Default Logic and purity of reasoning

Foreword Ray Reiter has worked out Default Logic during his sabbatical leave in 1979, at Imperial College in London. The paper on Default Logic then appeared in the landmark issue of the Artificial Intelligence Journal on " Commonsense Reasoning " [1]. A vast literature flourished from Ray's initial proposal, and after 20 years, it is still studied and used in many fields of Artificial Intelligence. Default Logic is one of the inventions of AI that will still be celebrated in the future. The secret of the success of Default Logic is in its ability to represent a good deal of the dynamics of reasoning, yet giving it the status of a logic. It has become paradigmatic for the evergreen challenge: how close can logic come to meet reasoning; how far can reasoning go to reach the abstract and formal sky of logic. Even more: Default Logic has challenged scientists interested in purity of reasoning, those interested in the mechanical and al-gorithmic nature of proofs that handle exceptions, and those interested in the logical foundations of cognitive agents. Default Logic has disclosed them both the possibility of making explicit aspects that were implicit or hidden in the initial formulation, and provided them the intuition to make a big step towards the formalization and mechanization of reasoning. We worked on Default Logic from 1989 till 1996. We are among those who believe in purity of reasoning. That is, reasoning mechanisms can be given a formal and symbolic apparatus independent of the instrument or tool that will make it work. This view allowed us to understand an important pattern that links the different forms of commonsense reasoning, i.e. the implicit and explicit contextual definability. We here briefly report our recent results on the logic of defaults, and dedicate them to Ray on his 60th birthday.