First-order Theory Change Systems and Their Dynamic Semantics

Systems of rst-order theory change are motivated, developed, and shown to be strongly connected to revisable systems of dynamic semantics. Such systems are applicable in the analysis of dialogue contributions in a full account of information exchange. They remedy the following problem; \our approach to information change is eliminative rather than constructive. And we don't know yet how to do better than that." A full theory of communication must incorporate the pragmatics of belief integration (or information change) in dialogue. Systems of theory change (TC) or \belief revision" ((Fuhrmann, 1994, GG ardenfors, 1986]) invoke transformations over sets of sentences (syntactic structures) expressed in propositional logic (the \AGM" approach). In contrast, systems of dynamic semantics (DS) typically employ transformations over sets of possibilities (semantical structures) of rst-order complexity. Cognitively relevant information processing is often argued to have at least the complexity of rst-order logic. Accordingly, predicates and variables are distinguished in the processing of information in natural language interpretation. In order to meet the demands of a model of communication , systems of theory change and their associated dynamic semantics must operate over structures expressed in rst-order logic.) has been connned to the dynamics of propositional systems. The paper is organized as follows. Intuitions are given about a set of rst-order rationality postulates for contraction, covering quantiiers, names, variables, and identity. Some of the intuitions are formalized. Following this, a family of systems for rst-order theory change (FOTC) are sketched, along with their