The Science of Language

by Noam Chomsky

  Because I was writing about it, I recently went through some of the records of the biolinguistics conferences in the sixties and seventies, and it's always schema, plan, position – which is what's wrong. It's just impassible. Language just has a highly specific, highly articulated format, and that's the only way you can account for language acquisition. That looked to me, and to everybody, like a convincing argument. Well, when the principles and parameters framework came along, it undercut that argument. It didn't answer the questions, but it undercut the argument, by looking at everything in a different way. Acquisition was disassociated from the format for grammar. Acquisition is fixing the parameters, and the grammar is whatever it is. It is no longer part of the acquisition process, so it is at least conceivable that it's a best possible solution to other conditions. Then you can start worrying about the third factor.

  JM: To pursue the parallel for the social sciences: at least in the case of linguistics, there was some target science to which linguistics could be accommodated. The thought was: well, it's got something to do with biology.

  NC: OK, but biology didn't help at all. You didn't get anything. The most that biology provided was comparative ethology – which amounted to little more than saying that all these guys who were saying that everything is stimulus-response are wrong. What could you find in biology?

  JM: Well, at least on the assumption that language is unique to human beings – this would incorporate the ethology facts too – it would seem that that must have something to do with the fact that it is a biological characteristic of the species . . .

  NC: . . . an organ of some kind . . .

  JM: an organ of some sort.

  NC: But nothing was known about [mental] organs. Remember that the beginnings of the understanding of the visual system that we have today came in around 1960. So that was coming out of the same background of interests.

  JM: Granted, you don't get any specific proposals of principles, or anything like that. But there was an appearance of universality and early acquisition, and a thought that that must be due to biology. Do you have anything like that for, say, economics? Again, in the case of the social sciences: I just don't know what would count as a target science . . .

  NC: I just don't think you can count much on borrowing from other sources. It's just never worked. If you can get some hints from something else, well, then, ok: but you're lucky.

  JM: But it did at least look like linguistics should be seen as a branch of biology. That's what posed the ‘what about a third factor contribution?’ problem. Why else would the format idea with its need for a high degree of language specificity pose a problem? . . .

  NC: . . . that it could be incorporated in biology; but that might require a change in biology.

  JM: . . . granted.

  NC: Biology itself provided almost nothing.

  JM: I must be very dense, but again: where would you go for the social sciences? Systems theory? What would you get?

  NC: Well, that's also a part of biology. You study ant colonies; that's a part of biology.

  You don't get free passes in this game.

  JM: Agreed – fully. Many thanks, Noam. I won't take up any more of your time.

  25 Linguistics and politics

  JM: What is the relationship between – I know you've been asked this several times (including by me), and you as often have for various reasons dodged the question –

  NC: . . . then I'll dodge it again; because I'm sure that the reasons still hold . . .

  JM: Well, I'll try anyway: what is the relationship between your linguistic work and your political work?

  NC: Well, it's principled, but it's weak . . .

  JM: You've said that there's no deep intellectual connection; I've always read that as saying that there's no way of deducing . . .

  NC: . . . there's no deductive connection. You could take any view on either of these topics, and it wouldn't be inconsistent to hold them . . . You know the line, and I don't have to repeat it. There's some point at which a commitment to human freedom enters into both. But you can't do much with that in itself.

  Appendices

  Appendix I: I-concepts, I-beliefs, and I-language

  Chomsky's notion of an I-language was introduced in part (in 1986) by appeal to a contrast with what he called an “E-language” approach to the study of language. An E-language approach is one that studies language that is ‘externalized.’ One form that externalization might take is found in a philosophers’ favorite, the notion of a public language. What is a public language? David Lewis and Wilfrid Sellars, among many others, assume that a language is an institution shared by individuals in a population, taught by training procedures with the aim of getting the child to conform to the rules for word and sentence usage (for Lewis, “conventions,” and for Sellars, “practices”) of the relevant population. This view turns out to be hopeless as a basis for scientific research for reasons taken up in appendices VI and XI. It does, however, conform quite nicely to a commonsense conception of language.

  Another version of an E-language approach is found in Quine, where he insists that there is no “fact of the matter” with regard to deciding between two grammars for ‘a language,’ so long as they are “extensionally equivalent.” To say that they are extensionally equivalent, each would have to generate all and only the same set of sentences, where a sentence is understood to be a ‘string’ of ‘words.’ To make sense of this, one must think that it is possible to identify a language for purposes of scientific investigation with a set – an infinite set – of strings. However, that belief is erroneous, for several reasons that become clear below; essentially, a language is a system in the head that has the competence to generate a potential infinity of sound–meaning pairs, where these pairs are defined by appeal to the theory, as is the recursive procedure that can yield them. What a person actually produces in various contexts during his or her lifetime is a very different creature: in Chomsky's terminology, it is an “epiphenomenon,” not a grouping of strings that can be the subject matter of a naturalistic scientific effort.

