Mermin on Spooky action at a distance

[Niles]

Mermin on "Spooky action at a distance"

Hi all.

I've read Mermin's article on EPR and non-locality: www.physics.iitm.ac.in/~arvind/ph350/mermin.pdf[/URL]

I don't understand what he write on page 12: "Did the particle at A have its 3-color prior to the measurement of the 3-color of the particle at B? The answer cannot be yes, because, prior to the measurement of the 3-color at B, it is altogether possible that the roll of the dice at B or the whim of the B-operator will result in the 2-color or the 1-color being
measured at B instead. Barring the most paranoid of conspiracy theories, “prior to the measurement of the 3-color at B” is indistinguishable from “prior to the measurement of the 2- (or 1-) color at B”. If the 3-color already existed, so also must the 2- and 1-colors have existed. But instruction sets (which consist of a specification of the 1-, 2-, and 3-colors) do not exist."

If our detector is set to position 3, it measures if our particle has that property "3". We know that the particles do not carry instruction sets, but surely the particles know what state they are in. So if e.g. particles with property "2" are sent out and the detectors are both at position 3, then we get red at both detectors.

But where does his explanation come into this?

 

[DrChinese]

If our detector is set to position 3, it measures if our particle has that property "3". We know that the particles do not carry instruction sets, but surely the particles know what state they are in. So if e.g. particles with property "2" are sent out and the detectors are both at position 3, then we get red at both detectors.

But where does his explanation come into this?

There are a couple of things going on. Clearly, if we ask the same question of both particles, we get the same answer.

The problem is that we might ask different questions, and the results are clearly affected - in some way - by the specific 2 different questions we ask. In the example, we would expect that we would get the same answer at *least* 1 in 3 times. But in practice, it is only 1 in 4. The implication is that particles take on properties in the context of how they are observed.

You mention that particles do not carry instruction sets. Yet you also mention that they must know what state they are in. And yet, that state is not determined prior to a measurement. So the idea that "particles with property "2" are sent out and the detectors are both at position 3" does not fit with the experimental facts. Unless, of course, they are in non-local communication or causal contact.

 

[ueit]

Barring the most paranoid of conspiracy theories, “prior to the measurement of the 3-color at B” is indistinguishable from “prior to the measurement of the 2- (or 1-) color at B”. If the 3-color already existed, so also must the 2- and 1-colors have existed. But instruction sets (which consist of a specification of the 1-, 2-, and 3-colors) do not exist.

There is nothing "paranoid" about the assumption that the "color" of the particle is related to the way this color is measured. For example, it might be that the detector itself produces a long range field that influence the colors of the entangled particles as they are "produced" at the source. If the detector is set to 3, you have a field; for a 2 setting you have a different field. Different fields, different colors.

Just ask the same question for a classical theory, like gravity. Say an object approaches the solar system from far away. Does the trajectory of that object (even when still far away) depend on the planets' configuration or not?

 

[DrChinese]

There is nothing "paranoid" about the assumption that the "color" of the particle is related to the way this color is measured. For example, it might be that the detector itself produces a long range field that influence the colors of the entangled particles as they are "produced" at the source. If the detector is set to 3, you have a field; for a 2 setting you have a different field. Different fields, different colors.

Of course, there are severe limits on this "field" explanation: it would need to be non-local, for one. And there appears to be no other evidence for such a field other than to explain entanglement.

Now, please note that Quantum Mechanics already fully accounts for the observed phenomena.

 

[ueit]

Of course, there are severe limits on this "field" explanation: it would need to be non-local, for one.

No, that's not true. You make a confusion between locality (which requires that no influence can propagate faster than light) and independent evolution of two distant systems , like the source of entangled particles and the detector (which requires that no influence exists whatsoever). Two systems that interact through a local field are still not independent of each other. As an example take classical electrodynamics or general relativity.

And there appears to be no other evidence for such a field other than to explain entanglement.

Such a field would explain not only entanglement but also all other non-intuitive aspects of QM. What evidence other than this would you expect to find for such a field?

Now, please note that Quantum Mechanics already fully accounts for the observed phenomena.

What do you mean by "Quantum Mechanics" and "fully accounts"? Do you have a speciffic interpretation in mind? Do you agree that the spot produced by a single particle in a double-slit experiment is an "observed phenomena"?

 

[DrChinese]

No, that's not true. You make a confusion between locality (which requires that no influence can propagate faster than light) and independent evolution of two distant systems , like the source of entangled particles and the detector (which requires that no influence exists whatsoever). Two systems that interact through a local field are still not independent of each other. As an example take classical electrodynamics or general relativity.

You know perfectly well that the settings of the detectors can be changed mid-flight without affecting the results in any way. If such a hypothetical field existed, it would need to transmit "something" to the other particle, and vice versa, so that they could yield "answers" that are consistent with the predictions of QM. So yes, there are severe restrictions and one of those is that any such mechanism must contain a non-local component.

