Semantic Category and Surface Form

Abstract

There is in philosophy of language a well known conflict concerning how to account for the meaning of expressions of the form 'the F'. On one side are those who adopt a 'unified' stance, believing all definite descriptions to be handled by a quantificational analysis;1 while on the other are those who think some definite descriptions must be handled as referential terms and who thus maintain that a single surface form may be correlated with two different semantic categories.2 Those who propose the latter approach, which I will call an 'ambiguity' account, might think to reinforce their claim by pointing to other types of noun phrase where a similar bifurcation between surface form and semantic category is called for. Prima facie, the most powerful evidence for ambiguity seems to come from the existence of 'anaphoric' terms: these are expressions which exist alongside potentially referential (or deictic) occurrences of pronouns and demonstratives, yet which seem to demand a semantic analysis akin to that given for the bound variables of quantification. So, the ambiguity theorist may argue, since we allow a 1:2 mapping between surface form and semantic category as far as deictic and anaphoric terms are concerned, we should be happy to allow the same for definite descriptions. My aim in this note, however, is to show that the premise here is mistaken: anaphoric expressions do not provide any evidence of disunity because they do not, in fact, share a single surface form with referential terms. To see this, let us begin by looking more closely at what anaphoric expressions are and why they might be thought to provide evidence for ambiguity, before looking at the kind of surface features such expressions display. An anaphoric context is one where the expression in question appears in a position which is bound by an earlier occurring quantifier phrase, such