Indefeasible Semantics and Defeasible Pragmatics

Abstract

An account of utterance interpretation in discourse needs to face the issue of how the discourse context controls the space of interacting preferences. Assuming a discourse processing architecture that distinguishes the grammar and pragmatics subsystems in terms of nontoxic and nonmonotonic inferences, I will discuss how independently motivated default preferences interact in interpretation of intersentential pronominal anaphora. 

In the framework of a general discourse processing model that integrates both the grammar and pragmatics subsystems, I propose a fine structure of the preferential interpretation in pragmatics in terms of defeasible rule interactions. The pronoun interpretation preferences that serve as the empirical ground draw from the survey data specifically obtained for the present purpose. A logical implementation of the preferential rule interactions is proposed using prioritized circumscription, a non-monotonic reasoning formalism in AI.