<?xml version="1.0" standalone="yes"?>
<Paper uid="N01-1021">
  <Title>A Probabilistic Earley Parser as a Psycholinguistic Model</Title>
  <Section position="8" start_page="8" end_page="8" type="concl">
    <SectionTitle>
Conclusion
</SectionTitle>
    <Paragraph position="0"> These examples suggest that a \total-parallelism&amp;quot; parsing theory based on probabilistic grammar can characterize some important processing phenomena.</Paragraph>
    <Paragraph position="1"> In the domain of structural ambiguity in particular, the explanation is of a di erent kind than in traditional reanalysis models: the order of processing is not theoretically signi cant, but the estimate of its magnitude at each point in a sentence is. Results with empirically-derived grammars suggest an a rmative answer to Gibson and Pearlmutter's ques- null The di erence in probability between subject and object rules could be due to the work necessary to set up storage for the ller, e ectively recapitulating the HOLD Hypothesis (Wanner and Maratsos, 1978, page 119) tion: phrase-level contingent frequencies can do the work formerly done by other mechanisms.</Paragraph>
    <Paragraph position="2"> Pursuit of methodological principles 1, 2 and 3 has identi ed a model capable of describing some of the same phenomena that motivate psycholinguistic interest in other theoretical frameworks. Moreover, this recommends probabilistic grammars as an attractive possibility for psycholinguistics by providing clear, testable predictions and the potential for new mathematical insights.</Paragraph>
  </Section>
class="xml-element"></Paper>