A Quantitative Analysis of Pragmatic Language in Adults with High-Functioning Autism

Background: Individuals with autism spectrum disorder often have impaired social communication (DSM-IV-TR; APA, 2000). These deficits, especially in pragmatics (i.e., the rules of conversation) have also been observed in individuals with high-functioning autism (HFA), who possess IQ and verbal skills within the normal range (Baron-Cohen, 1998). While many researchers have observed pragmatic language deficits in individuals with autism, few studies have been conducted in a conversational context.

Objectives: We aimed to identify differences in pragmatic language in conversations between typically developed individuals and individuals with HFA; we hypothesized that the HFA group would have fewer discourse markers, less supportive interruptions, and more off-topic responses compared to their typically developed peers.

Methods: In two studies, HFA (n=8) and typically developed (n=8) individuals participated in guided discussions that were recorded and transcribed. The data from Study 1 has been analyzed while the data from Study 2 is still being coded.

Results: Discourse markers are words that maintain grounding and listener understanding in conversations (Fox Tree, 2006). For example, the word like indicates ambiguity (e.g., they vary in price from five to like twenty cents) (Fox Tree, 2006). Usage of the discourse marker like was analyzed; preliminary results from Study 1 suggest that typically-developed adults were more likely to use like(n=100 times/30 minutes) compared to the HFA adults (n=2 times/60 minutes). Interruption events were analyzed for the outcome (i.e., who won the turn after the interruption), and categorized for their (non-)supportive role in conversation. In HFA speech, the interrupted person regained the turn in 34% of interruption events and the person interrupting regained the turn in 24% of interruption events, whereas in conversations of typically-developed participants, the interrupted person regained the turn in 49% of interruption events, and the person interrupting regained the turn in only 8% of the interrupted events. Moreover, over half of the HFA interruptions were non-supportive, commonly shifting the conversation away from the previous speaker, while over half of the interruptions were supportive in the typically-developed conversations. Lastly, participant responses were coded as appropriate or inappropriate to the topic. Results indicate that the HFA group was off-topic over five times more frequently than the typically-developed group.

Conclusions: These language patterns provide useful information about conversational interaction in adults with HFA, and have potential use for developing interventions to improve social communication skills in these individuals, fostering successful educational and occupational experiences as well as friendships and social competence.