I Why all the rejection of superdeterminism?

[greypilgrim]

Hi.

As far as I understand, superdeterminism (i.e. the experimentators are not free to choose the measurement parameters) allows the formulation of a local realistic quantum theory. But apparently physicists don't like the thought of not being in charge. Anton Zeilinger:
"[W]e always implicitly assume the freedom of the experimentalist... This fundamental assumption is essential to doing science. If this were not true, then, I suggest, it would make no sense at all to ask nature questions in an experiment, since then nature could determine what our questions are, and that could guide our questions such that we arrive at a false picture of nature."

I don't quite understand all that rejection. Up until the 20th century, physics was all about finding and describing the laws and mechanisms that underlie all things in nature, culminating in the idea of Laplace's demon. Certainly our minds would not be an exception. Surely there was quite a number of hardcore determinists in the physics community that did not question the usefulness of their work even if they were convinced that everything they did or thought was predetermined.

What has changed since those times that apparently now physicists reject the idea that their actions might not be based on free will? Especially since quantum mechanics shows that observers cannot be strictly separated from the system they're observing?

 

[_PJ_]

Philosophical and psychological.
People simply don't like to think they have no influence nor any means to influence anything. It makes their lives appear to be just as futile and unnecessary as they objectively are.

It's also unnecessary.
There are limits to understanding and measurement which may be limits to reality. The probabilistic notion of QM for example may be the bottom of that particularl rabbit hole.
Therefore determinism, isn't confined to one result, but a probability of results.

 

[A. Neumaier]

experimentators are not free to choose the measurement parameters

I think experimentators are indeed not free to choose the measurement parameters in the sense of indeterminism. They never choose absolutely freely but always determined by motives of various sorts. Without motives no incentive to perform a measurement.Thus there is no conflict with a deterministic universe.

 

[Demystifier]

The reason why physicists reject superdeterminism is not the conflict with free will. To have a conflict with free will, it's sufficient to have determinism (not superdeterminism), and physicists usually don't have a problem with this. Superdeterminism, unlike determinism, is problematic because it involves fine tuning of initial conditions. It means that not only your "choice" is predetermined by initial conditions (which is just determinism, and is not problematic), but is also strongly correlated with the measurement outcomes.

Here is an example which has nothing to do with quantum mechanics. Suppose that someone always correctly predicts the numbers that will be drawn in lottery. One possible explanation would be that she has supernatural powers, or more likely that she cheats somehow. But consider the following alternative explanation: She just makes guesses, but initial conditions in the universe are so finely tuned that her guesses (determined by deterministic processes in her brain and its environment) are perfectly correlated with deterministic chaotic processes that determine the lottery numbers. Such an alternative explanation would be - superdeterminism.

 

[stevendaryl]

Here is an example which has nothing to do with quantum mechanics. Suppose that someone always correctly predicts the numbers that will be drawn in lottery. One possible explanation would be that she has supernatural powers, or more likely that she cheats somehow. But consider the following alternative explanation: She just makes guesses, but initial conditions in the universe are so finely tuned that her guesses (determined by deterministic processes in her brain and its environment) are perfectly correlated with deterministic chaotic processes that determine the lottery numbers. Such an alternative explanation would be - superdeterminism.

That's exactly right. However, there is (in my mind) a bit of mystery involved in the rejection of superdeterminism. A completely normal sequence of events can look superdeterministic if you run it backwards: The newspapers print the winning lottery numbers and then later, someone generates random numbers, and they turn out to be exactly as predicted by the newspapers.

We think of the forward direction as unsurprising and the backward direction as really weird only because we are used to the (strange) fact that entropy was much lower in the early universe, which gives us a directionality to time. We don't really have a good explanation for that, other than: it's empirically true. So there is a sense in which superdeterminism would be a matter of making the forward direction of a sequence of events as weird as the time-reversed sequence.

 

[Strilanc]

To give an idea of *how much* fine tuning is required, consider that decoherence is happening basically everywhere always. Supposing for simplicity that decoherence happened as a single event, we can say that the number of decoherence events is proportional to both space and time: there's s^3*t decoherence events for a spacetime cube with width/height/depth=s and duration=t.

