Che cosa si aspettano le società di servizi di traduzione dai propri traduttori interni ed esterni, in qualità di 'fornitori'. Discussione sulle proposte per una formazione adeguata

Abstract

By "language service providers" we mean both freelance translators (and interpreters) or employees of companies, and companies offering translation (and interpreting) services on the free market (i.e., "translation agencies"). In this article, we will focus on the expectations that translation service providers have of their suppliers, i.e., translators, when they are asked to post-edit a target text after using Neural Machine Translation (NMT). Obviously, the expectations of translation companies are reflected in the training of the translators. The NMT programmes currently on the market are constantly improving and are increasingly being adopted by translation companies because their product, from a very "raw" and unusable material, is gradually becoming "tractable". However, this "tractability" is mainly a function of the correctness and reliability of the source text. Since translation companies working on the free market do not, and cannot, have control over the source text, it is difficult to establish precise rules for improving the NMT output upstream. Therefore, for now, we can for now only give post-editors indications about the typical errors that we have detected in years of experience working with NMT. The greater attention expected of post-editors, compared to translators working without NMT, goes hand in hand with the correct interpretation and understanding of the source text, despite the shortcomings it often presents. How these indications can be translated into proposals for the training of future translators is not obvious, but we wish to emphasise the fundamental contribution of practice and experience.