 

[RandallB]

Of course, there are severe limits on this "field" explanation: it would need to be non-local, for one.

No, that's not true. You make a confusion between locality …. and independent evolution of two distant systems , like the source of entangled particles and the detector (which requires that no influence exists whatsoever). Two systems that interact through a local field are still not independent of each other.

Here I can only assume you are using the interdependence of Super Determinism and that is well understood to be non-local.

As an example take classical electrodynamics or general relativity.

The fields of ‘classical electrodynamics’ must be assumed to be continuous meaning the effect caused by a field at different points is instantaneous which is just as non-local as Newtonian instantaneous gravity.

General Relativity has the same non-local problem with continuous gravitational fields – the main difference in GR is we usually do not call that problem a local vs. non-local issue; rather a “dependent vs. independent background” issue (see Smolin; Perimeter Institute). Here again there is nothing to refute the conclusion that GR requires an “independent background” which for the discussion here means non-local; as in unable to resolve things like entanglement or other non-intuitive aspects of QM.

Now, please note that Quantum Mechanics already fully accounts for the observed phenomena.

But I do disagree with this statement;
Niels Bohr nor QM claim a “Full Accounting” of observed phenomena.
QM starts from an assumption of HUP that specifically rejects considering or describing individual particle behaviors during the measurement process.
The claim for QM is that no realistic description of behaviors during the measurement process can produce such a more complete full accounting than that given by HUP/QM.

 

[ueit]

You know perfectly well that the settings of the detectors can be changed mid-flight without affecting the results in any way.

If by "results" you mean statistical results then I agree. We don't know if individual results are changed because QM doesn't say much about them.

If such a hypothetical field existed, it would need to transmit "something" to the other particle, and vice versa, so that they could yield "answers" that are consistent with the predictions of QM.

Both detectors produce a field. Their combined field has some value at the location of the source. The particle spin, on each axis, is a function of the local value of the field in that place. No non-locality required.

If you chose to change the detectors' settings then you have to provide some device that is capable of doing that (say an electric engine). That device itself produces a field at the particle source location so you will get different particle spins than in the case of a static detector. In order to find out what the spins are you need to calculate the field for each experimental setup.

So yes, there are severe restrictions and one of those is that any such mechanism must contain a non-local component.

There are restrictions, indeed, but non-locality is not one of them. The only way to impose such a restriction is to make the assumption that QM is fundamentally indeterministic. In this case you can argue that the same field could corresponds to different detector settings so a local field is excluded. However I see no need to make such an assumption. I choose determinism+ locality instead non-determinism+ non-locality combo.

 

[ueit]

Here I can only assume you are using the interdependence of Super Determinism and that is well understood to be non-local.

I don't think that there is a "interdependence of Super Determinism" as opposed to other types of interdependence. I'm trying to avoid the explicit use of the word "Super Determinism" in order not to deviate the discussion on pure philosophical notions like free will. So, if you want to use the term, I'm OK but let's discuss the assumption of a long-range local field that I've proposed. By "local field" I mean that a perturbation in that field propagates at a finite speed (c) and only the value of the field at the point where the particle is located exerts an effect upon that particle.

The fields of ‘classical electrodynamics’ must be assumed to be continuous meaning the effect caused by a field at different points is instantaneous which is just as non-local as Newtonian instantaneous gravity.

Why should a continuous field be necessarily non-local?

General Relativity has the same non-local problem with continuous gravitational fields – the main difference in GR is we usually do not call that problem a local vs. non-local issue; rather a “dependent vs. independent background” issue (see Smolin; Perimeter Institute). Here again there is nothing to refute the conclusion that GR requires an “independent background” which for the discussion here means non-local; as in unable to resolve things like entanglement or other non-intuitive aspects of QM.

Again, why is a continuous field be necessarily non-local? AFAIK GR also has a local character. A massive body "feels" the effect of the space curvature at its location and a perturbation of the curvature travels at a finite speed, c.

 

[Niles]

First I want to thank you all for participating.

My question is a lot more fundemental. This is how I have understood Mermin's article so far:

Two particles in a singlet spin-state are fired at the two detectors, which can measure the direction of spin at angles of 0, 120 and 240 degrees. Mermin then accounts for how why half the time the same color flashes, and this I understand.

Now look at this:

The problem is that we might ask different questions, and the results are clearly affected - in some way - by the specific 2 different questions we ask. In the example, we would expect that we would get the same answer at *least* 1 in 3 times. But in practice, it is only 1 in 4. The implication is that particles take on properties in the context of how they are observed.

