Social Engagement and the Pragmatics of Conversation in Autism

Background:

Socio-cognitive approaches to language acquisition emphasize the importance of social experience for a child’s ability to learn how to use words.  Given that social relations involve affective as well as cognitive processes, how far might some of the pragmatic language deficits of children with autism be grounded in affective or other intersubjective aspects of communication?

Objectives:

The aim was to examine whether among children with autism, there are relations among impaired social-affective and/or intersubjective processes and atypicalities in the pragmatics of language use in conversation.  We tested the specificity of such relations by assessing whether there were relatively spared structural aspects of language.

Methods: Participants were 12 children with autism between the ages of nine and 19 years with a mean verbal mental age of 6 years; 6 months (VMA range = 4 – 10 years) and 12 children without autism between the ages of 11 and 17 years with (VMA M = 6 years; 7 months, range = 4 – 9.5 years). Participants had each been videotaped in conversation with an adult for a previous study that yielded reliable ratings (on a five-point scale) of how ‘emotionally connected’ each participant was with the interviewer.  For the present studies, we transcribed a standard three minutes of the interview for ratings by judges who were ‘blind’ to diagnostic groups.

Firstly, we rated how far participants’ conversation manifested ‘linkage with speaker’s meanings’ (i.e. with what the speaker intended, ICC = .64) and ‘linkage with utterance meanings’ (i.e. with the literal meanings of the words spoken, but not necessarily the intended message, ICC = .68). 

Secondly, we employed a newly-devised form of Du Bois’ approach to rating transcripts of conversation for ‘dialogic resonance’, a measure focused upon intersubjectively configured aspects of dialogue.  Independent raters achieved satisfactory reliability (ICC range = .57 to .89) in rating atypicalities in three aspects of speaker-hearer resonance – that is, three features of speaker-hearer intersubjective alignment manifest in language – namely, frame resonance with missing/incoherent expansion (e.g. Interviewer: ‘What do you like most about yourself?’ Participant: ‘Most about myself is the teach’), taxonomic resonance (e.g. I: ‘What sort of person are you?’ P: ‘A girl’) and non-indexed alignment (e.g. I: ‘Is it important to control it?’ P: ‘It is’ [rather than: ‘Yes, it is’].

Results: As predicted, (a) the groups differed in linkage with speaker’s meanings, but not linkage with utterance meanings, and (b) for the children with autism, linkage with speaker’s meanings correlated with emotional connectedness (r = .54) but not verbal MA (r = -.09), whereas for children without autism, it correlated with verbal MA (r = .41) but not emotional connectedness (r = -.43).

In addition, the groups were significantly different on a composite measure of these three atypicalities of intersubjective, dialogic linkage – even though in other respects, participants with autism demonstrated abilities to register and elaborate upon what the conversational partner had said.

Conclusions:

Children with autism show atypicalities in pragmatic linguistic adjustment during conversation that are intimately related to the children’s limited affective/intersubjective engagement.