<?xml version="1.0" standalone="yes"?>
<Paper uid="W98-1114">
  <Title>Can Subcategorisation Probabilities Help a Statistical Parser?</Title>
  <Section position="6" start_page="124" end_page="124" type="concl">
    <SectionTitle>
4 Conclusions
</SectionTitle>
    <Paragraph position="0"> We surveyed recent work on automatic acquisition from corpora of subcategorisation and associated frequency information. We described an experiment with a wide-coverage statistical grammar and parser for English and subcategorisation frequencies acquired from 10 million words of text which shows that this information can significantly improve the accuracy of recovery of grammatical relation specifications from a test corpus of 500 sentences covering a number or different genres.</Paragraph>
    <Paragraph position="1"> Future work will include: investigating more principled probabilistic models; addressing immediate lower-level shortcomings in the current system as discussed in section 3.4 above; adding mod(ification) GR annotations to the test corpus and extending the parser to also return these; and working on incorporating selectional preference information that we are acquiring in other, related work (McCarthy, 1997).</Paragraph>
  </Section>
class="xml-element"></Paper>