Redefining Belief Change Terminology for Implemented Systems

Knowledge representation and reasoning systems run into danger when they attempt to implement traditional belief change theories intended for ideal reasoning agents. Resource limitations can cause a system that guarantees consistency to fail. We present a belief space whose formalization incorporates the fact that it is not guaranteed to be deductively closed (e.g. unknown implicit beliefs might indicate an undetected inconsistency). Using this formalization, we then define belief maintenance terminology that applies to this deductively open belief space. This new terminology is used to alter some traditional (ideal agent) belief change guidelines and postulates so that they are implementable. We complete our discussion by comparing some current KRR systems and theories in terms of these new concepts and guidelines.