Are You Conscious?

[westprog]

Ah, I see.

The fact that thoughts fly through your head an order of magnitude more rapidly than the diffusion of chemicals that any cells other than neurons use to propagate communication signals is entirely irrelevant, right?

I mean, when your hand is burned you feel it within 60ms or so despite the fact that it takes at least a second for any chemicals secreted into your blood by the cells at the site of the burn to reach the rest of your body, but who cares, right?

What good are numbers and actual measurements when they get in the way of our hypotheses, right? If reality disagrees with a hypothesis, we should force reality to change!

I thought the computational model was quite explicitly time independent. It's the physical model that allows for the possibility that consciousness might require certain time constraints. Turing machines don't have timing issues.

 

[!Kaggen]

The strong interpretation of Goethe's Metamorphosis of Plants demonstrates that that the technique of imaginative thinking as a form of introspection can be used to describe objective reality and hypothesize the homologous structures of plant organs. This same technique can be used to study consciousness.

This same technique can indeed be used to study consciousness. However, it gives answers that are now well established to be wrong.

Such as?

You asked Aku what else might be necessary for consciousness apart from computation. I have replied that consciousness is not just understood as a result of reasons, but the reason for results. In other words consciousness not only has necessary reasons, but also contingent results.

Evidence?

Ok, lets assume only computation is required. Computation is repeatable. Then an exact copy of a conscious computer is the same conscious computer. That is what you have been telling us, right. Therefore if I produce an exact copy of a human by cloning would I get the same person? No, behavior studies shows that the historical experience of the original person is a contingent part of his/her consciousness. Therefore even if computation is the necessary part of consciousness the contingent part (being the historical experience of the person) of consciousness is required to replicate a brain. Is this physically possible, no. Therefore the initial assumption is wrong and computation is not all that is required.

It requires imagination to predict the contingent results of consciousness.

Evidence?

Ask any person who makes a living from predicting the behavior of real people.

You missed the point.
Song writing develops the imagination.

In computers?

No, in those who want to build conscious computers.

 

[Belz...]

The problem is the wording of your analogy is as bad as, like, whatever.

The wording isn't "bad", it's just that you've just found that the analogy doesn't correspond 100% to the thing it was analogue to. This, of course, is true of all analogies, and if you had made an effort to understand how the analogy works rather than how it doesn't, we'd both have saved ourselves some time and effort.

 

[Belz...]

Nobody is putting forward a contrary view. I've already said so a few posts upthread.

Nobody is saying that the knowledge of the experiences gives one the experience. Knowing about millionaires doesn't make me one. I'm saying that the experience of pain IS pain. You're adding a useless layer to pain.

Any comment on this, Westprog ?

 

[Ichneumonwasp]

I originally said that approaching consciousness from the biological/neurological perspective was the way to go, and that computationalism is flawed because it ignores the significance of the physics that underpins consciousness. Everything I've said since then has been in defense of this statement since you seem to've taken umbrage at my suggestion that computation is not adequate to explain consciousness.



I already told you that my current point is not in reference to that hypothesis [though I don't rule it out], but to the fact that we need a solid understanding of what physical capacity allows our brains to produce consciousness.

I never argued against consciousness necessarily being instantiated in a physical system. I argued against your use of words. You made a few statements in defense of your position that were wrong -- namely that computation is only an abstract concept and the implication that if we can simulate what occurs in combustion on a computer that this implies that we can only simulate what a neuron does as well and the two senses of 'simulation' were somehow identical.

Both of these issues depend on equivocation over the words 'computation' and 'simulation'. The point of switching to 'duplication' was my attempt to highlight the equivocation.

While it is true that we use the word computation as an abstract concept and that almost every instance of computation we see in the real world is observer dependent, there is a special case of what has also been labelled computation -- namely the summation that neurons do -- that is an intrinsic property of neurons/neural networks.

So, to say that computation is only an abstraction is not correct since there is a physical process that does the same. The point behind current AI as I understand it is to recreate the salient feature of this process -- moving electrical charge through a network of nodes. The hypothesis is that a computer should be able to recreate consciousness if we follow the same wiring diagram as in a human brain and run electric charges through it -- though the engineering feat of doing what synapses do is going to be nigh on impossible, I think.

The same issue arises over the word 'simulation'. While some simulations are far from the original, others are so close to the process being 'repeated' that we can use the word 'duplicate', though strictly speaking that is not entirely correct either. The important point is that when it comes to artificial neural networks, we try to 'duplicate' the salient feature of what neurons do -- summate inputs and respond with a single output -- and that this is not the same use of the word 'simulate' as when we model physical processes in a computer -- like a simulation of combustion. Both examples use the same word, but they are different types of 'simulation' or 'duplication'. One represents a model that can help us understand a process (such as combustion) and the other is so close to the original that the actions performed have the possibility of reproducing the actual process under investigation.

Thats the central premise of computationalism and the position taken by all of the Strong AI proponents in this discussion. In their view, simulating a brain on any computer system is sufficient to produce consciousness. If this is not your position then why are you arguing against my objections to it?

See above. I'm arguing against the equivocations in your use of these words.

I said an operational definition is not possible until we have a scieintfic understanding of what consciousness is.

OK, I disagree, but if that's your stance, then fine. Of course we cannot have a final operational definition of consciousness until we understand it scientifically, but we can always work to refine the definitions that we do have based on our current state of knowledge and by 'playing the philosophy game'. That is what I was after in my thread and essentially what Pixy asked you.

When did I ever speak of "breaking down" awareness into "individual parts"? If you're going to continue basing your counter arguments on confabulated recollections we really have nothing to discuss.

I didn't say that you did. I said that you commented on my attempt to do so by saying the above -- that we can't get to any sort of operational definition, which requires understanding the 'parts' of consciousness. I disagree.

Alright. I see no reason to continue wasting my time trying to engage in a depth discussion with you when you're barely making a casual attempt to keep up with whats being argued. You've hardly even been following that I've said in the posts you're responding to. Your heart is clearly not in this.

Fine, have it your way; I have no particular desire to read insults hurled at me. But I caution you against continued equivocation with these words. And my comment above was directed at your earlier reference to your particular posting history. I'm sorry that I don't read all your posts. I don't read all of anyone's posts. That does not invalidate the criticism directed at your equivocation with these words, however.

I am also not at all certain that everyone with whom you are arguing holds the position that you think they hold. The general consensus seems to be that we have a chance to emulate consciousness in a computer system by re-enacting the movement of charge through a particular type of network. The old sense of 'computationalism' (like the old view of behaviorism), as I understand it, is dead, though there may be a few here who subscribe to that old view.

If by 'it's the physics' you simply mean the same thing -- understanding the way that electrical charge moves through the network to 'produce' consciousness -- then we have no basic disagreement except over the equivocal use of certain words. If you mean something more, then we do have a basic disagreement. We need to understand the networking, which is not a description at the level of physics, but at the level of biology. Of course, it would be great to understand this at the level of the physics, but there is nothing in the problem that screams that we could not arrive at an operational definition of consciousness with a mere biological understanding of the 'wiring diagram'.

 

[Ichneumonwasp]

I thought the computational model was quite explicitly time independent. It's the physical model that allows for the possibility that consciousness might require certain time constraints. Turing machines don't have timing issues.

I certainly don't keep up with the details of AI, but isn't that an old idea, not currently a part of models under investigation?

Neural processing is clearly time dependent.

 

[Dancing David]

Uncontrolled cellular growth, the invasion and destruction of adjacent tissue, the spread to other locations in the body via lymph or blood... is cancer a private behavior?

Yup, until made public by some means.

 

[AlBell]

I find it silly to believe that consciousness results from the electrical field over the brain because I know how to disrupt that field easily and it does not affect consciousness. I've done it repeatedly. I've seen it done by others. It doesn't affect consciousness.

