<?xml version="1.0" standalone="yes"?>
<Paper uid="H86-1008">
  <Title>A Terminological Simplification Transformation for Natural Language Question-Answering Systems 1</Title>
  <Section position="2" start_page="2238" end_page="2238" type="intro">
    <SectionTitle>
1. Introduction
</SectionTitle>
    <Paragraph position="0"> A common and useful strategy for constructing natural language interface systems is to divide the processing of an utterance into two major stages: the first mapping the utterance to a logical expression representing its &amp;quot;meaning&amp;quot; and the second producing from this logical expression the appropriate response. The second stage is not neccesarily trivial: the difficulty of its design is signifigantly affected by the complexity and generalness of the logical expressions it has to deal with. If this issue is not faced squarely, it may affect choices made elsewhere in the system. Indeed, a need to restrict the form of the meaning representation can be at odds with particular approaches towards producing it - as for example the &amp;quot;compositional&amp;quot; approach, which does not seek to control expression complexity by giving interpretations for whole phrasal patterns, but simply combines together the meaning of individual words in a manner appropriate to the syntax of the utterance. Such a conflict is certainly not desirable: we want to have freedom of linguistic action as well as to be able to obtain correct responses to utterances.</Paragraph>
    <Paragraph position="1"> This paper treats in detail the particular manifestation of these issues for natural-language systems which serve as interfaces to a database: the problems that arise in a module which maps the meaning representation to a second logical language for 1The work presented here was supported under DARPA contract ~Neee14-85-.c-0016. The views and conclusions contained in this document ore those of the author and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the Defense Advanced Research Projects Agency or of the United States Government. This paper was originally published in the Proceedings o\[ the 24th Annual Meeting o\[ the Association \[or Computational Linguistics, 10--13 June, 1986, Columbia University, New York. Requests for copies should be addressed to:  expressing actual database queries. A module performing such a mapping is a component of such question-answering systems as TEAM \[4\], PHLIQA1 \[7\] and IRUS \[1\]. As an example of difficulties which may be encountered, consider the question &amp;quot;Was the patient's mother a diabetic?&amp;quot; whose logical representation must be mapped onto a particular boolean field which encodes for each patient whether or not this complex property is true. Any variation on this question which a compositional semantics might also handle, such as &amp;quot;Was diabetes a disease the patient's mother suffered from?&amp;quot;, would result in a semantically equivalent but very different-looking logical expression; this different expression would also have to be mapped to this field. How to deal with these and many other possible variants, without making the mapping process excessively complex, is clearly a problem.</Paragraph>
    <Paragraph position="2"> The solution which this paper presents is a new level of processing, intermediate between the other two: a novel simplification transformation which is performed on the result of semantic interpretation before the attempt is made to map it to the database. This simplification method relies on knowledge which is stored in a taxonomic knowledge representation system such as NIKL \[5\]. The principle behind the method is that an expression may be simplified by translating its subexpressions, where possible, into the language of NIKL, and classifying the result into the taxonomy to obtain a simpler equivalent for them. The result is to produce an equivalent but syntactically simpler expression in which fewer, but more specific, properties and relations appear. The benefit is that deductions from the expression may be more easily &amp;quot;read off&amp;quot;; in particular, the mapping becomes easier because the properties and relations appearing are more likely to be either those of the database or composable from them.</Paragraph>
    <Paragraph position="3"> The body of the paper is divided into four sections. In the first, I will summarize some past treatments of the mapping between the meaning representation and the query language, and show the problems they fail to solve. The second section prepares the way by showing how to connect the taxonomic knowledge representation system to a logical language used for meaning representation. The third section presents the &amp;quot;recursive terminological simplification&amp;quot; algorithm itself. The last section describes the implementation status and suggests directions for interesting future work.</Paragraph>
  </Section>
class="xml-element"></Paper>