Speaker: Ilaria Canavotto
Date and Time: Thursday, November 8th 2018, 17:00-18:30
Venue: ILLC Seminar Room F1.15, Science Park 107.
Title: Introducing Causality in Stit Logic.
Abstract. Stit logic is a framework to reason about individual and group agency. More specifically, the central notion that stit theorists aim at modelling is the notion of seeing-to-it-that. Despite some variations, the basic idea is that an agent sees to it that a certain fact, A, is the case just in case her choice ensures that A is the case, no matter what all other agents do. This classic characterization of seeing-to-it-that is typically taken to capture the fact that an agent brings about A. However, the notion of bringing about modelled in stit logic turns out to be considerably narrower than an intuitive notion of bringing about: most of the cases in which we say that an agent brings about a certain fact are cases in which that agent causes that fact and, at the same time, what the other agents do matters. For instance, when we close our office’s door, we are typically prepared to say that we cause the door to be closed (or: we bring about the state that the door is closed), even though the fact that our office mates don’t keep us away from the door matters, since we could not close the door if they did. But, according to stit theory, if this is really the case, then we do not see to it that the door is closed. As we will see, this has important consequences on the possibility of taking standard stit operators as a tool to reason about responsibility.
In this talk, I will present a refinement of stit semantics suitable to represent, at least in a weak sense, the causal connections between an agent’s action and its consequences. I will do this step-wise, by including in the semantics, first, action types and, second, a relation of opposition between them. I will show that, in this way, we obtain a framework in which we can interpret new stit operators suitable to model basic degrees of responsibility of an agent. In general, familiarity with stit logic is not necessary for this talk.
(This is a joint work with Alexandru Baltag and Sonja Smets.)