Underinformative Implicature Derivation on the Broader Autism Phenotype

Abstract

Underinformative implicatures (UIs) are a type of inference that can give rise to two potential meanings. UIs can be either literal, which do not require the use of social inferences, or pragmatic, which necessitate social inferences. Given that individuals with autistic traits have marked deficits in communicative and social skills, there is reason to believe that these individuals approach deriving UIs differently than typically developing (TD) individuals. The current study is an attempt to replicate van Tiel and Kissine's (2018) online assessment of how individuals on the broader autism phenotype (BAPs) derive these inferences. The present study utilizes a previously untested technique to assess UI derivation; it employs a graded scale for participants to rate the appropriateness of stimuli rather than a binary response option as all prior research had done (Pijnacker, Hagoort, Buitelaar, Teunisse, & Geurts, 2009; Chevallier, Wilson, Happé, & Noveck, 2010; Su & Su, 2015; Van Tiel & Kissine, 2018). The current study found significant differences between how all BAPs and typically developing (TD) participants respond, but not between how low scoring individuals on the BAP (Low BAPs) compared to high scoring individuals on the BAP (High BAPs) based on autism spectrum quotient scores. The present study found nearly the exact opposite of what van Tiel and Kissine (2018) found in their study. Instead of only High BAPs responding more literally on distributivity implicatures, all BAPs responded more pragmatically on all implicature types except for distributivity. In addition, these results showed evidence of cognitive inflexibility via the tendency to answer using the ends of the scale on lexicalizable implicature types. The current study concludes that there are indeed significant differences in how BAPs respond to UIs in comparison to TDs. It is possible that such differences were revealed through the employment of a graded rating scale. These results are highly relevant to making educational spaces more linguistically inclusive for neurodiverse populations.