Abstract

It is argued that an adequate model-theoretic treatment of belief requires the devices of a "two-dimensional" intensional logic; a logic, that is, that can represent the relevant features of the epistemic situation of an agent in something like the way a logic of indexicals represents possible contexts of disourse. The general approach is illustrated by means of several examples, Kripke's "Puzzle About Belief" among them. A model theory is developed for a formal language containing a belief operator, quantifiers, description operator and identity predicate, and a truth predicate and reference operator. Certain properties of the model theory are demonstrated, and the application to "indexical belief" is sketched