h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Jordan:

"if consciousness is a holistic function of an entity or device, then
what is the cutoff point? In other words, at what level of complexity/sophistication is does consciousness become manifest?"

As you say, consciousness could be (and I think is) continuous. A cat or dog is less conscious than us, and so on down to insects and ameboae. I don't think this just shifts the question to 'at what point does something stop being conscious', or rather I don't think that that question really makes sense. I don't think consciousness is a fundamental property of the world, unlike mass or charge - its a high level concept and hence it requires interpretation to say if its there and to what degree. We can say 'that thing is definitely conscious' or not, but there will be fuzzy areas where we don't know and the word itself is not defined precisely enough to be sure - unless of course you can give me a definition of 'conscious' that would allow us to check

"But is my consciousness in the book? or is it in the brain (albeit separated) of
the person doing the algorithm?"

The consciousness is in the entire process. Just as in your brain, the consciousness is not in your neurons but in the process of their interactions. It doesnt feel like anything is conscious when flipping pages of the book, but then the time scale is so much longer than we are happy of thinking about - of course it would seem weird. Searle's Chinese room got pretty well attacked by others on h2g2 when an article went into the guide on it, I can't think of anything I'd want to add to what was already said on it right now.

What do you think of the Turing test as a test for consciousness? (or at least intelligence?)

Hi!

I've got to go home now, so I can't reply just yet ... I'll reply tomorrow. Masterstroke, moving the conversation to another thread... it was getting well off topic!

Thanks,

Jordan

Hi!


Sorry I took so long - I was a bit preoccupied with things! Too much to do... but I suppose I have less to do than most people .

I too think that animals possess consciousness, but to a lesser degree than humans. As to a definition of consciousness, I would say that the best, most operational, is that consciousness is what collapses the wavefunction of a particle... but I don't understand the arguments fully (yet!), so I can't really comment on it with any authority. You are obviously more advanced mathematically than I am, perhaps you ought to look into it? It's generally called the many-minds interpretation.

I suppose that, given this, you are justified in saying that consciousness is in the proccess, and not the elements or their arrangements... but if this is so, doesn't it seem to be tugging at the strings of Platonic Reality - and thus a form of dualism?

The Turing test... well, frankly I think it's a very clever but potentially flawed concept. Despite the obvious fact that communication (verbal or non-verbl) is the only tool that we have to identify the state of mind of another person, I think that the methods that are used to pass it (wiht increasing success) are becomeing successful to the degree where even a relatively 'stupid machine should be able to pass it.

Imagine that we have a machine that simulates 'Aunt Bubbles'. Now, imagine that this machine has a phenomenally huge memory, If this is so, we could concievably give it a response to every possible string of characters that could be typed in, say, an hour or so. The memory, strictly speaking, would not need to be infinite, but it would still be huge. We could economise considerably by grouping inputs that shared a common pattern (eg.allowing it to parse mathematical expressions that are not too complex, reserving responses for these such as "I can't answer that, ask my genius nephew!" ... "...Or HenryS!"). If we were able to do this to a significant degree, we might have a program (still huge, but perhaps not to large for a massively parallel supercomputer to run) that would almost definately pass the Turing test. Also, even if the computer is indeed conscious, we might not be able to tell - for instance, if we were to take a monkey, say, and train it to use a simple communication board, then (after observation) write a computer program to simulate this behaviour, it does not seem inconcievable that the computer might come out trumphs.

The point of this is simply that: (a) the computer might pass by virtue of clever programming, and not intelligence; (b) even if it were conscious, it might be to a degree that is so small (or, indeed, so much higher than our own) that we would might not credit it with such.

Frankly, I think that observation in its own right is more valuable than a 'game' such as the Turing test. It is a very clever idea, but it does not actually answer the question in every case, and I think that asimple 'test of understanding' might provide a better measure of its consciousness than even the best judge could give.

Anyway, I'm quite happy with the thought of having a computer that is not conscious yet can score 100 on a normal IQ test, and indeed I think it is possible. In other words, I would not be swayed by apparent intelligence as an indicator of consciousness (they can already beat grandmasters at chess!), and I also wonder if the difference between intelligence and genius is not algorithmic (someone who can add up quickly may not have the same insight into another problem than a slower, more original thinker). But please tell me what you think - I probably need some balance to my opinions, which have been raging out of control all on their own!


Yours Truly,


Jordan