Robot consciousness

[Piggy]

Proof?

You are giving some special quality to neurons that doesn't exist (to our knowledge).

If I built your exact brain out of computer chips, with exactly the same speed, is that not a brain that would generate consciousness?

If I built your exact brain out of metal cogs, with exactly the same speed, is that not a brain that would generate consciousness?

If I split that metal cog brain in half, separated by 1 mile, and connected the separated cogs with 1 mile long axles, is that not a brain that would generate consciousness?

If I remove those 1 mile long axles, and replace them with a computer that reads what one set of cogs do, and then transfers that data to the separated cogs via a network and some servos, is that not a brain that would generate consciousness?

If I replace that computer, network, and servos with somebody doing the calculation by pen/paper, and running back and forth to turn the cogs by hand, is that not a brain that would generate consciousness?

At that point, no, I don't see that consciousness would be possible.

And btw, I'm not giving any special properties to neurons. Quite the opposite. Neuronal activity is invisible to consciousness.

As you say, it doesn't matter that it's neurons doing what's being done. If the signals were carried some other way, the result would be the same, as long as it operated with the same thresholds and connections, and as long as the modules were arranged in the same configuration.

(Whether such an object could actually be made out of computer chips or cogs, I couldn't say. I seriously doubt it could be made with cogs, which are not very amenable to threshold-dependent response.)

But, as with your heartbeat, there must be a limit to how much the signaling can slow down before the higher-level phenomenon ceases.

I don't think that you or I or anyone alive is in a position to say where that point is. But I have no doubt that it's somewhere higher than a person running a mile between each synapse.

As steenkh noted above, consciousness is a rather fuzzy thing, smeared out over time. We're not aware of events as they happen, often not even in order, and by the time we are conscious of events, what we're aware of is not the raw input but something very highly processed, all filtered and chunked and lumped.

We act on perceptions that we're not conscious of. Most of what the brain does is independent of consciousness. Consciousness appears to be an after-effect, a downstream function.

 

[Piggy]

Okay, great!

Evidence?

How about logic?

What's happening when we write programs out on pen and paper?

Two things: First, a stylus is moving across a sheet of paper; second, we're doing calculations in our head to get to the next step.

We don't know exactly what creates consciousness, but we can be confident that dragging a stylus across a sheet doesn't.

Neither does doing math in my head. Yes, I'm conscious when I'm doing that, but it doesn't generate a new, discreet conscious entity.

So, since transcribing the Star Spangled Banner doesn't somehow generate consciousness from thin air, it doesn't matter if I change how I move my pen so that it writes equations instead.

Sitting and scribbling doesn't make consciousness, period.

So if we have our robot there, and at some point we do our dump-redump (take everything that's going on, route it out to a staff with pens and pads, have them do the next step(s), then dump the output back into the system at the appropriate points), we can be confident of two things:

1. The robot was not conscious during this pause. Why not? Because their was no activity in the robot brain.

2. The robot's consciousness was not transfered outside of his brain to the staff. Why not? Because dragging pens across paper is not what generates consciousness, no matter what those pens are writing.

 

[Piggy]

Why is it like that? What part of a TM-equivalent algorithmic process can I not emulate with paper and pencil?

This whole paper and pencil thing gets away from my original question, which is whether there is a speed at which the robot's brain would no longer be conscious. We agree that if the brain is TM-equivalent then there is not. If the brain is not TM-equivalent, then might there be aspects of it that depend on real time? For example, is it possible that neural oscillation times are not arbitrary, but related to some real-time "clocks"? Perhaps the neural oscillation times are related to the speed at which typical natural events occur.

Actually, that must be the case, but is rather trivially obvious.

By "TM-equivalent" do you mean it can pass a Turing test?

If so, then I'm not with you there.

Passing a Turing test doesn't mean that the machine that passed it is conscious. So it's irrelevant.

I'll stick with the original proposal, that we stipulate that we have a robot which is (regardless of how we know this) conscious like humans are.

If the question is "How slow can the processing go and still keep the robot conscious?", to be honest, I don't see the point of the question.

Why not simply ask that question about our brains?

But anyway, it seems clear to me that we don't have the knowledge or tools to answer that question. What in the world would you base any answer on?

But let's take your questions there:

"Is [there] a speed at which the robot's brain would no longer be conscious?"

