Abstract

Richard Jeffrey's Logic of Decision has come under fire on the grounds that it appears to prescribe irrational decisions under certain circumstances . A number of authors see the source of Jeffrey's difficulty as a lack of sensitivity to causal distinctions of a certain kind. They have proposed modifications of Jeffrey's theory to overcome this putative deficiency. David Lewis argues, convincingly, that these modified theories are all more or less the same. In essence, they all augment the Logic of Decision by the explicit introduction of causality. There is warrant, then, to classify them under the common heading of causal decision theory. I make use of Brian Skyrms' formulation; but my remarks apply to all variants. ;I give version of the argument which leads the causal decision theorists to reject Jeffrey's theory. I proceed, then, to modify this argument so as to yield what, at first glance, appears to be a counterexample to causal decision theory itself. Skyrms responded to this example in a way which brings into sharp relief both the causal decision theorists' plaint, and the essence of my defense of Jeffrey. There emerges, I claim, through a careful analysis of choice, action, and degree of belief, a clarification of the decision-theoretic enterprise, which, if I am correct, shows both Jeffrey's theory and causal decision theory to be prescriptively adequate; so that we then have no need of causal decision theory's encumbrances. I reject the "tickle" defenses. ;After addressing the prisoners' dilemma and Newcomb's paradox, I move on to discuss the descriptive pretensions, both of normative theories of rationality as a whole, and the Logic of Decision in particular, within the context of Donald Davidson's interpretative perspective.