<?xml version="1.0" standalone="yes"?>
<Paper uid="C86-1041">
  <Title>SITUAT\]Or~JA L</Title>
  <Section position="3" start_page="0" end_page="174" type="intro">
    <SectionTitle>
2, Presupposition
</SectionTitle>
    <Paragraph position="0"> Before formalizing pr(~supposition, we shah consider the important question: &amp;quot;what is a presupposition?&amp;quot; The answer to this question Is the kev to the construction of a formal theory of presupposition.</Paragraph>
    <Paragraph position="1"> We find in the literature- several definitions of presupposition. For example, many philosophers and linguists assume the definition generally represented as follows:  This definition leads the undesired conclusion that B Is a +.autology. Clearly an improved definl'tion is in order. Karttunen (1973) gives the following as an alternative: (IDef 2) A ~tically presupposesB relative tq a set of assumed facts C iff it is riot acceptable to utter A In the context C unless C entails B, This definition says that a presupposition is an entailment of the sentence in a context. Regretablv, nowever, there are no formal definitions for such terms as 'entails', 'relative to',  'context, in the above definition, Gazdar (1979),'on the other hand, gives the following definition.</Paragraph>
    <Paragraph position="2"> (Def 3) praBmatic presupposition is entailed bv the context in favour of the weaker requirement that they be consistent with the context.</Paragraph>
    <Paragraph position="3"> There he further gives the fallowing informal definitions of essential terms in the above definition @s follows,  (Def 3.1) A sentence A is entailed bv a set of sentence B Just in case A is true in every possible world in which all members of B are true.</Paragraph>
    <Paragraph position="4"> (Def 3.2) A sentence A is consistent with a set of  sentences 8 just in case A is true tn some possible world in which all members of B are true.</Paragraph>
    <Paragraph position="5"> Thus Gazdar's definition crucia}lv depends on the notion of consistency.</Paragraph>
    <Paragraph position="6"> Although llis definition seems to be more plausible than the other definitions based en entailments, as it enables us to accommodate the so-called projection prob!em at ease, yet It is not' entirely free from shortcomings, HIS theory is based on possible-world semantics, which is not quite adequate as a natural language semantics. The Inappropriateness of such a theory is discussed in BarwJse and Perry (1983), Akama (1986) in detail. From a computational point of view, especially many of its deficiencies can be stated, To say the least possible-world semantics seems to fail to deal with partial information in an effective way, And in theories in this tradition only restricted statements can be derived from its model that is, accessibility relations affect logical structures in the model. Moreover although unrealistic objects may be possible in a model, they are not suitable for a computational paradigm.</Paragraph>
    <Paragraph position="7"> In computer science Mercer and Relter (1982) formulated presupposition, more or less In Gazdar's spirit, as an inference generated from a pragmatic rule, namely, default rule. Since their formalism ts based on first-order theory, similar shortcomings as in the case of passible-world semantics can be pointed cut. It is, however, interesting to notice its flexlbiiity in the application to knowledge representation.</Paragraph>
    <Paragraph position="8"> To overcome the above mentioned difficulties in traditional theories, we introduces SS as an underlying theory for presupposition below,</Paragraph>
  </Section>
class="xml-element"></Paper>