The Rise and the Fall of Machine Translation

My father is a specialist in machine translation. He has dedicated his entire life to developing computer systems that translate from one language to another. However, machine translation as a field of knowledge is now effectively dead. It is one of the professions that are being destroyed every day by the advances in technology. Just imagine: my father saw the rise and the destruction of this field within his own lifetime. And he is still quite young.

The way our “ancient” machine translation systems worked was the following: we created algorithms that analyzed each sentence, broke it down into parts (subject, predicate, object, etc.), analyzed each part grammatically (number, gender, conjugation, etc.) and looked for an equivalent in the target language. This was a compex, extremely fascinating process.

And then Google Translator came and made all of this completely obsolete. It operates on an entirely different set of principles.  Google Translator does not analyze sentences or break them down into parts. It simply accesses its own incredibly huge amount of data and looks for the translation there. Since the Google search engine can access unimaginable quantities of information, it simply takes the sentence you ask it to translate and looks in translations of texts that are already available for something that looks close enough.

Of course, at the very beginning, Google’s translations were crap. But the beauty of the system is that every sentence you translate with it improves it. Every day, crowds of people work on improving the Google Translator completely for free. Isn’t that brilliant?

Technology changes our world so fast that only people who are flexible enough to drop their “old” professions and pick up new ones as soon as it’s needed will come out winning. This might sound unfair but who is prepared to give up on the benefits the ever-changing technology gives us?

Zygmunt Bauman, one of the world’s leading philosophers, suggests that people who never put down roots, who can move around the world freely will constitute today’s elites. But it isn’t just physical mobility that is of such a great value in today’s reality. It’s intellectual and psychological capacity to change, transform and move on that is so precious.

10 thoughts on “The Rise and the Fall of Machine Translation

  1. My wife’s cousin’s husband was an apprentice compositor. By the time he completed his apprenticeship, his trade was dead, and hot metal printing was on its way to the museum.

    Like

  2. My old roommate got his MS in machine translation–he started just before Google Translate, and by the time he finished and started working at a company, it was beating all the machines–luckily there are other applications of computational linguistics he could move on to. I’m frustrated that my students don’t use Google Translate properly though, so I’m incorporating a handout and mini-lecture on how to use it into my courses into the Fall.

    Like

    1. ” I’m frustrated that my students don’t use Google Translate properly though, so I’m incorporating a handout and mini-lecture on how to use it into my courses into the Fall.”

      – I only just started using it, so I’m still figuring it out. What a great idea to create a handout for students!!

      Like

  3. Don’t have much to add on the translation question, although any blog post that mentions Zygmunt B is great in my book.

    Like

  4. The decades of research, trial and error, within machine translation has, however, not been fruitless. We now have a much deeper understanding of the structure and acutal use of specific languages as well as many differences between languages.

    Like

Leave a comment