Humour Theory and Conversational Agents: An Application in the Development of Computer-based Agents

Much of the literature on the development of conversational agents comes from the domain of Human Computer Interaction (HCI) (Buchanan 2008; Cassell et al. 2000; Laurel 1993; Reeves and Nass 1996). The Computers Are Social Actors (CASA) paradigm suggests that a user will respond to a computer-based conversational agent in the same manner they would respond to a real person (Nass, Isbister, and Lee 2000). Shechtman and Horowitz, although critical of the conclusions of the CASA paradigm, suggest there is an “inextricable link between the use of natural language and social interaction. Perhaps relationship behaviors are simply difficult to filter out of communication and may arise as an artefact of using natural language in a conversational situation” (2003, p.288). Whether computers and humans are equal social actors or not may be an unnecessary distinction if the use of natural language is alone sufficient to generate the perception of social interaction and personality. If humour is one of the defining features of ‘human-ness’ then for a computer to truly be a social actor it must be able to engage in novel, surprising and humorous exchanges. This paper surveys a range of humour theories in search of a theory that can be applied to interactions of computer-based conversational agents