Hidden Variables
Bell's theorem - for or against Hidden Variables? - Printable Version

+- Hidden Variables (https://ilja-schmelzer.de/hidden-variables)
+-- Forum: Foundations of Quantum Theory (https://ilja-schmelzer.de/hidden-variables/forumdisplay.php?fid=3)
+--- Forum: The Violation of Bell's Inequalities (https://ilja-schmelzer.de/hidden-variables/forumdisplay.php?fid=7)
+--- Thread: Bell's theorem - for or against Hidden Variables? (/showthread.php?tid=8)

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28


RE: Bell's theorem - for or against Hidden Variables? - secur - 08-01-2016

To answer your question ... I don't know that aging process can't be reversed. In fact I'm quite confident it can. Within 100 years, assuming bioscience continues to progress at its current rate, "immortality" treatments will probably be available. But that has nothing to do with reversing "time" per se. When old bodies are rejuvenated - transformed into healthy 20-year old bodies - the medical treatment involved, whatever it may be, will proceed in the forward time direction and be explicable in scientific terms. No magic or time-reversal involved.

How do I know that time itself can't be reversed? That I can't travel back to 1963 and warn Kennedy to stay away from Dallas (for instance)? In fact I don't know it can't be done. I doubt it very much, for reasons of science, philosophy, logic. However it's a theoretical possibility. I can even imagine ways it might be done, which are of course extremely speculative.

Talking about time, I finally see the underlying "trick" of Hess and Philipp! I can explain how I figured it out, if anyone's interested. It's based on the fact that in actual lab experiments, it's not straightforward to determine which detection at station A pairs with which detection at station B. The way it's done in practice is, each station emits a "time-tag" along with the binary spin detection. If two time-tags are close enough - within a window - they are assumed to belong to entangled particles, otherwise the detection is invalid (discarded). So, by modifying the time-tags by a formula based on the detector settings, you can violate Bell inequality. Basically, if the angles are too far apart, discard that pair. They don't do it that simply, but it could be done that simply.

They never mention this in their paper, instead you have to dig it out. What Gill refers to as Hess and Philipp's "non-elements-of-reality" (tongue partly in cheek) amounts to a variable which operates in the post-detection selection process. The variable decides whether a pair of detections is valid based on the difference between the two time-tags. Which depends on the difference of the two detector settings! QED.


RE: Bell's theorem - for or against Hidden Variables? - secur - 08-01-2016

On further reading, I see this loophole is apparently called the "synchronization" loophole - if I understand correctly. In fact, you probably already know that Hess and Philipp must be based on this idea. But it's not at all obvious from reading the paper, and (I thought) worth pointing out.


RE: Bell's theorem - for or against Hidden Variables? - gill1109 - 08-01-2016

(08-01-2016, 07:14 PM)secur Wrote: On further reading, I see this loophole is apparently called the "synchronization" loophole - if I understand correctly. In fact, you probably already know that Hess and Philipp must be based on this idea. But it's not at all obvious from reading the paper, and (I thought) worth pointing out.

I call it the coincidence loophole https://arxiv.org/abs/quant-ph/0312035. I agree that is not at all obvious from reading the Hess and Philipp papers that this is what they were thinking of. Their paper did inspire Jan-Ake Larsson and myself to think about and analyse the loophole and show that it is worse than the well-known efficiency loophole.

The mathematics of the Hess-Philipp papers has not got anything to do with the coincidence loophole at all. They believed that they had built a genuine local hidden variables model which violated Bell inequalities. It was very difficult to track down the mistake in the mathematics, it was hidden very deeply and came down to forgetting an index (one of many) and forgetting to renormalise a measure to get a probability distribution, in the heart of their construction.

The latest experiments are not afflicted with either the detection loophole nor the coincidence loophole. A strict timing schedule means that the experimental unit is not "particle pair" but "time window". Per time window there are always two new settings and there are always two binary outcomes.


RE: Bell's theorem - for or against Hidden Variables? - secur - 08-01-2016

Gill1109 wrote: They believed that they had built a genuine local hidden variables model which violated Bell inequalities.

No doubt you're right, having so much experience with these things, but I wound up deciding they couldn't possibly think that. So, they were thinking of the "coincidence" loophole, and doing their best to hide it (perhaps only subconsciously). That "insight" (according to you, wrong) was, I thought, possibly a new contribution to the topic.

