<?xml version="1.0" standalone="yes"?>
<Paper uid="J95-3003">
  <Title>Collaborating on Referring Expressions</Title>
  <Section position="8" start_page="378" end_page="379" type="concl">
    <SectionTitle>
8. Conclusion
</SectionTitle>
    <Paragraph position="0"> We have presented a computational model of how a conversational participant collaborates in making and understanding a referring expression, based on the view that language is goal-oriented behavior. This has allowed us to do the following. First, we have accounted for the tasks of building a referring expression and identifying its referent by using plan construction and plan inference. Second, we have accounted for the conversational moves that participants make during the acceptance process by using meta-actions. Third, we have accounted for collaborative activity by proposing that agents are in a certain mental state that includes a goal, a plan that they are currently considering, and intentions. This mental state sanctions the acceptance of clarification plans, and sanctions the adoption of goals to clarify. Although our work has focused on referring expressions, we feel that it is relevant to collaboration in general and to how agents contribute to discourse.</Paragraph>
    <Paragraph position="1"> This paper is based on the model of collaboration proposed by Clark and Wilkes-Gibbs (1986). Their model makes two strong claims about how agents collaborate.</Paragraph>
    <Paragraph position="2"> First, it minimizes the distinction between the roles of the person who initiates the referring expression and the person who is trying to identify it. Both have the same moves available to them, for either can judge the description and either can refashion it. This allows both participants to contribute without being controlled or impeded by the other. Second, their model gives special status to the role of the current referring expression (current plan): participants judge and refashion the current referring expression directly, rather than recursively modifying modifications (e.g. Litman and Allen 1987; Chu-Carroll and Carberry 1994) or incrementally adding to the current plan with each accepted proposal (e.g. Traum and Hinkelman 1992; Sidner 1992). In our work, we have taken Clark and Wilkes-Gibbs's descriptive model and recast it into a computational one, thus demonstrating the computational feasibility of their work and its compatibility with current practices in artificial intelligence.</Paragraph>
    <Paragraph position="3"> There are many ways that this research could be extended. Perhaps the most obvious would be to extend the planning component of our model. First, our coverage of referring expressions could be extended to handle references to objects in focus and to descriptions that include a plan of physical actions for identifying the referent.</Paragraph>
    <Paragraph position="4"> Second, the treatment of clarifications could be improved; specifically, how plan failures are reasoned about, how plan failures affect the agent's beliefs, and how these failures are repaired. Third, this research needs to be integrated into a more complete plan-based approach to language, and needs to be extended so as to handle more general discourse plan failures (McRoy and Hirst 1993; McRoy and Hirst 1995; Horton and Hirst 1991; Heeman 1993; Edmonds 1994; Hirst et al. 1994). A benchmark for such future work could be dialog (8.1) below, from the London-Lund corpus (Svartvik and Quirk 1980, S.2.4a:1-8, which is the basis of the example used in Section 6. This dialog shows how collaboration on a referring expression can be embedded in other  Computational Linguistics Volume 21, Number 3 activities, how agents can return back to a collaborative activity, and even how agents can take advantage of a mistaken referent.</Paragraph>
    <Paragraph position="5"> (8.1) A'- 1 What's that weird creature over there? B: 2 In the corner? _~: 3 affirmative noise B.&amp;quot; 4 It's just a fern plant.</Paragraph>
    <Paragraph position="6"> A: 5 No, the one to the left of it.</Paragraph>
    <Paragraph position="7"> B-&amp;quot; 6 That's the television aerial. It pulls out.</Paragraph>
    <Paragraph position="8"> A second avenue for future work is to further investigate collaborative behavior and protocols for interaction. We need to formalize what it means for agents to be collaborating, in a theory that takes account of rational interaction and the beliefs and knowledge of the participants. Such a theory would do the following. First, it would give a more complete motivation for the processing rules that we used for how agents interact in a collaborative activity. Second, it would account for why agents would enter into such a mode of interaction, how it is initiated, how it is carried forward (especially how agents' beliefs and knowledge influence their actions), and how it ends. Third, it would be extendable to other forms of interaction, such as information-seeking dialogs. Fourth, it would specify how collaborative activity could be embedded in, or embed, other types of interactions. By answering these questions, we will not only have a better model to base natural language interfaces on, but we will also have a better understanding of how people interact.</Paragraph>
  </Section>
class="xml-element"></Paper>