The Science of Language

by Noam Chomsky

  These questions were throughout history typically discussed as metaphysical questions – so, is it the same ship, or thing in the world? Well, the ship is not a thing in the world to start with; it's a mental construction that, of course, relates to things in the world when we so use it. The thing that you're discussing is the mental construction. And that mental construction has some answers, but not other answers. And those answers don't really tell you a lot about metaphysics. They just tell us that that's the way we think about the world. We think about the world as involving certain kinds of continuities. They mean nothing to the physicist. As far as the physicist is concerned, if you take out a nail, it's a new ship.

  Well, the physicist's way, that's not the way we think about it. All this shows that we're just inquiring into our own internal conceptions, what we call common sense. It has a relation to the world, of course, but it's not the same thing. If these questions – what's the ship of Theseus, what's a person, what's a tree, and so on – if they're re-interpreted as they should be, cognitively, epistemologically rather than metaphysically, well, then, they can be explored as topics of cognitive science. What's the nature of our conceptual systems? And then we discover that it's true that persons have responsibilities and obligations, and trees don't. That's not because we made a discovery about the world. It's because that's the way we think about the world. Persons, apart from infants and the mentally ill, and so on, have responsibilities, deserve praise and censure, and so on, and a dog doesn't, unless you personify it. Is it because we've discovered something metaphysical? Well, there may be something metaphysical underlying that – in fact, there probably is – but that's not what we discover. We discover that this is the way we look at creatures. And in fact, if you go back in pretty recent history – say, again, the seventeenth or eighteenth century, when a lot of these topics really became alive – people are really confused about whether we regard, say, orangutans and negroes as persons. They seemed more or less the same, and they weren't like us; they're creatures of some kind. Were any of them persons? Were all of them persons? You get huge debates about all this because the internal concept PERSON just isn't going to carry you very far in figuring out how the world works – just as [Descartes's] contact mechanics carried you pretty far, but not past Newton, who showed you can't figure out how physical objects [as understood by contact mechanics] work.

  This goes right on into contemporary philosophy. A crucial, exciting topic in contemporary philosophy is [still] Kripkean essentialism – Putnam's version – which is based on questions like is water H2O? Well, it's like your intuitions about the ship of Theseus. You have whatever you have. It's not going to tell you anything about H2O (the stuff described by scientists) any more than talk about the ship of Theseus is going to tell you about ships from a physicist's point of view. It's telling you about how you look at and interpret the world. And these discussions are particularly odd, for one thing, because the alleged intuitions are mostly inside a philosophical cocoon. People have to be trained to have those intuitions by taking graduate courses in philosophy. And also it's very unclear what the intuitions are even supposed to be about. So take the sentence, water is H2O; that's the core sentence of the whole discussion. We've all learned – or maybe we've been taught – that a sentence has no meaning outside a language. So what language is this sentence in? It's got to be in some language. Well, it's not English. English doesn't have [the concept] H2O in it; that's an invented concept that you bring into English. It's not a sentence of physics, or chemistry, because they don't have the concept WATER. It's true that when a chemist writes an article, he or she will use the word water, but he or she's using informal discourse. You're not doing everything in a precise formalism, even if you're doing mathematics. But chemistry has no concept WATER; it has a concept of H2O, and you can informally call it “water,” if you like, but WATER is not a concept of chemistry. So the sentence is not chemistry, it's not English, it's not French, it's not German; in fact, it's no language at all. It's some amalgam of languages – or rather, symbolic systems and languages – that we pick up. But we can't have any intuitions about things like that. It's meaningless. It's like having intuitions about quantum physics. So far as I can see, the entire discussion on all sides is basically vacuous. And that's a primary theme in contemporary analytic philosophy. It's just not about anything.[C]

  JM: . . . You don't have to convince me of that. Returning to your characterization of cognitive science – a very nice characterization, I think, quite unlike Jerry Fodor's, where cognitive science is essentially representation-of-the-world – as an (internalist) investigation of our cognitive structures. Granted we can investigate conceptual structures where we're talking about concepts such as PERSON, but what about the concepts that appear in the sciences? Can you investigate them [as a part of an internalist cognitive science]?

  NC: Sure. You can investigate [the concept] H2O. We don't know the answers. We know how to do it. We try to place it in an explanatory framework that we make as precise as circumstances require. There's no point in formalizing beyond the level of understanding that we have. And then we work within that. That's the way the sciences have always been.

