• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Robot consciousness

Piggy said:
What is "consciousness" without "subjective experience"?
Click to expand...
For all I know we could be talking at cross purposes, this topic is terrible for that. To me, the answer to your question is "nothing". I just don't understand how to interpret Paul's comment about being able to turn off vision as demonstrating anything about consciousness if he shares my definition. My assumption is that he sees no distinction between being a soulless zombie/pod person and not being a soulless zombie/pod person so long as the information processing is the same, or rather I think he denies the possibility of a soulless zombie version of me.
 
Piggy said:
Regarding speed, when you're dealing with effects of physical systems, you can encounter windows of viability, and I believe that's the case w/ consciousness.

For example, a heart that beats too slow or too fast over a long enough period of time will fail to do what a heart is designed to do, resulting in failure of the entire biological system.
Click to expand...

That is true, but in those cases it isn't the entire system that is being slowed down, just part of it, and that is the cause of the problem.

For a heart, if you slow it down, the molecular motion of the blood eventually negates anything the heart itself does. If you could slow down those molecules as well, though, it would continue to work as before.

That is kind of the whole idea behind time diliation at high velocity, right? Even though a system is slowed to near stasis, it still functions normally, because everything is affected.
 
Piggy said:
"Working out" and "working" aren't the same thing.
Click to expand...
I'm talking about simulating the computer by hand, performing all the same calculations.

Pencil and paper are not the proper hardware.
Click to expand...
Why not?


If the only functioning example we can be sure about does not meet those requirements, then why are they relevant?
Click to expand...
To point out that the functioning example may indeed by affected by the speed of the underlying processor.

I've outlined a process that actually has a chance of working.

Your argument seems to be like that of the drunk who drops his keys into the weeds, but goes looking for them under the streetlamp because the light is better.
Click to expand...
Huh? Which argument of mine are you talking about?

~~ Paul
 
shuttIt said:
We're talking about different stuff. Of course you can turn off individual subsystems, or at least I'm happy enough assuming it. If you turn off the bits related to speech I won't be able to speak and it'll probably play havoc with my ability to think. How is this consciousness...? It's information processing. I can imagine that system working in the absence of me having a subjective experience of consciousness just as easily as with.
Click to expand...
Do you know what blindsight is? It's vision without consciousness of vision.

~~ Paul
 
Almo said:
I think clock speed is a big, fat red herring. If the algo works, it works. You might just not percieve its conciousness because your timescale is different.
Click to expand...
What if the algorithm includes various clocks and "measures" internal clock speed against wall clock speed?

rocketdodger said:
If you hold the view that consciousness is a property of information alone, that just happens to find a physical instantiation in human brains, then mathematically the "rate" is irrelevant.
Click to expand...
Ditto.

If you hold the view that the substrate itself is somehow intrinsic to consciousness, beyond the behavior it offers information that might be instantiated upon it, then "rate" could indeed be important.
Click to expand...
There is a big difference between a chemical and a silicon substrate, even if there is no magical way in which a chemical one attracts consciousness.

It seems that people are thinking about the purely computational aspects of consciousness without considering the subtleties of sensory input and other interactions with the external world.

~~ Paul
 
I couldn't get the quote function to work as I expected and I have no patience.

Why are pencil and paper not the correct tools to replicate how software performs?
A demonstration:
The code that causes your screen to display images. No matter how many times you replicate the process with pencil and paper it will not cause a crystal on a display to change its orientation, and it must be done correctly 400 times a second to not look blurry. The human eye can only catch about 25 times a second, but since the twist only lasts a very short time it must happen frequently to be seen.

Note that this is only one example of why pencil and paper execution of code is not the same as running it on the hardware it was intended for. Merely intended to illustrate the difference.
 
shuttIt said:
For all I know we could be talking at cross purposes, this topic is terrible for that. To me, the answer to your question is "nothing". I just don't understand how to interpret Paul's comment about being able to turn off vision as demonstrating anything about consciousness if he shares my definition. My assumption is that he sees no distinction between being a soulless zombie/pod person and not being a soulless zombie/pod person so long as the information processing is the same, or rather I think he denies the possibility of a soulless zombie version of me.
Click to expand...
As I clarified above, blindsight doesn't turn off vision, just consciousness of certain aspects of it.

