Harnessing locality for knowledge inconsistency management

Inconsistency is commonplace in the real world and in our long-term memory. Managing inconsistency is considered a hallmark of the plasticity of human intelligence which includes an important mental process: belief revision. To facilitate belief revision, a necessary condition is that we need to know the localities and contexts of inconsistency, and how different types of inconsistency are clustered. In this paper, we describe a formal definition of locality of inconsistency and use it to identify clusters of inconsistent circumstances in a knowledge base. The results in the paper will help pave the way for a disciplined way in managing knowledge inconsistency.