Modeling informal debates
Rachel Reichaan · International Joint Conference on Artificial Intelligence · 1981
Many rules of formal debate are documented, of common knowledge, and looked-up in preparation for planned debating. Informal debates, on the other hand, are highly dynamic, are complex, and are spontaneously generated with no prior rule-book preparation. They too, however, are rule-governed. In this paper I present an abstract process model capable of modeling well formed argument structures that occur in ordinary conversations. The formalization rests on a general theorettical framework for discourse engagement encapsulated in a discourse ATN grammar. A major feature of the system is its segmentation of discourse utterances into functionally related context spaces.