The initial state's information content doesn't grow with time. It has to be packed into a single instant. So for a spacetime cube with width/height/depth=s and duration=t we only have O(n^3) variables for the initial state but we must satisfy O(n^3 * t) constraints. The problem is massively overconstrained. Unless the constraints are almost all redundant, no solution will exist.

To my intuition, quantum measurement probabilities don't smell like a highly redundant constraint. And I bet superdeterminism has to introduce things that are weirder than quantum mechanics' no-signalling-style-non-locality in order to fix that problem.

Does anyone have a link to a toy superdeterminism model that's compatible with the Bell inequalities?

 

[stevendaryl]

Does anyone have a link to a toy superdeterminism model that's compatible with the Bell inequalities?

The famous physicist t'Hooft has a toy model for QM that is superdeterministic. I'm not convinced by it, but just as a data point that some have considered it seriously:

https://arxiv.org/abs/1405.1548

 

[DrChinese]

Superdeterminism is not a theory of a quantum mechanical world.

Superdeterminism is the hypothesis that local realism could be restored to a theory of a quantum mechanical world by the addition of system properties, initial conditions and other rules that are not only highly improbable, they defy any attempt to expose them. This hypothetical theory would feature no new predictive power, and would offer no superior scientific foundation than "Last Thursdayism" (or any variation on the Omphalos hypothesis).

On the other hand: there is no such theory to consider, critique or falsify. Were there such a theory, perhaps we could call it science and discuss it here. Certainly 't Hooft's paper could not be considered such a theory. It doesn't begin to explain how millions of independent measurements in a Bell test "conspire" to violate local realistic bounds. And how they do so in just the precise statistical manner so as to match the (right but for the wrong reasons) predictions of orthodox QM.

So my question back is: if you were to accept the premise of superdeterminism, how would you accept ANY experimentally observed outcome? Is the speed of light a constant c? Maybe it varies, but only appears to be c due to superdeterminism. Or what about atomic structure? Perhaps superdeterminism is fooling us there too. Maybe the Pauli exclusion principle is just an illusion, something that only appears when we attempt to look at it and is inactive at all other times. Just like with Bell tests.

Superdeterminism does not, in my book, qualify as science. It's more like a kind of religious belief. So, did I sufficiently convey how I REALLY feel?

 

[stevendaryl]

Superdeterminism does not, in my book, qualify as science. It's more like a kind of religious belief. So, did I sufficiently convey how I REALLY feel?

I share some of your dislike for superdeterminism, but your comments about it don't actually make sense to me. You're right that the claim: "There is some (unspecified) superdeterministic explanation for an experimental result is unfalsifiable, but so is the claim "there exists a non-superdeterministic theory". Until you specify a theory that makes specific predictions, you can't falsify it. If it does make specific predictions, then you can falsify it. Your examples are all examples of after-the-fact coming up with a superdeterministic theory to explain what has just been observed. That's not an argument against superdeterminism, it's an argument in favor of making predictions about future results, not just retrodicting past results. The superdeterministic theory that given a choice of M&Ms, nobody will choose a brown one is falsifiable.

 

[DrChinese]

Your examples are all examples of after-the-fact coming up with a superdeterministic theory to explain what has just been observed. That's not an argument against superdeterminism, it's an argument in favor of making predictions about future results, not just retrodicting past results. The superdeterministic theory that given a choice of M&Ms, nobody will choose a brown one is falsifiable.

A theory with the sole purpose of explaining quantum spin correlations as being the result of locally predetermined measurement outcomes synchronized with spacelike separated but nonetheless locally predetermined choices of observer measurement angles? Sounds like a brown M&M to me.

 

[stevendaryl]

A theory with the sole purpose of explaining quantum spin correlations as being the result of locally predetermined measurement outcomes synchronized with spacelike separated but nonetheless locally predetermined choices of observer measurement angles? Sounds like a brown M&M to me.

Well, there is no candidate superdeterministic theory, so it's weird to say, beforehand, that no such theory could possibly be scientific. I think that you could suspend judgment until such a theory was proposed.

As I said in another comment, the standard model of physics is superdeterministic if you run it backwards.

