In a democracy, social choices are supposed to be made on the basis of the various preferences of the members of a society. There is some sort of analogy with choice among models and theories on the basis of various theoretical values such as fit to data, simplicity, scope and so on. In the Seminar we will explore this analogy.
Systematic thinking about social choice goes back through the French Enlightenment to classical times. It has reached a very high level of technical sophistication in the last sixty years or so, as a branch of theoretical economics. One landmark is Arrow's "impossibility" theorem (1951), which sets a theoretical limit to the possibilities for democracy. Other work in welfare economics has isolated specific circumstances under which, in the nature of the social alternatives and the determination of individuals' preferences among them, some of Arrow's assumptions are false and democratic social choice is possible anyway. Intriguingly, analogous circumstances arise in theory choice as well. When they do, there can be procedures for theory choice that are "balanced" in the sense that they give due importance to the various dimensions of theoretical value.
The point of the Seminar is that the analogy between social choice and theory choice is pay dirt in which we can hope to find procedures for rational theory choice. There is some excitement in the field about the possibilities, and plenty of work still to be done. This Seminar will be away to learn about what has already been accomplished and to gain the necessary background for further explorations.
We will begin with technical preliminaries having to do with the philosophy of measurement, and the logic of relations. After that we will have a close look at several theoretical values, concentrating I expect on various conceptions of accuracy, or fit to the data, and simplicity. After that, we'll turn to the theory of social choice itself, with a technically rigorous development of Arrow's Theorem, and of its "single-profile" variants. We will consider various "escape routes" from the theorem, such as enriching the informational basis for social choice and domain restrictions.
With this background in place, we will be in a position to consider choice in science from a social- choice theoretic perspective. My thinking about what to look at is still developing. Among the topics I expect we'll cover are: (a) tradeoffs among the theoretical values of fit to data and simplicity in the Akaike Information Criterion (AIC) and in the Bayesian Information Criterion (BIC); (b) the availability of adequate notions of aggregate theoretical value in best Systems accounts of Laws and Probabilities; and (c) the existence of "single-peaked" domains for theory choice in normal science, which implies that very simple choice algorithms are available there.
Arrow, Kenneth 1951: Social Choice and Individual Values. New York: John Wiley.
Arrow, Kenneth J., Amartya K. Sen and Kotaro Suzumura (eds.) 2002: Handbook of Social Choice and Welfare vol. 1. Elsevier: Amsterdam.
Black, Duncan 1948: “On the Rationale of Group Decision Making,” The Journal of Political Economy 56: 23-34.
Bossert, Walter and John A. Weymark 2004: "Utility in Social Choice." In Salvador Barberà, Peter Hammond, and Christian Seidl (Eds) 2004: Handbook of Utility Theory, volume 2. Dordrecht: Kluwer, pp. 1099-177.