Philosophy Forums
Forums Links Articles Gallery
Style:
Language:


Calendar Quick Links open HelpRegister
Philosophy Forums » Philosophy of Language » How can Searle know that native Chinese understand chinese? » How can Searle know that native Chinese understand chinese?

How can Searle know that native Chinese understand chinese?

printPrint


Page: 1 2

How can Searle know that native Chinese understand chinese?
gabalus69
Initiate

Usergroup: Members
Joined: Jan 15, 2008
Total Topics: 3
Total Posts: 3
Posted 02/04/08 - 07:17 AM:
Subject: How can Searle know that native Chinese understand chinese? 		quote post
#1
Last week during the debate Turing vs Searle, the team defending Searle's position was asked:

how can you know that Chinese people understand chinese?

And the answer was far from good one, it was something like that: you know, supposing that Chines people don't understand Chinese at all, but they only have very good instruction how to combine the symbols is far-fetched.

For some who shere searleian point of view this question must be difficult. I'd like to ask you what ansewer shoud be given by the searle-defenders.


muxol
yuletide

Usergroup: Members
Joined: Dec 07, 2003
Total Topics: 1
Total Posts: 1812
Posted 02/04/08 - 09:28 AM:
quote post
#2
This seems to be in the wrong forum, but it's a little interesting.

There is an incredibly boring and general skeptic worry: How does one know *anything*? It seems to me as if you are almost asking this sort of question. Is that right? Otherwise your question seems to me misguided. I'll explain.

In the Searle thought experiment, we are asked (indirectly) to accept some basic notion of *understanding*. That notion has intuitive force, and so it grounds the plausibility of our accepting such a notion. Now *given this acceptance* it seems that the person in the Chinese room just doesn't understand, in the sense we have accepted.

It doesn't matter whether Searle knows *anyone* (I suppose 'someone' would be more accurate), let alone Chinese people, understand Chinese. We've accepted it by accepting some basic notion of understanding from which it follows that Chinese people understand Chinese. Plus the intuitive force of that basic notion grounds the plausibility that Chinese people do in fact understand Chinese, but this is somewhat beside the point. If we can't accept that Chinese people understand Chinese, then what sort of notion of understanding are we working with? Is it at all interesting?

Edited by muxol on 02/04/08 - 09:39 AM
jdrw
definitely ~d1

Usergroup: Moderators
Joined: Apr 11, 2004
Location: Pennsylvania
Total Topics: 11
Total Posts: 1490
Posted 02/04/08 - 12:44 PM:
quote post
#3
Seems to me that by "understand" one thing that can be involved is recognition--that is, re-cognition. If the sentence I am supposed to translate says in English "the cat is on the mat" I have re-cognitions of the referents as well as the syntactical knowledge by which to construct meaning. My re-cognition of the word cat involves neural firing patterns including memory experience of the phenomena I associate with that word. Similarly with mat. Additionally, via certain other neural firing patterns I have an experience I call understanding the spatial relationship described by the preposition and the syntax. Thus my knowledge of both the semantic and syntactic aspects result in certain neural firing patterns in my brain resulting in an experience I call “understanding the sentence.”

Whereas when I read a Chinese translation, there is no re-cognition of any referents. So since I have neither semantic nor syntactic knowledge, the patterns that fire up in my brain do not produce an experience I call “understanding the sentence.”

When people who actually understand both English and Chinese translate from one to the other, however, they can be said to “understand” both versions because both versions produce neural firings that produce a certain experience resulting from semantic and syntactic knowledge that we all have agreed to call “understanding the sentence.”


Cheers.
jd

_____________________
OTOH I might be exhaustively wrong about everything I've ever thought--with the possible exception of this sentence.
ctaylor
Initiate

Usergroup: Members
Joined: Jan 17, 2008
Total Topics: 1
Total Posts: 8
Posted 02/09/08 - 10:41 AM:
quote post
#4
I haven't read the relevant Searle here, but from a Wittgenstein perspective, it looks like the question arises from the more specific skeptical concern with our knowledge of other minds. This puts more pressure on the use of "know" in the question, and perhaps less pressure (or at least less immediate pressure) on the meaning of "understand". Then we have to ask under what conditions it would make sense to doubt such knowledge.