  Take, say, mathematics, the clearest case. We all know very well that up until the mid-nineteenth century, when a large part of the great mathematics was done, they didn't even define their concepts. No one had a clear notion of limit. Gauss was proving all these magnificent theorems where a limit just meant ‘getting closer and closer to.’ In fact, Berkeley was finding contradictions in Newton's proofs – a topic that English mathematicians took seriously, standing in the way of progress. One line of the proof treats zero as zero, and three lines later, it treats zero as something that you can get as close to as you want. Those are different concepts, so it's equivocation; the proof is based on an equivocation – it doesn't show anything. Well, on the continent, mathematicians knew it and didn't pay much attention to it and went ahead developing rich mathematics. Finally mathematics got to the point where you just had to understand what a limit is; you couldn't get along any more on these intuitive conceptions. So you get delta-epsilon definitions, Weirstrauss, Cauchy, and so on. OK, at that point, you know what a limit is. But you know it because it's been made explicit. And so it continues. Euclid in a sense had a real geometry. But it wasn't formalized until thousands of years later.

  JM: But then to investigate the concepts of science, in effect, you learn the science.

  NC: You learn the science, and you try to get as close as you can to a proper kind of scientific theory, whatever that means. We have all sorts of intuitive criteria that we use all the time to decide whether this is or is not a sensible scientific explanation.

  JM: Is there a science of scientific concepts in the way in which there is – or will be – a science of commonsense concepts?

  NC: I think that there should be. Here is where I tend to disagree with the line of thinking that Sue Carey (1987, 2009) is developing, although she could be right; I don't really know. Her basic position is that there's nothing new to say about how science is done; it's just more of the same. It's a more sophisticated version of what children are doing when they learn how to build houses out of blocks. Maybe; but my guess is that it's something quite different. There is a science-forming capacity that is – to some extent – put to use throughout human history when people make up mythological stories about creation or engage in magic – the transition between magic and science is not so clear. But it takes on a very different form in the modern period when it becomes a very self-conscious endeavor, trying to establish both empirical and epistemological criteria it's supposed to meet. It may change; it's not fixed. But at least it's being pursued as a systematic effort to gain a certain kind of explanation and insight. You can't just tell stories about something; you have to show that those stories have some substance. That's why so much talk about evolution is basically uninteresting; it's just stories. It could have happened that way; it could have
happened twenty different ways. You don't know how to formulate the question in such a way that you could answer it. That's storytelling – in the framework of scientific ideas, but still storytelling. If you're serious about it, you try to prove it. Instead of just concocting stories, try to figure out ways in which you can study it [and get evidence for it]. And it's not so simple.

  21 Philosophers and their roles

  JM: I'd like to better understand your view of what might – this is a question that is partly driven by what a graduate student asked me to ask you – of what you think a philosopher could plausibly contribute now. It seems that some philosophers – philosophers after Descartes's and Hume's time – have been behind the times. They have not fully comprehended how advanced the sciences [and in particular, linguistics] are . . .

  NC: There are some philosophers who know the sciences very well, and who have contributed to [the sciences]. They don't question the sciences; they try to clarify what they are doing and even contribute to them at some conceptual level. That's pretty much what Descartes and Kant and others did who were called philosophers. You can be connected to the sciences and know them extremely well. Take someone like Jim Higginbotham. He knows linguistics very well, and contributes to it.

  JM: Indeed . . .

  NC: and is doing it not the way technical linguists do, but with philosophical interests that relate to traditional questions of philosophy, and so on. I think that that's always a possibility. I suspect that John Austin was right when he said that philosophy should be the mother of the sciences. It's clearing away the thickets and the underbrush and trying to set things up in such a way that the sciences can take over.

  JM: So the job of philosophers is to beat around in the bushes and see if they can scare up any birds . . .?

  NC: Not only in the sciences, but in people's lives. . . . Take for example [John] Rawls[, the political philosopher]. He's not working in the sciences. He's trying to figure out what concepts of justice we have that underlie our moral systems, and so on. And it does verge on the sciences. So when John Mikhail [who has a degree in philosophy but is also developing a science of a moral faculty that distinguishes permissible from impermissible actions] picks it up, it becomes a science.

  JM: OK, that's a plausible suggestion. I wonder if there aren't other issues here that bear on the question of the nature of the task that these people conceive that they are undertaking. Take, for example, Jim Higginbotham. He's very much taken by Fregean notions of minds and thoughts and seems to want to cash out what he contributes in Fregean terms. And I'm not sure that I would take that part as a particularly useful contribution . . .

  NC: Which part of what he's doing? What do you have in mind?

  JM: Well, he at least used to in many cases talk about propositions and thoughts, and the like, and gave these a very Fregean kind of reading. The idea wasn't that these were somehow biological entities; at least, that didn't come out in what I've read of his.

  NC: His view, if I understand it correctly, is what he calls “weak conceptualism” – that these entities are independent of, but reflections of, internal mental events. But then the question comes, well, what function are they performing? If they're just a one-to-one image of what's inside, why not dispense with them? We could say the same about chemistry. There are the elements, and compounds of them; and then there are images of them in some Platonic universe that we could study, if we wanted. Unless they have some other properties that are not determined by the internal events that they are reflections of, then they're dispensable.

