<?xml version="1.0" standalone="yes"?>
<Paper uid="C04-1180">
  <Title>Wide-Coverage Semantic Representations from a CCG Parser</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> The levels of accuracy and robustness recently achieved by statistical parsers (e.g. Collins (1999), Charniak (2000)) have led to their use in a number of NLP applications, such as question-answering (Pasca and Harabagiu, 2001), machine translation (Charniak et al., 2003), sentence simplification (Carroll et al., 1999), and a linguist's search engine (Resnik and Elkiss, 2003). Such parsers typically return phrase-structure trees in the style of the Penn Treebank, but without traces and coindexation. However, the usefulness of this output is limited, since the underlying meaning (as represented in a predicate-argument structure or logical form) is difficult to reconstruct from such skeletal parse trees.</Paragraph>
    <Paragraph position="1"> In this paper we demonstrate how a wide-coverage statistical parser using Combinatory Categorial Grammar (CCG) can be used to generate semantic representations. There are a number of advantages to using CCG for this task. First, CCG provides &amp;quot;surface compositional&amp;quot; analysis of certain syntactic phenomena such as coordination and extraction, allowing the logical form to be obtained for such cases in a straightforward way. Second, CCG is a lexicalised grammar, and only uses a small number of semantically transparent combinatory rules to combine CCG categories. Hence providing a compositional semantics for CCG simply amounts to assigning semantic representations to the lexical entries and interpreting the combinatory rules. And third, there exist highly accurate, efficient and robust CCG parsers which can be used directly for this task (Clark and Curran, 2004b; Hockenmaier, 2003).</Paragraph>
    <Paragraph position="2"> The existing CCG parsers deliver predicate argument structures, but not semantic representations that can be used for inference. The present paper seeks to extend one of these wide coverage parsers by using it to build logical forms suitable for use in various NLP applications that require semantic interpretation. null We show how to construct first-order representations from CCG derivations using the l-calculus, and demonstrate that semantic representations can be produced for over 97% of the sentences in unseen WSJ text. The only other deep parser we are aware of to achieve such levels of robustness for the WSJ is Kaplan et al. (2004). The use of the l-calculus is integral to our method. However, first-order representations are simply used as a proof-of-concept; we could have used DRSs (Kamp and Reyle, 1993) or some other representation more tailored to the application in hand.</Paragraph>
    <Paragraph position="3"> There is some existing work with a similar motivation to ours. Briscoe and Carroll (2002) generate underspecified semantic representations from their robust parser. Toutanova et al. (2002) and Kaplan et al. (2004) combine statistical methods with a linguistically motivated grammar formalism (HPSG and LFG respectively) in an attempt to achieve levels of robustness and accuracy comparable to the Penn Treebank parsers (which Kaplan et al. do achieve).</Paragraph>
    <Paragraph position="4"> However, there is a key difference between these approaches and ours. In our approach the creation of the semantic representations forms a completely It could cost taxpayers 15 million to install and residents 1 million a year to maintain</Paragraph>
    <Paragraph position="6"> separate module to the syntax, whereas in the LFG and HPSG approaches the semantic representation forms an integral part of the grammar. This means that, in order for us to work with another semantic formalism, we simply have to modify the lexical entries with respect to the semantic component.</Paragraph>
  </Section>
class="xml-element"></Paper>