By the way, why wouldn't this be an appropriate topic for this forum?
jdrw
definitely ~d1

Usergroup: Moderators
Joined: Apr 11, 2004
Location: Pennsylvania
Total Topics: 11
Total Posts: 1490
Posted 02/09/08 - 07:26 PM:
quote post
#5
ctaylor wrote:


By the way, why wouldn't this be an appropriate topic for this forum?

It's in the appropriate forum now but originally it wasn't. I moved it here after muxol's comment.


jd

_____________________
OTOH I might be exhaustively wrong about everything I've ever thought--with the possible exception of this sentence.
modularsky
Dilettante
Avatar

Usergroup: Members
Joined: Sep 28, 2007
Location: Poughkeepsie, NY
Total Topics: 7
Total Posts: 59
1 of 1 people found this post helpful
Posted 02/10/08 - 07:27 AM:
quote post
#6
I'm not sure whether it was Searle who came up with the thought experiment really, but OP should have definitely elaborated on what exactly the problem is here. As I remember from my philosophy of Mind class readings, the premise is:

Given someone in a closed, windowless and doorless room who has a massive volume that has rules for each and every character and character combination in Chinese (that is, whenever the person sees a bunch of characters, she need only turn to the proper page and the appropriate response will be listed there). The question here is less a question of language than a question of minds. In this experiment, we are tempted to say that the person, no matter how proficient she gets at responding to outside questions, does not understand chinese. That is: even though we do not know whether the person inside is chinese or not, and in fact if we were to only go on her responses we would believe she is Chinese (because they are all very good responses to questions in Chinese).

This is effectively an objection to functionalist/behaviorist views of understanding. For under a functionalist picture, understanding is nothing more than being able to perform functions in such a way that your actions fall in line with the actions of someone who supposedly "understands" a concept as well. This is to say: if a computer was able to translate Chinese and respond to questions in Chinese perfectly, the functionalist will argue that the computer, for all purposes, understands Chinese.

Now, Being the sort of Wittgensteinian behaviorist that I am, I'm inclined to believe the person in the Chinese room understands Chinese just as much as a native speaker does, by dint of the fact that we are not privy to the Chinese speaker's mind-states (we are privy to no-ones, which is where the other minds question comes in). Anyway, hopefully this response helps to refigure the debate here towards something more constructive.

_____________________
"Es ist der geist, der sich den Körper baut!"
-Schiller
ctaylor
Initiate

Usergroup: Members
Joined: Jan 17, 2008
Total Topics: 1
Total Posts: 8
Posted 02/10/08 - 08:03 AM:
quote post
#7
What can "understand Chinese" mean except that one responds in such-and-such a way when Chinese is spoken to her? From this (Wittgenstein's) point of view, the question posed to Searle's defenders about this hypothetical room is the wrong picture of an other's mind. I'm not sure if Searle could/would have answered the question this way, however.
modularsky
Dilettante
Avatar

Usergroup: Members
Joined: Sep 28, 2007
Location: Poughkeepsie, NY
Total Topics: 7
Total Posts: 59
Posted 02/10/08 - 09:39 AM:
quote post
#8
Well, the Gedankenexperiment wasn't intended to sound plausible. The whole point (at least as far as I've seen it) of it is to show the ridiculousness of a functionalist/behaviorist picture of understanding. It is meant to suggest that understanding is something beyond the ability to perform in line with what empirical investigation would show to constitute 'understanding'.

_____________________
"Es ist der geist, der sich den Körper baut!"
-Schiller
Arlonian
Initiate

Usergroup: Members
Joined: Feb 14, 2008
Total Topics: 0
Total Posts: 1
Posted 02/14/08 - 06:17 PM:
quote post
#9
It is reasonable to conclude that people understand the language they are speaking when their responses depend on information from their other senses, and the responses they give tend to be accurate. You can have an excellent conversation with someone about which snow conditions you prefer, and they not understand a word. This may only become clear when you go out to an icy ski field and they comment without sarcasm on the lovely powder.

Oh - just an aside. It is begging the question against most AI to define understanding as "known patterns of brain states".
natutita
Initiate
Avatar

Usergroup: Members
Joined: Feb 15, 2008
Location: Buenos Aires, Argentina
Total Topics: 0
Total Posts: 12
Posted 02/19/08 - 05:16 AM:
quote post
#10
I thought one of the points Searle was trying to argue in favor of was that syntactical properties were neither necessary nor sufficient for semantics, and given that the mind is a semantic "mechanism", no machine will ever be able to replace it and "think" properly. You have to take into account that Searle has a very strong cartesian influence, he doesnt accept (as functionalism and behaviorism) the "3rd person" perspective.

Ill get into more detail later.
grin
Shalomon
Lantern holder

Usergroup: Members
Joined: Mar 15, 2008
Total Topics: 0
Total Posts: 16
Posted 03/17/08 - 04:17 PM:
quote post
#11
The Chinese room as stated in modularsky's post isn't a great thought experiment anyway, because no such mechanism would ever be able to respond in the same way that something that truly understood Chinese would. For one thing, it wouldn't be able to take context into account, so it couldn't understand references to previous exchanges. For another, the timing would be entirely off. I'm not talking about it being slow; that could be remedied by devising speedier mechanisms. What I mean is that, for example, one query which would take a true Chinese speaker a lot of thought to answer would be just as fast as a much simpler query.

I have heard the Chinese room phrased in other ways that overcome these deficiencies (for example, instead of a single static volume, the operator maintains a continually changing and inter-related system of notes and rules. But in such situations I think a very good case could be made that the room does understand Chinese at that point. The operator does not, of course; the understanding is rather embedded in the system composed by the whole. Even dropping the behaviorist perspective, any such system would have to be essentially as complex as the human brain (or at least the portions used for understanding speech), and thus capable of achieving any non-arbitrary definition of understanding.
modularsky
Dilettante
Avatar

Usergroup: Members
Joined: Sep 28, 2007
Location: Poughkeepsie, NY
Total Topics: 7
Total Posts: 59
Posted 03/17/08 - 06:57 PM:
quote post
#12
Shalomon wrote:
in such situations I think a very good case could be made that the room does understand Chinese at that point.


That's exactly it though, the answer to this thought experiment is not a choice between behaviorist or eliminative materialist philosophies, but is instead a functionalist philosophy. The way the experiment is set up necessarily means you either grant the room itself an understanding or you posit understanding as some non-quantifiable and indiscernible quality that only humans possess.

_____________________
"Es ist der geist, der sich den Körper baut!"
-Schiller
Shalomon
Lantern holder

Usergroup: Members
Joined: Mar 15, 2008
Total Topics: 0
Total Posts: 16
Posted 03/18/08 - 11:09 AM:
quote post
#13
Well, at that point you get into the "philosophical zombie" conunudrum: can you have something that is constructed exactly like a person, and acts just like a person, but isn't "conscious"? The only way you can do so is to assume that there is some supernatural extension of a person that's neither caused by material means nor has any effect on behavior; in short, that is entirely divorced from the material world. At which point you have to ask yourself whether such a thing matters. It's certainly not consciousness in the normal sense of the word, as something intimately tied to our sensation of the world and our reaction to it.
jaoman
On the Road...
Avatar

Usergroup: Sponsors
Joined: Mar 13, 2004
Location: Vancouver, Canada
Total Topics: 84
Total Posts: 1312
Posted 03/18/08 - 11:30 AM:
quote post
#14
Shalomon wrote:
Well, at that point you get into the "philosophical zombie" conunudrum: can you have something that is constructed exactly like a person, and acts just like a person, but isn't "conscious"?


But that's very crucial to Searle's overall argument. The Chinese room is not constructed exactly like a person. So question the Chinese room is really posing is whether something that is constructed differently from a person, but still demonstrates person-like behavior, would necessarily have the same mental states as a person. If the answer is no (and I don't see how it could be yes), then the argument effectively disproves the proof of AI presented by talking computers.