This is what I cannot understand. Let's look at the explanation Mermin gives with his "N-colors":

"Did the particle at A have its 3-color prior to the measurement of the 3-color of the particle at B? The answer cannot be yes, because, prior to the measurement of the 3-color at B, it is altogether possible that the roll of the dice at B or the whim of the B-operator will result in the 2-color or the 1-color being measured at B instead."

If we use spin-½ instead of N-colors, then spin in e.g. the z-direction is measured of particle B. But particle A cannot have a pre-determined spin in the z-direction, because if we chose to measure spin-½ in the x or y-direction at detector B, then ...This where I am stuck. I cannot see how we can rule out the possibility of particle A having a pre-determined spin-½ in a particular direction, because we are able to measure spin-½ of particle B in three different directions.

Thanks for enlightening me.

 

[RandallB]

Again, why is a continuous field be necessarily non-local? AFAIK GR also has a local character.

The nature of the fields are defined by the theories that use fields in their descriptions. It is not like a “field” has ever been directly physically observed only; their indirect effects as described by the theories that define the fields.

In order for GR to be “local in character” it needs to demonstrate a “dependent background”. And most if not satisfied with the Smolin point that GR is fundamentally a independent background description, certainly have not been able to demonstrate a dependent background for it.
It is not something you can just assume without foundation.
No more than you can assume the principles of Super Determinism as a local description by not mentioning the word “determinism”.

Only if you can demonstrate how General Relativity and the fields defined by GR can describe reality within a completely “dependent background” could I accept the opinion that GR is “local in character”.

 

[Peter Morgan]

Concerning Ueit's resort to fields, I refer you to my J. Phys. A paper "Bell inequalities for random fields", last pre-print at http://arxiv.org/abs/cond-mat/?0403692 (and thence a link to the published paper). It's not enough to use classical fields, probability also has to be introduced; mixing probability with classical fields results in a need for a mathematics called random fields that have quite intricate properties, which are rather similar to the mathematics of quantum fields. See also my just posted http://arXiv.org/abs/0810.2545 for an attempt at a broader discussion. The standard idea in Physics is that the papers of Bell's that I cite in "Bell inequalities for random fields" rule out classical fields; indeed the argument is not easy to gainsay definitively.

The question of non-locality is very delicate; for a random field model at equilibrium -- in the coarse-grained sense that the statistics of detector events are time-invariant even though in a fine-grained sense the detectors may be switching between different thermodynamic states thousands of times each second -- the global properties of the whole apparatus determine the detailed properties of the equilibrium, just as Bohr says they should. Hence, in a random field perspective it is more a question of holism at equilibrium -- which is reasonable in a classical Physics perspective -- than a question of nonlocality -- which is essentially not so reasonable.

It also should be noted that quantum fluctuations have to be introduced explicitly into random field models to give accounts for experiments in which quantum effects are seen. The import of the need for quantum theory is that the mean field approximation for quantum fluctuations is not always enough.

I note that Dr. Chinese says that "the settings of the detectors can be changed mid-flight without affecting the results in any way"; if there are no particles, however, only fields, there is no such thing as "mid-flight". The detector events are caused by the particular way that the "preparation device" drives the field, which has to be carefully tuned to obtain statistics that are incompatible with a particle property model. Provided the rapid random switching of the detectors is done in a way that does not change the essential right-left symmetries of the overall experiment, there is no reason to think that the statistics should change. If the statistics do change, then the random switching, a priori, does not preserve the symmetries of the apparatus.

I started writing this before Niles latest comment was posted. Seeing it's principal point, I note that in a field perspective, no event can possibly happen without a macroscopic detector being present. "The implication is that particles take on properties in the context of how they are observed", that is, only when a detector -- which is a delicately tuned macroscopic object that makes thermodynamic transitions repeatedly over time -- is placed in the right place. No detector, no events.

I hope this is useful to you, Niles, however this post is more comprehensible in terms of the discussion on fields between Ueit and Dr. Chinese. Thinking in terms of classical particles and their properties gets you into something of a mess, the thinking of the last 70 years pretty much shows (unless you want or are willing to go truly nonlocal with de Broglie-Bohm or similar models), so: go to fields. Sadly, that has to be random fields, not the more straightforward continuous fields. Good luck.

 

[Peter Morgan]

Regarding superdeterminism, I point out that quantum field theory is superdeterministic in the sense that a quantum state that determines the probabilities of various measurements now also determines the probabilities of whatever measurements we might make in the past.

If we restrict to probabilistic classical models, in other words to random fields, the state of the random field is no more superdeterministic than a quantum field state. Only probabilities in the past are determined by models in a random field formalism, not individual events. Indeed, we can present a random field as a commutative quantum field.

Most Physicists take quantum field theory to be the real thing, while nonrelativistic finite-dimensional Hilbert spaces are just approximations; but perhaps quantum field theory is unacceptable because our knowledge of the quantum state now partially determines how we suppose the world was, probably, in the past?

