Archived posting to the Leica Users Group, 2003/11/28
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]> Hello Daniel, > > > Turing machines are finite state machines > > Technically they are not (they have infinite storage). > Ironically, it could be argued that "natural language" is > finite in that every word ever spoken by anyone since the > beginning of time is a finite number, and if you bound > their existence in time (i.e., the species will cease > to exist one day), natural language is finite, and therefore > is a finite state machine. :-) But ... :) the possibilities (speaking of the future and not the only the past) is infinite. That's where the "generative" approach came it (it would generate _all and only_ grammatically correct "sentences". Still fascinating goal. If nothing else he has made it perfectly clear that no one has every produced a grammar to describe properly any natural language. And he's pointed out what such a grammar will have to do to succeed. Problem is that grammar is not very important for natural language. It doesn't take into account semantics (meaning). > I agree with you though on the little impact Chomsky's > work has made in natural language, as opposed to formal > language. He would be very sad to hear that :) Hope you still have a day off! (and enjoy it in that case) Daniel - -- To unsubscribe, see http://mejac.palo-alto.ca.us/leica-users/unsub.html