Autocircumscription

摘要

Reasoning can be used to select among various possible interpretations of events. But how are these possibilities determined? We isolate two key technical features of circumscription (consistency and minimization), and use the first as the basis for a reformulation of the circumscription principle in a way related to possibility, self-knowledge, and negative introspection. The second (minimization) then can be separately expressed on its own. Conceptual clarity and a kind of validity are results of this separation, as well as a computational means to determine (sometimes) when a wff is not among a reasoner's conclusions.