lkml.org 
[lkml]   [1998]   [Dec]   [14]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Internationalizing Linux
"H. Peter Anvin" wrote:
>
> > No the real problem is that: the second order predicative logic
> > isn't complete. So there is no way to map it properly into any kind
> > of formalized language... See: Tarski and Goedel :-). AI can't help
> > since it can't exist, at least on the kind of silicon based
> > computers we use those times, for the same reasons. However
> > amazingly enought this had been well known since already the
> > beginnings of this century... BTW. I'm quite happy with this
> > situation since it's implying as well that programmers are not going
> > to loose they jobs for a quite long time in the future ;-).
> >
> > Marcin
> >
>
> I consider that interpretation of Gödel's Incompleteness Theorem to be
> faulty. There is no credible evidence that humans need something
> beyond the purely computational to decode language, and it is pretty
> clear that humans are able to decode human language well; however,
> sometimes inherent disambiguities in the language force us to request
> clarification. Not perfect, but life goes on.

You say it yourself the main point here is *humans* are quite happy with
an *imperfect* world... BTW. I wouldn't call the process of decoding the
meaning of some words a computation at all, since a computation is in my
view something you can perfectly describe by a Turing machine and
this certainly isn't possible due to the simple fact that
there isn't something like a perfect understanding at all. Or I'm just
missing
something? Please give me the (formal!) definition of understanding...
Does the ticket automata equipped with speed recognition really
understand? I doubt it. Does the kernel understand if I'm telling him:

~# /sbin/shutdown -now

Certainly NOT! Otherwise he would try to resist/defend himself or at
least he
would be offended unless You tell him instead:

~# /sbin/shutdown -now -please

> Human language just *isn't* perfectly logical: a lot of the
> information is conveyed by choice of words, connotations, inflection
> , context and innuendo. This part of the information contents is
> mostly fuzzy and often ambiguous.

Yeep and there the requested credibility comes once again. "Fuzzy" is
the
phrase one should be looking for. If you are just looking for evidence
and no proof,
here it comes: Look at babelfish at altavista and enjoy :-). And of
course
most of the time there is no information contained in human words at
all.
(Quite common simple example: "I love you").

And certainly one quite important connotation which got lost in my last
post
was the big grin in my face during the writing of it :-))). I was just
amuesed by
the "AI-complete" and not *THAT* serious about it...

Marcin

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:46    [from the cache]
©2003-2011 Jasper Spaans