Banner
    God Does NOT Play Dice But Einstein Is Still Wrong
    By Sascha Vongehr | November 23rd 2012 01:01 AM | 22 comments | Print | E-mail | Track Comments
    About Sascha

    Dr. Sascha Vongehr [风洒沙] studied phil/math/chem/phys in Germany, obtained a BSc in theoretical physics (electro-mag) & MSc (stringtheory)...

    View Sascha's Profile

    Einstein infamously said, and he said so many times, that god does not play dice.  He said this in order to refute indeterminism, and therefore, we cannot say that he was correct:  The meaning he intended to communicate was probably wrong (depending on your definition of “god”).  However, the most relevant meaning of his statement is self-evidently true:  Randomness (indeterminism) is not satisfyingly explained by merely postulating some more mystical randomness (god’s dice).

     

     

    The modern version of god throwing dice is called “genuine stochastic” behavior or “Genuine Stochasticity” (GS):  Dice falling without gods throwing them.  Since GS is often presented as if such refutes many-world models, it is very popular even among otherwise relatively rational scientists, although GS explains nothing.  GS is usually used in order to desperately retain some sort of naïve realism, although some GS proponents vehemently refuse this charge.  Nevertheless, the core problem is usually a misunderstanding of what, for example, modal realism is and entails.

     

     

    Why is GS nonsense?  We want to explain stochastic physical processes, for example the random sequence of results when we play dice or observe simpler quantum systems.  For that reason, we want a fundamental theory that models satisfactorily how such random sequences occur in empirical data.  Now let us say that the theory needs some sort of god that throws dice, or say, because god is out of fashion, we postulate some fundamental (‘genuine’) dice.  The question is not answered at all, because the quest was to explain in a fundamental way why dice behave in such random ways.  The task was not to connect just my own dice to some mystical ‘genuine’ or ‘godly’ dice.  If I ask about a fundamental description of dice and their randomness, I want all dice and their randomness explained!  Throwing more dice at me is not going to cut it.

     

    In modern GS speak:  Stochastic behavior is fundamentally “explained” by ‘genuine’ stochastic behavior.  They say that your dice behave that way because there are a sort of fundamental dice, namely quantum particles or the quantum universe around you, which ‘genuinely’ behave that way.  Not only does the ‘genuine’ stay mysterious (what is that supposed to mean???), but such argumentation amounts to claiming that dice behave like dice because of there being another level where another set of dice is thrown:  The “fundamental”, or “godly”, or “genuine” level.

     

    To hide the main question on another ("deeper") level and then claiming to be finished is called “regress without definite termination”.  It explains nothing, but the strong influence of evolved emotions onto our thinking and our social endorsement selection mechanisms keeps regress errors like the “flow of time” and “conscious awareness” dominant in the public discourse.  Regress thinking is evolutionarily fit.

     

    Even many of those who think in terms of branching many-world models, do still not grasp that all randomness and time is already represented in such models.  They instead think that a ‘genuine’ flow, the “flow of time”, carries single souls to branching points, at which then quantum dice are thrown, deciding whether they drift this or that way.  However, time has no further ‘genuine’ time that allows it to flow.  The flow of time over time is precisely dt/dt = 1 second per second, an empty statement.  Similarly, randomness has no further ‘genuine’ stochastic randomness – there is no meta-randomness which ensures the fairness of lower level dice.  Randomness is already fully represented by the branching.  No g picks branches at random - neither god nor 'genuine'.

    Comments

    blue-green

    Fair dice are far simpler to handle (both mathematically and socially) than loaded dice. It is far simpler to let randomness be assumed a priori than not.

    At the bottom of the postulated randomness is the equal weighting given to each term in a Feynman sum. It is no more mysterious than treating each interval of the real number line equally when doing integration … or for that matter … using the same plus symbol when adding A+B+C. If the process of addition depended on the specific numbers or intervals being added, one would have to create entirely new methods of addition that are dependent on what is being added and also their order. One would have to be very specific about how one sets parentheses. In other words, if randomness was not fundamental, then (A+B) + C would not equal A + (B+C) … and that would take a lot of huff 'n puff to explain.