How does one measure consciousness precisely enough to know that it didn't change under various external emfs?

If you are trying to argue for something other than this, then please let me know, because I'm not at all sure what you are arguing for. You call it a physical process but you haven't defined it for me. To what physical process other than what neurons are doing do you refer? And if you have no issue with a computer doing it, why are we even having this conversation?

Will a conscious computer need to be designed such that brain-like emf fields are also measureable? Or do you contend that what is measured by EEGs and the like is only supervenient fluff and unneeded?

 

[PixyMisa]

How does one measure consciousness precisely enough to know that it didn't change under various external emfs?

Simple: Use coarse measures and ramp up the external field strength, as with MRI, TMS, or standing near a radar dish.

TMS affects mental function by actually inducing currents in the brain. If consciousness were related to an EM field of some sort, your mind would shut off at field strengths orders of magnitude lower than what is actually required to get any effect.

Will a conscious computer need to be designed such that brain-like emf fields are also measureable? Or do you contend that what is measured by EEGs and the like is only supervenient fluff and unneeded?

Is fluff. Anyway, computers have much stronger EM fields than humans.

 

[PixyMisa]

I certainly don't keep up with the details of AI, but isn't that an old idea, not currently a part of models under investigation?

Neural processing is clearly time dependent.

Order-dependent at the very least; it's one big race condition.

Which you can duplicate perfectly well in a Turing machine, of course.

 

[nescafe]

Its not so much that fields lend special explanatory power to consciousness, per se, but that it makes sense to try and understand exactly what field processes are the sufficient correlates of consciousness.

We can make EMF fields form a physical instantiation of a universal turing machine. Done. Nothing special there.

There must be some objective means of discerning whether a given physical system is conscious, or capable of supporting consciousness, without having to rely on guessing, indirect inference, or intuition.

Why must there be an objective means? Pixy's metric of self-referential information processing aside, I rely on my messy subjective intuition to do so -- I know it when I see it.

What does it mean, in physical terms, for there to be a subject experiencing things "the redness of red", "the bitterness of bitter", or "the loudness of loud"?

Dunno, I am one of those people who think that looking for a unique physical explanation for subjective experiencing is a fool's game -- I think the brain is a computational engine, and that the Church/Turing hypothesis applies.

What is it thats having those experiences?

The conscious system of course.

Physically speaking, when does processed information become experienced?

Good question. Let's build a nonhuman conscious entity and find out, although I suspect that the question itself may be incoherent -- you are implying that there is a separate part of a conscious system that does the experiencing, rather than the system as a whole.

We each personally know that these phenomena are real; the scientific challenge is to identify what they are exactly. What I'm proposing is that what are referred to as qualia must be physically identifiable in theory atleast, if not in practice.

If the phenomena you refer to as qualia are real and substrate independent, there may not be a physically identifiable thing that uniquely correlates to them at all.

And the question still remains: Exactly how does consciousness fit into that milieu?

And the question still remains: Why is this a question for physics and not biology? What makes consciousness so special that we have to address it in terms of EMF field interactions directly, unlike every other biological process?

The problem with that reckoning is that neurons themselves are just part of a larger network of cells we call the body.

I do not think it is fair to characterize the rest of the body as a network in the same sense we characterize the nervous system as one.

They are all "wired" together, they all network and process information collectively, and they do so to a degree that dwarfs even the most advanced technological systems we have available today. Yet, for some reason, it is only the activity of a certain neural cells that seems produce consciousness, and even then only for limited time periods.

I do not agree (how does a liver cell summate it inputs and decide to fire or not?), but giving it to you makes your emphasis on looking at the field processes look even sillier -- we don't need the full power of QED to examine how neural cells operate and what makes them different from the other cells, classical chemistry and neurobiology will do just fine

Neuroscience has not yet established what it is about these particular cells that gives them the capacity to produce consciousness while the rest of our body's cellular network doesn't.

Because neurons have evolved over hundreds of millions of years to specialize in summing their inputs and using that as a decision to fire or not fire as part of as elaborate network at high speed, instead just using chemical diffusion to communicate with their closest neighbors.

