Using Translation Research to Model Word-Image Interaction

Abstract

This paper sets out to examine the potential of translation research as a means to model multimodal meaning construction in an empirical manner. The paper introduces two classifications of word–image relationships put forward in previous multimodality research (Marsh and White 2003; Martinec and Salway 2005). These are compared to empirical analyses from a research project examining the translation of an illustrated technical text. The paper suggests that the reviewed classifications do not cover all of the possible ways in which words and images can co-construct meaning: the translators interpreted the combination of words and images in various ways, some of which are clearly not represented in the classifications. The paper concludes that word–image interaction is more complex than the classifications suggest and that relying on ready-made classifications in our analyses of word–image interaction may limit the way we perceive the manifold nature of multimodal meaning construction.