The Science of Language

by Noam Chomsky

  NC: My own interpretation of those proposals is that they're suggestions about our science-forming capacities. So these epistemic limits . . . your proposals should be consistent, try to avoid redundancy, try to unify different aspects of science – physical reductionism, say – I think that all of those can only be understood as explorations of the way that we, as particular creatures, try to proceed to gain our best understanding of the world in a systematic fashion. That's the way we do it . . . But if you want a proof that it's the right way, well, I don't see how that can be possible. All you can say is that it's the best that we can do. We may discover that we're always going off track, in which case maybe that's irremediable. If we can't find a different track, it's irremediable. And sometimes – if you look at history – humans have found a different track by lowering their sights. So, for example, lowering one's sights from understanding the world to understanding theories about the world led to a rather significant change, and it's a change – it's sort of symbolized by Newton – that took several centuries to become internalized.

  JM: It gives you a very interesting understanding of Wittgenstein's Tractatus and a number of other works in that genre. Was Russell engaged on that sort of project, at least as you understand him – in his early work?

  NC: Pre-Tractatus?

  JM: Russell before [Wittgenstein's] Tractatus.

  NC: He was engaged in a kind of conceptual analysis that I think he regarded as giving us insight into the nature of reality. But it was conceptual analysis – as, for example, the theory of descriptions.

  JM: By the time of Human Knowledge: Its Scope and Limits, he was quite explicit about making proposals, based on an understanding of human nature.

  NC: There it is explicit. It becomes a much more subtle and sophisticated approach that does appear to recognize – as the very title of the book indicates – that we're dealing with some organic phenomenon that is going to have its scopes and limits. He doesn't quite say it like that, but I don't know any other way of reading it.

  JM: You do get normative overtones in a lot of philosophical writing – this is the way you ought to proceed . . .

  NC: There's nothing wrong with that: you ought to do it by our lights, by the way we see things. It's the same with moral judgments.

  JM: How does that differ from saying, “this is the way we have to do it?”

  NC: It would be “this is the way we have to do it” if we knew enough about ourselves to say that there aren't any choices. Like “I have to fall off a cliff if I jump; I can't help it.” But we don't have that kind of understanding of much more complex things, like great areas of our lives.

  These kinds of questions come up in naturalistic moral theories and naturalistic epistemological theories, and in both – which are the traditional ones – you can try to work out what our moral instincts are and what our moral faculties are. But there's a gap between that and what's objectively right – an unbridgeable gap from the standpoint of some non-human creature that can be understood to be right, something of which our moral nature has only a partial grasp.

  And the same is true of epistemology. What makes the best theory? People use the term “best theory” freely, but what is the best theory? Well, we can try to sharpen up our criteria, as we understand them, but we're doing something analogous to investigating our own moral nature. We're investigating our epistemological nature, and within that framework you can come up with some notion of best theory that is on a par with our notion of right behavior. But again, from some point of view or standpoint that is external to us – which we can't take, because we're us, not that external thing – it could be evaluated in quite different ways.

  JM: You're not assuming that we can make sense of the notion of an objective right or an objective truth?

  NC: I do believe that there is an objective truth; ok, so I'm a naïve realist of sorts – I can't help it. But I think that if we think about ourselves, we will see that there is no way to have any confidence about it. We can have confidence about the fact that this is the best I can do with my cognitive capacities – and we can have less confidence, because we understand less, that this is the right way to behave in accordance with our moral nature. And I presume that we all have the same cognitive capacities and moral nature. But – and here we get to the last line of the Tractatus [of Wittgenstein] – beyond this, we just have to keep silent.

  JM: But there are people such as Peirce who tried to give some kind of content to the notion that there is an objective truth. Whatever an ideal science happens to discover . . .

  NC: He was putting it in the framework of an extremely poor evolutionary argument. It was a fallacious argument. If you take that argument away, then the conclusion collapses. His argument was that we were selected to attain the truth. We wouldn't have survived if we didn't have a truth-seeking capacity, and therefore if we just pursue it to the end, we would have the truth. That's the core of the argument. But it just doesn't work. Nothing in human evolution selected people who were good at quantum theory . . .

  JM: If one believed that the mind were – counterfactually – something like a universal device, that we have some kind of capacity to be able to solve every problem we might encounter . . .

  NC: . . . and to pose any kind of question . . .

  JM: . . . and to pose any kind of question . . .

  NC: I just don't know what that would mean. That's no organic entity that we can even conceive of.

  JM: Still, a person who held that kind of belief might have a different kind of view about human cognitive capacities and objective truth . . .

  NC: A person who held such a belief would be saying that we are somehow angels. There couldn't be a creature in the universe that would incorporate our cognitive capacities as a sub-part, maybe reject them the way we reject commonsense contact mechanics, and maybe go on to ask further questions that we don't know how to pose, and maybe find answers to those questions, without limit. How can we say that? How do we go beyond the limits of possible organic development?

