Has consciousness been fully explained?

[westprog]

And yet there are flying simulations. Simulated planes.

And yet there are automated pilots. Some day they'll lift off and land, too, if it's not the case already.

So we can simulate the plane AND the flying.

And a simulated plane and an automatic pilot are entirely different things. Yet in the world of artificial consciousness, they are treated as if they were the same thing.

It is not possible to plug a simulator into a real plane and fly it. The idea that you might be able to is an obvious category error.

If it were possible to produce an "automatic pilot" for a person - a device that replaces the brain and actually appeared to fullfil all its functions - then that would at least be evidence of the possibility of machine consciousness.

What's being proposed by the proponents of artificial intelligence, though, is as if flight were produced by the simulator.

 

[westprog]

Indeed. Otherwise NO simulation would be possible. Ever.

Maybe some clarity about emulation, simulation and what is actually being proposed would be helpful.

 

[Belz...]

If the simulation is required to interact with something outside itself, then a Turing machine is a bad model.

Who cares about fictional models ? We're talking about simulations in reality, are we not ?

All scientific models are a fiction. We choose them according to the extent to which they describe reality. We know that they aren't "real".

Theories are not simulations.

 

[Belz...]

And a simulated plane and an automatic pilot are entirely different things.

Indeed. Put both of them together, however, and you've just simulated flying.

Yet in the world of artificial consciousness, they are treated as if they were the same thing.

No, they're not. One is simulated (brain) in order for the other to emerge (consciousness).

It is not possible to plug a simulator into a real plane and fly it.

No but you can plug an automatic pilot into a simulated plane and it won't know the difference.

What's being proposed by the proponents of artificial intelligence, though, is as if flight were produced by the simulator.

Within the simulator, it is.

 

[rocketdodger]

Grandfather clocks can switch between two states. An input such as a sudden jolt, for instance, can trigger the switch between swinging and standing still, in either direction. So the whole thing is a switch. Therefore, by your definition, it's a simple computer.

The heart also can be switched among states by varying the input -- normal beat, tachycardia, fibrilation. This is because there is more than one possible pattern in which the group of muscle cells can contract.

So again, by your definition, the heart is a simple computer.

Therefore, to say that the brain is a computer, from this perspective, makes it no different from a heart or a grandfather clock.

"no different?"

wow

so by your logic a vacuum tube is "no different" from Deep Blue

 

[rocketdodger]

It's posts like this that I need to make a record of anytime someone enters these threads and says, surely the computationalists aren't making the claims you say they're making... you must have misunderstood them.

This post exhibits a profound lack of understanding of the brain and of consciousness.

Why?

 

[rocketdodger]

Difficult to know how to respond to a non sequitur like that, but I'll try.

We can be pretty darn sure that it is not true that all the functions of the human brain are necessary for consciousness.

But this has nothing to do with the question of whether you can make a set of tinker toys do what a human brain does.

Huh?

Let me formulate the statements in a way that might be more clear to you:

We are not sure what functions of the human brain are necessary for consciousness, call this set A.

Some of the functions of the brain can be done with a set of tinker toys, call this set B.

Therefore it is possible that the set of functions B is a superset of A. Until you know of a function in A that is not in B, it has very much to do with it. That was in no way a non sequitur.

 

[rocketdodger]

If you think the brain of a newborn is similar in its structure, functions, or capacities to a thermostat, then you've got a lot to learn about the brains of newborns.

You should also be much more circumspect about using overt bodily behavior as a marker of conscious awareness.

The brains of newborns are doing all kinds of amazing things.

But in any case, no, the current research gives us precisely 0 reasons to suspect that newborns are not conscious.

ETA: We also have no reasons from an evolutionary perspective to believe that they would not be, and every reason to believe that they are.

Why should I consider them conscious? They display less complex reaction to external stimuli than even the smallest mammals.

 

[rocketdodger]

It's posts like this that I need to make a record of anytime someone enters these threads and says, surely the computationalists aren't making the claims you say they're making... you must have misunderstood them.

This post exhibits a profound lack of understanding of the brain and of consciousness.

Why?

Newborns cry. Thats about it. They don't open their eyes, they can't move in any coordinated fashion, they certainly can't communicate, and a squirrel is more self-aware.

I don't see why the fact that a newborn develops into a fully conscious adult human should have any bearing on the question of their level of consciousness when they pop out of their mother.

And, note that I have never said newborns are not conscious. I just said their consciousness isn't very similar to an adult human. I know it isn't similar because I can't interact with a newborn in most of the ways I can interact with an adult.

What about this logic do you find incorrect?

 

[Frank Newgent]

Why?

Newborns cry. Thats about it. They don't open their eyes, they can't move in any coordinated fashion, they certainly can't communicate, and a squirrel is more self-aware.

I don't see why the fact that a newborn develops into a fully conscious adult human should have any bearing on the question of their level of consciousness when they pop out of their mother.

And, note that I have never said newborns are not conscious. I just said their consciousness isn't very similar to an adult human. I know it isn't similar because I can't interact with a newborn in most of the ways I can interact with an adult.

What about this logic do you find incorrect?

I'll bite. How is a newborn like a programmable thermostat?

 

[dlorde]

I'll bite. How is a newborn like a programmable thermostat?

Sounds like the Mad Hatter: "Why is a raven like a writing desk?"

 

[PixyMisa]

"no different?"

Well, no different from a hundred billion grandfather clocks interconnected by an amazing everchanging network of ropes and pulleys.

