Matchmaking: dialogue modelling and speech generation meet

This article is concerned with determining the constraints on the selection of appropriate intonation in speech generation in humanmachine information seeking dialogues. The two pillars of our sys tem--a state-of-theart computational dialogue model and a state-of-the-art NL generator--are presented. Based on this, we determine the kinds of linguistic and pragmatic knowledge needed to sufficiently constrain choice in the intonational resources of the system. We take into consideration factors such as dialogue history, speaker's attitudes, heater 's expectations, and semantic speech functions.