Language – far more complex than grammar and vocabulary, more intricate than what you see in language learning books and lessons, is a melting pot of expression and the ultimate tool for voicing one’s experience. It is our sole vessel of communication, creating a community and connecting people throughout the globe and allowing them to relate their opinions and emotions. The culture engrained in it and the different implications of certain phrases completely reject direct and literal translation – whether those be the slight differences in over 400 words related to snow with Scots or the total vocabulary of 123 words in the constructed Toki Pona, language is the carrier of the rich history of humanity. That immediately poses the question – can and should translation solely carry the meaning? Where do the lines between an exact translation which is misunderstood by the audience and a vague translation which carries the meaning forward merge? Are only words enough?
In the age of mass communication, the yearning for understanding makes translation a necessary part of reality, connecting various communities. It goes beyond the literal meanings of separate vocabulary and intertwined structures of grammar – it requires interpretation, cultural awareness, and thorough communicative ability.
The main components of translation in the 21st century are essential in understanding how a piece of text reaches and interacts with the reader. Those are namely the people behind the screens with lines and lines of text, the machine translation (MT) created by developers (mostly relying on AI and machine learning algorithms) and the translation memories (TM) that aid both and are assumed to be ‘correct’ as they have entered the final published stage.
MT has gained popularity through the immense progress made in technology and machine learning. It provides instant, free, and consistent translation. There is research to prove the need for human input, however, most of the theories are based on English-based translation and from a linguistic standpoint, English is not particularly the peak of difficulty. More complex languages, or those with further limited vocabulary, would require far more room for mistranslation. Machine translation strips the strand of language from culture, nuance, and overall complexity, which might lead to a total disconnect of meaning. Consistency in this case does not equate quality, and the easiest example of that are idioms or set phrases – the consistency of vocabulary awareness does not translate the meaning of them at all. Quality can be better maintained in TM, as it is created by translators for translators and refers to text, which is already published, hence has passed the post-editing stage.
To address the elephant in the room, from a corporate perspective it is cost-effective to use MT and support translators as post-editors, however, that would have to cut cost from the outsourced or in-house translators as they are no longer doing the “full work”. Does it ethically make sense to go cold turkey on human impact in translations and is it sustainable for the work force? Sure, it is statistically proven that the speed of translators is beneficially impacted using TM and MT, but they require the same amount of post-editing. The usage of such tools, however, is pushing boundaries of the openness of language, “challenging the dominance of single perspectives, languages and national identities”, and can thus limits the control and contribution of a translator. The switch from being a ‘translator’ to a ‘post-editor’ still comes in contrast with theories of translation being creative rather than mechanical work – communicating implications and nuances to the reader/listener can take shape in multitudes of ways that are rarely mechanically reproduced correctly. Translators know best that the creative impact they have on their work cannot be replicated as it follows no rules and is adaptable, individual, and ultimately irreplaceable.
References:
Gonzales, L. (2018) “A Revised Rhetoric of Translation,” in Sites of translation: What multilinguals can teach us about digital writing and Rhetoric. Ann Arbor: University of Michigan Press.
Slocum, J. (1985) “Machine translation,” Computers and the Humanities, 19(2), pp. 109–116. Available at: https://doi.org/10.1007/bf02259632.
Ive, J., Max, A. and Yvon, F. (2018) “Reassessing the proper place of man and machine in translation: A pre-translation scenario,” Machine Translation, 32(4), pp. 279–308. Available at: https://doi.org/10.1007/s10590-018-9223-9.
AUTHOR: Tsveta Georgieva, translator and transcreator at SLSP Ltd.
1 Comment