Goals PDF

# Goals Why study language? For Chomsky, the answer is that language is a mirror of the mind - by detailed study of language, we might hope to reach a better understanding of how the human mind produces and processes language. As Chomsky himself remarks: There are a number of questions that might l...

# Goals Why study language? For Chomsky, the answer is that language is a mirror of the mind - by detailed study of language, we might hope to reach a better understanding of how the human mind produces and processes language. As Chomsky himself remarks: There are a number of questions that might lead one to undertake a study of language. Personally, I am primarily intrigued by the possibility of learning something, from the study of language, that will bring to light inherent properties of the human mind. Chomsky seeks to attain two parallel, interrelated goals in the study of language - namely to develop (i) a Theory of Language, and (ii) a Theory of Language Acquisition. The Theory of Language will concern itself with what are the defining characteristics of natural (i.e. human) languages, and the Theory of Language Acquisition with the question of how children acquire their native language(s). Of the two, the task (i) of developing a Theory of Language is - in Chomsky's view - logically prior to the task (ii) of developing a Theory of Language Acquisition, since only if we first know what language is can we develop theories about how it is acquired; moreover, we shall see shortly that a Theory of Language is an important subpart of the Theory of Language Acquisition that Chomsky seeks to develop. So, the primary aim of Linguistics, for Chomsky, is to develop a Theory of Language. But what is it that such a theory seeks to characterise? The answer is that any adequate Theory of Language must provide answers to questions such as the following: * What is language? * What is it that you know when you know a language? What are the essential defining characteristics of natural languages which differentiate them from e.g. artificial languages like those used in Mathematics or Computing, or other forms of communication? Do languages differ from each other in unpredictable ways, or do they all share certain common, universal properties? But how do we attempt to develop a Theory of Language which will answer such questions? The first step is to formulate detailed descriptions (known technically as grammars) of particular languages (e.g. English): this is the study of Particular Grammar. The second step is to abstract from particular grammars common, universal properties that they all share: this is the study of Universal Grammar i.e. the search for linguistic universals. Consider first the study of Particular Grammar. What is a grammar of a particular language? Chomsky gives an essentially mentalist answer to this question: for him, a grammar is a model (= systematic description) of those linguistic abilities of the native speaker of a language which enable him to speak and understand his language fluently. These linguistic abilities, Chomsky terms the competence of the native speaker. Thus, a grammar of a language is a model of the linguistic competence of the fluent native speaker of the language. Competence (the fluent native speaker's knowledge of his language) is contrasted by Chomsky with Performance (what people actually say or understand by what someone else says on a given occasion): competence is 'the speaker-hearer's knowledge of his language', while performance is 'the actual use of language in concrete situations' (Chomsky, Aspects (1965), p. 4). Very often, performance is an imperfect reflection of competence: e.g. the fact that people make occasional 'slips of the tongue' in everyday conversation does not mean that they don't know their language, or don't have fluency (i.e. competence) in it. 'Slips of the tongue' and like phenomena are for Chomsky - performance errors, attributable to a variety of performance factors like tiredness, boredom, drunkenness, drugs, external distractions, and so forth. Linguistics is - for Chomsky - primarily concerned with competence, since a Theory of Competence will be a subpart of an eventual Theory of Performance: that is, you have to understand what a native speaker knows about his language before you can study the effects of tiredness, drunkenness, etc. on this knowledge. Chomsky distinguishes two types of competence: (i) pragmatic competence, and (ii) grammatical competence (see e.g. Chomsky, Essays (1977), p. 40). Pragmatics is concerned with the role played by nonlinguistic information such as background knowledge and personal beliefs in our use and interpretation of sentences. To take one of Chomsky's own examples (from Essays (1977), p. 40), suppose I have a friend who says to me 'Today was a disaster.' If I know (by way of background information) that he was giving a special lecture today, then on the basis of this background knowledge I infer that he probably means that his lecture went down very badly. It is the native speaker's pragmatic competence which enables him to bring into play nonlinguistic information in the interpretation of such sentences. By contrast, in the case of a sentence such as: (1) He thinks that John is wrong it is the native speaker's grammatical competence (his knowledge of the grammar of his language) which tells him that he cannot be interpreted as referring to the same person as John in (1). Chomsky's own work is almost exclusively concerned with the attempt to characterise grammatical competence: by contrast, his work in Pragmatics hitherto has been little more than anecdotal in nature. Grammatical competence in turn subsumes three primary types of linguistic ability - syntactic, semantic and phonological. The principal syntactic ability which forms part of the native speaker's grammatical competence is the ability to combine words together to form grammatical sentences in his native language, and to know which sequences of words form grammatical or ungrammatical sentences in his language. For example, any native speaker of English would intuitively recognise (leaving aside for the moment differences of style or dialect) that all of the examples in (2) below are grammatical (i.e. syntactically well-formed) sentences in English: (2) (a) I gave back the car to him (b) I gave the car back to him (c) I gave him back the car (d) I gave him the car back but that the following are ungrammatical as sentences of English: (2) (e) *I gave the car to him back (f) *I gave back him the car (an asterisk in front of a sentence means that it is ungrammatical i.e. syntactically ill-formed in some way; by convention, any sentence which does not have an asterisk in front of it is assumed to be grammatical). However, there is - for Chomsky a second aspect of the native speaker's syntactic competence which a grammar should characterise - namely the native speaker's intuitions about the syntactic structure of sentences in his language. For example, in a sentence like: (3) John likes very fast cars any native speaker would agree that very modifies fast (and not likes), and that very fast modifies cars and not John and so forth. Thus, the native speaker has two types of syntactic intuition: intuitions about well-formedness, and intuitions about structure. We should perhaps add that the word intuition here is used in a technical sense which has become standardised in Linguistics: by saying that a speaker has intuitions about sentence well-formedness, all we mean is that he has the ability to make judgements about whether or not a given sequence of words is grammatical in his native language. Among the semantic abilities which form part of the native speaker's grammatical competence are his intuitions about the semantic well-formedness or ill-formedness of sentences: thus, any native speaker of English would agree that (4) (a) below is semantically well-formed, but that (4) (b) is semantically ill-formed (i.e. 'odd' in some way by virtue of its meaning): (4) (a) I thought Mary was ill, but it turned out that she wasn't (b) ! I knew Mary was ill, but it turned out that she wasn't (! in front of a sentence means that it is semantically ill-formed). A second type of semantic intuition which native speakers have about their language concerns semantic structure, and semantic relations: for example, any native speaker of English will tell you that the fool can be interpreted as coreferential to (i.e. referring to the same individual as) Harry in (5) (a) below, but not in (5) (b): (5) (a) I don't like Harry, because the fool hates Linguistics (b) Harry says that the fool hates Linguistics Hence, intuitions about coreference relations in sentences are part of the set of intuitions we have about semantic relations in and between sentences. Among the phonological abilities subsumed under grammatical competence are the native speaker's intuitions about the phonological well-formedness or ill-formedness of sentences in his language. All speakers of English would agree, for example, that (6) (a) below is phonologically well-formed, but that (6) (b) is phonologically ill-formed (the syllables receiving primary stress are italicised): (6) (a) This is a grammatical sentence (b) This is a grammatical sentence A second type of phonological intuition which native speakers have in their language is intuitions about phonological structure: any English speaker intuitively feels, for example, that the sequence black bird can either be a single phonological word (blackbird, with primary stress on black = a species of bird, like thrush, robin, etc.), or two independent phonological words, each with its own primary stress (black bird = bird which is black, as opposed to white bird). Overall, then, we see that grammatical competence subsumes two types of intuition: (i) intuitions about syntactic/semantic/phonological well- or ill-formedness; and (ii) intuitions about syntactic/semantic/phonological structure. Before we go any further, however, it is useful to clear up a number of problems which arise with the notion of ill-formedness. One of these is that it is important not to confuse the descriptive notion of well-formedness with the corresponding prescriptive notion of correctness. For example, there are many dialects of English in which sentences like: (7) Mine is bigger than what yours is are perfectly grammatical, and for speakers of these dialects such sentences are perfectly well-formed. But at the same time, sentences like (7) are of a type stigmatised as 'incorrect', or 'bad grammar' by a certain self-styled socio-cultural elite (e.g. school-masters). This poses an apparent dilemma for the linguist: should their language concerns semantic structure, and semantic relations: for example, any native speaker of English will tell you that the fool can be interpreted as coreferential to (i.e. referring to the same individual as) Harry in (5) (a) below, but not in (5) (b): (5) (a) I don't like Harry, because the fool hates Linguistics (b) Harry says that the fool hates Linguistics Hence, intuitions about coreference relations in sentences are part of the set of intuitions we have about semantic relations in and between sentences. Among the phonological abilities subsumed under grammatical competence are the native speaker's intuitions about the phonological well-formedness or ill-formedness of sentences in his language. All speakers of English would agree, for example, that (6) (a) below is phonologically well-formed, but that (6) (b) is phonologically ill-formed (the syllables receiving primary stress are italicised): (6) (a) This is a grammatical sentence (b) This is a grammatical sentence A second type of phonological intuition which native speakers have in their language is intuitions about phonological structure: any English speaker intuitively feels, for example, that the sequence black bird can either be a single phonological word (blackbird, with primary stress on black = a species of bird, like thrush, robin, etc.), or two independent phonological words, each with its own primary stress (black bird = bird which is black, as opposed to white bird). Overall, then, we see that grammatical competence subsumes two types of intuition: (i) intuitions about syntactic/semantic/phonological well- or ill-formedness; and (ii) intuitions about syntactic/semantic/phonological structure. Before we go any further, however, it is useful to clear up a number of problems which arise with the notion of ill-formedness. One of these is that it is important not to confuse the descriptive notion of well-formedness with the corresponding prescriptive notion of correctness. For example, there are many dialects of English in which sentences like: (7) Mine is bigger than what yours is are perfectly grammatical, and for speakers of these dialects such sentences are perfectly well-formed. But at the same time, sentences like (7) are of a type stigmatised as 'incorrect', or 'bad grammar' by a certain self-styled socio-cultural elite (e.g. school-masters). This poses an apparent dilemma for the linguist: should he describe what people actually say, or should he attempt to prescribe what he or others think they ought to say? In other words, should Linguistics be descriptive, or prescriptive? In actual fact, it is hard to see how anyone could defend the prescriptive approach: in any other field of enquiry, it would be seen as patently absurd. What would we say of the sociologist who, instead of describing the way a given society is, sets about prescribing how he thinks it ought to be? And what would we think of the scientist who, regretting the unfortunate tendency for objects to fall downwards by gravity, instead proposes an alternative model in which everything is attracted upwards towards the sky, simply because he thinks things ought to be that way? No one these days would take any such enterprise seriously; and the same is true of Linguistics. Modern Linguistics is purely descriptive, not prescriptive. A more serious problem that arises with Chomsky's conception of a Grammar as a model of the linguistic intuitions of the average native speaker of a given language concerns what to do about disagreements among native speakers about the well-formedness or structure of particular sentences. One of the abstractions that Chomsky makes in studying language is to assume that speech communities are homogeneous: i.e. to assume that all native speakers of a given language will have essentially the same well-formedness intuitions: as Chomsky himself says (Aspects (1965), p. 3): 'Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech community...' But this is plainly not the case: all native speakers have to some extent their own individual way of speaking (or idiolect) which may not be exactly the same as that of any other member of the same speech community. There are, of course, larger linguistic groupings within society: speakers with a common geographical background may share a common dialect, while speakers from a common social background may share a common sociolect. We can illustrate the linguistic differences within a given speech community in terms of the examples in (8) below: each of these sentences would probably be accepted as well-formed by only a certain percentage of English speakers (hence the use of the % prefix): (8) (a) Your car wants mending (b) %That's to do tomorrow (c) %I gave it her (d) %There's a man delivers vegetables in the village (e) %It was me what told her The obvious question to ask is what the linguist is to do in such cases. The broad answer is that in general the problem of linguistic variation within a speech community is one which is more appropriately dealt with in a partially separate discipline (Sociolinguistics), and since it is not a problem which is essentially syntactic in nature, it is not the kind of problem which ought to be the primary focus of attention in the attempt to develop an adequate theory of Syntax. For practical purposes, most linguists describing a language of which they are native speakers rely on their own intuitions, and thus the grammar they devise is essentially a grammar of their own