>> With the chinese room, we arent doing any reasoning really, just looking up answers according to instructions.... but given that, how do we determine "understanding"?
>This was Searle's entire point. Mindless lookup is *not* understanding. It may *look* like >understanding from the outside (and if you have an infinitely large book that also has >mis-spellings, etc. -- you'll never be able to prove otherwise), but it is merely manipulation, >not intention. Thats my thought, but at what point is the manipulation responses just as good as "human" responses.... if they are both in a black box, to the outside observer, they are identical. Then they both could be said to have an equal "understanding" about what they are doing. Given that the grading is always done from an outside source. --- On Thu, 8/7/08, Mark Waser <[EMAIL PROTECTED]> wrote: From: Mark Waser <[EMAIL PROTECTED]> Subject: Re: [agi] Groundless reasoning --> Chinese Room To: [email protected] Date: Thursday, August 7, 2008, 2:49 PM :-) The Chinese Room can't pass the Turing test for exactly the reason you mention. >> Well in the Chinese Room case I think the "book of instructions" is infinitely large to handle all cases, so things like misspellings and stuff would be included.... and I dont think that was meant to be a difference. :-) Arguments that involved infinities are always problematical. Personally, I think that the intention was that you should accept the more reasonable and smaller infinity of just correct cases as being more in line with what Searle intended. This is obviously just speculation however and YMMV. >> With the chinese room, we arent doing any reasoning really, just looking up answers according to instructions.... but given that, how do we determine "understanding"? This was Searle's entire point. Mindless lookup is *not* understanding. It may *look* like understanding from the outside (and if you have an infinitely large book that also has mis-spellings, etc. -- you'll never be able to prove otherwise), but it is merely manipulation, not intention. ----- Original Message ----- From: James Ratcliff To: [email protected] Sent: Thursday, August 07, 2008 3:10 PM Subject: Re: [agi] Groundless reasoning --> Chinese Room >No, I said nothing of the sort. I said that Searle said (and I agree) that >a computer program that *only* manipulated formally defined elements without >intention or altering itself could not reach strong AIIs this part of the >Chinese Room? I looked and couldnt find that restriction. It would seem that to pass the Turing test, it would at least need to be able to add to its data, otherwise something as simple as the below would fail the Turing Test. Q: My name is James. AI: OK Q: What is my name? AI: *dont know didnt store it, or something like?* I read that the CR agent receives the input, looks up in a rulebook what to do, does it, and returns the output, correct? It seems that there is room for any action such as changing the rulebook in the middle of the process, maybe to add a synonym for a chinese word say. James _______________________________________ James Ratcliff - http://falazar.com Looking for something... From: Mark Waser <[EMAIL PROTECTED]> > So you are arguing that a computer program can not be defined solely > in terms of computational processes over formally defined elements? No, I said nothing of the sort. I said that Searle said (and I agree) that a computer program that *only* manipulated formally defined elements without intention or altering itself could not reach strong AI. > Computers could react to and interact with input back in the day when > Searle wrote his book. Yes. But the Chinese Room does *not* alter itself in response to input or add to it's knowledge. > A computer program is a computational process over formally defined > elements even if is able to build complex and sensitive structures of > knowledge about its IO data environment through its interactions with > it. Yes. This is why I believe that a computer program can achieve strong AI. > This is a subtle argument that cannot be dismissed with an appeal > to a hidden presumption of the human dominion over understanding or by > fixing it to some primitive theory about AI which was unable to learn > through trial and error. I was not dismissing the argument and certainly not making a presumption of human dominion over understanding. Quite the opposite in fact. I'm not quite sure why you believe that I did. Could you tell me which of my phrases caused you to believe that I did? Mark ----- Original Message ----- From: "Jim Bromer" <[EMAIL PROTECTED]> To: <[email protected]> Sent: Wednesday, August 06, 2008 7:32 PM Subject: Re: [agi] Groundless reasoning --> Chinese Room > On Wed, Aug 6, 2008 at 6:11 PM, Mark Waser <[EMAIL PROTECTED]> wrote: >> This has been a great thread! >> >> Actually, if you read Searle's original paper, I think that you will find >> that he... is *not* meaning to argue against the >> possibility of strong AI (since he makes repeated references to human as >> machines) but merely against the possibility of strong AI in machines >> where >> "the operation of the machine is defined solely in terms of computational >> processes over formally defined elements" (which was the current state of >> the art in AI when he was arguing against it -- unlike today where there >> are >> a number of systems which don't require axiomatic reasoning over formally >> defined elements). There's also the trick that Chinese Room is >> assumed/programmed to be 100% omniscient/correct in it's required domain. > > So you are arguing that a computer program can not be defined solely > in terms of computational processes over formally defined elements? > Computers could react to and interact with input back in the day when > Searle wrote his book. > A computer program is a computational process over formally defined > elements even if is able to build complex and sensitive structures of > knowledge about its IO data environment through its interactions with > it. This is a subtle argument that cannot be dismissed with an appeal > to a hidden presumption of the human dominion over understanding or by > fixing it to some primitive theory about AI which was unable to learn > through trial and error. > Jim Bromer > > agi | Archives | Modify Your Subscription agi | Archives | Modify Your Subscription ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=108809214-a0d121 Powered by Listbox: http://www.listbox.com
