<?xml version="1.0" standalone="yes"?>
<Paper uid="J95-4001">
  <Title>The Repair of Speech Act Misunderstandings by Abductive Inference</Title>
  <Section position="2" start_page="0" end_page="437" type="abstr">
    <SectionTitle>
1. Introduction
</SectionTitle>
    <Paragraph position="0"> Speech act misunderstandings occur when two participants differ in their understanding of the discourse role of some utterance. For example, one speaker might take an utterance as an assertion while another understands it to be a request. Although many researchers have considered the problem of avoiding misunderstanding (e.g., by correcting misconceptions), previously none has addressed the problem of identifying and repairing misunderstandings once they occurred. Here, we will consider a general model of dialogue that also accounts for the detection and repair of speech act misunderstandings.</Paragraph>
    <Section position="1" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
1.1 The difference between misunderstanding and misconception
</SectionTitle>
      <Paragraph position="0"> The notions of misunderstanding and misconception are easily confounded, so we shall begin by explicating the distinction. Misconceptions are errors in the prior knowl- null edge of a participant; for example, believing that Canada is one of the United States. * Department of Electrical Engineering and Computer Science, Milwaukee, WI 53201, mcroy@cs.uwm.edu t Department of Computer Science, Toronto, Canada M5S 1A4, gh@cs.toronto.edu @ 1995 Association for Computational Linguistics  Computational Linguistics Volume 21, Number 4 McCoy (1989), Calistri-Yeh (1991), Pollack (1986b), Pollack (1990), and others have studied the problem of how one participant can determine the misconceptions of another during a conversation (see Section 5.3 below). Typically such errors can be recognized immediately when an expression is not interpretable with respect to the computer's (presumedly perfect!) knowledge of the world.</Paragraph>
      <Paragraph position="1"> By contrast, a participant is not aware, at least initially, when misunderstanding has occurred. In misunderstanding, a participant obtains an interpretation that she believes is complete and correct, but which is, however, not the one that the other participant intended her to obtain. At the point of misunderstanding, the interpretations of the two participants begin to diverge. It is possible that a misunderstanding will remain unnoticed in a conversation and the participants continue to talk at cross-purposes. Alternatively, the conversation might break down, leading one participant or the other to decide that a misunderstanding has occurred and (possibly) attempt to resolve it.</Paragraph>
    </Section>
    <Section position="2" start_page="0" end_page="437" type="sub_section">
      <SectionTitle>
1.2 The use of repair in the negotiation of meaning
</SectionTitle>
      <Paragraph position="0"> Although they might not always recognize a misunderstanding when it occurs, discourse participants are aware that misunderstandings can occur. So, participants, rather than just passively hoping that they have understood and have been understood, actively listen for trouble and let each other know whether things seem okay. Each participant will use the subsequent discourse itself in order to judge whether previous discourse has been understood correctly. When one participant produces a response that is consistent and coherent with what the other has just said, then the other will take it as a display of understanding. Otherwise, it might be taken as evidence of misunderstanding. In either case, the response is used as an indication of how the second participant interpreted the first, as presumably his response must have some rational explanation; the indicated interpretation is called the displayed interpretation.</Paragraph>
      <Paragraph position="1"> When a participant notices a discrepancy between her own interpretation and the one displayed by the other participant, she can choose to initiate a repair or to let it pass.</Paragraph>
      <Paragraph position="2"> By their choice of repairing or accepting a displayed interpretation, speakers in effect negotiate the meaning of utterances. 1 Repairs can take many forms, depending on how and when a misunderstanding becomes apparent. Conversation analysts classify repairs according to how soon after the problematic turn a participant initiates a repair (Schegloff 1992). The most common type occurs within the turn itself or immediately after it, before the other participant has had a chance to reply. These are called first-turn repairs. The next most common type, second-turn repairs, occur as the reply to the problematic turn (e.g., as a request for clarification). We will not consider these two types of repairs further, because they do not involve misunderstanding per se. Rather, they are used to correct misconceptions, misspeakings, nonhearings, etc.</Paragraph>
      <Paragraph position="3"> Third-turn and fourth-turn 2 repairs address actual misunderstandings. If a display of misunderstanding occurs in the turn immediately following the one that was misunderstood, and the speaker notices the problem immediately and acts to resolve it, then we say that they have made a third-turn repair (see Example 1).</Paragraph>
      <Paragraph position="4">  1 Note that this choice allows for a speaker feigning the occurrence of a misunderstanding in order to achieve some social goal.</Paragraph>
      <Paragraph position="5"> 2 Schegloff (1992) distinguishes nth-turn repair from nth-position repair. The former corresponds to  repairs that begin exactly n - 1 turns after the problematic utterance, while the latter allows an arbitrary number of intervening pairs of turns. We shall use &amp;quot;nth-turn&amp;quot; to refer to both types, allowing intervening exchanges.</Paragraph>
      <Paragraph position="6">  If a display of misunderstanding occurs during a subsequent turn by the same speaker who generated the misunderstood turn, and the hearer then reinterprets the earlier turn and produces a new response to it, then we say that they have made a fourth-turn repair. The fragment of conversation shown in Example 2 (Terasaki 1976) includes a fourth-turn repair. Initially, Russ interprets T1 as expressing Mother's desire to tell, that is, as a pretelling or preannouncement, but finds this interpretation inconsistent with her next utterance. In T3, instead of telling him who's going (as one would expect after a pretelling), Mother claims that she does not know (and therefore could not tell).</Paragraph>
