Sing it, brother!
From SMBC, courtesy of reader Mark.
sub Ha! sub
And at the very back of the audience you can see a young Dan Dennett thinking: “What the hell is he on about?”.
Does this not hint at a different state of affairs being present in a parasite free person?
That’s what I was thinking — Shirley, that must be a reference to T. gondii, no?
My first thought too. Shirley, you like Gladiator movies.
No, and I’ve never been to a Turkish bathhouse, either!
You can run, you can run, tell my friend Free Willie Brown. And I’m standing at the crossroads, believe I’m sinking down.
I got the post hoc, ergo propter hoc blues.
‘Free choice is an illusion’ is a claim that is not identical to goddidit. It just isn’t. On the contrary, it is solidly established, based on easily verifiable empirical evidence.
There is a contingent of apostate non-believers. Something is wrong with them.
Okay, you can get off your pedestal now.
Move along now, nothing to see here.
Really? Can you point me to one such verification?
As I pointed out in a previous thread one has to believe the absurd proposition that an algorithm running on a register machine could be conscious in the same way that we are in order to believe that free will is an illusion.
Can’t do it, sorry. Not without some solid reason.
The solid reason is the universal speed limit.
Its existence means the most capable algorithmic machine is the universal machine of the Church-Turing thesis.
AFAIK it has been shown that neurons work to C-T capacity – no less, no more.
That doesn’t mean that any C-T machine will be conscious, it just means that some biological such are.
So long as you’re in the neighborhood, Torbjörn…I’m starting to suspect that Carroll’s observation on the complete understanding of the physics of everyday life may, in fact, be equivalent to a “proof” (such as that word applies to physics) of C-T.
Any thoughts on that?
Some people read more into C-T than is there.
It is talking about computations involving natural numbers.
It does not imply that a TM equivalent computer can do anything.
For example if reality is, as our physics suggests, continuous rather than discrete then everyday objects are performing operations that a Turing Machine cannot even in principle perform.
C-T, in its simplest form, just states that anything that is computable is also computable by a Turing Machine. I do not believe that I’ve ever characterized it otherwise. If you can point to such an example, please let me know so I can be sure to clarify and not make a similar mistrake in the future.
However, you are absolutely incorrect about the continuity of physics. Max Planck established otherwise long ago; see Planck Units and especially Planck length and Planck time.
Were the universe continuous, hypercomputation would be possible, and it is almost guaranteed that one could build a perpetual motion machine with the help of an hypercomputer. Therefore, unless you wish to advocate for the crankiest of all crank scams, you can be reasonably confident in the fundamentally discreet nature of physics.
Your mistake is in saying that physics is computable.
It is not. At least a vanishingly small proportion of it is computable if current models are correct.
Your information is sorely — nay, profoundly — out of date.
For a particular example, the computer on which you are reading this is capable of doing atomic-scale modeling of protein folding, and that’s far more than enough physics for brain simulations:
Our only limitation in simulating a brain at the molecular scale is the sheer volume of computer resources necessary. If we had a powerful enough computer, we’d already have such a simulation.
Ben Goren wrote: “And brains are, at least practically if not entirely, classical-scale phenomenon — and there’s nothing remotely approaching anything in the classical realm that can give you the kind of near-infinite precision necessary for the sorts of computation you’re hinting at. ”
Umm.. what computation are you claiming that I have hinted at?
I haven’t. I said what I said. The idea of the clockwork mind leads me to two possibilities, 1) that an algorithm running on a register machine can be conscious in the way that we are or 2) that the simulation of a human will appear to be incorrigibly convinced that it is conscious and yet not be conscious.
I cannot accept any theory which obliges me to be confidently sure that one or other of those propositions is true for reasons I have given already.
As I said before it seems perverse to try to believe that one of these propositions is true in order to believe that something else which seems to be the case is just an illusion.
Since that is the very epitome of a faith-based position, an absolute one based on neither evidence nor reason, it is clear that no amount of reason nor evidence will move you from it. Certainly, none I nor anybody else here can even theoretically provide.
For what it’s worth, your option 1) is the correct answer. Perhaps if, just as a thought experiment, you would assume it to be the correct answer and work through the consequences — all the consequences, including the obvious ones that your unshakable faith-based positions (along with their corollaries) are therefore unmistakably incorrect in this hypothetical alternate universe in which 1) is the correct answer — perhaps then you’ll manage to work yourself out of this hole you’ve dug yourself into.
Until then, I don’t think there’s anything further I can add to this conversation.
I am not sure what you mean by all that. It is a pity you cannot help Feynman out with this, I am sure he would be happy to know that his hypothesis turned out to be right.
I am sure he would be kicking himself to have missed that.
Ben Goren wrote: “C-T, in its simplest form, just states that anything that is computable is also computable by a Turing Machine. I do not believe that I’ve ever characterized it otherwise.