Gill1109 wrote: It was very difficult to track down the mistake in the mathematics, it was hidden very deeply and came down to forgetting an index (one of many) and forgetting to renormalise a measure to get a probability distribution, in the heart of their construction.

Yes I read your paper, that's what tipped me off to it, although I had noticed it a bit already. I just couldn't believe they could make such a mistake, and decided it was (more or less) deliberate. But again, I'll accept your judgment; compared to you I only glanced at the paper. Also I don't know the people involved as well.


RE: Bell's theorem - for or against Hidden Variables? - gill1109 - 08-02-2016

(08-01-2016, 10:19 PM)secur Wrote: Gill1109 wrote: They believed that they had built a genuine local hidden variables model which violated Bell inequalities.

No doubt you're right, having so much experience with these things, but I wound up deciding they couldn't possibly think that. So, they were thinking of the "coincidence" loophole, and doing their best to hide it (perhaps only subconsciously). That "insight" (according to you, wrong) was, I thought, possibly a new contribution to the topic.

Gill1109 wrote: It was very difficult to track down the mistake in the mathematics, it was hidden very deeply and came down to forgetting an index (one of many) and forgetting to renormalise a measure to get a probability distribution, in the heart of their construction.

Yes I read your paper, that's what tipped me off to it, although I had noticed it a bit already. I just couldn't believe they could make such a mistake, and decided it was (more or less) deliberate. But again, I'll accept your judgment; compared to you I only glanced at the paper. Also I don't know the people involved as well.
Well there are two authors of these papers: one is a mathematician, the other is a physicist. There is a mismatch between the physics ideas and the mathematics theorems.

Their insight that the coincidence loophole is important is absolutely correct. It seems someone had had it before (the paper is quite forgotten, and right now I can't recall the name of the author) but it was not widely known and appreciated till Hess and Philipp triggered Larsson and Gill to work on it. Our work became known and triggered improvement in the experiments.

PS: I looked up the reference: S. Pascazio (1986) Phys Letters A 118, 47-53 shows how the coincidence loophole can allow local realism to violate Bell inequalities.


RE: Bell's theorem - for or against Hidden Variables? - secur - 08-04-2016

Admittedly I still don't understand Hess and Philipp thoroughly and should study it more. But the conclusion seems to be, it does not present a valid alternative to QM interpretation of Bell experiments.

Christian's paper also fails. It's not in the same league as Hess, IMHO: Hess is a respectable attempt.

Bottom line: Bell seems to be right. Now, assuming that for the sake of argument, what conclusion can we draw?

Richard Gill says, in "Time, Finite Statistics, and Bell’s Fifth Position":

"Bell offered four quite different positions which one might like to take compatible with his mathematical results. They were:

1. Quantum mechanics is wrong.
2. Predetermination.
3. Nature is non-local.
4. Don’t care (Bohr).

In my opinion he missed an intriguing fifth position:

5. A decisive experiment cannot be done."

#5 is certainly possible. If so Quantum Computing will never work, I believe. This should be decided in a decade or two. #1 is rejected per assumption, #4 is negligible. #2 can't be proven one way or the other, so it's not really science.

That leaves #3, non-local, as (at least) a viable option. I don't see how anyone can reject it out of hand, as (for instance) Murray Gell-Mann does. He, and others with that opinion, just go over the basic facts of QM, then say "see, it's just QM!" - as though that answered the question.

There are more options, I think, which come under the heading of "non-realism". In particular EPR's "elements of reality" concept is clearly not the way the world works, if we accept QM. But most flavors of "non-realism" don't actually invalidate #3.

To emphasize, I'm making certain assumptions here. One, Hess and Philipp really are presenting a non-local scheme, as Gill says. (That remains an assumption, only because I haven't comprehended the argument entirely.) And, that no one else has a valid anti-Bell objection either - I certainly haven't studied all the literature. And, that the fifth position will be shown wrong in a decade or two.

So, this doesn't apply to Thomas Ray or others of his persuasion. He doesn't accept Bell in the first place, so naturally rejects non-locality, since Bell experiments are the only indication of it. I'm talking to those who accept Bell but still reject non-locality.

