Has consciousness been fully explained?

[Robin]

Now you're just being mean.

Do you mean me?

I wasn't trying to be mean.

 

[PixyMisa]

Well, by "mean" I mean "thoroughly deconstructing someone's argument".

 

[my_wan]

I was tempted to say that for you Pixy, but didn't on principle that it's dangerous to assume to know how somebody else meant something.

 

[Malerin]

Now remember, I am not asking if the simulation is conscious, I am asking if the simulation will produce the external behaviour that we observe in a conscious human. A lot of people here seem to avoid that question.

Will it, for, example, claim it has a Sofia?

It seems to me that if the matter in our brains acts as physics says it should, then it will make that claim.

This goes back to Piggy's point that if you arrange a bunch of neurons (matter) in a certain way, it doesn't just "act as physics says it should". Something new enters the picture: subjective experience (sofia).

As Westprog points out, beings like us report sofia events because we are experiencing sofia events. Pain feels bad is a true statement for us. We may be compelled by physics and biology to make such a statement, but it's still a true statement. A simulation can be programmed to report a sofia event, but it will be impossible to prove it's actually having a subjective experience.

Why/how sofia arises after a certain level of neural complexity is a mystery.

 

[Meed]

Are you kidding?

Enough with the incredulity.

This isn't some trick question. Rocks, unless they are melting or breaking apart after landing from a fall, just sit there. They might heat up in the sun, but that is a linear change. With few exceptions rocks display very little internal state change due to small changes in the environment around them.

Are they less susceptible to change than cells? Sure. But you can't classify things as either being susceptible or non-susceptible based on degree of change. That would be like defining the sun as hot and the human body as cold based on the sun being considerably hotter than the human body. However, both contain heat and a cutoff point for whether something can be considered "hot" or not has not been provided and justified.

As for the linearity of the change, that is a better strategy, because you're attempting to provide a distinction that isn't simply based on differing degrees. However, I'm not sure that a rock doesn't display non-linear changes to its environment (I'm not a material physicist). For example, doesn't a breach of the melting/crystalyzing point constitute a non-linear change in respect to the objects' behavior?

Because my definition of a "computation" has to do with such non-linear state changes. Large changes in internal state as a result of small changes in external state.

The subsystems of a cell compute -- alot. That is how a cell works. The external environment changes, even a little -- such as a chemical messenger molecle binding with a receptor on the cell surface -- and a whole slew of things happen in the cell as a result, because the little subsystems are busy reacting in non-linear ways to not only the environment but also the radically changed behaviors of their neighbors. Cascades of computations that are related to each other take place. And the cell does something different that what it was doing before.

The subsystems of a rock -- if it has any -- don't really do that. Yeah small groups of molecules in the mineral, or even the particles in a molecule, might exhibit nonlinear behavior all the time. But it is random with regard to the rock as a whole. The rock doesn't change as a result of those accidental computations in random molecules. You don't see a mineral pocket behaving differently because of X instead of Y, and you don't see another mineral pocket behaving differently because of the behavior of that first one, and so on. A rock just sits there.

See the difference?

Of course there's a difference. But I don't see how you've made the case that computation should be defined in terms of non-linearity.

And once you do see the difference, ask yourself whether the behavior of a transistor is something you can get out of a rock that just sits there.

I didn't claim a rock can behave as a transistor. I think that comes down to shape and ability to act as a semiconductor (i.e. material composition).

 

[my_wan]

This goes back to Piggy's point that if you arrange a bunch of neurons (matter) in a certain way, it doesn't just "act as physics says it should". Something new enters the picture: subjective experience (sofia).

