Mathijs Geurts
University of Salzburg
Speech act theory was developed with the idea that when we communicate we are not merely describing the world but acting on it. When we communicate, we assert, ask, command, promise and apologise. And in doing so, we make changes to the social sphere. But where does the social sphere end, and with it the applicability of speech act theory? In this workshop, we will examine the limits of speech act theory from various angles, tracing its development and its function across the private, communicative, and digital spheres. This gives rise to a range of interesting questions. Can LLM’s perform speech acts? Can we perform speech acts in the absence of others? What are the minimal requirements for performing a speech act? And how does our ability to perform and evaluate speech acts develop? These are the questions that will be tackled during the workshop, and together these talks will hopefully give a clearer picture of the aims of speech act theory in general.
Titles and Abstracts
Christopher Gauker (University of Salzburg): A Minimalist Theory of Assertion
In this talk, I will describe a simple language game and argue that it contains assertions that do not satisfy any of the conditions that are placed on assertions by the leading theories of assertion. I will offer independent criticisms of these leading theories, and I will show, by means of examples, that we actually make many assertions that are just as simple as the assertions in my example. Finally, I will take up the project of defining assertion in a general way. I will not complete that project, but I will explain how I think it can be completed.
Mathijs Geurts (University of Salzburg): How to perform speech acts without others
Speech acts are first and foremost social acts, and yet we appear to perform them when we are alone. When we talk to ourselves, we often ask ourselves questions, encourage ourselves, order ourselves and describe the world around us. This suggests that these speech acts have some function in our mental lives. But other speech acts, such as requests, promises and apologies, seem rather odd or even impossible to perform in private. In this talk, I present a pre-theoretic overview of private-apt and non-private-apt speech acts, discuss the differences between the two and offer some possible explanation for why some (apparently) social acts are problematic in private speech and some are not.
Neri Marsili (University of Turin): Rethinking Assertion in the Age of AI
Can LLMs make genuine assertions? Academics disagree. A key driver of this dispute is an underlying disagreement about what asserting requires. Rather than defending a specific characterisation of assertion, I will review different proposals, showing how each can be extended from human communication to machine communication. Artificial utterances, I will argue, satisfy different conceptions of assertoric force to different degrees. This pluralist approach better acknowledges our conflicting intuitions about machine assertion: it explains precisely why we feel that there is some sense in which LLMs can assert, and some sense in which they don’t.
Antonio Scarafone (LMU): Speech acts and normativity: The very first steps
In this talk, I investigate the co-development of linguistic communication and normativity in early childhood. Relying on both theoretical and experimental research, I outline some trajectories along which infants and adults become increasingly better at coordinating actions, start signaling acceptance and rejection of what the other is doing or projects to do, and intervene normatively in the face of incongruities. The core claim will be that infants need not have a prior, rational understanding of what they and others are doing. Rather, such an understanding is a late outcome of communicatively established, normative relationships.
Time: 15:50 – 19:30
Location: SR 1.005
SOPhiA 2026 