Demonstratives, Definite Descriptions and Definiteness∗

Kaplan (1989) showed us that demonstratives are special for two reasons. First, they are connected to the context: the “canonical” uses of demonstratives (to adopt Roberts’s (2002) terminology) are associated with demonstrations, extralinguistic acts that determine the reference. Second, they’re a particularly clear case of constituents with what Kaplan calls “prepropositional” content – that is, the content of a demonstrative, and in particular, the associated demonstration, doesn’t interact directly with the main assertion. For example, the referent of the demonstrative doesn’t seem to vary across worlds, even when we explicitly consider worlds in which the circumstances of the context are varied: (1) [Pointing at John throughout] If John and Mary switched places, that person would be a woman. (2) [Pointing at John throughout] If John and Mary switched places, the person I’m pointing at would be a woman. (Kaplan 1989; Roberts 2002) Example (1) is false, while example (2) is true for many speakers. So even though the phrase that person can be paraphrased as the person the speaker is pointing at, the two phrases don’t mean the same thing. The definite description can take narrow scope in this sentence, and the demonstrative appears to be scopally inert, taking widest scope only. This is the sort of fact that led Kaplan to argue that demonstratives have direct reference: that is, the referent of the demonstrative phrase is “loaded directly into the proposition,” and the content of the demonstrative does not play a direct role in the main assertion. ∗I am grateful for insightful comments from Sandy Chung, Donka Farkas, Michela Ippolito, Jim McCloskey, James Isaacs, Line Mikkelsen, the members of the UCSC Semantics Interest Group, Fall 2003, and the audience at NELS 2003. 1Those speakers who do not accept (2) appear to be interpreting the definite description as the person I’m pointing at now ; that is, they take the present tense to be indexical. If the present tense in the description requires the description to refer to something in the actual context, then of course the referent will not vary across worlds. Unlike the scopal inertness of demonstrative descriptions, however, the scopal inertness of the definite description in (2) is not an intrinsic feature of definite descriptions. As far as I know, all speakers judge (i) to be true in the relevant context, showing that definite descriptions can in principle take narrow scope for all speakers.