Abstract
In this paper, a formal model of communication in dialogues is described. Recently, a shift
seems to have taken place from the traditional perception of the computer as a mere tool for carrying out a task towards a view of the computer as an assistant with which one or more users work together - i.e., cooperate - on a task. Now, cooperation is rooted in coordination of actions, which in turn can be achieved through communication. The model for cooperative behaviour in dialogues that we propose is based on observational studies into human-human communication carried out at IPO and findings from Discourse Analysis and Conversation Analysis. Our main concern will be to explain the role of the basic constituents of conversation that are known as adjacency pairs. Central in our model are rules that describe how the commitments of the dialogue partners are updated during the course of a dialogue and how they constrain the possible moves of the dialogue participants. We take communication to be part of the overall activity in which the interlocutors are engaged. Such a model is needed if we want to account for the fact that information is often exchanged in dialogues in a sequence of alternating (combinations of) modalities (linguistic means, object manipulations and/or gestures). The model was used for the behaviour rules of an artificial assistant that is implemented as part of the DenK (Dialogue Modelling and Knowledge Acquisition) project.
Original language | English |
---|---|
Pages (from-to) | 89-97 |
Number of pages | 9 |
Journal | IPO Annual Progress Report |
Volume | 32 |
Publication status | Published - 1997 |