Reposted from:
http://ase.tufts.edu/cogstud/papers/valencia.htm
Also see:
http://kenyonreview.org/blog/?p=679 (thanks to Norm at onegoodmove.org)

INTRODUCTION

Weaver birds create intricate nests; sculptors and other artists and artisans also create intricate, ingenious constructions out of similar materials. The products may look similar, and outwardly the creative processes that create those processes may look similar, but there are surely large and important differences between them. What are they, and how important are they? The weaverbird nestmaking is 'instinctual,' and 'controlled by the genes' some would say, but we know that this is a crude approximation of a more interesting truth, involving an intricate interplay between genetic variation, long-term developmental and environmental interaction and short-term environmental variation—in opportunities and materials accessible at the time of nest building. And on the side of the human creator, a similarly complex story must be told. Genes play some role surely (think of the likelihood of heritable differences in musical aptitude, for instance), but so do both long-term and short-term environmental interactions. The myth of the artist "blessed" by a spark of 'divine genius' is even cruder and more distorted than the myth of the birdnest as a simple product of a gene—as if it were a protein.

Our thinking about human creativity is pulled out of shape somewhat by a famous contrast introduced to the world by Darwin. One of his earliest—and most outraged—critics summed it up vividly:

In the theory with which we have to deal, Absolute Ignorance is the artificer; so that we may enunciate as the fundamental principle of the whole system, that, IN ORDER TO MAKE A PERFECT AND BEAUTIFUL MACHINE, IT IS NOT REQUISITE TO KNOW HOW TO MAKE IT. This proposition will be found, on careful examination, to express, in condensed form, the essential purport of the Theory, and to express in a few words all Mr. Darwin's meaning; who, by a strange inversion of reasoning, seems to think Absolute Ignorance fully qualified to take the place of Absolute Wisdom in all the achievements of creative skill. (MacKenzie1868)

Darwin's 'strange inversion of reasoning' promises—or threatens—to dissolve the Cartesian res cogitans as the wellspring of creativity, and then where will we be? Nowhere, it seems. It seems that if creativity gets 'reduced' to 'mere mechanism' we will be shown not to exist at all. Or, we will exist, but we won't be thinkers, we won't manifest genuine 'Wisdom in all the achievements of creative skill.' Whenever we zoom in on the act of creation, it seems we lose sight of it, the intelligence or genius replaced at the last instant by stupid machinery, an echo of Darwin's shocking substitution of Absolute Ignorance for Absolute Wisdom in the creation of the biosphere. Many people dislike Darwinism in their guts, and of all the ill-lit, murky reasons for antipathy to Darwinism, this one has always struck me as the deepest, but only in the sense of being the most entrenched, least accessible to rational criticism. There are thoughtful people who scoff at Creationism, dismiss dualism out of hand, pledge allegiance to academic humanism—and then get quite squirrelly when it is suggested that a Darwinian theory of creative intelligence might be in the cards, and might demonstrate that all the works of human genius can be understood in the end to be products of a cascade of generate-and-test procedures that are, at bottom, algorithmic, mindless. Absolute Ignorance? Artificial Intelligence? Fie on anybody who would thus put 'A' and 'I' together!

Besides, wouldn't a Darwinian theory of human creativity be covertly self-contradictory? The Darwinian mechanism of natural selection is famously mindless, purposeless, lacking all foresight and intention—the blind watchmaker (Dawkins 1986) If natural selection is 'the opposite' of God, a strange inversion of the traditional vision of creativity, then it must be 'the opposite' of us, too, since God is made in our image! Human creative endeavors are obviously both foresighted and purposeful, so, then, they are Darwinian processes. What could be more obvious?

But there is a tension, isn't there? A key part of Darwin's great revolution is that we are part of it. Human beings are just one species among many, fully biological, and hence capable of no miracles, restricted to the same sorts of processes and methods as the other species. Our creative processes are surely natural (not supernatural!), so in that bland sense they are as biological as the creative processes of the weaverbird and the beaver.

William Poundstone (1985) puts the inescapable challenge succinctly in terms of 'the old fantasy of a monkey typing Hamlet by accident.' He calculates that the chances of this happening are '1 in 50 multiplied by itself 150,000 times.'

