The hard problem of consciousness

[Jimbo07]

At the moment no one can because we don't know what these phenomena are. It's that simple. We are confident (and for good reason) that they are emergent properties of our brain but we don't have a working theory (scientific theory) of them yet. We will.

I am worried that they are not, in fact, things, but are artifacts of language.

In the 30s, these were unknown, but regarding learning in the AI field:
- Chess playing is insufficient to define consciousness and support the existence of soul
- Self-reference seems to be insufficient to define consciousness and support the existence of soul
- Will experience be the next knocked down?

If, in fact, they are artifacts of language, my worry is that some people will never be satisfied with any level of detail about brain function. They will always be able to say, "You haven't disproved consciousness, nyah, nyah, nyah..."

 

[JoeEllison]

Nobody has given any plausible explanation.(not that I'm really expecting one)

So, there's no reason to ever discuss this with you again, is there? You have already decided that no explanation is good enough to explain something that has a perfectly rational explanation unless you choose to define it in a way that rejects all explanation. As Jimbo07, this "problem" seems to be a linguistic problem, and not a problem of brain function or even "consciousness", whatever that is.

 

[Ivor the Engineer]

Yes, and prior to understanding aerodynamics scientists could have simply defined the behaviors of flight.

<snip>

Isn't that what they did? Mathematical equations describe the behaviour of the various parts of the system and how they interact.

<snip>

What is emotion?

Emotions are defined by the behaviour (or modification of behaviour) associated with them. When I use the word 'behaviour', I'm including internal behaviour (i.e. thinking) as well as external behaviour. When someone is sad they think differently than when they are happy. E.g., a glass can go from half empty to half full as their mood improves.

What most computer systems lack are interfaces which allow them interact in a meaningful way with humans. Even if they do, the amount of computational resources available and the knowledge of how to write a program that complex, are currently insufficient to create conscious machines.

To separate consciousness from the behaviour of something considered conscious appears to me to be assuming consciousness is separable and distinct from behaviour.

 

[RandFan]

I am worried that they are not, in fact, things, but are artifacts of language.

I believe I understand your point. In an attempt to communicate to each other our experiences we have created words to do so and these words can give the appearance that something is real when in fact that "something" is nothing more than an abstract concept. While there is a word for unicorn the fact there is a word doesn't mean that there really are unicorns.

That I experience pain is real. You can dismiss it as an artifact of language but if the pain is bad enough it can lead a person to take his or her own life. I could be wrong but It would seem that such phenomena are more than mere artifacts of language. I think you would have a difficult time convincing a person who is suffering from chronic pain that his experience is simply an artifact of language.

In the 30s, these were unknown, but regarding learning in the AI field:
- Chess playing is insufficient to define consciousness and support the existence of soul
- Self-reference seems to be insufficient to define consciousness and support the existence of soul
- Will experience be the next knocked down?

If, in fact, they are artifacts of language, my worry is that some people will never be satisfied with any level of detail about brain function. They will always be able to say, "You haven't disproved consciousness, nyah, nyah, nyah..."

I really don't care. I find this entirely beside the point. I think we will understand consciousness and all of the wide range of experience and emotions. But even if we don't it is irrelevant whether people are satisfied or not.

 

[Ivor the Engineer]

Consciousness is a computer which fights you when it estimates you're trying to turn it off.

 

[rocketdodger]

If you know how objective reality leads to subjective experience then let the world in on it and pick up your Nobel prize. I assure you it is waiting for anyone who can adequately explain it and I suspect that the folks who are currently reverse engineering the brain may very well be the ones to take it home.

I have a very good idea. First, understand these concepts:

http://en.wikipedia.org/wiki/Inference_rule

http://en.wikipedia.org/wiki/Forward_chaining

http://en.wikipedia.org/wiki/Backward_chaining


Second, realize that neurons provide a perfect substrate for a physical representation of inference rules.

Imagine a whole bunch of neurons in your brain. Imagine you see an apple. Imagine that certain neurons are excited as a result. Since neurons that "fire together wire together," you end up with a physical data structure representing an apple.

