Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Local superdeterministic hidden variables - in Physical Review Letters

  1. Sep 22, 2011 #1


    User Avatar
    Science Advisor

    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:

    Any comments?
  2. jcsd
  3. Sep 22, 2011 #2

    Physics Monkey

    User Avatar
    Science Advisor
    Homework Helper

    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.
  4. Sep 22, 2011 #3


    User Avatar
    Science Advisor
    Gold Member

    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.
  5. Sep 22, 2011 #4
    What's the difference between local realism and local determinism?
  6. Sep 23, 2011 #5
    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 modelled 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.
  7. Sep 23, 2011 #6


    User Avatar
    Science Advisor

    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.
  8. Sep 23, 2011 #7
    What's being referred to as "initial conditions" in a quantum entanglement setup?
  9. Sep 23, 2011 #8


    User Avatar

    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.
    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.

  10. Sep 23, 2011 #9


    User Avatar
    Science Advisor

    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.
  11. Sep 23, 2011 #10


    User Avatar
    Science Advisor
    Gold Member

    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.
  12. Sep 23, 2011 #11
    Thanks for the feedback. I have more questions.

    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.

    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?
  13. Sep 24, 2011 #12
    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:
    The way that Bell describes it, superdeterminism does indeed seem to be determinism pure and simple?
    Last edited: Sep 24, 2011
  14. Sep 24, 2011 #13


    User Avatar

    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 belive 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?

  15. Sep 24, 2011 #14
    I concur.
  16. Sep 24, 2011 #15
    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.
  17. Sep 24, 2011 #16
    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.
  18. Sep 24, 2011 #17
    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.
  19. Sep 25, 2011 #18
    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?

    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?

    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?
  20. Sep 25, 2011 #19
    OK, but we're just nit-picking. An "initial condition" is a "boundary condition" in my book.
  21. Sep 25, 2011 #20


    User Avatar
    Science Advisor

    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.
    Last edited: Sep 25, 2011
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook