Broome and others have pointed out, very much as Jim argues, that it's always possible to redefine options in ways that can save the rationality hypothesis, even in the face of the aporia Peter mentions below. In order therefore for the hypothesis to "have any bite", as Broome puts it, there must be constraints on the interpretation of preferences. In her book Natural Reasons, Susan Hurley runs with this and argues that these constraints are, inter alia, evaluative and intersubjective. This has remarkable implications: first, formal rationality is either tautologous or incomplete and in need of a substantive rationality which rules out certain preferences as irrational; second, preferences are not subjective, since they can only be individuated by an appeal to evaluative, intersubjective standards; and, third, values cannot be reduced to preferences, as in neoclassical welfare economics, since the former are implicated in the individuation of the latter. Hurley's work lays out, persuasively, a *very* thick rationality. On Wed, 11 Jan 1995, Peter.Dorman wrote: > I must register a disagreement with Jim Devine on the question of rationality > in neoclassical economics. I DO think their view is non-tautological. > Rationality is generally defined as adherence to Von Neuman/Morgenstern (or > Friedman/Savage) expected utility, where the decision-maker selects the option > that maximizes the dot product of probabilities and values of potential > outcomes (all present value, of course). There is a large literature on the > discrepancy between this model and real-world observed behavior, beginning > with the Allais paradox and extending to Kahneman & Tversky, Thaler, Frank, > etc. (Simon should also be mentioned.) My personal view is that the weakness > of the rationality assumption is one of the soft underbellies of neoclassical > theory, particularly since "near-rational" behavior leads to significantly > different outcomes (e.g. Akerlof & Yellen). One of my interests is the > intersection of "thick" (non-NC) rationality and strategic behavior in > repeated games. I find this useful for understanding the emergence (or > eclipse) of solidarity, etc. > > Peter Dorman >
