Darwin's Dangerous Idea (24 page)

bet. So we would not be surprised to find that locomoting organisms on other This marriage of chance and necessity is a hallmark of biological regular-planets (with transparent atmospheres) had eyes. Eyes are an obviously good ities. People often want to ask: "Is it merely a massively contingent fact that solution to a very general problem that would often be encountered by circumstances are as they are, or can we read some deep necessity into moving metabolizers. Eyes may not always be "available," of course, for them?" The answer almost always is: Both. But note that the type of neces-QWERTY reasons, but they are obviously rational solutions to this highly sity that fits so well with the chance of random, blind generation is the abstract design problem.

necessity of
reason.
It is an inescapably teleological variety of necessity, the dictate of what Aristotle called
practical reasoning,
and what Kant called a
hypothetical imperative.

2. FORCED MOVES IN THE GAME OF DESIGN

If you
want to achieve goal G,
then this is what you
must
do, given the Now that we have encountered this appeal to what is obviously rational under circumstances.

some general set of circumstances, we can look back and see that our case of necessity, having an autonomous metabolism, can be recast as simply the The more universal the circumstances, the more universal the necessity.

only
acceptable solution to the
most general
design problem of life. If you That is why we would not be surprised to find that the living things on other wanna live, you gotta eat. In chess, when there is only one way of staving off planets included locomotors with eyes, and why we would be more than disaster, it is called a
forced move.
Such a move is not forced by the rules of surprised—utterly dumfounded—if we found things scurrying around on chess, and certainly not by the laws of physics (you can always kick the table various projects but lacking any metabolic processes. But now let us consider over and run away), but by what Hume might call a "dictate of reason." It is the difference between the similarities that would surprise us and the simply dead obvious that there is one and only one solution, as anybody with similarities that would not. Suppose SETI struck it rich, and established an ounce of wit can plainly see. Any alternatives are immediately suicidal.

communication with intelligent beings on another planet. We would not be In addition to having an autonomous metabolism, any organism must also surprised to find that they understood and used the same arithmetic that we have a more or less definite boundary, distinguishing itself from everything do. Why not? Because arithmetic is
right.

else. This condition, too, has an obvious and compelling rationale: "As soon Might there not be different kinds of arithmetic-like systems, all equally as something gets into the business of self-preservation, boundaries become good? Marvin Minsky, one of the founders of Artificial Intelligence, has important, for if you are setting out to preserve yourself, you don't want to squander effort trying to preserve the whole world: you draw the line"

(Dennett 1991a, p. 174). We would also expect the locomoting organisms on an alien planet to have efficiently shaped boundaries, like those of organisms 2. Are the constraints of pure logic deep or shallow? Some of each, I guess, depending on on Earth. Why? (Why on Earth?) If cost were no object, one might have no their obviousness. A delicious parody of adaptationist thinking is Norman Ellestrand's regard for streamlining in organisms that move through a relatively dense

"Why are Juveniles Smaller Than Their Parents?" (1983 ), which explores with a hero-fluid, such as water. But cost is
always
an object—the Second Law of ically straight face a variety of "strategic" reasons for JSS (Juvenile Small Size). It ends with a brave look towards future research: "In particular, another juvenile character is Thermodynamics guarantees that.

even more widespread than JSS and deserves some thoughtful theoretical attention, the fact that juveniles
always
seem to be younger than their parents."

130 THREADS OF ACTUALITY IN DESIGN SPACE

Forced Moves in the Game of Design
131

explored this curious question, and his ingeniously reasoned answer is No. In with five subunits. The "solution" of using-whatever-you've-got to count on

"Why Intelligent Aliens Will Be Intelligible," he offers grounds for believing is a fairly obvious one, if not quite in the forced-move category.3 It would not in something he calls the

be particularly surprising to find that our aliens had
a pair
of prehensile appendages, considering the good reasons there are for bodily symmetry, and Sparseness Principle: Whenever two relatively simple processes have prod-the frequency of problems that require one thing to be manipulated relative to ucts which are similar, those products are likely to be completely identi-another. But that there should be five subunits on each appendage looks like a cal! [Minsky 1985a, p. 119, exclamation point in the original.]

