2022, Vol.23, No.1

The Journal of Translation Studies 2023 KCI Impact Factor : 1.49

pISSN : 1229-795X
https://journal.kci.go.kr/kats

Drawing on Clifford’s concept of “museums as contact zones,” this paper examined a project by South Korea’s National Research Institute of Cultural Heritage (NRICH). In the project that lasted for almost four decades through to the 2010s, the NRICH conducted surveys on Korean collections at museums overseas. From the contact perspective, this paper argues that the NRICH project used the Korean collections as a medium to raise the country’s voice to be heard at the Western museums. The project featured diverse forms of “contact work,” including on-site surveys at the museums (making corrections to the identities of some objects), support for conservation treatment, and negotiations for repatriation. As Korea led the collaborative initiative, the two cultures involved have been able to build reciprocity and address “asymmetrical power relationships” to a certain extent. The survey results were published in the form of bilingual catalogues, which in turn function as texts, or subversive translations, “in response to or in dialogue with” metropolitan representations. Also noteworthy is the fact that the content was not translated uniformly into the lingua franca of English but into those languages in use in the countries involved, facilitating future contact between the cultures.

OTT (over-the-top) services have seen a significant increase in subscribers over the past few years. Their global expansion and an increasingly multilingual viewership fuel the need for translation into various languages. Netflix, for example, produces and offers subtitles in many different languages via pivot translation; the source text is first translated into an English template and then further produced into many languages. Although a widespread practice, there is scant research on pivot translation in translation studies. Yet, it deserves more attention given that it is likely to continue for the foreseeable future because of its many benefits such as streamlining processes and satisfying the increased need for multilingual translations. Against this background, this study explores Netflix’s subtitling guidelines to examine how Netflix sees pivot translation and to discuss how it may evolve in the future. First, Netflix’s subtitling guidelines for English pivot translation are compared to those of other individual languages for any differences. The guidelines are then applied to two programs—Terrace House and Kingdom. These two are chosen as representative cases where particularly speech levels of the former and cultural references of the latter can pose difficulties in translating when first going through the English meditating language. As a leader in OTT service, Netflix’s multilingual translation can have a significant impact on other streaming companies. This study could help to better understand the role of pivot translation and ways multilingual translation will unfold in the future.

In the automatic evaluation of translations, precision and recall are two indices that show how precisely (precision) and how much (recall) the system is able to recognize the well-translated portion in a translation. It would be ideal if two indices could be equally weighted in the evaluation system, since both accuracy and completeness are important criteria in evaluation of human translations (HT). This is, however, not easy, as both indices are negatively correlated. Papineni et al. (2002), for example, opted for precision, while Lavie et al. (2005) used both indices, giving recall nine times more weight than precision. The aim of this work is to examine which of the two indices correlates better with evaluation of professional evaluators and how much weight should be given each to precision and to recall. For this purpose, 459 translated texts were rated with precision, recall, F1 (harmonic mean of precision and recall) and Fmean (nine times higher weight on recall) as well as by professional evaluators. The results show that recall correlates better with human evaluation than precision in almost all cases, but not Fmean than F1, which were equivalent in all but one case. They indicate that recall is indeed a more important metric, but the weight as high as nine on recall is not ideal for HT evaluation.