idiolect, which they assume is representative of the language as a whole. An even more tricky problem which arises with the notion of ill-formedness concerns the attempt to identify in what way a given sentence which 'sounds odd' is ill-formed. Let's first draw a distinction between sentences which are 'pragmatically odd' in some way, and those which are 'linguistically ill-formed'. While the distinction may be clear enough in principle, it is often very hard in practice to decide which side of the dividing line a given sentence falls. For example, what is the status of sentences such as the following (taken from George Lakoff, 'Presupposition and relative well-formedness' (1971), p. 332): (9) (a) My uncle realises that I'm a lousy cook (b) My cat realises that I'm a lousy cook (c) My goldfish realises that I'm a lousy cook (d) My pet amoeba realises that I'm a lousy cook (e) My frying pan realises that I'm a lousy cook (f) My sincerity realises that I'm a lousy cook (g) My birth realises that I'm a lousy cook Intuitively, most people would regard (9) (a) as perfectly well-formed, (9) (b) as slightly less natural, (9) (c) as a little eccentric, (9) (d) as implausible, (9) (e) as just plain daft, and (9) (f) and (g) as absolutely inconceivable. But what precisely is the nature of the oddity in the more unusual sentences? The answer is that the oddity seems to be largely pragmatic (i.e. nonlinguistic) in nature. Thus, whether or not you find expressions like My goldfish thinks that... well-formed depends on whether or not you believe that goldfish do (or might) possess powers of thought; a sentence like (9) (c) presupposes that goldfish are capable of thought, and a person who rejects sentences like (9) (c) is in effect rejecting the proposition that goldfish are capable of thought. Why should he reject such a proposition? Presumably because it conflicts with his personal beliefs about the world. Thus, sentences like (9) (c) are not linguistically ill-formed in any way, but rather simply pragmatically odd, in the sense that they express ideas which do not conform to our view of the way the world is organised. But surely, (9) (f) and (g) are a very different case: surely nobody in any culture could accept such a sentence as well-formed? In this regard, Lakoff (ibid.) remarks: (9) (f) and (9) (g) are another matter. That properties and events have mental powers might seem to be an impossible belief, not just a strange one. If this were true, it would follow that (9) (f) and (9) (g) were universally impossible. However, Kenneth Hale informs me that, among the Papagos, events are assumed to have minds (whatever that might mean), and that sentences like (9) (g) would be perfectly normal. In other words, our judgements about the ill-formedness of sentences like (9) depend entirely on our cultural, religious, or personal beliefs not on any linguistic knowledge that we have about our language. Sentences like (9) are thus linguistically well-formed: they are neither ungrammatical (e.g. they don't involve a plural verb with a singular subject), nor semantically ill-formed (e.g. it is not the case that we find such sentences as completely incomprehensible or meaningless as a sentence from a language we do not speak). The distinction between sentences which are pragmatically or linguistically 'odd' in some way has been the cause of a great deal of confusion in the linguistic literature. For example, Katz and Postal (An Integrated Theory of Linguistic Descriptions (1964), p. 16) observe that a phrase like honest geranium is literally 'meaningless' (sic): this presupposes that the oddity of such a phrase is purely linguistic (and more particularly semantic) in nature. But this seems a remarkably obtuse statement, since most of us would have no difficulty in assigning a meaning to the phrase honest geranium: we might paraphrase it (in typical lexicographical jargon) as 'a flower of the genus pelargonium which is upright in word and deed'. Of course, those narrow-minded bigots (or linguists) who can't imagine geraniums having essentially human properties like 'not telling fibs' would doubtless feign to be mystified by such a phrase. They ought to stare in blank incomprehension at the following extract from a fictitious episode of The Magic Round-about: Dougal was looking for his dog-biscuits. 'Have you been nibbling my Kanine Krunchy-Munchies again?' he asked Gerald the geranium. 'I'm awfully sorry, old chap,' replied Gerald, 'But I was so hungry...' 'Oh well, never mind!' retorted Dougal, 'At least you're an honest geranium.' In such a context (i.e. a 'fairy story' context in which one's beliefs about the real world are suspended), such a phrase is clearly acceptable to anyone with a little imagination. But that is precisely the point: whether or not you find a phrase like honest geranium well-formed depends on your powers of imagination, not on any linguistic fact. Let's take another problematic case. What is the status of a sentence like: (10) Gary Gay never loses her temper with anyone Are we to say that such sentences are linguistically ill-formed - i.e. ungrammatical by virtue of the fact that the feminine pronoun her does not agree in gender with the masculine subject Gary Gay? Or are we to say that such sentences are semantically ill-formed – i.e. uninterpretable, and hence incomprehensible? Again, surely not. (10) is perfectly