Contributing to discourse

Abstract

For people to contribute to discourse, they must do more than utter the right sentence at the right time. The basic requirement is that they odd to their common ground in on orderly way. People take part in conversation in order to plan, debate, discuss, gossip, and carry out other social processes. When they do take part, they could be said to contribute to the discourse. But how do they contribute? At first the answer seems obvious. A discourse is a sequence of utterances produced as the participants proceed turn by turn. All that participants have to do to contribute is utter the right sentence at the right time. They may make errors, but once they have corrected them, they are done. The other participants have merely to listen and understand. This is the view subscribed to in most discourse theories in psychology, linguistics, philosophy; and artificial intelligence.A closer look at actual conversations, however, suggests that they are much more than sequences of utterances produced turn by turn. They are highly coordinated activities in which the current speaker tries to make sure he or she is being attended to, heard, and understood by the other participants, and they in turn try to let the speaker know when he or she has succeeded. Contributing to a discourse, then, appears to require more than just uttering the right words at the right time. It seems to consist of collective acts performed by the participants working together.In this paper we describe a model of contributions as parts of collective acts. We first describe the need for such a model, next present the model We thank many colleagues for discussions on the issues presented here and Eve V. Clark, Florence H. Edwards, and several reviewers for suggestions on the manuscript. The research was supported in part by Grant BNS 83-20284 Quirk, 1980). The empirical claim is that the model accounts for the bulk of the successful talk in these conversations. THE COURSE OF DISCOURSE "Models of discourse differ greatly depending on whether they originate in philosophy (e.g., Kamp, 1981;Lewis, 1979;Stalnaker, 1978), linguistics (Heim, 1983), artificial intelligence (Grosz & Sidner, 1986;Reichman, 1978;Polanyi and Scha, 1985), or psychology (Clark & Haviland, 1977; JohnsonLaird, 1983;van Dijk and Kintsch, 1983). Still, in one way or another, most of them make three assumptions.(1) Common ground: The participants in a discourse presuppose a certain common ground. (2) Accumulation: In the course of a discourse, the participants try to add to their common ground.(3) Unilateral action: The principal means by which the participants add to their common ground is by the speaker uttering the right sentence at the right time. To take Kamp's proposal as an example, the content of a discourse is accumulated in a Discourse Representation Model, or DRM, which is tacitly assumed to be common ground for the participants. With each new utterance, new structures get added to the DRM. These structures are simply assumed to be what the speaker intended; there are no special...

The influence of reference acceptance and reuse on conversational memory traces.

“…We thus suggest that in contrast to the common ground based on the information mentioned earlier during the current interaction, which is built in two steps (presentation and acceptance; Clark & Schaefer, 1989), the common ground based on the content of past interactions is built in three phases: presentation, acceptance and reuse.…”

“…Indeed, one of the central ideas in this approach is that any piece of information mentioned during an interaction is added to the speakers' common ground, which consists of the information that they share and are aware of sharing (Clark, 1996). Specifically, these pieces of information are added to the common ground through a joint contribution process (Clark & Schaefer, 1989). One of the speakers starts by presenting a piece of information (e.g., a reference) for the other speaker to understand; the latter then accepts it by showing that he or she has understood it well enough for current purposes (Clark & Brennan, 1991).…”

Incorporating a user model to improve detection of unhelpful robot answers

“…Where the units are not specified, the number represents the fraction of all the relevant dialogues where the respective dialogue pattern 2 We are loose in our interpretation of what constitutes a reasonable answer. We consider an answer to any plausible interpretation of the question as reasonable.…”

“…Misunderstood questions can also be thought of in terms of grounding [2], defined as the process of adding material to the common ground between speakers. A misunderstood question in these terms is a type of error in adding material to the common ground between the user and the robot.…”

Analysing Asynchronous Collaboration

“…The second feature of synchronous interaction for our purposes is that it provides an opportunity for individuals to align their interpretations of an artefact, whatever those interpretations may be, through grounding. Grounding is the basic process used to update common ground in communication and was originally designed to account for how individuals develop a common understanding of utterances or turns during a conversational exchange (Clark and Schaefer, 1989). We extend this approach to cover asynchronous interaction by treating artefacts as, in effect, superordinate 'turns' that are also grounded between individuals.…”