_____________________
"With no relation to class or social background, whether it suits them or not, people yearn for a dream. Sustained by a dream, hurt by a dream, revived by a dream, killed by a dream. And even after being abandoned by a dream, it continues to smolder from the bottom of one's heart... probably until the verge of death. A man should envision such a lifetime once. A life spent as a martyr to the god named "dream."
- Kentaro Miura
Shalomon
Lantern holder

Usergroup: Members
Joined: Mar 15, 2008
Total Topics: 0
Total Posts: 16
Posted 03/18/08 - 12:41 PM:
quote post
#15
Well, I am positing that any Chinese room that is able to handle the task of essentially passing a Turing test in Chinese will be so complex that no significant difference will be distinguishable between its workings and the workings of a brain. Sure, the superficial trappings are different; it uses paper notes instead of neurons, but the actual structure and operation of such a room would be essentially identical to a brain (to the degree that our brains are identical to each other, anyway). In other words, person-like behavior necessitates person-like construction, as long as you realize that what is meant by that isn't that it has to be biologically human, it just has to have the organizational patterns of a person. This isn't behaviorist because we're talking about the actual structure. It's just not a physical structure. Just like the same computer program could be run on different sets of hardware (or even, theoretically, be worked out by hand), because it's the pattern of information that's important.
jaoman
On the Road...
Avatar

Usergroup: Sponsors
Joined: Mar 13, 2004
Location: Vancouver, Canada
Total Topics: 84
Total Posts: 1312
Posted 03/18/08 - 12:57 PM:
quote post
#16
Shalomon wrote:
Well, I am positing that any Chinese room that is able to handle the task of essentially passing a Turing test in Chinese will be so complex that no significant difference will be distinguishable between its workings and the workings of a brain. Sure, the superficial trappings are different; it uses paper notes instead of neurons, but the actual structure and operation of such a room would be essentially identical to a brain (to the degree that our brains are identical to each other, anyway). In other words, person-like behavior necessitates person-like construction, as long as you realize that what is meant by that isn't that it has to be biologically human, it just has to have the organizational patterns of a person.


Not really. The Chinese room, as it stands now, has all the tools it needs to fake it. Give it enough of the right notes, and how would you tell the difference? Now you are positing that you would. Well, great, but you must specify your method of deduction.

And really, what could you look for? I can't even tell if my girlfriend is faking her orgasms without checking on her biology, a method which no longer applies. Good for my ego if I plan to sleep with Chinese rooms, but bad for intellectual progress if I take "Oh yes, I'm *****ing, jaoman!" as conclusive proof. sticking out tongue


Edited by jaoman on 03/18/08 - 01:04 PM

_____________________
"With no relation to class or social background, whether it suits them or not, people yearn for a dream. Sustained by a dream, hurt by a dream, revived by a dream, killed by a dream. And even after being abandoned by a dream, it continues to smolder from the bottom of one's heart... probably until the verge of death. A man should envision such a lifetime once. A life spent as a martyr to the god named "dream."
- Kentaro Miura
Shalomon
Lantern holder

Usergroup: Members
Joined: Mar 15, 2008
Total Topics: 0
Total Posts: 16
Posted 03/18/08 - 03:52 PM:
quote post
#17
jaoman wrote:

Not really. The Chinese room, as it stands now, has all the tools it needs to fake it. Give it enough of the right notes, and how would you tell the difference? Now you are positing that you would. Well, great, but you must specify your method of deduction.

I gave a bit of it above: two things to look for are that it handles context (ie. it can keep track of a conversation over multiple exchanges) and that it gets the relative timing "right" (ie. it pauses where a person would pause and answers quickly when required; of course both would happen much slower than normal). A simple lookup table architecture wouldn't be able to handle such things. Witness the difficulty we've had in creating AI: if it was simply a matter of gathering enough facts, we'd have AI easily capable of emulating a human by now. Instead, we've reached the limits of expert systems and are instead exploring other avenues such as neural networks (even those are really to simple to emulate higher cognition) which are (surprise!) intentionally modeled on the architecture of the brain.
jaoman
On the Road...
Avatar

Usergroup: Sponsors
Joined: Mar 13, 2004
Location: Vancouver, Canada
Total Topics: 84
Total Posts: 1312
Posted 03/19/08 - 09:59 AM:
quote post
#18
Shalomon wrote:
I gave a bit of it above: two things to look for are that it handles context (ie. it can keep track of a conversation over multiple exchanges) and that it gets the relative timing "right" (ie. it pauses where a person would pause and answers quickly when required; of course both would happen much slower than normal). A simple lookup table architecture wouldn't be able to handle such things. Witness the difficulty we've had in creating AI: if it was simply a matter of gathering enough facts, we'd have AI easily capable of emulating a human by now.


First of all, “by now”? AI has been a hot topic for what...20 years or so. Whereas, the greatest minds in history have been trying to figure out how the mind works for ages.

The problem is that what you describe does not pose much of a hardware problem. If we knew the markers to look for, the stuff that we respond to unconsciously, then we could isolate those from the input data and program an appropriate relationship. if a, do b. The Chinese room can totally do that. Admittedly, our greatest minds are apparently no sharper than the greatest minds of history, which is disappointing.

Here's the sticker: what do you suppose a neural network, or any other hardware arrangement, no matter how sophisticated, can do if the code isn't forthcoming? Neural networks are faster, they are better organized, but in the end of the day, they are a Turing machine just like your PC. That's how they process. So nothing they can do is theoretically impossible for your PC or the Chinese room. The smart boys are just stalling for time.

_____________________
"With no relation to class or social background, whether it suits them or not, people yearn for a dream. Sustained by a dream, hurt by a dream, revived by a dream, killed by a dream. And even after being abandoned by a dream, it continues to smolder from the bottom of one's heart... probably until the verge of death. A man should envision such a lifetime once. A life spent as a martyr to the god named "dream."
- Kentaro Miura
Shalomon
Lantern holder

Usergroup: Members
Joined: Mar 15, 2008
Total Topics: 0
Total Posts: 16
Posted 03/19/08 - 11:29 AM:
quote post
#19
jaoman wrote:

The problem is that what you describe does not pose much of a hardware problem. If we knew the markers to look for, the stuff that we respond to unconsciously, then we could isolate those from the input data and program an appropriate relationship. if a, do b. The Chinese room can totally do that. Admittedly, our greatest minds are apparently no sharper than the greatest minds of history, which is disappointing.

Here's the sticker: what do you suppose a neural network, or any other hardware arrangement, no matter how sophisticated, can do if the code isn't forthcoming? Neural networks are faster, they are better organized, but in the end of the day, they are a Turing machine just like your PC. That's how they process. So nothing they can do is theoretically impossible for your PC or the Chinese room. The smart boys are just stalling for time.

Right, that's exactly what I've been saying. Anything a neural net (ie., a brain) can do, a computer or Chinese room can also (theoretically) do, but only by mimicking its structure and processes. It is the software that composes understanding, the actual notes written on all those slips of paper and the manner in which they are manipulated, not the hardware. Thus to say that a Chinese room has no understanding because it is made of slips of paper is to miss the point of understanding. The reason we consider the Chinese room to be incapable of understanding is that we misconceive it, not being able to comprehend the sheer magnitude of emergent complexity that would truly be able to mimic the correct behavior.
jaoman
On the Road...
Avatar

Usergroup: Sponsors
Joined: Mar 13, 2004
Location: Vancouver, Canada
Total Topics: 84
Total Posts: 1312
Posted 03/20/08 - 01:49 PM:
quote post
#20
Shalomon wrote:
Right, that's exactly what I've been saying. Anything a neural net (ie., a brain) can do, a computer or Chinese room can also (theoretically) do, but only by mimicking its structure and processes.


Actually, it only needs to analyze input for appropriate markers and then produce an appropriate output. It does not need to mimic structure and process. It can do this using both different structure and different processes than the brain, it simply needs to be programmed with the logic of interaction.

