Majority Opinion on Interpreting Bell's Theorem

[Sherlock]

It is the essence of scientific work: put all you can at work to FALSIFY what you try to show :-)

You're missing the irony. As far as we know about what
experimenters *need* to prepare in order to get violations of
Bell inequalities (ie., entanglement) it's that paired detections
(ie., the result of what's been jointly analyzed in a given coincidence
interval) *must* be associated with polarizer-incident disturbances
that were emitted by the same oscillator. So, it's *necessary*
to jointly analyze emissions from the same oscillator in order
to show that these same emissions couldn't have come from
the same oscillator. :-) (Unless what's actually being jointly analyzed
isn't what's normally assumed as being jointly analyzed. And,
this is what I think is happening. Hence, violations of Bell
inequalities aren't actually telling us what we would conclude they're
telling us from the usual assumptions of what's being jointly
analyzed.)

... and you're confusing yourself, I think, by your "global context" semantics. But we've been through that already a few times...
cheers,
Patrick.

I think that looking just at the composite system, rather
than it's parts, is the key to understanding what's happening,
and why we shouldn't take these experiments as evidence
of a direct causal relationship between A and B.

 

[vanesch]

I think that looking just at the composite system, rather
than it's parts, is the key to understanding what's happening,
and why we shouldn't take these experiments as evidence
of a direct causal relationship between A and B.

Nobody said that these experiments are evidence of a direct causal relationship between A and B. However, they are also (as of Bell) not the result of a common cause *that didn't know yet of what was the final setup*, in that a (hidden or not) bag of information that each individual part carries with it, and has to face a final setup that is only determined after the little bags are under way and cannot adapt their contents anymore, cannot explain the result. That is exactly the contents of Bell's theorem.
Now, of course, as you suggest, that looking at the composite system, and *knowing what you are going to measure* is allowed for, this is sufficient to understand what's happening ; it is exactly what quantum mechanics does. But this implies using stuff that is only known over space-like intervals (indeed, as the setup can be decided over spacelike intervals, and if you have to use this global information to determine the outcomes in your "composite system" you are dealing with something that is highly non-local.
That's the entire mystery. You need, indeed, to use the composite system, including the source and the setup ; but the setup has only been decided for, independently, over spacelike intervals. If you do that, there's no problem. But you can hardly call then the theory "local". And people like local stuff, Einstein especially.

 

[Sherlock]

Nobody said that these experiments are evidence of a direct causal
relationship between A and B.

Isn't that how the term "nonlocality" is being used in this thread
(at least by some), that, at spacelike separations, events at A are
affecting events at B and vice versa?

However, they are also (as of Bell) not the result of a common cause *that didn't know yet of what was the final setup*, in that a (hidden or not) bag of information that each individual part carries with it, and has to face a final setup that is only determined after the little bags are under way and cannot adapt their contents anymore, cannot explain the result. That is exactly the contents of Bell's theorem.

If the *hidden* global information is always the same, then time-varying
the joint analyzer settings, or increasing the separation of the
analyzers should not affect the results, and this is what is observed.
The assumption is that this nonvarying hidden global information is produced
locally via emission, and the experiments don't contradict that assumption.

If the information that you're talking about is, say, the rotational information
that varies from pair to pair, and if that information is what is causing
the variable joint results, then (as of Bell, and experiments) the only alternative
is that there is a non-Lorentz-invariant, causal relationship between
events at A and events at B if the paired events are spacelike separated.

This view depends on the assumption that what is assumed
as being jointly analyzed in the formulation of the inequality
(eg., a specific, variable global rotational property of paired
disturbances incident on the polarizers) is actually what's
being jointly analyzed in the experiments.

However, if that's not what's being jointly analyzed in
the experiments, then violations of the inequality are
superfluous wrt a certain interpretation regarding the
physical meaning of the violations (the existence or
not of a local cause of the hidden global information).

Suppose that the hidden "bags of information" are carrying
rotational information that is relevant in determining individual
results, and that they are also carrying another sort of
information that's not relevant to individual results, but
is relevant to joint results. This *global* information is also
produced via emission, due to conservation of angular
momentum, but isn't varying from pair to pair. (This
is why a detection at one end *seems* to allow a
refinement of the probability of the individual result
at the other end for that coincidence interval. But, that
isn't really what's happening. The probability of individual
detection never changes. It's just that given a certain
Theta a certain percentage of the joint results will be
coincidences.)

From the assumption of a local hidden constant due to
conservation of angular momentum there would follow a
formulation *different* from Bell's.

That is, the joint results are due in part (ie., it's a necessary
prior condition) to a locally caused, common property of the
incident disturbances that doesn't change from emission enroute
to the polarizers, isn't changed at the other end after the first
result of a pair, and doesn't change from pair to pair.