I'm sure that's what Piggy meant.

 

[Piggy]

I was talking about neurons, and you know it. You're just being deliberately obtuse. Color me unimpressed.

I know precisely what you wrote. You wrote that we can program their building blocks, which are physical matter.

What I'm trying to get you to understand is that we can do no such thing, and that creating digital simulations is quite a different kettle of fish and not at all comparable.

 

[Piggy]

So what? Those play the same role in its consciousness as your eyes, arms, legs, mouth, etc. They are not essential. Sure, without any senses and any means to produce some effect on the outside world, consciousness cannot exist. But those parts are just the means. The only essential part is the brain.

Quite right about the only essential part being the brain.

I was not saying that arms and legs are necessary for consciousness.

I was drawing an analogy between their role in the behaviors they're involved in, and the brain's role in the behaviors that it is involved in.

If you want to create a machine that exhibits the behavior of moving across the floor, you can't do it by pure programming with only enough hardware to support the programming and no more. You have to have enough hardware to support running the program and to enable the movement.

Same thing for any other behavior you care to get a machine to perform, whether that's grasping, seeing, speaking, or being conscious (I still hate it that we have no verb for this).

There are no pure-programming solutions for behavior.

If a robot has real senses and real effectors it can be conscious the same way we are. If a program is running in a simulated environment, with emulated senses and effectors, it can be conscious in that simulated world.

It doesn't really matter whether a robot or a person is conscious of the external world, a simulated world, or a dream.

The fact remains that consciousness is a behavior locatable in real spacetime.

And programming alone is insufficient to produce such behavior.

You want a conscious robot, you're going to have to build it to be conscious.

We don't yet know how to do that, but we know that biological processes build our brains to do it.

 

[Piggy]

Could you provide a link to where he said that so I could have a context?

What I think based on the above is that for most neuron types we do know how they behave. While it might be possible that the remaining few that are not fully understood are critical for consciousness, it is unlikely, and furthermore, it is just a matter of time before neuroscience explains those as well. Either way, the probability of us not being able to explain how all types of neurons behave is so low that we can classify it in the same class as fairies or Santa existing.

I believe it's in "The Cognitive Neurosciences, 4th Ed." but I don't have a link. (ETA: It's either there or "Human" -- I'm reading both simultaneously.)

Anyway, yes, it's quite possible we'll figure it all out before long.

Doesn't change the fact that neurons are bits of matter, so if you want to build a functional equivalent it, you're not going to program your way there, for the same reason that you're not going to program your way to the functional equivalent of a heart or a leg.

Yes, computers and programming could well be involved, but there has to be a physical component.

 

[Piggy]

No, it's not. I don't have to be able to program matter to be able to emulate a system built out of matter.

Incorrect. You can represent the system. But that's not the same as creating something that can do what the system does in real 4-D spacetime.

You can't live in a blueprint of a house. You can't get electricity from a digital simulation of a hydroelectric station.

For precisely the same reason, creating a digital simulation of a brain will not generate a real-world instance of consciousness.

ETA: The folks who are in the process of creating a neural-level simulation of the brain agree that the machine running it will not be conscious, because a simulation can only represent a brain, it cannot be a brain.

 

[Piggy]

Except that there is no evidence against, and all existing evidence points to the fact that consciousness is indeed programmable.

What evidence is that?

Care to point to it?

I'm not aware of any, and the little that's been offered in this thread, and others on this forum, has been roundly debunked.

On the other hand, we can see that claiming a pure-programming solution is an attempt to generation real-world behavior with no real-world cause.

Since we can name no other behavior that anyone would claim can be purely programmed, what makes consciousness an exception? Why does this behavior violate the laws of physics?

An explanation would be very much appreciated.

 

[Piggy]

The right fuel ? A particular configuration ? I don't think those claims are justified.

They're not only justified, they're undergrad-level biology.

Don't feed your brain sugar and oxygen and see what happens to your conscious awareness.

The need for proper physical configuration is demonstrated by studies of emotional blindness, pain blindness, split brains, and so on.

 

[Piggy]

And yet there are flying simulations. Simulated planes.

And yet there are automated pilots. Some day they'll lift off and land, too, if it's not the case already.

So we can simulate the plane AND the flying.

So what?

The very reason we use flight simulations is so that no one gets hurt when the plane "crashes" -- i.e., it ain't real, it's just designed to trick a human brain into perceiving a phony reality that ain't there.

Yet for some reason, some folks want an exception whereby a simulation of a brain is real.

No one that I've seen -- here or anywhere else -- has offered any explanation of why that might be the case, or how.

 

[Piggy]

We're talking about simulations in reality, are we not ?

No.

We should be talking about brains, if we were to stick to the OP, but since this hijack is so long established, as it is we're talking about whether machines can be conscious, which is not a simulation of a brain but rather a functional model of a brain.

If we were talking about legs, we'd be talking about artificial legs, not computer simulations of legs. The two things are significantly different.

A robot could not walk on a digital simulation of a leg, only a functional equivalent of a leg.

For the same reason, a robot could not be conscious by running a simulation of a brain, but only by using a functional equivalent of a brain.

And a computer that runs simulations is not a functional equivalent of a brain. It doesn't do what brains do. It's not built to be conscious, and running a simulation of a brain won't make it conscious, for the same reason that running a simulation of a power plant won't generate electricity and allow the simulator to plug into itself.