3 ms·

Don't forget that English and French have much in common lexilogically and grammatically, and even some slang must be similar because of geographical proximity

Don't forget that English and French have much in common lexilogically and grammatically, and even some slang must be similar because of geographical proximity and cultural exchange. I am nearly certain that no machine will ever be able to accurately translate between languages with significant linguistic distance (e.g. Japanese and Swedish). I experience this first hand, whenever I translate from my native Greek to English, especially if there is slang involved. Whenever Google translate encounters long phrases in my texts, the result is comical, not to mention that the emotions are not properly conveyed. I can only begin to imagine the inaccuracies in translations from Mandarin. Nevertheless, I never expected even fairly accurate translations between even related languages such as French and English. It indeed sends chills down the spine. It feels like there is some form of actual intelligence involved.

tralarpa 3y ago
> I am nearly certain that no machine will ever be able to accurately translate between languages with significant linguistic distance (e.g. Japanese and Swedish). Is that really the reason? Or rather the fact that there is much less training data available?
- naniwaduni 3y ago
  Languages even moderately distant tend to strain the concept of an accurate translation in the first place for any nontrivial utterance.
  - bugglebeetle 3y ago
    Eh, I speak both English and Japanese and I would say that what counts as accurate translation is what is most proximate to that threshold of fundamental dissimilarity. Measuring translation accuracy for all languages the same way is more the problem here. There is no such thing as 1-1 to translation. It’s more like 1-(1+n) where n accounts for said distance. For languages with shared origins, n can be fairly small, while for those with entirely separate ones, it can be quite large. That being said, Japanese to English translation in things like popular culture tends to take far too many liberties, I expect because the culture around Japanese translation in America has a very annoying, Orientalist bent, with people getting off on their “expertise” about a fake exoticism.
    - naniwaduni 3y ago
      The fact that you have to have this conversation strongly implies that the notion of an accurate translation is, no surprise, already heavily strained; you've simply chosen to aim for/accept "most accurate possible translation" as the best you can do, while punting on choice of distance metric and its scale.
      - bugglebeetle 3y ago
        …or simply that words don’t have universal meanings, but general ones that vary by use and context.
        naniwaduni 3y ago
        Correct. The enumerable configurations of expressible shades of meaning are pretty sparse, and don't line up well between languages.
        bugglebeetle 3y ago
        …which means “accurate translation” has different meanings relative to the languages you’re referring to, as I first pointed out.
    - mcswell 3y ago
      All your base are belong to us.
acomjean 3y ago
in the late 90s my Mom (bilingual, having migrated to the US) did some work on the side for a translation agency (She worked doing internationalization for DataGeneral, parametrics and others) . She would evaluate translations for prospective translators. She got a batch that was terrible. Turns out they were machine translated. We have come a long way.

tralarpa 3y ago

> I am nearly certain that no machine will ever be able to accurately translate between languages with significant linguistic distance (e.g. Japanese and Swedish). Is that really the reason? Or rather the fact that there is much less training data available?

naniwaduni 3y ago
Languages even moderately distant tend to strain the concept of an accurate translation in the first place for any nontrivial utterance.
- bugglebeetle 3y ago
  Eh, I speak both English and Japanese and I would say that what counts as accurate translation is what is most proximate to that threshold of fundamental dissimilarity. Measuring translation accuracy for all languages the same way is more the problem here. There is no such thing as 1-1 to translation. It’s more like 1-(1+n) where n accounts for said distance. For languages with shared origins, n can be fairly small, while for those with entirely separate ones, it can be quite large. That being said, Japanese to English translation in things like popular culture tends to take far too many liberties, I expect because the culture around Japanese translation in America has a very annoying, Orientalist bent, with people getting off on their “expertise” about a fake exoticism.
  - naniwaduni 3y ago
    The fact that you have to have this conversation strongly implies that the notion of an accurate translation is, no surprise, already heavily strained; you've simply chosen to aim for/accept "most accurate possible translation" as the best you can do, while punting on choice of distance metric and its scale.
    - bugglebeetle 3y ago
      …or simply that words don’t have universal meanings, but general ones that vary by use and context.
      - naniwaduni 3y ago
        Correct. The enumerable configurations of expressible shades of meaning are pretty sparse, and don't line up well between languages.
        bugglebeetle 3y ago
        …which means “accurate translation” has different meanings relative to the languages you’re referring to, as I first pointed out.
  - mcswell 3y ago
    All your base are belong to us.

acomjean 3y ago

in the late 90s my Mom (bilingual, having migrated to the US) did some work on the side for a translation agency (She worked doing internationalization for DataGeneral, parametrics and others) . She would evaluate translations for prospective translators. She got a batch that was terrible. Turns out they were machine translated. We have come a long way.