Is there any hope at all for Locality?

[bohm2]

An interesting and useful paper I found on an overview on RBW and a somewhat related interpretation (Genuine Fortuitousness-GF) is the following paper by Daniel Peterson:

Genuine Fortuitousness, Relational Blockworld, Realism, and Time
http://www.johnboccio.com/research/quantum/Dan.pdf

He discusses some conceptual difficulties of RBW/GF. I found this paragraph interesting:

On the flip side, a world where the only real things are relations sinks into an infinite regress since everything exists only in terms of relations among other smaller entities. Such a view would be "turtles all the way down". However, by providing a ground for being in space-time and its geometry, RBW and GF succeed in accounting for a kind of realism (about objects now, not about scientific knowledge) concerning the world. Both are capable of "saving the appearances" while at the same time explaining the quantum world. However, calling both space-time itself and relations within it "real" seems contradictory; if reality is relational, how can a "thing in itself" like space-time exist without relation to anything else? And if space-time is allowed to exist as a "thing in itself", why should not certain entities within space-time be allowed to exist in the same way? There is a tension between these two views that neither GF nor RBW is fully able to resolve.

If I understand this point by the author, he is arguing that it's difficult to conceptualize how something can consist of nothing but “relational structure” all the way down to the "bottom". This is also a criticism of other relational interpretations like Rovelli's "relational QM". Don't relations need relata or intrinsic properties on some level, to ground them? It seems to me, that one can argue that things can't be relational all the way down? Then again, I might be misunderstanding the relationalism in RBW.

 

[stevendaryl]

If you're talking specifically about EPR or EPR-Bohm correlations (as opposed to, say, CHSH-type correlations), then they don't violate any Bell inequality and it's a fairly simple exercise to come up with a toy local model that reproduces them.

I'm not sure whether we're talking about the same thing, or not. The correlations that I'm talking about is the prediction, for spin-1/2 twin pair EPR experiments, that the correlation between Alice's measured result A and Bob's measured result B is given by:

\langle A B \rangle = - cos(\theta)

where \theta is the angle between the detector orientations of Alice and Bob. You're saying that there is a toy model that reproduces this? I was thinking there was a proof that there was not (I have proved it to myself to my own satisfaction).

 

[rkastner]

...

Einstein appears to conflate (or at least highlight) several different notions of “local” in said passage, including but not limited to, (1) local as localized in spacetime, (2) local as possessing primitive thisness with intrinsic properties, (3) local as in no faster than light interactions and (4) local as in being otherwise independent (e.g., statistically) of entities at other points in spacetime. Our beables (spacetimesources) are only local in the first and third senses. Our beables are of spacetime but not in spacetime. That is, spatiotemporal relations and observables (source values) are completely co-existent on our view. For us, classical spacetime also emerges from something more fundamental, but the sense of emergence isn't dynamical it's statistical, and the fundamental reality isn't a dynamical process. For us fundamental explanation is given in terms of adynamical global constraints-a self consistency criterion. As for your claim that dynamical interpretations of the quantum are more in keeping with the formalism of quantum mechanics, it depends on which of the many formalisms you have in mind. In our case it isn't Schrodinger dynamics and wave functions, but rather path integrals, discrete path integrals over graphs. ...

Thanks-- of course I don't take Einstein as the last word on a good definition of 'locality', but when people seek 'local realism', it's usually that sort of Einsteinian view they're talking about.
I should also clarify that PTI does not take the Schrodinger eqn as fundamental since position is not fundamental in PTI. If anything, the momentum representation is more fundamental because it describes the basic field excitations in a relativistically compatible picture.

I've given some reasons in my book as to why I don't think RBW does justice to the QM formalism. I don't think it really succeeds in solving the measurement problem except by a kind of fiat that is antirealist about much of the formalism and which has to create additional complicated formalism, which makes fundamental use of dynamical concepts like momentum/energy (as generators of spacetime translations) that according to RBW don't apply to anything real. I also haven't seen a physical account of the Born Rule except, as I recall, in terms of those symmetry operators (based on dynamical theoretical entities taken as not having dynamical referents). One can of course always choose to be antirealist about the formalism in this way, and nobody can say that you are incorrect to do so, but as you and Michael are aware, I don't think it's methodologcally straightforward nor the most natural way to interpret QM.

