Has consciousness been fully explained?

[rocketdodger]

But it could if we programmed it to do so. We don't because it is a tool and we want predictable performance, but there is no absolute barrier to a computer choosing one thing over another.

Westprog is trying to say that because computers ultimately come from human intelligence, they are equivalent to any other system that ultimately comes from human intelligence.

As a test I asked him if he would consider a human constructed biological person conscious. That is, if we built a person that was just a normal human -- other than the fact that he/she was constructed by human intelligence.

The answer? No answer -- as predicted.

 

[Drachasor]

Non sequitur. My post said nothing about how the brain works.

So can I take it you're either unable or unwilling to answer the question? What do you think the definition of "information" is?

It's not a non sequitur. I defined information as far as it matters to the brain, which is at the heart of the discussion.

 

[rocketdodger]

What do you think the definition of "information" is?

Given a system A and B, "information" is behavior of A (or a subsystem of A) that B interprets as referencing some other system C.

Meaning, the stuff inside a rock just sitting there isn't "information" until some other system interprets it as such.

 

[westprog]

Likewise I am amazed that you still think there is a difference, given that nothing in the universe functions outside of time. Where is this elusive "computational process" that is not constrained by the laws of nature, westprog?

Where?

Where?

When anyone who knows what they are talking about -- obviously you are not included -- speaks of a "computational process" they just mean a process that behaves in a way that can be described using the ideas of computation. Period. End of story.

The only person talking about some abstract Turing machine is you. You are alone in this.

Oh, I see. I just imagined the constant references to Church-Turing and the wonderful things that it supposedly proves. Just my imagination.

I've taken it on trust that you have some kind of knowledge and understanding of actual programming practice, but you've continually tried to demonstrate the opposite. It's not a matter of some obscure theoretical difference. It's a fundamental issue that real programmers working on real problems have to address.

Say, for example, a program to calculate the first hundred primes. Such a program would not be time dependent. It could take a microsecond, or a year, and it would produce exactly the same output. That kind of computation could be written by, say, a Pascal program.

Now say you want to write a program that switches a pump on when the level of a reservoir falls below one meter, and turns it on when it reaches four meters. Try to do that in Pascal. You cannot. The language specification won't allow for it. Pascal is designed to perform computations. That's not an accident. It's because Pascal (just like most programming languages) is designed to perform computations. It's a way of expressing a Turing machine. If you want to control the pump, and read the level of the water, and what's most significant, do it before the level of the water reaches five meters and overflows the reservoir, then you need to extend the language in some way with a way to interact with the environment in an asynchronous way.

This is why when it's asserted that according to Church-Turing (which apparently I'm the only person to refer to) all the functions of the brain can be performed by a computation. Now, I know what that means - it means a program such as the one that calculates prime numbers. It does not mean a program such as the one that controls the water works. So I know that the statement is not true.

If you are going to insist that it's the second type of program that you're referring to, then you have to abandon such theories as Church-Turing which refer to computation only.

You can continue to obfuscate and insist that asynchronous, event-driven programming is the same as Turing compatible computation, but that is demonstrably not the case.

Wrong. It tells us, at the very least, that it isn't some magic fairy dust

Oh, brilliant. You could publish that as a thesis.

You know, the constant obsession with magic doesn't really lend your ideas added significance.

limited to biological neurons that is responsible for consciousness.

In other words, a step in the right direction.

And furthermore, if the replacement neuron functions due to turing equivalent computation

Which it cannot do. Can. Not. You must know this. Surely you haven't gone all this time without understanding basic issues like this.

You can simulate a neuron using a turing equivalent computation. You can't replace one in a human nervous system.

it tells us, at the very least, that part of the "physical" function necessary for consciousness can be handled by turing equivalent computation -- of any kind.



What are you talking about? What issues are being dodged?

The only person dodging is you -- why can't you just answer the post?

Why does replacing a single neuron not destroy consciousness, while replacing all of them does?

Why?

Why?

Beats me. You're the one making the claim. I've certainly not said so.

 

[Drachasor]

I suppose that you could keep a rock on a totally stable platform at absolute zero, and minimise the amount of information being passed around. It would be much harder than say, unplugging a computer.

If you are dealing with a normal rock out in the sun, huge amounts of information are passing back and forth within it. How could there not be?

Remove the temperature gradient if you want, but the point is that if the forces are static, then the form of the object is stable and information inside it isn't changing.

 

[westprog]

Given a system A and B, "information" is behavior of A (or a subsystem of A) that B interprets as referencing some other system C.

