Chomsky Explores Origins of Language
(Institute Professor Noam Chomsky is the 1991-92 Killian Faculty Achievement Award recipient. He will present the Killian Lecture, Language and the "Cognitive Revolution," Wednesday, April 8, at 4:30pm in Kresge Auditorium. Naomi F. Chase of the News Office interviewed him recently about his work.)
Chase: How did you get into linguistics?
Chomsky: By a back door. The work I did as a graduate student was not really regarded as linguistics. Linguistics was part of behavioral science, what's called structuralism. You had an array of data and various methods for finding structural patterns in which they fit.
Chase: Is that phonetics? Could you give me an example?
Chomsky: A standard graduate student assignment would be to go to some American Indian community, take down texts, then try to determine from the text whether there are patterns in which the sounds appear. Two phonetically different sounds might really be the same element in the structural patterns of the language. First, you had to find the phonemes, the significant sounds of the language. It's not just phonetics, it's words and something in between them. Then you try to find meaningful sequences of phonemes, which are called morphemes, and then patterns in which they appear. This is done by essentially taxonomic techniques that reduce behavior, according to the texts, into organized and meaningful patterns. As an undergraduate, I tried to refine the techniques. But I was also doing work on my own that had no real academic home.
Chase: What kind of work?
Chomsky: I was trying to use language to get a theoretical understanding of what goes on in the mind of the speaker. What capacities does the speaker or hearer have that enable them to carry out the behavior and what are the fixed capacities of the species that allow these individual capacities to develop? It has no relationship to what used to be called linguistics.
Chase: Because it's biologically rather than behaviorally centered?
Chomsky: The question is: what do you think behavior is? You can regard behavior as the topic of inquiry and you're trying to get an organized account of it. Or you can regard behavior as just data. There's a crucial difference between data and evidence. Data is just phenomena around you. Evidence is a relational notion. Something is evidence for something else. The second approach would say that data, or behavior, is interesting in so far as it is evidence for theoretical understanding. For me, the topic of inquiry is not the noises that you and I are making. That's just a way of helping us get to the topic of inquiry and learn something about it. The topic of inquiry is: what's going on in our minds? Now, if the data of our interactions help you find out about that, good. If those data don't help you find out about that, throw them away. If some psychological experiment would help you find out about that, fine. If biochemistry would someday help you find out about that, that's fine too. The important thing is that the topic of inquiry has just radically shifted. It is no longer the texts. You could think of it as somewhat analogous to the difference between--roughly speaking--natural history and natural science. There are various ways in which you can look at flowers and animals. If they are the topic of your inquiry, you write Newcomb's Wild Flower Guide, which describes flowers and leaves so you can recognize them on a walk in the woods. Or you write the Audubon Guide to Birds.
Chase: They're both natural history.
Chomsky: Yes. Their concern is the phenomena themselves, the flowers, the birds, an organized, systematic account of the phenomena.
Chase: What does natural science do?
Chomsky: It's quite different. The phenomena are of interest insofar as they teach you something about the laws of nature. What you're interested in is the laws of nature, which could be very general, like the laws of mechanics. Or you could be interested in the nature of a particular organism. What makes the visual cortex work the way it does, for example. The phenomena are of interest only insofar as they guide you to that understanding. You might be interested in the difference between a robin and a cardinal because it tells you something about the nature of the birds.
Chase: You were interested in the laws of nature which govern speech?
Chomsky: I was interested in the laws of nature which govern human thinking.
Chase: What is the relationship of language to thinking?
Chomsky: We can't answer that because we have no clear conception of what thinking is apart from some of its manifestations. Some of the manifestations of thinking are language and, of course, there has to be a close relation. Introspectively we're often very conscious of thinking in words. But we're also conscious of thinking without words. For example, it's common experience to say something and retract it, to say, "No, that's not what I meant." Then I'll try saying it a different way and finally I'll formulate what I really meant. How do we interpret that? Presumably, I intended to express a certain thought--whatever that means. I tried, but the linguistic mechanism didn't quite work. I compared it with the thought I had in mind. Insofar as experiences like that are real, we can see that thought and language are not the same thing.
Chase: So you solve the problems without needing to articulate them?
Chomsky: Right. Here's another example. You recognize someone as the same person you saw before somewhere. There's plenty of thinking going on, but there's no way of describing it, and there is no reason to doubt that animals do something very much like what we do at some level.
Chase: I`ve read that psychologist Peter Marler who has done work on bird song says, "The ability to hear rhythm seems to be present early in animal evolution but the ability to hear pitch relations may be unique in humans." He also says that "For too long those studying language development in people have over-emphasized cultural and social influences at the expense of the biological side." Isn't that what you're saying, too?
Chomsky: Our work emphasized the biological side from the very beginning. We've assumed that the capacities you and I have, to speak and understand, extend vastly beyond our experience. Now, from a behavioral science point of view that comes as quite a shock. In the 1950's if you asked people in psychology or linguistics to comment on how language is acquired, they would say the big problem is why it's so hard to acquire such a simple thing. The standard problem was thought to be one of over-learning. Why does it take so much time for the child to pick up this trivial thing? As soon as we begin to look seriously at the capacities that you and I have, instead of just organizing data around us, we find that the problem is the exact opposite. How do we know so much, having had such impoverished exposure?
Chase: When you say "know so much," you're talking about language?
