<?xml version="1.0" standalone="yes"?>
<Paper uid="W94-0333">
  <Title>Cross Modal Comprehension in ZARDOZ An English to Sign-Language Translation System</Title>
  <Section position="4" start_page="250" end_page="250" type="intro">
    <SectionTitle>
~.4 Anaphora and Spatial Designation
</SectionTitle>
    <Paragraph position="0"> Anaphoric resolution is required in translation whenever source and target languages use different anaphofic discrimination systems. For instance, the English &amp;quot;They&amp;quot;, neutral in gender, can map onto either of &amp;quot;Elles&amp;quot; or &amp;quot;Ils&amp;quot; in French. It is thus necessary to resolve the reference of a pronoun before translation, so that the correct referring term can be generated in the target language. ZARDOZ employs the basic Hobbs algorithm for this task (see Hobbs 1978), augmented with discourse registers which track the movement of referents between peripheral and central focus.</Paragraph>
    <Paragraph position="1"> Sign language makes use of index locations in space to refer to entities in a conversation. Thus, locations should be allocated sign space in such a way that possible referential conflicts are minimised.</Paragraph>
    <Paragraph position="2"> ZARDOZ strives to allocate a different locations to the major cases of an utterance (e.g., agent, patient), and to maintain those assignments throughout a narrative as far as is possible.</Paragraph>
    <Section position="1" start_page="250" end_page="250" type="sub_section">
      <SectionTitle>
5.2 Word Order in Sign
</SectionTitle>
      <Paragraph position="0"> Word order, the dominant syntactic constraint in English, has a reduced role in ASL which can also employ the dimensions of space to indicate case roles (see Liddell 1980), The referents of a verb may be established at index locations in signing space, and the direction of movement of the verb between locations then indicates which is the agent and which the patient.</Paragraph>
      <Paragraph position="1"> For example, if B1LL is signed on the left and MARY on the fight, then a left to right motion while signing the verb ASL-CHASE, indicates BILL is the pursuer and MARY the pursued. Thus using the passive voice in ASL is simply a matter of reversing the order of agent and patient. Of course, the verb/predicate will now have to be signed after both agent and patient have been articulated. The SD-graph representing this transformation is presented in Figure 4(a).</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>