Rationally
Speaking
A monthly e-column by Massimo
Pigliucci
Department of Botany,
University of Tennessee
N. 8, March 2001: "Game Theory, Rational Egoism and the Evolution
of Fairness"
Is it rational to be ethical? Many philosophers have wrestled with
this most fundamental of questions, attempting to clarify whether
humans are well served by ethical rules or whether they weigh us down.
Would we really be better off if we all gave in to the desire to just
watch out for our own interests and take the greatest advantage to
ourselves whenever we can? Ayn Rand, for one, thought that the only
rational behavior is egoism, and books aiming at increasing personal
wealth (presumably at the expense of someone else's wealth) regularly
make the bestsellers list.
Plato, Kant, and John Stuart Mill, to mention a few, have tried to
show that there is more to life than selfishness. In the Republic,
Plato has Socrates defending his philosophy against the claim that
justice and fairness are only whatever rich and powerful people decide
they are. But the arguments of his opponents - that we can see plenty
of examples of unjust people who have a great life and of just ones
who suffer in equally great manner - seem more convincing than the
high-mindedness of the father of philosophy.
Kant attempted to reject what he saw as the nihilistic attitude of
Christianity, where you are good now because you will get an infinite
payoff later, and to establish independent rational foundations for
morality. Therefore he suggested that in order to decide if something
is ethical or not one has to ask what would happen if everybody were
adopting the same behavior. However, Kant never explained why his
version of rational ethics is indeed rational. Rand would object that
establishing double standards, one for yourself and one for the rest
of the universe, makes perfect sense.
Mill also tried to establish ethics on firm rational foundations,
in his case improving on Jeremy Bentham's idea of utilitarianism.
In chapter two of his book Utilitarianism, Mill writes: "Actions
are right in proportion as they tend to promote happiness; wrong as
they tend to produce the reverse of happiness." Leaving aside
the thorny question of what happiness is and the difficulty of actually
making such calculations, one still has to answer the fundamental
question of why one should care about increasing the average degree
of happiness instead of just one's own.
Things got worse with the advent of modern evolutionary biology.
It seemed for a long time that Darwin's theory would provide the naturalistic
basis for the ultimate selfish universe: nature red in tooth and claw
evokes images of "every man for himself," in pure Randian
style. In fact, Herbert Spencer popularized the infamous doctrine
of "Social Darwinism" (which Darwin never espoused) well
before Ayn Rand wrote Atlas Shrugged.
Recently, however, several scientists and philosophers have been
taking a second look at evolutionary theory and its relationship with
ethics, and are finding new ways of realizing the project of Plato,
Kant, and Mill of deriving a fundamentally rational way of being ethical.
Elliot Sober and David Sloan Wilson, in their Unto Others: the Psychology
and Evolution of Unselfish Behavior, as well as Peter Singer in A
Darwinian Left: Politics, Evolution and Cooperation, argue that human
beings evolved as social animals, not as lone, self-reliant brutes.
In a society, cooperative behavior (or at least, a balance between
cooperation and selfishness) will be selected in favor, while looking
out exclusively for number one will be ostracized because it reduces
the fitness of most individuals and of the group as a whole.
All of this sounds good, but does it actually work? A recent study
published in Science by Martin Nowak, Karen Page and Karl Sigmund
provides a splendid example of how mathematical evolutionary theory
can be applied to ethics, and how in fact social evolution favors
fair and cooperative behavior. Nowak and coworkers tackled the problem
posed by the so-called "ultimatum game." In it, two players
are offered the possibility of winning a pot of money, but they have
to agree on how to divide it. One of the players, the proposer, makes
an offer of a split ($90 for me, $10 for you, for example) to the
other player; the other player, the responder, has the option of accepting
or rejecting. If she rejects, the game is over and neither of them
gets any money.
It is easy to demonstrate that the rational strategy is for the proposer
to behave egotistically and to suggest a highly uneven split in which
she takes most of the money, and for the responder to accept. The
alternative is that neither of them gets anything. However, when real
human beings from a variety of cultures and using a panoply of rewards
play the game the outcome is invariably a fair share of the prize.
This would seem prima facie evidence that the human sense of fair
play overwhelms mere rationality and thwarts the rationalistic prediction.
On the other hand, it would also provide Ayn Rand with an argument
that most humans are simply stupid, because they don't appreciate
the math behind the game.
Nowak and colleagues, however, simulated the evolution of the game
in a situation in which several players get to interact repeatedly.
That is, they considered a social situation rather than isolated encounters.
If the players have memory of previous encounters (i.e., each player
builds a "reputation" in the group), then the winning strategy
is to be fair because people are willing to punish dishonest proposers,
which increases their own reputation for fairness and damages the
proposer's reputation for the next round. This means that - given
the social environment - it is rational to be less selfish toward
your neighbors.
While we are certainly far from a satisfying mathematical and evolutionary
theory of morality, it seems that science does, after all, have something
to say about optimal ethical rules. And the emerging picture is one
of fairness - not egotism - as the smart choice to make.
Next Month: "Red or Blue? What kind of life would
you choose?" © by Massimo Pigliucci, 2001
