Meeting of the Aristotelian Society held at Senate House, University of London, on 5 March 2007 at 4:15 pm. ©2007 The Aristotelian Society Proceedings of the Aristotelian Society, Vol. CVII, Part 2 187 VIII—EPISTEMIC DEFERENCE: THE CASE OF CHANCE JAMES M. JOYCE Epistemic deference is the phenomenon in which one person uses the de- liverances of some information source, perhaps the opinions of another person, as a model for what to believe. The paper aims to clarify the na- ture of epistemic deference in probabilistic contexts, to explain the condi- tions under which deference is appropriate, and to examine deference to objective chances, as epitomized in David Lewis’s Principal Principle. This latter analysis will show, in contrast with views that portray chance as an ideal inductive logician with total recall, that our deference to chance is grounded in contingent limitations on our ability to access in- formation and our recognition that the physical probabilities that instan- tiate the actual chances codify all the types of information that humans are able to possess. My broad topic is epistemic deference, the phenomenon in which one person uses the opinions of another, either a real person or some idealized information source, as a model for what to believe. Three questions will be addressed: How should we model epistemic deference within subjective probability theory? 1 What is it for a be- liever to regard another as being worthy of deference? Can we iso- late substantive conditions under which one believer should defer to another? In addressing the last question I will be considering the case of objective chance, which is often used as a paradigm for def- erence to an epistemic expert. I will conclude that, contrary to views that portray chance as a kind of super inductive reasoner, the basis of our epistemic deference to chance is primarily to be found in facts about our limitations as knowers. 1 I use the framework of subjective probability theory not because I think it psychologically realistic, but because it provides useful first step in evaluating claims about belief and justi- fication. Claims in general epistemology that are implausible in the idealized probabilistic framework are unlikely to hold up in more realistic contexts.