Brian Skyrms' Evolution of the Social Contract

Skyrms' book, Evolution of the Social Contract, offers a compelling explanation as to why individuals, when placed with one-shot prisoner's dilemmas, will often cooperate, or choose the equilibrium that will benefit both parties equally. He uses examples to outline how individuals of certain environments frequently engage in activities that benefit the group at their own personal expense. Using both game theory and decision theory, Skyrms explores problems with the social contract when it is applied to evolutionary dynamics. In the chapters of the book, he offers new insights into concepts such as sex and justice, commitment, and mutual aid.

Skyrms' writing goes beyond traditional game theory, and exposes some weaknesses in its application. He rejects the theory's traditional interpretation of rational actors and actions by discovering some glaring inconsistencies. Skyrms conducted a number of experiments using one-shot prisoners' dilemmas. The ultimatum the author introduces in the first chapter serves as a simple example of a one-shot prisoners' dilemma. In the initial form of the example, Skyrms proposes there is a cake that must be divided between two individuals. Each individual is looking to maximize his or her utility, and therefore, wants as much of the cake as possible. However, there is a third party, or what Skryms labels a "referee." The two individuals must determine the percentage or portion of the cake they want and summit these requests to the referee. The percentages must not exceed 100%, or the referee will consume all the cake. It is therefore not in either parties' best interest to request a significantly large portion. Additionally, if the total of the two requests is below 100% of the cake, the referee will take the left-over portion. The two parties will then aim to maximize their portion, however the best claim that an individual submits is dependent upon the other party's claim. There are two interacting optimization problems (Skyrms 3, 4).

An answer to the puzzle will be found in solutions that are in equilibrium. An equilibrium in informed rational self-interest, or a Nash equilibrium, is any solution to the problem whereby neither party could do better by altering its position. However, this is a general and broad definition. Further stipulations may be added. It could be further requisite that should a party alter its position, not only would it not do better, it would do worse than it would have at equilibrium. The inclusion of this additional provision is considered a strict Nash equilibrium. To arrive at such an outcome, given Skyrms example, each individual will request half of the cake (Skyrms 4, 5).

Such a result is frequent in laboratory experiments. The behaviors can be a bit puzzling when approaching the situation from the view of traditional game theory. One could argue that such a decision is quite irrational. If utility is measured in terms of material resources, the individuals are acting irrationally. However, it is not necessarily irrational for individuals to cooperate in situations where the individual's gain is so dependent

upon the other individual. It seems Skyrms is not attempting to explain this irrationality, but rather, is exploring how people can interpret utility in a broader context than personal self-interest. As he states in the text, "Equilibrium in informed rational self-interest, even when strictly construed, does not explain our conception of justice" (Skyrms 5).

Skyrms incorporates evolution and the theory of Ð''survival of the fittest' into his exploration of game theory and choice. Skyrms presupposes that our behavioral dispositions are inherited, and have evolved over time. He relates our sense of justice in certain situations to our set of behavioral dispositions. Given that behaviors are inherited, Skyrms proposes that certain notions of justice have come into existence by evolving over many years. Certain behaviors, or concepts of justice, were Ð''fitter' than others, and have survived through generations. In the cake example, Skyrms develops a concept of positive correlation between individuals, which he believes to be a result of evolved behaviors.

Skyrms uses further, more advanced examples to support the notion of evolved behaviors. He constructs an environment containing two categories of individuals: those that are "greedy" and those that are "modest." The greedy individuals claim 2/3 of the cake, while the modest only claim 1/3. Skyrms points out that the greedy individuals will benefit from being paired with the modest individuals, and may initially have an advantage. However, Skyrms also reveals that the greedy individuals will never gain anything when paired with one another (2/3 + 2/3 > 1). However, the modest individuals will gain 1/3 against the greedy individuals and 1/3 against other modest individuals. Given these circumstances, eventually the average gain for both greedy individuals and modest individuals will be 1/3, and the two groups will be at equilibrium. Skyrms further points out that ratio of greedy individuals to modest individuals is irrelevant, as negative feedback will eventually return the population proportions to equality (Skryms 12, 13).

To further support the theory of an individual's evolved tendency to cooperate in a manner of 50/50 when given the opportunity, one could further elaborate on the above-mentioned example. Skyrms used a computer to create a mock environment where 10 pieces of cake were available for consumption. The computer then created an initial combination of population proportions at random. Skyrms then ran 1,000 trials. According to his findings, fair division will result in the majority of the trials:

Total trials: 10,000

Fair division 6,198

4, 6 polymorphism 2,710

3, 7 polymorphism 919

2, 8 polymorphism 163

1, 9 polymorphism 10

(Skyrms 112, Note 23)

He noted that in further experiments, the more slices of cake available for division, the greater the chances for the population to evolve into the model of fair division. Thus, the tendency for a population to exhibit fair division is directly related to the granularity of the problem (Skyrms 13-16).

Skyrms goes on to use variations