<?xml version="1.0" standalone="yes"?>
<Paper uid="P04-1082">
  <Title>Using linguistic principles to recover empty categories</Title>
  <Section position="7" start_page="5" end_page="5" type="concl">
    <SectionTitle>
7 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper I have presented an algorithm for the recovery of empty categories in PTB-style trees that otherwise lack them. Unlike previous approaches, the current algorithm is rule-based rather than learning-based, which I have argued is appropriate for this task, given the highly theoretical nature of empty categories in the PTB.</Paragraph>
    <Paragraph position="1"> Moreover, the algorithm has no access to lexical information such as valency or verb class.</Paragraph>
    <Paragraph position="2"> Using the string-based evaluation metric proposed by Johnson (2002), the current system outperforms previously published algorithms on detection alone, as well as on detection combined with resolution, both on perfect input and in combination with parsing. In addition, we have performed additional evaluation using a tree-based metric, and including an evaluation of function tag assignment as well.</Paragraph>
  </Section>
class="xml-element"></Paper>