On Consciousness

[tensordyne]

That rather misses the point, though; it's still the Cartesian Theatre. What we need is to know the details of the process that is the experience.

Tell me your method in general terms on how you wish to do this, if you can.

Where does the rest of it reside? Or are you referring to the peripheral nervous system there, in which case I have no problem?

Yes.

 

[tensordyne]

Write a poll text that you would vote for, that hasn't a problem to you. Heck, let's see you write a set of poll choices on the topic that are free of problems.

Here is my list:

1. Computational Basis
2. Physical Basis
3. Some combination of physical and computational basis
4. Soul basis
5. Hamster on a running wheel basis.

You have to have something like the last item on the list for comedy purposes.

 

[PixyMisa]

Tell me your method in general terms on how you wish to do this, if you can.

Basically, this is computational neuroscience^WP. Examine the function of the brain at all levels, model it in detail, study the behaviours exhibited by the brain and the model.

A sufficiently detailed computer model will exhibit precisely the same behaviours as the brain, and a low-level model will map onto a medium level or high-level model (at least, it will if the respective models are accurate). But the model will offer more opportunity to examine the processes involved, and we can then take that data and refer back to the brain itself to confirm (or falsify) the validity of the models.

We could build a computer powerful enough to run a molecular simulation of the human brain today, but it would take so long to commission and test such a system (let alone co-ordinate all the software), and cost so much, that it's quite impractical. A decade ago it couldn't have been done at all; a decade from now it will be about average for a big science project (a billion dollars or so).

In the meantime, we should continue as we are: Learning everything we can about brain function and cognition, and modelling the brain in detail on a small scale (what Blue Brain is doing), and in less detail on a large scale.

What we shouldn't do, is introduce magic in an attempt to provide easy answers, whether in the mask of quantum mechanics (Penrose & Hameroff) or dualism (Searle, Chalmers).

Yes.

Okay, I certainly agree with that.

 

[tensordyne]

Mind is brain function.
The brain is a computer.
Computation is substrate-neutral.

Therefore, computer programs can be conscious.

None of this is the subject of rational dispute.

It is an assumption that the brain is only, or primarily, a computer.

 

[PixyMisa]

Here is my list:

1. Computational Basis
2. Physical Basis
3. Some combination of physical and computational basis
4. Soul basis
5. Hamster on a running wheel basis.

You have to have something like the last item on the list for comedy purposes.

The problem I have with that, again, is that computation is physical.

You could have:

1. Computational Basis
2. Physical Basis other than computation
3. Soul basis
4. Cupcakes

I think option 2 is still nonsense, but at least it's clear.

 

[PixyMisa]

It is an assumption that the brain is only, or primarily, a computer.

No. It's a conclusion. And it's a very, very, very well-supported conclusion.

Because we know what the brain does. We've poked around in it, and we can see what's going on. And it is clearly and obviously a computer.

Which doesn't prove that we haven't missed something, but there's no evidence in either form or function to indicate that, so there's no reason to give it serious consideration until such evidence is provided.

 

[tensordyne]

Thank you for giving your method.

Basically, this is computational neuroscience[/url]. Examine the function of the brain at all levels, model it in detail, study the behaviours exhibited by the brain and the model.
^WP

One level of the brain is the physics of the brain. I do not see that being addressed.

A sufficiently detailed computer model will exhibit precisely the same behaviours as the brain, and a low-level model will map onto a medium level or high-level model (at least, it will if the respective models are accurate). But the model will offer more opportunity to examine the processes involved, and we can then take that data and refer back to the brain itself to confirm (or falsify) the validity of the models.

There is no a priori reason to think that some physical aspect of the brain (what materials it is made of, how they move, that kind of thing) is not responsible for or directly related to consciousness.

I have no problem with computational neuroscience. I just do not think it will lead to the solution of consciousness by itself.

What we shouldn't do, is introduce magic in an attempt to provide easy answers, whether in the mask of quantum mechanics (Penrose & Hameroff) or dualism (Searle, Chalmers).

Sure, but physical and cognitive neuroscience is not magic, which are the main studies that I guess will lead to understanding consciousness. Psychophysics, not computational psychology.

A sufficiently detailed computer model will exhibit precisely the same behaviours as the brain, and a low-level model will map onto a medium level or high-level model (at least, it will if the respective models are accurate). But the model will offer more opportunity to examine the processes involved, and we can then take that data and refer back to the brain itself to confirm (or falsify) the validity of the models.

There is no a priori reason to think that some physical aspect of the brain (what materials it is made of, how they move, that kind of thing) is not responsible for or directly related to consciousness.

I have no problem with computational neuroscience. I just do not think it will lead to the solution of consciousness by itself.

What we shouldn't do, is introduce magic in an attempt to provide easy answers, whether in the mask of quantum mechanics (Penrose & Hameroff) or dualism (Searle, Chalmers).

Sure, but physical and cognitive neuroscience is not magic, which are the main studies that I guess will lead to understanding consciousness. Psychophysics, not computational psychology." target="_blank">WP

 

[PixyMisa]

Thank you for giving your method.

