This field has been examined from different points of view and according to different methodological perspectives: three paradigms. Nowadays: phase of transition. Innovative conceptions are needed, - theory and teaching certainly do not exclude each other, - several fundamental questions have not been answered as yet. Perspectives. Few descriptive data are available, usually gathered with precise but limited aims in mind. A great quantity and variety of corpus material is needed, a sort of data bank comprising many language combinations, as well as the methods used by numerous interpreters (whether professional or not). From the different kinds of research on the data base the author selected the following three: 1. analysis of functional style, 2. typology and frequency of the most common errors, 3. interpreters' strategies.