Bilingual concordancers: a new tool for bilingual lexicographers

BILINGUAL CONCORDANCERS: A NEW TOOL FOR BILINGUAL LEXICOGRAPHERS Lucie Langlois University of Ottawa ABSTRACT Concordances extracted from aligned bi-texts have become an extremely important tool for the bilingual lexicographer. This paper will show in a concrete way how bi-concordances are actually used in a bilingual dictionary project. 1. INTRODUCTION The Bilingual Canadian Dictionary Project, also called the BCD1, has as its primary objective the preparation of the first comprehensive Canadian French-English English-French general dictionary. Due for publication in 2003 in both paper and CD-ROM formats, it will contain a total of 80,000 entries. The BCD aims to be different from other English/French dictionaries in that it will present English and French as they are used in Canada, including words and word senses that are particular to Canadian usage. Since our typical users will be translators and writers, our entries will give these sophisticated users more equivalents for each headword, as well as more free combinations and collocations than existing bilingual dictionaries do. Another way in which the BCD hopes to improve on existing dictionaries is by its substantial use of corpora2. BCD lexicographers use both unilingual and bilingual corpora. These corpora are searched using concordance generating programs, which we call concordancers, the output of which are consulted by lexicographers in producing and revising entries. BCD lexicographers use various corpora in a multitude of ways. First, they analyze unilingual source language dictionary entries to get a feel for the semantic range of the word being worked on. They then check unilingual source language concordances to see if the semantic divisions suggested by the unilingual dictionaries actually reflect reality. This is not always the case; for instance, there have been many instances where the corpus data suggested more sense divisions than were actually found in existing dictionaries. Next, the lexicographer looks at bilingual dictionaries to see which equivalents have been proposed by other lexicographers. Finally, bilingual concordances are examined to validate and add equivalents. Bilingual corpora are used at different points of the lexicographic process. Although transfer dictionaries will not be mentioned explicitly here, techniques and tools that are useful in producing bilingual dictionaries should also be helpful in setting up dictionaries for machine translation systems. This paper deals solely with the use of bilingual corpora and bilingual concordances in bilingual lexicography. More specifically, it focuses on one particular type of 1 Funded by the Social Sciences and Humanities Research Council of Canada, the project includes three universities (University of Ottawa, Université de Montréal and Université Laval) and employs between 25 and 30 research assistants at any given time. 2 A complete account of the use of monolingual concordances in monolingual lexicography can be found in Sinclair (1987). Church and Gale (1991) suggest ways in which bi-concordancers could be useful in bilingual lexicography. 34 concordancer, namely one searching bi-text, to use Harris' (1988) term, to produce concordances in which a match consists of a source language segment together with its translation. Bilingual concordance generating programs will be referred to as "bi-concordancers" and …

Content truncated to 500 words