And why do you seem to presume physics is not what sophia is? The emotional content of our sense of sophia, which seems to provide our foundational motivations and gives us our sense of awe about sophia, predates our intellectual capacity to ponder it. I liken emotions to a more open ended version of instincts, more malleable to our circumstances and intellectual state. Though the intellectual control still seems a bit primitive, it seems to be steadily improving since the early 17th century in a fractal manner. Had the 100 millions deaths due to war and violence of the last century been in line with pre-17th century, that number would be more like 2 billion. Deification appears as if it was our first attempt at acquiring intellectual control of it. Funny that the Bible never made an issue of the uniqueness of our emotions, but rather our capacity to abide by provided laws.

Have you ever wondered what sophia would feel like without the emotional associations? The concept of ego, being trumped, pwned, etc., wouldn't exist. Even the notion of using our intellectual skills for recreation. Yet this same primitive motivational control mechanism is the very thing that gives you the illusion of a visual field you don't have, a connection to the past even when it's not valid, a sense of physical reality that seems uniquely real when it's merely a valid perspective (sometimes), a sense of awe and ownership of the intellectual skills you do have, and other perceptual qualia.

I've experienced myself, and retain memories of, lacking these connections. Even with the intellectual skills intact, I didn't even have the sophia awareness you have in dreams. The only remaining motivation seemed to be from a memory of connections you no longer have.

Why/how sofia arises after a certain level of neural complexity is a mystery.

Yes, it's still a mystery. The intellectual side is almost trivial compared to the motivational control system. I suspect that it's our emotional systems that both produce the most profound aspects of it, and our blindness to the actual structure. Much like certain illusions are effectively impossible to see through even when you know how it's done.

 

[Meed]

People who confidently predict that a computer simulation will definitely share all the properties of the thing it's simulating are too optimistic to be working in health and safety.

A simulation is, by definition, not the same as the thing it is simulating. Otherwise we wouldn't call it a simulation, but a replication.

It's possible that a computer based simulation of the brain could be created that approximates the functionality of the brain. But we should note that the processes carried out by the simulation are actually going on at the hardware level. That is, in the interaction of the billions (with present technology?) of CPUs and other computer hardware. The computer code is an input that causes the hardware to react in the intended manner. But the actual processing is going on in the hardware. So a computer simulation is a physical system, not just an abstract algorithm.

 

[Robin]

This goes back to Piggy's point that if you arrange a bunch of neurons (matter) in a certain way, it doesn't just "act as physics says it should". Something new enters the picture: subjective experience (sofia).

But whether or not something new comes into the picture - it acts as physics says it should right?

As Westprog points out, beings like us report sofia events because we are experiencing sofia events.

So you would think. So if the simulation of those same brain components doesn't experience Sofia events then it should not report Sofia events, right?

Pain feels bad is a true statement for us. We may be compelled by physics and biology to make such a statement, but it's still a true statement.

But not a true statement for the simulation if it is not experiencing pain.

But it is simulating the same biology that compels us to report pain when we are in pain.

So if the neurons are not enough to explain why we feel pain, why would they be enough to explain why we report pain?

 

[westprog]

E
As for the linearity of the change, that is a better strategy, because you're attempting to provide a distinction that isn't simply based on differing degrees. However, I'm not sure that a rock doesn't display non-linear changes to its environment (I'm not a material physicist). For example, doesn't a breach of the melting/crystalyzing point constitute a non-linear change in respect to the objects' behavior?

What do Intel and AMD do to the rock in order to make it into a computer chip? The first thing they* do is to extract the silicon and make it as linear as possible. The non-linear chaotic structure of the rock is precisely what makes it less useful. A linear response is what is needed from a device.

Yes, the device is then made capable of a tiny degree of well-controlled non-linear behaviour - which is still far less than the non-linear distribution of heat, say, in the highly complex and unpredictable and unknown internal structure of the rock.

The next aim is to make the changes in internal state as small as possible. The computations that took place in early computers resulted in huge changes in internal state, reflected in enormous consumption of energy. A modern computer aims to change its internal state as little as it possibly can in order to perform the computation. The aim is to have an effect that is nearly, but not quite, undetectable. A computer that uses half the energy is a better computer, because it's internal state changes less.