In contrast, PTI just says 'yes, the formalism describes reality, and this is the aspect of reality it describes that we weren't aware of previously; and this is why the Born Rule just drops out of it, and this is why you get collapse.'

So it's methodologically much simpler than RBW, and expands our view of reality, in line with previous 'revolutions' in science. That is, people used to think the world was flat and that's all there was. Then they found out the Earth was round and there are really other round objects out there. Then they found out that the Earth was just a planet in a solar system in a galaxy and ours was not the only galaxy. Science has been a continual expansion of our world view. It's a natural part of that progression to suppose that spacetime -- the world of appearance -- is not 'all there is' to reality, and QM describes aspects of reality that are deeper than the spacetime world of appearance. It seems to me that RBW goes in the wrong direction by calling a halt to this scientifically- based expansion of our concepts of reality: it says that spacetime is the only real thing there is) , and the weird phenomena we're seeing are due to weird properties of spacetime as a fundamental substance. But to do that, you have to invent a more complicated formalism to describe those weird aspects of a supposedly substantive spacetime.

We already have a complicated enough formalism -- I'm just suggesting a straightforward realist interpretation of the existing formalism (state vectors, not wave functions) in terms of a broadening of our concepts, which explains the Born Rule and collapse.

Best wishes
Ruth

 

[huelsnitz]

It depends on what you mean by locality. If you mean strange correlations can occur instantaneously then yes locality is dethroned. But that is not what is generally meant by locality which is the ability to actually send information. You can't use QM correlations to do that so locality is saved.

Thanks
Bill

I am still trying to clarify and solidify definitions in my head. But what I generally interpret locality as meaning is that something can only be affected by something in its immediate vicinity. For example, if the sun wants to affect the earth, then it needs to send something (i.e. photons, gravitons) to the earth. In the case of (some) entangled systems, this locality seems to be violated. What happens at one remote location seems to have an essentially immediate impact at another location. And we currently do not know of any mechanism for how that effect is transmitted.

Going a little further to test my understanding - it's my understanding that the inherent randomness in quantum mechanics is what leads to the preservation of causality. I.e., even with the immediacy of wave-function collapse or whatever happens to cause remote events to be entangled, information is still not transferred.


Warren

 

[bhobba]

But what I generally interpret locality as meaning is that something can only be affected by something in its immediate vicinity.

Hi Warren.

That's my view as well.

But the clanger is can that something be used to send information. If it can't then it can't be used to sync clocks which is what SR requires. The type of non locality they have in QM, if you think it's non locality that is, its hotly debated, can't be used to do that.

I can't say I am following too well some of the later posts in this thread. My gut tells me its simply this debate on what exactly non locality means at a more advanced level.

Thanks
Bill

 

[harrylin]

I'm not sure whether we're talking about the same thing, or not. The correlations that I'm talking about is the prediction, for spin-1/2 twin pair EPR experiments, that the correlation between Alice's measured result A and Bob's measured result B is given by:

\langle A B \rangle = - cos(\theta)

where \theta is the angle between the detector orientations of Alice and Bob. You're saying that there is a toy model that reproduces this? I was thinking there was a proof that there was not (I have proved it to myself to my own satisfaction).

Perhaps this one:

http://iopscience.iop.org/1063-7869/39/1/A06/

That paper emphasizes the difference between the non-linear correlation and Bell's theorem with a demonstrator. I had in mind to start a thread on that difference, and maybe that's still useful.

[Edit:] Note that it's a real classical demonstration which had a max. correlation of about 90%, but apparently ideally 100% as the authors comment it as follows:
'The absence of 100% correlations is explained by natural
thermodynamic fluctuations, particularly fluctuations of
the reference voltages of the comparators, which cause
fluctuations at the fronts of the signals being compared."

 

[stevendaryl]

Perhaps this one:

http://iopscience.iop.org/1063-7869/39/1/A06/

That paper emphasizes the difference between the non-linear correlation and Bell's theorem with a demonstrator. I had in mind to start a thread on that difference, and maybe that's still useful.

