Poetics of machine translation errors applied to artistic practices

Paper presented on March 12th 2015, at the conference on “Hybrid Practices in the Arts, Science and Technology from the 1960’s to today”, celebrated at the University of Kansas (USA), organized by the Spencer Museum of Art.

Early research into machine translation, especially by North American programmers and computer experts in the nineteen-fifties, gave birth to a new discipline called Computational Linguistics. At the same time, another applied branch was being developed, called Translation Studies. Despite their closeness not only in space and time but also in subject, it is paradoxical that even today each is still unaware of the way the other approaches the issue of translation. This paper surveys the different notions that these two disciplines have in common in order to ascertain how they could engage in dialogue with each other. In order to fulfil this goal, this research encompasses several disciplines, including Computational Linguistics and Translation Studies. It is based on a study of two contemporary artworks that deal with the poetics of machine translation errors, namely “automatic voice-over” (2013) by Jan Monsalvatje (SP) and “El 27” (2014) by Eugenio Tisselli (MEX). In the context of this investigation, art is seen as a link that reflects the relations between these fields and the poetic potential of machine translation errors.

Index Terms: machine translation, Computational Linguistics, Translation Studies, contemporary art, machine translation errors, poetics of error