<?xml version="1.0" standalone="yes"?>
<Paper uid="C88-2121">
  <Title>Example,~ of Adjuncts: S vp A A S VP PP S S A A A S ADV P NP$ SC S~</Title>
  <Section position="6" start_page="582" end_page="582" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper we presented a general parsing strategy based on 'lexicalized' grammar. We defined the notion of lexicalization of a grammar. We showed how a CFG can be 'lexicalized' by using only substitution. But the use of adjunction permits 'lexicalization' with linguistically motivated structures. TAGs have been shown to be naturally 'lexicalized'. Then we gave an overview of the specific lexicon of TAGs. The %xicalization ~ of grammar lead us to introduce a two step parsing strategy. The first step picks up the set of structures corresponding to each word in the sentence. The second step puts the argument structures into predicate structures. Therefore, the relationship between the morphological and syntactic analyses has been modified. In the first step, structures instead of categories are associated with lexical items. The strategy has been shown to be able to use non-local information in the input string. Also problems due to recursion are eliminated.</Paragraph>
    <Paragraph position="1"> The grammar of the parser has been reduced to a set of structures whose size is proportional to the length Of the input sentence. Furthermore, the parsing strategy applies to any parsing algorithm; in particular top-down. It can be formalized into a decidable deduction system that has finite search space for a sentence of finite length. The Earley-type parser for TAGs has been easily extended to take advantage of this strategy.</Paragraph>
  </Section>
class="xml-element"></Paper>