[Edit:] Note that it's a real classical demonstration which had a max. correlation of about 90%, but apparently ideally 100% as the authors comment it as follows:
'The absence of 100% correlations is explained by natural
thermodynamic fluctuations, particularly fluctuations of
the reference voltages of the comparators, which cause
fluctuations at the fronts of the signals being compared."

Sigh. The abstract seems to be claiming something that is provably false.

 

[harrylin]

Sigh. The abstract seems to be claiming something that is provably false.

Perhaps you misunderstand the abstract; I had to read both the abstract and the paper twice before I started to understand what they are saying (it was translated from Russian into English). What matters here is that they literally state what wle stated and also explain how to do it, although I don't understand that yet. As they apparently proved it right, both in theory and practice, and it was published in high quality journals, it will certainly be worth a discussion. I still intend to start a thread on that in a day of ten, which should also be helpful for those who cannot freely access the paper. It's quite possible that wle has another paper in mind.

 

[wle]

I'm not sure whether we're talking about the same thing, or not. The correlations that I'm talking about is the prediction, for spin-1/2 twin pair EPR experiments, that the correlation between Alice's measured result A and Bob's measured result B is given by:

\langle A B \rangle = - cos(\theta)

where \theta is the angle between the detector orientations of Alice and Bob.

No, then we're not talking about the same thing. I took EPR to refer specifically to the sort of correlations that appear in the EPR or EPR-Bohm argument: two measurements on each side with binary outcomes, perfect correlations (or anticorrelations) when the same measurements are performed, and no correlations when different measurements are performed. In other words,
<br /> \begin{eqnarray}<br /> \langle A_{1} B_{1} \rangle &amp;= \langle A_{2} B_{2} \rangle =&amp; \pm 1 \,, \\<br /> \langle A_{1} B_{2} \rangle &amp;= \langle A_{2} B_{1} \rangle =&amp; 0 \,.<br /> \end{eqnarray}<br />

You're saying that there is a toy model that reproduces this? I was thinking there was a proof that there was not (I have proved it to myself to my own satisfaction).

Well obviously there's no model that will work for all angles, since there's plenty of combinations of angles that will result in a CHSH violation.

 

[DrChinese]

What matters here is that they literally state what wle stated and also explain how to do it, although I don't understand that yet. As they apparently proved it right, both in theory and practice, and it was published in high quality journals, it will certainly be worth a discussion.

It is from 1996, which about says it all. This is far past the expiration date for such a speculative idea. I don't believe it has any citations either. I would not characterize this as suitable for discussion here.

 

[harrylin]

It is from 1996, which about says it all. This is far past the expiration date for such a speculative idea. I don't believe it has any citations either. I would not characterize this as suitable for discussion here.

We often discuss older papers, this one does not speculate but proves its points by derivation and a working physical demonstrator, and it has a number of citations. Thus you are completely off track here.

 

[wle]

Perhaps this one:

http://iopscience.iop.org/1063-7869/39/1/A06/

Sigh. The abstract seems to be claiming something that is provably false.

It is from 1996, which about says it all. This is far past the expiration date for such a speculative idea. I don't believe it has any citations either. I would not characterize this as suitable for discussion here.

I skimmed the paper. It doesn't seem to report anything particularly controversial.

They state (correctly) that the results of the EPR-Bohm thought experiment, involving either perfect correlations or no correlations depending on the settings being considered (as I clarified in [POST=4477984]post #159[/POST]), can be reproduced with a local model. For reasons best known to themselves, they see some need to "demonstrate" this with an experimental setup. They also seem to consider some more general local models which have no hope of ever violating a Bell inequality... and find that they don't get a Bell inequality violation.

