‘Pragmaticizing’ Kaplan: Flexible Inferential Bases and Fluid Characters

Abstract

It can easily be observed in analysing natural language discourse that the category of an indexical expression does not correspond to clearly delimited types of expressions: in spite of the philosophers’ rigid distinctions into directly referring expressions and expressions whose referring function is present on only some occasions of use (where, for example, indexicals and proper names belong to the first category and definite descriptions to the latter), the devices that natural languages use for the purpose of conveying reference cannot be easily put into this mould. At the same time, these rigid classifications, such as those informing Kaplan's content/character distinction, rely on a formal semanticists’ dubious assumption that expressions can be identified and classified prior to the function they have in a particular context. So, for example, for Kaplan, personal pronouns have characters (sometimes referred to as ‘linguistic meanings’) that are in need of contextual specification before the truth-conditional semantic analysis (in terms of a function from possible words/circumstances of evaluation to truth values) gives them content. However, what counts as character is arguably itself context-dependent: the unit that is relevant for such a contextual specification is in itself dependent on the particular situation of discourse, and even, arguably, on the speaker's intentions. It is therefore proposed that Kaplan's character/content distinction be ‘pragmaticized’ accordingly, to apply to such a flexible, use-driven category. In short, characters become ‘fluid’ because they correspond to ‘flexible’ inferential bases.