Text

Physicist, Startup Founder, Blogger, Dad

Saturday, September 25, 2010

Machines and bilingualism

I had a terrifying thought the other day. I would guess that at 90 percent confidence level machine translation and voice recognition will be good enough in 20 years that people will be able to communicate pretty well across most language barriers using cheap and unobtrusive devices. If so, is it worth all this effort to make sure my kids are bilingual?

I say it's terrifying because of the significant effort we're expending on our Bilingual Kids Project -- including relocating to Taiwan for this sabbatical. Another point of clarification: I'm not saying in 20 years we'll have AI (far from it). But something that translates basic phrases and simple content (surely we'll have that: Moore's law, massive corpora of translated text, statistical machine learning, yada yada) would reduce significantly the value of all but the most sophisticated language skills.

18 comments:

Fs said...

Terrifying thought?

This is amazing. I sucked in spanish class.

Jfmurphy said...

Yes.
Learning a second language changes the way a person thinks and what parts of their brains can be active in different situations. You're not just making sure that your kids are bilingual, you're making sure that their brains are bilingual and able to adapt to more situations.
Living as Americans in Taiwan for a year is also going to give them some of the good traits of third-culture kids, like closeness to family and acceptance of cultural differences.
They're also going to have touchpoints in time and culture of "Taiwan, 2010" versus "Oregon, 2009." In 20 years they will appreciate that.

anon said...

The Rapture seems more likely than what you expect with 90% confidence.

Underachiever said...

No, it is not worth it. I have never seen evidence that bilingualism, in and of itself, helps people. Also, why is this a terrifying thought?

Hao Ye said...

Agreed. Most linguists believe in at least a weak form of the Sapir-Whorf hypothesis: http://en.wikipedia.org/wiki/Linguistic_relativity.

Being multilingual also means that you learn to differentiate between those languages, which I think helps with being able to define abstract ideas and communicate with more precision, especially to people who are not native speakers!

Rasmus said...

It will probably still work as a status marker. Which should be a pretty big part of the value of being bilingual even today.

Yan Shen said...

I've read that it's harder to convey abstract and analytical concepts in Chinese than it is in Indo-European languages and that certain abstract concepts can only be talked about in Chinese in a roundabout way. One of the things I've heard is that it's harder in the Chinese language to express a general abstract quality as opposed to a specific instance of an abstraction. In Indo-European languages such as English, it's extremely easy to turn an adjective into a noun, by adding the suffix "ness" to the word. So for instance, dull can easily become dullness and it's easy for the English speaker to treat dullness as a Platonic abstraction. Also, it's easy in Chinese to talk about a horse, but harder to talk about horse-ness, or the Platonic essence of being a horse.

Have you also read about the controversy over the ability of Chinese speakers to explore counter-factuals? The original study by Bloom endorsed a strong Whorfian position and claimed that the lack of explicit counter-factual markers in the Chinese language impaired the ability of native Chinese speakers to explore counter-factuals. The later study by Au had Chinese bilinguals translate an idiomatic Chinese story in English and found that in this case, more Chinese speakers were able to identify counter-factual statements as opposed to English speakers. So the original Bloom study was criticized on the basis of a bad translation. I've read different studies which claimed that Chinese speakers were as good as speakers of English in identifying counter-factuals if they were clearly presented in a text, but that because Chinese has less clear markers of counter-factual statements in general, Chinese speakers identified them less often.

I know that Jospeh Needham also explores how language might have been a stumbling block for the development of Chinese science. In particular, Classical Chinese seems to have highly elliptical in nature.

Yan Shen said...

*In particular, Classical Chinese seems to have been highly elliptical in nature.

Hao Ye said...

I must admit that I have only dabbled a little bit in linguistics, so I don't have much in the way of actual examples, so thanks for that, Yan.

Brendanpenney said...

I think theres probably a strong value to not having your kids sound like the results of "Google Translate". Or, as google translate says:

"I think there may be a strong value Do not let your child's voice, such as "Google Translate" results".

Twenty years seems optimistic.

Anonymous said...

Precisely because only the most sophisticated language skills will matter 20 years from now, truly bilingual people will have an advantage over those who know a little bit of a second language but lack mastery, and those who rely on machine translations.

I doubt machine translations will be perfect in 20 years--one only has to look at translations between English and Chinese done by humans now to realize how many problems (nuance, puns, choice of register, slangs, parsing difficulties, etc.) remain. A translator is also an interpreter of subtext and context besides literal meanings. Can machines do that?

Seth said...

It isn't just bilingualism! You're giving your children a broader perspective on the world, and a big adventure for the whole family. All those friends of yours in the financial services industry are making you apologetic for the free-wheeling adventurousness that makes you unique. Shame on them. You are much wealthier than they are in the ways that matter. (And if you need an excuse, remind the greed-heads that Jim Rogers moved his family to Singapore to keep them from becoming poor along with the rest of lazy, culturally illiterate, fast-buck America. That may never come to pass, but the scary image will put them on the defensive ... ;)

steve hsu said...

Great advice, Seth. It isn't just the greed heads who are influencing my worldview, though. If I had chosen my sabbatical location purely on the basis of where I could do the most / best physics research, I probably would be in Berkeley or Pasadena or Cambridge or Princeton right now :-) That is, the opportunity cost for the language immersion extends to science, not just net worth...

steve hsu said...

What about a heads up display that offers a few translation possibilities that the user chooses from? I think a little mechanical turk input into the process would help a lot. The outcomes of these decisions will of course be recorded to train the algorithm and eventually you will have your "monumental project with a cast of thousands"!

Jay Cristo Lauren said...

Amazing write up. It's very informative. Thanks for posting this. norton utilities review

Web Design Melbourne said...

 

Chinese language is essentially composed of symbolic characters that represent letters as well as numbers. Before venturing into Chinese translation, it is essential to learn a few facts about the Chinese language. The traditional format of the language is more complex as it comes with more brush strokes for creating a character. But the simplified format of Chinese is easier to translate, as it is devoid of any complex characters.

translation to Simplified Chin said...

If you are interested in expanding your website to reach the "other half of the internet," one of the first things you probably wondered is what type of Chinese should be used. Mandarin or Cantonese? Traditional or Simplified? First I'll give you the simple answer, and then answer the most common questions I hear

kaye1270 said...

I would like to thank you for the efforts you have made in writing this article.Company Registry Hong Kong

Blog Archive

Labels