Recent experimental studies appear to discredit Gricean accounts of irony and metaphor. I argue that appearances are decidedly misleading here and that Gricean accounts of these figures of speech are actually confirmed by the studies in question. However, my primary aim is not so much to defend Gricean accounts of irony and metaphor as it is to motivate two related points: one substantive and one methodological. The substantive point concerns something Grice suggests in his brief remarks on irony: that the interpretation of an ironical (vs. metaphorical) utterance requires two distinct applications of second-order theory of mind (ToM). I argue that such a view has considerable explanatory power. It can explain an intuitive contrast between irony and metaphor, some interesting data on the ToM abilities of patients with schizophrenia, and some intuitive similarities between irony on the one hand and hyperbole and meiosis on the other. The methodological point concerns the relationship between the empirical psychologist’s (or experimental philosopher’s) experimental studies and the armchair philosopher’s thought-experiments. I suggest that the credibility of an experimentally supported claim is enhanced when it captures the reflective judgments captured in the armchair philosopher’s thought-experiments.
See for instance Hintikka (1999) and Weinberg et al. (2001).
He later discussed the phenomena in his (1978) and (1989).
Grice (1975) claims that conversational implicatures may be “grasped intuitively” rather than “worked out”. Nevertheless such implicatures must according to Grice be capable of being worked out because otherwise they would not be conversational but conventional implicatures.
Grice (1975) uses this locution in describing cases such as irony and metaphor where the speaker does not intend to communicate the literal meaning of her utterance. For Grice to say something is to mean it. In what follows I use the more natural if less precise terminology according to which speakers can and sometimes do say things that they do not mean.
Searle (1979) in contrast does develop full-fledged theories of non-literal language comprehension based on Gricean principles. Martinich (1984) does as well.
Searle (1979) however does advocate something very much like the Gricean account.
Wilson and Sperber also cite Winner (1988) and Langdon et al. (2002).