Copier Auto-Translates Japanese to English

StCredZero writes "Wild. Fuji has created a photocopier that automatically translates documents from Japanese to English. That's pretty nuts. Apparently, the copier can figure out what sections are text, OCR the text, send it to a translation engine, and put the english back into place."

Comments

Re:It's impossible...

[kklein]

Disclaimer: I'm a second language acquisition researcher and assessor.

I concur. Absolutely. Language is not pure information; it's information shorthand. It assumes a high degree of already-shared knowledge about the world. Some of these assumptions are near-universal; many are not.

Japanese and English (my languages) offer a great example, especially as it pertains to machine translation. Whereas English is a subject-predicate language, where basically all the information is encoded in the language stream, Japanese is a topic-comment language, where, once set, the "subject" is not re-stated until it changes. Beginning Anglophone learners of Japanese make the mistake of putting a "wa" to denote what they think of as the subject in every sentence, when it does not need to be there. "Wa" is a topic marker; not a subject marker.

This is a fundamentally different way of thinking about language and, therefore, about the world. Germanic languages seek to operate regardless of context; Asian languages seek to augment (or "comment on") it. If you've ever felt that Japanese people who speak English are beating around the bush or being vague, part of that is cultural, but part of that is the language of the culture that does not require explicitness. A big part of learning Japanese or, for Japanese people, of learning English is learning how to think about the world and about human interactions in a very different way.

Machines aren't human. They are information processors. They don't know what a "cat" is; they just know that it's a piece of code that can be slotted into a certain place in a set of syntax. Until machines are really intelligent (and I don't think that will be anytime soon), expect more crappy translation than useful. Anyone who tells you otherwise is probably selling something (a crappy machine translator, to be exact!).

Says someone who's never translated something.

[Valdrax]

Either you are old, or a bit naive. I think in the next 10 years we will see significant improvement.

Yeah, 'cause researchers have long promised us that AI will reach us in 10 years. <sarcasm>

Seriously, I think you underestimate the difficulty of translating. Have you done any major foreign-language translation -- especially of conversational speech? My experience has primarily been with Japanese and English, and I'll tell you right now that it can be nightmarish.

Sentence fragments are the worst part. Japanese has a completely different word order from English. All modifiers (including phrases and clauses) come before the word they modify, and the language has a Subject-Object-Verb order. "I just saw the man who stole my friend's watch last Tuesday" becomes "Just I Last Tuesday friend's watch stole man saw." Now try translating that from Japanese to English when the sentence is cut in half.

Worse, the language has very different levels of allowed vagueness. "Complete" sentences in Japanese can contain just a descriptor or an action without any specification of who did/was what. Conversely, translating "3 of them" in English to Japanese is hard because you have to know "3 of what?" to know what counting suffix to use.

Another problem is that many very different words sound exactly the same when conjugated to the gerund or perfective forms. English has a number of homonyms, but there are MANY more opportunities for mix-ups if you don't have access to kanji to tell the semantic meaning apart because Japanese has a much more limited range of phonemes. For example, take "katte" which is the gerund form of the verbs "kau" (buy), "kau" (keep/raise), "karu" (cut), "karu" (spur on), and "katsu" (win). That's 5 completely different verbs that conjugate to the same sound. If they're written phonetically or your going from speech, then you have to be able to understand the meaning behind the words to translate. (Did I mention earlier that you may not have an explicit subject and object to go off of?)

Then you get into issues of translating things like politeness levels, different ways of addressing people, and other concepts that don't translate well into English or concepts like singular vs. plural that are dropped in going to Japanese. Let's not even consider puns and poetry!

These are not trivial issues. An automatic translator would need to somehow be able to conceptualize what a person is trying to speak about, which would require understanding the story being told and an ability to predict where they are going with it. This will require strong AI.

Accurate and intelligible translation is an art -- not a science -- because it requires an intuitive and empathetic ability to understand the mind of the speaker well enough to map their thoughts into a different method of expression.