Argument that all computable universes are real

[StatusX]

argument that all computable universes are "real"

I haven't been on this forum in a while, and I'm sure this kind of thing has been talked about many times, but I thought I'd bring it up again so I could discuss it with you guys. Here's my argument.

Start with a human being named John. He is conscious, experiencing the world around him. In a certain sense, his experiences are all he knows to be real, although he believes there is a real world out there. Let’s assume he is right. Now let’s consider slowly replacing neurons in John’s brain with computer chips which carry out the same exact function. Would his experiences change? There doesn’t seem to be any essential difference between protein-based and silicon-based information processors, so let’s assume his experiences do not change. Now we feed input to his new brain that reflects a made-up, computer simulated world. Again, if this simulation were good enough, he would not know the difference. At this stage we basically have one big program running on a computer. Now let’s consider slowing down this computer. Since his perception of time is tied to processes in his brain, he would presumably not notice any difference.

Now, this computer is essentially just performing calculations, and if it’s going slow enough, these calculations could just as easily be done by a person. So let’s slowly replace this computer by a human operator. At first the computer could output some calculation that needs to be done, the operator would do it, and feed it back in. Eventually, all the calculations could be done by a human operator, perhaps using a hand-operated Turing machine with a gigantic piece of tape. It will probably take billions of years to run just a second of simulation time, so we will require many generations of operators, this doesn’t matter, John’s experience should be the same. Now let’s assume these operators get really good at what they’re doing, and rather than perform every step, they start noticing patterns, and can skip steps. Does John’s experience change? Does it skip ahead when the operator skips steps? What if the operator still goes through these steps in his head, so that they at least occur somewhere? What if he doesn’t? What if at some point he gives up? There’s still a well-defined sequence of states that would occur if he had continued, although they never end up actually being written on the tape. Does John ever experience these states?

At a certain point it starts to seem like the actual writing on the tape is not the important thing, what’s important is the program that’s being run. And it’s not important how it’s being run, on what machine, or how fast. All that’s important is the mathematical structure of the program itself. Thus the mere platonic existence of a program which implements a conscious observer seems to be sufficient for such an observer to experience things. There remains the important question of what precisely defines a conscious observer in a program, but certainly anything resembling human beings in this universe should qualify, and possibly things much simpler.

I’ll lay out the argument in a series of steps, and you can tell me at what point you think it breaks down:

1. A human brain can be slowly replaced by a computer program, and the subjective experience would not significantly change.
2. The input to this computer could be yet another computer program, so that the entire system is a program running on a computer, and still, there would be conscious experience.
3. Once we have such a program with conscious experience, the system used to implement it is not important, even a person manually operating a turing machine would suffice.
4. Since the implementation is not important, one might argue all that is important is the structure of the program itself. Even if it’s not actually running somewhere, there is still a subjective experience.

Potential objection: What if the operator makes a mistake, or even purposely modifies something against the rules of the program? It seems John would experience this, even though it doesn’t follow from the mere platonic existence of the program. All I can think of in response to this is that we can still imagine the operator-machine-program system itself is part of a more complicated program. John might wonder whether he is in the simpler program (where rules are never broken) or the more complicated, imperfect simulation. A priori, the former seems more likely by something like Occam’s razor.

 

[apeiron]

4. Since the implementation is not important, one might argue all that is important is the structure of the program itself. Even if it’s not actually running somewhere, there is still a subjective experience.

Don't you agree that you are just illustrating the limits of the computational paradigm using a reductio ad absurdum argument?

The real world is dynamical, developmental. Computation is a world designed to be without time, without dynamics or development.

Computation can be used to simulate dynamics and development with varying degrees of coarse grain crudity. So in practice, computational models can be useful.

But you are stripping down the paradigm to expose its unreal presumptions, showing the fundamental qualities that computation lacks.

This is why computational arguments make for bad metaphysics. Get down to basics and it is easy to see that computational principles do not encode some of the basic aspects of reality. Such as entropy gradients, such as indeterminancy, such as dynamical self-organisation.

 

[StatusX]

Don't you agree that you are just illustrating the limits of the computational paradigm using a reductio ad absurdum argument?

Maybe. Can you elaborate on how this argument would run?

In any case, I agree, it is more than possible that our universe cannot be simulated on a Turing machine. However, I still believe (hope?) that it is described in some mathematically precise way. Maybe everything follows from a simple set of laws, or maybe it turns out to be quite complicated and inelegant, but I still believe there is some logic to it, that there are rules that cannot be broken. Maybe "computable" is too strict a word, and we should allow a larger class of mathematical structures. This would hopefully address some of your other objections as well.

 

[apeiron]

Maybe. Can you elaborate on how this argument would run?.

I was saying you just ran the argument.

In any case, I agree, it is more than possible that our universe cannot be simulated on a Turing machine. However, I still believe (hope?) that it is described in some mathematically precise way. Maybe everything follows from a simple set of laws, or maybe it turns out to be quite complicated and inelegant, but I still believe there is some logic to it, that there are rules that cannot be broken. Maybe "computable" is too strict a word, and we should allow a larger class of mathematical structures. This would hopefully address some of your other objections as well.

Of course there are other descriptions possible. It is just important to understand what computation is not when it comes to making metaphysical arguments.

The value of the Turing machine is precisely that it is such a stripped down statement. You can see what must follow from its axioms. And also what is left out by those axioms.

It is not a bad model. Just a model. That some people may use badly.

 

[StatusX]

I was saying you just ran the argument.

What I mean is, what exactly is the absurd conclusion? And what does it imply, ie, which of the four steps in the argument are wrong? Are you saying human consciousness can't be simulated on a Turing machine? It doesn't seem necessary to have all the machinery of QM and GR to simulate a human brain, even a deterministic classical simulation (which a Turing machine could surely do, at least approximately), should do the trick. Or would you disagree with this?

 

[apeiron]

What I mean is, what exactly is the absurd conclusion?

As I cited in the first post, it is...

all that is important is the structure of the program itself. Even if it’s not actually running somewhere, there is still a subjective experience.

Clearly in some Platonic sense, the future state of a Turing machine is already fully determined and so "exists". The software is "everything that is real" so far as the paradigm goes. Running the program or not running the program is all the same so far as what the future state might be.

But meanwhile back in the real world, a Turing machine actually has to be run. There has to be some hardware it sits on. In fact the Turing machine reflects this inescapable fact. You need the tape, you need the gate. Otherwise nothing does exist except some fantasy about a series of marks and absences of marks - a pattern of information that floats about in Platonia.

The Turing argument is that "if you have this minimal contact with reality's material dynamics", then you can have a universal computational device. But in your argument, you want to throw away even the gate, and just have the tape. And then even just a notion about the information represented by the tape (as having an infinite tape is going to still take some physical explaining!).

You could also tackle the legitimacy of your very first presumption...

There doesn’t seem to be any essential difference between protein-based and silicon-based information processors, so let’s assume his experiences do not change.

I know enough about chips and neurons to beg to differ. But the problem is that arguments at this level can easily be turned into a "mere hardware implementation issue" because the computationally-minded are indeed already believers that whatever essential thing they are interested in, exists and is "real" at the level of software.

There are of course computational-looking aspects to neurons. We can model neurons in those terms with some success. But that is not an ontological argument, just an epistemic one.

To actually make a legitimate argument here, you would have to start by saying "we know that chips and neurons are functionally identical in these respects which are in turn known to be critical to consciousness in biological systems".

You just want to assume no essential difference. I would say you have to justify that by specifying the similarities known to be critical to reproducing biological intelligence.

 

[AlephZero]

A proof that all computable universes are real does not imply that computable universes exist, any more than a (logically correct) proof that all square circles have pink feet implies that square circles exist.

 

[apeiron]

A proof that all computable universes are real does not imply that computable universes exist, any more than a (logically correct) proof that all square circles have pink feet implies that square circles exist.

In what sense is something real if it does not also exist? Things can really not exist, and also merely potentially exist, but only by virtue of some greater context that itself is really existent. So I don't get the point of what you mean here.

Likewise the comment about square circles with pink feet. If B is a predicate of A, then you must have A (upon which B can be predicated)?

 

[Pythagorean]

In standard logic:

if p -> q

The above statement is always true when p is false.

 

[apeiron]

In standard logic:

if p -> q

The above statement is always true when p is false.

Yes, but this is not exactly an if/then statement is it?

"If x is real then x exists."

I was asking how it wasn't a tautology. The statement looks reversible. And you know that what is forbidden is for p to be true and q false.

So "If x exists then x is not real"? Doesn't compute.

 

[StatusX]

But meanwhile back in the real world, a Turing machine actually has to be run.

