The Science of Language

by Noam Chomsky

  [Keep in mind that these complex] problems are too hard for physics, even in its own domain. Physics does not study what goes on in the world – it studies what goes on under the highly controlled conditions of extremely artificial experiments. That's the Galilean revolution. What goes on in the world is just too complicated, so you try to discover some principles by studying artificially simple situations. And, of course, that [yields results that are] not obviously true for anything more than the highly simple entities that physics restricts itself to. It has been a non-trivial task to show that engineers, say, can substantially improve their practices by relying on fundamental principles of physics, or medical practitioners on biology, and so on. That's one of the reasons why I'm skeptical about work some friends of mine have done in designing an ideal society. A lot of it's interesting, but I don't think we have any idea of whether humans could live in that society.

  JM: Well, little progress has been made; but is the project itself still plausible?

  NC: I think it's plausible and everyone is engaged in it all the time. If you're even a moral agent – you care about the effects of your actions – you're presupposing something about human nature. It doesn't matter whether you're a revolutionary or a conservative, or a reformist, or whatever you are. If you regard yourself as a moral agent – you're trying to think about actions, or morals, or plans, or ideas that might make human life better – you're presupposing something about what is fundamental human nature. You can't articulate it. But that's not surprising – most of what we do we can't formulate the reasons for, but it's got to be there, otherwise human action makes no sense. If you deal with your children in a certain way, you're presumably trying to help them, not harm them. You must be presupposing something about what's good for them. Well, where does that come from? It comes from some intuitive conception of what their fundamental nature is. And that must be true of every human interaction.

  JM: You aren't going to deny that there are at least some kinds of distinctions that can be made even now?

  NC: There are very striking ones. I just gave a talk in England [at the Royal Institute of Philosophy (see Chomsky 2005b)] . . . It was a talk about truisms – moral truisms and why they're important. I started out by apologizing because I'm only going to bring up platitudes. But they must not be dismissed because they're always ignored – almost universally – and ignoring them leads to horrible effects. So take the most elementary truism we can imagine – the principle of universality. We apply to ourselves the same standards we apply to others – probably more rigorous standards if you're serious. If you accept that principle, practically the entire discussion and policy of international affairs collapses, because it's based on denial of the principle of universality. Sometimes it's explicit. I congratulate the people who say straight out that we've got to be similar to Nazis – like Henry Kissinger. So when the president's national security strategy came out in September 2002, there was a lot of criticism of it, although the criticisms were narrow, and most of them were of the form, “Well, the principles are right, but they're going about it the wrong way.” But Kissinger was straight about it. Good for him. He said yes, it's a good principle but – his words were – It must not be “universalized.” It's for us; we have the right of aggression. Maybe we delegate it to clients, but certainly nobody else does. And, as I say, he has to be congratulated for honestly saying what is always presupposed: that we must adopt the doctrines of the Nazis. We can't accept the most elementary moral truism. We have to firmly reject it. And if you look beyond, that's commonly true. So I think you can say a lot about moral truisms which people will accept abstractly, but reject instantly as soon as they are applied to their own actions and affairs.

  JM: So there are at least some fairly recognizable facts about our moral nature . . .

  NC: Well, if someone doesn't at least accept that, then they should just have the decency to shut up and not say anything. That does wipe out a good part of the commentary on international affairs, at least – and on domestic affairs too. So it's not a small point – even though it's a platitude.

  16 Human nature and evolution: thoughts on sociobiology and evolutionary psychology

  JM: Looking at how you have studied human nature now, could we begin by looking at some apparent differences between your approach to the biological study of human nature and that pursued by sociobiologists and evolutionary psychologists such as Wilson, Pinker, Dawkins, and many others. They seem to think that the science of human nature should aim to explain human behavior, and they conceive of the various biologically based affects and cognitive capacities on which they focus as biologically evolved in the long-term selectional way that neo-Darwinists favor. Your approach to the science of human nature pays scant attention to the explanation of behavior – in fact you've maintained since at least the late 1950s that while our everyday intentional, commonsense accounts of behavior might be successful enough for practical purposes, the scientific explanation of behavior is a hopeless project, except perhaps for extremely simple organisms. And while you have no objection to seeking selectional accounts of the development of affects and cognitive capacities, you emphasize in the case of human language – certainly the most distinctive and central mental faculty, one that no other creature has, and that seems to underlie many of our cognitive accomplishments – there is no evidence that a long-term selectional story will work. There are reasons to think that language was introduced at a single stroke with the introduction of Merge, perhaps some fifty or sixty thousand years ago. Could you comment on these apparent differences, and on what we might hope to accomplish in a naturalistic science of human nature?

