<?xml version="1.0" standalone="yes"?>
<Paper uid="J93-2005">
  <Title>Lexical Semantic Techniques for Corpus Analysis</Title>
  <Section position="8" start_page="353" end_page="353" type="concl">
    <SectionTitle>
6. Summary and Discussion
</SectionTitle>
    <Paragraph position="0"> In this paper we have presented a particularly directed program of research for how text corpora can contribute to linguistics and computational linguistics. We first presented a representation language for lexical knowledge, the generative lexicon, and demonstrated how it facilitates the structuring of lexical relations among words, looking in particular at the problems of metonymy and polysemy.</Paragraph>
    <Paragraph position="1"> Such a framework for lexical knowledge suggests that there are richer relationships among words in text beyond that of simple co-occurrence that can be extracted automatically. The work suggests how linguistic phenomena such as metonymy and polysemy might be exploited for knowledge acquisition for lexical items. Unlike purely statistical collocational analyses, the framework of a semantic theory allows the automatic construction of predictions about deeper semantic relationships among words appearing in collocational systems.</Paragraph>
    <Paragraph position="2"> We illustrated the approach for the acquisition of lexical information for several classes of nominals, and how such techniques can fine-tune the lexical structures acquired from an initial seeding of a machine-readable dictionary. In addition to conventional lexical semantic relations, we then showed how information concerning lexical presuppositions and preference relations can also be acquired from corpora, when analyzed with the appropriate semantic tools.</Paragraph>
    <Paragraph position="3"> In conclusion, we feel that the application of computational resources to the analysis of text corpora has and will continue to have a profound effect on the direction of linguistic and computational linguistic research. Unlike previous attempts at corpus research, the current focus is supported and guided by theoretical tools, and not merely statistical techniques. We should furthermore welcome the ability to expand the data set used for the confirmation of linguistic hypotheses. At the same time, we must remember that statistical results themselves reveal nothing, and require careful and systematic interpretation by the investigator to become linguistic data.</Paragraph>
  </Section>
class="xml-element"></Paper>