News Board

Research Theme

In this project, we will propose a model of conditionals and modals that (i) gives a basis for non-logical inference and (ii) provides a system that is robust for computation by machine.

Conditionals and modals, both of which are semantically close to each other, have been studied for decades, producing the enormous amount of papers. Since it was pointed out that conditionals cannot be properly treated with a classical logical system, several approaches such as a 'modal bases' approach have been proposed. They basically use a model theory in which a notion of similarity between possible worlds plays a central role or at least use a set of propositions. The method of dealing with a set of worlds or propositions, however, has been recognized as a potential problem for programming: it is not practical to incorporate such a set (theoretically infinite) in a machine computation. In order to provide a tool that causes no ontological explosion of objects, we will explore a method to keep those objects to a minimum, using a new approach that has been developed by the French team. They have studied a philosophical problem of properties, whose tradition goes back to the ancient and medieval (Aristotelian) philosophy, and been recently applying it to linguistics. Their approach is based on objects called tropes (manifestations of properties of objects), which require spatio-temporal location to be manifested. The French team is now embarking on a project to extend their approach to events. Our goal is to explore a method in which spatio-temporal manifestations of an event (expressed by a proposition) is integrated to the same event (which is conceptually close to an 'event-type') so that (ii) is achieved. As for (i), non-logical inference such as 'John might come (if he is not dead)' will be analyzed in connection with spatio-temporal manifestations of events under our approach.