Why this resistance to non-locality? Since it can't be used for signaling, it doesn't violate relativity or any other known physics.

I'd like to hear any logical reason for non-locality being "out of the question", as Gell-Mann asserts.


RE: Bell's theorem - for or against Hidden Variables? - Thomas Ray - 08-04-2016

" ... Hess and Philipp really are presenting a non-local scheme, as Gill says."

No they're not.  Gill's argument is supported by the belief that lambda can be "anything."  Hess and Phillip introduced a separate time  index.  (" ... the outcomes conditional to lambda may still depend on the space-time coordinates, and outcome independence conditional to lambda is not guaranteed.  For example, clocks in the two stations can show correlations no matter what was sent out from a source." ~ Einstein was Right! p. 80)

"I'm talking to those who accept Bell but still reject non-locality."

The better question for Gill is why he thinks non-locality is not a prior assumption of Bell's theorem.


RE: Bell's theorem - for or against Hidden Variables? - jrdixon - 08-04-2016

secur: I'm not familiar with the Gell-Mann quote, but my guess is that his doubt about nonlocality is akin to your doubt about retrocausality. Here I refer to an earlier post in this thread where you wrote: "How do I know that time itself can't be reversed? ... In fact I don't know it can't be done. I doubt it very much, for reasons of science, philosophy, logic." So you may be interested in this paper: arxiv.org/abs/1512.08275. It seems to me someone willing to accept nonlocality based on Bell's argument and related experiments would also be prone to accept retrocausality in light of the arguments in this paper.  On the other hand, doubt about the conclusions of this paper might perhaps cause a believer in nonlocality to reassess that belief, due to the similarity of the arguments. My own take is that with a wording change, this paper can be turned in to an argument for local realism with forecasts, which I argued above as an alternative to nonlocality.


RE: Bell's theorem - for or against Hidden Variables? - secur - 08-05-2016

@Thomas Ray,

You left out the sentences before and after the one you quoted:

I'm making certain assumptions here. One, Hess and Philipp really are presenting a non-local scheme, as Gill says. (That remains an assumption, only because I haven't comprehended the argument entirely.)  

For the time being, I've moved on to the question about non-locality. Main reason: "Einstein was Right" isn't available in my local library, so I've ordered it, and put off further study until it arrives. It sounds like a good read and no doubt will help me get Hess's ideas.

Having said that: all things considered, I think Gill is probably right. But, that's not to say he IS right, until I understand it myself. BTW note the difference between Hess and Christian. Christian's paper is simple enough that I can agree with Gill's conclusion - that Christian made mistakes - from my own knowledge. Anyway, in a week or two I hope to get back to Hess & Philipp. For the time being I'm interested in Gell-Mann and (what seems clearly to be) his mistake.

In a sane world I wouldn't have to work so hard! If the physics establishment all agree on some point - Hess is wrong, for instance - I should be able to simply accept the expert opinion. But in this "insane" world, the biggest names in physics, such as Gell-Mann, Witten, Weinberg, Deutsch, and so on, are clearly illogical. They make assertions which are prima facie unjustified; plus they disagree with each other, so someone has to be wrong (probably all of them). The assertions I'm referring to are of the sort "Theory XYZ is obviously right", where XYZ can be String Theory, Many Worlds, Bekenstein-Hawking entropy, etc. They also say "there is definitely no XYZ", where XYZ is "absolute reference frame", "ether", "non-locality", etc. But these theories and assertions not only have no experimental support, they arguably can never have such support! It's because of this that Hess and other dissenters from the mainstream (such as Schmelzer) may, pending my own study, very possibly be right. Actually, Bell's theorem is really a minor point compared to this overarching problem with theoretical physics: the inmates have taken over the asylum.

@jrdixon,

jrdixon wrote:  I'm not familiar with the Gell-Mann quote, but my guess is that his doubt about nonlocality ...

It's not a doubt, but a certainty, and that's the problem. If someone says "I doubt non-locality" - or, "I think String Theory is right", or "I think there's no ether" - that's fine. But certainty is totally unjustified. Many are guilty of this logical fallacy; probably, it's about funding. These stubborn stances make reasoned discourse, and the progress of physics, extremely difficult.

