Has consciousness been fully explained?

[Dancing David]

Penrose is not an expert in any relevant field.

His theory of microtubules is clownish, there are biochemical signals in the brain.

Whether he is a physicist or not is besides the point.

When Roger Penrose says "well, I worked on trying to produce conscious programs for years, and everything we thought would do it just failed" then we will listen to him. ?

I wouldn't unless they really tried something comparable to a mammalian brain, that is a lot of neurons and lot of connections.

 

[westprog]

Pretty much illustrates the level of purposeful stupidity we have to deal with in here.

If I said "a circle can roll well" people say "nu-uh, a circle can't do anything, a circle is a shape."

No-<rule8> sherlock, I obviously mean "a circular object" can roll well.

Hence, people are seriously disputing what a "turing machine" can and cannot do because a "turing machine" only exists in computer science textbooks. What a joke.

Lets just set the record straight, once and for all, eh?

A "Turing Machine," for the purposes of this discussion, is a system of physical entities that can exhibit a set behaviors called "turing completeness." Nothing more.

Thus any system of physical entities that features turing completeness is a turing machine. It might also be other things. Like a brain, or a calculator.

For one to say "A turing machine can't do what a brain does" is both stupid and obvious at once. It is like saying "a vehicle can't do what a car does."

Yes, and if that were what people were saying, then it would be stupid. However, that's not what people are saying.

The assertion is that the functionality of a conceptual Turing machine is completely sufficient to explain the functionality of the brain. This doesn't mean that the brain can be replaced by an abstract model. It means that any physical implementation of the same Turing machine will achieve the same functionality.

This leads directly to the claim that a Turing machine implementing the "algorithm of the brain" could be implemented, say, with a set of cards placed in cubby holes, or a pattern of stones on a beach, and that this implementation of the brain's functionality would be entirely equivalent. Indeed, PM has claimed that this equivalence is a proven fact.

I am asserting that such an implementation would not duplicate the functionality of the brain, because the brain carries out real-time operations (which may be essential for the creation of consciousness). It's absurd to claim simultaneously that it doesn't matter that the Turing machine definition doesn't allow for real-time processing because any practical implementation would have such a capacity built-in, and at the same time to insist that an entirely non-real-time version of the Turing machine would be entirely equivalent.

It's fairly obvious that somebody whose brain was implemented by a set of cards in cubby holes wouldn't be able to cross the road. It's therefore tendentious to claim that the experience of the set of cards would be the same as the person.

 

[PixyMisa]

His theory of microtubules is clownish, there are biochemical signals in the brain.

I wouldn't unless they really tried something comparable to a mammalian brain, that is a lot of neurons and lot of connections.

They're working on it - e.g. the Blue Brain project. We could probably simulate a whole human brain now - at least, construct a molecular simulation of our best knowledge of all the brain structures.

The reason no-one is even trying to do that is that it would cost as much as the LHC, and by the time you got the hardware up and running, much less the software, the price would have fallen by a factor of ten. So much saner to model the rat neocortex and try something bigger with each new generation of hardware.

 

[westprog]

Yep. The relative number of experts who know anything at all about this issue that side with the computational model is overwhelming.

Nobody should blindly accept argument from authority, but when so many experts can explain their reasoning so clearly and it makes sense to an educated individual, well that is saying something about the strength of the position.

So far this supposed consensus has been unsupported by evidence.

The Fathers of the field had been pretty confusing: John von Neumann speculated about computers and the human brain in analogies sufficiently wild to be worthy of a medieval thinker and Alan M. Turing thought about criteria to settle the question of whether Machines Can Think, a question of which we now know that it is about as relevant as the question of whether Submarines Can Swim.

The people who think that searching for consciousness via the medium of computer programs is futile and unproductive are unlikely to be carrying out such research. Naturally RD will regard their opinion as worthless, no matter what they've achieved elsewhere. It's notable that he regards physics as an irrelevant discipline.

 

[Meed]

That's the argument from ignorance.

I'm not saying that there is no evidence of experiential events being reducible to physical events, therefore they are not reducible. That would be the argument from ignorance. I'm saying that it appears to be a theoretical impossibility. Of course, we all have a certain level of ignorance on mental events given that it is a young area of science with a long way to go.

Except that we have every reason to believe that the mental reduces to the physical, and no reason at all to believe otherwise.

Obviously we disagree. They are causally linked. That fact doesn't provide any means for conveying what an experiential event is through physical description.

Nope.

Can you give a counter-example?

Nope.

Can you give a counter-example?

Nope.

Same question.

