Local superdeterministic hidden variables - in Physical Review Letters

[Demystifier]

It is an old idea that, at least in principle, hidden variables could be local if they are superdeterministic. However, so far this idea seemed too speculative for highly respectable journals such as Physical Review Letters to publish research on it.

But now it seems that it has changed. The following paper is recently accepted for publication in Physical Review Letters:
http://lanl.arxiv.org/abs/1103.2201

Any comments?

 

[Physics Monkey]

This reminds me of 't Hooft's cosmic conspiracy idea. The classical dynamical state of the world is just such that all the experiments we do, which are predetermined, can look consistent with the non-local correlations of quantum mechanics.

A technical point which I don't understand: normally restricting your measurement only has an effect if you also restrict the states you can measure. In other words, you can effectively measure different observables by first performing unitaries and then making your standard (predetermined?) measurement. I don't understand how this plays into the author's comments.

 

[DrChinese]

This concept does NOT lead to a local realistic theory, even if there is local determinism. And it leads to other theoretical issues, which I believe are clearly untenable.

 

[lugita15]

This concept does NOT lead to a local realistic theory, even if there is local determinism. And it leads to other theoretical issues, which I believe are clearly untenable.

What's the difference between local realism and local determinism?

 

[ThomasT]

It is an old idea that, at least in principle, hidden variables could be local if they are superdeterministic. However, so far this idea seemed too speculative for highly respectable journals such as Physical Review Letters to publish research on it.

But now it seems that it has changed. The following paper is recently accepted for publication in Physical Review Letters:
http://lanl.arxiv.org/abs/1103.2201

Any comments?

Imo, the cited paper adds to the confusion surrounding the interpretation of Bell's theorem.

Bell already demonstrated the compatibility of LRHV and standard QM wrt individual measurement scenarios.

The term 'superdeterministic' just means deterministic, which refers to LRHV formulations, which are ok wrt individual measurements.

The problem wrt modelling quantum entanglement is in requiring an experimental result (eg., coincidental photon flux) that isn't determined by the hidden variables which determine individual results to be modeled in terms of the hidden variables which determine individual results.

The underlying determinant of coincidental results in entanglement setups is a constant (because varying the global measurement parameter produces predictable variable results). This constant is the relationship (which doesn't vary from pair to pair) between the hidden variables (which vary randomly from pair to pair) which produce the individual results.

To say that hidden variables could be local if they're superdeterministic seems a bit silly. The underlying, variable, parameters that produce individual results are local. The underlying, nonvarying, global parameters that produce coincidental results are also, presumably, locally produced. And there's nothing in the literature that requires us to abandon that assumption -- except a misunderstanding of the meaning of Bell's theorem.

 

[Demystifier]

The term 'superdeterministic' just means deterministic

No it doesn't. In deterministic theory the initial conditions are arbitrary. In superdeterministic theory the initial conditions are "fine tuned" or "conspired" such that some additional regularity emerges. In the context of quantum mechanics, this regularity is correlations that cannot be explained by random initial conditions and local interactions between hidden variables.

 

[ThomasT]

No it doesn't. In deterministic theory the initial conditions are arbitrary. In superdeterministic theory the initial conditions are "fine tuned" or "conspired" such that some additional regularity emerges. In the context of quantum mechanics, this regularity is correlations that cannot be explained by random initial conditions and local interactions between hidden variables.

What's being referred to as "initial conditions" in a quantum entanglement setup?

 

[Fra]

I only read the abstract but I can weakly relate to one of their supposed keys, in an unexpected way since I'm probably as far away from seeking to restor realism as you can come. This is why it is funny. It's almost like a connection between two opposites.

It is based on the assumption that Alice and Bob can choose measurements from a measurement set containing multiple elements. We establish a new paradigm that departs from Bell’s paradigm by assuming that there are no choices for Alice and Bob and that the measurements Alice and Bob will make are fixed from the start.

This is an interesting concept to start with as it puts the focus on wether the observer A really have a CHOICE of what to measure, with respect to observer B? If not, why? Ie. what is it that constraints or tunes the questions that do ask?

This exact question appears in my view as well, but for other reasons. In my view, the choice of action and the choice of measurement is the same thing seen from an external observer. Because what A "chooses to measure" is exactly the same question how A behaves. Because the behaviour of A, seen from B is to just see what questions A asks.

So in that sense, the idea that two secondary observers (alice and bob) can FREELY choose what to measure actually doesn't quite make sense, because what questions they will ask are necessarily constrained by a rationality constraint.

For very small "observers" such as an atom, it seems more reasonable to think that their choice of measurement is "random" (seeing questioning as a random walk: one step = one question; question-answer is what maps the walk into hypothesis space).

But for classical or large observers, it in fact seems quite unresonable that their actions are random. They are more likely to follow quite precisely from the state of the observer (or measurement device). In this case the preparation of classical detectors etc. In the Alice/Bob exampl the HÚMAN decision is actually just introducing confusion. Technically different preparations are different observers! Thus we don't have one Bob, we has as many Bob's as there are "choices".

I'm not sure about the rest of the paper but I think there are good reasons to look close at exactly what the freedom toe choose what ot measure means. This is one thing that's overlooked, and one does not need to favour restoraion of realism to apprecaite this viewpoint.

/Fredrik

 

[Demystifier]

What's being referred to as "initial conditions" in a quantum entanglement setup?

Initial conditions of hidden variables. For example, in the Bohmian theory these are initial positions of all Bohmian particles, including the particles of the EPR pair, as well as particles constituting the brain which makes the "free decisions" of the direction in which the spin will be measured.

Bohmian theory is deterministic (and nonlocal), but not superdeterministic.

 

[DrChinese]

What's the difference between local realism and local determinism?

Counterfactual Definiteness (i.e. realism) requires that values be possible for other angle settings which the observer is unable to select (due to the predetermination). There are no such values for many settings except values which do not follow QM. So this is equivalent to denying the key EPR assertion of realism. Of course, I deny realism anyway, so in that I agree.

But Superdeterminism itself is a horrible crock of anti-science. You could use the same to explain ANY physical results, not just Bell test results.

 

[ThomasT]

Thanks for the feedback. I have more questions.

In deterministic theory the initial conditions are arbitrary. In superdeterministic theory the initial conditions are "fine tuned" or "conspired" such that some additional regularity emerges.

Ok, so a nonlocal deterministic theory can correctly predict the coincidental photon flux for any joint setting of the polarizers, while keeping lambda (the hidden variable which determines individual photon flux) random, because it allows paired (entangled) photons to communicate instantaneously or ftl.

On the other hand, a local deterministic theory cannot correctly predict the coincidental photon flux for some joint settings of the polarizers, while keeping lambda random, because it forbids paired (entangled) photons from communicating instantaneously or ftl.

But, a local superdeterministic theory can correctly predict the coincidental photon flux for any joint setting of the polarizers, because ... why? Because lambda is not described as varying randomly? If that's the case, then how is lambda described? Or is it something else?

I'm inclined to go with DrC's assessment that "Superdeterminism itself is a horrible crock of anti-science." But the fact is that I really don't know what it means (other than determinism with a superfluous prefix). So, I'm hoping that you or somebody will elaborate a bit.

In the context of quantum mechanics, this regularity is correlations that cannot be explained by random initial conditions and local interactions between hidden variables.

My understanding is that standard QM correctly predicts coincidental photon flux for any joint setting of the polarizers, while keeping lambda random, because it models coincidental photon flux in terms of the relationship between paired (entangled) photons. This relationship is an underlying, global parameter which doesn't require interaction between the paired (entangled) photons, and which is analyzed by the global measurement parameter of joint polarizer settings. The results are in line with the optics understanding of the behavior of photons locally interacting with polarizers. This can be illustrated by considering a simple optical Bell setup where, say, polarizer A is moved to the side with polarizer B -- then the coincidental photon flux follows the same cos^2 Theta angular dependency as when the polarizers are on opposite sides. In the case with polarizers A and B on the same side there's no need to posit nonlocal interactions to understand the observed angular dependency. So why should it be necessary when the polarizers are on opposite sides? But this is for understanding. In order to make an explicitly LRHV model of the coincidental photon flux with the polarizers on opposite sides, then the impossibility of instantaneous or ftl communication between the paired (entangled) photons has to be explicitly encoded into the model. And there doesn't seem to be any way of (clearly) doing that without skewing the predictions of such a model, even though the angular dependency remains essentially the same.

Anyway, I still have the tentative opinion that the cited paper doesn't improve our understanding of quantum entanglement or Bell's theorem ... that it basically just adds to the confusion surrounding the interpretations of these things.

You haven't yet said what you think of the paper. So, what do you think of it?

 

[moving-finger]

Initial conditions of hidden variables. For example, in the Bohmian theory these are initial positions of all Bohmian particles, including the particles of the EPR pair, as well as particles constituting the brain which makes the "free decisions" of the direction in which the spin will be measured.

If the world is deterministic, these "initial positions" are all determined (by prior states of the world) - including the "positions" of the particles in the brains of the experimenters - so one has to ask - what is a "free decision" in this context?

