<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-4016">
  <Title>Sydney, July 2006. c(c)2006 Association for Computational Linguistics TwicPen : Hand-held Scanner and Translation Software for non-Native Readers</Title>
  <Section position="3" start_page="0" end_page="61" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> As a consequence of globalization, a large and increasing number of people must cope with documents in a language other than their own. While readers who do not know the language can find help with machine translation services, people who have a basic fluency in the language while still experiencing some terminological difficulties do not want a full translation but rather more specific help for an unknown term or an opaque expression. Such typical users are the huge crowd of students and scientists around the world who routinely browse documents in English on the Internet or elsewhere. For on-line documents, a variety of terminological tools are available, some of them commercially, such as the ones provided by Google (word translation services) or Babylon Ltd. More advanced, research-oriented systems based on computational linguistics technologies have also been developed, such as GLOSSER-RuG (Nerbonne et al, 1996, 1999), Compass (Breidt et al., 1997) or TWiC (Wehrli, 2003, 2004).</Paragraph>
    <Paragraph position="1"> Similar needs are less easy to satisfy when it comes to more traditional documents such as books and other printed material. Multilingual scanning devices have been commercialized1, but they lack the computational linguistic resources to make them truly useful. The shortcomings of such systems are particularly blatant with inflected languages, or with compound-rich languages such as German, while the inadequate treatment of multi-word expressions is obvious for all languages. TwicPen has been designed to overcome these shortcomings and intends to provide readers of printed material with the same kind and quality of terminological help as is available for on-line documents. For concreteness, we will take our typical user to be a French-speaking reader with knowledge of English and German reading printed material, for instance a novel or a technical document, in English or in German.</Paragraph>
    <Paragraph position="2"> For such a user, German vocabulary is likely to be a major source of difficulty due in part to its opacity (for non-Germanic language speakers), the richness of its inflection and, above all, the number and the complexity of its compounds, as exemplified in figure 1 below.2  This paper will describe the TwicPen system, showing how an in-depth linguistic analysis of the sentence in which a problematic word occurs helps to provide a relevant answer to the reader.</Paragraph>
    <Paragraph position="3"> We will show, in particular, that the advantage of such an approach over a more traditional bilingual terminology system is (i) to reduce the noise with a better selection (disambiguation) of the source word, (ii) to provide in-depth morphological analysis and (iii) to handle multi-word expressions (compounds, collocations, idioms), even when the terms of the expression are not adjacent.</Paragraph>
  </Section>
class="xml-element"></Paper>