  JM: When you speak of investigating the limits of our cognitive capacities, I assume that you are allowing that there might be cognitive capacities that we simply have not been able to . . .

  NC: . . . have not produced thus far. Yes; that's not at all surprising. Take, say, your arithmetical capacity. That wasn't used throughout almost all of human evolutionary history. There's just a tiny little fleck of time during which that capacity has ever been used. This is what bothered Anthony Wallace in his debates with Darwin. He argued that things like a mathematical capacity couldn't have been selected, because they were never used. If you don't use it, it can't be selected. But they've got to be in there somehow. And, he suggested, there must be some other forces like gravitation, chemical reaction, and so on that entered into the development of what he called human moral and intellectual capacities. That was regarded at the time as a kind of mysticism. But we should regard it as just sane science. It's on a par with what Newton was unable to accept, but should have: there are forces in nature that are beyond interaction through contact. Newton said it [there are such forces – specifically, gravity], but he didn't believe it. But it was right.

  JM: If one had a view of human biology or perhaps biology in general rather more like Turing's or D'Arcy Thompson's, then you'd want to allow that proving useful is not a condition of a biological entity having some structure or being some kind of thing.

  NC: Take D'Arcy Thompson. If biophysical laws determine the general shape of the properties of creatures, it doesn't say that you can't build submarines.

  24 Studies of mind and behavior and their limitations

  JM: I wanted to ask you some questions about social science, but I'm reluctant to switch the topic that much. Well, maybe we can.

  NC: It should be a short conversation [both laugh].

  JM: The social sciences, and many philosophical approaches to mind, take very seriously the idea of mind as essentially a causal mechanism that is dr
iven by some sort of belief-desire psychology. That raises questions about the status of this particular kind of enterprise. It's very tempting to think of it as an outgrowth of folk science, never breaking with common sense as the serious sciences have done – hence, not a serious science. However, there are people such as Hilary Putnam – in his functionalist days, at least – who simply adopted the framework of belief-desire psychology, presented it in functionalist terms, and claimed that it could be conceived of as a science. Could I get your views on the status of this kind of exercise?

  NC: Let's take something concrete; let's take some of the standard examples. I look out the window, I believe it's raining, I desire to stay dry, I take my umbrella. So my belief and my desire caused me to take my umbrella.

  I think that that's just a description of what I did. There's no independent notion of belief, desire, or cause that enters into this discussion. It's just a way of describing what we regard as rational action. If instead of taking my umbrella I take my clothes off, we say it's irrational. But there's no more or less notion of cause, and we don't even know that there are such entities as beliefs and desires. In fact, plenty of languages don't have those words. What you would say is, well, I think it is raining, and I want to stay dry, and so I'm going to take my umbrella. There's no [mental representation] belief, there's no [mental representation] desire; just [my saying] here's what I want, here's what I think, here's what I do. What I think and what I want are probably related in some way or another to what I do, but that's not a sufficient basis for a science.

  You get the feeling that it might be a science because you nominalize. If you talk about beliefs, then ok [you think], there must be some sort of system of beliefs, and we can try to say something about it, and so on. But maybe that's just the wrong way of looking at it. English happens to be a highly nominalizing language, so we're led down that path very easily. But it doesn't tell you that this is the right thing; in most languages you just can't say that kind of thing. You might claim that it's right to say these kinds of things, as it's right to talk about tensors, molecules, and so on. But you've got to show that. You can't just tell a story using “tensor” and “molecules” and then say, ok, we did it. You have to say what they are, and what the theoretical framework is in which you embed them, and so on. In these [belief-desire] cases, it's just not done.

  Well, could it be done? You could have some kind of empirical study of what people believe and why they believe it, and so on. It might turn out – if that kind of study gets anywhere – that you could develop theories that postulate entities that they call “beliefs” and place them in some appropriate framework. Then you could talk about the belief component of the system. And the same with the desire component, perhaps. But even if you could do that, it's not at all clear that that's the right way to go. You would still have the non-trivial problem of bringing in causality. Now we're back to Descartes's problem. Does it cause you to do this? No; it just says that you're acting rationally, whatever that means. You can also choose to act irrationally; to take Descartes's example, you can choose to put your finger in the flame.

  JM: Davidson in his “Mental Events” (1970) rejects belief-desire psychology as a science, but he wants to insist that action is caused – it's just that it's caused ‘physically.’ I was reminded of another article of his while you were speaking – “Psychology as Philosophy” (1980) in which he reports carrying out an experiment on a class of students at Stanford, and effectively showed that there is no possible way of measuring beliefs, desires, and the like. If there's no way of measuring them, there's no possible way of putting them into a theory.

  NC: I don't know the article; what's a sketch of the proof?

