<?xml version="1.0" standalone="yes"?>
<Paper uid="W98-0315">
  <Title>Anchoring a Lexicalized Tree-Adjoining Grammar for Discourse</Title>
  <Section position="3" start_page="90" end_page="91" type="intro">
    <SectionTitle>
3 Cue Phrases as Feature Structures
</SectionTitle>
    <Paragraph position="0"> (Knott, 1996) Earlier we noted that there was benefit to be gained from taking the anchors of elementary trees to be feature structures into which discourse cues (whose semantics was also in terms of feature structures) could substitute. Here we briefly argue why we believe this is so.</Paragraph>
    <Paragraph position="1"> First, in viewing discourse cues in terms of feature structures, we are following recent work by Knott (1996; 1996). Knott's study of the substitutability patterns of discourse cues reveals that their four common patterns - synonymy, exclusivity, hypernymy/hyponymy and contingent substitutability -</Paragraph>
    <Paragraph position="3"> can, by assuming inheritance (that, except for contingent substitutability, a substitution pattern that holds for a discourse cue also holds for all its hyponyms), follow from interpreting cues in featuretheoretic terms: * If cue a is synonymous with cue 3, they signal the same values of the same features.</Paragraph>
    <Paragraph position="4"> * If a is exclusive with /3, they signal different values of at least one feature.</Paragraph>
    <Paragraph position="5"> * If a is a hypernym of ~3,/3 signals all the features that o~ signals, as well as some features for which a is undefined.</Paragraph>
    <Paragraph position="6"> * If a and ,3 are contingently substitutable, c~ and fl signal some of the same features, but a is also defined for a feature for which /3 is undefined and /3 is defined for a feature for which a is undefined.</Paragraph>
    <Paragraph position="7"> Drawing on the extensive literature devoted to individual cue phrases, Knott provided semantics for some of these features in terms of preconditions on their use and/or their communicative effects.</Paragraph>
    <Paragraph position="8"> Following Knott in treating discourse cues in terms of feature structures, it also appears beneficial to treat tree anchors as feature structures as well, distinct from those of discourse cues.</Paragraph>
    <Paragraph position="9"> The reason for treating the anchor of subordinate clause initial trees as feature structures is one of representational efficiency: we can posit fewer such trees if we take their anchors to be features structures that allow the (possibly contingent ) substitution of any subordinate conjunction with a compatible feature structure. For example, we can have one tree whose anchor has the feature restrictedsituation, that can be realized as either &amp;quot;if&amp;quot; or &amp;quot;when&amp;quot; in some texts, but only &amp;quot;when&amp;quot; in others -- e.g.: 15. Emergency parking regulations are in force \[when, if\] more than six inches of snow has fallen.</Paragraph>
    <Paragraph position="10"> 16. I found 30 new messages had arrived \[when, *if\] I logged on this morning.</Paragraph>
    <Paragraph position="11"> (Knott and Mellish, 1996) distinguish &amp;quot;if&amp;quot; and &amp;quot;when&amp;quot; by their different values for the feature modal status: &amp;quot;when&amp;quot; has the value actual, while &amp;quot;if&amp;quot; has the value hypothetical. One can therefore say that other semantic features in Ex. I6 conflict with the value hypothetical, only allowing &amp;quot;when&amp;quot;. (N.B. One could also take &amp;quot;'when&amp;quot; as being unmarked for modal status, its hypothetical reason begin synonymous with &amp;quot;whichever&amp;quot;. The conflict with &amp;quot;if&amp;quot; in Ex. 16 would still follow.) The argument for treating the pair of anchors of parallel structures as feature structures follows from the variability in the realization of the medial anchor noted in Section 2.1. One way to account for this is that the anchor has features separate from those of the discourse cues. Any cue can then be used to realize the anchor, as long as it is either * less specific than the anchor, as in Ex. 3 - &amp;quot;but&amp;quot; has few features in Knott's taxonomy; * more specific than the anchor, as in Ex. 5 &amp;quot;on the other extreme&amp;quot;, although it does not appear in Knott's taxonomy, intuitively appears to mean more than just &amp;quot;side&amp;quot;.</Paragraph>
    <Paragraph position="12"> * partially overlapping with the anchor, as in Ex. 4 - &amp;quot;at the same time&amp;quot; has temporal  features, but does not seem intrinsically contrastive. This corresponds to Knott's concept of contingent substitutability.</Paragraph>
    <Paragraph position="13"> It also appears as if the clause/discourse within the scope of an anchor can either reinforce its features (as in Ex. 17 below) or convey features of the anchor when it is not itself realized lexically, as in Ex. 18: 17. On the one hand, according to Fred, John is very generous, On the other hand, according to everyone else, he will only give if he sees an angle.</Paragraph>
    <Paragraph position="14"> 18. According to Fred, John is very generous. According to everyone else, he will only give if he sees an angle.</Paragraph>
    <Paragraph position="15"> But this part of our work is more speculative and the subject of needed future work.</Paragraph>
  </Section>
class="xml-element"></Paper>