Restricted Quantification

Abstract

In natural language you might meet a sentence like

• (1) Everybody is present. Now someone who utters (1) does not, as a rule, mean that everybody in the world is present. Presumably they mean that everybody in some intended set of people is present. But which set? Well, that is presumably to be determined by the context of utterance. At any rate one should not expect the semantics of everybody to tell you what set is intended. What you can expect the meaning of everybody to tell you is that the sentence is universally quantifying over the intended set. If the intended set can be specified, say it is the set of students enrolled in PHIL 783 at UMass in the fall of 1991 (for short ‘enrolled students’) we can replance (1) by

• (2) Every enrolled student is present. For that reason it may be possible to take a hard line about (1). Say that it really does mean that everybody in the world is present, and so it is false; but that what the speaker intended to say was (2), and we charitably assume that it was (2) that was ‘really’ uttered. Even if a hardline is possible it is surely unintuitve. Notice that in (2) I used ‘enrolled student’ as an abbreviation. The hardliner would have to say that practically none of the sentences we utter say what we mean, and the sentences which do will have to be enormously complicated in filling in all the contextually supplied information. At any rate I shall be looking at ways in which context can affect interpretation even though there may be no syntactic indication of the dependence. This work grows out of material in Cresswell 1990 and consists in putting together two ideas.