Darwin's Dangerous Idea (38 page)

copying—with occasional mutations—nothing
intrinsically
204 BIOLOGY IS ENGINEERING

Original Sin and the Birth of Meaning
205

counts as a canonical version of anything. That is, although we can identify never be pulled apart, that there could never be meaning where there was no mutations by simply comparing the "before" sequence with the "after"

mind, or mind where there was no meaning.
Intentionality
is the phi-sequence, there is no intrinsic way of telling which of the uncorrected losopher's technical term for this meaning; it is the "aboutness" that can relate typographical
errors
might more fruitfully be viewed as editorial
improve-one thing to another—a name to its bearer, an alarm call to the danger that
ments.9
Most mutations are what engineers would call "don't-cares," vari-triggered it, a word to its referent, a thought to its object.11 Only some things ations that make no discernible difference to viability, but as selection in the universe manifest intentionality. A book or a painting can be about a gradually takes its toll, the better versions begin to cluster. It is only relative mountain, but a mountain itself is not about anything. A map or a sign or a to a "wild type" (a center of gravity, in effect, of such a cluster) that we can dream or a song can be about Paris, but Paris is not about anything.

identify a particular version as a mistaken version, and even then there is the Intentionality is widely regarded by philosophers as
the
mark of the mental.

possibility, remote in practice but omnipresent in principle, that what seems a Where does intentionality come from? It comes from minds, of course.

mistake from the perspective of one wild type is a brilliant improvement from But that idea, perfectly good in its own way, becomes a source of mystery the perspective of a wild-type-in-the-making. And as new wild types emerge and confusion when it is used as a metaphysical principle, rather than a fact as the foci or summits of fitness landscapes, the direction of the steady of recent natural history. Aristotle called God the Unmoved Mover, the pressure of error
correction
can reverse in any particular neighborhood of source of all motion in the universe, and Locke's version of Aristotelian Design Space. Once a particular family of similar texts is no longer subject to doctrine, as we have seen, identifies this God as Mind, turning the Unmoved

"correction" relative to a receding or lapsed norm, it is free to wander into the Mover into the Unmeant Meaner, the source of all Intentionality. Locke took attractive basin of a new norm.10 Reproductive isolation is thus both a cause himself to be proving deductively what the tradition already took to be and an effect of the clumpiness of phenotypic space. Wherever there are obvious: original intentionality springs from the Mind of God; we are God's competing error-correcting regimes, one regime or the other will win out, and creatures, and derive our intentionality from Him.

hence the isthmus between the competitors will tend to dissolve, leaving Darwin turned this doctrine upside down: intentionality doesn't come from empty space between occupied zones of Design Space. Thus, just as norms of on high; it percolates up from below, from the initially mindless and pointless pronunciation and word use reinforce clustering in speech communities (a algorithmic processes that gradually acquire meaning and intelligence as they theoretically important point made by Quine 1960 in his discussion of error develop. And, perfectly following the pattern of all Darwinian thinking, we and the emergence of norms in language ), so norms of genomic expression see that the first meaning is not full-fledged meaning; it certainly fails to are the ultimate basis of speci-ation.

manifest all the "essential" properties of
real
meaning (whatever you may Through the same molecular-level microscope we see the birth of
meaning,
take those properties to be). It is mere quasi-meaning, or semi-semantics. It is in the acquisition of "semantics'' by the nucleotide sequences, which at first what John Searle (1980,1985,1992 ) has disparaged as mere
"as if
are mere syntactic objects. This is a crucial step in the Darwinian campaign intentionality" as opposed to what he calls "Original Intentionality." But you to overthrow John Locke's Mind-first vision of the cosmos. Philosophers have to start somewhere, and the fact that the first step in the right direction is commonly agree, for good reason, that meaning and mind can just barely discernible as a step towards meaning at all is just what we should expect.

