Belief Revision with Bridging Axioms

Belief revision deals with the problem of changing a declaratively specified repository under potentially conflicting information. Usually, the problem is approached by providing postulates that specify intended constraints for the revision and constructing concrete revision operators fulfilling them. In the last 30 years since the start of formal belief revision with the work of AGM (Aichourron, Gärdenfors, and Makinson) roughly four construction principles were investigated and mutually interrelated: partial-meet, epistemic entrenchment, safe/kernel, and the possible worlds (model based) construction. The aim of this paper is to raise into the focus another construction principle relying on the idea of reinterpretation: Conflicts are explained by different use of symbols and conflict resolution is handled by choosing appropriate bridging axioms that relate the different readings. The main purpose of the paper is to argue that the reinterpretation-based approach is sufficiently general by showing how to equivalently formulate classical revision operators such as the operators of Weber, a natural variant of Weber, the operator of Satoh (= skeptical operator of Del grande and Schaub) and the operator of Borgida with reinterpretation operators. Copyright © 2017, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.