I think it's safe to assume that there is a floor.

Obviously, in the human brain, if you had a single neural event per second in the areas where conscious awareness is handled, the effect would not be realized.

If anyone wants to propose an entirely different kind of robot brain that could produce consciousness at that speed, they'd have to explain in detail how it would work.

"Is it possible that neural oscillation times are not arbitrary, but related to some real-time 'clocks'?"

Possible, I suppose.

 

[Paul C. Anagnostopoulos]

Pedantic, but I think this is a different meaning of the word 'conscious'.

Really?

Oh, it's not that I believe in any mystical wotnot that is specific to consciousness.... it's just that consciousness has properties that are absolutely different to anything else in science. It's unique. I agree that it's pragmatic to assume that consciousness is caused by the same substances and processes that we know and to a fair degree understand, but that doesn't make it so. If it does, then those substances and processes have some very surprising properties.

Gravity is unique, too.

Really, who knows what consciousness is? All we're doing here is making a bunch of stated, and unstated, assumptions about it and then reasoning on that basis. We're never going to know if those assumptions are true. All of this is unfalsifiable.

You could falsify the claim that consciousness is a function of the brain by finding something else that is conscious.

I could perhaps suggest a way to falsify your hypothesis, but I don't know what it is.

~~ Paul

 

[Paul C. Anagnostopoulos]

We don't know exactly what creates consciousness, but we can be confident that dragging a stylus across a sheet doesn't.

Are you sure? What if the dragging is done is a certain way?

You appear to be treating consciousness as if it is a thing, rather than a process.

By "TM-equivalent" do you mean it can pass a Turing test?

No, I mean equivalent to a Turing machine.

Passing a Turing test doesn't mean that the machine that passed it is conscious. So it's irrelevant.

It better mean that, because you've only got two ways to tell if a machine is conscious: (a) witness its behavior; (b) analyze its structure.

If the question is "How slow can the processing go and still keep the robot conscious?", to be honest, I don't see the point of the question.

Why not simply ask that question about our brains?

Okay. I asked about robot consciousness because that's what came up on another forum.

~~ Paul

 

[Piggy]

Piggy,

I'm really curious to see what you come up with.

Can you really show this? I have always thought these questions came down to faith. I await your next post with baited keyboard.

I think that would be "with bated keyboard" -- but then again, I don't know your hobbies.

What I meant by that is that it's not neurological activity per se that creates consciousness. Rather, it's what happens at a higher level of organization.

We can look at two examples, one of a specific conscious function, and another of general conscious function.

I mentioned Melvin before, who lost conscious awareness of his emotional states after a stroke wiped out the neural bridge that feeds information to the part of the brain that handles awareness of emotions specifically.

His condition didn't arise because he lost a certain percentage of neurons, or a particular type of neurons. It came about because a connection was broken between brain modules.

If our brains were made of something other than neurons, and yet information still flowed among these larger structures in the same way, his stroke would have had the same effect.

There's also the case of Terry Schiavo, whose parents insisted she was conscious, but whose husband insisted she was not.

Despite the fact that she in some ways appeared to be aware but to have very limited functionality, there was a medical certainty that she was not in fact aware of anything, because of the macro-structures in her brain that had been destroyed.

What makes our brains work the way they do -- what makes them capable of posting on this forum, for example -- is the sequencing of the neurons and how they are grouped into larger structures which communicate among each other.

Rewire it in a way that allows the neurons to fire all happy, but that routes the information in a radically different pattern, and the brain will not do what it does.

There are simpler analogs in other organs. Take the heart, for instance. The muscular contractions follow a sequence that makes our hearts beat so that they pump blood. But there are other patterns that the contractions can follow along that network.

When there's a shock to the heart, it can "knock" the heart out of the normal pattern and into a different one. From the point of view of the muscle cells, everything's fine. They're doing what they have always done. But from the point of view of the organ, nothing's fine, because the new pattern of contraction doesn't make the heart beat so that it pumps blood.

That's why we use defibrillators -- to shock the heart again in an attempt to "knock" it back into the old pattern.

Similarly, it's not the flow of impulses from neuron to neuron that makes the brain work. Rather, it's a specific pattern of flow from neuron to neuron, and by extension from module to module, which creates higher-level phenomena that result in our being able to speak, read, type, even cough and breathe.