There seems to be some disagreement about just what "superdeterminism" is. In a quote attributed to John Bell in a BBC Radio interview with Paul Davies in 1985:

There is a way to escape the inference of superluminal speeds and spooky action at a distance. But it involves absolute determinism in the universe, the complete absence of free will. Suppose the world is super-deterministic, with not just inanimate nature running on behind-the-scenes clockwork, but with our behavior, including our belief that we are free to choose to do one experiment rather than another, absolutely predetermined, including the "decision" by the experimenter to carry out one set of measurements rather than another, the difficulty disappears. There is no need for a faster than light signal to tell particle A what measurement has been carried out on particle B, because the universe, including particle A, already "knows" what that measurement, and its outcome, will be.

The way that Bell describes it, superdeterminism does indeed seem to be determinism pure and simple?

 

[Fra]

I neither believe in realism nor determinism but when just trying to look at the reasoning, it seems to my that if you believe in complete determinism then it seems to me the superdeterminism is also required in order to have a coherent reasoning?

(My own comments above doesn't mean I believe in superdeterminism, but it does mean that I believe that "freed choices" of the secondary observers (alice and bob) is just regular physical actions from the point of view of the primary observer (the observer that later compares results from bob and alice), so the extend of "freedom" alice and bob has, from the point of view of the primary observer is nothing but the limited predictivity, so alice and bob free choices are IMO like gauge choices from the point of view of primary observer. )

If you believe in determinism, but not superdeterminism then the interesting question is where do you draw the line? If you adpot superdeterminism it seems to me that at least you are coherent in the reasoning?

/Fredrik

 

[moving-finger]

If you believe in determinism, but not superdeterminism then the interesting question is where do you draw the line? If you adpot superdeterminism it seems to me that at least you are coherent in the reasoning?

I concur.

 

[ThomasT]

There seems to be more than one sense in which the term 'superdeterminism' is being used.
There's the philosophical sense of superdeterminism, which doesn't seem to me to be any different from determinism.

Then there are the formal (model/theory) components referred to as superdeterministic replacements or supplements, which is what Demystifier seems to be talking about.

Then there are the elements of the experimental design/preparation referred to as superdeterministic, which is what the cited paper seems to involve, as well as the formal description of this.

What I don't understand wrt the paper is how does restricting the set of measurements enforce determinism/superdeterminism.

I also don't understand what Demystifier means by "fine tuning" a formalism so as to make the model/theory superdeterministic rather than just deterministic.

 

[moving-finger]

I also don't understand what Demystifier means by "fine tuning" a formalism so as to make the model/theory superdeterministic rather than just deterministic.

Demystifier seems to be suggesting that its the boundary conditions which make the difference - simple determinism could assume arbitrary boundary conditions, whereas superdeterminism (at least according to Demystifier) requires some special set of boundary conditions.

 

[lugita15]

Demystifier seems to be suggesting that its the boundary conditions which make the difference - simple determinism could assume arbitrary boundary conditions, whereas superdeterminism (at least according to Demystifier) requires some special set of boundary conditions.

Not boundary conditions, initial conditions. Like in classical mechanics you have Newton's 2nd law, which is a second-order differential equation, but to solve it in a particular situation requires initial conditions, typically the positions and velocities of all the particles at time t=0 (position and velocity would constitute the "hidden variables".

So in a "super-deterministic" or "conspiratorial" model, at time t=0 all the particles in the universe would have had a big meeting, where they all set their initial conditions just right in order to make it appear that local realism is wrong. This includes the particles that end up in the brain of the experimenter, or whatever determines the polarizer setting. So the EPR-pairs are given exactly the polarizations required to make the Bell inequality appear to be violated, because they know in advance what angles the polarizers will be turned to.

Of course, this is all unfalsifiable and hence non-scientific. If we could actually determine the values of the hidden variables, so that we had a viable scientific proposal, then Heisenberg's uncertainty principle, amply confirmed by untold numbers of experiments, would be wrong. So don't hold your breath.

 

[ThomasT]

Not boundary conditions, initial conditions. Like in classical mechanics you have Newton's 2nd law, which is a second-order differential equation, but to solve it in a particular situation requires initial conditions, typically the positions and velocities of all the particles at time t=0 (position and velocity would constitute the "hidden variables".

In an optical Bell test involving photons entangled in polarization, what does t=0 refer to? The time of emission of an entangled pair? What are the hidden variables? The polarizations of the paired (entangled) photons?

So in a "super-deterministic" or "conspiratorial" model, at time t=0 all the particles in the universe would have had a big meeting, where they all set their initial conditions just right in order to make it appear that local realism is wrong. This includes the particles that end up in the brain of the experimenter, or whatever determines the polarizer setting. So the EPR-pairs are given exactly the polarizations required to make the Bell inequality appear to be violated, because they know in advance what angles the polarizers will be turned to.

