The semantics of complex demonstrativesa defense of character theories

Resumen

This dissertation defends a traditional approach to the semantics of complex demonstratives, i.e., expressions like “this table” or “that man”. On the classical theory, complex demonstratives work just like any other indexical: pretty much like “I”, “she” or “here”, they are context-dependent terms endowed with a descriptive meaning, their so-called “character”, that helps fix their referent in a context but plays no truth-conditional role once this individual has been selected. Their content or truth-conditional contribution, in turn, is simply an object. This model of complex demonstratives was justified insofar as these expressions behave just like any other indexical in intensional environments and in phenomena like anaphora and ellipsis. Recent discussions, however, called into question the traditional picture on the basis of data about non-deictic uses of complex demonstratives. These include cases in which the nominal of a demonstrative is operated on by an external quantifier, instances of discourse anaphora, transparency under modals, Bach-Peters sentences, readings in which they take narrow scope under a quantifier, and constructions in which they function as bound variables. These problematic uses of complex demonstratives led many authors to develop alternative approaches that differ greatly from the traditional one, often viewing these expressions as quantifiers or as a special type of definite description. In this dissertation I argue that an independently justified revision of the classical theories of complex demonstratives suffices for accommodating all the counterevidence above. Namely, the required adjustments are two: the acknowledgment that natural language may contain hyperintensional operators (what Kaplan called “monsters”), and the adoption of a system with explicit quantification over times and worlds. The former idea is required by a proper treatment of variable binding in natural language, whereas the latter renders tense and mood on a par with pronouns – i.e., as variables that can be bound. Once these adjustments have been implemented, the above problems vanish. In the first place, it is no surprise that operating on the nominal of a demonstrative should result in non-deictic uses; after all, such operation implies meddling with the character – or hyperintension – of the expression. The same goes for bound variable uses of complex demonstratives. In the second place, the above-mentioned treatment of tense and mood bears an interesting result: complex demonstratives containing tensed relative clauses are also subject to external binding by modal operators. As a consequence, non-deictic demonstratives under modals can be reduced to a case in which the character of the expression is also being meddled with. Discourse anaphora is more complex, but it is no more problematic for complex demonstratives than it is for any pronoun, for it can be treated as a mere case of co-variation without c-commanding. In this dissertation I consider several solutions that have been put forward for dealing with this problematic type of anaphora, and I apply them to the particular case of complex demonstratives. Bach-Peters sentences, in turn, can be treated as a combination of both co-variation without c-commanding and operations on character. The central thesis of classical theories can therefore be retained: complex demonstratives should be understood on the model of pronouns