<?xml version="1.0" standalone="yes"?>
<Paper uid="W98-0709">
  <Title>I I I I I I I I I I I I I I I</Title>
  <Section position="7" start_page="68" end_page="71" type="concl">
    <SectionTitle>
6. Conclusions
</SectionTitle>
    <Paragraph position="0"> An approach for building in a fast and automatic way substantial fragments of WordNets have been presented. The method uses as skeleton English WordNet and extracts its knowledge from a variety of lexical sources (taxonomies, monolingual and bilingual dictionaries). Our approach makes extensive uses  of English WordNet in several steps of the building process. The system has been applied to build Spanish WordNet (within the framework of EuroWordNet) and Catalan WordNet. First, following \[Atserias et al. 97\], we applied a set of complementary techniques for linking Spanish and Catalan words collected from a bilingual MRDs (for nouns) and lexicons (for verbs) to English WordNet. Second, by applying the methodology described in section 4 we are able to build up accurate taxonomies from monolingual MRDs (see \[Rigau et al. 97\] and \[Rigau et al. 98\]). Third, taking profit of both lexical resources (the sparse connections produced by the first methodology and the taxonomies produced by the second) we have presented a novel bootstrapping methodology for covering substantial parts of the new WordNets not covered previously.</Paragraph>
  </Section>
class="xml-element"></Paper>