Shalomon wrote:
It is the software that composes understanding, the actual notes written on all those slips of paper and the manner in which they are manipulated, not the hardware. Thus to say that a Chinese room has no understanding because it is made of slips of paper is to miss the point of understanding. The reason we consider the Chinese room to be incapable of understanding is that we misconceive it, not being able to comprehend the sheer magnitude of emergent complexity that would truly be able to mimic the correct behavior.


I fear you are veering away from the question I posed to you earlier. Whatever processes the Chinese room goes through to produce its output are irrelevant. Their nature is mandated by the hardware that composes the Chinese room. We've agreed that the hardware does not, by itself, create intelligence. In other words, the construction of neural net networks that mimic the brain will not produce anymore or less intelligence in a computer than the regular CPU.

Well then, here we have the CPU, the Chinese room. Input goes in, output comes out. The program inside it is sufficient to give the appearance of understanding. Now, in order for us to assert that this appearance is, in fact, reality (the Chinese room really does understand) at least one more condition needs to be met. The Chinese room needs to be conscious of what it's doing. It must have a sense of itself, of the information its processing, and of the external reality to which it is returning output. And if we want human-like understanding, the Chinese room must furthermore have these sensations in a way that is similar to how a person has these sensations. The question is: can you demonstrate beyond even reasonable doubt (much less, all doubt) that these conditions have been met?

The genius of the Chinese room is that it demonstrates that such an assertion is untenable. The argument is inconclusive. We don't know whether any kind of magical duality has arisen out of the card shuffling going on in the Chinese room. And if we don't know for the Chinese room, neither do we know about the CPU. The AI therefore is reduced to nothing more than a really sophisticated mechanical process that produces the type of output that we are used to associating with understanding. However, whether that output has any more understanding than talking pictures in theaters remains questionable.

_____________________
"With no relation to class or social background, whether it suits them or not, people yearn for a dream. Sustained by a dream, hurt by a dream, revived by a dream, killed by a dream. And even after being abandoned by a dream, it continues to smolder from the bottom of one's heart... probably until the verge of death. A man should envision such a lifetime once. A life spent as a martyr to the god named "dream."
- Kentaro Miura
Shalomon
Lantern holder

Usergroup: Members
Joined: Mar 15, 2008
Total Topics: 0
Total Posts: 16
Posted 03/21/08 - 12:14 PM:
quote post
#21
jaoman wrote:
Actually, it only needs to analyze input for appropriate markers and then produce an appropriate output. It does not need to mimic structure and process. It can do this using both different structure and different processes than the brain, it simply needs to be programmed with the logic of interaction.

And I am asserting that this is not the case. At some level, in order to provide responses demonstrating consciousness, one must actually be conscious. I am not talking about behaviorism; consciousness is an internal pattern, not a series of external behaviors. However, to say that something can provide all the behaviors associated with consciousness and yet not have the internal pattern of consciousness implies that there is some other pattern that will do; but there isn't. If you contend that there is, I simply ask you to provide one. The evidence is against you: according to chaos theory, even small differences in complex systems get magnified until behavior is entirely different. And of course history is the same: expert systems have experienced diminishing returns as complexity increases.

jaoman wrote:
I fear you are veering away from the question I posed to you earlier. Whatever processes the Chinese room goes through to produce its output are irrelevant. Their nature is mandated by the hardware that composes the Chinese room. We've agreed that the hardware does not, by itself, create intelligence. In other words, the construction of neural net networks that mimic the brain will not produce anymore or less intelligence in a computer than the regular CPU.

Agreed, as long as you realize that what the regular CPU would have to do is mimic the internal information pattern of the neural net and brain.

jaoman wrote:
Well then, here we have the CPU, the Chinese room. Input goes in, output comes out. The program inside it is sufficient to give the appearance of understanding. Now, in order for us to assert that this appearance is, in fact, reality (the Chinese room really does understand) at least one more condition needs to be met. The Chinese room needs to be conscious of what it's doing. It must have a sense of itself, of the information its processing, and of the external reality to which it is returning output. And if we want human-like understanding, the Chinese room must furthermore have these sensations in a way that is similar to how a person has these sensations. The question is: can you demonstrate beyond even reasonable doubt (much less, all doubt) that these conditions have been met?