There are two paths to intentionality. The Darwinian path is diachronic, or historical, and concerns the gradual accretion, over billions of years, of the 9. Note the parallel here with my discussion of the false dichotomy between Orwellian sorts of Design—of functionality and purposiveness—that can support an and Stalinesque models of consciousness in
Consciousness Explained
(1991a). In that case as well, there is no
intrinsic
mark of the canonical.

intentional interpretation of the activities of organisms (the "doings" of

"agents"). Before intentionality can be fully
fledged,
it must go through its 10. Once again we see the tolerance for topsy-turvy imagery. Some theorists speak of awkward, ugly period of featherless pseudo-intentionality. The synchronic
basins of attraction,
guided by the metaphor of balls rolling blindly
downhill
to the local
minimum
instead of climbing blindly uphill to the local
maximum.
Just turn an adaptive landscape inside out and the mountains become basins, the ridges become canyons, and

"gravity" provides the analogue of selection pressure. It doesn't make any difference 11. The topic of intentionality has been written about extensively by philosophers of whether you choose "up" or "down" as the favored direction, just so long as you are many different traditions in recent years. For an overview and a general definition, see my consistent. Here I have slipped, momentarily, into the rival perspective, just to make this article "Intentionality" (co-authored with John Haugeland) in Gregory 1987. For more point.

detailed analyses, see my earlier books ( 1969, 1978, 1987b).

206 BIOLOGY IS ENGINEERING

The Computer That Learned to Play Checkers
207

path is the path of Artificial Intelligence, in an organism with genuine in-Darwinian thinking and Artificial Intelligence. Together they strike a funda-tentionality—such as yourself—there are, right now, many parts, and some of mental blow at the last refuge to which people have retreated in the face of these parts exhibit a sort of semi-intentionality, or mere
as if
intention-ality, the Copernican Revolution: the mind as an inner sanctum that science cannot or pseudo-intentionality—call it what you like—and your own genuine, fully reach. (See Mazlish 1993) It is a long and winding road from molecules to fledged intentionality is in fact the product (with no further miracle minds, with many diverting spectacles along the way—and we will tarry over ingredients) of the activities of all the semi-minded and mindless bits that the most interesting of these in subsequent chapters—but now is the time to make you up (this is the central thesis defended in Dennett 1987b, 1991a).

look more closely than usual at the Darwinian beginnings of Artificial That is what a mind
is
—not a miracle-machine, but a huge, semi-designed, Intelligence.

self-redesigning amalgam of smaller machines, each with its own design history, each playing its own role in the "economy of the soul." (Plato was right, as usual, when he saw a deep analogy between a republic and a 5. THE COMPUTER THAT LEARNED TO PLAY CHECKERS

person—but of course he had much too simple a vision of what this might mean.)

Alan Turing and John von Neumann were two of the greatest scientists of the There is a deep affinity between the synchronic and diachronic paths to century. If anybody could be said to have invented the computer, they did, intentionality. One way of dramatizing it is to parody an ancient anti-and their brainchild has come to be recognized as both a triumph of Darwinian sentiment: the monkey's uncle. Would you want your daughter to engineering and an intellectual vehicle for exploring the most abstract realms marry a robot? Well, if Darwin is right, your great-great-... grandmother
was
of pure science. Both thinkers were at one and the same time awesome a robot! A macro, in fact. That is the unavoidable conclusion of the previous theorists and deeply practical, epitomizing an intellectual style that has been chapters. Not only are you descended from macros; you are composed of playing a growing role in science since the Second World War. In addition to them. Your hemoglobin molecules, your antibodies, your neurons, your creating the computer, both Turing and von Neumann made fundamental vestibular-ocular reflex machinery—at every level of analysis we find contributions to theoretical biology. Von Neumann, as we have already noted, machinery that dumbly does a wonderful, elegantly designed job. We have applied his brilliant mind to the abstract problem of self-replication, and ceased to shudder, perhaps, at the scientific vision of viruses and bacteria Turing (1952) did pioneering work on the most basic theoretical problems of busily and mindlessly executing their subversive projects—horrid little au-embryology or morphogenesis: how can the complex topology—the shape—

tomata doing their evil deeds. But we should not think that we can take of an organism arise from the simple topology of the single fertilized cell comfort in the thought that
they
are alien invaders, so unlike the more from which it grows? The process begins, as every high-school student congenial tissues that make up
us.
We are made of the same sorts of automata knows, with an event of quite symmetrical division. (As Francois Jacob has that invade us—no halos of
elan vital
distinguish your antibodies from the said, the dream of every cell is to become two cells.) Two cells become four, antigens they combat; they simply belong to the club that is you, so they fight and four become eight, and eight become sixteen; how do hearts and livers on your behalf.

