AN EPIsTEmologIcAl APProAch To EssENTIAl INdExIcAlITy

The prevailing notion that the problem of essential indexicals must be solved through the theory of meaning of attitude ascriptions is incorrect. Well-known attempts to solve the problem along those lines, e.g., the proposals of lewis (1979) and Perry (1979), have rested on the overly optimistic assumption that there is no limit in principle to the access one may have to the contents of someone else’s thoughts, including their knowledge. That assumption is challenged in this essay. The hazards associated with limited accessibility can be and indeed must be alleviated by an epistemological theory in such a way that allows for a new nonsemantic approach to the problem. The following is an outline of the arguments in favor of this approach. In the first part, the problem of essential indexicals and the familiar corollary thesis that such indexicals are not reducible to nonindexicals is briefly reviewed. In the second part, indexical irreducibility is shown to hold for a range of explanations of theoretical rationality in addition to the basic explanations of action in which the irreducibility is typically illustrated. since these explanations must rely on normative features of a background epistemological theory, a minimally adequate epistemology must be able to support such explanations. This is the first premise of the central argument for the thesis that an adequate epistemology accepts limited accessibility. That argument is completed in the third part by establishing that no semantic alternative to the propositional attitude analysis of belief and knowledge offers a real escape from limited accessibility. This includes a detailed criticism of two prominent types of alternatives to the propositional view. In the fourth part, some of the epistemological implications of accepting limited accessibility are described. Tensions between limited accessibility and assumptions about the possibility of omniscience and public defensibility as necessary conditions of knowing are shown to be avoidable on particular epistemological theories, such as recent forms of foundationalism and coherentism. Although they may face other unrelated difficulties, such theories may allow for limited accessibility without loss of plausibility for that reason. All of this is in support of the point that the absorption of limited accessibility into an adequate epistemological theory yields a new resolution to the problem of essential indexicals.