Translation Skill-Sets in a Machine-Translation Age

The integration of data from statistical machine translation into translation memory suites (giving a range of TM/MT technologies) can be expected to replace fully human translation in many spheres of activity. This should bring about changes in the skill sets required of translators. With increased processing done by area experts who are not trained translators, the translator’s function can be expected to shift to linguistic postediting, without requirements for extensive area knowledge and possibly with a reduced emphasis on foreign-language expertise. This reconfiguration of the translation space must also recognize the active input roles of TM/MT databases, such that there is no longer a binary organization around a “source” and a “target”: we now have a “start text” (ST) complemented by source materials that take the shape of authorized translation memories, glossaries, terminology bases, and machine-translation feeds. In order to identify the skills required for translation work in such a space, a minimalist and “negative” approach may be adopted: first locate the most important decision-making problems resulting from the use of TM/MT, and then identify the corresponding skills to be learned. A total of ten such skills can be identified, arranged under three heads: learning to learn, learning to trust and mistrust data, and learning to revise with enhanced attention to detail. The acquisition of these skills can be favored by a pedagogy with specific desiderata for the design of suitable classroom spaces, the transversal use of TM/MT, students’ self-analyses of translation processes, and collaborative projects with area experts.