 

[Demystifier]

This hypothetical theory would feature no new predictive power, and would offer no superior scientific foundation than "Last Thursdayism" (or any variation on the Omphalos hypothesis).

Last Thursdayism has many similarities with Boltzmann brains, which are in fact quite popular in cosmology in a last couple of years. Here is one reasonable paper about Boltzmann brains
https://arxiv.org/abs/1702.00850

 

[ueit]

The reason why physicists reject superdeterminism is not the conflict with free will. To have a conflict with free will, it's sufficient to have determinism (not superdeterminism), and physicists usually don't have a problem with this. Superdeterminism, unlike determinism, is problematic because it involves fine tuning of initial conditions. It means that not only your "choice" is predetermined by initial conditions (which is just determinism, and is not problematic), but is also strongly correlated with the measurement outcomes.

Superdeterminism is the same thing as good old determinism. The word "superdeterminism" has been invented by Bell to make the fallacy involved in his theorem (circular reasoning) less obvious.

His theorem needs the instrument settings to be free parameters. Such a condition is already violated by classical determinism.

Let's try to analyze a Bell test from the point of view of classical electrodynamics. The whole experiment (source, detectors, experimenters, etc) is just a large system of charged particles (mainly electrons and quarks).

Once an initial condition (positions/velocities) is chosen everything becomes fixed. When the particles will be emitted by the source, what spin will they have, if and when they will be measured, what settings the detectors will have and what will be the result is a function of that particular initial condition. There is no need to use a "fine-tuned" initial condition. If you want the experimenters to "choose" a different setting you cannot implement that without changing the initial condition, and that change will also have an effect on the source and on the entangled particles themselves.

If you want to claim that the required change of the initial condition can be done without affecting the source and the particles I wish you luck. But be careful, we are dealing here with a field theory. If you move an electron around, the field of that electron will be modified at all places, no matter how far, so the trajectories of all particles will change.

Andrei

 

[A. Neumaier]

Once an initial condition (positions/velocities) is chosen everything becomes fixed. When the particles will be emitted by the source, what spin will they have, if and when they will be measured, what settings the detectors will have and what will be the result is a function of that particular initial condition. There is no need to use a "fine-tuned" initial condition.

Exactly!

 

[morrobay]

Superdeterminism is the same thing as good old determinism. The word "superdeterminism" has been invented by Bell to make the fallacy involved in his theorem (circular reasoning) less obvious.

His theorem needs the instrument settings to be free parameters. Such a condition is already violated by classical determinism.

Let's try to analyze a Bell test from the point of view of classical electrodynamics. The whole experiment (source, detectors, experimenters, etc) is just a large system of charged particles (mainly electrons and quarks).

Once an initial condition (positions/velocities) is chosen everything becomes fixed. When the particles will be emitted by the source, what spin will they have, if and when they will be measured, what settings the detectors will have and what will be the result is a function of that particular initial condition. There is no need to use a "fine-tuned" initial condition. If you want the experimenters to "choose" a different setting you cannot implement that without changing the initial condition, and that change will also have an effect on the source and on the entangled particles themselves.

If you want to claim that the required change of the initial condition can be done without affecting the source and the particles I wish you luck. But be careful, we are dealing here with a field theory. If you move an electron around, the field of that electron will be modified at all places, no matter how far, so the trajectories of all particles will change.

Andrei

Would your description above be a basis for a classical explanation for Bell inequality violations?

 

[ueit]

Would your description above be a basis for a classical explanation for Bell inequality violations?

I think we need to distinguish two different questions:

1. Are local, deterministic hidden variable theories possible?
2. Is Nature truly described by such a theory?

In order to answer affirmatively to the first question it is enough to show that all arguments excluding such theories are wrong. I think I have shown that above. Any field theory (classical electromagnetism, general relativity, fluid mechanics, etc.) has the generic properties required to put it outside the scope of Bell's theorem (and also Free-will theorem, and all variants), so, in principle is possible.