The Turing machine has to run if we want to know its state at some future point. But its future state is determined, independent of our knowledge. In any case, I'm not sure you accept my premise that a consciousness could be simulated on a Turing device. If you do accept it, we can continue down this path of whether the simulation has to actually run. But it's a moot point if you don't. Which brings me to...

I know enough about chips and neurons to beg to differ.

I'm certainly not an expert on neurons (or computer chips for that matter). I do know something about physics, however, and I do not believe any biological process requires anything beyond essentially classical physics to operate. Sure there are chemical reactions which cannot be understood without QM, but these could be hard-coded in. I don't think quantum superposition could be maintained at the scales and temperatures in a human brain, rather, we can model the neurons as a complicated, but classical, system of molecules. And in any case, it is probably not necessary to simulate every last detail of a neuron to replace it, one can probably do a reasonably good job with a fairly simple set of rules (which could consist of thousands, but still much less than a mole, of rules) for converting inputs to outputs.

 

[Pythagorean]

I'm certainly not an expert on neurons (or computer chips for that matter). I do know something about physics, however, and I do not believe any biological process requires anything beyond essentially classical physics to operate. Sure there are chemical reactions which cannot be understood without QM, but these could be hard-coded in. I don't think quantum superposition could be maintained at the scales and temperatures in a human brain, rather, we can model the neurons as a complicated, but classical, system of molecules. And in any case, it is probably not necessary to simulate every last detail of a neuron to replace it, one can probably do a reasonably good job with a fairly simple set of rules (which could consist of thousands, but still much less than a mole, of rules) for converting inputs to outputs.

Even in a "computational atoms" view, carbon is much more dynamical than silicon.

But more than that: silicon in computers are not even used at all like carbon atoms in life. We model things like differential equations with computers. Those differential equations can sometimes accurately represent the time-course of a system (even very complicated, nonlinear systems). But it's really a brute force method that's terribly inefficient when compared to the way life does it. We have engineers designing systems that purposefully filter out noise so that only intended signals make it through. Evolution/life on the other hand, is a random, chaotic process, full of spandels.

By the way, just in case:

http://www.scientificamerican.com/article.cfm?id=could-silicon-be-the-basi

 

[apeiron]

The Turing machine has to run if we want to know its state at some future point. But its future state is determined, independent of our knowledge.

Not sure what you are trying to argue then as the final step in your story was about consciousness being present even if a program is not run. Which to me seems to have clearly gone beyond the Turing machine story which does require a minimal amount of hardware to move things along. And while you say the conclusions are predetermined, we do know also that there is actually the halting problem issue - http://en.wikipedia.org/wiki/Halting_problem

This idea of programs that don't need to be run also features of course in Putnam's arguments on rocks and finite state automata -
http://consc.net/papers/rock.html

In any case, I'm not sure you accept my premise that a consciousness could be simulated on a Turing device. If you do accept it, we can continue down this path of whether the simulation has to actually run. But it's a moot point if you don't. Which brings me to...

I don't just simply accept the premise. I feel instead it is a highly complex issue.

Most people treat it as either/or - either the brain and neurons are just digital devices designed for Turing style computation, or they are something else (chaotic, non-linear, whatever). I instead see biology as poised between these two kinds of things - rate dependent dynamics and rate independent information - as the limiting states.

Thus I certainly don't believe that consciousness can be simulated on a Turing device is some metaphysically straightforward way. But I do believe that something like computation is key to controlling/constraining the local dynamics, channelling it usefully.

I'm certainly not an expert on neurons (or computer chips for that matter). I do know something about physics, however, and I do not believe any biological process requires anything beyond essentially classical physics to operate. Sure there are chemical reactions which cannot be understood without QM, but these could be hard-coded in. I don't think quantum superposition could be maintained at the scales and temperatures in a human brain, rather, we can model the neurons as a complicated, but classical, system of molecules. And in any case, it is probably not necessary to simulate every last detail of a neuron to replace it, one can probably do a reasonably good job with a fairly simple set of rules (which could consist of thousands, but still much less than a mole, of rules) for converting inputs to outputs.

Well, this is a mess of assumptions that are not completely wrong, but also not completely right.

People who are biologists and neuroscientists quite often believe that there is more to biology than a physicist might think. Indeed some pretty prominent physicists like Schrodinger have said the same thing.

QM is of course irrelevant. No issue there.