In view of this, it may seem remarkable that anything as complex as a text of Hamlet exists. The observation that Hamlet was written by Shakespeare and not some random agency only transfers the problem. Shakespeare, like everything else in the world, must have arisen (ultimately) from a homogeneous early universe. Any way you look at it Hamlet is a product of that primeval chaos.'

CREDIT ASSIGNMENT FOR CREATIVITY

Where does all that Design come from? What processes could conceivably yield such improbable 'achievements of creative skill'? What Darwin saw is that Design is always both valuable and costly. It does not fall like manna from heaven, but must be accumulated the hard way, by time-consuming, energy-consuming processes of mindless search through 'primeval chaos', automatically preserving happy accidents when they occur. This broadband process of Research and Development is breathtakingly inefficient, but—this is Darwin's great insight—if the costly fruits of R and D can be thriftily conserved, copied, stolen, and re-used, they can be accumulated over time to yield 'the achievements of creative skill.' 'This principle of preservation, I have called, for the sake of brevity, Natural Selection.' (Darwin 1865)

There is no requirement in Darwin's vision that these R and D processes run everywhere and always at the same tempo, with the same (in-)efficiency. If we think of design work as lifting in Design Space (an extremely natural and oft-used metaphor, exploited in models of hill-climbing and peaks in adaptive landscapes, to name the most obvious and popular applications), then we can see that the gradualistic, frequently back-sliding, maximally inefficient basic search process can on important occasions yield new conditions that speed up the process, permitting faster, more effective local lifting ( Maynard Smith and Szathmary 1995). Call any such product of earlier R and D a crane, and distinguish it from what Darwinism says does not happen: skyhooks (Dennett 1995). Skyhooks, like manna from heaven, would be miracles, and if we posit a skyhook anywhere in our 'explanation' of creativity, we have in fact conceded defeat.

What, then, is a mind? The Darwinian answer is straightforward. A mind is a crane, made of cranes, made of cranes, a mechanism of not quite unimaginable complexity that can clamber through Design Space at a giddy—but not miraculously giddy—pace, thanks to all the earlier R and D, from all sources, that it exploits. What is the anti-Darwinian answer? It is perfectly expressed by one of the 20th century's great creative geniuses (though, like MacKenzie, he probably didn't mean by his words what I intend to mean by them).
Je ne cherche pas; je trouve.
—Pablo Picasso

Picasso purports to be a genius indeed, someone who does not need to engage in the menial work of trial and error, generate-and-test, R and D; he claims to be able to leap to the summits of the peaks—the excellent designs—in the vast reaches of Design Space without having to guide his trajectory (he searches not) by sidelong testing at any way stations. As an inspired bit of bragging, this is non pareil, but I don't believe it for a minute. And anyone who has strolled through an exhibit of Picasso drawings (as I recently did in Valencia, while attending the conference that led to this volume) looking at literally dozens of variations on a single theme, all signed--and sold—by the artist, will appreciate that whatever Picasso may have meant by his bon mot, he could not truly claim that he didn't engage in a time-consuming, energy-consuming exploration of neighborhoods in Design Space. At best he could claim that his own searches were so advanced, so efficient, that it didn't seem—to himself—to be design work at all. But then what did he have within him that made him such a great designer? A skyhook, or a superb collection of cranes? (I have been unable to discover the source of Picasso's claim, which is nicely balanced by a better known remark by a more down-to-earth creative genius, Thomas Edison: 'Genius is one per cent inspiration and ninety-nine per cent perspiration.' (1932))

We can now characterize a mutual suspicion between Darwinians and anti-Darwinians that distorts the empirical investigation of creativity. Darwinians suspect their opponents of hankering after a skyhook, a miraculous gift of genius whose powers have no decomposition into mechanical operations, however complex and informed by earlier processes of R and D. Anti-Darwinians suspect their opponents of hankering after an account of creative processes that so diminishes the Finder, the Author, the Creator, that it disappears, at best a mere temporary locus of mindless differential replication. We can make a little progress, I think, by building on Poundstone's example of the creation of the creator of Hamlet. Consider, then, a little thought experiment.

