Darwin's Dangerous Idea (48 page)

promise each other to hang tough? (In the standard jargon of the Prisoner's Dilemma, the hang-tough option is called
cooperating.
) You could promise, 13. 1 sometimes wonder if there is any important advance in thinking in the second half of this century that von Neumann is
not
the father of. The computer, the model of self-but you would each then feel the temptation—whether or not you acted on replication, game theory—and if that weren't enough, von Neumann also made major it—to
defect,
since then you would go scot free, leaving the
sucker,
sad to contributions to quantum physics. For what it is worth, however, I suspect that his say, in deep trouble. Since the game is symmetrical, the other person will be formulation of the measurement problem in quantum mechanics is his one bad idea, a just as tempted, of course, to make a sucker of you by defecting. Can you risk sleight-of-hand endorsement of a fundamentally Cartesian model of conscious observa-life in prison on the other person's keeping his promise? Probably safer to tion that has bedeviled quantum mechanics ever since. My student Turhan Canli first defect, isn't it? That way, you definitely avoid the worst outcome of all, and opened this door in his (undergraduate!) term paper for me on the problem of Schrodinger's cat, in which he developed the sketch of an alternative formulation of quantum might even go free. Of course, the other fellow will figure this out, too, if it's physics in which time is quantized. If I ever master the physics (a very remote prospect, such a bright idea, so he'll probably play it safe and defect, too, in which case sad to say ), 1 will tackle this hunch, which might extend in wildly ambitious ways my you
must
defect to avoid calamity—unless you are so saintly that you don't theory of consciousness (1991a); more likely, however, is the prospect that I will be a mind spending your life in prison to save a promise-breaker!—so you'll both semi-comprehending but enthusiastic spectator of this development, wherever it leads.

wind up with medium-length sentences. If only you could overcome this 14. For a fascinating account of the history of game theory and its relation to nuclear reasoning and cooperate!

disarmament, see William Poundstone's 1992 book,
Prisoner's Dilemma John von Neumann, Game Theory, and the Puzzle of the Bomb.

254 SEARCHING FOR QUALITY

Playing with Constraints
255

The logical structure of the game is what matters, not this particular set-could avoid the trouble of building those ridiculous and expensive trunks, ting, which is a usefully vivid imagination-driver. We can replace the prison stay low and thrifty shrubs, and get just as much sunlight as before! [Den-sentences with positive outcomes (it's a chance to win different amounts of nett 1990b, p. 132.]

cash—or, say, descendants ) just so long as the payoffs are symmetrical, and ordered so that lone defection pays more than mutual cooperation, which But they can't get together; under these circumstances, defection from any pays each more than mutual defection does, which in turn pays more than the cooperative "agreement" is bound to pay off if ever or whenever it occurs, so sucker payoff one gets when the other is a lone defector. (And in formal set-trees would be stuck with the "tragedy of the commons" (Hardin 1968) if tings we set a further condition: the average of the sucker and mutual-there weren't an essentially inexhaustible supply of sunshine. The tragedy of defection payoffs must not be greater than the mutual-cooperation payoff.) the commons occurs when there is a finite "public" or shared resource that Whenever this structure is instantiated in the world, there is a Prisoner's Di-individuals will be selfishly tempted to take more of than their fair share—

lemma.

such as the edible fish in the oceans. Unless very specific and enforceable Game-theoretic explorations have been undertaken in many fields, from agreements can be reached, the result will tend to be the destruction of the philosophy and psychology to economics and biology. The most influential resource. Many species, in many regards, face various sorts of Prisoner's of the many applications of game-theoretic thinking to evolutionary theory is Dilemmas. And we human beings face them both consciously and Maynard Smith's concept of an
evolutionarily stable
strategy,
or ESS, a unconsciously—sometimes in ways that we might never have imagined strategy that may not be "best" from any Olympian (or Fujian!) standpoint, without the aid of adaptationist thinking.

but is unimprovable-upon and unsubvertible under the circumstances. May-Homo sapiens is not exempt from the sort of genetic conflict David Haig nard Smith (1988, especially chh. 21 and 22) is an excellent introductory postulates to explain genomic imprinting; in an important new article (1993) account of game theory in evolution. The revised edition of Richard Daw-he analyzes a variety of conflicts that exist between the genes of a pregnant kins'
The Selfish Gene
( 1989a) has a particularly good account of the de-woman and the genes of her embryo. It is in the embryo's interests, of course, velopment of ESS thinking in biology during the last decade or so, when that the mother bearing it stay strong and healthy, for its own survival large-scale computer simulations of various game-theoretic models revealed depends on her not only completing her term of pregnancy but tending for her complications that had been overlooked by the earlier, less realistic versions.

newborn. However, if the mother, in her attempt to stay healthy under trying circumstances—famine, for instance, which must have been a common I now like to express the essential idea of an ESS in the following more circumstance in most generations of human existence— should cut down on economical way. An ESS is a strategy that does well against copies of itself.

the nutrition she provides her embryo, at some point this becomes more of a The rationale for this is as follows. A successful strategy is one that dom-threat to the embryo's survival than the alternative, a weakened mother.

inates the population. Therefore it will tend to encounter copies of itself.

