Activities will resume in Autumn 2026.

If you would like to lead a session, please get in touch.

You may also be interested in a panel we are organising for IPrA 2027 (Helsinki) – description below.

Abstracts for oral presentations should be around 300 words (minimum 250; maximum 500). Panel slots will be 30 minutes (20 minutes presentation with 10 mins Q&A).

Full details on how to submit your abstract can be found on the IPrA webpage: IPC20 Call for Papers – International Pragmatics Association

Deadline for Submissions to Panel: 15 October 2026.

Inference, Cognition and Human-AI Interaction

The emergence, development and widespread use of Artificial Intelligence (AI), Large Language Models (LLMs) raises important questions for the field of pragmatics and the study of communication. These issues are particularly significant for those approaches to communication and interaction which focus on the role of human cognitive processes and intentions (such as Gricean and relevance-theoretic analyses).

We invite papers which engage with issues, challenges, possibilities and applications that arise from the interaction between inferential and cognitive approaches to pragmatics and artificial intelligence. This might include theoretical approaches to human-AI interactions, experimental and empirical work on AI discourse, and analyses of AI generated texts and responses. Isses that might be considered include (but are not limited to) the following:

How might theories of speaker production and hearer interpretation processes apply in contexts where one (or more) of the interlocutors is non-human, and, therefore, has no communicative intentions or expectations? Do human cognitive processes function differently when processing language-based outputs from AI than they do when interpreting human utterances (and other ostensive communication)? Do pragmatic inferential processes such as context selection, abductive reasoning, and the construction of mental representations operate differently when hearers or readers know that a text has been generated by AI rather than by a human communicator? How are cues that would normally support the attribution of mental states to an author, such as style, hedging, emotion, or expressions of personal experience, interpreted when produced by systems that lack beliefs, intentions, and subjective experience? What are the implications for applied pragmatic fields (for example, language learning and acquisition, translation and interpreting, or clinical applications) of AI generated texts and human-AI interactions? How have human speakers adapted their utterances to accommodate the “abilities” of LLM programmes in human-AI interactions? To what extent can LLMs be viewed as modelling or approximating pragmatic inferential processes, and what can their successes and failures reveal about the nature of human pragmatic competence?

Other areas of inferential and cognitive pragmatics that deserves to be revisited in light of human-AI interaction include (but are not limited to):

• Characterisation of the speaker and/or addressee/audience.
• The role of intentions in the interpretation of non-human generated text.
• Emotional responses to AI generated utterances
• Implicature generation and understanding in human-AI discourse
• Epistemic vigilance
• Cognitive pragmatics and the ethics of human-AI interaction.
• Attribution of mental states, emotions, and experiential knowledge to AI systems.
• The interpretation of expressive, experiential, and non-propositional effects (e.g. onomatopoeia, affective language, impressionistic expressions, and other forms of weak communication) in AI-generated discourse.
• Mutual manifestness, common ground, and the dynamics of salience in interactions where communicators do not share persistent cognitive environments
• Irony, metaphor, and other forms of figurative language in human-AI communication.

We anticipate that the papers included in this panel and the discussions that are generated from it will support the future development of a cognitive pragmatic special interest group with a focus on human-AI interaction

Latest from the Blog