earticle

영어

These days, advances in Neural Machine Translation (NMT), Artificial Intelligence (AI), and the development of the Fourth Industrial Revolution have perplexed professional translators and their teachers, because it is predicted that their work environments and education systems will change drastically, even if they are unprepared for such changes. However, and perhaps unbeknownst to them, Machine Translation (MT) has been constantly developed since the 1950s and was eventually implemented based on AI, i.e. NMT by Google Translate in 2016. In the early days of MT, the consensus on cooperation between human translators and machines was optimistic. People thought that machines would become servants rather than enemies; computers would serve human translators, rather than replacing them; and that machines and computers would eventually raise the status of the translator’s profession. In this presentation, we examine if the optimistic views on human and MT interaction from the 1980s still exist even in today’s digital era. For this purpose, we reconsider Bar-Hillel’s assertion in 1960, in which Fully Automatic High Quality Translation (FAHQT) is not impossible, and we arrive at the same conclusion; that is to say, a machine cannot perfectly translate without human intervention, because a machine cannot yet discern ambiguous meanings of words. Therefore, humans and machines must cooperate to produce high quality translations. However, it is evident that the quality of MT will continue to improve and the translation profession and its training will be implemented on basis of the MT. This is particularly true of technical translation and, with continued technological progress, even literary translation. If so, what should human translators do in the era of MT? MT will expand the roles of human translators. As language service providers, and to improve the quality of MT, human translators will participate in establishing corpora of human linguistic content; simplify sentence structures for machine comprehension (pre-editing); and correct errors in the raw output of MT (postediting). In particular, these roles differ from those of traditional translators. The latter carry out a creative role, creating translated texts with their own expressions and interpretations; however, it seems that translators who perform pre-editing and post-editing tasks are servants to the machines and are secondary in the translation process. Such a view may sadden human translators. However, if looked at from a different angle, MT is just another tool helping to translate considerable volumes of text more quickly. Human translating tools have developed from pen on paper to word processors, and then, to CAT(Computer-Assisted Translation) tools; automatic translation is simply the next step. To achieve FAHQT without human intervention will take considerable time and may ultimately be impossible.