Grounding in Multi-modal Task-Oriented Collaboration

1 Introduction

This paper describes the first results of a series of experiments on multi-modal computer-supported collaborative problem solving. This research was motivated by our previous work on systems in which a rule-based agent collaborated with a human agent [Dillenbourg92;Dillenbourg94]. Roschelle and Teasley [Roshelle95] defined collaboration as a "Coordinated, synchronous activity that is the result of a continued attempt to construct and maintain a shared conception of a problem" (p. 70). This definition, which is now widely accepted in psychology, is difficult to translate in terms of human-computer collaboration. Precisely, when we experimented with the human-computer collaborative systems mentioned above, we observed various problems due to the difficulty for a human and a machine to share conception of a problem [Dillenbourg95a]. We hence decided to study the social grounding process, i.e. the mechanisms by which common ground is updated and maintained between two collaborating human agents. The experimental setting, as described below in Section 3, does not include audio and video communication in order to adapt the bandwidth to current widely available interfaces for human-computer collaboration. The final goal is to design more powerful collaboration interfaces between a human user and a knowledge-based system, not necessarily by imitating human-human collaboration, but by designing functionally equivalent grounding mechanisms.

This goal gives a special colour to our study of grounding. We are less concerned by the quality of communication between agents, than by the cognitive effects of their interactions. From an efficiency of communication point of view, it would seem to be more advantageous to minimize the necessary effort for actions (such as repairs) aimed primarily at grounding. However, as articulated by [ClarkWilkes-Gibbs86], what is important is not individual effort by the producer or receiver of a communicative act, but the overall Least Collaborative Effort. The cost of producing a perfect utterance may be higher (if it is even possible) than the cost of collaboratively repairing those problems which do arise. Conversely, given our concern for cognitive effects, we would rather talk in terms of Optimal Collaborative Effort. When two partners misunderstand, they have to build explanations, justify themselves, often make explicit some knowledge which would otherwise remain tacit, monitor each other and therefore reflect on their own knowledge, and so forth. These mechanisms are crucial for showing that collaborative learning is sometime more effective than learning alone [Dillenbourg95b]. Thus, the extra effort for grounding, even if it does slow down interaction, may lead to better results in the task which motivated the communication episode, particularly when the task involves learning or negotiation. However, as suggested by the word 'optimal', those grounding efforts have to remain subordinated to the accomplishment of the task, i.e. to the effective need for grounding knowledge.

Grounding in Multi-modal Task-Oriented Collaboration - 3 SEP 1996

Generated with Harlequin WebMaker