Now imagine that you see a fire engine. Imagine that certain neurons are excited as a result. Most will be different. But a few -- namely, the ones in your visual cortex responsible for discriminating color, or whatever -- will be the same. But, those neurons are also part of the apple data structure, and thus can implicitly realize the inference rule "apple->red" in reverse.

Thus in my model our brains are just massive assemblies of these data structures whose interlinkage is a physical representation of the rules of inference we derive from our experiences. In other words, they perform logical chaining naturally -- indeed, automatically -- and that is what we call "thinking."

Now, I might be wrong and this isn't the way our minds operate. But I guarantee you that I could make a conscious mind exactly according to this model, given the proper tools. And, more importantly, the numerous features of consciousness encapsulated by the hard problem all follow in a natural and easily explainable way from this model.

Why haven't I, or any of the other people who share these ideas, won a nobel prize? Probably because we are just speculating. But it is damn good speculation -- all available evidence agrees with it so far and it doesn't rely on any unconfirmed mechanisms.

 

[RandFan]

Isn't that what they did? Mathematical equations describe the behaviour of the various parts of the system and how they interact.

No. I wouldn't say that at all. Perhaps we are just arguing semantics but I don't think so. Prior to understanding aerodynamics humans couldn't build flying machines. We could have reverse engineered birds and eventually built flying machines but understanding is more than simply describing the behavior.

Emotions are defined by the behaviour (or modification of behaviour) associated with them. When I use the word 'behaviour', I'm including internal behaviour (i.e. thinking) as well as external behaviour. When someone is sad they think differently than when they are happy. E.g., a glass can go from half empty to half full as their mood improves.

This tells us nothing. I can describe the behavior of gravity without understanding it.

What most computer systems lack are interfaces which allow them interact in a meaningful way with humans. Even if they do, the amount of computational resources available and the knowledge of how to write a program that complex, are currently insufficient to create conscious machines.

We are getting there.

To separate consciousness from the behaviour of something considered conscious appears to me to be assuming consciousness is separable and distinct from behaviour.

I make no such assumptions. I'm simply saying that we don't know what phenomena like emotion is at the moment. That we don't know how it emerges from brain processes. We only know that it does (or we are very confident that it does and that there is zero evidence and no reason to suppose that it doesn't).

 

[Jimbo07]

I believe I understand your point. In an attempt to communicate to each other our experiences we have created words to do so and these words can give the appearance that something is real when in fact that "something" is nothing more than an abstract concept. While there is a word for unicorn the fact there is a word doesn't mean that there really are unicorns.

Not quite, but this is a good point on its own. Wizards exist. They have an existence to which we've attached the label of fiction, as opposed to real, but they exist all the same. This is a definite artifact of language, and should probably be expected where abstraction is possible (I'm not confident about this last point).

That I experience pain is real. You can dismiss it as an artifact of language but if the pain is bad enough it can lead a person to take his or her own life. I could be wrong but It would seem that such phenomena are more than mere artifacts of language.

but, according to my post above, dogs have the appearance of experiencing pain. My point about the artifact of language lies here. You said earlier:

I can build a machine that can detect the difference between red and blue but I can't build one that can experience the colors.

I believe the valid question to be, "Is there any genuine difference between the words, experience, phenomena and detection?" How could we know?

I think you would have a difficult time convincing a person who is suffering from chronic pain that his experience is simply an artifact of language.

The pain, as a nervous reaction, exists. Is it an experience, a detection, or a phenomenon? Why do you give special meaning to the word detection over experience?

 

[RandFan]

Thus in my model our brains are just massive assemblies of these data structures whose interlinkage is a physical representation of the rules of inference we derive from our experiences. In other words, they perform logical chaining naturally -- indeed, automatically -- and that is what we call "thinking."

But it is not what we call emotion or pain. Experiencing blue or hearing Mozart. Experiences of emotion or sensation might emerge from these "interlinkages" but we don't know that and we don't have any reason to suppose that they do other than we don't have any other theory.

