| |
|
Science Forum Index » Cognitive Science Forum » Gold's Theorem and Natural Language Learning
Page 1 of 1
|
| Author |
Message |
| LauLuna |
 Posted: Wed Apr 23, 2008 12:55 pm |
|
|
|
Guest
|
I've recently known of Gold's theorem (1967).
It states that any class C of languages containing an enumerable
number of them L1, L2, L3,... and still a language L-omega, such that
1.for any n, Ln is strictly included in Ln+1 and in L-omega;
2. whatever is in L-omega is in some Ln;
is unlearnable, i.e there is no learner able to learn each L-alpha in
C given any environment of sentences of L-alpha.
But I don't see exactly what bearing the theorem should have on
natural language learning. Is there any result to the effect that
natural language can be conceived of as an unlearnable class of
languages?
Thanks |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Sat May 17, 2008 12:10 am
|
|
