Applying machine translation (MT) to literary texts involves the same domain shift challenges that arise for any sublanguage (e.g. medical or scientific). However, it also introduces additional challenges. One focus in the discussion of translation theory in the humanities has been on the human translator's role in staying faithful to an original text versus adapting it to make it more familiar to readers. In contrast to other domains, one objective in literary translation is to preserve the experience of reading a text when moving to the target language. We use existing MT systems to translate samples of French literature into English. We then use qualitative analysis grounded in translation theory and real example outputs in order to address what makes literary translation particularly hard and the potential role of the machine in it.
展开▼
