIT Press

Chomsky, N. (1982) Some Concepts and Consequences of the Theory of Government and Binding, Boston , Mass.: MIT Press

Chomsky, N. (1986) Knowledge of Language: Its Nature, Origin and Use. New York : Praeger.

Chomsky, N. (1991) Some notes on economy of derivation and representation. In R. Freidin (ed) Principles and Parameters in Comparative Grammar (pp. 417-545). Cambridge , Mass.: MIT Press.

Chomsky, N. (2000) The Architecture of Language. New Delhi: Oxford University Press.

Commission of the European Communities (1987) Young Europeans in 1987. EC Commission.

Cook, V.J. (1990) Observational evidence and the UG theory of language acquisition. In I. Roca (ed), Logical Issues in Language Acquisition, Foris. online version

Cook, V.J. (2002) Background to the L2 user. In V.J. Cook (ed) Portraits of the L2 User (editor and contributor) (pp. 1-28). Clevedon: Multilingual Matters.

Cook, V.J. (2003) Introduction: the changing L1 in the L2 user's mind. In V.J. Cook (ed) Effects of the Second Language on the First (pp. 1-18). Clevedon: Multilingual Matters. online version

Cook, V.J. (2007) The nature of the L2 user. In L. Roberts & A. Gurel (eds) EUROSLA Yearbook. Vol. 7 (pp. 000-000). Amsterdam : Benjamins. online version [Reprinted in L. Wei (ed.) (2011), The Routledge Applied Linguistics Reader, Routledge 77-89]

Cook, V.J. (submitted). The second language in second language acquisition.

Cook, V.J., Bassetti, B., Kasai, C., Sasaki, M. & Takahashi, J.A. (2006) Do bilinguals have different concepts? The case of shape and material in Japanese L2 users of English. International Journal of Bilingualism 10 (2), 137-152.

Cook, V.J., Iarossi, E., Stellakis, N. & Tokumaru, Y. (2003) Effects of the second language on the syntactic processing of the first language. In V.J. Cook (ed) Effects of the Second Language on the First (pp. 214-233). Clevedon: Multilingual Matters.

Cook, V.J. & Newson, M. (2007) Chomsky’s Universal Grammar: An Introduction. 3rd edition. Oxford : Blackwell.

Curtiss, S. (1977) Genie: A Psycholinguistic Study of a Modern-Day `Wild child’. New York: Academic Press.

De Angelis, G. (2005) Interlanguage transfer of function words. Language Learning, 55, 379-414

DeKeyser, R. and Larson Hall, J. (2005) What does the critical period really mean?. In J. Kroll & A. De Groot (eds) Handbook of Bilingualism: Psycholinguistic Approaches (pp. 88-108). Oxford: Oxford University Press.

Duncan, D.M. (1989) Working with Bilingual Language Disability. London : Chapman and Hall

European Union (2005) The Luxembourg Presidency – Online document:

Graddol, D. (2006). English Next. London: the British Council.

Hyltenstam, K. & Abrahamsson, N. (2003) Maturational constraints in SLA : a test of the critical period hypothesis. In C. Doughty & M. Long (eds) The Handbook of Second Language Acquisition (pp. 359-588). Rowley MA: Blackwell.

Kasper, G. & Kellerman, E. (eds) (1997) Communication strategies: psycholinguistic and sociolinguistic perspectives. London : Longman

Kato, K. (2004) Second language (L2) segmental speech learning: Perception and production of L2 English by Japanese native speakers. Unpublished Ph.D. dissertation, University of Essex , U.K.

Labov, W. (1969) The logic of non-standard English. Georgetown Monographs on Language and Linguistics, 22, 1-31.

Lenneberg, E. (1967) Biological Foundations of Language. New York: Wiley & Sons.

Leung, Y.-k. I. (2005) L2 vs. L3 initial state: a comparative study of the acquisition  of French DPs by Vietnamese monolinguals and Cantonese-English bilinguals. Bilingualism: Language and Cognition, 8 (1), 39-61

Levinson , S.C. (1996) Relativity in spatial conception and description. In J.J. Gumperz & S.C. Levinson (eds) Rethinking Linguistic Relativity (pp. 177-202). Cambridge : Cambridge University Press.

Milroy, L. & Muyskens, P. (eds) (1995) One Speaker, Two Languages. Cambridge: Cambridge University Press.

Olson, D.R. (1996) Toward a psychology of literacy: on the relations between speech and writing. Cognition, 60, 83–104.

Pavlenko, A. and Jarvis, S. (2007) Cross Linguistic Influence. Mahwah , NJ: Lawrence Erlbaum

Roeper, T. (1999) Universal bilingualism. Bilingualism: Language & Cognition, 2, 169-86.

Said, E. (1999) Out of Place: A Memoir. New York: Vintage Books.

Sapir, E. (1921) Language: An Introduction to the Study of Speech. New York: Harcourt, Brace and Company.

Satterfield, T. (1999) Bilingual Selection of Syntactic Knowledge: Extending the Principles and Parameters Approach. Kluwer

Satterfield, T. (2003) Economy of interpretation: patterns of pronoun selection in transitional bilinguals. In V.J. Cook (ed), L2 Effects on the L1 (pp. 214-233). Clevedon: Multilingual Matters.

Selinker, L. (1972) Interlanguage. International Review of Applied Linguistics, X(3), 209-231.

Smith, N. (1973) The Acquisition of Phonology: a Case Study. Cambridge: Cambridge University Press.

Stow, C. and Dodd, B. (2003) Providing an equitable service to bilingual children in the UK: a review. International Journal of Language and Communication Disorders 38 (4), 351-377.

Trudgill, P. (1978) Accent, Dialect and the School. London: Edward Arnold.

Tsimpli, T., Sorace, A., Heycock, C. and Filiaci, F. (2004) First language attrition and syntactic subjects: a study of Greek and Italian near-native speakers of English. International Journal of Bilingualism 3, 257-278.

Weinreich, U. (1953) Languages in Contact. The Hague: Mouton.

Yelland, G.W., Pollard, J. and Mercuri, A. (1993) The metalinguistic benefits of limited contact with a second language. Applied Psycholinguistics 14, 423-444.