Yes, you can, if the Chinese room operates in the same way that human consciousness does. In other word, it uses the same pattern of information, it's just stored as words on paper slips instead of chemical potentials in neurons. In such a case, you would have to have the same certainty of consciousness in the Chinese room that you would in a human.

jaoman wrote:
The genius of the Chinese room is that it demonstrates that such an assertion is untenable. The argument is inconclusive. We don't know whether any kind of magical duality has arisen out of the card shuffling going on in the Chinese room. And if we don't know for the Chinese room, neither do we know about the CPU. The AI therefore is reduced to nothing more than a really sophisticated mechanical process that produces the type of output that we are used to associating with understanding. However, whether that output has any more understanding than talking pictures in theaters remains questionable.

If what you say is true, it is true about the brain just as much as it is the Chinese room or the CPU.
jaoman
On the Road...
Avatar

Usergroup: Sponsors
Joined: Mar 13, 2004
Location: Vancouver, Canada
Total Topics: 84
Total Posts: 1312
Posted 03/21/08 - 02:15 PM:
quote post
#22
Shalomon wrote:
And I am asserting that this is not the case. At some level, in order to provide responses demonstrating consciousness, one must actually be conscious. I am not talking about behaviorism; consciousness is an internal pattern, not a series of external behaviors. However, to say that something can provide all the behaviors associated with consciousness and yet not have the internal pattern of consciousness implies that there is some other pattern that will do; but there isn't. If you contend that there is, I simply ask you to provide one. The evidence is against you: according to chaos theory, even small differences in complex systems get magnified until behavior is entirely different. And of course history is the same: expert systems have experienced diminishing returns as complexity increases.


Great answer. I agree completely!

But here's the kicker: the nature off the internal processes and structure will always be influenced by the components of which they are made. That is scientific fact. So why stop at process and structure of the brain? Why not go all the way and import the same components? Because if we don't, we will always be making compromises to account for the different properties of our synthetics.

Shalomon wrote:
Yes, you can, if the Chinese room operates in the same way that human consciousness does. In other word, it uses the same pattern of information, it's just stored as words on paper slips instead of chemical potentials in neurons. In such a case, you would have to have the same certainty of consciousness in the Chinese room that you would in a human.


The problem with this is that patterns are subjective. We perceive them. They do not exist outside out perception. Material composition is the only quality proven to have objective existence.

It's true, hardware without software is useless. But it is also true that software cannot run properly on hardware that can't support it. It may do something, but no power insures that it what you desire. To create sound, you need hardware that can support the output of sound. It follows that to create consciousness you need hardware that can generate that as well. Just anything will not do.

Shalomon wrote:
If what you say is true, it is true about the brain just as much as it is the Chinese room or the CPU.


It would be, if I didn't already have irrefutable proof that brains work: my brain.

_____________________
"With no relation to class or social background, whether it suits them or not, people yearn for a dream. Sustained by a dream, hurt by a dream, revived by a dream, killed by a dream. And even after being abandoned by a dream, it continues to smolder from the bottom of one's heart... probably until the verge of death. A man should envision such a lifetime once. A life spent as a martyr to the god named "dream."
- Kentaro Miura
Shalomon
Lantern holder

Usergroup: Members
Joined: Mar 15, 2008
Total Topics: 0
Total Posts: 16
Posted 03/21/08 - 03:57 PM:
quote post
#23
jaoman wrote:

It's true, hardware without software is useless. But it is also true that software cannot run properly on hardware that can't support it. It may do something, but no power insures that it what you desire. To create sound, you need hardware that can support the output of sound. It follows that to create consciousness you need hardware that can generate that as well. Just anything will not do.

But that's true only for I/O. It's not true of the underlying algorithms. When you're talking software, any sufficiently advanced computer can emulate any other, and thus any computer can run any software. Essentially, it is possible to mimic the functions of hardware in software. Software programs can even be worked out, theoretically, by hand, which is pretty much exactly what the Chinese room is doing. If software relied on hardware, a Chinese room wouldn't be able to emulate a computer, either, but we know that it can despite the vast differences in hardware.

jaoman wrote:
It would be, if I didn't already have irrefutable proof that brains work: my brain.

