What is it that gives people a consciousness?

Post Reply New Topic

But, but.... you are talking to one now. Or... are you suggesting that I do not exist?

But you're not an AI.

Seriously, I am fascinated with the mathematics of making good choices (AI, OR, economics, game theory, etc.), to the point that I have to 'hold myself back' and really focus on the other essential subjects that I tend to overlook. I imagine that robots are capable of everything up to and including building a fully autonomous spacefaring 'civilization'. By almost any measure, they'd probably do s**t better than we ever did. But the whole thing would be a very, very elaborate trick. They'd be no more conscious than the AMD processor that's handling the characters in this post...

While wrongheaded from many philosophical viewpoints, I think p-zombies at least make the case fairly clear, and there is sufficient literature on the subject so that allegations of not having "clear reasoning" doesn't seem very sensible to me, even if you disagree with the thought experiment.

Thanks for bringing up 'p-zombies'. I never heard of them before.

In that context, fully autonomous weak AIs would be more or less 'neurological p-zombies'.

But there you go... quoting the exact thing I was talking about. I object to every premise in the p-zombie "thought experiments" - and I find it tedious to list even some of them:

• That our current languages are good enough to talk about "<whatever>".
• That one could even approach the mimicry, without already having created an entity just as "<whatever>" as a human.
• That anyone has come remotely near a good description of "<whatever>".
• That even those who think they have defined "<whatever>" have yet to come up with any reason why the p-zombie would not have it. (With the exception of invoking some "higher being", using much hand waving and obfuscation.)

Substitute your particular "experiment of the month" for <whatever>. E,g, self-aware, consciousness, sentience and so forth.

You're missing the point. The primary purpose isn't "why shouldn't X have <whatever>". It's "Is it logically possible for p-zombies to exist". If the latter holds, we have some kind of dualism.

I find the 'Chinese room' argument pretty convincing tbh. It does not necessarily suggest dualism outright, but can be taken that way

Last edited by chever on 22 Sep 2008, 8:50 am, edited 1 time in total.

dualism? in OUR universe? say it ain't so!

Merle

As I say, the premises are all in your third sentence.

• What do you mean by "logically possible"? I suspect you are overworking the phrase. If you are asking whether the TE can be imprecisely expressed in words, I suppose the answer is you want is "yes". If you are asking for all the implied meanings behind those words to be automatically accepted as "a priori" absolute truths, the answer is certainly "no way".
• A p-zombie only becomes a meaningful concept if you will clearly define that particular attribute which somehow distinguishes it from a non-zombie.
• "Existence" comes with a lot of baggage.

Make your postulate some kind of dualism and you can certainly (rather trivially) seem to derive a proof of some kind of dualism.

you know when your in a room full of uber AS when the discussion flicks from p-zombies to dualism.
the thread title cannot be mentioned lest it defies definition.
artificial inttelligence gets mentioned yet only to debate if it is possible.
some one asks if he does or doesnt exist, and gets no answer.

one of the thread fonts has a half pint of beer as his avatar.

then suddenly we all end up asking "wtf is a chinese room argument"?

this is responded to by a distinction between OUR universe and another. with an automatic AS presumtion that everyone knows there is another universe, presumably it has to be paralle to ours otherwise ours woudl bump into it.

some where in all of this, we have to avoid an upset but persistent NT Claimant who seeks out weak arguments and flames random individuals for being..........for just being!

did someone mention n****r?

no, not yet, these things usually devolve into someone being called 'worse' than Adolf Hitler, first.

Merle

Yes. Another of those thought experiments that I've been happy to dismiss.

The basis of the whole idea is that you are clear on what "understands Chinese" means. Once that concept has been snuck in, under the radar, to start with, any argument is over.

Actually, I'd argue for a greater mind, if the "program" used to generate correct responses could do so without merely "understanding Chinese", but actually having such a "meta-understanding" of Chinese, that it would be almost instantly be applicable to a whole range of languages, cultures, worlds, and so on.

'able to attach semantic value to symbols'

Machines do not do this.

And, of course, dictionary['双射'] = 'bijection' does not count as 'attaching semantic value'.

Another can of worms. Which particular definition of "semantic value" would you be using?
I write a "C" program. The lexical symbols are one thing, the syntax is another. The semantics? Well, if the program does what I intended it to do, I guess I have imbued it with sufficient "ability" to "attach" "semantic" "value" to "symbols" for my purpose. If my purpose were to create a peer intelligence, I see no inherent reason why I should not be able to do that.


Back to bald statements, again? Why do you think this statement is true? Or rather, if they don't (can't) do it, then neither can I, as I am a machine.

But that would just be more symbolic manipulation.

The dictionary example I gave was a little trite, but it serves a purpose: if you dig deep enough, any emulation of sentience breaks down into more syntactic content. That's why Lisp atoms evaluate to themselves ... there's nowhere else to go. They're syntactic data that become syntactic data. They don't have meanings.

Now I'm not saying that you couldn't make robots and computers capable of exceeding human intellectual ability in effect. In fact, I wouldn't be surprised at all if they started making necklaces of our wisdom teeth as souvenirs once they start to go into battle against us. AI's will make great mathematicians, great scientists, great engineers, great soldiers and ... who knows ... maybe even artists?

But the critical distinction is that the wires, the circuit boards, the ANNs, all the little ins and outs ... wouldn't be 'aware' of a damn thing. Even if they had their own 'civilization'. In theory, it's possible to create a general AI with strictly mechanical components. It would be very hard and expensive and huge, but it's possible. But no one would claim that such a thing, made of gears and levers, had thoughts. The same applies to digital computers.



Obviously you're not.

I'm glad you finally see my point. You have placed the word 'aware' in quotes, thus acknowledging the fact that you have no more idea how to define that than I do.

I fully admit that I find "awareness" a fascinating subject. To be more exact, in a sense, it is the only really interesting subject. It is my opinion that any sufficiently complex system (be it built of cells, transistors or mud) is likely to become demonstrably aware, and self-aware - at least to the same degree as I am.

And I still insist - I am a machine. I cannot see why you believe that I might not be.