Neural Machine Translation Systems and Chinese Wuxia Movies: Moving into Uncharted Territory

Abstract

AbstractThis chapter contributes to the discourse on the fansubbing of wuxia movies using neural machine translation (NMT) systems. The chapter is informed by the current popularity of wuxia movies among global audiences, increase in fansubs and user-generated translation agents, and ubiquity of free open-source MT systems. Wuxia movies are deeply rooted in the ancient culture and history of China, making their translation particularly challenging for MT systems. Therefore, this chapter attempts to answer the following question: How can fansubbers make the best use of NMT systems in translating culture-bound elements of wuxia movies? To answer this question, we adopt a corpus-based approach consisting of MTs of culture-bound excerpts of the movie Crouching Tiger, Hidden Dragon (2002) by Ang Lee. Any MT errors identified are extracted and categorized. Thereafter, we suggest competencies, translation strategies, and approaches likely to help fansubbers improve MT output quality. In addition to furthering the discourse on the use of MT and the intrusion of non-professionals in the translation profession, the chapter provides an opportunity for the industry, especially training institutions, to create targeted short-term training programs that offer fansubbers the basic skills and competencies they may need.