QWERTY phenomenon that has been deeply rooted for hundreds of millions of years—a mere historical happenstance that has restricted
our
options, but Consider the set of
all possible processes,
which Minsky interprets
a la
the should not be expected to have restricted theirs. But perhaps we Library of Babel as all permutations of all possible computers. (Any com-underestimate the Tightness, the rationality, of having five subunits. For puter can be identified, abstractly, as one "Turing machine" or another, and reasons we have not yet fathomed, it may be a Good Idea in general, and not these can be given unique identifying numbers, and then put in numerical merely something we are stuck with. Then it would not be amazing after all order, just like the alphabetical order in the Library of Babel.) Except for a to find that our interlocutors from outer space had converged on the same Vanishing few, the Vast majority of these processes "do scarcely anything at Good Idea, and counted in tens, hundreds, and thousands.

all." So if you find "two" that do something similar (and worth noticing), they We would be flabbergasted, however, to find them using the very symbols are almost bound to be one and the same process, at some level of analysis.

we use, the so-called arabic numerals: "1," "2," "3" ... We know that right Minsky (p. 122) applies the principle to arithmetic: here on Earth there are perfectly fine alternatives, such as the Arabic numerals, " I," "v," " v,"
"i"
... as well as some not-so-viable alternatives, such From all this, I conclude that any entity who searches through the simplest as roman numerals, "i," "ii,"
"iii," "iv"
... If we found the inhabitants of processes will soon find fragments which do not merely resemble arith-another planet using our arabic numerals, we would be quite sure that it was metic but
are
arithmetic. It is not a matter of inventiveness or imagination, no coincidence—there
had
to be a historical connection. Why? Because the only a fact about the geography of the universe of computation, a world far space of possible numeral shapes in which there is no
reason
for choosing more constrained than that of real things.

one over the others is Vast; the likelihood of two independent "searches"

ending up in the same place is Vanishing.

The point is clearly not restricted to arithmetic, but to all "necessary Students often have a hard time keeping clear about the distinction be-truths"—what philosophers since Plato have called
a priori
knowledge. As tween numbers and numerals. Numbers are the abstract, "Platonic" objects Minsky (p. 119 ) says, "We can expect certain
'apriori'
structures to appear, that numerals are the names of. The arabic numeral "4" and the roman almost always, whenever a computation system evolves by selection from a numeral "IV" are simply different
names
for one and the same thing—the universe of possible processes." It has often been pointed out that Plato's
number
4. (I can't talk about the number without naming it in one way or curious theory of reincarnation and reminiscence, which he offers as an another, any more than I can talk about Clinton without using some word explanation of the source of our
a priori
knowledge, bears a striking resemblance to Darwin's theory, and this resemblance is particularly striking from our current vantage point. Darwin himself famously noted the resemblance in a remark in one of his notebooks. Commenting on the claim that 3. Seymour Papert (1993, p. 90) describes observing a "learning disabled" boy in a Plato thought our "necessary ideas" arise from the pre-existence of the soul, classroom in which counting on your fingers was forbidden: "As he sat in the resource Darwin wrote: "read monkeys for preexistence" (Desmond and Moore 1991, room I could see him itching to do finger manipulations. But he knew better. Then I saw p. 263).

him look around for something else to count with. Nothing was at hand. I could see his frustration grow. What could I do?... Inspiration came! I walked casually up to the boy We would not be surprised, then, to find that extra-terrestrials had the and said out loud: 'Did you think about your teeth?' I knew instantly from his face that he same unshakable grip on "2 + 2 = 4" and its kin that we do, but we would be got the point, and from the aide's face that she didn't. 'Learning disability indeed!' I said surprised, wouldn't we, if we found them using the decimal system for to myself. He did his sums with a half-concealed smile, obviously delighted with the expressing their truths of arithmetic. We are inclined to believe that our subversive idea." (When considering using-whatever-you've-got as a possible forced fondness for it is something of a historical accident, derived from counting move, it is worth recalling that not all peoples of our Earth have used the decimal system; the Mayans, for instance, used a base-20 system.)