But when it comes to coarse graining the dynamics of neurons, you are really in murky territory there. You just won't find many neuroscientists agreeing with the kind of blanket statement you want to make.

 

[Maui]

People who are biologists and neuroscientists quite often believe that there is more to biology than a physicist might think. Indeed some pretty prominent physicists like Schrodinger have said the same thing.

QM is of course irrelevant. No issue there.

I agree completely, but I would word it more mildly, to keep the physicists happy - perhaps we would one day find the new laws that govern emergent behavior.

That being said, i'd add my own speculation(the thread is poised toward speculation either way) that because mind/consciousness(choose another label) is vastly different from anything else in reality, i would expect that it cannot be possibly retained and simulated on a digital device(computer, chip, etc.). That is, until some new laws are discovered that guide consciousness(if they exist and if they are accessible). I am not an optimist(not that it matters much on such remote possibilities).

 

[Maui]

1. A human brain can be slowly replaced by a computer program, and the subjective experience would not significantly change.

I would consider awareness and free-will inseparable(two sides of the same coin) and would expect that some aspects of the human brain could be simulated(stimuli, pain, etc.), while others would not(meaning, logic, etc.). But i don't consider the human mind totally deterministic, maybe that explains my position better than anything i could propose.

2. The input to this computer could be yet another computer program, so that the entire system is a program running on a computer, and still, there would be conscious experience.

What happened to 'intelligence'? It got lost in the software and is now just a code, right? Yet, i would consider 'intelligence' a pretty fundamental aspect of this comprehensible reality.

3. Once we have such a program with conscious experience, the system used to implement it is not important, even a person manually operating a turing machine would suffice.

This makes no sense to me, but it could be that i am misreading your thoughts.

4. Since the implementation is not important, one might argue all that is important is the structure of the program itself. Even if it’s not actually running somewhere, there is still a subjective experience.

What does a "conscious zombie" mean? Isn't that supposed to be us, humans?

 

[StatusX]

It seems like we're getting stuck pretty early on in the argument. Namely, people don't seem to agree that the human brain could be simulated by a computer. That is, it seems like you don't even agree that a simulation of the functional properties of the brain is possible, let alone that such a simulation would have experiences of its own. Further breaking this down, I think there are three steps in this argument:

1. Any (classical) physical system can be simulated to arbitrary precision by a sufficiently powerful computer.

2. The brain is nothing more than a (very complicated, but essentially classical) physical system. If one were to simulate every molecule in the brain according to the laws of physics, one would find identical behavior to a real brain. In particular, it would behave in a way we would consider intelligent.

3. This brain would have conscious experience. Let me pause first and make it clear that we can't talk about this unless you accept the above two points (if you don't, that's fine, but then we should talk about them), since my argument here is dependent on them. Now, obviously this is tricky to argue for, since consciousness isn't understood in any deep sense. But the gradual replacement procedure outlined in my original post is meant as a way to motivate it. It seems likely that, if all functional properties of the brain were not changing during this process, that experience would not change either. If you disagree with this, you have to explain how it could change. Would it gradually fade away? Would it suddenly switch off? Keep in mind that, during this entire time, the subject would continue saying things like "Everything's fine, I'm still experiencing things like before," as must be the case since, by assumption, the brain is functionally the same. And another thing that you would have to explain is what, exactly, is special about biological systems that entitles them to consciousness, but not other systems which are, in every relevant way, functionally the same (again, if you think non-biological systems can't even be functionally the same, we need to go back to the first two points).

One more thing. I've seen the word "emergent" mentioned (or hinted at) a few times, and I'm not sure what people mean by this. Are you suggesting there could be additional physical laws governing very complicated systems? I find this hard to accept. Namely, if the microscopic laws of physics are correct, then the behavior of every individual atom is completely determined. There's no additional freedom. Are these emergent laws supposed to supersede the microscopic ones, so that the microscopic ones can sometimes be broken? Such a claim would require a great deal of evidence for me to take it seriously.

 

[apeiron]

It seems like we're getting stuck pretty early on in the argument. Namely, people don't seem to agree that the human brain could be simulated by a computer.

It would certainly be nice if you could prove this to be the case . Is simulating the brain Church-Turing computable? How do we know?

We certainly have a good prima facie case that the neurology is based on nonlinear processes. The question then becomes whether this nonlinearity is something the neurology wants to filter out (coarse grain away in an attempt to be digital, computational) or whether instead it is the driving force that the neurology wants to harness (and so it remains intrinsic and non-computable - it would take infinite time to calculate anything with the infinite accuracy required)?