    A bit of a stretch there in the logic, yet that's my thesis in a nutshell.

    vongehr
    Correct, and close to tautological modal realism, except for the nagging question of whether your number model is applicable to what fundamentally is simply uncertainty due to the trouble of getting self-identification of the ground.
    blue-green
    Can you spell that out better? Whether you mean getting self-identification of or off the ground, it is cryptic. Do you want to address issues of personal identity? Helium atoms have no personal identity or individuality. Self-identification? I don't think you are referring to spectral lines, although that would be a great use of self-identification from afar.
    vongehr
    Sorry about the spelling mistake - I always forget which one it is.  I mean that probability quantifies the uncertainty about my own situation, or in MWI/multiverse language, the uncertainty about locating myself (in which of the possible worlds am I now, in which one will I find myself?).  This makes probability dependent on a well defined "I" that tells what belongs to "me". I suspect it to be the key problem in deriving for example Born probabilities from tautological modal realism. Given nothing but my uncertainty, what is the rational for applying the mathematical model that you propose? Since the quantum probabilities depend on there being (stronger than classical common cause) correlations between different "worlds", this should somehow be inside the argument, otherwise one may as well end up with classical probability laws (no Bell violation).
    Ah, well, don't lose sight of your initial thesis!

    To say there is no genuine stochasticity is to say there is no fundamentally random process. It does not rule out equal probabilities for "being in" equivalent possible worlds. This question is present in a classical model too: even with infinite time and an ergodic system, why are we more likely to find ourselves in a "typical" state rather than a freak one? So it seems to me that you are now struggling with justifying the use of statistical arguments to cover ignorance (about ones microstate or world). This is a problem which applies in very simple classical statistics - like a genuinely random toss of a coin :) - hardly in the same league as Born probabilities.
    vongehr
    Well, is this not basically what I said? Let me try to be careful:
    Ah, well, don't lose sight of your initial thesis!
    Not sure which one you mean.
    To say there is no genuine stochasticity is to say there is no fundamentally random process.
    1.) I would not say there is no GS rather than that saying X is simply "genuine" is not an acceptable further illumination of X.
    2.) Yes, GS holds time (and thus processes) as fundamental/a priory, which is not a satisfying fundamental description.
    It does not rule out equal probabilities for "being in" equivalent possible worlds.
    As far as GS has been argued to me, and also because of the time-aspect just discussed, GS has equal probabilities for "finding myself later on" in possible worlds that are now equivalent in that very way (having the same probability), however, GS is what those people cling to who hate the idea of now "being in" equivalent worlds in the sense of relative/indexical actualization.
    This question is present in a classical model too:
    Yes - that is what 99.9% of people do not understand.  Most of this stuff has nothing to do with QM, many worlds models are not all QM.
    why are we more likely to find ourselves in a "typical" state rather than a freak one?
    Because that is the definition of "typical".
    So it seems to me that you are now struggling with justifying the use of statistical arguments to cover ignorance
    No - just that I would like the argument to lead to QM probabilities rather than classical ones.
    This is a problem which applies in very simple classical statistics - like a genuinely random toss of a coin :) - hardly in the same league as Born probabilities.
    Yes precisely, if you mean by "classical" that probabilities are not violating the Bell inequality (with classical determinism, it is not "genuinely random").
    Did I get what you wanted to say or am I all in a parallel universe again?

    Not exactly parallel but not perfectly orthogonal either - coming at the subject from a different basis, it could hardly be otherwise! :)
    Because that is the definition of "typical".
    Since when have definitions been reliable? :)

    I don't see how you can create an operational procedure for taking a typical sample in an experiment without assuming a genuinely random process to create the sample. It's not a problem if you assume a God's-eye view (since you can then take the responsibility of asserting that the situation is probably typical and justify it using the mathematical axiom of choice), however that's no help to Wigner who must merely assume that he has probably been placed in a typical world.