meaningful: it carries the implication that Gary Gay is effeminate. Of course, a person who finds any such insinuation distasteful may well object to such a sentence - but for nonlinguistic reasons. Similar problems arise with phrases like: (II) (a) the tree who we saw (b) the man which we saw An obvious initial reaction to such phrases would be to regard them as syntactically ill-formed: but this is surely a very myopic view. (11) (a) is neither syntactically nor semantically ill-formed: the use of who referring back to tree carries with it the implication that the tree concerned is thought of as having e.g. humanlike qualities (i.e. in traditional terminology, it is personified). As such, (11) (a) would be fully acceptable in a 'fairy story' context (where beliefs about the real world are suspended) - as in the following fictitious extract from The Magic Forest: The naughty tree who we passed waved his cheery branches at Mary and said 'You can tickle my leaves any time, darling.' Support for this way of looking at phrases like (11) (a) (i.e. as pragmatically, rather than linguistically 'odd'), comes from the title of Pamela McCorduck's (1979) book Machines who Think (W.H. Freeman and Co., San Francisco). Likewise, (11) (b) is again both grammatical and meaningful: but the phrase carries the clear implication that the man referred to is somehow thought of as being an inferior being of some sort. Hence, (11) (b) might be appropriate if uttered by an arrogant Martian or Dalek, or even if uttered by a human in relation to a plastic model of a man, or a robot-man, etc. Thus, it is easy to be deceived into thinking that a sentence is linguistically ill-formed in some way, when in fact very often the sentence is merely 'pragmatically odd' i.e. it expresses an idea which (to people with certain beliefs, culture, education, etc.) may seem unacceptable. But if we now turn instead to sentences which do appear to involve some kind of linguistic ill-formedness, we again find many uncertainties: in particular, we find that it is sometimes very difficult to distinguish between sentences which are semantically ill-formed (or unsemantic, to use more recent terminology introduced by Chomsky), and those which are syntactically ill-formed (or unsyntactic). There are, of course, relatively clear cases of semantic ill-formedness, like: (12) (a) !I killed John, but he didn't die (b) !All my friends are linguists, but I have no friends and relatively clear cases of syntactic ill-formedness, like: (13) (a) *John very Mary much loves (b) *All are linguists friends my But there are also awkward intermediate cases, where it is clear that a sentence is linguistically ill-formed, but less clear whether it is syntactically or semantically ill-formed: consider, for example, the sentence: (14) We respect himself A traditional way of looking at such sentences is to say that they are ungrammatical (i.e. syntactically ill-formed), by virtue of the fact that himself is a third person masculine singular reflexive pronoun, and does not agree in number or person with the first person plural nonreflexive pronoun subject we. But there is another, less traditional but equally plausible way of looking at such sentences: we might argue that they are syntactically perfectly well-formed: after all, a reflexive pronoun like himself can occur as the direct object of a verb like respect in a sentence like: (15) John respects himself and hence there is clearly no overall syntactic restriction in English against using himself as the object of a transitive verb, and hence no reason to suppose that (14) is any more ungrammatical than (15). Instead, we might argue that (14) is in fact semantically ill-formed - i.e. simply uninterpretable. What's wrong with (14) might say is that reflexive pronouns like himself have no independent reference, but rather must take their reference from some antecedent in the same sentence which is compatible in number, gender, and person with the reflexive. But in (14) - unlike (15) there is no compatible antecedent for himself; hence, himself cannot be assigned any reference, and the resulting sentence as a whole is uninterpretable (or unsemantic, to use Chomsky's term) i.e. 'meaningless'. Faced with two equally plausible ways of looking at the ill-formedness of a sentence like (14), the obvious question to ask is how we decide whether to analyse (14) as syntactically, or semantically ill-formed. It might seem that the natural answer to this question is to rely on the intuitions of the native speaker. 'Do native speakers feel that such sentences are syntactically, or semantically ill-formed?' we might ask. However, as Chomsky has repeatedly emphasised, this is an unrealistic and unreasonable question to ask of an informant (= a native speaker who you are questioning about his language): for the simple fact is that very often all an informant can tell you is that one sentence 'sounds OK', and another 'sounds odd' - without being able to say how or why it sounds odd. As Chomsky himself remarks: We may make an intuitive judgement that some linguistic expression is odd or deviant. But we cannot in general know, pretheoretically, whether this deviance is a matter of syntax, semantics, pragmatics, belief, memory limitations, style, etc., or even whether these are appropriate categories for the interpretation of the judgement in question. It is an obvious