Suppose Dr. Frankenstein designs and constructs a monster, Spakesheare, that thereupon sits up and writes out a play, Spamlet. Who is the author of Spamlet? First, let's take note of what I claim to be irrelevant in this thought experiment. I haven't said whether Spakesheare is a robot, constructed out of metal and silicon chips, or, like the original Frankenstein's monster, constructed out of human tissues—or cells, or proteins, or amino acids, or carbon atoms. As long as the design work and the construction were carried out by Dr. Frankenstein, it makes no difference to the example what the materials are. It might well turn out that the only way to build a robot small enough and fast enough and energy-efficient enough to sit on a stool and type out a play is to construct it from artificial cells filled with beautifully crafted motor proteins and other carbon-based nanorobots. That is an interesting technical and scientific question, but not of concern here. For exactly the same reason, if Spakesheare is a metal-and-silicon robot, it may be allowed to be larger than a galaxy, if that's what it takes to get the requisite complication into its program—and we'll just have to repeal the speed limit for light for the sake of our thought experiment. These technical constraints are commonly declared to be off-limits in these thought experiments, so so be it. If Dr. Frankenstein chooses to make his AI robot out of proteins and the like, that's his business. If his robot is cross-fertile with normal human beings and hence capable of creating what is arguably a new species by giving birth to a child, that is fascinating, but what we will be concerned with is Spakesheare's purported brainchild, Spamlet. Back to our question: Who is the author of Spamlet?

In order to get a grip on this question, we have to look inside and see what happens in Spakesheare. At one extreme, we find inside a file (if Spakesheare is a robot with a computer memory) or a basically memorized version of Spamlet, all loaded and ready to run. In such an extreme case, Dr. Frankenstein is surely the author of Spamlet (unless we find there is a Ms. Shelley who is the author of Dr. Frankenstein!), using his intermediate creation, Spakesheare, as a mere storage-and-delivery device, a particularly fancy word processor. All the R and D work was done earlier, and copied to Spakesheare by one means or another. Now look at the other extreme, in which Dr. Frankenstein leaves most of the work to Spakesheare. The most realistic scenario would surely be that Spakesheare has been equipped by Dr. Frankenstein with a virtual past, a lifetime stock of pseudo-memories of experiences on which to draw while responding to its Frankenstein-installed obsessive desire to write a play. Among those pseudo-memories, we may suppose, are many evenings at the theater, or reading books, but also some unrequited loves, some shocking close calls, some shameful betrayals and the like. Now what happens? Perhaps some scrap of a 'human interest' story on the network news will be the catalyst that spurs Spakesheare into a frenzy of generate-and-test, ransacking its memory for useful tidbits and themes, transforming—transposing, morphing—what it finds, jiggling the pieces into temporary, hopeful structures that compete for completion, most of them dismantled by the corrosive processes of criticism that nevertheless expose useful bits now and then, and so forth, and all of this multi-leveled search would be somewhat guided by multi-level, internally generated evaluations, including evaluation of the evaluation . . . .of the evaluation functions as a response to evaluation of . . . the products of the ongoing searches.

Now if the amazing Dr. Frankenstein had actually anticipated all this activity down to its finest grain at the most turbulent and chaotic level, and had hand-designed Spakesheare's virtual past, and all its search machinery, to yield just this product, Spamlet, then Dr. Frankenstein would be, once again, the author of Spamlet, but also, in a word, God. Such Vast (not literally infinite, but Very much more than Astronomical—Dennett 1995, p109) foreknowledge would be simply miraculous. Restoring a smidgen of realism to our fantasy, we can consider a rather less extreme position and assume that Dr. Frankenstein was unable to foresee all this in detail, but rather delegated to Spakesheare most of the hard work of completing the trajectory in Design Space to one literary work or another, something to be determined by later R and D occurring within Spakesheare itself.

REAL ARTIFICIAL CREATORS

We have now arrived in the neighborhood of reality itself, for we already have actual examples of impressive artificial authors that vastly outstrip the foresight of their own creators. Nobody has yet created an artificial playwright worth serious attention, but an artificial chess player—IBM's Deep Blue—and an artificial composer—David Cope's EMI—have both achieved results that are, in some respects, equal to the best that human creative genius can muster.

