Findings from five experiments support the view that negation generates sarcastic utterance-interpretations by default. 1 When presented in isolation, novel negative constructions ("Punctuality is not his forte," "Thoroughness is not her most distinctive feature"), free of semantic anomaly or internal incongruity, were interpreted sarcastically and rated as sarcastic compared to their novel affirmative counterparts (Experiments 1 and 3). In strongly supportive contexts, they were processed faster when biased toward their noncoded (nonsalient) sarcastic interpretation than toward their noncoded but (salience-based) literal interpretation (Experiments 2 and 4). Experiment 5 reduces the possibility that it is structural markedness rather than negation that prompts nonliteralness. Such findings, attesting to the priority of sarcastic interpretations, are unaccountable by any contemporary processing model, including the Graded Salience Hypothesis.

Sarcastic irony, uttered in four (within and between) gender-based settings, is used here as a tool to diagnose affective attitudes toward women. The kind of sarcasm tested here is an aggressive type of humor, whereby a speaker derides another individual, turning her or him into the victim of the humorous utterance. Finding this kind of irony less or more pleasing allows indexing between-and within-group attitudes. Participants were overall nonsexist, scoring low on sexism scales, but male participants were still more sexist than female participants. Results show that, as predicted by Ariel and Giora (1998), female participants fully adopted a feminine point of view, enjoying sarcastic irony best when it was directed by women at men and least when it was directed by women at women. Being more sexist, our male participants adopted a feminine point of view only partially, enjoying sarcastic irony more when directed at men than directed at women, regardless of the speaker's gender.

Based on natural language use, we examine the contextual environment of some negative constructions (e.g., Punctuality is not her forte/best attribute). Previous findings show that, as predicted by the view of default nonliteral interpretations, such negative constructions are interpreted nonliterally by default: (a) when presented in isolation, they are interpreted sarcastically and rated as sarcastic compared to affirmative counterparts; (b) when embedded in equally strongly biasing contexts, they are processed faster in sarcastically than in literally biasing contexts (Giora et al., 2013; Giora, Drucker et al., 2014). Here we test a third prediction that, unlike affirmative sarcasm, (c) such negative utterances will convey a sarcastic interpretation and their natural environment will echo their nonsalient (sarcastic) interpretation rather than their salience-based (literal) interpretation (Giora et al., 2010, 2013). Findings from 2 corpus-based studies of (Hebrew and English) negative constructions lend usage-based support to the view of default nonliteral interpretations (Giora et al., 2010, 2013; Giora, Drucker et al., 2014). They show that when occurring in natural discourses, such utterances communicate sarcasm significantly more often than their alternative affirmatives. Their neighboring utterances further reflect their nonsalient sarcastic interpretation rather than their salience-based nonsarcastic interpretation.

No abstract