Communication and Translation in a Multilingual Corporate Environment.

In the paper an attempt is made to find a unifying approach to the study of the translator’s praxis, assuming that translation is guided by certain, recognizable, semiotic processes. Computational, corpus-based methods intended to aid in the research of large text bases are introduced. Alignment of text segments from files in different languages contained in a corpus, where these text files are known to be mutual translations is described. Text encoding in order to allow comparison of the results of translation studies performed by different scholars is also demonstrated. One goal is to establish qualitative and quantitative variables, on the sentential as well as the textual level, which would permit generalizations about the concrete procedures performed by professional translators in authentic work situations, e.g. in multi-lingual corporate environments. Empirical, descriptive methods Today large amounts of texts sit on the hard disks of computers in companies and organizations, but exact, empirical, detailed, descriptive information telling us what translators actually do when they translate is not abundant. A natural solution to this dilemma is the collection of evidence from existing texts included in aligned bilingual corpora. The purpose of text alignment is to establish version complexes1, i.e. sets of corresponding elements in the source and target texts