Backchannels (BCs; listener signals such as ‘mmhm’ or ‘okay’) are a ubiquitous and essential feature of spoken interaction. They are used by listeners predominantly to support the ongoing turn of their interlocutor and to signal understanding and agreement. Listeners seem to be highly sensitive to the exact realisations of BCs and to judge deviations from typical forms as negative. Very little is known about the use of BCs by speakers diagnosed with autism spectrum disorder (ASD). In dialogue recordings of 28 German adults in two groups of disposition-matched dyads (i.e., both interlocutors were either autistic or non-autistic), we found that the ASD group was characterised by (1) a lower rate of BCs per minute (particularly in the early stages of conversation), (2) less diversity in the lexical realisation of BCs and (3) a less diverse and flexible mapping of different intonation contours to different BC types. We interpret these results as reflecting more general characteristics of autistic as compared to non-autistic individuals, namely different strategies in signalling attention towards an interlocutor and less flexible behaviour in social interaction.

It is proposed that the response token Right, in one important use, is a marker of epistemic dependency between two units of talk by a prior speaker, and that this talk has progressed the understanding by the Right producer of a complex activity involving much information transfer. Two other Rights as response tokens are considered: as an epistemic confirmation token similar to That's right, and as a change-of-activity token similar to Alright/Okay. In addition, Right is shown to be different from other response tokens, including the news receipt Oh, newsmarkers such as Really?, and continuers and acknowledgment tokens such as Mm hm and Yeah. The primary data consist of a fully transcribed dietetic consultation in an Australian hospital between a dietician and a client.This article has been evolving for nearly ten years, and I wish to thank numerous members of a range of audiences for their stimulating questions. I also wish to thank four anonymous reviewers whose penetrating questions led to two fairly radical rethinks of the arguments and organization. I would particularly like to thank Keith Abbott, who was in the audience for what I hope was the last oral presentation of this paper. He provided extensive written feedback and really put me on the track of epistemic progression marking as being the best characterization of this response token.