    Anyway, from your following comment it would seem you are not concerned with that rather subtle point...
    No - just that I would like the argument to lead to QM probabilities rather than classical ones.
    Okay, but this is the reason I have taken early retirement - to brush up of some pretty basic QM. I still am not sure where the problem lies - I have the impression that, if we forget the magical observer and any Wheeler-esque "completion of an observation", environmental decoherence accounts for how the Born rule arises without making it a "quantum postulate".

    Be that as it may, what intrigues me is this: decoherence is specific to QM as we know it, so when you say "rather than classical ones", you must have in mind how classical models would create emergent probability. So I presume you want firstly to free your Description from the assumptions of those models and secondly to derive a particular quantum law. Do you think the latter is ever going to happen - "Cogito ergo sum ergo Schrodinger ergo Born"? Or would you be happy at this stage to show that classical probabilities are not implied if you jettison some assumptions (preferably ones now known to be wrong!)? 
    vongehr
    operational procedure for taking a typical sample in an experiment without assuming a genuinely random process to create the sample
    Indeed we are rotating orthogonal rather than parallel. All I ever know is my so called "empirical" records.  Even if I just won the jackpot, I still find that most others that are in whatever way similar to me, have not won.
    environmental decoherence accounts for how the Born rule arises without making it a "quantum postulate".
    So you put already environment and entanglement (which then can decohere) and all that in?  How do you start refusing to be in a classical environment or one like the Game of Life in the first place?
    decoherence is specific to QM as we know it
    Probably my point.
    how classical models would create emergent probability.
    I guess classical actually have "objective" probability, which is why so many people desperately defend such classicality.  That probability is "really out there" or does make no sense at all is their charge.
    So I presume you want firstly to free your Description from the assumptions of those models and secondly to derive a particular quantum law. Do you think the latter is ever going to happen - "Cogito ergo sum
    That is pretty much it, because all other ways leave the origin of "reasonable" stochastic laws for ever mysterious.  As I wrote in "utterly simple idea that demands the quantum" and other places, there is nothing left but god painting isolated block universes a certain way if you refuse any logical reason for the laws that restrict stochastic behavior.  The dynamics of a certain substance simply making the future out of its present shape, that was a great classical idea that almost satisfied.  Now we know for sure that there is no such substance (was philosophically always clear anyway).
    I suspect that one way toward progress is the EPR "three body problem" rather than starting with assuming quantum environment.  By "three body problem" I mean Charlotte, Alice and Bob thinking through the mutual consistency of all their worlds/minds matching up.  If the sine squared violating Bell falls out from that, we have derived "the quantum".  Or perhaps Alice and Bob plus considering how they communicate the relative angle is enough.  That has never been analyzed properly.  In order to get to the empirical laws that dependent on the angles, the relative angle must enter the empirical records, too!  There is no actual angle but just its probability, say due to a sequence of polarized photons that are communicated and measured via the polarizing crystals, all outcomes (100101101) being another equivalent world (as typical as 00000000).  Finite angle resolution is given via the operational protocols (In QM given by the total possible spin of the setup, but of course QM should not be assumed, just classical "photons").  You may not find the last answer, but there is something deep there that you can find relatively fast.  I wish I could retire.
    Where exactly do you divert from Everett's approach, or are you on the same course almost 60 years later? One of the working titles for Everett's thesis was 'Wave Mechanics Without Probability'. Methinks that the objections to Everett by Rosenfeld and Bohr are still valid today.

    blue-green
    Let's step back a bit and critically review whether an all encompassing universal wave function Psi can be a valid concept in quantum mechanics QM. In everyday operational QM, Psi comes with an explicit delineation of the type of experimental setup and context in which it is being used.  It does not exist a priori of any initial or boundary conditions. The great many quantum amplitudes that can be in Psi are like the harmonics in a tone ~ they need a body ~ a chamber with boundary conditions ~ and something to strike a tone, for them to be, even if it is just vacuum energy. The specific decomposition of Psi into harmonics (orthogonal functions) is dependent on the coordinate system one chooses to use, just as it is with the components of a vector. Be that as it may, the abstract existence of a vector (tensor or matrix operator) without coordinates is no license for treating Psi in the abstract divorsed from any particular experimental frame.  A Psi without a body or observational framework is required if one applies it across an entire universe. This is a questionable stretch of the quantum paradigm.

