Since the 1970s the notion of "translation competence" has been viewed as at least 1) a mode of bilingualism, open to linguistic analysis, 2) a question of market demands, given to extreme historical and social change, 3) a multicomponent competence, involving sets of skills that are linguistic, cultural, technological and professional, and 4) a "supercompetence" that would somehow stand above the rest. The general trend among theorists has been to expand the multicomponent model so as to bring new skills and proficiencies into the field of translator training. This trend may be expected to continue with the increasing use of electronic tools. Here it is argued, however, that the multicomponential expansions of competence are partly grounded in institutional interests and are conceptually flawed in that they will always be one or two steps behind market demands. On the other hand, a simple minimalist concept of translation competence, based on the production then elimination of alternatives, can help orient translator training in times of rapid technological and professional change.