1. Any (classical) physical system can be simulated to arbitrary precision by a sufficiently powerful computer.

Have you dealt with the computability issue yet?

2. The brain is nothing more than a (very complicated, but essentially classical) physical system. If one were to simulate every molecule in the brain according to the laws of physics, one would find identical behavior to a real brain. In particular, it would behave in a way we would consider intelligent.

Is reality actually computable? Is that what Church-Turing says (or are you invoking Church-Turing-Deutsch?)?

Then as I said in an adjacent thread, life/mind is defined by a matter~symbol epistemic cut. Or a hardware~software divide as computer science would see it. And can you in principle see the software from a full description of the state of the software? If you can't, then how do you simulate what you are not even able to describe.

It seems likely that, if all functional properties of the brain were not changing during this process, that experience would not change either. If you disagree with this, you have to explain how it could change.

If nothing changes, then nothing changes. No one disagrees about that. It is indeed the first two steps that already run into many difficulties.

One more thing. I've seen the word "emergent" mentioned (or hinted at) a few times, and I'm not sure what people mean by this. Are you suggesting there could be additional physical laws governing very complicated systems? I find this hard to accept. Namely, if the microscopic laws of physics are correct, then the behavior of every individual atom is completely determined. There's no additional freedom. Are these emergent laws supposed to supersede the microscopic ones, so that the microscopic ones can sometimes be broken? Such a claim would require a great deal of evidence for me to take it seriously.

Again, I refer to Pattee's epistemic cut and matter~symbol distinction (many theoretical biologists say the same thing in different ways, I just find him the clearest to cite).

So the evidence is certainly there if you want to take it seriously.

 

[Q_Goest]

Hi StatusX, apeiron... Apeiron, to your credit, you are an indefatigable debator but I think we have to be honest and forthright in presenting the model of consciousness you espouse. That conceptual model of the brain is not the mainstream paradigm. That isn't to say it's not valid, but I personally don't agree with it and it wouldn't be considered mainstream in the neuroscience community today. I think the primary issues that prevent it from becoming widely accepted is that it first requires a type of holistic, nonseparable model of nature at a classical scale that isn't widely accepted and because it requires some type of downward causation. I don't think arguing these points in this thread would be appropriate because that kind of discussion does not respond to the OP. I'd suggest opening a second thread if you'd like to discuss, I'd like to stay on topic here.

One more thing. I've seen the word "emergent" mentioned (or hinted at) a few times, and I'm not sure what people mean by this. Are you suggesting there could be additional physical laws governing very complicated systems? I find this hard to accept. Namely, if the microscopic laws of physics are correct, then the behavior of every individual atom is completely determined. There's no additional freedom. Are these emergent laws supposed to supersede the microscopic ones, so that the microscopic ones can sometimes be broken? Such a claim would require a great deal of evidence for me to take it seriously.

The concept of emergence has a lot of folks confused. There certainly are different definintions but they all generally boil down into one of 2 varieties. The kind of emergence you (and I) would accept is called weak emergence. That kind of emergence doesn't require new physical laws or any explanations of how microscopic elements of a system depend on the macroscopic description of that system. Weak emergence is what I'd call the mainstream view of emergence, that low level physical laws such as Van Der Waal forces are simply physical laws taken at one level that 'create' the physical laws such as those described by the Navier Stokes equations for example, at a higher level. We sometimes say that there are 'bridging laws' between the higher and lower level descriptions or physical laws that allow one law to be described by another. In this example, the Van Der Waals forces are not over-ridden by the higher level description, they are simply a description of nature at a given level. In comparison, strong emergence, or downward causation is the concept that these microscopic physical laws can be influenced in some way by the macroscopic description of the system. This is the type of concept that others espouse but it isn't as widely accpeted. It requires some type of causation which is created by the system at large that (somehow) influences the time evolution of said system at a microscopic scale. In this case for example, the description of the molecules bouncing around in a gas that can be described using Van Der Waals forces are insufficient to describe those interactions and other, higher level causes sometimes called "formal causes" or "top down constraints" intercede in or mediate the local interactions.

Potential objection: What if the operator makes a mistake, or even purposely modifies something against the rules of the program? It seems John would experience this, even though it doesn’t follow from the mere platonic existence of the program. All I can think of in response to this is that we can still imagine the operator-machine-program system itself is part of a more complicated program. John might wonder whether he is in the simpler program (where rules are never broken) or the more complicated, imperfect simulation. A priori, the former seems more likely by something like Occam’s razor.

