Comprehension and production of referential expressions across autism spectrum conditions

Resumen

Abstract Reference is a phenomenon both fundamental to communication and inherent to every act of language use. Autism spectrum conditions (ASC) are marked by problems of both communication and language. This thesis presents three studies of production and comprehension of reference in ASC. Study 1 profiled the use of referential devices in narratives as part of a broader profiling of narrative language in ASC. Studies 2 and 3 tracked comprehension of grammatically different referential devices in a picture book task (the ‘book task’) and an interactive task which specifically tested for the comprehension of definite determiners (the ‘box task’). Study 1 investigated autism without intellectual disabilities (ASC–ID, n=18) in comparison to typically developing controls (TD, n= 18) while in studies 2 & 3 recruitment was broadened to include the entirety of the verbal spectrum, including individuals with intellectual disability (ASC–ID, n= 34; ASC+ID, n=14) as well TD (n=34) controls and a control group of children with intellectual disability without ASC (ID, n= 9). Results of study 1 showed that individuals with ASC–ID more frequently misused referential devices that imply specificity in comparison to TD controls, while the quantitative distribution of Noun Phrase (NP) types did not distinguish groups. The broader grammatical profile showed differences in grammatical complexity as well as rate of lexical errors while mental state content did not differ. Study 2, the book task, showed a relative weakness in comprehension of inflectional pronouns and clitics among participants with ASC. While the differences between the ASC–ID and TD groups only approached statistical significance, both ASC groups(+/-ID) showed a similar pattern of having significantly poorer comprehension of pronouns/clitics in comparison to non- anaphoric reference types, a pattern not found in the TD controls or the control group with ID without ASC. The ASC+ID group showed a significantly poorer comprehension of inflectional pronouns/clitics in relation to both the ASC–ID group as well as the ID group, suggesting that such increased difficulties may not be explainable by presence of intellectual disability alone. Study 3, the box task, assessed the comprehension of definite anaphora in an interactive task using pairs of objects that belonged to the same kind category. In this task, the participant and experimenter interacted with an object (say, a ball) and then, after a moment of having their eyes closed, the participant pulled an object out of a box (either the same object or a noticeably different objects belonging to the same kind category, e.g. a different ball). The test condition assessed whether the participant interpreted correctly that in this context the definite-anaphoric NP the ball necessarily links to the ball that was interacted with previously rather than a novel exemplar of the same kind. The results showed that the ASC–ID group had a significantly poorer performance in the definite-anaphoric condition in comparison to TD controls. The ASC+ID group performed significantly worse than the ASC–ID group, yet similar to the ID group. These results show across comprehension and production that handling referential devices is an area of vulnerability in autism against the background of a broader spectrum of linguistic impairments.