 

[Niles]

Although I really appreciate your posts, my question is a lot more fundemental. I simply do not understand (litteraly!) his explanations.

 

[DrChinese]

First I want to thank you all for participating.

My question is a lot more fundemental. This is how I have understood Mermin's article so far:

Two particles in a singlet spin-state are fired at the two detectors, which can measure the direction of spin at angles of 0, 120 and 240 degrees. Mermin then accounts for how why half the time the same color flashes, and this I understand.

Now look at this:This is what I cannot understand. Let's look at the explanation Mermin gives with his "N-colors":

"Did the particle at A have its 3-color prior to the measurement of the 3-color of the particle at B? The answer cannot be yes, because, prior to the measurement of the 3-color at B, it is altogether possible that the roll of the dice at B or the whim of the B-operator will result in the 2-color or the 1-color being measured at B instead."

If we use spin-½ instead of N-colors, then spin in e.g. the z-direction is measured of particle B. But particle A cannot have a pre-determined spin in the z-direction, because if we chose to measure spin-½ in the x or y-direction at detector B, then ...This where I am stuck. I cannot see how we can rule out the possibility of particle A having a pre-determined spin-½ in a particular direction, because we are able to measure spin-½ of particle B in three different directions.

Thanks for enlightening me.

Yes, if you stick to angles that are perpendicular for spin 1/2, then the problem is not apparent. But that is not what makes Bell's Theorem work, and same for Mermin's version. I might recommend we stick to the photon (spin 1) version because it is easier to see. But it would also work for spin 1/2, as that is what Bell used.

Bell noted that at certain detector angles, you would be observing a mix of the degrees of freedom. At 3 angles spread across 2 degrees of freedom, he discovered that it would not be internally consistent as to a "simultaneous realistic" answer for all 3 settings. So that is the problem: you cannot have a logically consistent set of answers that match experiment.

So again, let's look at Mermin's 3 measurement positions. A measurement of one particle at position 1 guarantees that the OTHER particle's answer at position 2 or 3 will be "as if" it had already been measured at position 1. What's so weird about that? Well, that implies that there might be some kind of spooky action at a distance. And how does that manifest itself? It shows up as a statistical relationship (.25) that is lower than the lower bound should "logically" be (.33).

 

[DrChinese]

Although I really appreciate your posts, my question is a lot more fundemental. I simply do not understand (litteraly!) his explanations.

Not sure if that will make things any easier, but I have a web page that is devoted to explaning Bell using Mermin's ideas:

Bell's Theorem with Easy Math

I hope that helps too. Keep your questions coming...

 

[JesseM]

If it helps, I came up with another analogy similar to Mermin's on this thread:

The key to seeing why you can't explain the results by just imagining the electrons had preexisting spins on each axis is to look at what happens when the two experimenters pick different axes to measure. Here's an analogy I came up with on another thread (for more info, google 'Bell's inequality'):

Suppose we have a machine that generates pairs of scratch lotto cards, each of which has three boxes that, when scratched, can reveal either a cherry or a lemon. We give one card to Alice and one to Bob, and each scratches only one of the three boxes. When we repeat this many times, we find that whenever they both pick the same box to scratch, they always get opposite results--if Bob scratches box A and finds a cherry, and Alice scratches box A on her card, she's guaranteed to find a lemon.

Classically, we might explain this by supposing that there is definitely either a cherry or a lemon in each box, even though we don't reveal it until we scratch it, and that the machine prints pairs of cards in such a way that the "hidden" fruit in a given box of one card is always the opposite of the hidden fruit in the same box of the other card. If we represent cherries as + and lemons as -, so that a B+ card would represent one where box B's hidden fruit is a cherry, then the classical assumption is that each card's +'s and -'s are the opposite of the other--if the first card was created with hidden fruits A+,B+,C-, then the other card must have been created with the hidden fruits A-,B-,C+.

The problem is that if this were true, it would force you to the conclusion that on those trials where Alice and Bob picked different boxes to scratch, they should find opposite fruits on at least 1/3 of the trials. For example, if we imagine Bob's card has the hidden fruits A+,B-,C+ and Alice's card has the hidden fruits A-,B+,C-, then we can look at each possible way that Alice and Bob can randomly choose different boxes to scratch, and what the results would be:

Bob picks A, Alice picks B: same result (Bob gets a cherry, Alice gets a cherry)

Bob picks A, Alice picks C: opposite results (Bob gets a cherry, Alice gets a lemon)

Bob picks B, Alice picks A: same result (Bob gets a lemon, Alice gets a lemon)

Bob picks B, Alice picks C: same result (Bob gets a lemon, Alice gets a lemon)

Bob picks C, Alice picks A: opposite results (Bob gets a cherry, Alice gets a lemon)

