Re: Off Topic: NLP, le to'e melbi glico bo bangu

[Christopher Palmer <reid@pconline.com>]

On Fri, 19 Feb 1999, xod wrote:

> Do you doubt the appearance of real-time text language translation, to
> 95% accuracy, in 50 years?

Yes.

As for why: 50 years ago they were predicting 9x% accurate machine
translation 'by 1960' or some other it-didn't-happen date. Computer
technology has not changed significantly (=revolved, as opposed to
evolved) in the last 50 years. No theory of language has been able
accurately and completely to describe the human language faculty, and
computational theories are even less on the mark.

Which is not to say that (computational) linguistics is worthless, just
limited.

> > What are your sources for this? How are measuring difficulty? What are
> > the language-internal implications of difficulty (on scope of
> > expression, flexibility of syntax, ...) Who is it so difficult for?
> 
> So I've been told, repeatedly. I thought it was commonly understood and
> agreed that English and Chinese are the two most difficult languages for
> non-natives to achieve fluency in.

'Some guy that told me; I forgot who it was' is my source too, which is
why I doubted it. L2 acquisition is a far more problematic and tangled
issue than merely rating the 'difficulty' of acquisition: it's a whole
field of study with piles of nasty, unasnwered questions. A blanket
statement like 'English is hard' is meaningless.



---------(( Christopher Reid Palmer : www.pconline.com/~reid/ ))---------

              the characters i am, made into a word complete
                              -- Meshuggah