<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-0909">
  <Title>Surfaces and Depths in Text Understanding: The Case of Newspaper Commentary</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Generally speaking, language understanding for some cognitive agent means reconstructing the presumed speaker's goals in communicating with her/him/it. An application-specific automatic system might very well hard-wire some or most of the aspects of this reconstruction process, but things get more interesting when the complexity is acknowledged and paid attention to. When moving from individual utterances to understanding connected discourse, an additional problem arises: that of partitioning the material into segments (usually at various levels) and that of inferring the connections between text segments (or between their underlying illocutions).</Paragraph>
    <Paragraph position="1"> In recent years, some surface-based approaches to &amp;quot;rhetorical parsing&amp;quot; have been proposed, which try to recover a text's discourse structure, following the general layout of Rhetorical Structure Theory (Mann, Thompson, 1988). Starting from this idea, in this paper, we imagine to push the goal of rhetorical parsing a bit further.</Paragraph>
    <Paragraph position="2"> The idea is that of a system that can take a newspaper commentary and understand it to the effect that it can, amongst other things, produce the &amp;quot;most concise summary&amp;quot; of it: a0 the topic of the commentary a0 the position the author is taking toward it This goal does not seem reachable with methods of shallow analysis alone. But why exactly is it not, and what methods are needed in addition? In the following, we work through a sample commentary and analyse the steps and the knowledge necessary to arrive at the desired result, i.e., a concise summary. Thereafter, we sketch the state of our implementation work, which follows the goal of fusing surface-based methods with knowledge-based analysis.</Paragraph>
  </Section>
class="xml-element"></Paper>