Bob picks C, Alice picks picks B: same result (Bob gets a cherry, Alice gets a cherry)

In this case, you can see that in 1/3 of trials where they pick different boxes, they should get opposite results. You'd get the same answer if you assumed any other preexisting state where there are two fruits of one type and one of the other, like A+,B+,C-/A-,B-,C+ or A+,B-,C-/A-,B+,C+. On the other hand, if you assume a state where each card has the same fruit behind all three boxes, like A+,B+,C+/A-,B-,C-, then of course even if Alice and Bob pick different boxes to scratch they're guaranteed to get opposite fruits with probability 1. So if you imagine that when multiple pairs of cards are generated by the machine, some fraction of pairs are created in inhomogoneous preexisting states like A+,B-,C-/A-,B+,C+ while other pairs are created in homogoneous preexisting states like A+,B+,C+/A-,B-,C-, then the probability of getting opposite fruits when you scratch different boxes should be somewhere between 1/3 and 1. 1/3 is the lower bound, though--even if 100% of all the pairs were created in inhomogoneous preexisting states, it wouldn't make sense for you to get opposite answers in less than 1/3 of trials where you scratch different boxes, provided you assume that each card has such a preexisting state with "hidden fruits" in each box.

But now suppose Alice and Bob look at all the trials where they picked different boxes, and found that they only got opposite fruits 1/4 of the time! That would be the violation of Bell's inequality, and something equivalent actually can happen when you measure the spin of entangled photons along one of three different possible axes. So in this example, it seems we can't resolve the mystery by just assuming the machine creates two cards with definite "hidden fruits" behind each box, such that the two cards always have opposite fruits in a given box.

I also showed how this example could be applied to a different Bell inequality in post #8 of this thread if you're interested.

 

[DrChinese]

This where I am stuck. I cannot see how we can rule out the possibility of particle A having a pre-determined spin-½ in a particular direction, because we are able to measure spin-½ of particle B in three different directions.

JesseM's example above is good too.

Returning to your question above, consider this... suppose you measure an electron at 0 degrees (call that x-axis), and follow it with another measurement at 30 degrees. The second measurement would be a mix of the x and y axes presumably. In other words, we had earlier established that there is an element of reality (by EPR's definition) associated with any possible angle setting because there are "perfect" correlations between entangled particles at any identical angle setting.

So if we are measuring a mix of axes, what does that mean? In your example, you picked the extremes where there is no connection between measurement results (axes are separated by 90 degrees, for spin 1/2 particles). For a spin 1 photon, that would be a separation of 45 degrees to get analogous results instead of 90 degrees. But the question you need to ask is: if you measure in between, a mix of 2 degrees of freedom (i.e. 2 axes), what happens then? That is what Bell asked. He realized that QM's predictions were internally inconsistent IF there were simultaneous reality to all possible observable angle settings. That led to Bell's Theorem. Mermin's presentation just makes the math easier to see. Or maybe not... LOL.

 

[Niles]

JesseM's example is very good.

Mermin's explanation of the math is not too bad, I think. So now I fully agree that there is no way that pre-determined instruction sets can explain the results that we get. Also I fully understand Mermin's example, where he uses spin-½ to explain the results.

So is the conslusion: We can show that there are no instruction sets, because then the same color (I am using Mermin's example) woud flash 5/9 of all runs, whereas they only flash 1/2 of all runs. So no instruction sets.

Also, one way of building the device is to use spin-½. This configuration also accounts for the data. So since there are no instruction sets, the particles do not have pre-determined spin?

 

[DrChinese]

So since there are no instruction sets, the particles do not have pre-determined spin?

That is safe to say. Even in non-local theories, the results are contextual and therefore not predetermined. The results are somehow shaped by the specific observation settings chosen, which is done at a later time.

 

[Niles]

Great, I understand the words now and the math. Thanks for that, everyone.

 

[Autochthon]

That is safe to say. Even in non-local theories, the results are contextual and therefore not predetermined. The results are somehow shaped by the specific observation settings chosen, which is done at a later time.

This is exactly what I've been puzzling over lately (the perspective of the observation effects the result). I somehow miss the connection between Bell's Inequality and this conclusion. I'd always thought that Bell himself concluded that the inequality did no more than exclude locality. By contextual do you mean you are restricting the result to a single outcome rather than the family of outcomes encompassed by the possible perspectives? Why might not the full set of possible observations describe the object?

 

[DrChinese]

This is exactly what I've been puzzling over lately (the perspective of the observation effects the result). I somehow miss the connection between Bell's Inequality and this conclusion. I'd always thought that Bell himself concluded that the inequality did no more than exclude locality.

