Default Reasoning

Abstract

Artificial Intelligence [Barr, Feigenbaum and Cohen 1982] [Charniak and McDermott 1985] is undeniably concerned with the design of sophisticated electronic agents, in the form of computer systems that people could regard as “intelligent”. Since reasoning appears to be the most sparkling manifestation of intelligence, Artificial Intelligence requires programs with a capacity to reason. Accordingly, the study of models of reasoning is a major theme in Artificial Intelligence research. By a model of reasoning we mean the complete, though not necessarily detailed, description of the process which allows an agent (electronic here) to reason. So, a prior endeavour is to determine the forms of reasoning that such a process (hence its model of reasoning) should display if the corresponding agent is to be considered intelligent.