Has consciousness been fully explained?

[Orbini]

A 'relica' of human consciousness is not a structural replica of the human brain. Consciousness is a phenomena tied in with the activity of our organic brains -- and possibly other biological systems. We need a physical understanding of the phenomena in-and-of-itself, not just the structural plumbing of the nervous system.

It looks like you're saying that the brain replica would be the equivalent of a p-zombie. In other words, we may get a perfect, indistinguishable replica of all the behavior, but there's no real consciousness inside.

Is that what you're saying ?

 

[AkuManiMani]

It looks like you're saying that the brain replica would be the equivalent of a p-zombie. In other words, we may get a perfect, indistinguishable replica of all the behavior, but there's no real consciousness inside.

Is that what you're saying ?

Simply having a brain replica would not even function as a p-zombie. It would essentially be the same as the brain of a cadaver. Consciousness =/= brain structure.

 

[Orbini]

Simply having a brain replica would not even function as a p-zombie. It would essentially be the same as the brain of a cadaver. Consciousness =/= brain structure.

I'm not talking (yet) about consciousness, just functional emulation of the behavior. Are you saying that, no matter how long we study function and structure, we'll never be able to build a working replica ?

 

[AkuManiMani]

Simply having a brain replica would not even function as a p-zombie. It would essentially be the same as the brain of a cadaver. Consciousness =/= brain structure.

I'm not talking (yet) about consciousness, just functional emulation. Are you saying that, no matter how long we study function and structure, we'll never be able to build a working replica ?

A computational simulation of a generator will not produce electricity, and a structural replica of a radio will not pick up stations. We need to grasp the physics of what living brains are doing with regard to consciousness in order to know how to create artificial systems that have the same capabilities.

 

[westprog]

The ECG can only measure the gross external, objective correlates of my experiences. The map is not the territory.


They are the same thing from the inside, when I am that brain.

The feeling might be what is created by the functioning of the neurons, but it isn't the same thing.

 

[westprog]

And while you're at it, please explain why some qualia appear to be reducible to more than one process if they are supposed to be the quanta of mental life.

There's nothing in qualia that requires them to be irreducible. Many feelings are complex.

 

[westprog]

Another straw man analogy. Cargo cult thinking involves imitating the appearance and ritual of the technology without the function. If the cults had copied the airstrips, aircraft, and radios using functional components, things would have been different. We're talking about copying the functional components, structure, and architecture of the brain, not making something that looks like a brain.

That might be what you plan to do - it's not what the computationalists plan to do. They want to make a computer model of the brain, not a functioning copy.

A functioning copy of the brain would be able to control a human body (or equivalent robot) in real time. A computer model would not. This is a vital issue, and should not be blurred.

 

[Ichneumonwasp]

There's nothing in qualia that requires them to be irreducible. Many feelings are complex.

That would take them out of the realm of being like quanta, though, wouldn't you agree?

 

[Ichneumonwasp]

The feeling might be what is created by the functioning of the neurons, but it isn't the same thing.

Why not? If a feeling is created by the functioning of the neurons, why isn't it just the functioning of the neurons?

 

[dlorde]

No, these analogies are metaphors to help convey a conceptual problem that you're consistently side-stepping.

In that case, I guess they are failing to convey this conceptual problem you claim I'm side-stepping. Can you express it clearly without using such metaphors?

Such emulations ARE merely imitating appearance if they do not incorporate the relevant physical processes.

If the physical components we use for our copy have the same functions as the physical components in the brain, and we structure and connect them in the same way, we will get a functioning brain.

Or are you suggesting that there is some physical process involved in the neural networks in the brain that we are as yet unaware of? If so, what makes you think so? if not, what are you talking about?

Nature IS the physics of those neurons and what they are doing. You must understand that physics in order to artificially reproduce the same physical results -- which includes conscious experiences. This is not a difficult concept to grasp; why are you go to such lengths to resist acknowledging the obvious?

It's not obvious to me. I can assemble a working radio by copying the architecture of a working radio using equivalently functional components, without understanding the physics of radios. Why should a biological structure be any different, in principle?

We already have fairly complex artificial neural networks that produce the same outputs as the biological neural networks they were copied from - and there are ongoing projects to emulate whole brains (relatively simple arthropod brains). Do you feel the whole brain projects will fail to perform like natural brains because there's some physical process they're not accounting for?

Or is it that you believe that there is some as yet undiscovered physical process that is unique to consciousness?

 

[Orbini]

A computational simulation of a generator will not produce electricity, and a structural replica of a radio will not pick up stations. We need to grasp the physics of what living brains are doing with regard to consciousness in order to know how to create artificial systems that have the same capabilities.

Obviously it wouldn't be identical, but we're only looking at the externally observable behavior.

We could take a person, remove their brain, and insert the computer running a simulation, carefully attaching it to all the nerve endings, through suitable converters.

In theory, we should be able to create the impression of a perfectly normal human being, capable of conversation, and enjoying a walk on the beach.

Agreed ?

 

[AkuManiMani]

Obviously it wouldn't be identical, but we're only looking at the externally observable behavior.

We could take a person, remove their brain, and insert the computer running a simulation, carefully attaching it to all the nerve endings, through suitable converters.

In theory, we should be able to create the impression of a perfectly normal human being, capable of conversation, and enjoying a walk on the beach.

