Game theory as a common language

The common language of social science and evolution for studying cooperation

Karl Sigmund and Martin Nowak offer a perspective on evolution and social science in this week's _Science_, 3-Nov-2000, 290:5493, called "A Tale of Two Selves."

The "Two Selves" are the 2 different versions of selfish theories: the economists' selfish human, Homo economicus, who supposedly acts rationally for their own benefit,
and the biologists' selfish genes, which construct vehicles for the benefit of their own replication.

The article focuses on what the authors consider an impending merger of social sciences and natural sciences through better understanding of human nature, heralded by
the gathering of specialists for European Science Days on "The Evolution of Cooperation and Communication" held in July, 2000.

The central thrust is represented by evolutionary game theory, which is described as the common language of the meeting. While human beings superficially resemble insect colonies, the same genetic explanations don't work because our social structures don't reflect genetic ties. While it is a powerful, robust, and widely applicable concept, even reciprocal altruism doesn't carry the day because of problems with group size. Reciprocal altruism becomes less likely as the group size increases, and group size has often been theorized as an important factor in human evolution.

Evolutionary biologist John Maynard Smith proposes that the fates of genes become linked in progressively larger and stronger competing entities, causing natural selection to act on a number of different kinds of units, not just gene frequencies. With the help of game theory, we can construct experiments that demonstrate where human decisions
become irrational for the individual, yet contribute to a positive outcome in some sense.

In the "Public Goods" game for example, studied by Ernst Fehr, four players are given $10 each and asked individually to invest any part into a common pool. The experimenter doubles the pool and distributes it evenly. If everyone invests their money, they all get back double.

The temptation to freeload in "Public Goods" is great since they can get back more relative to their investment if others invest heavily. As is often found in these kinds of games, a significant number of real humans initially tend to invest heavily rather than act strictly according to rational selfishness. However after a few rounds, they all learn to contribute less and less, until all contributions become negligible, and we have a sort of Tragedy of the Commons manifest itself.

But then surprisingly, everything changes if an option for punishment is introduced into the game. If the less generous players are fined by more generous players for not
investing (even though the fines don't go to any player), players begin to invest in fear of punishment, to the benefit of all. Remarkably, even if the punisher
is themselves fined (half the fine assessed) for the privilege of punishing, they still tend to do it. Fehr points out that punishment in similar situations is widespread and joins kinship and reciprocal altruism as factors influencing the evolution of cooperation.

Replies:

The Way of the Loner, by Todd, 05/09/02

| Behavior OnLine Home Page | Disclaimer |

Copyright © 1996-2004 Behavior OnLine, Inc. All rights reserved.