<?xml version="1.0" standalone="yes"?>
<Paper uid="C04-1174">
  <Title>Automatic Construction of Nominal Case Frames and its Application to Indirect Anaphora Resolution</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> What is represented in a text has originally a network structure, in which several concepts have tight relations with each other. However, because of the linear constraint of texts, most of them disappear in the normal form of texts.</Paragraph>
    <Paragraph position="1"> Automatic reproduction of such relations can be regarded as the first step of &amp;quot;text understanding&amp;quot;, and surely benefits NLP applications such as machine translation, automatic abstraction, and question answering.</Paragraph>
    <Paragraph position="2"> One of such latent relationship is indirect anaphora, functional anaphora, or bridging reference, such as the following examples.</Paragraph>
    <Paragraph position="3">  (1) I bought a ticket. The price was 20 dollars. (2) There was a house. The roof was white.</Paragraph>
    <Paragraph position="4">  Here, &amp;quot;the price&amp;quot; means &amp;quot;the price of a ticket&amp;quot; and &amp;quot;the roof&amp;quot; means &amp;quot;the roof of a house.&amp;quot; Most nouns have their indispensable or requisite entities: &amp;quot;price&amp;quot; is a price of some goods or service, &amp;quot;roof&amp;quot; is a roof of some building, &amp;quot;coach&amp;quot; is a coach of some sport, and &amp;quot;virus&amp;quot; is a virus causing some disease. The relation between a noun and its indispensable entity is parallel to that between a verb and its arguments or obligatory cases. In this paper, we call indispensable entities of nouns obligatory cases.</Paragraph>
    <Paragraph position="5"> Indirect anaphora resolution needs a comprehensive information or dictionary of obligatory cases of nouns.</Paragraph>
    <Paragraph position="6"> In case of verbs, syntactic structures such as subject/object/PP in English or case markers such as ga, wo, ni in Japanese can be utilized as a strong clue to distinguish several obligatory cases and adjuncts (and adverbs), which makes it feasible to construct case frames from large corpora automatically (Briscoe and Carroll, 1997; Kawahara and Kurohashi, 2002).</Paragraph>
    <Paragraph position="7"> (Kawahara and Kurohashi, 2004) then utilized the automatically constructed case frames to Japanese zero pronoun resolution.</Paragraph>
    <Paragraph position="8"> On the other hand, in case of nouns, obligatory cases of noun Nh appear, in most cases, in the single form of noun phrase &amp;quot;Nh of Nm&amp;quot; in English, or &amp;quot;Nm no Nh&amp;quot; in Japanese. This single form can express several obligatory cases, and furthermore optional cases, for example, &amp;quot;rugby no coach&amp;quot; (obligatory case concerning what sport), &amp;quot;club no coach&amp;quot; (obligatory case concerning which institution), and &amp;quot;kyonen 'last year' no coach&amp;quot; (optional case). Therefore, the key issue to construct nominal case frames is to analyze &amp;quot;Nh of Nm&amp;quot; or &amp;quot;Nm no Nh&amp;quot; phrases to distinguish obligatory case examples and others.</Paragraph>
    <Paragraph position="9"> Work which addressed indirect anaphora in English texts so far restricts relationships to a small, relatively well-defined set, mainly part-of relation like the above example (2), and utilized hand-crafted heuristic rules or hand-crafted lexical knowledge such as WordNet (Hahn et al., 1996; Vieira and Poesio, 2000; Strube and Hahn, 1999). (Poesio et al., 2002) proposed a method of acquiring lexical knowledge from &amp;quot;Nh of Nm&amp;quot; phrases, but again concentrated on part-of relation.</Paragraph>
    <Paragraph position="10"> In case of Japanese text analysis, (Murata et al., 1999) proposed a method of utilizing &amp;quot;Nm no Nh&amp;quot; phrases for indirect anaphora resolution of diverse relationships. However, they basically used all &amp;quot;Nm no Nh&amp;quot; phrases from corpora, just excluding some pre-fixed stop words. They confessed that an accurate analysis of &amp;quot;Nm no Nh&amp;quot; phrases is necessary for the further improvement of indirect anaphora resolution.</Paragraph>
    <Paragraph position="11"> As a response to these problems and following the work in (Kurohashi and Sakai, 1999), we propose a method to construct Japanese nominal case frames from large corpora, based on an accurate analysis of &amp;quot;Nm no Nh&amp;quot; phrases using an ordinary dictionary and a thesaurus.</Paragraph>
    <Paragraph position="12"> To examine the practical usefulness of the constructed nominal case frames, we also built a system of indirect anaphora resolution based on the case frames.</Paragraph>
  </Section>
class="xml-element"></Paper>