For most devices that humans make, the aim of the design is to ensure a linear response. There also has to be the on-off option. That is as far as the non-linearity goes. The reason that a chisel is useful, say, as opposed to a random chunk of metal that's lying around, is that it provides a linear distribution of force. Hit it twice as hard, you exert twice the force at the end of the chisel. The lump of metal cannot be relied on for this. Nature falls into non-linear behaviour all the time. Producing linear effects requires enormous effort.

 

[westprog]

But whether or not something new comes into the picture - it acts as physics says it should right?

So you would think. So if the simulation of those same brain components doesn't experience Sofia events then it should not report Sofia events, right?

As I've said before - the simulation will report seeing and hearing things that we know are not present. We don't believe it is seeing a real tree. We do believe that we are seeing a real tree.

The simulation is a pretence. We don't regard anything it reports as reliable. So when it claims to have experienced a sensation, we cannot view that as indicating that there is anything there.

Of course, this is piling one hypothetical situation upon another. But at least if such a computer simulation were to be built, we could quickly tell exactly why it claimed, or did not claim to be conscious.

But not a true statement for the simulation if it is not experiencing pain.

But it is simulating the same biology that compels us to report pain when we are in pain.

So if the neurons are not enough to explain why we feel pain, why would they be enough to explain why we report pain?

 

[PixyMisa]

What do Intel and AMD do to the rock in order to make it into a computer chip? The first thing they* do is to extract the silicon and make it as linear as possible. The non-linear chaotic structure of the rock is precisely what makes it less useful. A linear response is what is needed from a device.

No.

Yes, the device is then made capable of a tiny degree of well-controlled non-linear behaviour - which is still far less than the non-linear distribution of heat, say, in the highly complex and unpredictable and unknown internal structure of the rock.

No.

The next aim is to make the changes in internal state as small as possible. The computations that took place in early computers resulted in huge changes in internal state, reflected in enormous consumption of energy. A modern computer aims to change its internal state as little as it possibly can in order to perform the computation. The aim is to have an effect that is nearly, but not quite, undetectable. A computer that uses half the energy is a better computer, because it's internal state changes less.

Yes.

For most devices that humans make, the aim of the design is to ensure a linear response.

No.

There also has to be the on-off option. That is as far as the non-linearity goes. The reason that a chisel is useful, say, as opposed to a random chunk of metal that's lying around, is that it provides a linear distribution of force. Hit it twice as hard, you exert twice the force at the end of the chisel. The lump of metal cannot be relied on for this. Nature falls into non-linear behaviour all the time. Producing linear effects requires enormous effort.

No.

 

[Malerin]

But whether or not something new comes into the picture - it acts as physics says it should right?

What does physics have to say about how my subjective experience should act? What laws does it have to follow? Gravity? Speed of light?

So you would think. So if the simulation of those same brain components doesn't experience Sofia events then it should not report Sofia events, right?

No, it will report whatever it's programmed to report.

But not a true statement for the simulation if it is not experiencing pain.

Right

But it is simulating the same biology that compels us to report pain when we are in pain.

Key word: simulating

So if the neurons are not enough to explain why we feel pain, why would they be enough to explain why we report pain?

Personally, I don't think they are. I was assuming Piggy's POV: we may eventually know enough about neurons and their arrangements and interaction to explain subjective experience, but we're not there yet. Simple chat-bot simulations programmed to report sofia don't tell us anything.

 

[my_wan]

Here's a "Robot Scientist" system that has discovered 12 new gene functions. Some of which has previously been discovered, so the robot independently verified some of that 12. The system uses abductive reasoning.

Here is a more detailed description, which includes in the discussion the use of the label "Robot Scientist".
http://www.aejournal.net/content/2/1/1

Here's a news article about it.
http://dsc.discovery.com/news/2009/04/02/robot-scientist.html

That system lacks the intellectual continuity needed for complete autonomy. Yet I wouldn't think intelligence alone will ever produce a sense of sophia, regardless of how intelligent or autonomous. The emotion, qualia and motivational control, aspect is at least as complex as the intelligence issue.

