Reflections on Language
On Noam Chomsky: Critical Essays
Since the publication of Syntactic Structures nineteen years ago the general shape of Chomsky’s position in linguistic theory has become familiar. The subject, as he conceives it, is a branch of cognitive psychology; its basic problem is posed by the human capacity to acquire a natural language, something which Chomsky has insisted we should see as remarkable, with regard both to what the child experiences and to what he acquires. What he acquires is an indefinitely extensive creative capacity to produce and to understand an openended set of sentences that he has never heard before. What he is offered by his elders (or rather from them, since Chomsky thinks little importance can be attached to directed language teaching) is evidence, as he has put it, “not only meager in scope, but degenerate in quality.” The actual performances the child is exposed to are fragmented and distorted relative to his recognition, apparent in the competence he acquires, of what would be an acceptable sentence of his language.
To explain the gross disproportion between what is acquired (in the form of competence) and what is experienced (in the form of speech) we need to posit a strongly constrained, internal, innate mechanism which, when triggered by the experience of speech, builds a cognitive structure, a grammar of the language, within limits set by very specialized schemata. Any human child, moreover, can learn naturally any human language, so the schemata must be universal, and when Chomsky refers to the properties of the innate mechanism, he often indicates that each of us possesses, indeed knows, the principles of a universal grammar. His model, though cognitive, is also biological, and in the present book, which consists of three lectures given in 1975, together with a long paper which is a revision of one submitted for a Festschrift, he particularly favors an embryological analogy, in which development of language is compared to the genetically controlled development of an animal.
As Chomsky has tirelessly pointed out to his critics, the mere idea of an innate component in learning a language is undisputed and uninteresting: the blankest theory of behaviorism requires some innate mechanism, however minimal. The important question concerns how complex and how specific to language acquisition the mechanism is supposed to be. In particular, Chomsky has differed from the empiricist tradition in regarding the mechanism as not simply one that applies a general learning strategy to language. Discussion over the past years, however, has made it clear that this particular difference between Chomsky and the empiricists is ambiguous, and some recognition of the ambiguity can be traced in the present book.
A “general learning capacity” might be defined in terms of some very simple learning theory, such as the traditional empiricist theories of “association” or of “inductive” generalization. In this sense, Chomsky convincingly insists that no one has offered a plausible or even coherent way of representing the learning of language by such empiricist learning theories. But it might also be true that very little that is learned can be represented in these simple empiricist terms: maybe most learning requires innate mechanisms more complex, and with more defined limits, than empiricism traditionally has allowed. If this is so, then the important question for language concerns how specific the capacity for language acquisition is, not the extent to which it is, peculiarly, innate.
The general issue is to some degree, but only to some, independent of what exactly the principles of the grammar of a human language have to be. Chomsky has retained, of course, his original picture of grammatical rules as being “generative.” They include transformational rules which turn abstract “deep structures” into “surface structures.” These surface structures take on a particular phonological form and emerge as the sentences one actually hears.
But many other more particular aspects of the theory have changed over the years. Above all on the matter of the relations between syntax (how and why a sentence is well formed or, in most everyday senses, “grammatical”) and semantics (what a sentence means, what it refers to, what has to obtain for it to be true), Chomsky has abandoned the “standard” theory of Aspects of the Theory of Syntax (1965), by which semantic interpretation was applied to deep structure. He now applies semantic interpretation to surface structure, a notable modification of his earlier views. On a verbal matter, Chomsky in these pages proposes giving up the wellknown phrase “deep structure” for those initial abstract base sentences to which the transformations are applied. His grounds for doing so are revealing of what he cares most about. He argues that not only these base sentences, but processes applied at the level of surface structure as well, are “deep” in the only interesting sense—namely, expressive of important and hidden human powers.
On the questions of the relations of syntax and semantics, many other positions are possible and have been vigorously discussed. Some expressions of these disagreements about the place of semantics in transformational grammar are to be found in some of the papers collected in the book edited by Gilbert Harman; other papers take the discussion further, into questions about how semantics in general is to be understood and pursued—by using a theory of truth, for instance, as proposed by Donald Davidson, or by certain abstract structures of “possible worlds,” advocated here by David Lewis. It is exceptionally difficult for someone like myself, who is not engaged full-time in the technical literature of these subjects (difficult, also, I suspect, for those who are), to have any full sense of how these various approaches relate to one another and to generative grammar, or to understand how far these and other semantic theories exclude one another, or are rather dealing in complementary questions.
Harman’s collection, though it contains much good material, is not going to help anyone with this problem. His introduction is useful as far as it goes, but it does not go nearly far enough; and while the book is called On Noam Chomsky, in the case of one or two papers it would take someone with a sophisticated understanding of the subject to grasp why the matters discussed bear on issues raised by Chomsky at all.