Note that this is not "my" method. This is mainstream science.

One level of the brain is the physics of the brain. I do not see that being addressed.

If you're modelling at the molecular level, what physics exactly are you missing?

There is no a priori reason to think that some physical aspect of the brain (what materials it is made of, how they move, that kind of thing) is not responsible for or directly related to consciousness.

Yes there is, and it is this: That sentence doesn't actually mean anything.

I have no problem with computational neuroscience. I just do not think it will lead to the solution of consciousness by itself.
[snip]
Sure, but physical and cognitive neuroscience is not magic, which are the main studies that I guess will lead to understanding consciousness.

Those are included in computational neuroscience.

I mangled the link earlier, so here it is again: computational neuroscience^WP.

Psychophysics, not computational psychology.

Psychophysics is one part of computational neuroscience. In itself it cannot produce the full story.

 

[Belz...]

Been away, just saw this thread.

I'm not going to get into another who-shot-John with the "Flock of Seagulls" here, but I do want to point out that the poll presents a false choice between, essentially, the soul model and the data-processing model.

Both are wrong.

Well, that was easy ! I'm convinced, piggy. Thanks !

 

[Belz...]

You ask some of the most interesting questions Belz.... The topic of this thread is consciousness, the topic at that point in the discussion was philosophy of science.

The point of your POST was the philosophy of science, I reckon because you were trying to introduce some concepts into the discussion that would make consciousness something more... magical.

Depending on your philosophy of science you will investigate consciousness in different ways.

I doubt it. The method is the same. Your interpretation may vary.

 

[Belz...]

It certainly should be.

The absence of any reference to "data" and the lack of a claim that "general purpose computers" might be conscious stand out.

Again, I'm not sure it adds to the options in the poll.

 

[Piggy]

I have no problem with computational neuroscience.

Well, you should, since nobody currently studying consciousness in the brain -- which is the only place to study it -- believes that computation is responsible for consciousness, except in the sense that all physical actions are computations.

The computational literalists here are not concerned with the brain. They find it irrelevant.

And truth be told, they're not concerned with consciousness, either. They're concerned with intelligence, but they don't recognize the fact that intelligence and consciousness are not the same thing. (They have simply taken definitions of intelligence and pasted them onto consciousness.)

Yes, computer models of real brains are helpful... in fact, they're necessary to the study of the brain and of consciousness. But this is not at all unique, because computer models are helpful, even necessary, for the study of climate, the cosmos, the functions of the heart, and so forth. That fact does not justify a claim that the heart is a computer. Ditto for the brain.

These guys think they can study machines which are not based on the brain, then come to conclusions which are not verified against observations of the brain, and somehow come to valid conclusions about consciousness.

This is like saying I can come to conclusions about what's going on inside the sun by studying what happens in a light bulb, without bothering to verify my conclusions against observations of the sun.

PixyMisa in particular is prone to outlandish statements, including the claim that it is known that consciousness is caused by self-referential information processing.

In fact, however, if you read standard texts such as The Cognitive Neurosciences, edited by Michael Gazzaniga (one of the many luminaries in the field dismissed as a hack by the computational literalists), you will discover that there is unequivocal and unanimous agreement among scientists who actually study the brain and consciousness that the mechanism for consciousness is not known, and that in fact we do not currently have a theoretical basis upon which to base a working hypothesis.

Real researchers recognize this fact. The flock of seagulls who regularly clog up every thread about consciousness on this board with irrelevant discussions of AI and as-yet nonexistant machine consciousness do not.

Not only that, but you will quickly discover why Pixy's theory is completely useless for doing real work.

First of all, there are all sorts of feedback loops (self-referential processing) going on in the brain, and we know without a doubt that a huge number of them have nothing to do with generating conscious experience.

Furthermore, if you look for "information" in the brain, you will only find the sort of "information" that's present in any other part of your body, the kind Wolfram describes when discussing the computational power of, say, hurricanes.

In many ways, the functioning of the brain resembles sand on a beach more than it does the machine I'm using to type this post.

If you read texts like The Cognitive Neurosciences, you will certainly find people talking about information and data with reference to brain activity -- but this is a rhetorical convenience, a kind of shortcut to keep discussion of the brain from becoming impossibly complex.

And you will find people talking about the computational power of the brain, or computational demands of certain tasks, which is simply a way of notating the number of state-changes involved, but has nothing to do with equating a brain to a computer.

There is one article explicitly on information theory in TCN, and it directly contradicts PixyMisa's claims about conscious machines.

Nothing in that text -- or any other contemporary texts I've seen, or any current popular books by real researchers like Gazzaniga -- implies in any way that brains are computers, or that consciousness is the result of anything but physical computation (the same type of computation your calf muscle does). In fact, the only time that TCN bothers to mention computational literalism is to refute it.

