Inconsistency and its handling are very important in the real world and in the fields of computer science and artificial intelligence. When dealing with inconsistency in a knowledge base (KB), there is a whole host of deeper issues we need to contend with in order to develop rational and robust intelligent systems. In this paper, we focus our attention on one of the issues in coping with KB inconsistency: how to measure the information content and the significance of inconsistency in a KB. Our approach is based on a fixpoint semantics for KB. The approach reflects each inconsistent set of rules in the least fixpoint of a KB and then measures the inconsistency in the context of the least fixpoint for the KB. Compared with the existing results, our approach has some unique benefits.