<?xml version="1.0" standalone="yes"?>
<Paper uid="C86-1091">
  <Title>Towards the Automatic Acquisition of Lexical Data</Title>
  <Section position="11" start_page="388" end_page="388" type="concl">
    <SectionTitle>
6. Conclusion
</SectionTitle>
    <Paragraph position="0"> We have presented a sysLem which automates acquisition of lexical data for a natural language understanding system to a large extent. Knowledge acquisition takes place in graceful interaction with a human who is not supposed to have specific linguistic knowledge. The system relies on the existing natural language system VIE-LANG containing among other sources of knowledge a lexicon with a basic vocabulary such that acquisition does not start from scratch but can be seen as an iterative process. The acquisition system is based on a small rule based system in which three different sorts of knowledge - inflections\], morphonological and heuristic are distinguished and processed differently. As for der\[vational endings as well as compound words the system heavily relies on existing lexicon entries to form its hypotheses.</Paragraph>
    <Paragraph position="1"> The described system forms part of an integrated system for the acquisition of different sorts of knowledge for natural language understanding. An outline of the overall system is to be found in Trost and Buchberger (\]985). The final goal will be a system which augments its knowledge automatically in every interaction with the user in a practical and comfortable way.</Paragraph>
  </Section>
class="xml-element"></Paper>