Pragmatics for Propositional Attitudes

Suppose sentences (1a) and (1b) are true. What is the truth value of (1c) under the assignment that maps x to the individual d which is Cicero and Tully? As Quine observes, the ordinary notion of belief seems to require that although (1c) holds when x is specified in one way, namely as Cicero, it may yet fail when the same x is specified in some other way, namely as Tully. Classical modal predicate logic, in which variables are taken to range over plain individuals, fails to account for this ordinary sense of belief. In a possible world semantics, these ‘ways of specifying objects’ can be characterized by means of the notion of an (individual) concept, i.e. a function from the set of worlds to the set of individuals. Many authors have observed that if we let variables range over concepts rather than plain individuals, we manage to account for Quine’s intuition about example (1). An obvious problem arises though if we let variables range over all concepts. The following classical example due to Kaplan illustrates why. Suppose Ralph believes there are spies, but does not believe of anyone that (s)he is a spy. Believing that spies differ in height, Ralph believes that one among them is shortest. Ortcutt happens to be the shortest spy. The de re interpretations represented in (2b) and (3b) are deviant in this situation.