I. Discussion Question for Monday, Feb. 9.

On Monday, we will play a fictional n-person PD supergame. Each member of the class will play the role of a farmer in a small farming community with no government institutions. It will be common knowledge that there is a potentially indefinite series of public projects that would further the public good (e.g., dikes, roads, pest control, etc.). For convenience, we suppose that the individual benefit to each farmer from any individual project is 2; and that the total cost of any individual project is n; so that the cost of any project to each contributor depends on the number of contributors m, and equals n/m. Note that if less than half of the farmers contribute in any particular subgame, those who contribute suffer a net loss and those who do not contribute have a gain of 2. If more than half of the farmers contribute in any particular subgame, both the contributors and the non-contributors will be better off than if none had contributed, though the non-contributors will do better than the contributors. (Thus, Schelling's k = [(n/2) + 1].) Come to class prepared to discuss which strategy you (as one of the farmers) would adopt for the supergame. For example, there are the unconditional strategies C- and D-; and there is a family of conditionally cooperative strategies Bi = Cooperate on the first subgame; in each succeeding subgame, cooperate if at least i other players cooperated in the preceding subgame, otherwise defect. (Note that because i refers to the number of other cooperators, it can be any number from 0 to (n-1).) You are not limited to these strategies. You are free to devise one of your own. Come to class with a strategy that you want to begin with. You are not required to stick with your initial strategy. You will be given an opportunity to change your strategy after you find out what strategies the other players favor.

II. Homework Assignment #5, Due in Class on Wed., Feb. 11:

1. (a) On pages 119-120 [Reader pp. 78-79], Elster argues that some sanctions must be performed for motives other than the fear of being sanctioned. He expresses his conclusion by saying that chains of sanctions require an unmoved mover. What does he mean by these claims?

2. What are the empirical premises of his argument for the claims stated above?

3. Are the empirical premises always true for human groups? To help answer this question, consider the following example: Suppose that a group, P (for "patriots") regards another group E (no member of P is also a member of E) as its enemy. The members of P could define "traitor" inductively as follows: First, a level-1 traitor is any member of P (the patriots) who does business with or provides aid to a member of E; a level-2 traitor is any member of P who does business with or provides aid to a level-1 traitor; etc. Suppose that there is full information, so that all traitorous acts of every level are known by all. Explain why Elster's assumptions would imply that there is some relatively low number n, such that patriots would not be motivated to sanction traitors of level-n or above. Is this necessarily true? If your answer is negative, explain how one or more of Elster's empirical assumptions might fail to hold in this sort of case. If your answer is affirmative, explain why the patriots would not be motivated to sanction traitors of level-n or above, even when something that the members of P value very highly (e.g., their liberty or their survival as a group) depends on defeating the enemy E.