Reasoning with inconsistent knowledge base

Abstract

AbstractThere are several approaches to reasoning from an inconsistent knowledge base, such as the consistency‐based method and the argument‐based method, from the point of view of the definition of conclusions derived from an inconsistent knowledge base. This paper proposes a treatment focusing on the condition for assuring the validity of conclusions derived from an inconsistent knowledge base. A consistency‐based method performs deductive reasoning with consistent subsets selected from an inconsistent knowledge base. There exist maximal consistent sets which derive different conclusions inconsistent with each other. Thus, we propose a condition to distinguish these sets in order to assure the validity of conclusions. On the other hand, an argument‐based method takes an argument that consists of a conclusion and a consistent knowledge base which derives the conclusion. This method selects an acceptable argument according to the alternative relation of possible arguments, as a conclusion has arguments from which it is derived and alternative arguments. Thus, we propose a condition which undercuts alternative arguments, in order to assure the validity of a conclusion. We show that these two methods are essentially identical with regard to the validity of a given conclusion by proving these conditions' equivalence. © 2006 Wiley Periodicals, Inc. Syst Comp Jpn, 37(3): 41–48, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.20391