If the embryo were "given a choice" between being spontaneously aborted Therefore it won't stay successful unless it does well against copies of early in the pregnancy or being stillborn or of low birth weight on the one itself. This definition is not so mathematically precise as Maynard Smith's, hand, versus being born at normal weight of a weak or even dying mother on and it cannot replace his definition because it is actually incomplete. But the other, what would (selfish) reason dictate? It would dictate taking it does have the virtue of encapsulating, intuitively, the basic ESS idea.

whatever steps are available to try to ensure that the mother does not cut her

[Dawkins 1989a, p. 282.]

losses (she can always try to have another child later, when the famine is over), and this is just what the embryo does. Both embryo and mother can be There can be no doubt that game-theoretic analyses work in evolutionary entirely oblivious of this conflict—as oblivious as the trees rising theory. Why, for instance, are the trees in the forest so tall? For the very same competitively in the forest. The conflict plays out in the genes and their reason that huge arrays of garish signs compete for our attention along control of hormones, not in the brains of mother and embryo; it is the same commercial strips in every region of the country! Each tree is looking out for sort of conflict we saw between maternal and paternal genes in the mouse.

itself, and trying to get as much sunlight as possible.

There is a flood of hormones; the embryo produces a hormone that will enhance its own growth at the expense of the mother's nutritional needs, her If only those redwoods could get together and agree on some sensible body responds with an antagonist hormone that attempts to undo the effect of zoning restrictions and stop competing with each other for sunlight, they the first; and so on, in an escalation that can produce 256 SEARCHING FOR QUALITY

Playing with Constraints
257

hormone levels many times higher than normal. This tug-of-war usually ends take? Isn't this Panglossian optimism? (As we have just seen, this sometimes in a mutually semi-satisfactory standoff, but it produces a host of by-products looks more like Panglossian
pessimism.
"Darn—organisms are 'too smart' to that would be utterly baffling and senseless were they not the predictable cooperate!"16)

effects of such conflict. Haig concludes with an application of the The standard assumption of game theory is that there will always be fundamental game-theoretic insight: "Maternal and fetal genes would both mutations that have the "right" phenotypic effects to rise to the occasion, but benefit if a given transfer of resources was achieved with a lesser production what if the right move just doesn't "occur to Mother Nature"? Is this ever or of... hormones and less maternal resistance, but such an agreement is often very likely? We certainly know of cases in which Mother Nature
does
evolutionarily unenforceable" (Haig 1993, p. 518).

take the move—to make the forests, for instance. Are there perhaps just as This is not, in many regards, welcome news. Von Neumann's all-too-casual many (or more) cases in which some sort of hidden constraint prevents this remark on the inevitability of human selfishness epitomizes the Darwinian from happening? There may well be, but in every such case, adaptationists mind-set that many people view with loathing, and it is not hard to see why.

will want to persist by asking the next question: And is
there a reason
in this They fear that Darwinian "survival of the fittest" would
entail
that people are case why Mother Nature doesn't take the move, or is it just a brute, nasty and selfish. Isn't that just what von Neumann is saying? No. Not quite.

unthinking constraint on Mother Nature's rational gamesmanship?

He is saying that it is indeed entailed by Darwinism that such virtues as Gould has suggested that a fundamental flaw of adaptationist reasoning is cooperation should be
in general
"evolutionarily unenforceable" and hence the assumption that in every fitness landscape, the way is always shown as hard to come by. If cooperation and the other unselfish virtues are to exist, clear to the tops of the various summits, but there might well be hidden
they must be designed
—they do not come for free. They
can
be designed constraints, rather like railroad tracks lying across the landscape. "The conunder special circumstances. (See, for instance, Eshel 1984, 1985, and Haig straints of inherited form and developmental pathways may so channel any and Grafen 1991) After all, the eukaryotic revolution that made multicelled change that even though selection induces motion down permitted paths, the organisms possible was a revolution that began when an enforceable truce channel itself represents the primary determinant of evolutionary direction"

was somehow engineered between certain prokaryotic cells and their (Gould 1982a, p. 383). Populations, then, do not get to spread
ad lib
across bacterial invaders. They found a way of joining forces and submerging their the terrain, but are forced to stay on the tracks, as in figure 9.4.

selfish interests.

Suppose this is true. Now, how do we locate the hidden constraints? It is Cooperation and the other virtues are, in general, rare and special prop-all very well for Gould and Lewontin to point to the possibility of hidden erties that can only emerge under very particular and complex R-and-D cir-constraints—every adaptationist already acknowledges this as an omnipres-cumstances. We might contrast the Panglossian Paradigm, then, with the ent possibility—but we need to consider what methodology might be best for Pollyannian Paradigm, which cheerfully assumes, with Pollyanna, that discovering them. Consider a curious variation on a standard practice in Mother Nature is Nice.15 In general, she isn't—but that isn't the end of the chess.

world. Even in the present case, we can see that there are other perspectives When a stronger player plays a weaker opponent in friendly matches, the to adopt. Aren't we really rather fortunate, for instance, that trees are so in-stronger player often volunteers to take on a handicap, to make the game superably selfish? The beautiful forests—to say nothing of the beautiful more evenly matched and exciting. The standard handicap is to give up a wooden sailing ships and the clean white paper on which we write our po-piece or two—to play with only one bishop or one rook, or, in a really etry—could not exist if trees weren't selfish.

extreme case, to play without a queen. But here is another handicapping There can be no doubt, as I say, that game-theoretic analyses work in system that might have interesting results. Before the match, the stronger evolutionary theory, but do they
always
work? Under what conditions do player writes down on a piece of paper a hidden constraint ( or constraints ) they apply, and how can we tell when we are overstepping? Game-theory calculations always assume that there is a certain range of "possible" moves, from which the selfish-by-definition contestants make their choices. But how realistic is this
in general!
Just because a move in a particular circumstance is 16. The Panglossian pessimist says, "Isn't it a shame that this is, after all, the best of all possible worlds!" Imagine a beer commercial: As the sun sets over the mountains, one of the move that
reason dictates,
is it the move nature will always the hunks lounging around the campfire intones, "It doesn't get any better than this!"—at which point his beautiful companion bursts into tears: "Oh no! Is that really true?" It wouldn't sell much beer.

15. For a powerful antidote to the Pollyannian Paradigm, see G. Williams 1988.