Page 28 - Encyclopedia of Philosophy of Language

Introduction
conversational implicatures, developed in his William James Lectures of 1967, that any systematic attempt was made to identify principles underlying the whole class of phenomena of this supposedly nonsemantic kind.
Grice (1989) proposes a distinction between (a) what is said in an utterance, as determined by the semantic properties of the words uttered, (b) what is conventionally implicated, those implications which although not strictly semantic can nevertheless be drawn from the conventional meanings of the words, and (c) what is conversationally implicated, those impli- cations which arise not from conventional meaning but from certain general features of discourse.
Imagine the following exchange, drawn from Grice's discussion:
A. I am out ofpetrol.
B. Thereisagarageroundthecorner.
According to Grice, there is a conversational implica- ture in B's reply that B thinks the garage has petrol, is open, etc., even though that is not strictly part of what B said, nor a consequence of the meanings of the words. Grice suggests that such implicatures arise from a tacitly accepted 'cooperative principle' gov- erning conversation, in conjunction with con- versational maxims of the kind be as informative as is required, say only what you take to be true,be relevant, be perspicuous. Conversational implicatures charac- teristically arise where these maxims appear to have been flouted but have not in fact been flouted; on the basis of that assumption a hearer is forced to construe the utterance (what the speaker intended) such that it conforms to the cooperative principle.
Two prominent applications for the idea of con- versational implicatures have been to the meaning of logical connectives and to presuppositions. Grice (1989) seeks to explain the apparent divergences between the truth-functional definitions of the logical particles (&, v , z>, ~ ) and the meanings of and, or, if ... then, and not by appeal to implicatures. His general strategy is to relocate the distinctive features of the natural language particles from semantic con- tent to pragmatic conditions of context-based speech. Thus, for example, the temporal connotation of and (with the implication and then) is not, he argues, part of the semantics of am/but arises from the orderliness of discourse. The second application, to pre- supposition theory, is associated with a more general program to define presuppositions independently of truth-conditions. A standard account of pre- supposition is that if p presupposes q then q is a necess- ary condition for the truth-or-falsity of p and the negation of p also presupposes q. A radical proposal is to supplant this conception and explainthe phenom- ena to which it is standardly applied with the twin notions of logical entailment and conversational implicature (see Kempson 1975). This serves, as in the
case of the logical particles, to sustain a relatively uncomplicated truth-conditional semantics. Needless to say, these proposals remain highly controversial and face serious difficulties in their implementation. For an indication of some of the difficulties and fur- ther applications, see Davis (1991).
Philosophers of language have explored other linguistic phenomena whose location within the prag- matics/semantics axis is problematic. One such is metaphor (see Metaphor), which has been the subject of more or less every standard approach, from truth- conditional semantics to speech acts to pragmatic implicatures (Ortony 1979). Another is fictionality. which can be viewed either as a kind of utterance, subject to speech act conditions (Searle 1979), or as a degenerate case of reference (Donnellan 1974), or indeed in other ways besides.
For more detail on Grice's theory of implicature, see Grice, H. P.; Conversational Maxims; Cooperative Principle. On presupposition theory, see Pre- supposition and Presupposition, Pragmatic. Also Metaphor; Fiction, Logic of.
3. Reference
The final major area of concern within philosophy of language is reference in its many forms. One issue which reflects, though is not entirely coincident with, the debate between semantics and pragmatics is whether reference is best understood as a relation between symbols and objects or between speakers, objects, and hearers. According to the latter view, it is speakers who refer to things, while in the former it is expressions in a language which refer (or denote).
3.1 Definite Descriptions
This issue is most famously associated in philosophy of language with the debate over Russell's Theory of Definite Descriptions. According to Russell, definite descriptions of the kind thefather of Charles II or the fastest man on earth should not be treated logically as naming expressions but rather as 'incomplete symbols,' which acquire meaning only in the context of a proposition and can be contextually paraphrased such that in a fully analyzed sentence they give way to quantifiers and predicates. Thus the sentence The father of Charles II was executed analyzes into There is one and only one person who begot Charles II and that person was executed. Russell's account was praised as a 'paradigm of logical analysis' and indeed it neatly solved several problems for the formal representation of referring expressions, not least by snowing how meaningfulness could be retained for sentences con- taining definite descriptions which failed to refer (the present King of France, and so on). However, it came under attack, notably by P. F. Strawson (see essays in Strawson 1971), for treating reference as a property of expressions rather than as something that speakers perform in an utterance characteristically for the pur-
6