To answer the second question one has to show that the proposed theory gives quantum mechanics in some limit. Progress has been made starting from Yves Couder oil-drops experiments. It has been shown that many properties assumed to be uniquely quantum can be reproduced by such fluid-mechanical systems. Stable and quantified orbits, tunneling, single-particle interference have been shown to appear in such classic experiments. It has been shown that the oil drops can even be described by an analog of Schrodinger's equation. You can take a look at the article below:

Why bouncing droplets are a pretty good model of quantum mechanics:
https://arxiv.org/pdf/1401.4356.pdf

In conclusion my answer will be "certainly yes" for the first question and "probably yes" to the second.

I would like to stress here that a direct explanation of a Bell test in terms of a classical theory seems unlikely because of the great complexity of the system. Performing a simulation of 10^30 or so particles will probably be for ever outside our computational possibilities. But showing that QM's formalism could be deduced from a more fundamental classical field theory should be possible.

Andrei

 

[stevendaryl]

Superdeterminism is the same thing as good old determinism.

No, it's not. Regular determinism is the claim that the history for all time is uniquely determined by the initial state. Superdeterminism imposes an additional constraint on the initial conditions to insure that something happen in the future.

Demystifier explained, in a non-quantum way, the distinction. Let me try another example: Suppose your physical theory is just Newtonian mechanics, plus you have an additional rule that says that there is a special coin that can be flipped to determine whether it's going to rain. That is a superdeterministic theory. It's not an additional force or equation of motion, it's a fine-tuning constraint on the initial conditions of the universe. The coin flip doesn't cause the rain, and the rain conditions don't cause the coin to flip in any particular way. It's just that the initial conditions of the universe were set up to make this correlation work.

 

[stevendaryl]

I think we need to distinguish two different questions:

1. Are local, deterministic hidden variable theories possible?
2. Is Nature truly described by such a theory?

In order to answer affirmatively to the first question it is enough to show that all arguments excluding such theories are wrong.

I'm not going to say that you are wrong (even though I think you are), but what you're saying is not mainstream physics. It is an unorthodox personal theory. Get it published, and then we can discuss it in Physics Forums.

 

[stevendaryl]

Exactly!

No, not exactly. What he's saying is wrong---or at best, is an unorthodox, non-mainstream opinion.

 

[ueit]

I'm not going to say that you are wrong (even though I think you are), but what you're saying is not mainstream physics. It is an unorthodox personal theory. Get it published, and then we can discuss it in Physics Forums.

What I am saying is basic knowledge regarding Maxwell's theory. The theory is generally agreed to be deterministic, therefore any future state follows uniquely from the initial state. The theory is also reversible, a future state cannot be obtained from two distinct initial states. Do you really want a reference for that?

Then I referred to the fact that the electric field at a certain location depends on the position of its source. Do you need a reference for that?

Couder's experiments and their analogy to QM have been published in several papers. Here is the one reproducing single-particle interference:

Single-Particle Diffraction and Interference at a Macroscopic Scale
http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.97.154101So, please be more specific what is the statement in my post that is controversial and needs to be supported and I will try to provide a published reference.

Andrei

 

[stevendaryl]

What I am saying is basic knowledge regarding Maxwell's theory.

The claim that there is no difference between superdeterminism and determinism is not mainstream (it's false). The claim that Bell's reasoning is fallacious is not mainstream. The claim that Maxwell's equations predict violations of Bell's inequality for an EPR experiment is not mainstream.

 

[DrChinese]

Well, there is no candidate superdeterministic theory, so it's weird to say, beforehand, that no such theory could possibly be scientific. I think that you could suspend judgment until such a theory was proposed.

My point is precisely that there is no theory to suspend judgment on. What I am not suspending judgment on is the hypothesis that such a theory COULD be presented that could begin to make sense as a candidate. I say its hogwash, there is no scientific purpose to pursuing such. Those are my opinions, which are in fact far more reasonable than the counter.

 

[DrChinese]

I think we need to distinguish two different questions:

1. Are local, deterministic hidden variable theories possible?
2. Is Nature truly described by such a theory?

...

Sorry, these elements of the discussion have nothing to do with this thread. Start a new one, and please do not further derail this to discuss your personal ideas about Bell and local realism. I think you and I have covered this in the past.

 

[DrChinese]

Superdeterminism is the same thing as good old determinism.