  JM: As I recall, he asked students to rank preference order for objects – some twenty or so – and then asked them to go over the list again and, pairwise, rank preference of one object over another: do they prefer a to b, b to c, and so on . . .

  NC: . . . and it doesn't come out consistent.

  JM: . . . it doesn't come out consistent.

  NC: There's a lot of work on that; for example, Tversky and Kahneman (1974) and Kahneman, Slovic, and Tversky (1982). There are a lot of strange phenomena. Massimo Piattelli-Palmarini (1994) has an interesting book in which he includes many paradoxical conclusions that people come up with. But that doesn't tell us that there's no belief system, just that they won't be consistent.

  JM: That's right. Davidson used this in an argument to show that psychology is essentially philosophy, not science, because . . .

  NC: It would say that a scientific psychology shows that people's belief systems are not consistent. That doesn't surprise me; I'm sure it's true of mine. And in fact, we know it's true. We do certain things because we feel, somehow, that that's what's happening, while some other part of our mind tells us that that's not what's happening. I can't imagine anyone not having that experience.

  This is interesting work because Kahneman and Tversky spell out what happens. In fact, they come up with criteria that show why certain kinds of questions will give one kind of ranking, while another kind of question will give a different ranking. You're looking at the questions in some different kind of framework.

  As for irrational beliefs: it's in front of us every moment. Take a look at the election that's coming this year [2004]. Probably large numbers of the people who are voting for Bush are doing it on the basis of irrational beliefs. These are poor working people who are getting completely shafted and have objective facts right in front of them that tell them that that's what's happening to them, and can with a trivial argument see that it's the result of these policies [of Bush]. But they nevertheless accept the view that this guy is standing up for us against the rich and powerful elitists. You couldn't have more obvious irrational beliefs.

  JM: Is that really what they believe? I would have thought that the Republicans had managed to press the ‘family values’ button very hard, and also somehow in addition managed to mobilize the jingoist attitudes . . .

  NC: Partly; but somehow, the end result is that many people – perhaps the majority – have the feeling that this guy is defending us against the liberal elitists . . .

  JM: If so, it's even more irrational than I already imagined . . .

  NC: But apparently that is the range of possible, testable attitudes. Maybe there are all kinds of reasons for it. But if you explore it a bit, it's clearly a system of extremely irrational beliefs. And that happens all the time.

  JM: Returning to Davidson very quickly . . . He in “Mental Events” used what he took to be the non-scientific character of our psychological explanations – basically, belief-desire folk psychology – and used this as the basis for claiming the anomalism of the mental, and this in turn to claim token (not type) mind-brain identity. We speak of causing actions and perceptions being caused, so mental events cause and are caused by physical events. Causation is deterministic. Science as he understood it seems to require some kind of deterministic, law-like principles . . .

  NC: Well, we don't know that the mind does not work by deterministic, law-like principles, we just know that the belief-desire type of description doesn't have those properties. You can say the same about the anomalism of the folk ‘physical.’ The ways we talk about the world – our intuitive understanding of the world in folk physical ways – is also not deterministic in the sense of Newtonian (or later) science.[C]

  JM: Yes. But now, in your view of science, not folk science, are we committed to determinism?

  NC: Well, if we are, that's a comment on our cognitive capacities. There's no external criterion that requires it.

  JM: So causality has the status of a Kantian regulative principle . . .?

  NC: . . . except that it's human-specific . . .

  JM: . . . human-specific, not rational beings in general – whatever they are?

  NC: Human-specific, unless we take ourselves to be the criterion for a rational being, in which case we end up being logical. If we h
ave some other concept of rationality, we may conclude that humans are inherently irrational.

  Suppose it's true. There's a lot of work now that suggests that religious beliefs of one kind or another – that's a broad category, belief in some supernatural force or whatever it may be – is inherent in human nature. You can imagine how that might be true. Take children; you can show experimentally that if there's something moving over here and something else moves over there in a systematic way unconnected to it, they will assume that there is some lever or something or other that's connecting them that they can't see. You look for mechanical causes; it's just in our nature. That's why it was so hard even for Newton himself to accept his law of gravitation; it's in our nature to look for mechanical causes. Well, you take a look at the world around you; the number of things that you can account for in terms of mechanical causes is infinitesimal, and if that is the way you have to look at things, you're going to have to look for some other cause. So you go to a supernatural cause, not a natural. So it could be that our cognitive capacities – I'm not suggesting it, but it could be – that the cognitive capacities of this creature will compel it in the actual world to conjure up supernatural forces. “There must be a mechanical causal law that explains this” happens to be an irrational belief. But it might be that we are just destined by our very nature to have that belief; that's just the way we are. I can certainly imagine a creature that worked like that, and it could well be us.

  JM: There are any number of examples of individuals – like myself – who don't accept those beliefs in supernatural causes, meaning by that ones that don't accord with science as we have developed it . . ..