We simply do not know what the necessary physical conditions are for consciousness so what in the world makes you believe that it's justified to assume that we can produce consciousness in any substrate just by employing network computation?

Because neural networks are a tighter necessary precondition than the Standard Model and GR are.

What makes you think that the current efforts of the field of AI to produce consciousness are anything more than a cargo cult enterprise?

Substrate independence of information processing (as proven by Church and Turing) coupled with the observation that we can emulate neural networks using universal turing machines.

 

[rocketdodger]

I thought the computational model was quite explicitly time independent.

Yes. But "independence" doesn't mean timing is irrelevant, it just means that an algorithm could be run at any rate relative to the outside world.

The relative timings within the algorithm are a totally different matter.

It's the physical model that allows for the possibility that consciousness might require certain time constraints. Turing machines don't have timing issues.

Since abstract turing machines don't exist, and every turing equivalent machine is physical, this isn't a problem is it.

yet another strawman from westprog...

EDIT: Furthermore, this has nothing to do with the point I was making, which is that if chemical diffusion happens slower than a thought reaches your mind it clearly cannot be responsible for said thought. Do you disagree with that?

 

[rocketdodger]

What's more, other cells do not form switched networks. It's just equivocation.

Actually, they do.

Sorry!

But Aku is still wrong, since the timescale of such communications is too high to account for thought.

 

[PixyMisa]

Actually, they do.

Sorry!

Hmm?

Example or link? It's not my field, so maybe this is something I missed completely, or it's something I know about but wasn't considering as such.

 

[rocketdodger]

How is it "woo" to suggest that consciousness must be understood in terms of physics, yet proposing that strong AI ideology is the way to the afterlife isn't?

Because the strong AI ideology is founded upon KNOWN physics while your hypothesis is explicitly defined to be dependent upon UNKNOWN physics.

That is woo.

If you would just say "I think we have all the fundamentals, we just need to figure out how they interact to make consciousness" then nobody would take exception. But you don't say that. You keep asserting that we are lacking some fundamental.

That is woo.

My bad. Its strait-up religious faith.

Ok, so again, can you make a logically valid argument to support this claim?

How is your hope in a computer afterlife any different than any other afterlife belief? You sit there accusing other people of "woo" and having ulterior ideological motives, yet the rationale behind your own arguments are not much different than that of the theists you like to decry. I don't have a problem with you having your own metaphysical beliefs but your hypocrisy is really off putting.

Well, first, I didn't form that ideology until after I received an education in computer science and neurobiology, and after I realized how much better the computational model predicts human and animal behavior than any other existing model.

Second, if you really think that the rationale behind my arguments, and those of Pixy, or yy2bggggs, or Paul, or I-wasp, or anyone else here who subscribes to the computational model, are "not much different than the theists I like to decry," then there is even more standing in the way of you understanding this issue than I thought.

It was you who turned this into a discussion about ulterior ideological motives when the original topic was on the role of physics in consciousness.

O I see.

So although I haven't dodged any of your questions, and I haven't ignored any relevant issues that anyone brings up, by virtue of the fact that I made comments about the behavior of forum members I am "changing the subject?"

I wasn't aware that it was possible to "change the subject" while making posts explicitly pertaining to the subject, but whatever.

I've never proposed such a thing. All I'm saying -- all I've been saying -- is that our knowledge concerning consciousness is incomplete and that we need to understand it a lot better than we do before we can propose ways to reproduce it artificially. The only point where I really differ from you is that I think consciousness is a substrate dependent phenomenon rather than an abstract computational process.

But you just don't get it -- there is no such thing as a substrate dependent phenomenon in the sense that you are thinking.

There are no known physical phenomena that cannot be duplicated on any substrate featuring the required attributes.

This is trivially true because every physical system is merely a collection of fundamental particles -- the only difference between any two systems is the number of particles and how those particles are arranged. So the most a phenomena can require is just a specfic configuration of specific types of particles. This is a scientific fact.