Thanks for the clarification, I was unsure of the point you were making.
So would you agree that the C-T thesis does not in any way imply that everything that happens is a computation?
No, your conclusion does not follow.
To near-completeness (and to completeness over significantly more than human scales), the laws of physics are completely understood and accounted for, and they are the very epitome of computability. (And the physics we haven’t yet understood and accounted for we can know for certain is utterly irrelevant to human cognition, so whether that’s computable or not is irrelevant.)
Since it’s all computable, it is, in a very real sense, a computation. That is, there exists a (more-than-mind-bogglingly-complex) single equation that is a perfect description of the (human-scale-and-then-some) universe. And there is no logical difference between an equation and a Turing-comparable machine computing it or any equivalent equation.
In other words, whether or not the Universe is some type of Matrix-style simulation (I strongly leads towards, “no,” but it’s impossible to rule out such a possibility), the universe is itself a Rube-Goldberg computer in the process of computing itself.
And, if the Universe, against all odds, is a Matrix-style simulation, then the program running said simulation is also a perfectly equivalent analogue of the exact same equation / computation.
It’s no different from how these two equations are the exact same thing:
x + y = 0 x = y * -1
If you agree with me that there’s no difference between those two, then the rest follows inescapably.
The fact of Planck units does not imply that the Universe is discrete.
You may have noticed that we use real numbers to model the universe, even at quantum levels.
We also use calculus. We would not need to if the Universe was fundamentally discrete.
And the Universe being continuous does not imply that there could be a hypercomputer.
Hey, but don’t believe me, believe Richard Fenyman who in a 1965 lecture explicitly says that the current laws of the Universe implies that it is continuous:
“It always bothers me that, according to the laws as we understand them today, it takes a computing machine an infinite number of logical operations to figure out what goes on in no matter how tiny a region of space, and no matter how tiny a region of time. How can all that be going on in that tiny space? Why should it take an infinite amount of logic to figure out what one tiny piece of space/time is going to do? So I have often made the hypothesis that ultimately physics will not require a mathematical statement, that in the end the machinery will b
e revealed, and the laws will turn out to be simple like the cheq
uer board…”—Feynman, The Character of Physical Law”
Eh, the fact that you can perform calculations outside of their suitable scope and come up with garbage answers — which is what you’re suggesting by applying continuous math to somehow probe smaller than Planck scales — should hardly be surprising. We already know that applying relativistic equations to quantum-scale phenomenon or vice-versa produces garbage answers.
Incidentally, that’s proof that physics is incomplete. However, any new, more complete physics, will have to reduce to quantum and / or relativistic physics at suitable scales, just as quantum and relativistic physics both reduce to Newtonian physics at human scales.
…and that leads right into my next point. Even if the universe as a whole is somehow continuous at small scales (never mind that we already know that it’s not continuous at large scales — see the speed of light, for example), that still doesn’t get you anywhere, for we know that it’s not continuous at the quantum scale. That’s why Quantum Physics is named, “Quantum”! Once anything smaller than quantum physics is scaled up to quantum physics, even if it was somehow continuous at the smaller scale, we know it’s discrete at the quantum scale.
And brains are, at least practically if not entirely, classical-scale phenomenon — and there’s nothing remotely approaching anything in the classical realm that can give you the kind of near-infinite precision necessary for the sorts of computation you’re hinting at. Brains — and even silicon chips — are already pushing the limits where the noise floor is perilously close to the signal, no matter how sophisticated your engineering gets. Because, remember? Much smaller and there’s the quantum world.
I don’t think anybody ever asked Feynman about hypercomputation and the computational implications of his hunch that space is ultimately continuous…but I suspect that, if somebody like you were to present ideas such as yours to him, he’d smile and laugh and gently explain to you the differences in scales and everything that stands between the scales you’re trying to bridge.
…on that topic, to help put it in perspective, the period at the end of this sentence is as much larger than Planck Length as the entire visible universe is bigger than the period.
I ask you: of what even remotely hypothetically potentially imaginably possible relevance could anything so small have to human cognition, in its world of neurons that you can see under a cheap microscope?
I am not sure how that changes what I say. I never suggested that neurons could out compute a Turing Machine.
And I am not saying that any algorithm ought to be conscious.
In fact I am doubting that an algorithm can be conscious.
What do you say about the hypothetical algorithm running on a register machine and modelling our neural processes, demonstrating the behaviour that those neural processes produce – would it be conscious, in the way that you are conscious?
Let me lay this out again.
Imagine a time when the available computing power has increased and also the human brain has been reverse engineered.
It should then be possible to have an algorithm running on a register machine which models a complete human being and can demonstrate any behaviour we find in a real human.
But this model cannot be conscious in the way we are, for reasons I have set out before and am happy to do again.
But on the other hand this leaves us with a sort of digital p-zombie. A simulation of a human brain and body which seems to be incorrigibly convinced that it is conscious, but is not. I screams in pain but feels no pain.
