Abstract: Inheritance reasoning has frequently been characterized by algorithms designed to operate on inheritance networks, or collections of links. While the intended meaning of links in a network is understood, formal semantic accounts of such networks are somewhat troublesome, as are semantic accounts of the inference process. We suggest that links be interpreted as sentences in the conditional logic E, providing a formal interpretation for such networks. Furthermore, we develop a semantic characterization of inheritance reasoning based on the technique of minimal (or preferred) models. In the process, we identify a key difference between this characterization of inference in networks and those based on the notion of inferential distance, specifically with respect to stability.
0 Replies
Loading