Goal. To become familiar with different approaches of nonmonotonic reasoning and studying nonmonotonic reasoning from an abstract perspective, in terms of consequence relation.
Classical logic systems are in general monotonic in the sense that if the premise set increases, the conclusion set either remains the same or may increase. Nonmonotonic logic systems are logics where introduction of new premises may invalidate old conclusions. There are many such existing logic systems, proposed by Reiter, McCarthy and others. A large portion of the work in artificial intelligence on commonsense reasoning involves the development of such formal systems.We will study some of these systems, namely, default reasoning, circumscription and autoepistemic logic. We will investigate the notion of nonmonotonic reasoning from a more general perspective, in terms of consequence relations and consider the profile of some nonmonotonic logic systems with regard to the behaviour of their consequence relations. We will also discuss some applications of nonmonotonic reasoning.