Who beat Garry Kasparov, the reigning World Chess Champion? Not Murray Campbell or any of his IBM team. Deep Blue beat Kasparov. Deep Blue designs better chess games than any of them can design. None of them can author a winning game against Kasparov. Deep Blue can. Yes, but. Yes, but. I am sure many of you are tempted to insist at this point that when Deep Blue beats Kasparov at chess, its brute force search methods are entirely unlike the exploratory processes that Kasparov uses when he conjures up his chess moves. But that is simply not so—or at least it is not so in the only way that could make a difference to the context of this debate about the universality of the Darwinian perspective on creativity. Kasparov's brain is made of organic materials, and has an architecture importantly unlike that of Deep Blue, but it is still, so far as we know, a massively parallel search engine which has built up, over time, an outstanding array of heuristic pruning techniques that keep it from wasting time on unlikely branches. There is no doubt that the investment in R and D has a different profile in the two cases; Kasparov has methods of extracting good design principles from past games, so that he can recognize, and know enough to ignore, huge portions of the game space that Deep Blue must still patiently canvass seriatim. Kasparov's 'insight' dramatically changes the shape of the search he engages in, but it does not constitute 'an entirely different' means of creation. Whenever Deep Blue's exhaustive searches close off a type of avenue that it has some means of recognizing (a difficult, but not impossible task), it can re-use that R and D whenever it is appropriate, just as Kasparov does. Much of this analytical work has been done for Deep Blue by its designers, and given as an innate endowment, but Kasparov has likewise benefitted from hundreds of thousands of person-years of chess exploration transmitted to him by players, coaches and books. It is interesting in this regard to contemplate the suggestion recently made by Bobby Fischer, who proposes to restore the game of chess to its intended rational purity by requiring that the major pieces be randomly placed in the back row at the start of each game (random, but mirror image for black and white). This would instantly render the mountain of memorized openings almost entirely obsolete, for humans and machines alike, since only rarely would any of this lore come into play. One would be thrown back onto a reliance on fundamental principles; one would have to do more of the hard design work in real time—with the clock running. It is far from clear whether this change in rules would benefit human beings more than computers. It all depends on which type of chess player is relying most heavily on what is, in effect, rote memory—reliance with minimal comprehension on the R and D of earlier explorers.

The fact is that the search space for chess is too big for even Deep Blue to explore exhaustively in real time, so like Kasparov, it prunes its search trees by taking calculated risks, and like Kasparov, it often gets these risks pre-calculated. Both presumably do massive amounts of 'brute force' computation on their very different architectures. After all, what do neurons know about chess? Any work they do must be brute force work of one sort or another.

It may seem that I am begging the question in favor of a computational, AI approach by describing the work done by Kasparov's brain in this way, but the work has to be done somehow, and no other way of getting the work done has ever been articulated. It won't do to say that Kasparov uses 'insight' or 'intuition' since that just means that Kasparov himself has no privileged access, no insight, into how the good results come to him. So, since nobody knows how Kasparov's brain does it—least of all Kasparov—there is not yet any evidence at all to support the claim that Kasparov's means are 'entirely unlike' the means exploited by Deep Blue. One should remember this when tempted to insist that 'of course' Kasparov's methods are hugely different. What on earth could provoke one to go out on a limb like that? Wishful thinking? Fear?

But that's just chess, you say, not art. Chess is trivial compared to art (now that the world champion chess player is a computer). This is where David Cope's EMI comes into play (Cope, 2001 ; Dennett 2001 c). Cope set out to create a mere efficiency-enhancer, a composer's aid to help him over the blockades of composition any creator confronts, a high-tech extension of the traditional search vehicles (the piano, staff paper, the tape recorder, etc.). As EMI grew in competence, it promoted itself into a whole composer, incorporating more and more of the generate-and-test process. When EMI is fed music by Bach, it responds by generating musical compositions in the style of Bach. When given Mozart, or Schubert, or Puccini, or Scott Joplin, it readily analyzes their styles and composes new music in their styles, better pastiches than Cope himself—or almost any human composer—can compose. When fed music by two composers, it can promptly compose pieces that eerily unite their styles, and when fed, all at once (with no clearing of the palate, you might say) all these styles at once, it proceeds to write music based on the totality of its musical experience. The compositions that result can then also be fed back into it, over and over, along with whatever other music comes along in MIDI format, and the result is EMI's own 'personal' musical style, a style that candidly reveals its debts to the masters, while being an unquestionably idiosyncratic integration of all this 'experience.' EMI can now compose not just two-part inventions and art songs but whole symphonies—and has composed over a thousand, when last I heard. They are good enough to fool experts (composers and professors of music) and I can personally attest to the fact that an EMI-Puccini aria brought a lump to my throat—but then, I'm on a hair trigger when it comes to Puccini, and this was a good enough imitation to fool me. David Cope can no more claim to be the composer of EMI's symphonies and motets and art songs than Murray Campbell can claim to have beaten Kasparov in chess.