We either have to accept the conclusion of the conscious algorithm on a register machine, or the digital p-zombie and I suggest that both are absurd conclusions.
But we can only avoid one or the other by abandoning the idea of the clockwork mind that is inherent in the claim that free will is an illusion.
The “free will is an illusion” claim is far from the obvious conclusion that some say it is.
“But this model cannot be conscious in the way we are, for reasons I have set out before and am happy to do again.”
You say that with authority and conviction, but I don’t understand why you think your claim is any more supported by the current evidence than “free will is an illusion.” Since we are at the early stages of figuring out some useful tools to merely begin to directly investigate the functioning of the brain, and given the meager understanding gleaned so far, your assuredness seems premature.
Fine, you suggest that a “conscious algorithm on a register machine” and a “digital p-zombie” are both “absurd conclusions.” You have demonstrated your incredulity and your distaste, but beyond that you have not given any convincing reasons to support your position.
Your posited scenario makes no sense to me. If the brain has been reverese engineered, and the available computing technology is capable of fully modeling it, then by that very definition we would be capable of creating a mind that has all the features of the human mind. Unless you posit magic. It is either magic, or we have not accurately reverse engineered the human brain. So far there is zero evidence for magic, and lots of evidence against it.
“Authority and conviction”? No, I have stated an opinion and have given the reasons I think it is true in an earlier post. I have pointed out that I am happy to do so again if someone thinks an algorithm running on a register machine can be conscious as we are conscious.
We have but little understanding of a brain. On the other hand we know exactly how a register machine works.
My argument was based entirely on the knowledge of how a register machine works.
Again, “incredulity and distaste”? Incredulity about what? Distaste about what?
Where, exactly, have I expressed these things?
“If the brain has been reverese engineered, and the available computing technology is capable of fully modeling it, then by that very definition we would be capable of creating a mind that has all the features of the human mind”
That is an evasive answer. I am not talking about “fully” modelling it.
We can reproduce any particular behaviour of a system without “fully” modelling it.
I am talking about a model, running on a register machine, that is sufficient to produce the externally observable behaviour of a human.
In the “free will is an illusion” scenario this should be possible.
So my argument treats of the two possibilities of this simulation which produces the externally observable behaviour of a human.
Either it would be an algorithm running on a register machine that was conscious as we are.
Or else it would be a simulation of a human which appeared to be incorrigibly convinced that it is conscious but was not.
Which do you think it would be?
darelle wrote: “but beyond that you have not given any convincing reasons to support your position.
You mean that I have not given any reasons that have convinced you.
What part are you not convinced about?
Do you agree that the simulation described ought to be possible under the “free will is an illusion” scenario?
Do you agree with the two possibilities I stated about the simulation?
Do you think that an algorithm running on a register machine could, in principle, be conscious in the way that we are conscious?
Or do you disagree with me that the digital p-zombie option is absurd?
And I still have not hear any inference that actually supports the thesis that free will is an illusion.
So far I have heard some general arguments like “because of physics”, or that neurons cannot do any computation that a TM cannot do.
But how exactly do these support the conclusion that free will is an illusion?
Our host posted on this subject some time ago.
Yes I am familiar with the experiments – I am afraid these “searchlight among the voxels” exercises do not impress me as they do some.
The first point is that, despite how it was spun in the media, no actual prediction took place, all the data was analysed long after the experiments were concluded.
As I recall the hit rate for the left/right guess was about 57%, which is just on the cusp of statistical significance (p=0.05) for that sample size.
It would not surprise anybody if our conscious short term decisions were somewhat influenced by unconscious factors, especially ones taken under those circumstances (have you ever been in one of those machines?)
I don’t think that even the most ardent supporter of free will says that the decisions are completely uninfluenced by unconscious factors.
But the results could also be explained by conscious factors (people having a count of the number of left presses and thinking that they should balance it out) or mis reporting. for example the experiments appear almost designed to elicit late reporting.
The data was analyzed after the experiments concluded. This is unusual in what way?
I don’t recall saying it was unusual.
I said the media line on this was that there was an actual prediction of the choices.
Allow me to quote:
“The first point is that, despite how it was spun in the media, no actual prediction took place, all the data was analysed long after the experiments were concluded”
I thought it was the first point.
Which is exactly what I said that I said.
And nothing like what you said that I said.
Please point out the part where I say that this is unusual.
Perhaps we’ve fallen down a rabbit hole?
Over here, pointing out that data is analyzed after experiments are concluded is rather like saying travel was accomplished after distance has been traversed.
How is it possible to analyze data before it is collected? In what universe do people analyzed data before generating it?
The Republican and Christian universes, for starters….
Enter your email address to subscribe to this blog and receive notifications of new posts by email.
Join 37,159 other followers
Get every new post delivered to your Inbox.