and legs and brains get started under such a regime?12 Turing saw the Can it be that if you put enough of these dumb homunculi together you continuity between such molecular-level problems and the problem of how a make a real conscious person? The Darwinian says there could be no other poet writes a sonnet, and from the earliest days of computers, the ambition of way of making one. Now, it certainly does not follow from the fact that you those who saw what Turing saw has

are descended from robots that you are a robot. After all, you are also a direct descendant of some fish, and you are not a fish; you are a direct descendant of some bacteria, and you are not a bacterium. But unless dualism or vitalism is true (in which case you have some extra, secret ingredient in you ), you are 12. Two highly accessible accounts of Turing's work on morphogenesis are Hodges (

made of
robots—or what comes to the same thing, a collection of trillions of 1983, ch. 7), and Stewart and Golubitsky ( 1992), which also discusses their relation to macromolecular machines. And all of these are ultimately descended from the more recent theoretical explorations in the field. Beautiful as Turing's ideas are, they original macros. So something made of robots
can
exhibit genuine probably have at best a very attenuated application to real biological systems. John consciousness, or genuine intentionality, because you do if anything does.

Maynard Smith (personal communication) recalls being entranced by Turing's 1952

paper (which his supervisor, J. B. S. Haldane, had shown him), and for years he was No wonder, then, that there should be so much antagonism to both convinced that "my fingers must be Turing waves; my vertebrae must be Turing waves"—

but he eventually came to realize, reluctantly, that it could not be that simple and beautiful.

208 BIOLOGY IS ENGINEERING

The Computer That Learned to Play Checkers
209

been to use his wonderful machine to explore the mysteries of thought.13

lectual species, AI, but also of its more recent offshoot, AL, Artificial Life.

Turing published his prophetic essay, "Computing Machinery and Intel-Legendary though it is, few people today are familiar with its remarkable ligence," in the philosophical journal
Mind
in 1950, surely one of the most details, many of which deserve to be more widely known.14 Samuel's first frequently cited articles ever to appear in that journal. At the time he wrote it, checkers program was written in 1952, for the IBM 701, but the learning there were no Artificial Intelligence programs—there were really only two version wasn't finished until 1955, and ran on an IBM 704; a later version ran operating computers in the world—but within a few years, there were enough on the IBM 7090. Samuel found some elegant ways of coding any state of a machines up and running twenty-four hours a day so that Arthur Samuel, a checkers game into four thirty-six-bit "words" and any move into a simple research scientist at IBM, could fill the otherwise idle late-night time on one arithmetical operation on those words. (Compared with today's prodigiously of the early giants with the activities of a program that is as good a candidate wasteful computer programs which run on for megabytes, Samuel's basic as any for the retrospective title of AI-Adam. Samuel's program played program was microscopic in size—a "low-tech" genome indeed, with fewer checkers, and it got better and better by playing against itself through the than six thousand lines of code—but, then, he had to write it in machine small hours of the night, redesigning itself by throwing out earlier versions code; this was before the days of computer programming languages.) Once that had not fared well in the nightly tournament and trying out new he'd solved the problem of representing the basic process of legal checkers mutations that were mindlessly generated. It eventually became a much better play, he had to face the truly hard part of the problem: getting the computer checkers-player than Samuel himself, providing one of the first clear program to
evaluate
the moves, so it could
select
the best move (or at least counterexamples to the somewhat hysterical myth that "a computer can really one of the better moves) whenever possible.