<?xml version="1.0" standalone="yes"?>
<Paper uid="H89-1043">
  <Title>INTEGRATING SPEECH AND NATURAL-LANGUAGE PROCESSING</Title>
  <Section position="6" start_page="246" end_page="246" type="concl">
    <SectionTitle>
SUMMARY
</SectionTitle>
    <Paragraph position="0"> We believe the results reported in this paper have demonstated:  with only minimal modifications * A clear path toward including important semantic constraints on recognition, and possibly prosodicphonological/syntactic constraints as well These results suggest that the dynamic-grammar-network approach to the integration of speech and NL processing may well be the key technology needed to produce true spoken-language systems with real-time performance.</Paragraph>
  </Section>
class="xml-element"></Paper>