on our two five-digit hands. But suppose they, too, have a pair of hands, each 132 THREADS OF ACTUALITY IN DESIGN SPACE

Forced Moves in the Game of Design
133

or words that refer to him, but Clinton is a man, not a word, and numbers there was a perfectly good engineering
reason
for the initial choice—it was aren't symbols either—numerals are.) Here is a vivid way of seeing the just a reason whose supporting circumstances had long ago lapsed.) importance of the distinction between numbers and numerals; we have just Design work—lifting—can now be characterized as the work of discov-observed that it would
not
be surprising at all to find that extra-terrestrials ering good ways of solving "problems that arise." Some problems are given at used the same
numbers
we do, but simply incredible if they used the same the outset, in all environments, under all conditions, to all species. Further
numerals.

problems are then created by the initial "attempts at solution" made by In a Vast space of possibilities, the odds of a similarity between two different species faced with the first problems. Some of these subsidiary independently chosen elements is Vanishing
unless there is a reason.
There problems are created by the other species of organisms (who must make a is for numbers (arithmetic is
true
and variations on arithmetic aren't) and living, too), and other subsidiary problems are created by a species' own there isn't for numerals (the symbol "§" would function exactly as well as the solutions to its own problems. For instance, now that one has decided—by symbol "5" as a name for the number that follows 4).

flipping a coin, perhaps—to search for solutions in
this
area, one is stuck with Suppose we found the extra-terrestrials, like us, using the decimal system problem B instead of problem A, which poses subproblems p, q, and r, for most informal purposes, but converting to binary arithmetic when doing instead of subproblems x, y, and z, and so forth. Should we personify a computation with the aid of mechanical prosthetic devices (computers). Their species in this way and treat it as an agent or practical reasoner (Schull 1990, use of 0 and 1 in their computers (supposing they had invented computers!) Dennett 1990a)? Alternatively, we may choose to think of species as would not surprise us, since there are good engineering
reasons
for adopting perfectly mindless nonagents, and put the rationale in the process of natural the binary system, and though these reasons are not dead obvious, they are selection itself (perhaps jocularly personified as Mother Nature). Remember probably within striking distance for average-type thinkers. "You don't have Francis Crick's quip about evolution's being cleverer than you are. Or we may to be a rocket scientist" to appreciate the virtues of binary.

choose to shrink from these vivid modes of expression altogether, but the In general, we would expect them to have discovered many of the various analyses we do will have the same logic in any case.

ways things have of
being the right way.
Wherever there are many different This is what lies behind our intuition that design work is somehow in-ways of skinning a cat, and none is much better than any other, our surprise tellectual work. Design work is discernible (in the otherwise uninterpret-able at their doing it
our
way will be proportional to how many different ways we typography of shifting genomes) only if we start imposing
reasons
on it. (In think there are. Notice that even when we are contemplating some Vast earlier work, I characterized these as "free-floating rationales," a term that number of
equivalent ways,
a value judgment is implicit. For us to recognize has apparently induced terror or nausea in many otherwise well-disposed items as things falling in one of these Vast sets, they have to be seen as readers. Bear with me; I will soon provide some more palatable ways of equally good ways, as ways of
performing the function x.
Function-alistic making these points.)

thinking is simply inescapable in this sort of inquiry; you can't even So Paley was right in saying not just that Design was a wonderful thing to enumerate the possibilities without presupposing a concept of function. (Now explain, but also that Design took Intelligence. All he missed—and Darwin we can see that even our deliberately antiseptic formalization of the Library provided—was the idea that this Intelligence could be broken into bits so tiny of Mendel invoked functional presuppositions; we can't identify something and stupid that they didn't count as intelligence at all, and then distributed as a
possible genome
without thinking of genomes as things that might serve through space and time in a gigantic, connected network of algorithmic a particular function within a reproductive system.) process. The work must get done, but which work gets done is largely a So there turn out to be general principles of practical reasoning (including, matter of chance, since chance helps determine which problems (and in more modern dress,
cost-benefit analysis)
that can be relied upon to subproblems and subsubproblems) get "addressed" by the machinery.