Meta-argumentation frameworks for multi-party dialogues

Abstract

The conclusions drawn from a dialogue depend both on the content of the arguments, and the level of trust placed in the arguments and the entity advancing them. In this paper, we describe a framework for dialogue where such trust forms the basis for expressing preferences between arguments, and in turn, for computing conclusions of the dialogue. Our framework contains object and meta-level arguments, and uses ASPIC+ to represent arguments, while argument schemes capture meta-level arguments about trust and preferences.