Literary machine translation

Since the mid-2010s, there has been renewed interest in the automation of literary translation, with both neural machine translation (NMT) and large language models (LLMs) currently applied to the task. Much research in the field has been concerned with technical issues such as how generic systems can be customized for literary translation and how outputs should be evaluated. Other sources focus on the impact of automatic literary translation on translated texts, the translation process, readers’ experience, and the translation market. The consequences for human literary translators of the increased uptake of NMT and LLMs in the profession have also received attention, with translators’ associations particularly vocal in their opposition to the perceived misuse of the technology. This chapter surveys these developments, ultimately arguing that despite the controversy it engenders, research into what we label “literary machine translation” still has much to offer, given its potential to inform critiques of the practice, to propose ways in which machines might assist rather than replace humans in literary translation, and to mould public opinion in a context in which literary machine translation is already a commercial reality.