The flock of seagulls claim that running a computer simulation of a brain would make the simulator conscious. Well, there are people attempting to do just that, and they state very clearly that the resulting machine will NOT be conscious.

The flock of seagulls claim that Church-Turing proves that a computer can do anything a human brain can do, but they can't cite anything because that's actually nowhere in the hypothesis, or even remotely implied by it.

Whenever that angle is debunked, they pipe down for a while, but when things die down and they think the coast is clear, they start harping on it again.

They make all sorts of errors in their thinking, such as asserting that the neural activity in your optic nerves resulting from light bouncing off a tree and onto your eyeballs is an "image of a tree" in your brain.

This makes as much sense as saying that the waves caused by tossing a rock into a pond are an "image of a rock".

The question of machine consciousness is simply this: How is it possible to make a machine hallucinate?

Because all consciousness, waking or dreaming, is hallucination. Nothing in your conscious world corresponds to anything that exists outside that hallucination -- not color, not sound, not texture, not pain or any other bodily sensation, not smell, not taste, nothing in your conscious experience has any existence outside of the hallucination (yours or someone else's).

It is in no way an image of the outside world. It is a hallucination which bears no resemblance to reality, but which nevertheless allows your body to successfully interact with the world when it's functioning properly.

(If you don't believe me, I have a little thought experiment we can do.)

In short, there is no computational theory of consciousness. It does not exist. And even the computational model of the mind is recognized as simply a useful metaphor.

Those who take it literally are simply not up on the current brain research, are sloppy in their thinking about consciousness and the brain (their posts are rife with anthropomorphisms), confuse intelligence with consciousness, entify abstractions such as information and relationships, ignore the importance of semantics (it's all syntax for them), don't bother with evidence, and ignore the fact that their hypotheses lead to absurd conclusions such as the idea that a conscious brain could be made out of rope or that the actual physical configuration of the brain has no effect on consciousness.

It's a bankrupt idea, and all of their carrying on about AI and conscious robots is 100% totally irrelevant to the actual discussion of how the brain does consciousness.

In short, you should have some very serious problems with so-called computational neuroscience when it comes to understanding consciousness.

 

[PixyMisa]

Well, you should, since nobody currently studying consciousness in the brain -- which is the only place to study it

No.

 

[Piggy]

No.

And that's all you got.

Not one single cite.

Not one single researcher studying consciousness in the brain who agrees with you.

You assert over and over again that your ideas are in line with mainstream brain science, and when that's shown not to be true, you respond with junk like that.

I asked you for a cite, and you cited yourself, asserting that "the brain is a computer" without one shred of evidence -- which you've never had for all your hundreds of posts -- despite the fact that nobody working in the field today believes that.

So you're free to respond with "No" and "Wrong" and "Framing error" all you like. It's all you got.

Which is fine. Good luck with that. May it serve you well.

 

[PixyMisa]

And that's all you got.

That's all I need, Piggy. Your position is nonsensical and your assertions untrue.

I have far more. I have all of modern science. The only problem is, all the facts disagree with you. As does basic logic.

 

[Piggy]

That's all I need, Piggy. Your position is nonsensical and your assertions untrue.

Well, if that's all you need, OK.

Anyone can pick up TCN or any other standard text and verify that what I'm saying is accurate.

But if all you need is "no" and "wrong" to keep your worldview in tact, more power to you.

 

[Piggy]

I have far more. I have all of modern science. The only problem is, all the facts disagree with you. As does basic logic.

Assertions, assertions.

In other threads, I've shown that science and logic disagree with you, and you have yet to offer any counterargument except "no" or telling me to read Godel-Escher-Bach.

So I'm done responding to you for now until you actually pony up. Which I don't expect will ever happen.

 

[PixyMisa]

Well, if that's all you need, OK.

Anyone can pick up TCN or any other standard text and verify that what I'm saying is accurate.

Even if what you are saying about TCN is accurate, that merely means that it's wrong. Brains are computers, Piggy. It's what they do. Anyone who asserts otherwise is wrong. Either they don't understand what the brain does, or they don't understand what a computer is, or both.

Here, for example, is an article pointing out 10 important differences between brains and computers. Anyone who actually knows the definition of computer will run out of hands with which to facepalm before they reach the end, because none of those so-called differences are actually differences.

If the article had been titled "10 differences between the brain and a low-end desktop PC that don't affect the computational model one iota", then it would be fine. Just not terribly relevant.

 

[PixyMisa]

In other threads, I've shown that science and logic disagree with you

Not even remotely. You can't even define your disagreement with me, much less show any evidence for it.

 

[PixyMisa]

Anyway, it has been shown that it is impossible - logically impossible - that computers can't become conscious, and that the only question is, what is required to achieve this. My position is that a very high-level approach is perfectly sufficient, and no-one has presented any evidence to the contrary.

Even if I'm wrong in that respect (implausible, but let's consider it), all it means is that a lower-level, more detailed approach is required. But it if you accept that brains produce conscious function, then there is nothing that can restrict this behaviour - and it is a behaviour - from being duplicated by computers.