<?xml version="1.0" standalone="yes"?>
<Paper uid="H94-1035">
  <Title>Information Based Intonation Synthesis*</Title>
  <Section position="5" start_page="195" end_page="195" type="metho">
    <SectionTitle>
4. Results
</SectionTitle>
    <Paragraph position="0"> The IBIS system produces distinct intonational differences in minimal pairs of queries like those in examples (10)-(13) below. These minimal pairs illustrate the system's capability for producing appropriately different intonation contours for a single string of words under the control of discourse context. If the responses in these</Paragraph>
  </Section>
  <Section position="6" start_page="195" end_page="195" type="metho">
    <SectionTitle>
4 Full descriptions of the CCG generation algorithm are given
</SectionTitle>
    <Paragraph position="0"> in \[17\].</Paragraph>
    <Paragraph position="1"> examples are interchanged, the results sound distinctly unnatural in the given contexts.</Paragraph>
    <Paragraph position="2">  Examples (10) and (11) illustrate the necessity of the theme/theme distinction. Although the pitch accent locations in the responses in these examples are identical, occurring on lacerations and bleeding, the alternation in the theme and theme tunes is necessary to convey the intended propositions in the given contexts. Examples (12) and (13) show that the system makes appropriate distinctions in focus placement within themes and themes based on context. More complex examples, like those shown in (14)-(16), illustrate the usefulness of the contrastive stress algorithm for assigning pitch accents in appropriate locations. 5</Paragraph>
  </Section>
class="xml-element"></Paper>