CUI 2022 features two keynotes from internationally-renowned experts on conversational user interfaces, which will be delivered at the start of the conference on Wednesday and Thursday.

Style matters: Insights for conversational user interfaces from sociolinguistic theory Socially-Aware Conversational Interfaces: Challenges and Perspectives

Download the slides from Jane’s talk »

On Wednesday, the conference will open with a keynote by Professor Jane Stuart-Smith, Professor of Phonetics and Sociolinguistics, at the University of Glasgow.

When humans respond to digital agents, it seems likely that they extend and adapt behaviours evolved for interpersonal interaction, to interaction with non-human agents (Reeves & Nass, 1996; Staum Casasanto, L Jasmin & Casasanto, 2010). Sociolinguistics is the branch of linguistics which considers linguistic variability in social context, and provides a wealth of information about the range of socially-driven behaviours and motivations for how humans respond to each other (Tagliamonte, 2012; Schilling, 2013). This talk will consider two aspects of sociolinguistic research which may be useful in helping to inform the development of effective conversational user interfaces: 1) evidence and theories accounting for intraspeaker variation, or ‘speech style’ (Coupland, 2007); and 2) the role of socially-based attitudes to linguistic variation, and in particular, accent bias, on cognitive and linguistic processing (Staum Casasanto, 2009).


Coupland, N. (2007). Style: Language Variation and Identity. CUP.
Reeves, B., & Nass, C. I. (1996). The media equation: How people treat computers, television, and new media like real people and places. Cambridge university press.
Schilling, N. (2013). Investigating stylistic variation. In The Handbook of Language Variation and Change.
Staum Casasanto, L Jasmin, K., & Casasanto, D. (2010). Virtual accommodating: Speech rate accommodation to a virtual interlocutor. In S. Ohlsson & R. Catrambone (Eds.), Proc. 32nd Ann Conf Cog Sci Soc., Austin (pp. 127–132).
Staum Casasanto, L. (2009). How do listeners represent sociolinguistic knowledge? Proceedings of the Annual Meeting of the Cognitive Science Society, Volume 31.
Tagliamonte, S. (2012). Variationist Sociolinguistics: Change, Observation, Interpretation. Wiley-Blackwell.

What insights from conversation analysis can, should, and should not be leveraged when collaborating in conversation design?

On Thursday, we will have a keynote from Cathy Pearl (Conversation Design Lead, Google), Dr Saul Albert (Lecturer in Social Sciences/Social Psychology, Loughborough University) and Professor Elizabeth Stokoe (Professor of Social Interaction, Loughborough University).

We have three aims in our presentation. First, we reflect on the different fundamental assumptions underpinning conversation analysis (CA) and conversation design (CD), on the Gricean underpinnings of the standard processing pipeline for most conversational user interfaces, and how this infrastructure makes some natural conversational things easy, and some impossible. Second, we examine examples of social interactional phenomenon to consider the potential reasons and need, as well as challenges, of leveraging CA research for CD. These include non-lexical turns and silence, request design, and self-initiated self-repair. Finally, we discuss the rhetoric and reality of conversational AI and the benefits and challenges of collaborating across disciplines and industries. Throughout the presentation, we address the question asked in our title, about what the aims might and might not be of embedding insights from CA into CD and conversational AI.