This topic contains 5 replies, has 3 voices, and was last updated by 
-
Posts
-
Hi!
I just watched the TED talk about “Duolingo”:
http://edition.cnn.com/2012/01/08/opinion/von-ahn-translate-web/
As I found it extremely interesting, I thought I should share it with you. The essential part starts around minute 12 or 13.Basically it’s a new site for learning languages by translating sentences. You’re learning literally by doing and do not pay money, but your time (so to speak). The cool thing is that the translations are actually put to use (see TED talk). The point of it is mainly to capitalize on human learning to translate big parts of the web which otherwise would take tremendous amounts of money to pay professional translators, since machine translation simply isn’t good enough today.
There’s also a super funny example of an automated translation of a Japanese forum post about JavaScript:
“This is question, English is faulty. Thank computer to translate to help. SORRY!!!!!
At often, the goat-time install a error is vomit.
How many times like the wind, a pole, and the dragon?
This insult to father’s stones?
Please apologize for your stupidity. There are a many thank you”Hilarious xD I wonder what the original post meant to say ^_^
Anyway, do you think this way of learning actually works? I’m not sure, but I registered for the private beta already. :)
I think yes and no. It will be helpful learning grammar and new words but it will not train you to think in your target language. Also this will difficult with Japanese where there is no correct english equivalent for some terms that are tied into some idiosyncrasies of the culture like よろしくお願いします、ごちそうさま、 正座 and you can only just describe the meaning. I had my time as a scanlator and translating entire song lyrics so I know from experience
In other words I am rather skeptical about DuoLingo. I am worried about how much it will treat you as a language learner rather than human capital and that its screwing with translating work, something that is not easy. In ways this almost pisses me off because although translating web articles and factual information may be ok because its objective most of the time, but what about subjective material such as novels (and/or manga)? How many language learners will be able to accurately translate not only the words but understand the exact feelings the author wants to express whilst trying to make it as natural as possible in one’s own native language?
I’ve looked at Duolingo before, and I think it’s a cool concept that could work well for some languages.
However I have a hard time seeing how it would work for a language as heavily context based and inherently difficult to translate as Japanese – and I suspect the creators probably know that, which is why Japanese isn’t even currently listed as one of the “coming later” languages.
I see. I didn’t take a look at the site though
I agree with you, missingno15. I suppose it really only makes sense in objective texts, such as found on Wikipedia. But generally I really like the idea that this way big parts of the web could be translated in different languages. People, who previously didn’t have access to that information suddenly could read and understand it. Also given a sufficient number of people taking part in the project (i.e. millions) the translation will be done much faster than if performed by paid individuals.
Regarding the missing context, that is a worry of me too. The meaning of a single sentence can vary drastically depending on its context, and that is of course not only in Japanese, but really in any language. I’m looking forward to see how they want to approach this problem.
Finally, I think this could be a fun way of learning a new language. But only as long as they don’t exploit the users. It would kind of have an akward feeling to it, if you knew that they were earning tons of money with your work, even though technically you learned something and have no fees as in normal language learning programs.
You must be logged in to reply to this topic.