To a Darwinian, this new element in the cascade of cranes is simply the latest in a long history, and we should recognize that the boundary between authors and their artifacts should be just as penetrable as all the other boundaries in the cascade. When Richard Dawkins (1982) notes that the beaver's dam is as much a part of the beaver phenotype—its extended phenotype—as its teeth and its fur, he sets the stage for the further observation that the boundaries of a human author are exactly as amenable to extension. In fact, of course, we've known this for centuries, and have carpentered various semi-stable conventions for dealing with the products of Rubens, of Rubens' studio, of Rubens' various students. Wherever there can be a helping hand, we can raise the question of just who is helping whom, what is creator and what is creation. How should we deal with such questions? To the extent that anti-Darwinians simply want us to preserve some tradition of authorship, to have some rules of thumb for determining who or what shall receive the honor (or blame) that attends authorship, their desires can be acknowledged and met, one way or another (which doesn't necessarily mean we should meet them). To the extent that this is not enough for the anti-Darwinians, to the extent that they want to hold out for authors as an objective, metaphysically grounded, 'natural kind', they are looking for a skyhook.

DOES THE AUTHOR DISAPPEAR?

There is a persistent problem of imagination management in the debates surrounding this issue: people on both sides have a tendency to underestimate the resources of Darwinism, imagining simplistic alternatives that do not exhaust the space of possibilities. Darwinians are notoriously quick to find (or invent) differences in genetic fitness to go with every difference they observe, for instance. Meanwhile, anti-Darwinians, noting the huge distance between a beehive and the St. Matthew Passion as created objects, are apt to suppose that anybody who proposes to explain both creative processes with a single set of principles must be guilty of one reductionist fantasy or another: 'Bach had a gene for writing baroque counterpoint just like the bees' gene for forming wax hexagons' or 'Bach was just a mindless trial-and-error mutator and selector of the musical memes that already flourished in his cultural environment.' Both of these alternatives are nonsense, of course, but pointing out their flaws does nothing to support the idea that ('therefore') there must be irreducibly non-Darwinian principles at work in any account of Bach's creativity. In place of this dimly imagined chasm with 'Darwinian phenomena' on one side and 'non-Darwinian phenomena' on the other side, we need to learn to see the space between bee and Bach as populated with all manner of mixed cases, differing from their nearest neighbors in barely perceptible ways, replacing the chasm with a traversable gradient of non-minds, protominds, hemi-demi-semi minds, magpie minds, copycat minds, aping minds, clever-pastiche minds, 'path-finding' minds, 'ground-breaking' minds, and eventually, genius minds. And the individual minds, of each caliber, will themselves be composed of different sorts of parts, including, surely, some special-purpose 'modules' adapted to various new tricks and tasks, as well as a cascade of higher-order reflection devices, capable of generating ever more rarefied and delimited searches through pre-selected regions of the Vast space of possible designs.

It is important to recognize that genius is itself a product of natural selection and involves generate-and-test procedures all the way down. Once you have such a product, it is often no longer particularly perspicuous to view it solely as a cascade of generate-and-test processes. It often makes good sense to leap ahead on a narrative course, thinking of the agent as a self, with a variety of projects, goals, presuppositions, hopes, . . . . In short, it often makes good sense to adopt the intentional stance (Dennett, 1971, 1987) towards the whole complex product of evolutionary processes. This effectively brackets the largely unknown and unknowable mechanical microprocesses as well as the history that set them up, and puts them out of focus while highlighting the patterns of rational activity that those mechanical microprocesses track so closely. This tactic makes especially good sense to the creator himself or herself, who must learn not to be oppressed by the revelation that on close inspection, even on close introspection, a genius dissolves into a pack rat, which dissolves in turn into a collection of trial-and-error processes over which nobody has ultimate control.

