> I don't think this is the operational sense of NLP as pursued by > applying linguistic theories in narrow AI setting. (e.g. Dynamic > Syntax, DRT, HPSG, ...)
but we want to apply NLP generally (i.e. not just in a narrow AI setting) > I was writing in context of Mark Waser language-specific solutions (as > I understand them), which if wished could be later reused in boarder > contexts. Actually, to convert from one language-specific version to another would be pretty trivial I believe. I think that all it would require would be tagging each word with a language, a languageA to languageB dictionary, and a quick overhaul of the parser and generator to make the link types be language specific. And yes, I *am* saying/claiming that I believe that this approach will pretty much automatically give you natural language translation. >> In other words, although there is enough special-purpose hardware in there to >> make it make sense to call language a "module", the full capability is so >> interwoven with general cognition that it can't be separated across a >> bottleneck. I agree strongly with Josh here . . . . > We stumble here on the meaning of capacity in this context. For > example, a general GUI library is not expected to be generally > intelligent. because I don't think that a general NLP is possible without general intelligence (again, of a *very* specific level). = = = = = = = = = = One of the things that I may have not been clear about saying (that is absolutely critical) is that I don't believe that NLP is possible *unless* you have a built-in world model. And, I don't know how you would have a world model without a certain amount of general intelligence. Fortunately, as I've stated in a previous email -- "I'm also getting the impression (or developing a stronger opinion) that there is far more declarative knowledge than actual rules at this level" -- with the importance of this being that I believe that declarative knowledge can be harvested while (initially, at least) rules are going to have to be hand-coded. Mark ----- Original Message ----- From: "Lukasz Stafiniak" <[EMAIL PROTECTED]> To: <agi@v2.listbox.com> Sent: Saturday, April 28, 2007 1:15 PM Subject: Re: [agi] rule-based NL system > On 4/28/07, J. Storrs Hall, PhD. <[EMAIL PROTECTED]> wrote: >> I disagree with this two ways. First, it's fairly well accepted among >> mainstream AI researchers that full NL competence is "AI-complete", i.e. that >> human-level intelligence is a prerequisite for NL. > > I don't think this is the operational sense of NLP as pursued by > applying linguistic theories in narrow AI setting. (e.g. Dynamic > Syntax, DRT, HPSG, ...) > >> Secondly, even the parsing >> part of NLP is part of a more general recursive sequence >> understander/generator, which is used for doing complex tasks with the hands >> (and the conjecture is that language bootstrapped itself on this capability). >> > I was writing in context of Mark Waser language-specific solutions (as > I understand them), which if wished could be later reused in boarder > contexts. > >> In other words, although there is enough special-purpose hardware in there to >> make it make sense to call language a "module", the full capability is so >> interwoven with general cognition that it can't be separated across a >> bottleneck. >> >> Josh >> > We stumble here on the meaning of capacity in this context. For > example, a general GUI library is not expected to be generally > intelligent. > > ----- > This list is sponsored by AGIRI: http://www.agiri.org/email > To unsubscribe or change your options, please go to: > http://v2.listbox.com/member/?& > ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415&user_secret=fabd7936