How old is the Earth?

Elaborate?

 

[westprog]

Elaborate?

It's the "If you don't agree with me, you must be crazy" argument. Park it with "all the experts agree with me on this, and if they don't, they aren't experts".

It's not the sign of a strong argument, in either case. A good argument wouldn't need recourse to ridicule or appeal to authority.

 

[Meed]

Then you're not only wrong, you're professing to a logically incoherent belief system.

It's either a precise simulation or it isn't.

If it is, the simulation produces all the behaviours of the system being simulated.

If it's not, you contradicted the premise you had just accepted.

If you're asserting that simulated oranges don't contain physical Vitamin C, then you're committing a category error.

So if I simulate an orange with my Turing Machine I can actually eat it and reduce my risk of gout?

 

[Malerin]

I don't agree that the mental always reduces to the physical, for the reason I gave in my response to Piggy. That is, physical description cannot capture the subjective experiential quality of a mental state. For example, an individual who has never experienced pain could study in great detail the neurological processes that correspond to what we call pain, but doing so could not lead him or her to know what it is like to feel pain. This is different than saying the mental is not causally determined by the physical.

This.

 

[westprog]

So if I simulate an orange with my Turing Machine I can actually eat it and reduce my risk of gout?

There's no such thing as a simulation that simulates all the behaviours of the system being simulated. Such a thing would be impossible, in practice and probably in theory as well.

We know that the simulated orange doesn't produce real vitamin C, but if we don't believe that the simulated brain produces real consciousness, then we are guilty of dualism, and acting like walls.

 

[rocketdodger]

The reason that this is relevant is that the brain is clearly, without any doubt, performing real-time operations. Such operations cannot be performed according to the Turing model, and hence a Turing machine cannot perform the function of the brain.

But we have already been over this, and you are wrong, and you don't understand why you are wrong because your understanding of the universe around you is lacking.

Remember? Time reducing to order dependence? The time between two events is merely another way to label the set of events that happen between them? How it is impossible to determine "time" without referring to ordered events?

Did you forget all of that?

 

[rocketdodger]

It's absurd to claim simultaneously that it doesn't matter that the Turing machine definition doesn't allow for real-time processing because any practical implementation would have such a capacity built-in, and at the same time to insist that an entirely non-real-time version of the Turing machine would be entirely equivalent.

But again, "real-time" means nothing mathematically. Time is just the order of events. Period.

 

[rocketdodger]

So far this supposed consensus has been unsupported by evidence.



The people who think that searching for consciousness via the medium of computer programs is futile and unproductive are unlikely to be carrying out such research. Naturally RD will regard their opinion as worthless, no matter what they've achieved elsewhere. It's notable that he regards physics as an irrelevant discipline.

As a matter of fact, I do consider Dijkstra's opinion on the subject to be as irrelevant as I do Penrose's.

Why? Well let me ask you this -- are you familiar with a single program that Dijkstra wrote that is even remotely similar to the things we are talking about?

I am not. That doesn't mean he didn't, but just that I haven't heard of it. The man specialized in low level theory and language development, not the emergent behavior of complex systems.

 

[westprog]

But we have already been over this, and you are wrong, and you don't understand why you are wrong because your understanding of the universe around you is lacking.

Remember? Time reducing to order dependence? The time between two events is merely another way to label the set of events that happen between them? How it is impossible to determine "time" without referring to ordered events?

Did you forget all of that?

Yeah, that was funny.

Now, back to the serious business- it's not possible to carry out the functionality of the brain using only the functionality of a Turing machine.

 

[rocketdodger]

Your delusions of superiority are laughable. If you do possess some in depth knowledge of the subject due to expertise in some area, please share it. Don't divert from real discussion with vague allusions to being more knowledgeable and educated than others.

I have.

I am the only person on these forums to ever speak about what the actual mechanics of the brain might be, in particular how series of perceptron layers fed into boltzmann machine like associative networks are a very likely architecture, and that is supported by evidence from both research on artificial networks of those types and actual brain topography.

I am the only person to painstakingly come up with a "physical theory of computation" for westprog, which despite the peanut gallery's opinions is quite sound and refutes any arguments westprog has had thus far in that context. Everyone else just ignored him, because actually the argument is quite stupid to begin with, but I did the science anyway and got it right.

I am the only person to bother to address the "turing model can't account for time" nonsense that is being ressurected as of late, taking the time to make many posts regarding how time reduces to event order.

And like everyone else who gives a hoot I have tried to patiently explain the fundamentals of both computer science and neuroscience that people refuse to accept, never mind actually understand, in post after post.