I don't believe the scientific community in general addresses all the philosophical issues such as this one that you've raised, but certainly the philosophical community has considered these kinds of problems and logical objections while debating over the viability of computationalism. Some for example would contest that Searle's Chinese room thought experiment fails because the room, the person and the books used by that person all together equate to a system that harbors conscious experience. The machine you refer to for example, that consists of a person operating a machine is very similar and some would similarly argue that the machine and person constitute a conscious system. Where the system may break down and not give reliable results as you've pointed out has also been considered. Systems that may or may not support 'the proper counterfactuals' as it's been talked about in the literature, may or may not present a problem for computationalism. I call this the "special signal problem" which we had some discussion on https://www.physicsforums.com/showthread.php?t=460323".

I think Ferris nicely summarized the arguments around these problems here:

From http://www.macrovu.com/CCTGeneralInfo.html" :

Three premisses of computationalism (Maudlin, 1989):
1. Computational condition: Any physical system running an appropriate program is sufficient for supporting consciousness.
2. Non-triviality condition: It is necessary that the system support http://en.wikipedia.org/wiki/Counterfactual_conditionals" - states the system would have gone into had the input been different.
3. Supervenience condition: Two systems engaged in the same physical activity will produce identical mentality (if they produce any at all).

Imagine 2 pinball machines, the second of which has had exactly those pins removed that are never touched by the ball. We now have 2 different machines but it doesn't make a difference to the paths traced by the pinballs. The counterfactuals are different in each machine (the pinballs would behave differently if the first were to hit a "counterfactual" pin), but the physical activity of the 2 systems is, as it happens, identical. So, counterfactual differences in 2 systems are irrelevant to differences in the physical activity of those systems.

It's easy to imagine two machines with identical physical activity running a "consciousness program", one of which supports counterfactual states, while the other doesn't. The computationalist must claim, based on the non-triviality condition, that one is conscious and the other isn't. But this contradicts the supervenience thesis.

Suggestions for avoiding this paradox:
1. Rejecting supervenience.
2. Assigning no consciousness to either machine (Eric Barnes, 1991):
Neither of the imagined machines is conscious, because neither of them can causally interact with the environment. According to the computationalists, causal interaction is a necessary component for any model of consciousness. The object of any cognitive act must play a causally active role for it to be true that one cognizes that object.

I think by rejecting supervenience, you actually reject computationalism. By accepting causal interaction you are forced to accept downward causation, which doesn't sound good too. So what's left is rejecting the non-triviality condition.

I think the problem with any philosophical discussion around these issues is very similar to the problems of explaining any complex subject without the background in it, a lot of the concepts presented simply go over our heads. But this debate around counterfactual conditionals is what I think goes to the heart of what it is you see as being paradoxical about computationalism. There's a large number of philosophers that have argued on both sides of this debate. More about that if you'd like to discuss.

 

[apeiron]

In comparison, strong emergence, or downward causation is the concept that these microscopic physical laws can be influenced in some way by the macroscopic description of the system. This is the type of concept that others espouse but it isn't as widely accpeted. It requires some type of causation which is created by the system at large that (somehow) influences the time evolution of said system at a microscopic scale.

Hi Q Goest - So what is your argument against Pattee's epistemic cut then? His matter~symbol distinction? This is a "strong emergence" claim. But not of the sort you keep setting up as a straw man here.

You keep telling me what you believe, but you have yet to argue the case against what I've actually said.

 

[Q_Goest]

hi apeiron

Hi Q Goest - So what is your argument against Pattee's epistemic cut then? His matter~symbol distinction? This is a "strong emergence" claim.

I think having a discussion around this would be both impolite to the OP and a case of 'hijacking' the thread. If you're interested in discussing Pattee's epistemic cut, by all means, please create a new thread. Thanks.

 

[apeiron]

hi apeiron

I think having a discussion around this would be both impolite to the OP and a case of 'hijacking' the thread. If you're interested in discussing Pattee's epistemic cut, by all means, please create a new thread. Thanks.

Very cute. If you have a counter argument to any of the points I have raised in this thread, then now is your chance to present it.

Instead, you persist in attempting to derail the thread with your straw man version of the "emergence" argument.

