The search for technical collocations and subtechnical vocabulary in native-language corpora as an aid to LSP translation

In LSPs terms combine with other words to create phrases, sentences and finally texts. The percentage of terms compared to other words is relatively low. Producing an adequate LSP text does not only presuppose a knowledge of terms, but also implies reproducing the conventiopnal usage of the domain in question. Paraphrases and unusual word combinations cause competent readers to regard the text as lacking in quality, no matter whether its contents are correct. Therefore, translators need to refer to a wealth of documents to study terms in context so as to produce technical translations which are lexically and grammatically correct. The objective of this paper is to show how analysis of KWIC concordances of key terms and subtechnical vocabulary in computerised corpora can be used to speed up this work to obtain more reliable and systematic data on term usage. Technical collocations and subtechnical vocabulary identified in this way can then be used to integrate information on terms in term collections thus providing valuable help in the production of high-quality translation.