I am simply echoing stevendaryl correct comments above on this and ueit's line of discussion.

 

[ueit]

The claim that there is no difference between superdeterminism and determinism is not mainstream (it's false). The claim that Bell's reasoning is fallacious is not mainstream. The claim that Maxwell's equations predict violations of Bell's inequality for an EPR experiment is not mainstream.

If you agree to the basic facts about field theories presented above you need to accept, by logical deduction, that such theories do not allow for "free decisions", so they cannot be in the scope of Bell's theorem.

Bell's theorem in itself is not fallacious, it only becomes so when applied to theories that are not within its scope. So, I agree that hidden variable theories that do allow the detector settings to be freely chosen cannot agree with QM. If you claim that Maxwell's theory does comply to the requirements of Bell's theorem then I would want a published reference for that. The "mainstream" literature is curiously devoid of references to field approaches.

I have not claimed that "Maxwell's equations predict violations of Bell's inequality". It might do so, or it might not.

Andrei

 

[A. Neumaier]

No, not exactly. What he's saying is wrong---or at best, is an unorthodox, non-mainstream opinion.

Then please point out where the error is. If one assumes that the universe (including all detectors and experimentors) is described by classical physics, the initial conditions determine everything including the choices of the experimentors, and there is no freedom. This is orthodox thinking since Laplace.

 

[stevendaryl]

Then please point out where the error is. If one assumes that the universe (including all detectors and experimentors) is described by classical physics, the initial conditions determine everything including the choices of the experimentors, and there is no freedom. This is orthodox thinking since Laplace.

Determinism is not the issue; superdeterminism is. In order for a deterministic local variables theory to reproduce the predictions of QM, the hidden variable must be chosen taking into account the future settings of the experimenters. Yes, if everything is deterministic, this is a theoretical possibility. However, the experiment can take place arbitrarily far in the future, and the setting can be made to depend on arbitrarily large regions of the universe, including regions that were initially outside of causal contact with the source of the twin pair.

Maybe it can all be made to work, but it is certainly not obvious that it is possible in any non-ad hoc way; in a way that doesn't seem as if the universe were constructed specially to make Alice and Bob have the right statistics.

 

[A. Neumaier]

the hidden variable must be chosen taking into account the future settings of the experimenters.

If there is nothing to choose except the initial state of the universe there is no difference. Given the initial conditions that actually describe the classical universe in which the experiments are performed, the laws of the universe will automatically conspire to force the experimenter to make exactly the choice actually made in the future settings. The reason is that to have a deterministic universe means that everything is determined by the initial conditions, including what appears to the experimenters as their free choices.

This is completely independent of Bell's arguments or of whether the classical universe in question matches our real universe and its quantum mechanical properties.

 

[DrChinese]

Bell's theorem in itself is not fallacious, it only becomes so when applied to theories that are not within its scope.

One last time: start a new thread to tout your personal theories about Bell. This thread is about superdeterminism.

 

[Haelfix]

If there is nothing to choose except the initial state of the universe there is no difference. Given the initial conditions that actually describe the classical universe in which the experiments are performed, the laws of the universe will automatically conspire to force the experimenter to make exactly the choice actually made in the future settings. The reason is that to have a deterministic universe means that everything is determined by the initial conditions, including what appears to the experimenters as their free choices.

This is completely independent of Bell's arguments or of whether the classical universe in question matches our real universe and its quantum mechanical properties.

The difference is there is a conspiracy taking place with superdeterminism and not with regular determinism. It's completely isomorphic to saying 'God put the Dinosaur fossils on Earth for humans to find them so as to convince them that the Earth is older than 5000 years'. It's a logically consistent possibility, but it renders the pursuit of science a moot point. Alternatively you can make it so that a 'fluctuation' in your brain triggers the second you want to measure Newtons constant. Everything is the same as far as the evolution of the laws of physics, the latter just seems fundamentally ridiculous as the set of initial conditions with precisely that property is tuned to such a ridiculous degree (amongst the set of all plausible initial conditions that produces the conditions that we observe). Its difficult to make that probabilistic statement precise of course (b/c of the various measure infinities) but at the very least it should simply be taken as an axiom in order to safely go about our day jobs.