I've forgotten. Are you a panpsychist as well? Or are you some kind of epiphenomenalist? What is your actual position on consciousness?

Is the mind "just physics"? Is the mind "just computation"? Is computation completely reducible to physics? What is it that you believe?

[Edit] To be clear, Pattee talks about the causal relationship between computation and material events. So if the OP is about "what kind of causality must be preserved" in a replacement scenario, Pattee it telling you precisely what kind.

Others like Searle show there is a symbol grounding issue. That is very easy to demonstrate. But Pattee offers the solution.

And the solution accepts that there is a computational aspect to life and mind. (So the thought experiment is not proven wrong by QM or chaos or other evidence of physical indeterminancy). But the computational aspect is top-down. It is a downwards imposition of constraints on dynamics.

 

[Pythagorean]

If you guys interested in the medical science of replacing neurons, it's been done with Parkinson's patients:

http://stemcells.nih.gov/info/scireport/chapter8.asp

If you get impatient, you can scroll down to:
Fetal Tissue Transplants in Parkinson's Disease Research

Here's one of the key papers:

Freed, C.R., Greene, P.E., Breeze, R.E., Tsai, W.Y., DuMouchel, W., Kao, R., Dillon, S., Winfield, H., Culver, S., Trojanowski, J.Q., Eidelberg, D., and Fahn, S. (2001). Transplantation of embryonic dopamine neurons for severe Parkinson's disease. N. Engl. J. Med. 344, 710–719.

There was no subjective change in experience for this test:

The results of one of these trials, led by Curt Freed, were published recently (Freed, 2001). Compared with control, patients who received the fetal-tissue transplant showed no significant benefit in a subjective assessment of the patient's quality of life, which was the study's primary endpoint. Moreover, two years after surgery, 5 out of 33 treated patients developed persistent dyskinesia—uncontrolled flailing movements—that had not been observed in the open-label work described above.

 

[StatusX]

Q_Goest, you raise some interesting points, but I'm not completely sure what your position on the argument in the original post is. Specifically, if you disagree with the conclusion, it would be helpful to me if you could explain at what point you think the argument breaks down.

As far as the material in some of the links you've provided, I'm not sure whether I think it is necessary that a system support the correct counterfactuals to be conscious. Why, for example, is it obvious that the system of individual neurons hooked up to impulse generators is not conscious? The pattern the neurons are implementing is due to a strict set of rules (a program), even if those rules aren't directly reflected in the physics of the system. They were relevant, for example, in designing the impulse generators. I'm agnostic on the issue of "when" the experiences are felt, if such an idea even makes sense. There's more I could say here, but I'll wait until you reply since I'm not sure what we agree on at this point.

 

[Q_Goest]

Hi StatusX,

Q_Goest, you raise some interesting points, but I'm not completely sure what your position on the argument in the original post is. Specifically, if you disagree with the conclusion, it would be helpful to me if you could explain at what point you think the argument breaks down.

I'm not trying to give you my position. I hardly ever do that here, but I'll let you know if I have a position on it and what that is. Rather, generally prefer to point out what the two sides of the argument have to say. Chalmers, Cristley, Endicott, Copeland and many other philosophers would argue that once you break down a system so that the computation is no longer reliable, once your computational string of events is such that it no longer is capable of reproducing a given computational algorithm reliably, including all counterfactuals, then phenomenal experience will either fade slowly or pop out of existence altogether just as you've suggested. Maudlin (and others) calls this requirement for a reliable, computational algorithm the "non-triviality condition". In other words, the computational network has to be non-trivial. There are others for example, Puttnam, Bishop, etc.. that suggest that such 'counterfactual conditionals' can't be necessary because there is essentially no physical reason for this condition. Puttnam points out that functionalims reduces to behaviorism if we reject the non-triviality condition which means that the system can behave as if it were conscious, and can even pass a Turing test, but not actually be conscious. I'm not sure where Maudlin stands exactly, but it seems like he rejects the non-triviality condition. If you want my personal belief, I'd also say I reject the non-triviality condition. In other words, there are no special signals.

As far as the material in some of the links you've provided, I'm not sure whether I think it is necessary that a system support the correct counterfactuals to be conscious. Why, for example, is it obvious that the system of individual neurons hooked up to impulse generators is not conscious?

Interesting that you'd say that. I don't feel it's necessary that a system support the correct counterfactuals to be conscious either. But clearly there are arguments on both sides of the fence.