A Look at Human vs. Machine Translations

Tags

Top Highlights

• The purpose of machine translation,” writes one such author, “is, on the one hand, to relieve the human translator of the need to work with tedious, repetitive and aesthetically unsatisfying material and, on the other, to speed up and facilitate worldwide information dissemination

• ranslating fiction and poetry is not a task for machine translation.”

• n favor of a more systematized, normative process, in which efficiency and accuracy are the goal.

• it’s not inconceivable that as artificial intelligence matures, more sophisticated algorithms could enable it to make the kinds of distinctions on which literary translations depend

• Words do not always have direct equivalents in other languages, and even when they do, they’re not always consistent. T

• ntextual shadings that can cause an identical semantic structure to mean very different things are instinctively perceived by the human brain

• quality of the resulting translations sharply declined

• Naturally, in the “that was then, this is now” world of technology, we can expect that computers will eventually learn to distinguish between soft fruit and hardware. But in the near term, at least, some form of human post-editing seems required.

• he more one moves away from translation as equivalence and paraphrase — that is, the more one enters the domain of literature, with its myriad shadings, innovations, and ambiguities — the less applicable it becomes, as even its boosters freely recognize

• MT is still mainly intended for utilitarian forms of translation,

• Running more complicated, and less famous, quotes through Google yields a predictable patchwork of accurate phrases, unnatural syntax, and gibberish that sounds like the user’s guide to a foreign-made appliance

• his book invites us to sympathize with the translator not as a “traitor” but as the author’s creative partner.