Linguistics, MT, and AI...

alex gross language at
Sat Oct 30 03:25:47 UTC 2010

Thanks to Yuri Tambovtsev for the following:

> There is a crisis in linguistics, though linguists do not care to notice 
> it.

I don't intend to push the points I have already made too much further, 
since I respect all of you for the difficult situation in which you find 
yourselves, even if you "do not care to notice it."

If it's any consolation, you're not alone in this situation, it's one shared 
by everyone in the AI field, of which MT is just one branch (and yes, I'm 
aware there was an MT movement before AI which would later merge). Up there 
on that Tokyo Prize stage with Chomsky was John McCarthy, the inventor of 
the phrase Artificial Intelligence and often named as its father.  (If 
memory serves, you'll find a photo of them both on that Tokyo platform in 
Barsky's bio).

It was a heady year, 1988, when AI was at its summit of prestige, when we 
were told repeatedly that it was only a matter of time before KEs (Knowledge 
Engineers) would be arriving in every area of business, education, and 
industry to consult with DEs (Domain Experts), and once the KEs had absorbed 
what the DEs knew, they would digitize it all, and DEs would no longer be 

But it all worked out rather differently, just as in MT, where it was widely 
boasted in 1988 that human translators, the DEs in their field, were soon 
due to vanish. What actually turned out is that almost all the MT companies 
of the 80s & 90s went under, while remaining MT/TM companies are virtually 
begging human translators to use their systems, since only human translators 
are capable of coping with them.

Back during those ambitious 80s & 90s, it was widely boasted that the Austin 
CYC project would soon solve all of AI and language by joining vast numbers 
of computers together to simulate the human mind & human language. We hear 
little about CYC these days, though according to the Times it has just been 
reborn at Carnegie Mellon under its new name of NELL (Never Ending Language 
Learning).  I'm willing to bet that my 1994 criticisms of CYC will still 
apply to NELL, you'll find them at:

Just a few weeks ago the New York Times published another pro-AI article 
claiming that robotically driven cars are nearly here.  None other than John 
McCarthy made this claim to me on-line back in 1998, and once again I 
believe that most of the counter-arguments I made to him then still hold. 
You'll find them at:

I could go on and on with further AI examples but will spare you. The future 
of linguistics is your problem, and sooner or later you will need to 
confront it.

Very best to everyone!


PS--I realize that "culture" is a dirty word around here, but the granting 
of the Tokyo Prize to Chomsky and McCarthy by Japan was scarcely a 
disinterested one devoid of cultural and political overtones. This was the 
period when the Japanese were staking their own future on AI with their 
so-called Fifth Generation drive, and they saw these two men as providing 
key guidance in their campaign, Chomsky for language, McCarthy for 
everything else.  I discussed all of this in some detail at the time in a 
review you can find here: 

More information about the Funknet mailing list