This seems to be what qm assumes, and what experimenters
are looking to produce in their preparations.

Now, of course, as you suggest, that looking at the composite system, and *knowing what you are going to measure* is allowed for, this is sufficient to understand what's happening ; it is exactly what quantum mechanics does. But this implies using stuff that is only known over space-like intervals (indeed, as the setup can be decided over spacelike intervals, and if you have to use this global information to determine the outcomes in your "composite system" you are dealing with something that is highly non-local.

Well of course determining joint results (or system-dependent, or
composite, or global, or however one might say that we're
dealing with correlating the behavior of two or more spatially
separated events) requires composite (global) information.
That by itself doesn't mean that "something nonlocal is
going on", if by "nonlocal" we mean that the separated
events are causally related *to-each-other*. (And, if we're
not saying *that*, then what's the problem?)

In the typical optical Bell experiments, it's assumed that what is going
to be (jointly) measured is not varying from pair to pair.
This is the entanglement at the submicroscipic level. It's due to
conservation of angular momentum (so long as they're dealing
with paired photons associated with disturbances emitted from the same
oscillator).

The fact that the relationship defined by the conservation
law doesn't change over spacelike separations, and that
this is what the observational variable, Theta, is actually
analyzing, jointly, at spacelike separations,
means that a causal relationship between events at
A and events at B is not necessary to explain the results.

If nonlocality means that events at A and B are causally
related to each other (which is how at least some of the
respondents in this thread have been using the term),
then, per a correct interpretation of the meaning of
Bell inequalities, there's no evident nonlocality. If
nonlocality just means system-dependent then of course
*all* systems exhibit nonlocal behavior. The mystery lies
in the fact that wrt some systems (eg., gravitational ones) the
local causal determinants are hidden from us. In Bell tests they're
also hidden, but in a way that is at least a bit less
mysterious than with gravitational systems.

That's the entire mystery. You need, indeed, to use the composite system, including the source and the setup ; but the setup has only been decided for, independently, over spacelike intervals. If you do that, there's no problem. But you can hardly call then the theory "local". And people like local stuff, Einstein especially.

And none of this is nonlocal in any sense that that means that
the relationship or entanglement being analyzed by the separated
polarizers wasn't produced via a local common cause.

The assumption that's being contradicted (in Bell tests) is not the
assumption of a local common cause (ie., emission from the same oscillator)
of the shared and invariant property (a global constant) of the incident
disturbances. (In fact, this assumption is actually supported by
experimental violations of Bell inequalities, insofar as they're
used as evidence of the presence of entanglement.)

The assumption that's being contradicted is the assumption
that the specific variable rotational properties of pairs are determining
the joint results. Bell has convincingly ruled *this* assumption out.

In summary, if the statement "nature is nonlocal" just means that
nature is comprised of systems, then considerations regarding
the *existence* of local hidden variables or non-Lorentz-invariant
causal relationships are (wrt what is currently known) obviated.

If the statement "nature is nonlocal" means that nature is
non-Lorentz-invariant, then, since there isn't any evidence for
that, the statement is (wrt what is currently known) wrong.

Or maybe better to say, in the words of Pauli, not even wrong,
since it is primarily due to an incorrect assessment of the
relevance of Bell inequalities to this consideration..

 

[NateTG]

To be fair, I think every permutation of the situation is equally preposterous and that is what makes the debate so interesting.

Very nicely put. Next week we might discuss how many angels dance on the head of a pin.

 

[Sherlock]

To be fair, I think every permutation of the situation is equally preposterous and that is what makes the debate so interesting.

Very nicely put. Next week we might discuss how many angels dance on the head of a pin.

The permutations of the situation aren't equally preposterous, and it's
not a "how many angels dance on the head of a pin" sort of discussion.
However the relevant point is subtle because we've gotten used to
focusing on some assumptions that aren't relevant wrt joint results.

It's pretty well supported that what spacelike separated polarizers are jointly analyzing is a relationship (an 'entanglement') between paired incident disturbances that's due to conservation of angular momentum. This is *not* a hidden *variable*. If a formulation assumes that a global hidden *variable* is determining joint results, then that formulation will *not* agree with qm or experiment, and it's *that* (the variability of the hidden parameter) assumption (not the assumption of local reality) that is being contradicted by qm and experiment.

 

[owl3951]

Information Source Request

Does anyone know where I can download the original raw observations database, and a complete description of the physical layout of the experiment? Within complete description, I include how far the "detector" machines are from each other, how far they are from the "emitter" machine, the electrons' velocity with respect to the detector machines, with respect to the emitter machine, and, if it has been worked out, the electrons' velocity with respect to one another, and the verified angle of the electron's travel with respect to the emitter and the detector machines?

In other words, everything needed to follow and analyze the experiment...I understand this has been done by lots of people, but I would like to do it for myself.