<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-2033">
  <Title>Sydney, July 2006. c(c)2006 Association for Computational Linguistics Conceptual Coherence in the Generation of Referring Expressions</Title>
  <Section position="6" start_page="260" end_page="261" type="concl">
    <SectionTitle>
5 Conclusions and future work
</SectionTitle>
    <Paragraph position="0"> This paper started with an empirical investigation of conceptual coherence in reference, which led to a definition of local coherence as the basis for a new greedy algorithm that tries to minimise the semantic distance between the perspectives repre- null sented in a description. The evaluation strongly supports our Coherence Model.</Paragraph>
    <Paragraph position="1"> We are extending this work in two directions.</Paragraph>
    <Paragraph position="2"> First, we are investigating similarity effects across noun phrases, and their impact on text readability. Finding an impact of such factors would make this model a useful complement to current theories of discourse, which usually interpret coherence in terms of discourse/sentential structure.</Paragraph>
    <Paragraph position="3"> Second, we intend to relinquish the assumption of a one-to-one correspondence between properties and words (cf. Siddharthan and Copestake (2004)), making use of the fact that words can be disambiguated by nearby words that are similar.</Paragraph>
    <Paragraph position="4"> To use a well-worn example: the 'financial institution' sense of bank might not make the river and its bank lexically incoherent as a description of a piece of scenery, since the word river might cause the hearer to focus on the aquatic reading of the word anyway.</Paragraph>
  </Section>
class="xml-element"></Paper>