 

[Piggy]

Are you sure? What if the dragging is done is a certain way?

You appear to be treating consciousness as if it is a thing, rather than a process.

No, of course I'm not treating it as a thing.

We know of one object that produces consciousness. We know of other similar objects we can be pretty confident about.

None of them bear any resemblance to dragging a pen across paper.

It's an indefensible leap. It would be like saying that a falling leaf makes volcanoes erupt.

 

[Piggy]

It better mean that, because you've only got two ways to tell if a machine is conscious: (a) witness its behavior; (b) analyze its structure.

Thisis what I meant by the drunk looking for his keys under the streetlight because the light is better.

You can't simply corral the Turing test because you got nothing else.

For the thought experiment to be of any value, we just have to stipulate a conscious robot.

There's simply no reason to believe that the Turing test is capable of discerning consciousness.

 

[Paul C. Anagnostopoulos]

No, of course I'm not treating it as a thing.

We know of one object that produces consciousness. We know of other similar objects we can be pretty confident about.

None of them bear any resemblance to dragging a pen across paper.

Lack of resemblance is irrelevant.

Okay, so do you agree that the brain must not be equivalent to a Turing machine? If so, what sort of additional mechanisms are there that preclude the process of consciousness being generated by a Turing machine-equivalent computer such as pencil and paper?

You can't simply corral the Turing test because you got nothing else.

For the thought experiment to be of any value, we just have to stipulate a conscious robot.

But if it's not conscious in some manner similar to humans, then the thought experiment is also useless because we can't even begin to consider the question of whether processing speed matters. And one of the two ways we know other humans are conscious is by using some variant of a Turing test: I interact with you and you appear to be conscious.

~~ Paul

 

[roger]

I think that would be "with bated keyboard" -- but then again, I don't know your hobbies.

What I meant by that is that it's not neurological activity per se that creates consciousness. Rather, it's what happens at a higher level of organization.

We can look at two examples, one of a specific conscious function, and another of general conscious function.

I mentioned Melvin before, who lost conscious awareness of his emotional states after a stroke wiped out the neural bridge that feeds information to the part of the brain that handles awareness of emotions specifically.

His condition didn't arise because he lost a certain percentage of neurons, or a particular type of neurons. It came about because a connection was broken between brain modules.

If our brains were made of something other than neurons, and yet information still flowed among these larger structures in the same way, his stroke would have had the same effect.

There's also the case of Terry Schiavo, whose parents insisted she was conscious, but whose husband insisted she was not.

Despite the fact that she in some ways appeared to be aware but to have very limited functionality, there was a medical certainty that she was not in fact aware of anything, because of the macro-structures in her brain that had been destroyed.

What makes our brains work the way they do -- what makes them capable of posting on this forum, for example -- is the sequencing of the neurons and how they are grouped into larger structures which communicate among each other.

Rewire it in a way that allows the neurons to fire all happy, but that routes the information in a radically different pattern, and the brain will not do what it does.

There are simpler analogs in other organs. Take the heart, for instance. The muscular contractions follow a sequence that makes our hearts beat so that they pump blood. But there are other patterns that the contractions can follow along that network.

When there's a shock to the heart, it can "knock" the heart out of the normal pattern and into a different one. From the point of view of the muscle cells, everything's fine. They're doing what they have always done. But from the point of view of the organ, nothing's fine, because the new pattern of contraction doesn't make the heart beat so that it pumps blood.

That's why we use defibrillators -- to shock the heart again in an attempt to "knock" it back into the old pattern.

Similarly, it's not the flow of impulses from neuron to neuron that makes the brain work. Rather, it's a specific pattern of flow from neuron to neuron, and by extension from module to module, which creates higher-level phenomena that result in our being able to speak, read, type, even cough and breathe.

Well, see, I agree with all of that. Consciousness (so far as we can tell) arises from how neurons (or whatever) interact. A brain made of computer chips could be conscious. So could one with mechanical gears.

But yet in another post you say there is a speed limit. That's the problem, you say it, without any evidence or reasoning. it makes no sense. I postulate our universe stops 1 billion times a second, with what we consider 1 billion years during the pause. That's completely unfalsafiable, as there is nothing we could observe to contradict that. Yet your conscious continues unabated. Based on everything you've typed you say there is nothing special or 'extra', so you'd have to conclude that if the universe worked that way, you'd still be conscious.

