A Model for Graded Levels of Generalizations in Intensional Query Answering

Contents

We present in this paper a model for graded levels of generalizations, within a cooperative questionanswering framework. We describe how intensional answers descriptions can be generated when the set of extensional answers set, for a given natural language question, is very large. We develop a variable-depth intensional calculus that allows for the generation of intensional responses at the best level of abstraction.