I think Bell envisioned a non-local theory as the likely outcome originally. But there has been a lot happen since then. GHZ is one example. The Weihs (and Aspect as well, which Bel knew about) experiment too. But there has been no really non-local model to provide any reasonable description which is non-contextual (i.e. realistic). They all talk about a mixed system in which the measuring apparatus influences the results. Well, that is contextual and not "realistic" in the EPR sense.

A realistic theory, on the other hand, will be able to provide answers for the question: what are the correlation probabilities for THREE simultaneous measurement positions. But we already know there are no such non-negative solutions. So I have no issue with a non-local theory, I just assert that a non-local realistic theory is not yet presented for out consideration that can meet this criterion.

The dBB-type solutions claim to be deterministic, but I am not sure that is really a fair characterization consistent with experiement. There must be a non-local signal (between Alice and Bob) in such a theory to account for the fact that the measurement settings can be changed mid-flight. I would not call that "pre-determined", quite the opposite.

 

[Ken G]

A realistic theory, on the other hand, will be able to provide answers for the question: what are the correlation probabilities for THREE simultaneous measurement positions. But we already know there are no such non-negative solutions. So I have no issue with a non-local theory, I just assert that a non-local realistic theory is not yet presented for out consideration that can meet this criterion.

Can you expound on this point? It sounds like you mean we have three entangled particles, and want the correlations among three measurements. Certainly quantum mechanics allows us to calculate those probabilistic correlations for a given set of observational settings. So what do you mean that there are no non-negative solutions?

The dBB-type solutions claim to be deterministic, but I am not sure that is really a fair characterization consistent with experiement. There must be a non-local signal (between Alice and Bob) in such a theory to account for the fact that the measurement settings can be changed mid-flight. I would not call that "pre-determined", quite the opposite.

I think you are raising an important issue about what "pre-determined" should mean. The most natural interpretation is what you are using, which sounds like, it is pre-determined if you can know in advance (in principle, even if our science cannot actually get these answers), and at the same time, what will be the answer to any set of questions you can put to the system, whether or not those questions are ever actually posed to the system. But there is a more flexible version of pre-determined which might just assert that we cannot know (even in principle) simultaneously the answer to all the possible questions, but we can know in advance (in principle) the answer to anyone given set of questions. Knowing the answer (in advance) to those questions could preclude knowing the answer to certain other questions, but that can still be a version of "determinism", albeit a much more subtle one.

In other words, we must ask what is actually being "determined" in a "deterministic" reality. Does it have to be all things to all people? Or can we allow reality to only be able to determine the answer to questions that are actually posed. By which I mean, it is quite possible (and seems likely) that the posing of questions is not a purely hypothetical exercise, it is itself part of the reality we are trying to understand. We should not conclude reality cannot be deterministic simply because we have not yet determined what questions are going to be posed, that is not reality's fault. It can only determine what answers must appear to whatever questions are actually posed, as the latter is part of reality too, and reality cannot be self-consistently required to address unreal events that do not in fact occur.

 

[DrChinese]

(Originally Posted by DrChinese

A realistic theory, on the other hand, will be able to provide answers for the question: what are the correlation probabilities for THREE simultaneous measurement positions. But we already know there are no such non-negative solutions. So I have no issue with a non-local theory, I just assert that a non-local realistic theory is not yet presented for out consideration that can meet this criterion.)

1. Can you expound on this point? It sounds like you mean we have three entangled particles, and want the correlations among three measurements. Certainly quantum mechanics allows us to calculate those probabilistic correlations for a given set of observational settings. So what do you mean that there are no non-negative solutions?


2. I think you are raising an important issue about what "pre-determined" should mean. The most natural interpretation is what you are using, which sounds like, it is pre-determined if you can know in advance (in principle, even if our science cannot actually get these answers), and at the same time, what will be the answer to any set of questions you can put to the system, whether or not those questions are ever actually posed to the system. But there is a more flexible version of pre-determined which might just assert that we cannot know (even in principle) simultaneously the answer to all the possible questions, but we can know in advance (in principle) the answer to anyone given set of questions. Knowing the answer (in advance) to those questions could preclude knowing the answer to certain other questions, but that can still be a version of "determinism", albeit a much more subtle one.

In other words, we must ask what is actually being "determined" in a "deterministic" reality. Does it have to be all things to all people? Or can we allow reality to only be able to determine the answer to questions that are actually posed. By which I mean, it is quite possible (and seems likely) that the posing of questions is not a purely hypothetical exercise, it is itself part of the reality we are trying to understand. We should not conclude reality cannot be deterministic simply because we have not yet determined what questions are going to be posed, that is not reality's fault. It can only determine what answers must appear to whatever questions are actually posed, as the latter is part of reality too, and reality cannot be self-consistently required to address unreal events that do not in fact occur.

