Avoiding logical omniscience and perfect reasoning: a survey

Article type: Research Article

Authors: Moreno, Antonio

Affiliations: Escola Tècnica Superior d’Enginyeria, Departament d’Enginyeria Informàtica, Universitat Rovira i Virgili (URV), Carretera de Salou, s/n. 43006‐Tarragona, Spain E‐mail: [email protected]

Abstract: Beliefs have been formally modelled in the last decades using doxastic logics. The possible worlds model and its associated Kripke semantics provide an intuitive semantics for these logics, but they seem to commit us to model agents that are logically omniscient (they believe every classical tautology) and perfect reasoners (their beliefs are closed under classical deductive closure). Thus, this model does not seem to be appropriate to model non‐ideal agents, that have resource limitations that can prevent them from attaining such levels of doxastic competence. This article contains a statement of these problems and a survey of some of the most interesting approaches that have been suggested to overcome them.

Keywords: Doxastic logics, possible worlds, Kripke semantics, logical omniscience, perfect reasoning, non‐ideal agents

Journal: AI Communications, vol. 11, no. 2, pp. 101-122, 1998