I do not believe we have any sort of soul that could be disembodied from our brains. If that is the case, then p-zombies are an incoherent concept.

~~ Paul
 
LissaLysikan said:
Why are pencil and paper not the correct tools to replicate how software performs?
A demonstration:
The code that causes your screen to display images. No matter how many times you replicate the process with pencil and paper it will not cause a crystal on a display to change its orientation, and it must be done correctly 400 times a second to not look blurry. The human eye can only catch about 25 times a second, but since the twist only lasts a very short time it must happen frequently to be seen.

Note that this is only one example of why pencil and paper execution of code is not the same as running it on the hardware it was intended for. Merely intended to illustrate the difference.
Click to expand...
When I talk about hand-simulating the algorithms of a robot brain, I'm talking about a (difficult to create) situation where the inputs are from the robot's senses and the outputs are fed back into the robot's servo system. I'm simply trying to picture off-loading te computations onto pencil and paper.

If the computations are purely internal and deterministic, it shouldn't make any difference. However, if part of the computation of consciousness involves something like comparing various clocks, some of which are circadian, then it could make a difference. The human brain synchronizes ensembles of neurons in certain timeframes, so perhaps real time matters.

I'm not claiming to be asking a particularly profound question here, but the topic came up on another forum and I knew I'd get good responses here.

~~ Paul
 
Paul C. Anagnostopoulos said:
When I talk about hand-simulating the algorithms of a robot brain, I'm talking about a (difficult to create) situation where the inputs are from the robot's senses and the outputs are fed back into the robot's servo system. I'm simply trying to picture off-loading te computations onto pencil and paper.

If the computations are purely internal and deterministic, it shouldn't make any difference. However, if part of the computation of consciousness involves something like comparing various clocks, some of which are circadian, then it could make a difference. The human brain synchronizes ensembles of neurons in certain timeframes, so perhaps real time matters.

I'm not claiming to be asking a particularly profound question here, but the topic came up on another forum and I knew I'd get good responses here.

~~ Paul
Click to expand...

But how would you handle the multiple processes? Even current, non-intelligent software has multiple processes working both synchronously and asynchronously that merge at some point/points to produce a result. The concept of trying to replicate that for a consciousness simulation is mind-boggling, to put it mildly. How do you account for the time that might be spent in the "emotion" thread running async with the "logic" thread running async with the "primitive hardcoded response" (instinct) thread? You would need to figure out a timing that worked for every one of them, then mesh them at the correct places.
BUT - if you could do that, then the clock/rate/time would be irrelevant - as long as all threads perform at the same relative rate, it shouldn't matter. Unless the hypothetical robot was about to be eaten by something big.
 
Paul C. Anagnostopoulos said:
When I talk about hand-simulating the algorithms of a robot brain, I'm talking about a (difficult to create) situation where the inputs are from the robot's senses and the outputs are fed back into the robot's servo system. I'm simply trying to picture off-loading te computations onto pencil and paper.
Click to expand...

Ah. Now I achieve clarity.

It would break the system.
 
LissaLysikan said:
But how would you handle the multiple processes? Even current, non-intelligent software has multiple processes working both synchronously and asynchronously that merge at some point/points to produce a result. The concept of trying to replicate that for a consciousness simulation is mind-boggling, to put it mildly. How do you account for the time that might be spent in the "emotion" thread running async with the "logic" thread running async with the "primitive hardcoded response" (instinct) thread? You would need to figure out a timing that worked for every one of them, then mesh them at the correct places.
Click to expand...
Are you talking about this being a problem for hand-simulation, or even on a real computer?