  Still another way to construe a language ‘externally’ is to conceive of it as an abstract entity of some sort, located not in the head, but in some abstract realm. This view raises many questions unique to it – what is this abstract entity, where is it, how do people acquire it, how does it play a role in speech/sign production, how does it differ from the naturalistic description of the language organ, among others. Moreover, given a choice between an existing natural science of language with a good track record (Chomsky's) and a proposal that lacks evidence, serious theoretical proposals, or plausible answers to any of the questions mentioned, there is nothing to recommend it.

  An I-language approach, in contrast, is a study of language that is “in the mind/brain.” I here stands for “individual, internal, and intensional” and – one could add – “innate and intrinsic.” This approach assumes that the target of the science of language is a system in a person's head that is a (developed, grown) state of a “mental faculty,” a mental faculty that can be investigated using the methods of the natural sciences, which – among other things – idealize and offer naturalistic and empirically supportable hypotheses concerning the natures of their subject matters. To offer an hypothesis is to provide a theory of what that internal system is and – since any individual's I-language is a developed state of an initial, universal state (called Universal Grammar) – one is committed to thinking of an I-language as a grown/developed biological ‘organ’ in a person's mind/brain, and to conceiving of the science of language in the form of a computational theory of system ‘in the head’ as an abstract version of biology – essentially, as an I-language. A satisfactory naturalistic theory of an I-language cannot limit itself to a single person's head at a specific stage of development and lexical store. Rather, a descriptively and explanatorily adequate science of language is a science that describes and explains the growth and biophysically possible final sta
tes of a system inside the head. To develop a theoretical under-standing of any such system – of any specific I-language – the only way is to construct a theory that will encompass it and all others. That requires a theory that hypothesizes a biophysically fixed “initial state” (Universal Grammar). With this theory and adequate theoretical descriptions of how, given experience (input), biology and other natural systems constrain growth/development from an initial state to a stable, final state, we would have a way to describe the growth of any I-language. That makes the I-language approach to language intensional: the theory says what a language can be/is, and any biophysically possible language is a recursive system in the head, not a set of sentences ‘out there’ in some sense, whether a set of practices of a population, a set of strings, or an abstract entity. The science of language offers a formal theory of the system and its possible states; in doing so, it specifies with a formal or mathematical function the developed state of any specific person's language faculty, a faculty that takes words (theoretically defined) as input and yields expressions/sentences (formally defined) as outputs. The function is specified “under intension,” not (per impossibile) via enumerating its (infinite) output. Thus, a language is internal; it is also individual (think of each person's state of their universal language faculty as something like an idiolect, even though that is a dubious notion); and it is intensional. It is also innate, by virtue of an assumption – that language is a biophysical system – that is warranted because the biophysical theories of it are successful. And it and its properties are studied not in their relationship to something else ‘outside,’ but in terms of their intrinsic properties. That is part of what Chomsky means when he says that his study of language, including the study of linguistic meaning, is “syntactic.” The point is explained in Appendix XI.

  In addition to I-languages, one can speak of I-concepts and I-beliefs. Chomsky explains in the interchange below: JM: I-concept and I-belief: what are they?

  NC: Well, [I-] internal, individual, and intensional where intensional amounts here to ‘theory-defined’ . . . take me. I have some way of interpreting and thinking about the world, applying my actions, and so on and so forth, and we don't really know how it's done, but there are presumably some elements that enter into shaping ways of thinking about the world. Whatever they are, we call them concepts. It's like Newton saying that the smallest elements of the world are corpuscles. I don't know what they are, but there have got to be some. So whatever is there in the head that is used for shaping the ways we conceive and perceive the world, there are concepts. And we can presumably make more complicated constructions out of them, and those are what we call thoughts. Now some of them we have a degree of confidence in, and those we call beliefs. But as to what they are, that is a scientific question; it's a matter of finding out what the world is made up of.

  I, then, just means ‘I.’

  Now, the standard view of this is that these things – concepts, beliefs, and so on – are outside of people's heads in some kind of universe . . .

  JM: Fregean abstract entities, perhaps . . .

  NC: I don't see a reason to believe any of that. In fact, I think that historically a lot of those ways of thinking about things come from the fact that the work is mostly done in English and German, which happen to be very nominalizing languages. English in particular – take the word belief. It can't be translated into many other languages, except with a paraphrase involving a verb. These are English conceptions which have rough counterparts in other languages, but in most languages, there's no such thing as belief, and there's no such word as believe, it's think. Take Hebrew, for example. There is a word for I believe, and it means ‘I have faith.’ If you want to say I believe it's raining, you say I think it's raining. And most other languages are like that. Now, English goes beyond that and even nominalizes the notion of [thought]. If Wittgenstein and John Austin and so on taught us anything, it's at least not to do that.

