<?xml version="1.0" standalone="yes"?>
<Paper uid="W91-0216">
  <Title>For the Lexicon That Has Everything</Title>
  <Section position="9" start_page="184" end_page="184" type="concl">
    <SectionTitle>
9 Summary
</SectionTitle>
    <Paragraph position="0"> We are trying to build a big lexical database that contains detailed information about all its entries. We argue in support of this enterprise that much information often classified as encyclopedic is needed by natural language processing programs trying to carry out tasks in parsing, generation, and information retrieval. In particular, the need for thesaurus information (information about lexical and semantic relationships between words) is becoming increasingly clear. We are convinced that phrasal entries need information that is at least as rich and detailed as that provided for individual words.</Paragraph>
    <Paragraph position="1"> We are trying to make this lexicon usable by many different people for many different tasks, with the goal of providing it free to anyone who can use it. This means that we must build it out of pieces that are readily available to the research community, that we must be able to provide subsets of many different kinds, and that we must provide tools so that others can access these files and add to them whatever further entries and information they need for particular applications.</Paragraph>
  </Section>
class="xml-element"></Paper>