Speaker Intentions in Context

Speaker Intentions in Context Jeffrey C. King Rutgers University It is generally believed that natural languages have lots of contextually sensitive expressions. In addition to familiar examples like ‘I’, ‘here’, ‘today, ‘he’, ‘that’ and so on that everyone takes to be contextually sensitive, examples of expressions that many would take to be contextually sensitive include tense, modals, gradable adjectives, relational terms (‘local’; ‘enemy’), possessives (‘Annie’s book’) and quantifiers (via quantifier domains). With the exception of contextually sensitive expressions discussed by Kaplan [1977], there has not been a lot of discussion as to the mechanism whereby contextually sensitive expressions get their values in context, aside from vague references to speakers’ intentions. My main task here will be to propose a candidate for being this mechanism and defend the claim that it is such. Because, as I suggested, these issues have been most extensively discussed in the case of demonstratives, I’ll focus on these expressions by way of contrasting the mechanism I will propose with others in the literature. However, as I’ll discuss further below, the reader should bear in mind that I am inclined to think that the mechanism I’ll discuss applies in the case of (almost—see below) all contextually sensitive expressions. At least since the work of David Kaplan [1977], many philosophers and linguists distinguish two kinds of contextually sensitive singular terms. 1 On the one hand, there are pure indexicals: expressions that take on different values in different contexts and whose conventional meanings by themselves suffice to secure values in contexts. One could argue over particular cases, but ‘I’, ‘here’, ‘now’, ‘today’ and ‘tomorrow’ are often