1. Einstein said the moon was there even if we are not observing it. EPR says that observable attributes that can be predicted with certainty without disturbing a system have an element of reality. I think these can be combined reasonably to conclude that a photon - per Einstein - has a definite polarization at all possible angles even if they cannot be known simultaneously. So let's just say (to be specific) that we want to assert that a photon has a definite polarization value at 3 angle settings. We know we can used an entangled photon pair to learn 2 of the 3 possible values. But Bell's Theorem shows us that there is an internal inconsistency in that argument (that there are 3), because the quantum mechanical predictions lead to predictions of negative probabilities at some angle settings.

Unfortunately, there is currently no 3 particle entanglement scenarios in which the so-called "perfect correlations" possible. There is 3 particle entanglement but different statistics apply.

So what I am saying is that any theory which claims to be "realistic" or "hidden variable" or similar should come out of the closet and tell us what the statistics are for one of these "forbidden" angle combinations: say A=0, B=120 and C=240 degrees for example. I would like someone to explain a scenario in which the correlation is 25% between any 2 of these 3 settings while at the same time the perfect correlations hold as well.

The internal (realistic) inconsistency is: IF the AB Correlation percentage is .25 (matching experiment and QM predictions), and the BC Correlation percentage is .25, THEN the AC Correlation percentage would be .56 (.75 * .75); YET all 3 should be the identical (.25) due to considerations of rotational invariance (and since they are all 120 degrees apart).

2. Some of the hypothetical solutions have asserted that the observer must be added into the system. I think the problem with the "pre-determined" scenario is that the results appear clearly contextual, yet the measurement apparati (essentially the observer or at least the observer's choice of settings) can be changed mid-flight. So this precludes predetermination, I would say.

In my opinion, this causes problems even for non-local theories. Because now we have the apparati (which now becomes the variable) in some kind of mutual communication. How is this possible, when such communication apparently cancels out to zero conveniently for everything else in the universe for this measurement?

In other words, I think it is pretty amazing that even trying to come up with an explicit mechanism which is either non-local or contextual (or both!), it isn't easy to construct a plausible scenario. The reason you have trouble with these mechanisms is that they seem to quickly run afoul of everything else we know about physics! You end up with new theory widgets which ONLY seem to apply to entangled particles.

I can picture another way out of things (this is merely speculation mind you!): relax the requirement that the past cannot be influenced by the future. If you drop that requirement, then it seems natural to me that a particle's history includes its future. There would be communication lines between entangled particles that way, and there would be the opportunity for all histories to potentially interfere with each other (as they appear to do). Not that this makes any more sense than any other explanation, but to me it is one more possibility to consider. So in this scenario, a measurement of Alice retro-influences Bob. This allows for the correct statistics, is fully contextual (and non-realistic) and is also local and time-symmetric. But you pay a strange price for it!

The only other thing that seems to make any consistent sense is simply the mathematical formalism, which is what seems so empty - at some level anyway - to many.

 

[Autochthon]

I can picture another way out of things (this is merely speculation mind you!): relax the requirement that the past cannot be influenced by the future. If you drop that requirement, then it seems natural to me that a particle's history includes its future. There would be communication lines between entangled particles that way, and there would be the opportunity for all histories to potentially interfere with each other (as they appear to do). Not that this makes any more sense than any other explanation, but to me it is one more possibility to consider. So in this scenario, a measurement of Alice retro-influences Bob. This allows for the correct statistics, is fully contextual (and non-realistic) and is also local and time-symmetric. But you pay a strange price for it!
.

I suspect my question arises from thinking along these lines . Basically that there is a family of observations each of which includes all of the possible Alice-Bob correlations and the measured result is the outcome with the perspective that matches the measurement or perhaps; The family of outcomes includes all possible outcomes but we only "see" the one that includes the measurement from one perspective. Time linkages would be part of the perspective. The time component would necessarily limit information transfer to "c" but the interaction between Alice and Bob wouldn't be held to that restriction only Alice, Bob and "the Measurement".

Are you saying something like this when you mean contextual?

 

[Ken G]

The internal (realistic) inconsistency is: IF the AB Correlation percentage is .25 (matching experiment and QM predictions), and the BC Correlation percentage is .25, THEN the AC Correlation percentage would be .56 (.75 * .75); YET all 3 should be the identical (.25) due to considerations of rotational invariance (and since they are all 120 degrees apart).

I suspect the basic source of this apparent inconsistency is the logic that allows us to imagine that reality must account for the answers to questions that are not actually being posed in that reality. The 0.56 calculation requires such a hypothetical aspect-- it is a logical analysis of certain observations that does not take into account in any explicit way whether or not those observations have actually occurred. In other words, we can no longer assume observations are purely hypothetical-- if they are going to be used in describing a certain reality, they have to be included in the reality we are describing.

2. Some of the hypothetical solutions have asserted that the observer must be added into the system. I think the problem with the "pre-determined" scenario is that the results appear clearly contextual, yet the measurement apparati (essentially the observer or at least the observer's choice of settings) can be changed mid-flight. So this precludes predetermination, I would say.