 

[RandFan]

The pain, as a nervous reaction, exists. Is it an experience, a detection, or a phenomenon? Why do you give special meaning to the word detection over experience?

Yes, that is the crux of the discussion. Dennett first said that a thermostat is aware. He then said that it is not aware that it is aware. Does the thermostat become uncomfortable as the temperature rises? Why or why not? What is discomfort? We understand what it is from an evolutionary POV. Pain and discomfort are very powerful motivators. An entity capable of such is more likely to avoid falling into a fire than ones that are not. Is there a difference between detection and experience? I would have to say yes. Experience is subjective. A person who is paralyzed could have electrodes implanted in his or her skin to detect changes in temperature but still not be able to experience the sensation of heat and cold. He or she could also theoretically have electrodes implanted to stimulate the sensation of heat and cold.

 

[Ivor the Engineer]

No. I wouldn't say that at all. Perhaps we are just arguing semantics but I don't think so. Prior to understanding aerodynamics humans couldn't build flying machines. We could have reverse engineered birds and eventually built flying machines but understanding is more than simply describing the behavior.

It comes down to the level of understanding of behaviour. Scientists didn't (and still don't) need to understand quantum physics to build flying machines. They observe the behaviour of (particles of) air flowing over wings of various shapes and the forces produced.

This tells us nothing. I can describe the behavior of gravity without understanding it.

But all understanding is ultimately limited to observation of behaviour. If you believe consciousness cannot be understood sufficiently from ever more detailed observations of behaviour, then you will never consider it to be understood sufficiently!

I make no such assumptions. I'm simply saying that we don't know what phenomena like emotion is at the moment. That we don't know how it emerges from brain processes. We only know that it does (or we are very confident that it does and that there is zero evidence and no reason to suppose that it doesn't).

I still don't see how you can logically separate an emotion from the brain processes which generate it. I.e. "Happy" is a behaviour, a brain state, not a separate entity.

 

[Jimbo07]

Is there a difference between detection and experience? I would have to say yes. Experience is subjective.

Apart from me wondering if it really is as subjective as people like to believe... if I give you that "detection" is different from "experience," is then "detection with subsequent review and abstraction" different from "experience"?

A person who is paralyzed could have electrodes implanted in his or her skin to detect changes in temperature but still not be able to experience the sensation of heat and cold.

People like analogies, so I would say that that is a sensor failure. A proper patch would have to go from the skin to the same signal processing. Otherwise, it is an alternate detection system, not a repaired one.

He or she could also theoretically have electrodes implanted to stimulate the sensation of heat and cold.

Yes, although this would be like signal testing, and might speak more strongly to the similarity between detection and experience!

 

[rocketdodger]

But it is not what we call emotion or pain. Experiencing blue or hearing Mozart.

How do you know? First, define exactly what those experiences are. Then you can make a claim like this. I suspect, however, that when you actually define these experiences you will suddenly see that they can quite easily be modeled by an inference based reasoning system.

Experiences of emotion or sensation might emerge from these "interlinkages" but we don't know that

Agreed. I am pretty sure of it, though. At least, nothing in my knowledge base leads me to conclude the opposite.

Furthermore it doesn't even matter -- you can't be sure of this regardless. Everyone you interact with could be a p-zombie or walking chinese room. The best you can do is take their word for it if an entity claims to be self conscious.

and we don't have any reason to suppose that they do other than we don't have any other theory.

No. We have reason to suppose that they do because it makes sense.

 

[Beerina]

* "Why should physical processing give rise to a rich inner life at all?"

Why not? The question is a non-starter. The word "should" isn't likely to lead to any understanding regardless of whether or not reality is monist, materialistic or whether or not there is some god or a homunculus (soul/mind/whatever).

It's more along the lines of why does anything exist at all, or why is the universe this way instead of that.

Looking at the question from the other end, i.e. the physics to support it is there, and evolution latched onto it and used it, it is no longer a mystery.

The only question is why such a facility exists to begin with, and that's like asking why is an atom the way it is.

