[quoting Alessi, 27 Feb 96.a] Steve says my dictation prediction is overly optimistic. Perhaps, but again, its just a matter of when. Also, his main objection to my dictation prediction occurs because he is, I believe, confusing transcribing and understanding. Understanding is necessary for language translation (my third prediction) but NOT very necessary to take a speakers speech and transcribe it electronically.
Steve, You, like most people working outside linguistics, greatly underestimate the problem. Transcribing [real world language], is ABSOLUTELY impossible without understanding. You, like most people, think we speak in words that are made out of "letters." In fact we speak in blurs. There are no white spaces between words. The "d" of dog is different from the "d" of dig. The "h" of hit is entirely different from the "h" of human. There is no "n" in can't. Can a computer tell the difference between I scream and ice cream? Try transcribing a language that you don't know. If it is continuous speech, you can't begin to tell what the words are. If it is single words with "white spaces" between, you can't tell what the "letters" are.
Granted, if you allow a long enough time frame, these problems will be solved, but ten to twenty years is optimistic, in my opinion.
Steve makes the interesting comment that Japanese depends on its written form. I admit I know nothing about Japanese linguistics, but I find that claim fascinating. Did the Japanese language spring into existence in written form? (Said tongue in cheek.)
The Japanese language today is not the Japanese language of ancient times. When the Chinese writing system was borrowed, it afforded a new way of creating words. The bulk of the Japanese language is now a product of the written forms.
Steve says my translation prediction is also overly optimistic. Here I can agree more because translation does depend on understanding. But again, its just a matter of time.
Just like higher salaries for professors.
I did not understand the last point, that written versus aural communication differences contradict Richard Clarks "mere vehicles" hypothesis. Would you explain a bit more, Steve? I doubt that Dick Clark considers human modes of communication as equivalent to different media, even if related.
But if one is educated from books, that is a different medium from a "talking computer." Clark (amazingly) does not define medium, but surely this qualifies. Raising questions about the "effects" of learning in this way is taboo if you are a Clarkian.
I was NOT thinking of Americans and forgetting the Europeans. Europeans are a small part of the world population. How many people are in rural China (over a billion?), India (800 million?) Indonesia (200 million?) and other parts of Asia, Africa, and South America. I was thinking of those huge rural and third-world populations when making my statement. If I am wrong, I hope someone with data will correct me.
Most countries are multi-lingual. China (multi-dialectal, where the dialects are mutually unintelligible) [Actually the difference between language and dialect is political, not linguistic. An old linguistics joke is that a language is a dialect with an army and a navy.], India, and Indonesia and all multilingual. The degree to which the populations (especially the lower classes) are bilingual is problematical. Even though a country is multilingual it is possible to be monolingual and live happily, especially if one lives in a rural area. There is probably a general drift towards monolingualism, if enough of the population can communicate in that language. In Singapore, where there are four official languages (Tamil, Malay, Chinese, and English), young people who marry outside their ethnic group will tend to raise their children speaking English. Others will do so simply for economic advantages.
You can communicate through a translator if you have time, but it is a slow and painful process. My university is in Japan, but its official working language is English. Our faculty meetings are all translated. If you think your faculty meetings are tedious you should come to ours. Even with skilled translators, it is often the case that the translation is botched. Some languages (like English) express hypotheticals more easily than others. This often produces great misunderstandings (remember, we are not even using unreliable machines). Another problem is that some (most) speakers don't really know what they are saying (they are composing as they speak), so it is no wonder that the translation is wrong. If you really want to communicate, it is better to be bilingual. If machine translations are capable of 95% of what human translators can do, we are still in trouble. Note that a specialized machine (Deep Blue), in a highly rule-bound and constrained domain (chess), cannot beat a Grand Master. We are all Grand Masters of language--a domain that is far messier than chess. I wouldn't hold my breath waiting for a Deep Blue of language. I'll bet your computer can't even display the Japanese characters in my signature.