And don't go talking about the pauses. None of of think the consciousness continues during the pause.

As for the pencil special case, I don't see why you draw it. You admit a computer chip brain would be conscious, as would a cog brain, yet for some weird reason a pencil brain wouldn't be. Yet we are talkiing about a pencil brain that creates "a specific pattern of flow from neuron to neuron, and by extension from module to module," Exact same patterns, exact same modules, just running on a different substrate.

x = x + 1 doesn't run differently on silicon vs galium arsenide. The brain patterns and modules you mentioned don't run differently on human meat, silicon, the difference engine, a TM, or pen/paper. Same algorithms, same organization, same data input, same data output. Same.

 

[Piggy]

Lack of resemblance is irrelevant.

Of course it's relevant.

The bottle of white-out on my desk doesn't resemble my car's engine in any way, therefore I have no reason to believe it could run my car.

Okay, so do you agree that the brain must not be equivalent to a Turing machine? If so, what sort of additional mechanisms are there that preclude the process of consciousness being generated by a Turing machine-equivalent computer such as pencil and paper?

That's not what I said.

What I said was that there's no reason to believe that the Turing test can distinguish between conscious and unconscious machines/beings.

Therefore, it is a fatal flaw to rely on it to do so.

But if it's not conscious in some manner similar to humans, then the thought experiment is also useless because we can't even begin to consider the question of whether processing speed matters.

That's been my point for some time now.

And one of the two ways we know other humans are conscious is by using some variant of a Turing test: I interact with you and you appear to be conscious.

I don't buy that.

We know other people are conscious because they have human bodies (which means human brains) and they interact with us.

They've got the same mechanism we do.

 

[Piggy]

Well, see, I agree with all of that. Consciousness (so far as we can tell) arises from how neurons (or whatever) interact. A brain made of computer chips could be conscious. So could one with mechanical gears.

Not so fast with the mechanical gears there.

But yet in another post you say there is a speed limit. That's the problem, you say it, without any evidence or reasoning. it makes no sense. I postulate our universe stops 1 billion times a second, with what we consider 1 billion years during the pause. That's completely unfalsafiable, as there is nothing we could observe to contradict that. Yet your conscious continues unabated.

No, because if my consciousness continued unabated, I'd be aware of the pauses. Maybe my consciousness flickers in and out with the universe. I kinda think there'd be some fingerprint if that actually happened, but that's not relevant here.

As for the speed floor, there's every reason to believe it exists. We know that the brain works at the current speed, and that it relies on rapid coordination of inputs, including its own outputs as inputs.

We also know that there's no consciousness when there's no activity.

And based on what we can deduce from how the brain is built, and the rather fuzzy nature of consciousness, it's just impossible to believe that one signal per second would be sufficient to maintain the effect.

Although I freely admit that I don't know whereabouts the floor would be. Neither does anyone else right now, I suspect.

Based on everything you've typed you say there is nothing special or 'extra', so you'd have to conclude that if the universe worked that way, you'd still be conscious.

And that's consistent with what I've been saying. If we do the dump-redump, our robot is not conscious in the interim.

And don't go talking about the pauses. None of of think the consciousness continues during the pause.

Too late. I just don't get your point here. We seem to agree.

As for the pencil special case, I don't see why you draw it. You admit a computer chip brain would be conscious, as would a cog brain, yet for some weird reason a pencil brain wouldn't be.

I don't know that a cog brain would work. I suspect it would not. How are you going to reproduce threshold sensitivity with cogs?

I'm not 100% certain a chip brain can be conscious. I suspect it can mimic the modular structure of the brain. Not sure, though.

The pencil brain is sheer lunacy.

Yet we are talking about a pencil brain that creates "a specific pattern of flow from neuron to neuron, and by extension from module to module," Exact same patterns, exact same modules, just running on a different substrate.

No, it's not.

Doing the equations on paper is, as I've pointed out numerous times, removing the hardware.

When you get down to it, the brain is nothing but hardware. It's all physical reaction.

I can sit down and, theoretically, if I had some life-extending drug that gave me the time, describe on paper every chemical process that occurred in John Doe's brain in his life.

Yet this does not constitute a pencil-and-paper brain.