We a priori expect continuity in all things, and it forms the "truther" foundation of a lot of the new age crap. It's the same thing with CT'ers, in which continuity is provided through secret organizations. Much of the continuity we perceive in the structure of the room we are in is provided by our minds as a form of qualia. Luckily it tends to corresponds to a valid structural perspective, usually. Our qualia becomes much more suspect in issues of science and truth.

CT'ers, new agers, and religious freaks, feel this 'belief' continuity as a qualia, much like we see a square box even when we view it at an angle and squareness isn't explicitly in the sensory data. This qualia gives it a guttural sense of truth. Belief, in the woo sense, then feeds and is defined by qualia. The notion of a loss of continuity at death is more than some people can accept. Whatever we accept as truth based on sensory qualia is highly suspect, this includes our qualia sense of sophia.

 

[AlBell]

"Whatever we accept as truth based on sensory qualia is highly suspect, this includes our qualia sense of sophia. "

So true, and none of us have ever demonstrated we have any other connection between "me" and 'all else' other than that link. Some photon interactions, and a bit from gravitons.

 

[my_wan]

"Whatever we accept as truth based on sensory qualia is highly suspect, this includes our qualia sense of sophia. "

So true, and none of us have ever demonstrated we have any other connection between "me" and 'all else' other than that link. Some photon interactions, and a bit from gravitons.

In principle, an AI with a real sense of qualia, based on the same principles ours is, would be just as real if that AI's sensory data was completely defined by a virtual world. However, in principle it would be possible for a scientific method to extract out that abstraction layer, if the virtual tools, program access, were not artificially restricted at a level above bitwise manipulation. The biggest roadblock would be the qualia of that AI. The AI "believers" would go absolutely bonkers over such a notion.

Can you imagine the power if you learned to maniplulate the world to the degree a programmer can manipulate a virtual world? Conservation laws would take some very weird forms.

 

[rocketdodger]

It's not possible to calculate the behaviour of three bodies and their gravitational attraction. It's possible to approximate it using numeric methods.

http://en.wikipedia.org/wiki/Three-body_problem

Approximating to an arbitrary degree of precision is not considered "calculating?"

 

[rocketdodger]

Enough with the incredulity.

Stop saying absurd things, then.

Are they less susceptible to change than cells? Sure. But you can't classify things as either being susceptible or non-susceptible based on degree of change. That would be like defining the sun as hot and the human body as cold based on the sun being considerably hotter than the human body. However, both contain heat and a cutoff point for whether something can be considered "hot" or not has not been provided and justified.

Well, a rock rolling down a hill and a mouse running down a hill are both moving collections of particles. So any classification regarding their "movement" is going to be based only on the various degrees of their movement.

Yet I don't know of anyone who would claim a live mouse rolls and a rock runs. Why is that?

As for the linearity of the change, that is a better strategy, because you're attempting to provide a distinction that isn't simply based on differing degrees. However, I'm not sure that a rock doesn't display non-linear changes to its environment (I'm not a material physicist). For example, doesn't a breach of the melting/crystalyzing point constitute a non-linear change in respect to the objects' behavior?

Yes, the transition from solid to liquid is definitely the kind of non-linear change I am speaking of. That is why I have always said, from day one, that you could build a computer out of melting rocks.

But a rock sitting in the sun is not displaying such behavior. Neither is a cell that is not doing anything -- if it was frozen, for example. But when cells are not frozen, or in some kind of stasis, they are doing things that a rock simply does not do on a regular basis. That is why you see living organisms running around talking to each other being made of cells and not rocks. Someday they might be made of transistors. Still not rocks.

Of course there's a difference. But I don't see how you've made the case that computation should be defined in terms of non-linearity.

It is not just non-linearity. It is non-linearity mapping to a completely different behavior. That is what all that stability mumbo jumbo I explained was pertaining to.

