Daniel Dennett: How Skyhooks Hoist Only Their Own Petards Interview by Larry Hardesty Featured in BBR December 1995 LH: I wanted to start off with a quote, because I thought it might describe the approach in Darwin's Dangerous Idea. "The correct method in philosophy would really be the following: to say nothing except what can be said, i.e. propositions of natural science...and then, whenever someone else wanted to say something metaphysical, to demonstrate to him that he had failed to give a meaning to certain signs in his propositions. [Wittgenstein, Tractatus Logico-Philosophicus]" DD: Well, that's old positivism, and if only life were that easy. No, I think that one of the big mistakes of philosophical thinking that came out of logical positivism was this silly idea that scientists are somehow immune to philosophical confusion themselves, that they can tell us the true facts about the things that are in their purview and we can just be scolds about the rest of it. Whereas, in fact what's really wonderful is that science itself is infested with interesting philosophical problems that really get in the way of the work of the scientists. And some of them are very dismissive of philosophical puzzlement about these-at their peril-and they end up saying all kinds of foolish things and wasting a lot of time on research that is ill-conceived. Others sort of see the point. Some of them are naturally adroit about philosophical issues, whether they read any philosophy or not. And what's nice right now is that more than a handful of philosophers who have done their homework in these areas are taken seriously by the scientific workers and are treated as equal partners, to same degree, in these research areas. I love to see that development, and I hate to see the events that threaten it, which jeopardize that understanding, which are all too frequent when philosophers mouth off before they know what the facts are, for instance. LH: Well, then, could we amend the Wittgenstein quote to say instead of, "when someone else wanted to say something metaphysical," "when the scientists themselves allow metaphysics to creep into their thinking?" DD: There's a sense of metaphysics where, yes, they should keep it out. But there's another sense, a more positive sense, in which indeed there's metaphysics in scientific thinking. After all, if you want to know about such grand metaphysical issues as "what is time," "what is cause," "what is effect," where do you look? You look in the sciences. That's because there really is some metaphysical content to science pure, not science done grinding a metaphysical axe, but just science itself. You know, Dick Rorty in that lovely review in Lingua Franca of Darwin's Dangerous Idea says that I'm showing that science can't answer certain sorts of philosophical questions. And that's right, but I would hate to see that interpreted as, "And so, of course, if we're interested in philosophical questions, we shouldn't be interested in scientific questions about those same topics." No, I would interpret Rorty to be saying we are right to be cautious, or even sceptical, or even suspicious when a scientist proclaims, "This is the metaphysical or philosophical import of my work." But it's not because the scientist's work doesn't have metaphysical or philosophical import. It's just that, the scientist may not be the best one to say what it is. LH: What particular training makes the philosopher better qualified? DD: I think that training in philosophy is almost a misnomer, in that it's not like training in any of the technical fields, where there are these fundamental techniques and a certain array of concepts you've mastered, and proved yourself in the lab, it isn't like that. It couldn't be. But probably, by seeing lots of plausible mistakes that have been made in the past, seeing why they're wrong, and why they're plausible, you become particularly sensitive to not making those mistakes again. The philosophical mistakes that are perennials are perennials for a very good reason: they are really tempting. And it's not easy to see what's wrong with them. Nobody's hit upon a decision procedure or a formula, or a formal way of protecting yourself against these. It seems that the only way you can do that is just get burned a lot and then learn not to be burned in those ways again. So that philosophical training is really exposure to lots of tempting bad theory, and learning how to say what's wrong with it. LH: It seems that both Consciousness Explained and Darwin's Dangerous Idea attack particular bad theories: first, Cartesian dualism and second, well, an aversion to natural selection. DD: Of course, dualism was not the target in Consciousness Explained, it was Cartesian materialism. Cartesian materialism is the view that we're left with when we've admitted, as just about everybody has, "Oh yes, Cartesian dualism, false, false, false." But they haven't thrown out enough. They've left a Cartesian way of thinking, and they've tried to simply plaster that onto modern materialism. And it doesn't work. LH: But it's still