Parallellkorpusen som resurs i lexikografiskt arbete

Resumé

This paper discusses the use of parallel corpora as a tool in bilingual lexicography. After a brief background sketch with reference to literature and (lack of) reported ongoing projects, the text briefly considers the notion of a parallel corpus, before it reaches the main subject: a hands-on study of the opportunities connected to the use of parallel corpora in lexicography. The corpus used for the study is a relatively small, about 115,000-token, ad-hoc corpus compiled of texts from the Nordic Council. The languages investigated are Icelandic on the one hand, and Danish, Norwegian and Swedish on the other. One advantage of parallel corpora seems to be the possibility to compare concordances of n-grams between the studied languages. The two Icelandic verbs probed as examples were shown to be part of hitherto unnoticed equivalent pairs, where one or both sides turned out to be MWEs. Co-occurring patterns between active or passive voice and certain translation equivalents have also been noticed.