Meaning, the stuff inside a rock just sitting there isn't "information" until some other system interprets it as such.

There's usually a weasel word in these supposed "objective definitions". Where is it now? Ah yes. "Interpret". What does it mean, in physical terms? Why, nothing at all! Can we provide an objective set of rules so some third party scientist in China can determine when B is "interpreting" the information, and when it isn't? No, it's just supposed to be obvious.

 

[westprog]

Remove the temperature gradient if you want, but the point is that if the forces are static, then the form of the object is stable and information inside it isn't changing.

"Information" has nothing to do with "form". Does the "form" of the computer change?

 

[Drachasor]

This is why when it's asserted that according to Church-Turing (which apparently I'm the only person to refer to) all the functions of the brain can be performed by a computation. Now, I know what that means - it means a program such as the one that calculates prime numbers. It does not mean a program such as the one that controls the water works. So I know that the statement is not true.

Have the computer simulate the solar system/galaxy/universe. All things are dependent only on other internal factors. It can setup an internal timing scheme that doesn't need to interact with the outside world. People can exist inside such a simulation, and they'd have consciousness.

 

[Drachasor]

"Information" has nothing to do with "form". Does the "form" of the computer change?

Sure it can. There are lots of definitions of information. I picked one of them, I could pick another if you want.

Do you even read your own links?

 

[Meed]

But now a person is involved in the detection process.

A cell doesn't need a person to be involved -- it does the detection all by itself.

Do you dispute this?

This is a trite objection, since it's obvious a person does not need to be included in the example. If a leaf and a boulder naturally fall into a rock then the rock can "detect the difference". It's irrelevant whether a human (or a bunny rabbit or a thunder storm) is involved.

A running computer is a closed system. An intelligent entity gives it an initial state and everything else takes place inside the computer.

If a computer is a closed system then it cannot detect the difference between a rock and another computer.

Your rock + chisel + celery system is not closed. The person must continue to contribute the whole time. They need to pick up the chisel, strike the rock, put the chisel down, pick up the celery, strike the rock, etc.

No.

If you want to rig up some stupid contraption where the person can set up an initial state and everything else happens automatically, like in a computer, then you will have departed from your example of a rock "detecting" anything and just proven my point. Congratulations.

No. The rock still detects differences between things in the sense that it responds in different ways physically to contact between different things. This is the same thing we mean when we say a computer detects differences between rocks and other computers. It responds (physically) in different ways to contact with different things in its environment.

 

[westprog]

A running computer is a closed system.

Really? What does it run on?

 

[rocketdodger]

Really? What does it run on?

What does your laptop run on, genius?

 

[rocketdodger]

Oh, I see. I just imagined the constant references to Church-Turing and the wonderful things that it supposedly proves. Just my imagination.

I've taken it on trust that you have some kind of knowledge and understanding of actual programming practice, but you've continually tried to demonstrate the opposite. It's not a matter of some obscure theoretical difference. It's a fundamental issue that real programmers working on real problems have to address.

Say, for example, a program to calculate the first hundred primes. Such a program would not be time dependent. It could take a microsecond, or a year, and it would produce exactly the same output. That kind of computation could be written by, say, a Pascal program.

Now say you want to write a program that switches a pump on when the level of a reservoir falls below one meter, and turns it on when it reaches four meters. Try to do that in Pascal. You cannot. The language specification won't allow for it. Pascal is designed to perform computations. That's not an accident. It's because Pascal (just like most programming languages) is designed to perform computations. It's a way of expressing a Turing machine. If you want to control the pump, and read the level of the water, and what's most significant, do it before the level of the water reaches five meters and overflows the reservoir, then you need to extend the language in some way with a way to interact with the environment in an asynchronous way.

This is why when it's asserted that according to Church-Turing (which apparently I'm the only person to refer to) all the functions of the brain can be performed by a computation. Now, I know what that means - it means a program such as the one that calculates prime numbers. It does not mean a program such as the one that controls the water works. So I know that the statement is not true.

If you are going to insist that it's the second type of program that you're referring to, then you have to abandon such theories as Church-Turing which refer to computation only.

You can continue to obfuscate and insist that asynchronous, event-driven programming is the same as Turing compatible computation, but that is demonstrably not the case.

The argument that if a neuron simulation calculated the proper state 1000 times too slow or too fast it wouldn't "work" with other neurons is both trivial and obvious.

Coupling.

If the whole thing is simulated, time in our frame is no longer relevant -- only time in the simulation. Church-Turing is correct.