vestigially dualistic... DD: It is vestigially dualistic, but you see, you call someone who's doing that a dualist, and they know they're not a dualist, so they think you just haven't understood them. They say, "Pah! I give the back of my hand to all of that dualism stuff." But they still have an allegiance to something else which is Cartesian, which isn't dualism, but which is this idea that, "there's got to be a place where it all comes together." There's got to be a central boss, in an inner sanctum. And that's a tremendously attractive idea, which is quite independent of dualism. And the main point of that book, I suppose, is to show people that materialism is actually harder than you might have thought. You can't just get rid of the dualism part of Descartes' vision. You've got to adjust some other habits of thought as well. And then in Darwin's Dangerous Idea, I wanted to set up a framework in which the debates could be couched so you could see what the alternatives were, because I think that's been fuzzed over by a lot of the controversy. And so I present a rather stripped-down, clean Darwinist position. It is the position that I think is pretty much routinely accepted in the field. But it is also a position that is called hyper-Darwinist, or ultra-Darwinist by some. The trouble is that they very often caricature ultra-Darwinism. So, one wants to get behind that and say, "All right. Here's the view. This is the basic Darwinian framework." Now, what's wrong with it? I don't know that there's anything wrong. It's wonderful. Against that framework, then, we can look at objections, problems, and we can sort them into those that are really serious and those that are interesting but not revolutionary-they could go either way. Because a lot of people are very eager to see as watered-down a version of Darwinism as possible survive. LH: Francis Crick, in his book The Astonishing Hypothesis, faults you for a slight neuroanatomical oversight in Consciousness Explained. But as far as I can tell, that detail has no bearing on your theory whatsoever. DD: You can't win. I mean, you can, because you just don't worry about it. But a lot of people in the neuroscience business criticized Consciousness Explained for not going out on enough limbs, for being too neutral with regard to, say, neuroanatomical locations. You know, where do these multiple drafts things go? I deliberately didn't talk about the role of frontal or pre-frontal structures. I didn't talk about the role of the hippocampus. I didn't talk about thalamic pathways. I didn't for a very good reason: I didn't want to be hung for getting something factually wrong which was not the point. But whenever anyone does come along with good fillings for that I'm happy to go along. It's not that I'm saying there isn't such a story, it's just that I didn't want to tell it. So, for instance, Tony Damasio's new book [Descartes' Error] has some bold postulations about how these things happen, and he says at the end of his book that he thinks that his account of consciousness is not the same as mine. But it's consistent with mine, I think, and with a few friendly amendments, there's no reason I can't adopt, ultimately, some of his stuff. He's in a better position than I am to put forth those details, and I will be instructed. LH: So does that suggest a continuum between philosophy and science, and you're just working in different regions? Or is philosophy a more general, or theoretical, approach? DD: I think there is a sort of continuum. Crick, thank goodness, is ambitious, he doesn't just want to fill in blanks in a table of things we don't know yet about various parts of the brain. He wants to have a whole theory of consciousness. He wants to explain it all. Great! I want to see that happen. I'd love to play a big role in it. So in that regard we're both playing the same game, but with somewhat different tools. And of course, that didn't used to be the case. It used to be very bad form for scientists to hold forth on these topics unless they were retiring and gave their farewell lectures and talked about the soul or something. And I'm glad that now some pioneers have made it respectable to worry about these problems publicly and actually address them. LH: A number of well-regarded scientists are targets of criticism in Darwin's Dangerous Idea. You say, for example, that Roger Penrose's "exposition of [the idea that Gödel's Theorem proves AI impossible] is so clear that it amounts to exposure." DD: Yeah. Although I disagree so vigorously with a lot of what Roger stands for, I have tremendous respect for his intellectual honesty, his trying to get it clear; he just lays it right out there for you. And if it's wrong, he'll help you see why it's wrong. And that's really admirable. LH: And, in this case, how is he wrong? DD: The glaring problem in Penrose is simply that he attacks a doctrine of artificial intelligence that has never been held by artificial intelligence. For AI, we've always been looking for so-called heuristic programs for intelligence, and those are simply not covered by