  JM: But wouldn't it then be at least useful for philosophers to reconceive of themselves as engaged on an internalist project as opposed to the kind of externalist one that most of them imagine that they are engaged upon?

  NC: Well, can you make up a sensible externalist project? There certainly are externalist projects – when you and I are talking, it's not just what is going on in your head and what's going on in my head; we're interacting. So the study of how parts of the world interact, depending on their internal natures and lots of other things – that's a topic, but I don't understand why it's a particular topic for philosophers. That's just another topic. Maybe philosophers have something interesting to say about it; that's fine. There's interesting work on pragmatics. But what doesn't seem to exist, so far as I can see, is externalist semantics. Did you take a look at Tyler Burge's latest book?[C]

  JM: No, I haven't . . .

  NC: You should; this comes up. There are a lot of critical essays on his work; one of them is mine [see “Internalist Exploration,” in Chomsky 2000]. What I wrote is mainly on the externalism, and he has interesting responses (Hahn & Ramberg 2003). You can see if you can make something of them. I don't personally think that they come to anything. He's an intelligent person trying to engage with the issues; most philosophers don't even engage with them; and – of its type – it's as good a job as I know of. Paul Pietroski is writing about these things now.

  JM: Yes, he is; I admire his work. I particularly like some of his contributions to semantics (Pietroski 2005); a good internalist understanding of it too . . .

  NC: and a good critical analysis of what's going on in the field, which I think is rare.

  22 Biophysical limitations on understanding

  JM: Incidentally, your LSA paper and the emphasis on the third factor threw a bit of a monkey wrench into my efforts to write a chapter on innateness as a contribution to a book on cognitive science . . .1

  NC: Well, you just don't know . . . The more you can attribute to the third factor – which is the way that science ought to go; the goal of any serious scientist interested in this is to see how much of the complexity of an organism you can explain in terms of general properties of the world. That's almost like the nature of science. Insofar as there is a residue, you have to attribute it to some specific genetic encoding; and then you've got to worry about where that came from. Obviously, there's got to be something there; we're not all amoebas. Something has got to be there; so, what is it?

  JM: It might be nice to have answers.

  NC: I'm not sure; I like the edges of the puzzle.

  JM: OK, you're right. They're much more fun.

  NC: Think of how boring the world would be if we knew everything we can know, and even knew that we can't understand the rest.

  JM: Yes, Peirce's millennial form of science does sound boring.

  NC: Well, the nice thing about it is that his view can't be true, because he was making a serious error about evolution – assuming that we're basically angels by natural selection. But you could have something like it. You could imagine that the species would reach the point that everything knowable is known, including the limits of knowledge. So you could know that there are puzzles out there that can't be formulated. That would be ultimate boring.

  JM: Yes, worse than heaven.[C]

  1 The chapter that eventully (2005) took into account third factor considerations in what it had to say about the innateness of language.

  23 Epistemology and biological limits

  JM: You've suggested many times that human cognitive capacities have limitations; they must have, because they're biologically based. You've also suggested that one could investigate those limitations.

  NC: in principle.

  JM: . . . in principle. Unlike Kant, you're not going to simply exclude that kind of study. He seems to have thought that it's beyond the capacity of human beings to define the limits . . .

  NC: . . . well, it might be beyond a human capacity; but that's just another empirical statement about limitations, like the statement that I can't see ultraviolet light, that it's beyond my capacity.

  JM: OK; but is the investigation of our cognitive limitations in effect an investigation of the concepts that we have?

  NC: Well, it may be contradictory, but I don't see any internal contradiction in the idea that we can investigate the nature of our science-forming capacities and discover something about their scope and limits
. There's no internal contradiction in that program; whether we can carry it out or not is another question.

  JM: And common sense has its limitations too.

  NC: Unless we're angels. Either we're angels or we're organic creatures. If we're organic creatures, every capacity is going to have its scope and limits. That's the nature of the organic world. You ask “Can we ever find the truth in science?” – well, we've run into this question. Peirce, for example, thought that truth is just the limit that science reaches. That's not a good definition of truth. If our cognitive capacities are organic entities, which I take for granted they are, there is some limit they'll reach; but we have no confidence that that's the truth about the world. It may be a part of the truth; but maybe some Martian with different cognitive capacities is laughing at us and asking why we're going off in this false direction all the time. And the Martian might be right.

  JM: . . . assuming that the Martian could understand our cognitive capacities.

  NC: . . . right.

  JM: The project of investigating the limits of our cognitive capacities seems to me to be quite different from the kinds of projects that philosophers are fond of, or have been fond of, when they have introduced various kinds of epistemic constraints on what counts as meaningful or sensible or whatnot. Investigating the limits is a scientific project, not what too often amounts to a stipulative one.