  JM: Not to mention some seventeenth- and eighteenth-century philosophers. And I take it that the assumption is also that individuals – who happen to be, say, English speakers – don't necessarily share the same I-concepts.

  NC: Any more than that they share the same language, or visual system. There's no English vision, or American visual system. There's my [I-]visual system.

  JM: Again, what are I-beliefs?

  NC: Whatever thoughts we have that we have some confidence in . . . whatever they might turn out to be. We don't know what they might turn out to be. I mean, we make the best guess we can, but we can't take it too seriously, any more than you can take the domain of corpuscles too seriously. There's something – there are some building blocks – but who knows what they are?

  I was reading Science magazine last night; it has an interesting article on water. Apparently the H2O molecule is one of the most complicated molecules there is because of its different states. There are different angles, different bonds, so different structures . . . [See Ruan et al. (2004) and Wernet et al. (2004), plus a follow-up article to both: Zubavicus & Grunze (2004).] Water assumes different configurations depending on different chemical environments, with different bond angles, and different bond lengths; it forms and reforms in various configurations while undergoing changes in temperature, substrate, pressure, etc. Its surface or ‘interface’ properties – those that play a crucial role in all sorts of processes, including critical biological and physiological ones – are variable and depend on a large number of internal and external factors. It's a very complex object which, the better techniques there are for studying it, the less confident we become that we understand it. So, what philosophers speculate about when they for various reasons identify water with H2O . . .

  JM: . . . well, nothing will stop them.

  NC: No.

  Appendix II: The several uses of “function”

  To understand better what is at stake, it helps to taxonomize some of the several uses of the term “function.” Along the way, I speak to what appear to be features of Chomsky's views of science, of common sense, and of our access to the world through the concepts that we have available in common sense and that we can create in the sciences. While I cover these topics in other appendices in more detail, I raise them here too because the term “function” has different uses in the sciences than it does in the commonsense framework, and conflating the frameworks risks conflating the uses of the term. The two frameworks have very different orientations: they serve different kinds of human cognitive project. The kind of understanding one gets in common sense serves the interests of agents who must act to satisfy their needs and desires. Because of this orientation, it is no surprise that common sense and the concepts it makes available to understand the world and others – the concepts expressed in natural languages – have an anthropocentric focus. Nor should it be a surprise that the various metaphysical and epistemological ‘theories’ offered by the majority of philosophers from Plato on are anthropocentric. Consider, for example, Aristotle's universe with the earth at its core, Moore's commonsense philosophy, Wittgenstein and “ordinary language philosophy,” and the currently still-popular view of language as a social institution, made by humans to serve their epistemic interests, and transmitted in training procedures. The sciences – at least, the advanced mathematical ones – have developed slowly over centuries and increasingly won a struggle against the pull of common sense and its anthropocentric orientation. They are oriented toward objective description and explanation and, as scientists early in the seventeenth century soon found out, the picture the sciences paint of the world and of human beings is very different from that depicted in common sense.

  Biology, a science of particular interest in the study of language, seems still to be in transition; it seems still to owe some allegiance to commonsense understanding. Darwin's view of natural selection (even supplemented with genes in what is now called “neo-Darwinism”) and the concept of adaptation built on it remains indebted to what Alan Turing and Richar
d Lewontin call “history,” not to mathematical formal theories of structure and form and to the constraints they impose on both potential modifications in and growth/development of organisms. Indeed, there are some naïve versions of selection and adaptation that appear in evolutionary discussions that are difficult to distinguish from historicized versions of behaviorism, a point B. F. Skinner emphasized in his Beyond Freedom and Dignity (1971). Chomsky remarks on this in the main text. But the explanatory role of ‘history’ in biology is likely to diminish. “Evo-devo,” discoveries of a massive degree of conservation in genetic materials across all species, recognition of the crucial role of gene placement and of their timing mechanisms in explaining structure and its development, plus other work in biology – including Chomsky's contributions to the biology of language – have emphasized the role of other naturalistic sciences such as physics and chemistry, and accounts of formal constraints on development and growth that constrain naturally possible forms. In the case of biology, these impose constraints on biological structure and possible variations and modifications in biological systems, including individual mental ‘organs.’ This has had the effect of reducing the role of selection in the explanation of structure, making it less important: it ‘chooses’ which structures will survive, and does not always do so efficiently – as Chomsky points out in our 2009 update discussion on pages 51–52. It does not create or build structure in any interesting way; that is in large measure due to contributions to possible form mentioned above. For an overview of the many issues involved in what seems to be at least the partial dethroning of the role of natural selection in biological explanation, see Stephen J. Gould's monumental (2002) The Structure of Evolutionary Theory. See also Chomsky's views on the topic in the main text and writings, related to those found in Part I of Jerry Fodor and Massimo Piattelli-Palmarini's What Darwin Got Wrong (2010).