Agreed ?

I very much doubt that

I do not think a non-conscious system can convincingly generate the same repertoire of behaviors that a conscious individual can. Like I said earlier I think the closest thing possible to what you describe would be a system with some rudimentary level of consciousness and the intellectual capacity to emulate observed behaviors, or a system being puppeted by an individual(s) who is conscious.

 

[dlorde]

That might be what you plan to do - it's not what the computationalists plan to do. They want to make a computer model of the brain, not a functioning copy.

I think accepting the possibility of an artificial brain with functionally equivalent components is a step on the way.

A functioning copy of the brain would be able to control a human body (or equivalent robot) in real time. A computer model would not. This is a vital issue, and should not be blurred.

If one can copy a simple biological neural network and duplicate its activity (i.e. outputs for given inputs) in a computational neural network, and interface both to physical effectors so they can have physical effects, would it not be possible to scale that up, in principle?

Or would you say that it is not possible to hook a computational neural network up to a physical effector?

 

[Orbini]

and a structural replica of a radio will not pick up stations.

By the way, this isn't even true. Some people are working on software defined radio's, that consist of a small RF section + Analog/Digital converter that feeds into a computer, and then the rest of the radio functions are all done in software. You can then listen to the stations on your sound card.

In theory, you only need an antenna and an A/D converter, but with current state of the art technology, a bit of RF filtering is required for good results.

 

[AkuManiMani]

By the way, this isn't even true. Some people are working on software defined radio's, that consist of a small RF section + Analog/Digital converter that feeds into a computer, and then the rest of the radio functions are all done in software. You can then listen to the stations on your sound card.

In theory, you only need an antenna and an A/D converter, but with current state of the art technology, a bit of RF filtering is required for good results.

But we know what materials are needed to pick of the radio signals and undertsand the physics that underlie those materials' capabilities. We cannot say the same yet with regard to brains and consciousness.

 

[Orbini]

I very much doubt that

Excellent! That's good news.

I do not think a non-conscious system can convincingly generate the same repertoire of behaviors that a conscious individual can.

If there's an observable difference in external behavior, we should be able to figure out where the difference is coming from, purely by tracing the signals in the brain, tracing the same signals in the computer, and see where they start to differ.

By using suitably advanced technology, we could follow all individual impulses between neurons until we get from the ear to the mouth, and see exactly how an answer to our question is produced.

Then we do perform the same produce in our computer (much easier), and compare the results. If we find the difference, we fix that part of the software.

The beauty is that all of this can be done objectively, just by comparing function vs function, on a suitable level of detail.

Please forget about consciousness for a moment. Think about emulating an individual brain cell, and then groups of brain cells, until we've covered the whole brain. Where would it go wrong ?

 

[Orbini]

But we know what materials are needed to pick of the radio signals and undertsand the physics that underlie those materials' capabilities. We cannot say the same yet with regard to brains and consciousness.

Obviously, and that's why we need to study the structure of the brain, until we see how it works.

 

[AkuManiMani]

If there's an observable difference in external behavior, we should be able to figure out where the difference is coming from, purely by tracing the signals in the brain, tracing the same signals in the computer, and see where they start to differ.

By using suitably advanced technology, we could follow all individual impulses between neurons until we get from the ear to the mouth, and see exactly how an answer to our question is produced.

Then we do perform the same produce in our computer (much easier), and compare the results. If we find the difference, we fix that part of the software.

The beauty is that all of this can be done objectively, just by comparing function vs function, on a suitable level of detail.

Please forget about consciousness for a moment. Think about emulating an individual brain cell, and then groups of brain cells, until we've covered the whole brain. Where would it go wrong ?

[I was thinking of a scenario somewhat similar to this when I gave the example of a puppet-zombie.]

The end result would be a system that rigidly produces a particular range of outputs according to what was recorded in the particular brain being modeled. This would be the equivalent of recoding an archive of canned responses that is linked to an algorithm the plays the responses when certain cues are given. As long as a given interaction is sufficiently similar to the circumstances in which the original model was recorded (lets say the brain scan of a human having a conversation in French), the automaton may be able to produce passable responses. However, as soon as it encounters situations that diverge from the range of conditions in the original model conscious designers will have to go in and update it accordingly.

Without consciousness, the system itself has no native capacity to create meaningful links between certain external patterns of stimuli and it's responses to them, or to generate meaningful behaviors in the absence of external prompting from entities that are able to make and break semiotic connections. Computation without conscious direction is blind. No matter how sophisticated the computational system in question is it will eventually come up against a metaphorical brick wall that it lacks the awareness, directionality, or ingenuity to intelligently circumvent. It must be directed by innate motivation(s) of some kind and possess some degree of lucidity regarding it's circumstances; in order to accomplish that it must have some subjective dimension to it's functioning. I know you said to try and keep consciousness out of the equation but I do not see how that is possible when trying to produce convincing behaviors >_<

 

[westprog]

That would take them out of the realm of being like quanta, though, wouldn't you agree?

Yes, but that's never been an essential part of the non-theory.

 

[westprog]

Why not? If a feeling is created by the functioning of the neurons, why isn't it just the functioning of the neurons?

Because we don't know what function of the neurons causes the feeling. It's unlikely to be the entirety of what the neurons do.