Discourse structure and performance efficiency in interactive and non-interactive spoken modalities

Near-term spoken language systems will likely be limited in their interactive capabilities. To design them, we shall need to model how limitations on speaker interaction influence spoken discourse patterns in different types of tasks. Speaker interaction is a central feature of human dialogue, one presumed to have a powerful influence on its discourse structure and performance efficiency. The present study examined two speech modalities that represent opposites on the spectrum of speaker interaction-the telephone dialogue and audiotape monologue. Experts provided spontaneous instructions by either telephone or audiotape as their novice partner completed an assembly task. Within this task framework, a comprehensive analysis is provided of the basic differences in discourse organization, referential characteristics and performance efficiency for these two spoken modalities. The outlined distinctions are interpreted with special reference to the role of confirmation feedback in promoting dialogue efficiency. Implications are discussed for the design of non-interactive speech technology and limited interaction spoken language systems.