Why? Because what's happening in the real world is that I am sitting at a table moving a pencil.

It doesn't matter that the symbols on the paper represent (to me) the chemical reactions in John's brain. The fact remains that they are not, in reality, those reactions.

No chemical reactions are happening. No sort of brain is operating, on any sort of substrate.

I may think in my head "These symbols represent the actions of John's brain to me" while I'm doing it... but that doesn't change the physical reality.

x = x + 1 doesn't run differently on silicon vs galium arsenide. The brain patterns and modules you mentioned don't run differently on human meat, silicon, the difference engine, a TM, or pen/paper. Same algorithms, same organization, same data input, same data output. Same.

Pen/paper is entirely different, for the reason I just described.

 

[shuttlt]

We are talking about the entire process. Are you conscious while you sleep? No.

Pedantic, but I think this is a different meaning of the word 'conscious'.

Really?

When I am asleep I am unconscious. When I am asleep and dreaming I am still unconscious, but clearly in another sense I am conscious.

Gravity is unique, too.

Not in the same sense. Gravity is a force, there are many forces. Pick something else that exists in the same sense that consciousness clearly exists but that can't be measured, or detected (outside our subjective experience of it).

You could falsify the claim that consciousness is a function of the brain by finding something else that is conscious.

And how would you know if it was conscious. For all I know, my blackberry is conscious. For all I know the woos are right and the whole universe is one big consciousness. I don't believe any of that stuff, but I don't see any way that one could find out.

I could perhaps suggest a way to falsify your hypothesis, but I don't know what it is.

Let me know when you think of a way. I don't believe any such test could be performed.

 

[shuttlt]

I think that would be "with bated keyboard" -- but then again, I don't know your hobbies.

Damn my rubbish spelling. I knew it wasn't a good idea to use Google to check my spelling.

I mentioned Melvin before, who lost conscious awareness of his emotional states after a stroke wiped out the neural bridge that feeds information to the part of the brain that handles awareness of emotions specifically.

His condition didn't arise because he lost a certain percentage of neurons, or a particular type of neurons. It came about because a connection was broken between brain modules.

Perhaps he was only passing-a-Turing-test conscious after his stroke. Is there a way to tell?

If our brains were made of something other than neurons, and yet information still flowed among these larger structures in the same way, his stroke would have had the same effect.

OK

There's also the case of Terry Schiavo, whose parents insisted she was conscious, but whose husband insisted she was not.

Despite the fact that she in some ways appeared to be aware but to have very limited functionality, there was a medical certainty that she was not in fact aware of anything, because of the macro-structures in her brain that had been destroyed.

When they say conscious, don't they mean... aware of what's going on, having thoughts....

What makes our brains work the way they do -- what makes them capable of posting on this forum, for example -- is the sequencing of the neurons and how they are grouped into larger structures which communicate among each other.

OK

Rewire it in a way that allows the neurons to fire all happy, but that routes the information in a radically different pattern, and the brain will not do what it does.

Wasn't there a thread about a girl born with half a brain a week or two ago? It's amazing how much rewiring can be done and still be conscious.

There are simpler analogs in other organs. Take the heart, for instance. The muscular contractions follow a sequence that makes our hearts beat so that they pump blood. But there are other patterns that the contractions can follow along that network.

You are treating consciousness and information processing as if they had a 1-1 relationship. If that is true, then I agree with you.

When there's a shock to the heart, it can "knock" the heart out of the normal pattern and into a different one. From the point of view of the muscle cells, everything's fine. They're doing what they have always done. But from the point of view of the organ, nothing's fine, because the new pattern of contraction doesn't make the heart beat so that it pumps blood.

This is a plausible theory, but you don't have a test to confirm it.

Similarly, it's not the flow of impulses from neuron to neuron that makes the brain work. Rather, it's a specific pattern of flow from neuron to neuron, and by extension from module to module, which creates higher-level phenomena that result in our being able to speak, read, type, even cough and breathe.

Why are being able to speak, read, type, cough and breathe necessarily related to consciousness? I can imagine being unable to do any of those things and still be conscious. I can imagine being unable to sense, or interact with the outside world, or the rest of my body and still be conscious. I can imagine being unable to make memories and living in an ever present now and still be conscious. It's harder to imagine, but why couldn't I be conscious even if I wasn't actually having any thoughts any longer? Perhaps I would be, perhaps I wouldn't. Is there any test we can do to determine whether at any of these points I would cease to be conscious?