But didn't Demystifier indicate, or at least suggest, that the predictions of local superdeterministic models (as opposed to the predictions of local deterministic models) agree with QM? That is, aren't local superdeterministic models enhanced in some way so as to predict (correctly) results that local deterministic models can't? This is what I'm asking about. What makes a model of a particular experimental preparation superdeterministic as opposed to merely deterministic?

If we could actually determine the values of the hidden variables ... then Heisenberg's uncertainty principle ... would be wrong.

Why/how would that invalidate the uncertainty relations? They're based on the assumption that a fundamental quantum is a fact of the 'resonances' of nature, aren't they?

Suppose that we had the ability to actually qualitatively determine what QM already assumes wrt, say, the photons entangled in polarization in the preparations of the Aspect 1982 experiment, which is that opposite moving photons emitted during the same atomic transition (ie., from the same atom) are identically polarized. What would that entail? More accurate predictions of coincidental photon flux?

Suppose that we could determine the actual polarization axes, prior to filtration, of the entangled photons in that experiment? What would that entail? The prediction of nonrandom sequences wrt individual detection attributes? Would it change the predictions/observations of photon flux at detector A or B. Would it change/enhance the predictions/observations of coincidental photon flux? Would it invalidate the assumption of the existence a fundamental quantum?

 

[moving-finger]

Not boundary conditions, initial conditions. Like in classical mechanics you have Newton's 2nd law, which is a second-order differential equation, but to solve it in a particular situation requires initial conditions, typically the positions and velocities of all the particles at time t=0 (position and velocity would constitute the "hidden variables".

OK, but we're just nit-picking. An "initial condition" is a "boundary condition" in my book.

 

[kith]

As Bell indicates, the 'super' in superdeterminism just refers to deterministic theories in the absence of free will.

In a world with deterministic laws (like Schrödinger's equation, Newton's second law, etc), you can still choose initial conditions in experiments, if there is free will. If there is no free will, your choice of initial conditions is also determined. In this sense, the second case is more deterministic than the first, hence the phrasing 'superdeterministic'. Confusingly, this is exactly what's called 'determinism' in philosophy.

 

[lugita15]

In an optical Bell test involving photons entangled in polarization, what does t=0 refer to? The time of emission of an entangled pair? What are the hidden variables? The polarizations of the paired (entangled) photons?

The time t=0 is some hypothetical time in the past when all of the particles in your system, or worse yet all the particles in the universe, communicated with each other and set the initial values of their hidden variables. This include the particles, or the ancestors of the particles, which will eventually end up in the brain of the experimenter, or whatever device he uses to choose the polarizer setting. It also includes the photons, or the ancestors of the photons, which will be measured in the Bell test. Presumably t=0 occurred long before the emission of your entangled pair, because it had to be a time when all of the particles were within a small distance of each other, so that they could communicate without FTL signals (otherwise we would have a nonlocal realist theory).

As to what the hidden variables are, they need to come in two kinds:
1. The particles whose descendants will be the photons in the Bell test will need to have information about whether a photon should go through or not when it encounters the polarizer, knowing in advance what the angle will be.
2. The particles whose descendants will (for instance) be in the brain of the experimenter need to have information about which setting the polarizer should be set to, knowing in advance whether the photon will go through or not.

But didn't Demystifier indicate, or at least suggest, that the predictions of local superdeterministic models (as opposed to the predictions of local deterministic models) agree with QM? That is, aren't local superdeterministic models enhanced in some way so as to predict (correctly) results that local deterministic models can't? This is what I'm asking about. What makes a model of a particular experimental preparation superdeterministic as opposed to merely deterministic?

Yes, a local superdeterminist model would make the same predictions as quantum mechanics. In a standard local realist model, Bell's inequality would be satisfied, whereas in quantum mechanics it is violated. In a superdeterminist model, the particles would set their initial conditions, knowing in advance what the polarizer settings will be, in order to make Bell's inequality appear violated. In other words, they are conspiring in order to make local determinism seem false when it is really true.

Why/how would that invalidate the uncertainty relations?

If you know both whether the photon will go through a 0 degree polarizer and whether it will go through a 45 degree polarizer, both with 100% certainty, then you have more knowledge than is allowed by the uncertainty principle. To take a simpler example, if you do an experiment which determines the momentum of a particle exactly, and position is the hidden variable (like in Bohmian mechanics), then if you also found out the value of the hidden variable, you would again be violating the uncertainty principle.

 

[Demystifier]

I don't like superdeterminism for essentially the same reason as Dr Chinese, so I have no intention to further defend it. Instead, I give links to some superdeterministic papers by the Nobel prize winner 't Hooft:
http://xxx.lanl.gov/abs/quant-ph/0701097
http://xxx.lanl.gov/abs/0908.3408

But I find very interesting that PRL publishes a paper that seems to advocate superdeterminism, even if not using that word explicitly.