    The great error in Hugh Everett's state vector for Psi and so much that has come afterword involving whole universes or totalities is the mistake of trying to free Psi from the contexts that give it meaning in the first place. This is why Everett's thesis was dead on arrival in Copenhagen in spite of Wheeler's valient efforts to give it a hearing. It is Psi's dependency and encoding of the whole experintial arrangement in which it lives and breathes that gives the quantum amplitudes and subsequent probabilities their holistic and nonlocal flavor. Psi may feel like a global all encompasing object, yet it never exists without all the details about what exactly is bolted down to a lab bench, both when and where without violating Special or General Relativity.

    Psi never exists in isolation or pure abstractedness like a Platonic solid. It is always anchored in a material situation or at least represents such a case. There can be an array of representative circumstances which can be encapsulated by the same Psi. They fit in an equivalence class signified by Psi, each of whose members give the same experimental data.

    A universal Psi would be ill defined ~ essentially ambiguous ~ a misuse of the fundamental tenants of QM. In Bohr's words, the "uncontrollable interactions" would not be present to create the "conditions which define the possible types of predictions regarding the future behaviour of the system."

    ((Hopefully this comment is not too out of place. I tried to get in earlier when I was in Zurich and Bern ... I really would like for you to answer the question I posed when I could not log in.  
    As I noted then, One of the working titles for Everett's thesis was 'Wave Mechanics Without Probability'. Is this what you are trying to do? Where exactly do you divert from Everett's approach, or are you on the same course almost 60 years later? Methinks that the objections to Everett by Rosenfeld and Bohr are still valid today. I have tried to spell this out in the above. Maybe it will tickel Mr. Potter's fancy.))
    vongehr
    This is not on topic.  I do not want to do wave mechanics.  Everett is dead and we know more today than he ever did, so I don't care about his working title for some thesis.  I sometimes say "Everett relativity" instead of things like "relative actualization with quantum correlation" in order to talk to people who are hung up on famous names.  It is a compromise in an environment of academia and science (writing) that is 100% marketing.  If you understand that all this now fashionable psi talk is nonsense, why you talk about it?  I never mentioned psi.
    blue-green
    Yes, it is true, I looked up your more technical arxiv paper on slicing sausages before making my comment, and sure enough, it doesn't mention Psi. So why do I bring up Psi? Because what would QM be without it?

    We are familiar with you distancing yourself from others. When you refer to branches or many-worlds or dice, your take is unique and independent and unchained from ghosts of the past. Fresh air is good. Let's have some more. Eventually you'll get around to what Psi means to you. It can't be irrelevant.
    vongehr
    Eventually you'll get around to what Psi means to you. It can't be irrelevant.
    I see your point now.  I think you give implicitly the answer in your question, which makes it a good question.  Psi developed as a description of parts of totality and always presupposes a split of totality into some system that an observer/describer, who is split from it, is uncertain about.  It is constructed by the boundary put between them via the self-identification of the observer.  Thus psi does not work for a totality (that contains all observers).  The question we are after is:  Why is that boundary quantum, for example, why is it holographic in case the boundary goes through an assumed spacetime?  Now your version of the question is:  Why does Psi make sense on that boundary?  David Deutsch asks the same question via asking for why Born probabilities make sense to a rational decision maker who has to decide for self-consistent expectation values attached to different outcomes of interaction measurement changing that boundary.  I ask the question at present by asking for the self-consistency that demands quantum correlations between different situations in which I can find myself in, which are the different ways in which "I" the observer/describer can self-identify, which includes uncertainty about that cutting apart of totality, like how much of totality and its multiplicity I internalize by identifying.  One could label such a description "solipsist" in some way, and constructionist of course.
    Why should GS require an explanation? GS itself is the ultimate explanation. Any other explanation begs the question: why this explanation and not some other one? In other words: why should the world be like this? And any answer to this question invites new questions. Only when we recognize GS as the ultimate answer will this regression stop. Let'saccept that the world is random at the most basic level. The only alternative would be a God with a lot of phantasy.

    vongehr
    I almost agree if you mean to define GS simply as fundamental uncertainty, and "explanation" in a rather non-functional way (there should be some sort of transformation into a fitting model that makes whatever question "answered"), however, that is not what GS is used for in as far as I have encountered it. It is applied as if such mere "I stop here, period" somehow is a valid refutation of, among others, many world models having applicability anywhere and apparent quantum non-locality disproving naive realism.
    Well, it's not just a matter of explaining one sort of randomness with another. It's useful to chase the appearence of randomness down to where it might originate. A finite, not infinite, regress. I doubt whether you would be able to speak convincingly about "genuine randomness" being superfluous if the right statistics were not naturally emergent from Schrodinger in MWI.
     




    vongehr
    That is why I always go carefully with "regress without definite termination". The regress as terminated via GS is not satisfyingly terminated. Regress is fine, even infinite regress could conceivably be a satisfying solution to some sort of regress problem, but if we have finite regress, the termination has to be something more than stamping one's foot on the ground and saying "this level now is the genuine one".
    As you may remember, in my tautological modal realism, future indeterminism comes from a more fundamental uncertainty. That uncertainty is coming straight from the trivial fact of that there are more than a single possible situation ("world") in order to have any description at all. So I terminate at my starting point, namely totality (of possible worlds), which is not something that one can question in any description. GS would never know whether it is not apparent/emergent like relativity in an Einstein-ether.
    I would accept your axiom after you explain to me how the determinism of the macrocosm arises from the indeterminism of the most basic level of yours. Otherwise, keep quite. My cat is in the box alive. I can hear the noise.

    I really dont understand this urge to find words to capitalize. Science is about prediction and data compression and nothing more. Assigning metaphysical significance the various latent variables in your model, might be satisfying in some sense, but it doesn't actually accomplish anything. In this context, I would simply say that, noise (stochasticity) is simply how we model the effects of our ignorance. That ignorance can have structure (fair vs unfair coins). Now no matter how clever we are and how much data we acquire some ignorance will always be with us and, as a result, some noise. Is that Genuine Stochasticity? Should we care? I think not. Rather, and I think the author hits on this, declaring something to be GS should be avoided both because it metaphysical BS and because its giving up, something which ought not set well with a scientific mind.

    vongehr
    I can almost support your point about "ignorance" being the core - in my tautological modal realism I employ the term "uncertainty" in a somewhat similar way, and it is not what some people call GS. I disagree however with your statement in so far as it reduces science to empirical science. There is more than data, especially after quantum mechanics seems to quite seriously tell us that data can fundamentally not be relied on to support even quantum mechanics itself (freak branches). So, logic and construction of consistent terminology belong to science, or else, science is dead.
    damonisherwood
    I just 'corkboarded' this but I think you might find it of interest too.A letter from Einstein answering a letter from a school girl about whether scientists pray. 
    "Dear Phyllis, I will attempt to reply to your question as simply as I can. Here is my answer: Scientists believe that every occurrence, including the affairs of human beings, is due to the laws of nature. Therefore a scientist cannot be inclined to believe that the course of events can be influenced by prayer, that is, by a supernaturally manifested wish. However, we must concede that our actual knowledge of these forces is imperfect, so that in the end the belief in the existence of a final, ultimate spirit rests on a kind of faith. Such belief remains widespread even with the current achievements in science. But also, everyone who is seriously involved in the pursuit of science becomes convinced that some spirit is manifest in the laws of the universe, one that is vastly superior to that of man. In this way the pursuit of science leads to a religious feeling of a special sort, which is surely quite different from the religiosity of someone more naive. With cordial greetings, your A. Einstein"

    damonish
    vongehr
    I think you might find it of interest too.
    Thank you, but no, not really.  Einstein is here just a hook in order to dismiss the atheistic update "God's dice2.0" called "genuine stochasticity" that people desperately throw against modal realism.