  NC: First of all, just to be historically accurate, what's now called evolutionary psychology or sociobiology begins with Kropotkin in his critique of social Darwinism. But he was a Darwinist. He argued on selectional grounds that you would expect what he called “mutual aid” – cooperative societies that happened to be his vision of communitarian anarchism. Well, why is Kropotkin disregarded? – because people don't like his conclusions. If you mention him to Dawkins and so on, they'll just ridicule him. Now is it because he didn't have any evidence? Well, he didn't have any relevant scientific evidence. Does anyone have such evidence for anything else? Barely. They don't like his conclusions.

  There are several different questions here. What is the nature – what is human nature, or ant nature? Well, those are scientific questions. Separate from that is the question of what role selection had in developing them. And those are just two different scientific questions, not to be intermingled. Part of the problem with the kind of pop biology that's common today is they're just intermingled – it's assumed that if there's a nature, it's got to be selected. It doesn't make any sense – it doesn't make any sense for the kidney or the visual system or anything else, and it doesn't make any sense here.

  There is a nature, undoubtedly. People who argue against it for a blank slate – that's just puffery for a popular audience. It doesn't mean anything – nobody ever believed in that who was sensible. So yes, there's a fixed nature and it developed somehow, but we do not know for any aspect of it (whether it's the chin or the visual system or the bones in the ear or whatever it may be) – we don't know the answer to how it evolved until we know the answer. And the answers when we find them will often be very surprising. So there's no issue here about whether natural selection operates – obviously it does – but there are some very big questions about the channel within which it does, and about other factors involved in evolution, of which many are known. So we have to separate totally the kind of rhetorical posturing about selection and the question of intrinsic human nature.

  Well, there is recent work – like, say, on kin selection, Hamilton's work – which suggests some plausible evolutionary basis for certain kinds of what appear to be altruism. But it's pretty narrow. If you pursue kin selection to its limits, you're going to have a hard time explaining why humans devote enormous energy and take tremendous
risks to save dolphins but don't care how many children are dying in Africa. Something else is happening. It's interesting work; I don't want to denigrate it. But the results that have some human application come down to the fact that I'm going to pay more attention to my children than to my nephews. We didn't need biology to tell us that, and it doesn't really tell us much beyond that. And it also doesn't tell you why I'm going to pay just as much attention to an adopted child as to my own children – and take the same attitude towards it, even though I know it's adopted. [Nor does it tell you] why many people care more about their cats and dogs than they do about their children – or take dolphins, which are the classic case. So it just doesn't get us very far. It's interesting work and we learn something about insects and other organisms and something about social behavior, but very little about humans that has any implications.

  We know in advance that that's going to be true. Science deals with simple questions. It can't deal with questions that are beyond the borders of understanding. We kind of chip away at the limits.

  Take the evolution of language. It's a question; and so is the evolution of bee communication a question. But just compare sometime the literature on one with the literature on the other. There are libraries of material on the evolution of human language and some scattered technical papers on the evolution of bee communication, which mostly point out that it's too hard to study, although it's vastly easier to study than evolution of human language. This is just irrational. Not that it's wrong to study the evolution of human language and I think there are some directions to look, like the one you mentioned. There's comparative evidence about the sensory-motor system, which may turn out to be very peripheral – but it's there. So sure, study it to the extent you can, but sensibly – knowing when you're talking and producing serious science and when you're gesturing rhetorically to a general public who you're misleading. Those are important distinctions, and I think if we make those distinctions, a lot of this literature pretty much disappears.

  To some extent, there are other factors that enter into it which cause it to be misleading. Many of these people, like Dawkins, regard themselves very plausibly as fighting a battle for scientific rationality against creationists and fanatics and so on. And yes, that's an important social activity to be engaged in, but not by misleading people about the nature of evolution – that's not a contribution to scientific rationality. Tell them the truth about evolution, which is that selection plays some kind of a role, but you don't know how much until you know. It could be small, it could be large; it could [in principle even] be nonexistent. We have to find out. In the few cases where something has been learned, it's often very surprising, like the evolution of the eye. What appears to be the case is completely different from what had been speculated for centuries in biology, and the same is true of many other things – it could be true of human language. So there's nothing wrong with sociobiology or evolutionary psychology – the field that Kropotkin basically invented – but it has to be done seriously and without pretense.

  JM: There are some specific hypotheses – let me just pursue one. Robert Trivers suggested back in the seventies that cooperative behavior could have evolved among biological creatures that are often conceived, where biologically unrelated, to be essentially selfish. He assumed that cooperation could have evolved among biologically selfish creatures if it were generally to involve reciprocity – when x does something for y, x can expect something in return. The result is a back-scratching conception of cooperation and social behavior. Trivers's work has been given center stage by sociobiologists and evolutionary psychologists. They have suggested – in a way reminiscent of utilitarian thinking – that his form of reciprocal altruism offers the key to understanding the biological basis of morality. As I understand your view, you do not think of humans as essentially selfish. You think of people as capable of – and getting satisfaction from – aiding others who are not kin and not even tribe, and cannot be expected to reciprocate. Furthermore you do not think that neo-Darwinian selectional stories exhaust what can be said about the biological foundations for language and perhaps for other domains. Are there alternative, non-Triversian biological grounds for morality, and does something like Humean sympathy offer such an alternative ground?