Finally, because as far as I know I am the closest thing to an actual A.I. researcher here, I do think I am qualified to at least tell people when they are flat out wrong about what machines can and cannot do, especially in the cases where I myself have written A.I. that can do it.

The one time someone who actually did research on this specific subject -- modeling of portions of mamallian brains using ANNs -- came in here, he got banned shortly after. Go figure.

There really is a genuine difference between the intelligence displayed on one side vs. the other.

Yeah, that was funny.

Now, back to the serious business- it's not possible to carry out the functionality of the brain using only the functionality of a Turing machine.

See what I mean?

Instead of actually addressing the issue, we get a nyah-na-na-na-nyah-na.

 

[AlBell]

I have.

I am the only person on these forums to ever speak about what the actual mechanics of the brain might be, in particular how series of perceptron layers fed into boltzmann machine like associative networks are a very likely architecture, and that is supported by evidence from both research on artificial networks of those types and actual brain topography.

I am the only person to painstakingly come up with a "physical theory of computation" for westprog, which despite the peanut gallery's opinions is quite sound and refutes any arguments westprog has had thus far in that context. Everyone else just ignored him, because actually the argument is quite stupid to begin with, but I did the science anyway and got it right.

I am the only person to bother to address the "turing model can't account for time" nonsense that is being ressurected as of late, taking the time to make many posts regarding how time reduces to event order.

And like everyone else who gives a hoot I have tried to patiently explain the fundamentals of both computer science and neuroscience that people refuse to accept, never mind actually understand, in post after post.

Finally, because as far as I know I am the closest thing to an actual A.I. researcher here, I do think I am qualified to at least tell people when they are flat out wrong about what machines can and cannot do, especially in the cases where I myself have written A.I. that can do it.

The one time someone who actually did research on this specific subject -- modeling of portions of mamallian brains using ANNs -- came in here, he got banned shortly after. Go figure.

There really is a genuine difference between the intelligence displayed on one side vs. the other.



See what I mean?

Instead of actually addressing the issue, we get a nyah-na-na-na-nyah-na.

Now if you could just convince anyone you were right.

 

[Meed]

Finally, because as far as I know I am the closest thing to an actual A.I. researcher here, I do think I am qualified to at least tell people when they are flat out wrong about what machines can and cannot do, especially in the cases where I myself have written A.I. that can do it.

You've written A.I. that can do what (for example)?

There really is a genuine difference between the intelligence displayed on one side vs. the other.

Do intelligent people usually spend a lot of time telling people how intelligent they are?

 

[rocketdodger]

Now if you could just convince anyone you were right.

Whether I am in fact right or not depends only on the strength of the arguments I make.

Whether or not someone is convinced I am right depends, unfortunately, on their intelligence as well.

I can only do so much.

 

[rocketdodger]

You've written A.I. that can do what (for example)?

Be aware of itself and its environment, to start with. Just the fundamentals of an intelligent agent that anyone with any knowledge of A.I. should be familiar with.

Look I am not an A.I. researcher, I wouldn't claim as much. But do you know that people here have seriously questioned notions as simple as self reference?

People like nick227 claim there is no self, so how can self reference exist in code. People like westprog claim that computation isn't different from cheese, so how can self reference exist in machines?

And in the face of clear examples to the contrary, their tune is always the same -- well, that isn't "real" self reference. Or "that just begs the question of what the self is to begin with."

That is how it goes down in here.

Do intelligent people usually spend a lot of time telling people how intelligent they are?

Not relatively, no. What they prefer to do instead is provide clear examples of why their arguments are correct, and then hope for responses that address the specific issues in question.

Why don't you go over my post history and determine the percentage of statements to the effect of how much smarter I am vs. statements where I provide evidence to support my arguments?

Why don't you go over the responses to such posts, and determine the percentage of relevant replies vs. issue dodging and outright omission?

 

[AlBell]

Whether I am in fact right or not depends only on the strength of the arguments I make.

No, it also depends on the axioms you take as fact being true.

Whether or not someone is convinced I am right depends, unfortunately, on their intelligence as well.

Given the caveat above, that may also a factor.

 

[Piggy]

Are those "Sofia events" anything other than stuff getting stored in long term memory?

Absolutely. In fact, they're separate processes, although tangled up the way brain processes tend to be. I remember seeing some research recently studying the differing ways in which consciously experienced events (i.e., whatever the output of Sofia is) are impressed on memory versus events we're never consciously aware of.

We remember all sorts of stuff that never gets routed over to the Sofia process(es).