<?xml version="1.0" standalone="yes"?>
<Paper uid="P02-1012">
  <Title>Pronominalization in Generated Discourse and Dialogue</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
2 Background and Related Work
</SectionTitle>
    <Paragraph position="0"> Because most NLG systems have focused on linguistic phenomena at the paragraph level and below, there has been intensive investigation into the core areas of generation that are required to produce them: discourse planning, sentence planning and surface realization. Since pronouns are more likely to be a multiparagraph, discourse-level phenomenon, it has been possible to ignore their inclusion into working NLG systems which are not called upon to generate lengthy passages.</Paragraph>
    <Paragraph position="1"> Indeed, most work on pronouns in computational linguistics has come under the heading of anaphora resolution as an element of parsing rather than the heading of pronominalization as an element of generation. Since discourse anaphora resolution was first studied theoretically (Grosz, 1977; Webber, 1979; Sidner, 1983; Grosz and Sidner, 1986), it has come to be dominated by Centering Theory (Grosz et al., 1995; Di Eugenio, 1998; Walker, 1998) which proposes rules for the determination of focus and salience within a given segment of discourse. Relatively little work has been done on alternate approaches to pronoun resolution (Hobbs, 1976; Baldwin, 1995).</Paragraph>
    <Paragraph position="2"> While many NLG researchers have attempted to transfer the ideas of Centering Theory to generation (Not, 1996; Yeh and Mellish, 1997; McCoy and Strube, 1999; Henschel et al., 2000; Kibble and Power, 2000), there has yet been no substantial return contribution to the field of anaphora resolution. There are two principal reasons for this.</Paragraph>
    <Paragraph position="3"> First, it is extremely difficult to create an NLG system that generates the large quantity of texts needed to exhibit discourse-level phenomena while consistently employing the deep linguistic representations needed to determine appropriate pronominal forms.</Paragraph>
    <Paragraph position="4"> Second, Centering Theory is still vague on the exact definition of terms such as &amp;quot;segment&amp;quot; (Poesio et al., 1999a), making it difficult to create a mutually agreeable implementation.</Paragraph>
    <Paragraph position="5"> An additional area of NLG research that deals with pronouns is that of referring expression generation (Appelt, 1985; Heeman and Hirst, 1986; Claassen, 1992; Dale, 1992), which attempts to find the optimal noun phrase (whether full description, definite description, deixis, pronoun, or reduced noun phrase) to enable a reader to mentally select the intended referent from the set of possible referents (Reiter and Dale, 1997). Comparatively, referring expression generation is a process for local disambiguation and is not generally concerned with single phenomena spanning multiple paragraphs. Because of this, and because the domains and genres we have studied typically do not involve sets of very similar referents, we concentrate on discourse-motivated sources of pronominalization.</Paragraph>
  </Section>
class="xml-element"></Paper>