Common Grounds as Multiple Information States

The main goal of this paper is to show how to derive a common ground In order to do this, first we need to know what a common ground should be like. In traditional dynamic semantics, a common ground is given as a single information state which is shared by the people in the context. In this paper I am going to show that a common ground must include more than one information state, and that this is necessary to deal with conversational implicatures, which require distinction between speakers and audience. In this sense, the new common ground will be the basis for a conversational model of dynamic semantics. The traditional dynamic semantics is not appropriate for conversations, but rather a model for text processing.