<?xml version="1.0" standalone="yes"?>
<Paper uid="P05-2010">
  <Title>Using Readers to Identify Lexical Cohesive Structures in Texts</Title>
  <Section position="6" start_page="58" end_page="58" type="evalu">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> This paper presented a reader-based experiment on finding lexical cohesive patterns in texts. As it often happens with tasks related to semantics/pragmatics (Poesio and Vieira, 1998; Morris and Hirst, 2005), the inter-reader agreement levels did not reach the accepted reliability thresholds. We showed, however, that statistical analysis of the data, in conjunction with a subsequent validation experiment, allow identification of a reliably annotated core of the phenomenon. null The core data may now be used in various ways.</Paragraph>
    <Paragraph position="1"> First, it can seed psycholinguistic experimentation of lexical cohesion: are anchored items processed quicker than unanchored ones? When asked to recall the content of a text, would people remember prolific anchors of this text? Such experiments will further our understanding of the nature of text-reader interaction and help improve applications like text generation and summarization.</Paragraph>
    <Paragraph position="2"> Second, it can serve as a minimal test data for computational models of lexical cohesion: any good model should at least get the core part right. Much of the existing applied research on lexical cohesion uses WordNet-based (Miller, 1990) lexical chains to identify the cohesive texture for a larger text processing application (Barzilay and Elhadad, 1997; Stokes et al., 2004; Moldovan and Novischi, 2002; Al-Halimi and Kazman, 1998). We can now subject these putative chains to a direct test; in fact, this is the immediate future research direction.</Paragraph>
    <Paragraph position="3"> In addition, analysis techniques discussed in the paper - separating interpretation disagreement from difference in consistency, using statistical hypothesis testing to find reliable parts of the annotations and validating them experimentally - may be applied to data resulting from other kinds of exploratory experiments to gain insights about the phenomena at hand.</Paragraph>
  </Section>
class="xml-element"></Paper>