Enseigner la traduction humaine en s’inspirant de la traduction automatique

Translation is an activity, which can be performed, today, by human beings and by machines as well. However, the question is: how? How can human beings learn to translate? Which are the strategies chosen in relation to different textual, lexical, syntactic and semantic difficulties? And, how can machines, in their turn, manage a translation? Which operations can they perform to translate a text? Can there exist a possible interaction between human translation learning model and machine translation learning model?
The project presented in this article aims at answering to those questions by means of a pedagogical experimentation, which is founded on the adaptation of an rule-based translation system. The objective is to study the pedagogical interest of “human-machine” interaction in the context of translation learning.