The time is 2029. Google and Amazon have emerged as two behemoths that handle about 96% of the world’s translation needs between them for all the major languages in the world.
Last week they announced this year’s winner of Nobel prize for literature. I have never heard of this guy before, so I did a search online and found one of his books. The company my friend Xiao works for is a subscribed user of Google’s Premium Translation service. So I badgered him into providing me with the password to his company’s Google Translation Service account and I got the book translated last Friday and spent the weekend reading the book.
I met Xiao for lunch today and we talked about the book. He asked me which style I chose for translation. I told him I selected “the Nostalgic 1960’s” style.
“No, no, that’s the wrong style to use for this writer’s books. He was born in 1975. So you should’ve chosen the “2010 Rebellion” style! That would tell the Google machine to only use the post-millennium corpus of literary text for source vocabulary when producing your translation!” He declared.
We debated over the merits and demerits of the dozens of styles offered by Google and I promised to try out the style recommended by him. I then lamented that none of these styles offers the same satisfaction as reading a book translated by Ru-Long, my favourite translator of Russian literature in the 1960s. “You are so old-fashioned!” He admonished me, “language has evolved. Nowadays young kids have adapted to new styles churned out by machines, your language taste is horribly antiquated!”
He is right, of course. Although some books are still being translated by human translators, in this age where people live fast-paced lives, by the time the translation is completed, the readers’ interest has moved on to the next sensation. The handful of human translators are just relics of a bygone era and they represent less than 0.3% of the total translation turnover in the world, They are not in the business for the money (there is no money to be made anyway), they are doing it to prove a point: that human translators are better than machines, for whatever it is worth.
This Gotterdammerung for translators happened about 3 years ago when quantum computer was introduced which enabled machines to learn things very fast. They have discarded the dictionary-based system they used in the early days. I remember back in 2019 many of my fellow translators used to ridicule the translation produced by machines because human translators were able to make better judgement regarding the choice of words/tone/register based on their knowledge and nous which was the result of having read hundreds or even thousands of books over a lifetime, which went far beyond the scope of dictionaries. But even in those days, I had a premonition that machines were going to outflank us. At that time I was learning English at advanced level, aiming for native-level competency. During this process, I realised that we humans are not as free-thinking as we thought we where, we are slaves to a language pattern that has been developed by a society collectively over the years. As an outsider who was not familiar with this pattern, I was liable to make language mistakes that any native-speaker would immediately detect, just like a system can detect any alien body who behaves differently. That means that a computer system based on artificial neural networks that is able to detect patterns can do a better job at language translation than we humans can! The only thing that was holding them back was the slow processing speed of computers.
And then the quantum computer was introduced. Now machines can “read” millions of books and analyse patterns in a few seconds, they are more learned than any living or dead people in the world. Now, instead of limiting their scope to dictionaries, they can compare similar situations contained in the numerous books and magazines stored in their memory and come up with appropriate choice of words/sentences that wow us.
But they still need humans though. Some agricultural company from Australia approached Google last month requesting help to translate into Chinese some esoteric agricultural stuff for which there is a woeful lack of existing database to work with. My friend Xiao’s company is a consulting firm contracted by Google to provide linguistic service. He and another colleague, along with two professors from a Chinese agricultural university, will be sent to spend two months with that company in Australia to learn about their requirements, and then they will spend another two months in an agricultural area in northern China. The outcome of their work will enrich Google system so that it can handle this kind of requirements in the future, so the bill will be footed by Google. During this process, the team will create a nomenclature that will be the gold standard for this business in the future. The costly confusion caused by 100 human translators inventing 100 different wheels when translating the same new technical jargon, a situation that we used to see a lot back in 2018/19, is now history.
Have machines replaced human translators? Some people insist they have not and will never do, pointing to that handful of human translators translating literary works as well as United Nations which still uses human interpreters when world leaders meet. But some say these leaders are actually getting information from the earpiece that each of them wears in their ears and according to some reports, Google machine is linked to the other end. If that is true, then these human interpreters are just for decoration rather than serving any practical purpose. Anyway, I think whether that replacement has happened or not is only an academic point because as far as my translator colleagues from our Facebook group are concerned, I haven’t heard from them for a long while. And the rumour says that Google is working on an update to its machines, the next version, version 375.14159260 will be the definitive version that will send all of us to enjoy a life on the breadline.