For this case, we could only conclude that non-sentient intelligence could (probably) have evolved, but that the subjective perceptual experience is useful enough it's been used and integrated into it. Personally I doubt they're the same thing; consciousness seems to be more of a motivator than a thinker in and of itself.

 

[skiba]

So, there's no reason to ever discuss this with you again, is there?

The point is to discuss. If we disagree, we can always explain why. Thats how we learn something from each other.

You have already decided that no explanation is good enough to explain something that has a perfectly rational explanation unless you choose to define it in a way that rejects all explanation. As Jimbo07, this "problem" seems to be a linguistic problem, and not a problem of brain function or even "consciousness", whatever that is.

No need to twist my words. I said I wasn't expecting one bc it's an age old question. I have explained myself why I disagree with the prevailing theorys, or should I say materialistic interpretation of consciousness. Hope you do the same or just ask elaborate on any part, and will try to do so with my somewhat limited english.

 

[idunno]

My thoughts exactly.

Computation and transfering information based on the physical properties of the media, explains nothing about consciousness.



I really don't see how you could explain that experience is not real?
This is one of the arguments Dan Dennett suggests. "consciousness is just a trick of the mind". Well, again something is experiening that trick. This is a form of circular logic. It's a way to scrub the HPC under the rug.


@RandFan

Totally agree

is it possible that its just the body that experiences the trick?

 

[Roboramma]

It goes deeper than that: You lack an explanation of why there needs to exist such a mechanism, as a discrete thing. Consciousness may just be an emergent property of our already known mechanisms, provided sufficient complexity.

Well, I think it necessarily is an emergent property of those mechanisms. But that doesn't mean we understand why that's the case.

No one here (as far as I know) thinks a rock experiences something when you throw it against another rock. That water experiences something when it falls as rain. But do dogs experience something when they eat their dinner? I think almost certainly, yes.
I just don't see how that experience is a necessary byproduct of the information processing that goes in to being a dog. There's no way in which, if we were computers that didn't have 'experience' we would predict it's existence based on what we currently know.

So, I think it's a problem. I also think it's one with a solution, but we don't have that yet.

I don't think the technology must necessarily be different, but the architecture must; present computers are specifically designed to be deterministic. As long as you have an intrinsically deterministic architecture, there is no particular reason to expect that it well become more conscious with increased complexity.

What has determinism got to do with consciousness? And how will we know if and when computers do become conscious?

 

[Jimbo07]

What has determinism got to do with consciousness? And how will we know if and when computers do become conscious?

If I'm correct in my worries about the language problem, some people will still be navel-gazing about these questions when the robots are violently demanding their rights!

 

[Ivor the Engineer]

Yes, that is the crux of the discussion. Dennett first said that a thermostat is aware. He then said that it is not aware that it is aware. Does the thermostat become uncomfortable as the temperature rises? Why or why not? What is discomfort? We understand what it is from an evolutionary POV. Pain and discomfort are very powerful motivators. An entity capable of such is more likely to avoid falling into a fire than ones that are not. Is there a difference between detection and experience? I would have to say yes. Experience is subjective. A person who is paralyzed could have electrodes implanted in his or her skin to detect changes in temperature but still not be able to experience the sensation of heat and cold. He or she could also theoretically have electrodes implanted to stimulate the sensation of heat and cold.

Check out phantom limbs:

http://en.wikipedia.org/wiki/Phantom_limb

 

[JoeEllison]

The point is to discuss. If we disagree, we can always explain why. Thats how we learn something from each other.



No need to twist my words. I said I wasn't expecting one bc it's an age old question. I have explained myself why I disagree with the prevailing theorys, or should I say materialistic interpretation of consciousness. Hope you do the same or just ask elaborate on any part, and will try to do so with my somewhat limited english.

Ahhh, yes... the limited English. Gets them every time.

I don't feel as though I've twisted your words at all. The problem seems to be that to treat the "HPC" as being a legitimate problem means to accept a whole lot of assumptions, some of them having to do with the words we use, and the slightly different ways that each of us define and understand those words.