Even when proposals about the nature of semantics are explicitly related to Chomsky, the size, weight, and exact location of disagreement can remain obscure. Harman valuably reprints John Searle’s admirable piece “Chomsky’s Revolution in Linguistics,” which appeared in this journal in 1972. In it Searle criticizes Chomsky for not associating semantic study with the notion of communicative intention; he urges his own and Grice’s theories of speech-acts, which link the meaning of the sentence to the intentions of the speaker, to fill what he sees as a void in the Chomskyan system—the point of language. Chomsky, in the present book, replies to Searle, but chiefly to deny, once more, the necessity of communicative intent, and to insist that uses of language can be thoroughly and seriously meant though not intended to influence any hearer.
As a graduate student, I spent two years writing a lengthy manuscript, assuming throughout that it would never be published or read by anyone. I meant everything I wrote, intending nothing as to what anyone would believe about my beliefs, in fact taking it for granted that there would be no audience.
But how could the existence of such cases possibly be the main issue? One would like to know whether, if what Searle says about the connections of meaning and intention were true, much or any of Chomsky’s views would be upset. It might be true that, despite exceptions, language is primarily or centrally connected with intentions to influence hearers’ attitudes; and yet some Chomskyan account of the capacities to produce and understand those utterances could also be true.
All through this field, as in other places where exciting work is going on, views which may well be compatible nevertheless struggle with one another. This is because they struggle for attention; the research programs may be compatible in content, but the thought that goes with each of them, “this is the way to go on,” excludes the others. This thought may be essential to the researcher—the synoptic peacemaker who pleads for compatibility is up in the observer’s balloon, not engaged at the scientific front. In Chomsky’s own case, however, a further dimension is involved, and a more important one. He associates his own approach with an affirmation of the depth of the human mind and the value of the individual, and he is suspicious on more than theoretical grounds of many styles of opposing theory.
Chomsky’s arguments with his opponents are painstakingly reasoned and academic in tone (though he is unduly given to that polemicists’ put-down, “unfortunately,” as in, “Unfortunately, X is rather careless in his references,” page 218). It is only after technical argument—but still too soon, granted some complexities we shall come to—that he falls back on ideological explanation, associating empiricist opposition to his views with social reaction. But through the laborious and sometimes peripheral self-defense, one can see that he is deeply distrustful and disapproving of some other ways of doing linguistics and the other human sciences, and that his linguistic theories, in their central contentions, have an ideological significance which relates them in some unclear but powerful way to his political and social outlook. This comes to the surface in a few pages of Chomsky’s present book, and it gets very brief consideration in the last item in Harman’s collection, Dell Hymes’s informative review of John Lyons’s book on Chomsky in the Modern Masters series (Viking, 1970). It is worth exploring further.
What exactly is involved in the innate component, what principles the language-acquisition device is armed with, are of course technical matters. But whatever they may exactly turn out to be, the more general question comes up of how to describe their presence. Chomsky has favored the terminology of unconscious, innate, knowledge, and this has helped to connect his theories with those Rationalist thinkers of the seventeenth century and later whom he has claimed, always with some caution, as his intellectual ancestors. Whether “knowledge” is the right concept, however, is a hard question. Certainly many objections which have been put to Chomsky, such as that, on his principles, a falling stone must know how to fall, entirely miss the point, and ignore the cognitive character of the states governed by Chomsky’s innate schemata, as Thomas Nagel well argues in Harman’s book. Yet, as Nagel also points out, there remains a long step to accepting the concept of knowledge as applying to the presence of the schemata themselves, and there are theoretical embarrassments in the use of that concept.
There is, for example, the question (raised elsewhere by Harman) of the vehicle by which this knowledge is represented in the mind, a vehicle looking suspiciously like another already mastered language. There is also the related problem that knowledge which is more than merely skill should imply the possession of concepts, and we have no reason to ascribe to the language-learner, at any level, the theoretical concepts of universal grammar. Chomsky’s own embryological analogy hardly points unwaveringly in the direction of a model that uses the notion of knowledge. In the present book he suggests that whether we call these potentialities “knowledge” is a verbal question, and he is prepared to let the word go; but how slight a concession this is becomes clear when he agrees to put, in place of “know,” the word “cognize.”
Chomsky’s insistence on a cognitive vocabulary to describe the presence of the innate schemata seems to be sustained by one of his strongest convictions, the power of linguistic theory to reveal the depth of the human mind. It implies an uncovering of lower, but continuously related, levels of human thought, and historically it helps to tie his theory to earlier Rationalist speculations. These, however, usually took a less naturalistic view of the mind than Chomsky does—the idea of a cognitive study as a branch of human biology is for Descartes’s own system unintelligible. But their views can, like his, be handily opposed to an empiricist outlook, which takes a shallower and more mechanical view of the psychological. That empiricist outlook can, moreover, Chomsky believes, be easily associated with a denial that there is a human nature, and with a manipulative and authoritarian conception of what can be done to human beings.