Does this realization amount to a loss—an elimination—of selfhood, of genius, of creativity? Those who are closest to the issue—the artistic and scientific geniuses who have reflected on it—often confront this discovery with equanimity. Mozart (in an oft-quoted but possibly spurious passage--see Dennett 1995, p346-7) is reputed to have said of his best musical ideas: 'Whence and how do they come? I don't know and I have nothing to do with it.' The painter Philip Guston is equally unperturbed by this evaporation of visible self when the creative juices start flowing:

When I first come into the studio to work, there is this noisy crowd which follows me there; it includes all of the important painters in history, all of my contemporaries, all the art critics, etc. As I become involved in the work, one by one, they all leave. If I'm lucky, every one of them will disappear. If I'm really lucky, I will too.

Resistance to extending Darwinian thinking into human creativity and human culture is not restricted to closet Creationists and anti-scientific humanists. Two highly visible Darwinian spokespersons—Stephen Jay Gould and Steven Pinker—who agree on precious little else, find common ground in their doubts about this:

I am convinced that comparisons between biological evolution and human cultural or technological change have done vastly more harm than good--and examples abound of this most common of intellectual traps. . . . . Biological evolution is powered by natural selection, cultural evolution by a different set of principles that I understand but dimly.
(Gould 1991, p.63.)

To say that cultural evolution is Lamarckian is to confess that one has no idea how it works. The striking features of cultural products, namely their ingenuity, beauty, and truth (analogous to organisms' complex adaptive design), come from the mental computations that 'direct'--that is, invent--the 'mutations,' and that 'acquire'--that is, understand--the 'characteristics.' (Pinker 1997, p209)

Pinker has imputed the wrong parallel; it is not Lamarck's model, but Darwin's models of unconscious and methodical (artificial) selection (as special cases of natural selection) that accommodate the phenomena he draws to our attention in this passage (Dennett 2001b). And it is ironic that Pinker overlooks this, since the cultural phenomena he himself has highlighted as examples of evolution-designed systems, linguistic phenomena, are almost certainly not the products of foresightful, ingenious, deliberate human invention. Some designed features of human languages are no doubt genetically transmitted, but many others--such as changes in pronunciation, for instance--are surely culturally transmitted, and hence products of cultural, not genetic, evolution.

CONCLUSION

The cranes of human culture didn't just open up Design Space; they opened up perspectives on Design Space that permitted 'directed' mutation, foresighted mutation, reflective mutation, both in cultural and, most recently, genetic innovation. This nesting of different processes of natural selection now has a new member: genetic engineering. How does it differ from the methodical selection of Darwin's day? It is less dependent on the pre-existing variation in the gene pool, and proceeds more directly to new candidate genomes, with less overt trial and error. Darwin (1865, p38) had noted that in his day, Man can hardly select, or only with much difficulty, any deviation of structure excepting such as is externally visible; and indeed he rarely cares for what is internal. But today's genetic engineers have carried their insight into the molecular innards of the organisms they are trying to create. There is ever more accurate foresight, but even here, if we look closely at the practices in the laboratory, we will find a large measure of exploratory trial and error in their search of the best combinations of genes. (In fact, biochemists and molecular biologists are finding that artificial evolutionary processes are more efficient R and D procedures than their foresightful hand-work efforts by orders of magnitude. In other words, they are finding that the breeding of domesticated micro-organisms and polymers is the best way to conduct their creative searches.)

Are the products of genetic engineering 'Darwinian' products? They are produced not by blind or random trial-and-error variation, but by highly intelligent, guided, foresightful processes. Nevertheless these processes are themselves the products of earlier design work accomplished by Darwinian R and D, and if we look closely at the microprocesses that compose their current, local search, we will still find plenty of instances of random (undesigned, chaotic) generation of candidates for further scrutiny.