and uncontroversial fact that informant judgements and other data do not fall neatly into clear categories: syntactic, semantic, etc. Indeed, it is not surprising that informants should not be able to tell you whether a sentence is syntactically or semantically ill-formed; for the very notions 'syntactically ill-formed' and 'semantically ill-formed' are terms borrowed from linguistic theory: and like all theoretical terms, they are meaningless to those not familiar with the theory. An informant simply gives judgements about ill-formedness: it is up to the linguist to decide, on the basis of the internal organisation of his own theory, what category a particular type of ill-formedness represents within his theory. So now let's ask a rather different question about sentences like (14): 'Within Chomsky's theory, would (14) be analysed as syntactically, or semantically ill-formed?' Before we can answer this question, however, we should first say something about Chomsky's view of the overall organisation of a grammar. Chomsky takes an essentially modular approach to Linguistics: as we have already seen, he believes that Pragmatics should be separated from (and studied independently of) Grammar; within the domain of Grammar, Chomsky also believes that Syntax, Semantics, and Phonology should all three be treated as autonomous of each other, and studied independently. What is of more direct concern to us here is his insistence that Syntax should be studied independently of Semantics: this is known as the Autonomous Syntax thesis, and is defended at length in his Essays (1977), pp. 36-59 (not for beginners). Chomsky repeatedly emphasises that his claim that a distinction can and must be drawn between syntactic and semantic phenomena is an empirical one i.e. a working hypothesis to be judged solely on the basis of whether it leads to interesting generalisations about Syntax on the one hand, and Semantics on the other. As he himself remarks: . The thesis constitutes an empirical hypothesis about the organisation of language, leaving ample scope for systematic form-meaning connections while excluding many imaginable possibilities. In principle, the dividing line between Syntax and Semantics seems clear enough: Syntax studies how words can be combined together to form sentences, what positions in the sentence a given word or phrase can occupy, and so forth; Semantics studies the meaning of words, phrases and sentences. Syntax answers the question 'Is such-and-such a sentence grammatical, and, if so, what is its syntactic structure?'; whereas Semantics answers the independent question 'Is such-and-such a sentence meaningful, and, if so, what does it mean?' But in practice, the issues may not be as clearcut: returning to our earlier problem, are we to treat (14) as syntactically or semantically ill-formed? In pre-1970 work in TG, sentences like (14) would undoubtedly have been considered ungrammatical; but in post-1970 work, they would be considered grammatical, but uninterpretable (i.e. unsemantic). The difference in attitude to such sentences reflects a difference in the overall organisation of grammars between Chomsky's earlier, and later work. Many phenomena dealt with as syntactic in earlier work have been reanalysed as semantic in more recent work. Subsequent discussion will give a fairly clear idea of where the dividing line between Syntax and Semantics is currently drawn, so we shall not dwell on the issue here. Hitherto, we have been concerned with a variety of problems which arise with the notion of sentence ill-formedness (ill-formed relative to whose idiolect, and in what way(s)?). But there is one further distinction which we should draw - namely between sentences which are ill-formed (e.g. unsemantic, unsyntactic (= ungrammatical), etc.), and those which are unacceptable. This distinction which Chomsky draws is related to the distinctions which he draws between Competence and Performance on the one hand, and Pragmatics and Grammar on the other. Native speakers give judgements about the acceptability of sentences (i.e. they indicate whether they find them acceptable or unacceptable); acceptability is a Performance notion. If a native speaker says that a sentence is unacceptable in his language, this may be for a variety of reasons: one possibility (discussed earlier in relation to (9)) is that the sentence is simply pragmatically odd in some way (i.e. it expresses an idea which the speaker finds distasteful, or otherwise unacceptable); as we have already seen, sentences which are pragmatically odd are not necessarily linguistically ill-formed. However, it may be that a speaker will find unacceptable a sentence which is not pragmatically odd in any way: does this mean that the sentence concerned is ill-formed? Not necessarily. It may be that the informant is confused, or tired, or bored with all those questions about his language, and hence makes a hasty, and perhaps erroneous judgement: in such a case, Performance is a poor reflection of Competence. Or it may be that the informant is influenced by prescriptive notions inculcated at school: thus, some English speakers asked about a sentence like: (16) Who did you meet at the party? would reply that such a sentence is unacceptable because it is 'bad grammar', and should be 'corrected' to: (17) Whom did you meet at the party? But of course this is nonsense: a sentence like (16) is perfectly well-formed, and is characteristic of everyday conversation. In this case, a performance factor (prescriptive education) is interfering with the natural competence of the native speaker, with the result that the acceptability judgements which he gives are not an accurate reflection of the well-formedness or otherwise of sentences in his language. To summarise: native speakers of a language have the ability to make performance judgements about sentence acceptability, on the basis of which the linguist seeks to capture the intuitions about sentence well-formedness which form the basis of their competence. But the remarkable fact about these abilities - Chomsky argues - is that they hold not only for familiar sentences that we have heard before, but equally for sentences that we have never heard before or, as Chomsky calls them in his earlier work, novel utterances: herein lies what Chomsky refers to as the essential creativity of language: cf. The most striking aspect of linguistic competence is what we may call the 'creativity of language', that is, the speaker's ability to produce new sentences, sentences that are immediately understood by other speakers although they bear no physical resemblance to sentences which are familiar. To cite one of Chomsky's own examples (from Logical Structure (1975), p. 132), you have probably never encountered any of the following sentences before: (18) (a) Look at the cross-eyed elephant (b) Look at the cross-eyed kindness (c) Look at the cross-eyed from And yet if you are a speaker of English - you intuitively know that (18) (a) is acceptable in English, whereas (18) (b) is odd, and (18) (c) almost inconceivable. Any native speaker is capable of producing, and understanding 'new sentences' like (18) (a), or making judgements about their well-formedness. What is the significance of the fact that all native speakers have the ability to form, interpret and pronounce sentences that they have not come across before? Chomsky argues that this essential creativity of language shows that language can't simply be learned by imitation: i.e. learning a language doesn't simply involve learning a list of sentences produced by others, and repeating them parrot-fashion. On the contrary, argues Chomsky, very few of the sentences we produce or hear are repetitions of previous utterances: The normal use of language is innovative in the sense that much of what we say in the course of normal language use is entirely new, not a repetition of anything that we have heard before, and not even similar in pattern - in any useful sense of the terms 'similar' and 'pattern' - to sentences or discourse that we have heard in the past. The novelty of most utterances we produce or hear provides a strong argument against the claim made by behavioural psychologists that language-learning simply involves the acquisition of a set of 'linguistic habits'. As Chomsky remarks: This creative use of language is quite incompatible with the idea that language is a habit-structure. Whatever a habit-structure is, it's clear that you can't innovate by habit, and the characteristic use of language, both by a speaker and a hearer, is innovative. You're constantly producing new sentences in your lifetime - that's the normal use of language. So, if language isn't learned by imitation, how is it acquired? Chomsky argues that in order to account for the native speaker's ability to produce and understand new sentences, we must postulate that the child learning a language and faced with a certain set of data (i.e. the speech of people around him) abstracts from the data a set of general principles about how sentences are formed, interpreted and pronounced. These principles (or rules) must be of a sufficiently general nature to allow the child to form, interpret and pronounce new sentences that he hasn't come across before. In other words, acquisition of a language involves acquisition of (at least) the following: (i) a set of rules of sentence-formation (which specify how to combine words together to form grammatical sentences) (ii) a set of rules of sentence-interpretation (which specify how to interpret what sentences mean) (iii) a set of rules of sentence-pronunciation This is what Chomsky means by saying that Language is rule-governed. The task of the linguist seeking to account for this creative aspect of grammatical competence is to formulate a set of rules of sentence-formation (= syntactic rules), rules of sentence-interpretation (= semantic rules), and rules of sentence-pronunciation (= phonological rules). A grammar of a language will thus comprise three interrelated components: a syntactic component, a semantic component, and a phonological component. In Chomsky's view, then, learning a language involves learning a set of syntactic, semantic and phonological rules. But is there any evidence that learning a language actually does involve learning a set of rules? Some evidence in support of this assumption comes from studies of child language acquisition. Jean Berko in her (1958) paper 'The child's learning of English morphology' describes a simple experiment designed to prove the point: she drew a picture of an imaginary animal on a piece of paper, and told a child that it was called a wug; then she drew a picture of two of the same animals on another piece of paper, and asked the child what they were. The child replied: