Customised Corpora – a source of information for translators
To translate is to make choices. The quality of the final text is directly related and to a great extent depends on the quality of these choices. Every time a translated text moves away from the reality of its target public, the result is considered "strange" or "unnatural". The translator seeks precision, accuracy, and can find it in glossaries, encyclopaedias, textbooks and grammars. However, these sources merely bring "what it should be like" and fail to emphasize "what it is like". Naturalness is the result of the use of phrases, terms and expressions which sound natural to those who will have contact with the translated text.
Ready corpora, which are available on CD-ROMS and on the Internet, usually deal with general language and mix different registers and types of documents.
In order to establish a customized corpus which attends to specific research needs (register, linguistic variants, types of documents, target public), the translator should follow certain basic steps, which, as they are incorporated into their work routine, will become automatic, quick and will help to make the final text more natural and precise.
Customized corpora may be modified, enriched or altered whenever necessary, in order to be used in a new translation project, or even to continue a current project.
In the initial stage of data collection, translators may have the feeling that they have used more time than is necessary for conventional research, but this feeling is unfounded. When using the customized corpus, translators will realize that the responses are much more precise, quick and coherent.
In the research carried out with customized corpora, the results are homogeneous, there are no suspicious terms, and all the phrases come from pre-selected texts. The translator will economize time, security is increased, and there is a guarantee of a more coherent and precise result. Thus, in order to reach a quality translation, translators will not have to abandon their research on the web but rather will use the most suitable texts they find, storing them in a programme which has the required corpora research tools, and, whenever necessary, will use this safe source which they themselves have organized - their customized corpus.
Click for presentation – PDF file, 3.2MB