Here is the simple explanation:

You have system A and B, that interact with each other. For a wide range of states of A, B remains at a local minima within the state transition space of B. That is, regardless of what state B is put in by a change in A, it will converge over time (all else being equal) back to this local minima of sorts. That is where "stability" comes into play -- there needs to be somewhat reliable local minima in the state space.

At a cetain point, the transitions of the states of A become non-linear with respect to B. When A transitions from state aX to aY, it puts B in a *different* subspace with a *different* local minima of sorts. That is, bX converges to a *different* local minima than bY.

In such an example, system A might be the environment and B might be the melting rock.

Now I call such a thing a computation because conceptually system B has partitioned the universe into two sets -- one, the set of all states where the local minima of B is downstream of bX, and the other, the set of all states where the local minima of B is downstream of bY. Any other system can then use system B to determine if the state of the universe is in the first partition or the second. A computation has taken place, regardless of whether it is used somehow or not.

"Stability" comes into play in two places. First, B needs "stable" local minima with respect to A. Second, B needs "stable" local minima with respect to any system that would use the results of B's computation -- such a third system would require some way to itself determine which state B is in, similar to how B needs a way to determine which state A is in. Note that in practice this is taken care of by the laws of physics.

 

[rocketdodger]

But we should note that the processes carried out by the simulation are actually going on at the hardware level. That is, in the interaction of the billions (with present technology?) of CPUs and other computer hardware. The computer code is an input that causes the hardware to react in the intended manner. But the actual processing is going on in the hardware. So a computer simulation is a physical system, not just an abstract algorithm.

There is only one person here who has ever said anything to the contrary. I won't name names.

 

[Meed]

Stop saying absurd things, then.

You laid out your proclamation of incredulity toward disagreement well before I even entered the topic. From your first post introducing your attempt at a physical definition at computation:

"the following brief should be enough to convince anyone that respectably has business being here discussing it in the first place."

Since you've pre-labeled all disagreement as absurd, it seems the only way not to "say absurd things" in your view would be to agree with whatever you say.

Well, a rock rolling down a hill and a mouse running down a hill are both moving collections of particles. So any classification regarding their "movement" is going to be based only on the various degrees of their movement.

Of course, unless the cutoff point is made between movement and complete stillness. I'm not sure what your point is.

Yet I don't know of anyone who would claim a live mouse rolls and a rock runs. Why is that?

I'm not sure what your point is or how that relates to anything I said.

Yes, the transition from solid to liquid is definitely the kind of non-linear change I am speaking of. That is why I have always said, from day one, that you could build a computer out of melting rocks.

So we agree that a rock is capable of both stability and non-linear changes in reaction to the environment.

But a rock sitting in the sun is not displaying such behavior. Neither is a cell that is not doing anything -- if it was frozen, for example. But when cells are not frozen, or in some kind of stasis, they are doing things that a rock simply does not do on a regular basis. That is why you see living organisms running around talking to each other being made of cells and not rocks.

And we agree that cells do "things" that rocks do not do.

It is not just non-linearity. It is non-linearity mapping to a completely different behavior. That is what all that stability mumbo jumbo I explained was pertaining to.

Different than...?

Here is the simple explanation:

You have system A and B, that interact with each other. For a wide range of states of A, B remains at a local minima within the state transition space of B. That is, regardless of what state B is put in by a change in A, it will converge over time (all else being equal) back to this local minima of sorts. That is where "stability" comes into play -- there needs to be somewhat reliable local minima in the state space.

Can you provide an example?

At a cetain point, the transitions of the states of A become non-linear with respect to B. When A transitions from state aX to aY, it puts B in a *different* subspace with a *different* local minima of sorts. That is, bX converges to a *different* local minima than bY.

In such an example, system A might be the environment and B might be the melting rock.

But what is the "local minima" in this example?

 

[Meed]

There is only one person here who has ever said anything to the contrary. I won't name names.

Do you mean Robin?