[DiEugenio et al.2000] DiEugenio, B., Jordan, P. W., Thomason, R. H., and Moore, J. D. (2000). The agreement process: an empirical investigation of human-human computer-mediated collaborative dialogues. International Journal of Human Computer Studies, pages 1–57.
The main contribution of the work is the definition of a framwork which combines the IRMA framework for rational problem solving (Bratman, Israel and Pollack, 1988) with Clark’s work on language as a collaborative activity (1992, 1996). To the authors, what is missing in IRMA is an explicit link to collaboration, particularly in dialogue. IRMA cannot explain how negotiation unfolds in a dialogue, how conversant come to agree on a certain solution. Clark’s contribution model seems, for them, to bridging the gap, as it explains how the mutual belief needed for an agreement can be reached.
Mainly, in their studies they try to set the framework to be able to recognise the function of each utterance within the agremment process. In their adaptation of the contribution model, a proposals correspond to situations where an agent can deliberate to the point of making a commitment. Therefore in the light of their corpus study they conclude that the notion of commitment (Cohen and Leversque 1990, Bratman 1992) is more useful than that of acceptance / rejection in order to model the agreement process. This is because tracing how commitment changes with respect of certain proposal, they account for how negotiation unfolds over several turns, so to overcome the problem of recognising implicit or passive acceptances.
This exend on my idea that we have to look for tangible signs of the grounding process instead of referring to beliefs that are hidden and difficult to trace.
In their work they used the COCONUT environment, claiming that the basic agreement process is not affected by the particular modality of the dialogue.