Regarding the reproduction of EPR-Bohm correlations, Bell already gave a local model reproducing them in his 1964 paper (and Evdokimov et. al. explicitly cite his paper as such). An even simpler model is to take a hidden variable of the form \lambda = (s_{z}, s_{x}) with s_{z}, s_{x} \in \{-1, +1\}, with outcomes defined by:
<br /> \begin{eqnarray}<br /> A_{1}(\lambda) &amp;= B_{1}(\lambda) =&amp; s_{z} \,, \\<br /> A_{2}(\lambda) &amp;= B_{2}(\lambda) =&amp; s_{x} \,.<br /> \end{eqnarray}<br />
The EPR-Bohm correlations are given by
<br /> \begin{eqnarray}<br /> P(A_{1} = B_{1}) &amp;=&amp; 1 \,, \\<br /> P(A_{2} = B_{2}) &amp;=&amp; 1 \,, \\<br /> P(A_{1} = B_{2}) &amp;=&amp; 1/2 \,, \\<br /> P(A_{2} = B_{1}) &amp;=&amp; 1/2 \,,<br /> \end{eqnarray}<br />
with completely random marginals on each side. To reproduce them, you just draw \lambda = (s_{z}, s_{x}) uniformly at random:
p(\lambda) = p_{z}(s_{z}) p_{x}(s_{x}) = 1/4 \,, \quad \forall \lambda \,.
The whole point of Bell's theorem, of course, is that this sort of strategy won't work for more general correlations predicted by quantum physics.

 

[stevendaryl]

Well obviously there's no model that will work for all angles, since there's plenty of combinations of angles that will result in a CHSH violation.

Oh, okay. I misunderstood what was being claimed.

 

[RUTA]

An interesting and useful paper I found on an overview on RBW and a somewhat related interpretation (Genuine Fortuitousness-GF) is the following paper by Daniel Peterson:

Genuine Fortuitousness, Relational Blockworld, Realism, and Time
http://www.johnboccio.com/research/quantum/Dan.pdf

He discusses some conceptual difficulties of RBW/GF. I found this paragraph interesting:

If I understand this point by the author, he is arguing that it's difficult to conceptualize how something can consist of nothing but “relational structure” all the way down to the "bottom". This is also a criticism of other relational interpretations like Rovelli's "relational QM". Don't relations need relata or intrinsic properties on some level, to ground them? It seems to me, that one can argue that things can't be relational all the way down? Then again, I might be misunderstanding the relationalism in RBW.

You understand the criticism and it’s a valid point. The way we stop the infinite regress in RBW’s version of theory X is by defining observables relationally. Thus, an observation ends the regress, unless you want to define consciousness relationally, but that’s not the realm of physics See Figures 1 and 2 of http://users.etown.edu/s/stuckeym/FOP2013.pdf. [You can use this link until the revised version shows up at http://arxiv.org/abs/0908.4348.] This is the paper we presented at Foundations of Physics 2013 in Munich last month and it will appear in an IOP collection on quantum gravity later this year or early next year.

 

[audioloop]

I am still trying to clarify and solidify definitions in my head. But what I generally interpret locality as meaning is that something can only be affected by something in its immediate vicinity. For example, if the sun wants to affect the earth, then it needs to send something (i.e. photons, gravitons) to the earth. In the case of (some) entangled systems, this locality seems to be violated. What happens at one remote location seems to have an essentially immediate impact at another location. And we currently do not know of any mechanism for how that effect is transmitted.

Think first in Contextuality. It will help you a lot.
https://www.physicsforums.com/showthread.php?t=619905
contextuality is broader, subsumes nonlocality.
same thing in quantum information (nonlocality is a generic feature of non-signaling).

 

[RUTA]

I am still trying to clarify and solidify definitions in my head. But what I generally interpret locality as meaning is that something can only be affected by something in its immediate vicinity. For example, if the sun wants to affect the earth, then it needs to send something (i.e. photons, gravitons) to the earth. In the case of (some) entangled systems, this locality seems to be violated. What happens at one remote location seems to have an essentially immediate impact at another location. And we currently do not know of any mechanism for how that effect is transmitted.

The way we get around this dilemma is to first, think in terms of a blockworld and second, decompose that blockworld relationally rather than dynamically. An analogy would be a spacetime picture of some phenomenon decomposed via a jigsaw puzzle. The shape of the pieces has nothing to do with the picture that is created when it's assembled, but each piece does contain some portion of the finished picture and that information can be valuable in assembling the puzzle. In a dynamical decomposition of the picture, the pieces would be the 3D objects involved in the phenomenon and they would be placed *into* a pre-existing spacetime frame. Thinking dynamically, QM phenomena are mysterious, but if you accept that the fundamental rule of physics is like that of the jigsaw puzzle, then QM phenomena are no problem.