Chomsky: Yes. You hear sentences now that are radically different from those you've heard in the past. But you can understand them. And if I shifted a couple of words around, you'd recognize it to be gibberish. That's what goes on in normal speech and when you look at it closely, you quickly discover that the properties of language that you and I have mastered, or any child has mastered--even the most simple properties--involve things for which we have had minimal, if any, exposure. That's true with the simplest part of language, like word learning. You learn the meaning of a word--but meaning is a very complicated thing. A dictionary doesn't come close to defining any word. It only gives hints that you can use as an intelligent human. If you really try to give a precise characterization of a word like "house" or "chase," you'll find it's remarkably intricate.
Chase: Could you give me an example.
Chomsky: Take, say, the word "building." If I say that somebody sees Building 20, the building we're in, you know they are outside. Right?
Chomsky: The only way you and I could see Building 20, sitting where we are, is if there were a mirror outside that window, and we looked at the mirror and saw Building 20 reflected. That already tells us that a building, or a box, or a house, is not just the physical entity itself. Rather its exterior surface has a special status. There's other evidence that supports the same conclusion. Words like "house," "building," "airplane," and "box" have very strange properties which we come to know without any experience. Nobody ever taught us that if a house is brown, that means its exterior is brown. There's an immense amount that we know about words that we were never taught.
Chase: What does this mean?
Chomsky: What you know can only come from two sources--inside and outside. If it doesn't come from what's outside us, from our experience, it's got to come from our inner nature. Having a language is something like having arms, not wings. Or like undergoing puberty. These properties are not simply due to the nutritional environment of the embryo.
Chase: You're saying it's part of our biological design.
Chomsky: Yes. Contrary to what people thought, language is not taught, not even learned. It's something that your mind grows in a particular environment, just the way your body grows in a particular environment. Now your shape, how heavy you are, or the onset of puberty may vary depending on environment, nutritional level, etc. Normal physical development is always triggered by external phenomena, and also responds in some modified way to external phenomena. But its major properties are built in. And so is language. A child can't help acquiring it, though we can improve the way it's acquired as we can improve the way a child walks, marginally. Maybe it will walk two weeks earlier but it's going to walk, because it's a human.
Chase: What is the relationship to hearing, though? Because it's said that birds that are deaf do not learn to sing.
Chomsky: A child who's deaf will not learn to speak. But it still has the language capacity and it will learn how to sign.
Chase: I've read that deaf babies make up their own sign system.
Chomsky: Laura Pettito who was working in Cambridge and has been studying deaf children for many years, shows that deaf children go through early stages of acquisition of signing that parallel hearing children with speech. And Susan Golden-Meadow's work shows that children without any linguistic environment at all may nevertheless develop their own sign languages which have the basic properties of spoken language. That's the extreme case of deprivation of experience. So with zero experience, the system may still manifest itself in a particular way.
Chase: You've said that the everyday phenomena of language acquisition is just the growth of capacities stimulated by the environment. And the same is true of grammar?
Chomsky: Yes. If I say, "He thinks that John is intelligent," you know that "he" is somebody other than John. If I say "his mother thinks John is intelligent," "his" may refer to John. Those are things we aren't taught. The basic features of language structure are built into our biological nature. They're like the principles that determine why mammals see things in a particular way and insects see things in a different way.
Chase: What are the big questions left to answer in linguistics?
Chomsky: The traditional questions. The big one is what is it about you and me that distinguishes us from a rock or a bird?
Chase: The old question: What is man?
Chomsky: Right. What are humans? One distinctive feature about humans is the language capacity. It's central for our present existence and it doesn't seem to have anything analogous or homologous to other organisms. It seems unique to the human species, it's essential and it's also uniform across the species. A sub-question of that is: what are the varieties of this characteristic. In what ways can human languages vary? Then there are questions about how these systems grow in the individual, how they are used when we express our thoughts, how they are realized in what we assume to be brain mechanisms. These are open questions. In principle there are questions about the evolution of these capacities. Whether those questions are realistic or not we do not know.
Chase: Like when did humans start to speak?
Chomsky: In principle you can ask about the evolutionary history of any capacity. But the fact that you can ask a question doesn't mean it's answerable. Opinions differ. You can find people right on this campus who think it's a very answerable question.
Chase: What do you think?
Chomsky: I don't think we have a prayer of answering it on the basis of anything that's now understood.
Chase: What is the connection between your interest in linguistics and your interest in politics?
Chomsky: The connections are very abstract. But take what I said about our ability to use language and make judgements about things vastly beyond our experience. Well, the same logic applies in other areas. You and I are capable of making moral judgments that go vastly beyond any direct experience we have. We make moral judgments in new cases and we often know how to make them in quite intricate cases. You have a certain conception of how a society ought to be organized, you can see things that you don't like and you can change them. That must come from some concept of human beings and what's right for them. What needs they have, what rights they have and so on.
Chase: So certain ideas are innate.
Chomsky: Here we are turning to speculation. Perhaps our nature has at its core what has sometimes been called an instinct for freedom. If so, perhaps this has some relation to what is found in the study of human cognitive abilities, the possibility that at the core of your use of language is a certain striking and dramatic kind of freedom that was traditionally taken to be the basic difference between humans and machines.
A version of this article appeared in the April 1, 1992 issue of MIT Tech Talk (Volume 36, Number 25).