The Elitzur et al paper is good. My dismissal of "time reversal" referred to "macro" time reversal, such as going back to 1963; not this subtle effect. I have no problem with it. Previously I defined "non-local" in terms of what one would have to do with a computer simulation to achieve QM results. As I defined "Property X", a computer program can't be retrocausal, so must use an orthocausal, simulated FTL signal. But that doesn't mean Nature must use that mechanism. I agree that retrocausality, as defined here, is another way to do it. Off the top of my head there's no way to distinguish between the two alternatives, but Elitzur et al mention such experiments might be possible.

It also helps me understand what you're getting at with your "forecast" idea. Without examining it in detail it seems another valid ontology, as I said before. But I think it's confusing to call it "local realism with forecasts" because the word "realism" is being stretched too far.

Until there's an experiment such as Elitzur et al hint at the distinction between these ontologies is philosophical and I lump them all under "Property X". It's very important to note: this QM weirdness is implicit in basic QM (two-slit experiment, for instance) without Bell, GHZ, or whatever. For QM to work at all requires something like Property X. The question addressed by Aspect experiment etc, is whether it still works when past light cones are appropriately separated. Back in 1927 Einstein noted this, and concluded there was a problem with QM. Evidently he was wrong, but there are still details to nail down.

So I consider Elitzur et al's retrocausality (and, probably, your "forecasting") as philosophical variants of "property X". Non-locality seems, to me, the simplest way to imagine Nature's mechanism. There are others, and they're worth considering; but until experiments can discriminate, they're FAPP equivalent.


RE: Bell's theorem - for or against Hidden Variables? - Thomas Ray - 08-05-2016

secur,

"@Thomas Ray,

You left out the sentences before and after the one you quoted:

I'm making certain assumptions here. One, Hess and Philipp really are presenting a non-local scheme, as Gill says. (That remains an assumption, only because I haven't comprehended the argument entirely.)  

For the time being, I've moved on to the question about non-locality. Main reason: "Einstein was Right" isn't available in my local library, so I've ordered it, and put off further study until it arrives. It sounds like a good read and no doubt will help me get Hess's ideas.

Having said that: all things considered, I think Gill is probably right. But, that's not to say he IS right, until I understand it myself. BTW note the difference between Hess and Christian. Christian's paper is simple enough that I can agree with Gill's conclusion - that Christian made mistakes - from my own knowledge. Anyway, in a week or two I hope to get back to Hess & Philipp. For the time being I'm interested in Gell-Mann and (what seems clearly to be) his mistake.

In a sane world I wouldn't have to work so hard! If the physics establishment all agree on some point - Hess is wrong, for instance - I should be able to simply accept the expert opinion. But in this "insane" world, the biggest names in physics, such as Gell-Mann, Witten, Weinberg, Deutsch, and so on, are clearly illogical. They make assertions which are prima facie unjustified; plus they disagree with each other, so someone has to be wrong (probably all of them). The assertions I'm referring to are of the sort "Theory XYZ is obviously right", where XYZ can be String Theory, Many Worlds, Bekenstein-Hawking entropy, etc. They also say "there is definitely no XYZ", where XYZ is "absolute reference frame", "ether", "non-locality", etc. But these theories and assertions not only have no experimental support, they arguably can never have such support! It's because of this that Hess and other dissenters from the mainstream (such as Schmelzer) may, pending my own study, very possibly be right. Actually, Bell's theorem is really a minor point compared to this overarching problem with theoretical physics: the inmates have taken over the asylum."

I sincerely apologize.  I don't often cut comments short -- in this case, I wanted to emphasize that the problem of non-locality goes away with the realization that it is no more than an assumption to prop up Bell's theorem.  That's why Gill wants to do away with it, and can't, because it lies at the heart of any particle theory.  If the quantum wave function is probabilistic -- linear superposition is real, and if superposition is real, events are non-local.

Taking space-time as real, superposition is a mathematical artifact, not real.  Where the wheels went off the track, long ago, is in separating Einstein realism from locality -- normalizing time and making its role disappear.  We are learning, though, how the non-linearity of time contributes to the illusion of entanglement and superposition.

I'm going to reserve comment on Joy's framework for the time being. Except to say that it is analytical and non-linear -- often misrepresented and criticized for what it is not.

Have I told you how delightful it is to have a reasonable discussion?