Perry argues that certain sorts of indexicals are 'essential', in the sense that they cannot be eliminated in favor of descriptions. This paper also introduces the influential idea that certain sorts of indexicals play a special role in thought, and have a special connection to action.

Frege and quine notwithstanding, Some singular terms in belief contexts have normal reference but do not admit truth-Preserving substitution of co-Referential terms. The conditions of a sentence's being true of a sequence of referents may be partially determined by its singular terms; substitution may change those conditions, While preserving genuine reference. On one reading, 'n believes that f is g' is true iff n believes of the f that it is the f and is g.

Many different analyses of the concept of de re belief have been proposed in recent years. Most of these analyses may be called ‘reductionist’ since they attempt to “reduce” de re belief to de dicta belief or to analyze de re belief in terms of de dicta belief. Some reductionist analyses are extremely liberal in their attribution of de re beliefs — they imply that people have de re beliefs in a variety of situations in which more restrictive analyses have (...) no such implication. In this paper I will show that the most liberal of the reductionist theories, those Roderick Chisholm calls “latitudinarian theories”, are unacceptable.Latitudinarian analyses have been proposed by many philosophers, including Ernest Sosa, Mark Pastin, and also, perhaps, W. V. Quine and Wilfrid Sellars. (shrink)