"The sliced raw fish shoes it wishes" - New York Times article on Google Translate

Today's New York Times has an article on the progress made in automatic translation by Google through the using of mass amounts of comparative data instead of trying to teach computers to translate languages by themselves. For language students it has also become an invaluable tool, especially for minor languages like Icelandic, Maltese and Estonian where resources online are otherwise quite scarce and finding the meaning of one word or another used to be a nightmare at times. Here is the video I recently made on using Google Translate on Ted.com videos to learn languages.



The biggest problems in the future will certainly lie in areas where languages simply don't match up with each other, no matter how much sample material there is between the two. Using a language like English to translate a letter or personal message into French or any other language with a t-v (tu - vous) distinction simply doesn't work properly as only the speaker alone knows how close he is to the person he is talking to. The only way to surmount that would be to have a list of options where the person requesting the translation is able to fill out closeness to the speaker, but that brings us back into the realm of computers having to understand languages again.

With English to Korean it's even worse, because not only is there a t-v distinction but there are other distinctions as well, such as using informal speech with politeness markers to talk to friends about people you respect (e.g. two kids talking about their grandparents) or talking to someone older than you about people your age (now the speech itself is polite but the person you're talking about doesn't receive a politeness marker). Add to that the dropping of pronouns all the time and translating the other way around is a nightmare too.

그래서 어제 오셨다고? - So (grandmother or some other respected person) came yesterday? (to friend or someone younger than you). No pronoun here either. The sentence is just 그래서 (so) plus 어제 (yesterday) plus 오다 (come) with a 시 politeness marker making it 오시다, and then put in the informal past tense 오셨다 and then followed by a 고, used in quoting. I.e. (you) said that (grandmother or some other person) came yesterday?

Luckily, these distinctions are almost the same as Japanese so translating between those two alone is really easy. It's the translating between two semantically different languages that presents such difficulties.