Using Parallel Corpora for Translation-Oriented Term Extraction

In many scientific, technological or political fields terminology and the production of up-to-date reference works is lagging behind, which causes problems to translators and results in inconsistent translations. Parallel corpora of texts already translated can be used as a resource for automatic extraction of terms and terminological collocations. Especially for smaller languages where existing resources are scarce, collecting and exploiting parallel corpora may be the chief method of obtaining terminological data. The paper describes how a methodology for multi-word term extraction and bilingual conceptual mapping was developed for Slovene-English terms. We used word-to-word alignment to extract a bilingual glossary of single-word terms, and for multi-word terms two methods were tested and compared. The statistical method is broadly applicable but gives results of very limited use, while the method of syntactic patterns extracts highly useful terminological phrases, however only from a tagged corpus. A vision of further development is given and how these methods might be incorporated into existing translation tools.