<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0904">
  <Title>OntoSem and SIMPLE: Two Multi-Lingual World Views</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Ontological Semantics (OntoSem) is a multi-lingual text processing environment that takes as input unrestricted text and, using a suite of static resources and processors, automatically creates text-meaning representations (TMRs) which can then be used as the basis for any NLP application, including MT, question answering, summarization, etc. OntoSem knowledge resources are developed in coordination with each other and with the processors they serve. Some of the resources are fully language independent while others are readily parameterizable, wherein lies the cross-linguistic portability of the system. Although in this paper, we focus on OntoSem lexicons, a crucial point is that they are not built in isolation but, rather, in an integrated environment where their utility can be tested and evaluated in a variety of practical applications.</Paragraph>
    <Paragraph position="1"> The SIMPLE project takes a different approach to achieving the dual goals of multilinguality and resource utility across applications. It aims to develop compatible (they use the term &amp;quot;harmonised&amp;quot;) lexicons for 12 European languages, attempting to foresee what will be most useful for applications but without referring to any particular processors that will use the information and without building any other knowledge resources to share the burden of semantic specification. As we will show, these different points of departure lead to quite different realizations of cross-lingual lexicons for NLP.</Paragraph>
  </Section>
class="xml-element"></Paper>