Dealing with Exceptions in Textual Databases

This paper addresses the semantic issue of exceptions in natural language texts. After observing that rst-order logic is not adequate for semantic representations of natural language , we make use of extended logic programs for meaning representation and inference engine in natural language processing. We show how to generate semantics of texts through deductive parsing by using deenite clause grammars with focus on meanings of exceptions. Given a textual database, its semantics can be produced as an extended logic program, and the query against the textual database is thus reduced to a query in an extended logic program. The non-monotonicity of extended logic programs makes our method to be especially applicable to incomplete textual databases, but it is also generally applicable to the analysis of natural language descriptions in view of conceptual modelling and to natural language interfaces for database querying. Generally speaking, it also provides a method for connict resolution when the connict arises from exceptions.