In this article the author attempts to show that translators' selections, just as text producers' decisions in all acts of communications, are never random but rather 'rhetorically motivated'. In doing so, the author proposes to review Vinay and Darbelnet's original techniques and the incisive criticism that has been made of them. This discussion raises fundamental questions of methodology and some methods in translation studies are reviewed, followed by an attempt to resurrect some of the techniques within a text-linguistic framework. In discussing a set of attested examples of translation practice, the author concentrates on 'borrowing' and, to a lesser extent, 'modulation. The author's fundamental position is that all matters affecting the texture of texts are deemed inevitably to be affected by users' perceptions of the communicative, pragmatic and semiotic dimensions of context.