Translating what the image conveys or what it arouses?

An ideal strongly anchored in the realm of translation is that of “translating without additions or modifications.” However, with multimodal texts, one is confronted with the problem posed by the image, its reading, and its interpretation. This article aims to better delineate the interpretative threshold between the global meaning that a still image might convey in and of itself, on the one hand, and the more personal interpretations that this image can arouse in its receiver (including the translator) on the other. In passing, the article also aims to suggest new ways of sensitizing translation students to the existence of such a threshold. The principle, based on Sperber and Wilson’s relevance theory and its strong vs. weak implicature continuum, is that the visual and compositional cues of the image, their relative salience, and their eventual semantic convergence, combined with contextual factors in the initial production of the visual document, would constitute the fundamental semiotic data to be considered in the translation process. Conversely, any projection of meaning external to the image, and emanating from the translator-interpreter himself, would have to be treated with more circumspection within that process, since it would amount to recontextualizing the original visual message by coloring it with a particular meaning, in other words modifying it through added meanings. The corpus used for the observations consists of analyses of a photographic image made by MA-level students in Translation Studies.