<?xml version="1.0" standalone="yes"?>
<Paper uid="W01-1627">
  <Title>Dialogue tagsets in oncology</Title>
  <Section position="4" start_page="0" end_page="0" type="metho">
    <SectionTitle>
3 Dialogue tagsets
</SectionTitle>
    <Paragraph position="0"> Not surprisingly, the actual tagsets developed in oncology re ect their domain more closely than the parameter sets do. In comparison with NLP work, syntactic classi cation is minimal and functionally oriented, while communication management and psychological / emotional loading receive prominent, ne-grained analysis.</Paragraph>
    <Section position="1" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
3.1 Form
</SectionTitle>
      <Paragraph position="0"> Although all four oncology systems encode the form of an utterance in some way, the classi cations have a strong pragmatic bias. Questions are distinguished, not in traditional syntactic terms as yes-no or wh-, but according to their e ect on the ow of the dialogue. The simplest set is that  of Butow et al: Open Question, Closed Question, Response to Question, Statement, Other. PMG add Directive Question (open), Directive Question (closed), Screening Question, Leading Question, Multiple Question.</Paragraph>
      <Paragraph position="1">  Ford et al distinguish \modes&amp;quot; from \content codes&amp;quot;, but even the modes encode coarse-grained content information as well as a ective classi cation. The form categories of Ong et al are \instrumental&amp;quot; (Directions, Question-asking, Information-giving, &amp;c), and they specify that \if a decision must be made between categorizing an utterance in an instrumental or a ect category, the a ect category should be used&amp;quot; - quite reasonably, given the purpose of their analysis. Even with a prior commitment to maintaining separate and independent levels of analysis, some leakage between levels can occur. (The set of forty-two Dialogue Act labels used by Stolcke et al (2000) shows some similar mixing of levels, including both purely syntactic tags (such as Declarative Yes-No Question) and a ective tags (such as Appreciation).) null</Paragraph>
    </Section>
    <Section position="2" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
3.2 Content
</SectionTitle>
      <Paragraph position="0"> The content of an utterance is also encoded in all four systems, and the tagsets on this level are the most domain-speci c. Butow et al cite seven content categories: Treatment, Diagnosis, Prognosis, History, Other medical matters, Social matters.</Paragraph>
      <Paragraph position="1"> Ford et al, with 15 content codes, and PMG, with 38, are the most fully developed. Both include Medical (further distinguished by PMG, with four categories for diagnosis and two for prognosis), Treatment, Psychological, Social, Lifestyle, &amp;c. PMG are particularly detailed in their categories for psychological and emotional issues, shading into the a ect level: Concerns, Feelings, Emotions, Religion, &amp;c. Again, this is what one would expect, given that their reason for carrying out the analysis is to assess the health professional's success in getting the patient to talk about exactly these issues.</Paragraph>
      <Paragraph position="2"> Both Ford and PMG also include the opening and closing of the interview under this heading, where it sits oddly. A separate level of communication management, as in DAMSL, would accommodate these and the open/ closed/ directive question distinction currently made in the Form tagsets, clarifying all three.</Paragraph>
    </Section>
    <Section position="3" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
3.3 Pragmatics
</SectionTitle>
      <Paragraph position="0"> As noted above, the Form classes used in the four coding schemes express more pragmatic than syntactic information. Ong et al's \instrumental clusters and categories&amp;quot; (Directions, Question- asking, Information-giving, Counselling) can be considered pragmatic. So can PMG's \Function&amp;quot; codes: eliciting, checking, acknowledgement (psychological, general, cognitions); reassurance, negotiation, information giving. These are similar to some of the Dialogue Act labels used in NLP work: Stolcke et al's (2000) agreement, response acknowledgement, summarize, or VERBMOBIL's suggest, con rm, clarify (Jekat et al 1995).</Paragraph>
    </Section>
    <Section position="4" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
3.4 A ect
</SectionTitle>
      <Paragraph position="0"> Cognitive a ect - the psychological force, for a patient, of an utterance or a complete dialogue is the focus of interest in oncology and thus the most highly developed area. Ford et al pick out eight of their \modes&amp;quot; as a ective, including the expression of irritation, gratitude, apology, and concern.</Paragraph>
      <Paragraph position="1"> Ong et al rate both doctor and patient, by coding their utterances, on ve distinct \global a ect&amp;quot; scales: Anger/ irritation, Anxiety/ nervousness, Dominance/ assertiveness, Interest/ engagement, Friendliness/ warmth. Their \a ective clusters and categories&amp;quot; comprise (with subheadings) social behaviour, verbal attentiveness, showing concern, and negative talk.</Paragraph>
      <Paragraph position="2"> PMG do not represent a ect as a separate parameter, as such. Their function codes include a ective functions such as Empathy and Reassurance. Many of their content codes can also represent a ect, as noted above. Topics such as Concerns, Feelings about health care, Religion / spiritual issues can be addressed at any level from simply factual to deeply emotional, blurring the picture: this would be clari ed if the a ect level were explicitly factored out. The most direct representation of a ective level comes in the two codes Psychological explicit and Psychological implicit.</Paragraph>
      <Paragraph position="3"> Each utterance in a dialogue can be given several content codes, commonly including one of these two, as seen in the sample dialogue below.</Paragraph>
      <Paragraph position="4"> Cognitive a ect has barely been touched on by NLP research in dialogue tagging. It is clearly more subtle and di cult than syntactic, semantic, or pragmatic analysis, and also less signi cant in instructional or service dialogues than in the highly charged, life-critical domain of cancer care.</Paragraph>
      <Paragraph position="5"> It is, however, an important aspect of dialogue and speaker modelling, and of the design of appropriate responses. In this area, NLP could learn some valuable lessons from oncology.</Paragraph>
    </Section>
  </Section>
  <Section position="5" start_page="0" end_page="0" type="metho">
    <SectionTitle>
4 An example
</SectionTitle>
    <Paragraph position="0"> Here is a brief typical example from a PMG annotated dialogue. Notice the multiple and somewhat diverse content codes, and the classi cation of cue management (somewhat counter-intuitively attached to the cue utterance itself, not the response).</Paragraph>
    <Paragraph position="1"> P26: I said there's only another thing that I hope I never have to have and that's selectron treatment.</Paragraph>
  </Section>
class="xml-element"></Paper>