A Framework for Belief Update

Belief update and reasoning about actions are two well studied areas of research about the evolution of knowledge over time. The similarities between these two fields have already been pointed out by some researchers: for example del Val and Shoham [4] use a theory of action to derive a semantics for belief update; Li and Pereira [8] use a Ginsberg-like semantics for updating a theory of actions. In this paper we present a very simple action description language [6] with narratives that allows expressing several different update semantics. The basic principles of this language has already been investigated in the literature. Indeed, the basic semantics of this language can be seen as a proper restriction of the language L by Baral et al. [1]. What is new in this paper is not the language itself, but rather the way it is able to express update semantics. To introduce the language, we consider an example similar to the evergreen Yale Shooting Problem.