Gödel's theorem at all, so the criticism is just irrelevant. And I thought this had been pretty well realized by everybody in the field for twenty years, but somehow Roger didn't pick up on it. LH: Someone else whose views you find fault with, but whom you approach with great respect, is Chomsky. DD: Noam Chomsky is the founder of a very important scientific discipline and has had major, incisive things to say about every aspect of cognitive science, and linguistics, and, for that matter, philosophy. He's always taken philosophy very seriously. He's been a model of the scientist taking philosophy seriously. At the same time, he's had a polemic going against a certain way of viewing language and a certain way of viewing the mind, which is really the Darwinian way. And since he's so influential, that anti-Darwinian attitude has been tremendously widespread among people who read linguistics but haven't really studied evolution at all. And so, it was very important to me simply to present that fact so that people could realize that if they were getting their deep suspicion or outright dismissal of Darwinian thinking in linguistics and philosophy of mind from Chomsky, that wasn't a good source. However great he is on everything else, this is a bugbear of his that they should not take at face value. LH: Your charge against him-as against the philosophers Jerry Fodor and Colin McGinn, as well as Stephen Jay Gould-is that they're looking for "skyhooks." DD: I am hoping that "cranes" and "skyhooks" as a way of characterizing these issues will catch on because I think that it really does illuminate a lot. Ever since Darwin put forward his idea of natural selection, the quite natural, plausible, reasonable and responsible suspicion has been, "Gosh, all that work done in so little time. Only four billion years to do all that." My God, when you think of all the design that has happened, it's just too much work to have been done by brute, mechanical, mindless, purposeless, unforesightful processes. Well, that's the challenge: has there been time or hasn't there? The skeptics have tried to show that there were phenomena in nature, something-the human mind, language, the bird's wing, sex-that defied explanation from nothing but natural selection. What they wanted to have was a "you-can't-get-here-from-there" case. If they got one, they'd have a skyhook. That's something where there's lifting done in design space that had to be done from on high because you couldn't build up to it from below, from this purely mindless process. And again and again, people looking for such a skyhook have found cranes, that is, when they got clear about the issue, they saw how Darwin could do it after all. And so, one of the main themes for the book is that the skeptical search for skyhooks has been a very positive thing in a way, in that it's helped us discover a lot of neat cranes. By recursion, by building on top of itself, it's just amazing what it's been able to create. Some people, though, still just cannot bring themselves to believe that it's possible. And to me, the case that's in a way most amusing is human creativity. They think, "you just can't give a mechanistic, Darwinian explanation of human creativity." And I say, "Wait a minute. You're quite prepared to allow a Darwinian explanation of a nightingale, but not an ode to a nightingale?" You think that an ode to a nightingale is so much more wonderful than a real nightingale that it couldn't be explained by the same sort of process? Come on! What kind of hubris is that? LH: Another common objection to Darwinism, which Richard Dawkins mentions in his latest book [River Out of Eden, reviewed BBR, volume 2, issue 7], is that certain adaptations are too perfectly balanced to have evolved by half-measures. It seems that your account of intentionality and meaning are sometimes criticized in the same way. DD: I suppose the parallel is to attempt to contrast human reason at its most wonderful with the clunky, half-measures approximations of something like an artificial intelligence program. It's just inconceivable that by gradual, incremental improvement and building up you could ever achieve, by the artificial intelligence route, the straightforward intuitive genius of the former. Well, there's the suspicion, and Darwin had a lovely twist to his reply. In effect, Darwin said, "Well, if things were really that perfect, maybe you'd have a case." But, of course, the closer we look at nature, the more we see these curious copyings, these curious re-usings of things. This doesn't look like the work of an omnipotent artificer at all. This looks like...a bit of a hodge-podge, a bit of a kluge, as we would say in AI circles. And, of course, the same thing is true of the mind. However wonderful our minds are, we are subject to all sorts of embarrassing shortfalls. Even the brightest are susceptible to these, which begins to open up the perspective from which the human mind can look like a bunch of tricks after all. But good tricks. LH: That uniformity of process between the development of species and of meaning or intelligence suggests that the simple algorithm of natural selection is applicable across a wide range of media. DD: That's certainly the claim that I want to make. Same process continued by other means at a different time-scale. LH: Which brings us to one of Gould's purported revolutions: that evolution proceeds by selection at the level of the species, not the organism or the genome. Why are they mutually exclusive? DD: I don't know. It is obviously true that when you have a computer program running, it's the sum total of all the little bit swaps and computations that it's composed of. There's nothing extra going on. But of course, you'd be insane to try to explain lots of the features of that program at the electron level. If you want to know the difference between Microsoft Word and WordPerfect, the level at which you do that is a high one, not the level of the electronics, not even at the level of the machine language. You want to talk about high level structures. And the same is true of evolutionary theory, certainly, and the same is true of the social sciences in general-that there are patterns that are particularly visible to particular scientific approaches, and those are the right levels at which to explore those patterns. That's obvious if the question is, "Should we reduce economics to biology?" The answer is, well, of course not, because the regularities of economics are plenty robust on their own, and they're much more readily seen and explained at the level of economics than at the level of the metabolism of the creatures that are the agents of economics. And so, it's also obviously true that if you want to look at populations, and at diversity, and at geographic spread, you want to look at high level patterns, you don't just want to look at genomes. LH: Does that mean that genomes are not the ultimate source for those higher level variations? DD: Ultimate source, or ultimate and sole source would be too strong. Because, of course, while they are the ultimate repository-even there, that's a bit too strong, because one of the things that each generation bequeaths to the next generation, in addition to its genes, is the world as it's left it: the streets, and the dust in the corners, and the buildings, the debts, and the residual problems are also bequeathed. And that's as true in non-human species as in us. Birds leave their nesting sites for the next generation. LH: Darwin's Dangerous Idea moves, ultimately, into a discussion of ethics. What in your approach to the topic is particularly Darwinian, or informed by Darwinism? DD: The reason the book has the title it does is that I do think that Darwin's idea overturns some traditional notions, which have been seen, rightly or wrongly, to be foundational to ethics. Well, they aren't. Help! What are we going to do? We're going to have to replace them with something else. Older but wiser, we're going to have to replace myths that we've outgrown with other views. We do have to draw lines, so that we can distinguish between, say, contraception and infanticide. We have to draw lines, and we have to defend them. But if we thought that there was a way of just asking science to tell us where the boundary was, where life began, or something, science isn't going to give us that answer. And that's a very Darwinian result. We're going to have to defend that answer on terms that we develop ourselves. It's a very liberating idea, and I think this is what Nietzsche saw too. After Darwin, we couldn't be content with this kind of trickle-down theory of ethics anymore. We were going to have to be, in effect, existentialists, and create our own ethics, and figure out what our values were. And I think it's very important to expose and confront that side of Darwin, because I think that's what people see in there, dimly, and that's why there's so much anxiety about Darwinism. They see that, "Gosh, it shows that certain doctrines are really just myths." Well, yes, so now let's think about what we're going to put in place of those myths. LH: In the last section of the book, you give an example of moral reasoning in action: the selection of a fellowship recipient from a field of candidates far too large to be evaluated exhaustively. DD: I meant that example to show how, in a realistic case, this unprincipled truncation of exploration actually happens. We don't consider all things; we can't. We know better than to think we can, and so we muddle along. Here are the principles we seem to rely on. That structure is going to face us everywhere. We shouldn't have any false aspirations about what we can get from an ethical theory. Then maybe we can really make some progress. LH: Towards what? DD: Well, to take a term from John Rawls, reflective equilibrium. It would be nice to achieve. I think that's a realistic goal, just barely. But anything else would be terribly pessimistic. Larry Hardesty is Editorial Assistant at the Boston Book Review. &copy1996. Boston Book Review. All rights reserved.

BookWire Home | Search BookWire | BBR Home | BBR Fiction | Subscribe to the BBR! | Send Mail