This opposition between rationalist and empiricist approaches (though, as Hymes says, Chomsky’s own theoretical work has transcended it) has great ideological significance for Chomsky. The empiricist conception of human beings as unpredisposed objects for conditioning he associates with potentialities for technological oppression. Chomsky admits that empiricist systems of ideas do, as a matter of historical fact, strongly resist being categorized in the way he requires, since the idea that there is no fixed human nature, but that man is a social product, has very often been associated (for instance, by many Marxists) with “progressive and even revolutionary social thinking,” while the opposite view has supported conservative and pessimistic outlooks.
“But a deeper look,” he goes on (page 132), “will show that the concept of the ‘empty organism,’ plastic and unstructured, apart from being false, also serves naturally as the support for the most reactionary social doctrines.” “Serves naturally” here is pure ideologists’ sticky tape, no better than “goes with.” Similarly unreliable connections are made when Chomsky goes into historical interpretation:
Empiricism rose to ascendancy in association with a doctrine of “possesive [sic] individualism” that was integral to early capitalism, in an age of empire, with the concomitant growth (one might almost say “creation”) of racist ideology. [Page 130]
In so far as this offers anything except evasive insinuation, it invites a quick answer: in any sense in which classical empiricism was “in association with” early capitalism, slavery, etc., so was its near contemporary, classical rationalism. No historical speculations of this sort can effect anything, and Chomsky’s other admissions show that he is, or at least has good reason to be, uneasy with them. He is likely to find better ground in certain features of the present. Now, when psychological technology is a conscious weapon of political power, and the notion of human needs as opposed to contingent preferences has been moved by many forces into the center of serious social thought, there really is a case for saying that the spirit of empiricist and, above all, behaviorist outlooks must, apart from their intellectual inadequacies, turn us in the wrong direction. Here Chomsky’s negative view, at least, seems to have real power.
But even so, it is more doubtful whether Chomsky’s own innatist doctrines can turn us in the right direction: they might even help to do the opposite. For here the question of the specificity of the language capacity, mentioned earlier, takes on a considerable and unexpected ideological significance. Chomsky’s claims have always been for the special character of man’s capacity for linguistic learning, and part of his evidence for this has precisely been that, the grossly defective apart, men are equal in this capacity, while differing in general intelligence and in their capacities to learn other things, such as physics. But now why should these distributions of innate capacities have any tendency at all to encourage belief in the foundations of libertarian socialism? Chomsky speaks of his hopes for progress toward human self-determination and genuine freedom; but the basic linguistic competence, as he describes it, has no connection with notions of progress at all—it is perfect as it is. What does leave room for progress, and indeed progress toward self-determination and freedom, is man’s lexical sophistication and conceptual grasp—but that, precisely, is a linguistic dimension in which men do differ, and in which the results of learning are not the same for all, and the innatist element correspondingly weaker.
Again, why should Chomsky’s theory have any power against racism, association with which was one of his more sinister charges against empiricism? In so far as racism has any coherent relation at all to opinions about different intellectual capacities, why should the fact of an equal innate capacity for language acquisition be thought to help against it? No theorist of apartheid is likely to be daunted by being reminded that the African child can effortlessly acquire Xhosa—or, come to that, Afrikaans.
If, on the other hand, the innatist claims are substantially extended from the field of language acquisition into other forms of learning (for instance, as Chomsky seems to speculate at one point, into moral capacities), the charge of their irrelevance to racist issues may decline, but the possibility of an unfortunate and destructive type of belief in their “relevance” might increase. It seems odd that anyone should need reminding at the moment that it is the environmentalist view on matters of “intelligence” which has been identified as the liberal one. The entire question of the ideological significance of such studies is now a notorious moral and intellectual mess, but to accept their ideological significance while cleaving to innatist styles of explanation may not necessarily be the most progressive way out of it.
An important point here is that in the field of language acquisition, Chomsky has good reason to equate what is genetically determined with what is common to the species. That equation may very well hold for all innate human cognitive capacities, but there is no a priori guarantee from the nature of genetics or anything else that it must be so.
The ideological implications of Chomsky’s theories are by no means straight-forward or unambiguous, and Chomsky himself moves with dangerous speed and simplicity between his theoretical preoccupations and the political ideals for which he has so conspicuously stood up. In fact, the ideological effect of Chomsky’s work in language seems to me not so much to support or express distinctively socialist aspirations for society, or opposition to oppression, but rather, a stage further back, to assist a humane revaluation of tough-minded inquiry in the psychological sciences. His work, apart from its spectacular and ongoing effect in linguistics, constitutes the most powerful and encouraging reassurance that a psychological science which is recognizably continuous with the natural sciences does not have to treat human beings as very boring machines.
The recognition that human beings might be scientifically understood but are yet not just machines could well coexist with more than one kind of social or political view, not all equally liberal. But as well as being a vital truth in itself, it is a necessary step to any adequate views on these issues at all, including any adequate liberal views. It is here, in the humane understanding of science itself, rather than in more direct ideological interpretation, that the most general significance of Chomsky’s deeply impressive work is likely to be found.