It may seem, however, that we have now passed the Pickwickian limits of Darwinian orthodoxy. Does a Darwinian gloss actually supplement or adjust the traditional intellectualist ways of thinking? I think it does, because without the steady pressure of the Darwinian 'strange inversion of reasoning,' it is all too tempting to revert to the old essentialist, Cartesian perspectives. For instance, there is always the temptation, often succumbed to, to establish 'principled' boundaries, or to erect a polar contrast between insightful and blind processes of search, as we saw in the unsupportable assertion that Kasparov's methods are fundamentally unlike Deep Blue's. If Deep Blue's methods are ultimately 'blind and mechanical,' then so, ultimately, are Kasparov's—his neurons are as blind and mechanical as any circuit board. The foresighted, purposeful breeding of domesticated plants and animals is obviously not a damning counterexample to Darwin's theory of natural selection as a foresightless, purposeless process, because his theory shows (as we are beginning to learn) how such foresight and purpose could itself evolve by blind natural selection. Kasparov's creative genius (or Bach's or Shakespeare's) is for the same reason no counterexample to the Darwinian theory of creativity, but rather one of the most recent blooms on the tree of life that we still need to account for in Darwinian terms.

FURTHER READING

Campbell, D.T. 1960. Blind variation and selective retention in creative thought as in
other knowledge processes. Psychological Review 67: 380—400.
Mithen, S. J. 1998, ed. Creativity in human evolution and prehistory. London; New York: Routledge, 1998.
Mithen, S. J. 1996. The Prehistory of the Mind: The Cognitive Origins of Art, Religion and Science, Cambridge: Cambridge University Press.
Toulmin, Stephen. 1972. Human Understanding. Volume 1: The Collective Use and Development of Concepts. Oxford: Clarendon Press

ACKNOWLEDGEMENTS
Portions of this paper are drawn from Dennett, 2001a. I have been unable to locate the source of Philip Guston's quote, but I have found much the same remark attributed to the composer, John Cage, a close friend and contemporary of Guston's, who [is said to have] said this about painting: 'When you are working, everybody is in your studio‑the past, your friends, the art world, and above all, your own ideas‑all are there. But as you continue painting, they start leaving, one by one, and you are left completely alone. Then, if you are lucky, even you leave.' Like all other creators, Guston and I like to re-use what we find, adding a few touches from time to time.

REFERENCES

Cope, D., 2001, Virtual Music: Computer Synthesis of Musical Style. Cambridge, MA: MIT Press.
Darwin, C. 1859. On the Origin of Species by Means of Natural Selection. London: Murray. (page numbers as in Harvard Univ. Press facsimile edn.)
Dawkins, R. 1982. The Extended Phenotype. Oxford and San Francisco: Freeman.
Dawkins, R. 1986. The Blind Watchmaker, London: Longmans.
Dennett, D. C. 1971, Intentional Systems, Journal of Philosophy, 68, pp87‑106.
Dennett, D. C. 1987, The Intentional Stance, MIT Press, Cambridge, MA.
Dennett, D. C. 1995. Darwin's Dangerous Idea. New York: Simon & Schuster.
Dennett, D. C. 2001a. In Darwin's Wake, Where am I? (Presidential Address, Eastern Division, American Philosophical Association, December 29, 2000) in Proceedings of the APA, and also in Cambridge Companion to Darwin, ed. Radick et al., Cambridge: Cambridge Univ. Press.
Dennett, D. C. 2001b, The Evolution of Culture, in The Monist 84, no. 3, Peru, Illinois (special issue on cultural evolution, ed. D. Sperber), pp 305‑324.
Dennett, D. 2001c, Collision Detection, Muselot, and Scribble: Some Reflections on Creativity, in Cope 2001, pp283-291.
Edison, T. 1932, interview in Life, ch. 24 (according to the Oxford Dictionary of Quotations).
Gould, S. J. 1991. Bully for Brontosaurus, New York: Norton.
Mackenzie, R. B. 1868. The Darwinian Theory of the Transmutation of Species Examined. (published anonymously 'By a Graduate of the University of Cambridge') London: Nisbet & Co. (Quoted in a review, Athenaeum, no 2102. Feb 8, p217.).

Maynard Smith, J. and Szathmary, E. 1995. The Major Transitions in Evolution. Oxford: Freeman.
Pinker, S., 1997. How the Mind Works. New York: Norton.
Poundstone, W. 1985. The Recursive Universe: Cosmic Complexity and the Limits of Scientific Knowledge. New York: Morrow.