<?xml version="1.0" standalone="yes"?>
<Paper uid="P96-1017">
  <Title>Coordination as a Direct Process</Title>
  <Section position="4" start_page="124" end_page="124" type="metho">
    <SectionTitle>
2 A brief description of Basic Data
</SectionTitle>
    <Paragraph position="0"/>
    <Section position="1" start_page="124" end_page="124" type="sub_section">
      <SectionTitle>
and Constraints
</SectionTitle>
      <Paragraph position="0"> The classical typology of coordination, i.e. coordination of constituents (1) and of non-constituents, hides some regularity of the phenomenon as it focuses on concepts of constituent and syntactic category. null A coordination of constituents is interpreted as one phrase without any gap. The constituents may be of the same category (1) as well as of different categories (2)-(3). However, this last case is constrained as examplified hereafter 2.</Paragraph>
      <Paragraph position="1">  (2) Je sais son gge et qu'elle est venue ici.</Paragraph>
      <Paragraph position="2"> (I know her age and that she came here.) (2a) Je sais son PSge et son adresse.</Paragraph>
      <Paragraph position="3"> (I know her age and her address.) (2b) Je sais qu'elle a 30 ans et qu'elle est venue ici. (I know that she is 30 and that she came here.) (2c) *Je sais PS Marie et qu'elle est venue ici. *(I know to Marie and that she came here.) (2d) 3e demande l'addition et que quelqu'un paie. (I ask for the bill and for someone to pay.) (2e) *\]e rends \]'addition et que quelqu'un paie.  *(I give back the bill and someone to pay.) In these examples, the coordinate structure acts as the argument of the verb. This verb must subcategorize for each constituent of the coordination and this is not the case in example (2c)-(2e). Note that modelizing coordination of different categories as the unification (i.e. underspecification) of the different categories would lead to accept the six examples or wrongly reject (2d) according to the descriptions used 3.</Paragraph>
      <Paragraph position="4"> Coordination of more than one constituent are often classified as Conjunction Reduction (4), Gapping (la-lb) and Right Node Raising (5) (Hudson, 1976). (la) Jean danse la valse et Pierre, le tango.</Paragraph>
      <Paragraph position="5"> (Jean dances the waltz and Pierre the tango.) (lb) Hier, Jean a dans~ la valse et aujourd'hui, le tango.</Paragraph>
      <Paragraph position="6"> (Yesterday, Jean danced the waltz and today, the tango.) In the case of Gapping structures, the subject (la) and/or an extracted element (lb) is present in the two sides. The only allowed coordinated structure is \[Jean danse la valse\] et \[Pierre le tango\] for (la) and \[Hier, Jean a dansd la valse\] et \[aujourd'hui, le tango\] for (lb) as wh-sentences on other parts (\[la valse\] el \[Pierre\]or \[la valse\] el \[Pierre le langoj~ are impossible.</Paragraph>
      <Paragraph position="7"> A contrario, in the case of Conjunction Reductions, wh-sentences as well as cliticization are al- null lowed referring to what follows the verb (as for coordination of constituents) and treating the arguments simultaneously on the two parts of the coordination: (4a) Je sais k qui demander un v~lo etune canne p~che.</Paragraph>
      <Paragraph position="9"> (4c) Je leur demande un v~lo etune canne ~ p~che.</Paragraph>
      <Paragraph position="10"> (I ask them for a bike and for a fishing rod.) (4d) Je les leur demande.</Paragraph>
      <Paragraph position="11"> (I ask them for them.) Let us remark that a comma is inserted between Marie and sa canne ~ p~che in case of extraction before el as in (lb), indicating the two sentences have not necessarily to be analyzed in the same way: (4e) Je demande PS Pierre son v~lo et PS Marie sa canne ~ p~che.</Paragraph>
      <Paragraph position="12"> (I ask Peter for his bike and Marie for her fishing rod.) (4f) A Pierre, je demande son v~lo et PS Marie, sa canne ~ p~che.</Paragraph>
      <Paragraph position="13"> (Peter, I ask for a bike and Marie, for a fishing rod.) Two structures are available in case of Conjunction Reductions. One structure corresponds to a co-ordination of sentences with a gap of the verb after el, the other one consists in taking the coordinate parallel sequence of constituents as only one structure. The previous facts argue for the second possibility (see also section 3 for criticism of deletion approach).</Paragraph>
      <Paragraph position="14"> Last, note that gapping the verb is less compatible with head-driven mechanisms (and the comma in (4f) could be such a head mark, see (BEF, 1996) for an analysis of Gapping coordinations). It seems then that the structure needed for Conjunction Reduction is some generalization of the standard structure used for coordination of constituents. Our proposal is then focused on this extension. We do not care of Gapping cases as their linguistic properties seem to be different.</Paragraph>
      <Paragraph position="15"> It remains to integrate Right-Node Raising and to extend these cases to more complicated ones. Section 4 includes examples of such cases and shows that our proposal can manage them adequately.</Paragraph>
    </Section>
  </Section>
  <Section position="5" start_page="124" end_page="125" type="metho">
    <SectionTitle>
3 Previous Approaches
</SectionTitle>
    <Paragraph position="0"> There exists a classical way to eschew the question &amp;quot;what can be coordinated ?&amp;quot; if one assumes a deletion analysis. Indeed, according to this approach (Chomsky, 1957; Banfield, 1981), only coordination of sentences are basic and other syntagmatic coordinations should be considered as coordinations of reduced sentences, the reduction being performed by deleting repeated elements. This approach comes up  against insurmountable obstacles, chiefly with the problem of applying transformation in reverse, in the analysis process (Schachter, 1973).</Paragraph>
    <Paragraph position="1"> A direct approach has been proposed at once by Sag &amp; al. (Sag et al., 1985) within the framework of Generalized Phrase Structure Grammar (GPSG), by (Pollard and Sag, 1994) within HPSG, and (Bresnan, 1986) within Lexical Functional Grammar (LFG). These approaches have tried to account for coordination of different categories in reducing the constraint from requiring the same category for conjuncts to a weaker constraint of category compatibility. Whatever the nature of subcategorization information may be, syntactical in GPSG, hybrid in HPSG, functional in LFG, two categories are compatible if they subsume a &amp;quot;common denominator&amp;quot;, in this case a common partial structure.</Paragraph>
    <Paragraph position="2"> Technically, the compatibility is checked by computing a &amp;quot;generalization&amp;quot; of categories and imposing the generalization comprises all features expected in the given context. For example, the context in (6), that is, the verb ~tre (to be), expects a predicative argument and both categories NP and AP are just predicative categories.</Paragraph>
    <Paragraph position="3"> (6) I1 est le p~re de Marie et tier de l'~tre.</Paragraph>
    <Paragraph position="4"> (He is Mary's father and proud of it.) However, this solution cannot be applied generally because all coordinations have not such &amp;quot;natural&amp;quot; intersection (see (2)). So we claim that we have nothing else to do but explicitly enumerate, within the head subcategorization feature, all the structures allowed as complement.</Paragraph>
  </Section>
class="xml-element"></Paper>