 

[shuttlt]

Surely you don't believe that full-blown human consciousness exists at the level of the neuron.

Helpfully, I don't have a theory of my own to add.

Even if there is some mysterious mindon particle, completely hidden from our particle theories and our accelerators but required to endow zombie matter with consciousness, it's pretty clear that human consciousness is an emergent property of the brain. Otherwise we should be having metaphysical conversations with the wind.

Additional particles just move the problem back. It's the old elephants all the way down problem. As for metaphysics, if you want to insist that consciousness fits into a particular mode of explanation then that's fine, I'm sure you'll find, given that assumption, that consciousness does indeed fit into that mode of explanation.

I suppose you might believe that full-blown human souls exist separately and somehow invade the brain and animate it. But then we should be having metaphysical conversations with thin air. And not the kind John Edward has.

I'm not about to claim on the JREF that souls flit about the place inhabiting brains like wandering Thetans.

 

[Piggy]

This is a plausible theory, but you don't have a test to confirm it.

Regarding heart contractions, are you saying you know of someone who disputes that?

 

[Piggy]

Why are being able to speak, read, type, cough and breathe necessarily related to consciousness?

They're not. I was speaking, at that point, of the the functioning of the brain. Consciousness is simply one of those functions.

 

[Dan O.]

I can imagine being unable to sense, or interact with the outside world, or the rest of my body and still be conscious.

In the definition of "conscious", the ability to sense and interact with the outside is a necessary component. However, there is a related state that occurs during lucid dreams where one has the same sense of thought as when awake. Is this lucid state "conscious"?

If we call the lucid state "self conscious", is there a state of consciousness that does not include being self conscious (ie, can you be conscious and not aware of it)?

 

[Paul C. Anagnostopoulos]

The bottle of white-out on my desk doesn't resemble my car's engine in any way, therefore I have no reason to believe it could run my car.

You know they don't resemble each other because you know exactly what those two things do. You do not know how the brain works nor whether a black-box program is emulating one. Therefore you cannot "calculate their resemblance."

That's not what I said.

What I said was that there's no reason to believe that the Turing test can distinguish between conscious and unconscious machines/beings.

I wasn't responding to any comment about a Turing test. You said:

We know of one object that produces consciousness. We know of other similar objects we can be pretty confident about.

None of them bear any resemblance to dragging a pen across paper.

And I responded:

Okay, so do you agree that the brain must not be equivalent to a Turing machine? If so, what sort of additional mechanisms are there that preclude the process of consciousness being generated by a Turing machine-equivalent computer such as pencil and paper?

I don't buy that.

We know other people are conscious because they have human bodies (which means human brains) and they interact with us.

Yeah, that's what I said:

And one of the two ways we know other humans are conscious is by using some variant of a Turing test: I interact with you and you appear to be conscious.

They've got the same mechanism we do.

Yes, that is the other way you assume a human is conscious: He has the same sort of brain that you do.

~~ Paul

 

[Paul C. Anagnostopoulos]

When I am asleep I am unconscious. When I am asleep and dreaming I am still unconscious, but clearly in another sense I am conscious.

I think you just have altered consciousness when you are dreaming. If that's what you meant by "different meaning," then fine.

And how would you know if it was conscious. For all I know, my blackberry is conscious. For all I know the woos are right and the whole universe is one big consciousness. I don't believe any of that stuff, but I don't see any way that one could find out.

Just try to find something that is conscious in a manner similar to humans, but without a brain. That would falsify the claim that consciousness is a function of the brain.

Let me know when you think of a way. I don't believe any such test could be performed.

I don't know what you think consciousness is, so I can't propose a way to falsify it.

Additional particles just move the problem back. It's the old elephants all the way down problem. As for metaphysics, if you want to insist that consciousness fits into a particular mode of explanation then that's fine, I'm sure you'll find, given that assumption, that consciousness does indeed fit into that mode of explanation.

I have no plans to bring up metaphysics when trying to explain consciousness. I just figure it fits into science somewhere. Is that a mode of explanation?

I'm not about to claim on the JREF that souls flit about the place inhabiting brains like wandering Thetans.

Good decision.

~~ Paul