When is irony effortful?

Whereas some studies indicate that ironic-as opposed to literal-readings of utterances take longer to process, others indicate that the 2 are processed at comparable speeds. We propose that mindreading processes are at least partly responsible for the mixed results, as we present 3 experiments that include stories having a target utterance with either an Ironic or Literal reading. Experiment 1 replicates earlier findings (Spotorno, Koun, Prado, Van Der Henst, & Noveck, 2012) showing that ironic readings take longer than literal ones when fillers include decoys, stories that call for an ironic remark but present a banal utterance instead and thus diffuse participants' expectations for irony. Starting with Experiment 2, decoys are removed, leading to effects that are arguably revealing of Theory of Mind processes. One is an Early-Late effect, which occurs when ironic utterances are read as readily as literal ones in the 2nd half of an experimental session, creating "mixed" results in the laboratory. In Experiment 3, we further added antecedents before a critical event so that, later, the target utterance would be echoing an explicitly stated thought and would facilitate irony comprehension (Gibbs, 1986; Sperber & Wilson, 1981). Results reveal an Early-Late effect here, too. Further evidence of Theory of Mind activity follows from analyses of participants' Social Skill subscale scores in the Autism-Spectrum Quotient (Baron-Cohen, Wheelwright, Skinner, Martin, & Clubley, 2001). Socially inclined participants are more likely than the socially disinclined to use a story's negative event to portend the arrival of an irony; in contrast, socially disinclined participants appear more reactive than the socially inclined to explicit antecedents.