If only part is simulated, then time in the simulation must couple correctly with time in our frame. Obviously this requires something in addition to a "Turing machine" just like any two systems require an interface in order to interact with each other.

You can simulate a neuron using a turing equivalent computation. You can't replace one in a human nervous system.

Durr.

Thats why I was careful to say "and a suitable interface" in every relevant statement of that post.

Try reading, westprog. It helps.

Beats me. You're the one making the claim. I've certainly not said so.

Well Al Bell said so, and you jumped in.

You contend that if we replaced a single neuron with a simulated one + interface, the brain would cease to be conscious?

 

[Drachasor]

No. The rock still detects differences between things in the sense that it responds in different ways physically to contact between different things. This is the same thing we mean when we say a computer detects differences between rocks and other computers. It responds (physically) in different ways to contact with different things in its environment.

Eh, so do people.

What's your point again?

 

[rocketdodger]

There's usually a weasel word in these supposed "objective definitions". Where is it now? Ah yes. "Interpret". What does it mean, in physical terms? Why, nothing at all! Can we provide an objective set of rules so some third party scientist in China can determine when B is "interpreting" the information, and when it isn't? No, it's just supposed to be obvious.

The biochemical subsystems of a cell that are involved with the central dogma of molecular biology "interpret" the sequence of molecules in a DNA strand.

The biochemical subsystems of a cell that are involved with intracelluar communication "interpret" chemicals in the environment.

Is that clear enough? Objective enough? Do you need more examples?

 

[Drachasor]

The biochemical subsystems of a cell that are involved with the central dogma of molecular biology "interpret" the sequence of molecules in a DNA strand.

The membrane receptors of a cell that are involved with intracelluar communication "interpret" chemicals in the environment.

Is that clear enough? Objective enough? Do you need more examples?

Much like how say...a Turing machine, interprets the string of symbols.

 

[rocketdodger]

This is a trite objection, since it's obvious a person does not need to be included in the example. If a leaf and a boulder naturally fall into a rock then the rock can "detect the difference". It's irrelevant whether a human (or a bunny rabbit or a thunder storm) is involved.

That is a special case. Normally rocks just sit there. That is the whole point.

I have said over and over that a rock can do many things -- what makes cells and computers different is that cells and computers do those things all the time, as the norm.

If a computer is a closed system then it cannot detect the difference between a rock and another computer.

I was obviously talking about the system of computer + whatever it is detecting.

The system of rock + chisel + celery is not capable of the same class of behaviors a computer + chisel + celery system is capable of.

No. The rock still detects differences between things in the sense that it responds in different ways physically to contact between different things. This is the same thing we mean when we say a computer detects differences between rocks and other computers. It responds (physically) in different ways to contact with different things in its environment.

So do cells.

What does that mean?

What does it mean that rocks, cells, and computers respond physically to contact with different things in their environment?

You tell me. Does this mean cells and rocks have the same behavior?

 

[rocketdodger]

Much like how say...a Turing machine, interprets the string of symbols.

Try to stay away from any man-made thing.

Westprog won't ever drop the stupid man-made argument. So try to stick to examples that have nothing to do with humans.

 

[Meed]

It's not a non sequitur. I defined information as far as it matters to the brain, which is at the heart of the discussion.

You didn't define it as far as it matters to the brain. This is what I think you must be referring to:

What does the brain do?

How does it interact with the rest of the body?

Treat it as a black box (initially).

Blood brings it what it needs to run. Nerves feed it information and output information (the blood does a little bit of that too).

We can easily define the information the brain receives in terms of nerve impulses and the bits of hormones and the like. All quite quantifiable.

It takes that stuff, processes it, and the output comes out. The output tells the body what to do in terms of information. Neurological studies on the brain further reinforce this view that it is an information processing center (complete with areas dedicated to certain kinds of processing). All this can be perfectly modeled in principle.

This isn't like a plane or a boat or water. Model the brain and you copy the most essential thing it does; process information.

I see no definition in there. "We can define the information the brain receives in terms of nerve impulses and the bits of hormones and the like" isn't a definition.

Is your definition of information "something that can be taken, processed and result in output"? If that's the case, then everything is information and all physical interactions are information processing.

BTW, why did you link me to the wiki for Shannon's information theory when I initially asked the question? I was expecting some sort of elaboration.

 

[Meed]

Could I take a crack at it? I think in most general terms, information is "not something else". It's weird, but much of it seems to be defined negatively; at least that's the way the nervous system seems to be set up, with surround inhibition and the like.

Under that definition everything is information, isn't it? There is nothing that is not "not something else".