Thus the assertion that consciousness is a neuron phenomenon, rather than a behavior-of-neuron phenomenon, is not grounded in known science. Because neurons are merely the same types of particles as everything else in the universe and the only difference between a neuron and a rock is how those particles are arranged, which affects how they behave.

Saying that there is something special about neurons that isn't just a behavior of particles is outright dualism. And woo.

 

[Ichneumonwasp]

How does one measure consciousness precisely enough to know that it didn't change under various external emfs?

The same way we always do with consciousness. Ask the person questions.

Will a conscious computer need to be designed such that brain-like emf fields are also measureable? Or do you contend that what is measured by EEGs and the like is only supervenient fluff and unneeded?

I don't see any way we could see anything like EEG activity with a computer. EEGs don't 'sample' all brain activity but only summated post-synaptic potentials, and only those that are properly oriented so that they can summate for the recording. Granted we can put electrodes on the surface of the brain or depth electrodes in the brain.

There is no way to answer a question such as "is EEG activity just supevenient fluff" since it exists only as a measure of brain activity. The emf field certainly doesn't seem to do anything, it's just something that we can measure and use clinically (and sometimes investigationally), and we can interfere with it in many different ways with no clear effects on the person, which is why I say that it doesn't seem to do anything. So, if that is what you mean by fluff, then sure, it's fluff.

 

[rocketdodger]

Hmm?

Example or link? It's not my field, so maybe this is something I missed completely, or it's something I know about but wasn't considering as such.

The working of any hormone is a perfect example. Compounds are released by certain cells (usually in the brain) and travel throughout the bloodstream. When receptors for those compounds are filled on cellular surfaces, those cells change their behavior I.E. are switched.

And yes that is only a single switch deep, but there are mechanisms (particularly in growth, like Aku suggested) that are much more complex. In fact initial fetal development is determined in a large part by an initial chemical gradient set up in the egg after fertilization, before the first division. They have done fascinating research about this in fruit flies, in particular.

Another example is what happens when a blood vessel is injured. Cells near the injury send out chemical signals, platelets clot the wound until a certain concentration is met and negative feedback starts to occur and shuts the process down.

To be fair, there is nothing nearly as complex as the switching that occurs in neural networks. But biological life uses every available means to communicate between cells and cause switching, slow chemical diffusion included.

I will try to find some links later.

 

[westprog]

I certainly don't keep up with the details of AI, but isn't that an old idea, not currently a part of models under investigation?

Neural processing is clearly time dependent.

Yes, it obviously is, which is why the position of Pixy and Rocketdodger that consciousness is pure Turing, with no time dependencies, is not supported by the facts. Though since RD has now brought the issue of timing in, that might well mean a different position.

However, if consciousness is purely due to the operation of a Turing machine, then it is timing independent. No point in quibbling with Aku and me. We are looking for a physical solution, and timing may well be critical. If consciousness is physical, then it might well not happen at all at half speed - and hence the slower interactions of the body will simply not be relevant. If consciousness is computational, the speed of the interactions simply doesn't matter - except in the interaction with the external world.

I get the constant impression of a theory being made up as it goes along.

 

[westprog]

Actually, they do.

Sorry!

But Aku is still wrong, since the timescale of such communications is too high to account for thought.

You've been arguing for a Turing model of consciousness throughout. The Turing model is time independent.

 

[PixyMisa]

The working of any hormone is a perfect example. Compounds are released by certain cells (usually in the brain) and travel throughout the bloodstream. When receptors for those compounds are filled on cellular surfaces, those cells change their behavior I.E. are switched.

And yes that is only a single switch deep, but there are mechanisms (particularly in growth, like Aku suggested) that are much more complex. In fact initial fetal development is determined in a large part by an initial chemical gradient set up in the egg after fertilization, before the first division. They have done fascinating research about this in fruit flies, in particular.

Okay, sure. Definitely communication and co-ordinated activity, but not really a switched network.