Investigating Language Processing of Social and Indirect Communication During Reading in An Autistic and Non-Autistic Population.

Abstract

Autism spectrum condition (ASC) is characterised by impaired social communication as well as interaction. Individuals with ASC have been reported to have difficulties in processing figurative language when compared to non-autistic controls, suggesting autistic people may have difficulty extracting hidden meaning. Using a combination of eye-tracking and reading time data two main aspects of socially motivated language processing were investigated: (1) indirect replies and (2) potential processing strategies used in order to understand socially motivated dialogue, specifically integrative (bottom-up) and predictive (top-down) accounts. Also investigated was the influence of individual variability on processing. Results from Studies 1-5 suggest autistic readers are sensitive to the types of social communication and processing strategies often labelled 'impaired' in autistic individuals. In two reading time studies (Chapter 2), indirect replies were investigated in an autistic and non-autistic sample. Results suggest that both autistic and non-autistic readers are sensitive to politeness conventions that mediate the use of indirect replies in order to convey negative meaning. When an indirect reply ('The exams are not fair') is uttered, both autistic and non-autistic readers, find positive contexts (Mary had done well on the exam) difficult to comprehend compared to negative contexts (Mary had done poorly on the exam) indicated by significantly longer reading times for positive conditions. This suggests autistic individuals are not only sensitive to the face-saving function of an indirect reply but perform comparable to non-autistic controls. Studies 3 and 4 (Chapter 3), investigated the likely processing strategies involved when parsing written dialogue. Using reading times, both autistic and non-autistic readers show a preference for bottom-up over top-down processing strategies. This suggests that readers regardless of group are not pre-activating any semantic representation of the likely reply to a question based on contextual cues that facilitate prediction. There was marginal evidence to suggest autistic readers may be masking any predictive effects in the non-autistic population. Converging evidence from Study 5 (Chapter 4), confirms this may be the case. Using eye-tracking, in the non-autistic population we found clear evidence that contextual information that facilitates prediction is utilised quickly and prior to the replies encounter with the visual system. Additionally, the contextual information continues to facilitate processing of written discourse aiding comprehension via both top-down and bottom-up strategies. Results from Studies 1-5, indicate that individual variability associated with language ability is predictive of overall reading times as well as regressive eye-movements. The reported thesis aims to firstly address theoretical and experimental areas of interest but also to combat the 'ableist' language present in current research. Overall the reported thesis, provides firstly clear evidence that individuals with a diagnosis of autism are sensitive to unconventional figurative language use (indirect replies). This is inconsistent with both ASC diagnostic criteria and previous research overemphasizing impairments. We suggest integrative processing strategies better account for the way in which autistic individuals process socially motivated language. Unlike non-autistic readers, who are happy to make predictive judgments based on contextual information, that may turn out to be wrong- it seems ASC readers require more certainty when parsing dialogue and therefore rely on a more time-consuming integrative process to understand the language presented. Importantly, this research provides convincing evidence that phrasing group differences as 'deficits' does not benefit the clinical community under investigation nor the academic advancement of knowledge and oversimplifies what proves to be a complex and often misr

Date of Award	1 Aug 2024
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Emma Gowen (Supervisor), Andrew Stewart (Supervisor) & Jennifer Mcbride (Supervisor)