Irony-understanding and theory of mind. A conceptual and experimental inquiry from a realist perspective

Abstract

The main aim of this work is to investigate the prospects for a realist view of irony that take ironic assertions to concern ¿worldly¿states of affairs and individuals ¿just as literal assertions do¿, as opposed to an attributive conception (adopted, for instance, bythe relevance theoretic echoic account (Sperber 1984; Sperber & Wilson 1981, 1995 [1986]; Wilson 2006, 2009; Wilson &Sperber 1992, 2012) and the pretense theory¿Clark & Gerrig 1984) that claim that ironic utterances always concern utterances orbeliefs. The adoption of one view or the other has consequences for the selection of the target/victim of the negative attitudegenerally associated with irony, and the cognitive abilities presupposed in the understanding of ironic assertions. First, for a realistview, the speaker adopts a negative stance towards the state of affairs or individual the ironic utterance concerns; for an attributiveone, the speaker always criticises the utterance or belief the ironic utterance is about and the utterer or thinker of it. This leadssometimes to the wrong prediction of target/victim. Second, realist irony-understanding should require a lower-order of the theoryof mind ability ¿ToM¿ (the ability to attribute and reflect about beliefs of others) than attributive irony-understanding, since, inattributive instances, the hearer has to attribute to the speaker a belief about another belief or utterance. On an initial conception ofthe realist view, realist and attributive were taken to be complementary and mutually exclusive types of irony.In order to assess the existence of realist irony, I conducted two empirical studies (in collaboration with Elise Clin andMikhail Kissine). Study 1 was done with autistic participants, a population that have their ToM ability impaired and should,supposedly, have fewer problems to understand realist irony than attributive. Study 2 was done with a larger sample in order to testthe effect of traits normally associated with autism and of different languages on the understanding of irony. The results of theempirical studies did not support the conception of the initial realist view, but a stronger one. The data better fit the view thatattributive and realist irony are not two different types of irony, but that irony is always realist, and that attributive irony is a subtypeof it. The results have also served to clarify a misguided assumption in the literature, that irony requires a high-order ToM thanliteral assertions. Understanding both literal and ironic assertions requires the attribution of complex beliefs and intentions to thespeaker. The difference in difficulty between ironic and literal assertions seems to lie in the fact that the speaker of ironic assertionsdoes not express the belief that corresponds to the truth-condition of the utterance, and that she necessarily adopts of a negativestance.