<?xml version="1.0" standalone="yes"?>
<Paper uid="P90-1002">
  <Title>I Logical Form = Argument Structure Z Surface Structure -- Intonation Structure = Information Structure I Phdegndegldeggi Pdegrm I</Title>
  <Section position="7" start_page="14" end_page="15" type="concl">
    <SectionTitle>
CONCLUSION
</SectionTitle>
    <Paragraph position="0"> The pathway between phonological form and interpretation can now be viewed as in Figure 2:  Such an architecture is considerably simpler than the one shown earlier in Figure 1. Phonological form maps via the rules of combinatory grammar directly onto a surface structure, whose highest level constituents correspond to intonational constituents, annotated as to their discourse function. Surface structure therefore subsumes intonational structure. It also subsumes information structure, since the translations of those surface constituents correspond to the entities and open propositions which constitute the topic or theme (if any) and the comment or rheme. These in 11The inclusion in the full grammar of further roles of type-raising in addition to the subject rule discussed above means that the set of categories over which ~ ranges is larger than it is possible to reveal in the present paper. (For example, it includes object complements). See the earlier papers and \[17\] for digcussion.  turn reduce via functional application to yield canonical function-argument structure, or &amp;quot;logical form&amp;quot;. There may be significant advantages for automatic spoken language understanding in such a theory.</Paragraph>
    <Paragraph position="1"> Most obviously, where in the past parsing and phonological processing have tended to deliver conflicting structural analyses, and have had to be pursued independently, they now are seen to be in concert. That is not to say that intonational cues remove all local structural ambiguity. Nor should the problem of recognising cues like boundary tones be underestimated, for the acoustic realisation in the fundamental frequency F0 of the intonational tunes discussed above is entirely dependent upon the rest of the phonology that is, upon the phonemes and words that bear the tune. It therefore seems most unlikely that intonational contour can be identified in isolation from word recognition. 12 What the isomorphism between syntactic structure and intonational structure does mean is that simply structured modular processors which use both sources of information at once can be more easily devised.</Paragraph>
    <Paragraph position="2"> Such an architecture may reasonably be expected to simplify the problem of resolving local structural ambiguity in both domains. For example, a syntactic analysis that is so closely related to the structure of the signal should be easier to use to &amp;quot;filter&amp;quot; the ambiguities arising from lexical recognition.</Paragraph>
    <Paragraph position="3"> However, it is probably more important that the constituents that arise under this analysis are also semantically interpreted. The interpretations are directly related to the concepts, referents and themes that have been established in the context of discourse, say as the result of a question. These discourse entities are in turn directly reducible to the structures involved in knowledge-representation and inference.</Paragraph>
    <Paragraph position="4"> The direct path from speech to these higher levels of analysis offered by the present theory should therefore make it possible to use more effectively the much more powerful resources of semantics and domain-specific knowledge, including knowledge of the discourse, to filter low-level ambiguities, using larger grammars of a more expressive class than is currently possible. While vast improvements in purely bottom-up word recognition can be expected to conrinue, such filtering is likely to remain crucial to successful speech processing by machine, and appears to be characteristic of all levels of human processing, for both spoken and written language.</Paragraph>
    <Paragraph position="5"> 12This is no bad thing. The converse also applies: intonation contour effects the acoustic rcalisation of words, particularly with respect to timing. It is therefore likely that the benefits of combining intonational recognition and word recognition will be mutual.</Paragraph>
  </Section>
class="xml-element"></Paper>