Which is why I don't think it is true.
jaoman
On the Road...
Avatar

Usergroup: Sponsors
Joined: Mar 13, 2004
Location: Vancouver, Canada
Total Topics: 84
Total Posts: 1312
Posted 03/22/08 - 09:57 AM:
quote post
#24
Shalomon wrote:
But that's true only for I/O. It's not true of the underlying algorithms. When you're talking software, any sufficiently advanced computer can emulate any other, and thus any computer can run any software. Essentially, it is possible to mimic the functions of hardware in software. Software programs can even be worked out, theoretically, by hand, which is pretty much exactly what the Chinese room is doing. If software relied on hardware, a Chinese room wouldn't be able to emulate a computer, either, but we know that it can despite the vast differences in hardware.


Try playing World of Warcraft on the Chinese room and see how much gold you can earn. An algorithm simple represents your understanding, it doesn't mean anything for the Chinese room, even if it's implementing the algorithm, if it doesn't have the composition to process that algorithm to the desired effect.

What does “only for I/O” supposed to mean? I/O is nine tenth of the point. If there is no conscious output from the Chinese room, if only to itself, the exercise is a failure. In order of us to proclaim AI, the Chinese room needs provable self consciousness. The Chinese room has no such capacity.

_____________________
"With no relation to class or social background, whether it suits them or not, people yearn for a dream. Sustained by a dream, hurt by a dream, revived by a dream, killed by a dream. And even after being abandoned by a dream, it continues to smolder from the bottom of one's heart... probably until the verge of death. A man should envision such a lifetime once. A life spent as a martyr to the god named "dream."
- Kentaro Miura
modularsky
Dilettante
Avatar

Usergroup: Members
Joined: Sep 28, 2007
Location: Poughkeepsie, NY
Total Topics: 7
Total Posts: 59
Posted 03/31/08 - 10:33 AM:
quote post
#25
jaoman wrote:

What does “only for I/O” supposed to mean? I/O is nine tenth of the point. If there is no conscious output from the Chinese room, if only to itself, the exercise is a failure. In order of us to proclaim AI, the Chinese room needs provable self consciousness. The Chinese room has no such capacity.


I'm gonna step in at this point. What you guys have all been riffing on has been (more or less) what was discussed earlier on in this thread (though more thoroughly fleshed out). What is at stake here, and has not been (nor will ever be, in my mind) solved is the argument between essentialism and anti-essentialism. You argue that the chinese room is not self-conscious. Well, what makes you think that you yourself are self-conscious? Of course, you can respond: "I know because I am" - similar to your response to whether or not you know a brain works - "my brain". I am inclined to agree with you, because the other side represents a slightly-developed version of Wittgensteinian meaning-skepticism, and we can't accept skeptical arguments if we are to philosophize (though they will help test the viability of our theories).

In short, what is at issue is: Does something that is "designed" by humans to behave in every way that a human does--this includes giving the proper outputs so as to convince an interrogator that this 'chinese room' has self-consciousness (when it is asked, "do you care if your girlfriend has orgasms when you have sex" it responds with a nervous "yes", etc.) For any way that you respond to me (any rule you make), an exception exists. If you posit that it is some innately human characteristic, then what if we simply fabricate a human? If we create a carbon (forgive my pun) copy of a human being using completely artificial means, is this entity a human? You will be impelled to say no. Well, if we create an environment that simulates the earth 4.3 billion years ago, and then are somehow able to speed up evolution, are the life-like entities that result 'human'? I think once again you are impelled to say no, because at stake is whether or not 'consciousness' can be fabricated.

_____________________
"Es ist der geist, der sich den Körper baut!"
-Schiller
Download thread as

Page: 1 2



You don't have permission to post.

Please login or register.

Contact the Administration

25 total queries
This page was created in 4.72 seconds
Memory used: 6777576 bytes
Server Status: time since last reboot is 47 days, 14:40, load average: 3.24, 2.38, 1.98