  NC: Trivers's work is quite interesting. I don't think it gets us very far. I don't think it explains why people are willing to support the system of social security that's going to give a disabled widow across town enough food to survive – or the fact that we care more about dolphins than we do about people pretty near us who could help us. It just deals with a very small topic. It's interesting; and there are game-theoretic approaches that try to work out the consequences. All that's fine and should be done. But does it yield conclusions of human significance or scientific interest? Well, not of human significance as far as I can see. Of scientific interest, yes, but within a very narrow domain. Are there other human capacities that enter into our moral nature? I just don't see how that can be doubted. We know too much from our own experience and intuition; it shows that there's a huge domain – in fact, virtually all of human action, thought, and interpretation that doesn't fall within this category.

  Is there some other evolutionary explanation for the rest of it? No. But that's true of almost everything. There's no evolutionary explanation for bee communication, or the most elementary questions of how simple organisms function – nematodes, for example. So yes, of course, science shines often-penetrating light on extremely simple questions. One of the reasons that physics is such a successful science is that it is granted the unique privilege of keeping to questions that are extremely simple. If the helium atom is too hard to study, you give it to the chemists. Other fields don't have that privilege but deal with the level of complexity that they're presented with, and as a result they're very shallow by comparison. In these areas – evolutionary explanations – we're just groping in the dark for the moment – there aren't good ideas, even for much simpler organisms. So all this work is fine. If you can achieve some plausible confirmed scientific results, everybody applauds; and there're no issues. What are its implications for human life and society? – well, that you have to investigate, and I think when you do, you find them extremely limited. Hume's and Smith's assumptions are, I think, much more plausible and lead to suggestions about how to behave in the world that are far more reasonable – and in fact that we adopt all the time.

  JM: If we're decent, anyway.

  NC: If we're decent. And that's why people like Kissinger are important – to tell us that we must reject our fundamental moral nature for reasons of power and so on. OK; that's interesting to hear. Now tell me something about moral human nature.

  JM: Kissinger does strike me as pathological.

  NC: . . . or maybe honest. I kind of prefer him to people who look at that view with horror and then presuppose it.

  17 Human nature again

  JM: Could we get back to human nature again? I'm still trying to figure out just what is distinctive about human nature. What I mean by ‘distinctive' is: distinguishing us from other sorts of primates, or apes. Clearly Merge – some kind of recursive system – human conceptual systems; in that, we are distinct. Is there anything else you've thought of?

  NC: If you look at language, you can find a thousand things that look different. If you look at a system you don't understand, everything looks special. As you begin to understand it, things begin to fall into place, and you see that some things that look special, really aren't. Take Move – the displacement phenomenon. It's just a fact about language that displacement is ubiquitous. All over the place, you're pronouncing something in one position, and interpreting it in some other position. That's the crude phenomenon of displacement – it's just inescapable. It's always seemed to me some kind of imperfection in language – a strange phenomenon of language that has to be explained somehow. And now, I think, we can see that it's an inevitable part of language: you'd have to explain why it isn't around. Because
if you do have the fundamental recursive operation which forms hierarchic structures of discrete infinity, one of the possibilities – which you'd have to stipulate to eliminate – is what amounts to movement – taking something from within one of the units you've formed and putting it at the edge; that's movement. So what looked like a fundamental property of language and also looked like a strange imperfection of language turns out to be an inevitable property of language – and then the question is, how is it used, how does it work, and so on and so forth. That's a serious rethinking of perspective. And that's what happens when you learn something about what looks like a chaotic system.

  It's been the same throughout the history of the sciences. Everything you look at appears to be completely chaotic and inexplicable. You begin to tease a few principles out of it, and some of it – maybe a lot – falls into place; and the rest looks chaotic. How far can you go in this? You don't know. But the question of what is unique to humans as distinct from other primates arises right there – how much of what we see in language and human theory of mind and the curious nature of human concepts, which are not referential in the animal sense – how much of this is unique to humans, how much isn't? For example, the special properties of phonological rules – are they unique to humans, or will they be shown some day to be just an optimal way of meeting conditions that phonology has to meet? Phonology has to meet the condition that it relates the sensory-motor system to objects created by the computational operations of hierarchic recursion. Well, maybe phonology is an optimal way of meeting those conditions, and its various apparent complexities arise as a solution to that problem. Nobody knows enough to know whether that's true; but it wouldn't be very surprising. In fact, I strongly suspect that something like that must be true – otherwise, how would anybody ever pick it up?