Belief Consolidation for Description Logic

The Aim Of The Research Our motivation is based on the study about the check on consistency in municipal bylaw enacted by Toyama Prefecture in Japan (Hagiwara & Tojo 2006). Instead of corporeal documents, the prefecture office permitted the resident to submit the various kinds of application electrically in 2002. However, this new bylaw might be inconsistent with the old administrative procedure act, since digital documents are not considered as corporeal. Thus, there is a problem: According to the new act for administrative procedure with telecommunication, can we submit the digital documents instead of the corporeal documents? In (Hagiwara & Tojo 2006), the method for automatic discordance detection was shown. In the study, inconsistency does not only arise from the logical one (A and ¬A), but also antonyms (‘liquid’ and ‘solid’, ‘vice’ and ’virtue’, etc.). Whether two concepts are antonyms or not is decided by an ontology. However, the experimental result showed that there was no conflicting concepts except for several loops. That is, although we introduce the ontology, we can find few discrepancies in the law. We consider that the cause is not only due to a lack of the valid legal codes for detecting inconsistency, but also the poverty of the ontology. As shown in the study of LRI Core, predominant common-sense characters should be presupposed by all legal domain ontologies, but such a core ontology was not assumed in the domain ontology for the municipal bylaw. Therefore, we come to rewrite the ontology with some generic ontology like LRI Core. Our main problem is whether there is inconsistency in the ontology or not. For example, we often consider that the following conditional sentences are valid. example 1 1. If a is a document, then a is corporeal. 2. If a is a digital document, then a is a document. 3. If a is a digital document, then a is not corporeal. When we suppose there is a digital document, we can conclude from 1 and 2 that it is corporeal, but we can also conclude from 3 that it is not corporeal. When that such sentences are included in our ontology, we want the ontology to be consistent. How do we solve the problem?