W

What Is Said Within the classical Gricean paradigm, a distinction is made between what is said and what is implicated (e.g. Grice 1989: 25). However, as pointed out by Levinson (2000: 170), Grice’s characterization of what is said is quite complex, though it may roughly be presented as follows:

(1) Grice’s concept of what is said U said that p by uttering x if and only if: a. x conventionally means p b. U speaker-meant p c. p is the conventional meaning of x

minus any conventional implicature

where U stands for the speaker/utterer, p for a proposition, and x for a linguistic expression. Given the definition in (1), what is said is

generally taken to be (i) the conventional meaning of a sentence uttered with the exclusion of any conventional implicature, and (ii) the truth-conditional propositional content of the sentence uttered (e.g. Grice 1989: 25; Levinson 1983: 97; Levinson 2000: 170; Neale 1992: 520-21; Clark 1996: 141). According to Bach (2004b), the Gricean

notion of what is said is needed in order to account for three kinds of cases. In the first, the speaker means what he or she says and something else as well, as in conversational implicatures and indirect speech acts. The second of these cases is where the speaker says one thing and means something else instead, as in nonliteral utterances such asmetaphor, irony and hyperbole. Third and finally, there is the case where the speaker says something and does not

mean anything at all by it. Cases in point may include translating, reciting or rehearsing in which the speaker says something with full understanding but does not use it to communicate (Bach 1994a; see also Huang 2007). On Grice’s (1989: 25) view, before one works

out what is said, one has (i) reference to identify, as in (2); (ii) deixis to fix, as in (3); and (iii) ambiguity and ambivalences to resolve, as in (4).