      <Paragraph position="7"> Russ recovers by reinterpreting T1 as an indirect request, which his T4 attempts to satisfy. Fox (1987) points out that such repairs involve, in effect, a reconstruction of the initial utterance. From an AI perspective, these reconstructions resemble the operation of a truth-maintenance system upon an abductive assumption that has proved to be incorrect. 3  of the teachers.</Paragraph>
      <Paragraph position="8"> 1.3 The need for both intentional and social information The problem of interpreting an utterance involves deciding what actions the speaker is doing or trying to do. This process involves not only looking at the surface form of an utterance--for example, was it stated as a declarative?--but also at the context in which it was uttered. This context includes the tasks that the participants are involved in, the prior beliefs that they had, and the discourse itself. Context is important because it allows speakers to use the same set of words, for example, &amp;quot;Do you know what time it is?&amp;quot;, to request the time, to express a complaint, or to ask a yes-no question. Intentional information can rule out some of these readings; for example, a belief that the speaker already knows the time might rule out the 'request' interpretation.</Paragraph>
      <Paragraph position="9"> The difficulty in considering misunderstandings in addition to intended interpretations is that it greatly increases the number of alternatives that an interpreter needs to consider, because one cannot simply ignore the interpretations that seem inconsistent. However, predominant computational approaches to dialogue, which are based solely on inference of intention, already have difficulty constraining the interpretation process. Sociological accounts suggest a more constrained approach to interpretation 3 This is distinct from the kind of plan repair described by Spencer (1990), which he models using an assumption-based truth-maintenance system. In his work, &amp;quot;repair&amp;quot; addresses the problem of incompleteness in a taxonomy of plans, rather than errors in interpretation.</Paragraph>
      <Paragraph position="10">  Computational Linguistics Volume 21, Number 4 and the recognition of misunderstanding, but none are computational. Our model extends the intentional and social accounts of discourse, combining the strengths of both.</Paragraph>
      <Paragraph position="11"> In the intentional accounts, speakers use their beliefs, goals, and expectations to decide what to say; when they interpret an utterance, they identify goals that might account for it. For example, a speaker who wants someone to know that she lacks a pencil might say &amp;quot;I don't have a pencil.&amp;quot; A hearer might then interpret this utterance as an attempt to convey the information. However, for any goal that would explain an utterance, the reasons for having that goal would also be potential interpretations of the utterance. Thus, for the above utterance, intentional accounts might also consider interpretations corresponding to an attempt to express a need for a pencil, a request to be given the pencil, an incomplete attempt to fill out a questionnaire, and so on. 4 The inherent difficulty with this approach is thus knowing when to stop searching for potential meanings.</Paragraph>
      <Paragraph position="12"> According to the ethnomethodological account of human communication known as Conversation Analysis (CA), agents design their behavior with the understanding that they will be held accountable for it. Agents know that their utterances will be taken to display their understanding of some (culturally determined) rules of conversation and the situation prior to the utterance. Agents, aware of some rule or norm that is relevant to their current situation, choose to follow (or not follow) the rule, depending on how they view the consequences of their choice. One important convention is the adjacency pair. Adjacency pairs are sequentially constrained pairs of utterances, (such as question-answer), in which an utterance of the first type creates an expectation for one of the second. A hearer is not bound to produce the expected reply, but if he does not, he must be ready to justify his action and to accept responsibility for any inferences that the speaker might make (Schegloff and Sacks 1973). Where the CA approach is weakest is in its explanation of how the recipient of an utterance is able to understand an utterance that is the first part of an adjacency pair. For this, an agent needs linguistic knowledge linking the features of an utterance to a range of speech acts that can form adjacency pairs. Agents also need to have some idea of the beliefs and intentions that particular actions express, so they can make judgments about their appropriateness in the context.</Paragraph>
    </Section>
    <Section position="3" start_page="437" end_page="437" type="sub_section">
      <SectionTitle>
1.4 Overview
</SectionTitle>
      <Paragraph position="0"> The aim of our research is to construct a model of communicative interaction that will be able to support the negotiation of meaning. In particular, we want to develop a general model of conversation that is flexible enough to handle misunderstandings.</Paragraph>
      <Paragraph position="1"> To support this degree of flexibility, the agents that we model form expectations on the basis of what they hear, monitor for differences in understanding, and, when appropriate, change their own interpretations in response to new information. The model specifies the relationship between this reasoning and discourse participants' beliefs, intentions, and previously expressed attitudes, as well as their knowledge of social conventions.</Paragraph>
      <Paragraph position="2"> In the account, speakers select speech acts on the basis of both their goals and their knowledge of which speech acts are expected to follow upon a given speech act. They must select an utterance form that both parties would agree (in the current</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>