BUT - if you could do that, then the clock/rate/time would be irrelevant - as long as all threads perform at the same relative rate, it shouldn't matter. Unless the hypothetical robot was about to be eaten by something big.
Click to expand...
You're not addressing my issue of comparing neural clock rates with circadian clocks. You're assuming real-time independency of all algorithms.

Piggy said:
I know. But performing the calculations by hand isn't the same as running the program on the hardware.
Click to expand...
Why not? That's the question.

It would break the system.
Click to expand...
Why?

~~ Paul
 
Paul C. Anagnostopoulos said:
Why?
Click to expand...

Well, we're back to what I pointed out earlier.

The only working model we can be sure about relies on rapid parallel processing and some series processing among modules.

It appears to arise out of a coordination of inputs in real time.

Now, if you had one person for each neuron -- because you can't say in advance which sets of neurons are going to be firing at any time -- and each person worked at exactly the same speed, and all they had to do was convey a yes (1) when any neuron fired, I suppose you could route brain activity through a very large number of human calculators (who would not be doing much calculation) and back in again, and the only effect would be a delayed consciousness, which would simply confuse the being and make it uncoordinated.

But one person would not be able to take all the "input" -- which gets real complex b/c in the brain much output=input -- for all the modules (assuming it can even be defined clearly enough for that) and feed them back into the system.

If you're going to propose a robot that is conscious by some other means (tho you've stipulated that it's conscious like we are) then you'll have to specify those means before you can ask the question.
 
circadian clocks are only relevant to living things that depend on time-of-day due to evolutionary pressures. For an AI they wouldn't matter. All that would matter then is whatever pressures caused the breakthrough to consciousness and if it continued in the simulation (the pencil/paper recreation).
 
The rate of computations wouldn't matter at all - as long as all parts of the system are working at the same relative rates, absolute time wouldn't matter, and intuitively shouldn't. Why would it matter if it takes four seconds or 1/4 second for part A to respond to part B, if part B is working on the same time scale as part A?
 
Last edited:
Paul C. Anagnostopoulos said:
As I clarified above, blindsight doesn't turn off vision, just consciousness of certain aspects of it.
Click to expand...
I know what blindsight is. It's like I'm blind, but if you throw me a ball I can catch it....

Paul C. Anagnostopoulos said:
I do not believe we have any sort of soul that could be disembodied from our brains. If that is the case, then p-zombies are an incoherent concept.
Click to expand...
I think you've left wiggle room by including the bolded part of the first sentence. Forget woo souls that can float off and talk to angels. Lets say 'souls' are grounded in brains. What then? Can you imaging a brain with a 'soul' and one without?
 
Last edited:
From taking his sentence in context, I gather he's saying he doesn't believe in the concept of a supernatural part of the mind. "soul" is just a convenient term for that.
 
LissaLysikan said:
The rate of computations wouldn't matter at all - as long as all parts of the system are working at the same relative rates, absolute time wouldn't matter, and intuitively shouldn't. Why would it matter if it takes four seconds or 1/4 second for part A to respond to part B, if part B is working on the same time scale as part A?
Click to expand...

The only working model we have is a biological one that operates in a real-world context.

If we're talking about some other type of contraption that produces consciousness in a different way, we need to describe it before we can answer the question.

In the case of the human brain, conscious experience appears to be a downstream effect that depends heavily on parallell processing and feedback loops. The brain is taking its own outputs as inputs.

What's analogous to programming in the brain is likely very simple. What creates the effect of consciousness is the activity of the brain running at speed.

So the question here is, how much could the neuronal activity be slowed down before consciousness ceased?

It's analogous to asking how low blood pressure can go or how slow the heart can beat before the system fails.

Clearly, the heart can't beat once per minute and supply oxygen to cells fast enough to maintain life.

Presumably, there's some threshhold for neuronal activity as well.

I don't know what it is, but it must certainly be higher than hand-calculation speed, if we're talking about one person trying to work out everything the brain is doing.
 
You must log in or register to reply here.

Back
Top Bottom