This I believe is the crux of the issue. I'm not concerned with the problem of changing the settings mid-flight, because it doesn't matter what the settings were when no observation occurred, it only matters what the settings were when observations did occur. Only the latter is part of the reality that we are trying to describe, the reality we are wondering whether or not is deterministic. We never asked determinism to explain things that don't happen, only things that do. In other words, we can no longer think of an observation as a hypothetical interaction, we cannot say "because I can set the instrument, it doesn't matter if I actually execute the setting or not", we have to actually ask, "what observation are you actually going to execute, that is the only reality that I need to understand". If you take that approach, then you do not run into any internal inconsistencies.

The question you pose, then, is whether or not this can count as determinism. I believe it can, just a more subtle version where we do not say that any hypothetical result is determinable, but merely that the outcome of experiments that are actually executed have to be determinable. Specifically, that frees reality from needing to be able to logically combine any two observations without explicitly allowing the presence of either observation to alter the reality that is being described. No logical chains of hypothetical realities are quantum mechanically meaningful, that's the bottom line.

In other words, I think it is pretty amazing that even trying to come up with an explicit mechanism which is either non-local or contextual (or both!), it isn't easy to construct a plausible scenario. The reason you have trouble with these mechanisms is that they seem to quickly run afoul of everything else we know about physics! You end up with new theory widgets which ONLY seem to apply to entangled particles.

Yes, this is indeed the rub. Even the idea that we should not need to be able to describe logical chains of hypothetical realities is something of a widget-- we certainly get away with that classically. But maybe it's not a widget after all-- because it's all the information we are throwing away in classical experiments that allows us to get away with hypothetical logic. When, in quantum experiments, we do not throw away information, we are held to a higher standard. In short, it's not that there is something special about quantum mechanics that means we cannot use hypothetical logic, it's that there's something special about classical mechanics that means we can. The widget is what we've been using all this time-- quantum mechanics is just the wakeup call that we've been using a widget we virtually took for granted. We permitted ourselves a type of logic we never really knew could be counted on, and entanglement tells us when we can't count on it.

I can picture another way out of things (this is merely speculation mind you!): relax the requirement that the past cannot be influenced by the future.

But that really sounds like a widget. You might make the same argument I just did, that the widget is the idea that the future does not influence the past, and it is only a widget in classical systems. But the difference is, I argued that entanglement rules something out, whereas you need it to permit something (reverse causality) that is ruled out normally. So that seems a bit more problematical, though I will admit that the arrow of time does seem to have some aspects that stem from classical effects, like huge numbers of virtually indistinguishable states, aspects that don't exist in quantum mechanics. So I don't think your suggestion is out of bounds, even as a widget.

Maybe the future can affect the past in purely quantum systems, but all traces of that back-reaction must be wiped out when we couple them to the classical world. Since entanglement experiments do involve such classical coupling (all experiments do), one must take pains to insure that no trace of your back-reaction survives at the level of the instrument readings. It might only exist when you imagine what the quantum systems are "actually doing", but not what the classical instruments are doing (it would seem important, for example, to insure that classical systems cannot send a message backward in time).

So in this scenario, a measurement of Alice retro-influences Bob. This allows for the correct statistics, is fully contextual (and non-realistic) and is also local and time-symmetric. But you pay a strange price for it!

I agree with all of that except I'm not wild about the idea that the measurements themselves are retro-influenced, because that is at the classical level where there must be an arrow of time, and I'm not sure why you call it local, as it seems the retro-communication is nonlocal. I would say that there are no influences between the Alice and Bob measurements themselves, so no signal propagations at the macro level, but to understand their correlations, we need to postulate retro-communication at the quantum level. Again, the kind of message would not be of the form "a measurement is saying such-and-such", as that is at the macro level, but rather "I am a quantum system about to undergo a certain type of influence from a macro instrument, and I can retro-communicate that fact to my entangled partner, and that communication will alter the context of whatever hidden variables are influencing whatever measurements are being made on my partner". In short, the hidden variables are affected at the quantum level, but there is no actual communication between the macro observables, and hidden-variable-free theories don't even need to recognize that there is any communication at all, retro or otherwise.

 

[DrChinese]

I suspect my question arises from thinking along these lines . Basically that there is a family of observations each of which includes all of the possible Alice-Bob correlations and the measured result is the outcome with the perspective that matches the measurement or perhaps; The family of outcomes includes all possible outcomes but we only "see" the one that includes the measurement from one perspective. Time linkages would be part of the perspective. The time component would necessarily limit information transfer to "c" but the interaction between Alice and Bob wouldn't be held to that restriction only Alice, Bob and "the Measurement".

Are you saying something like this when you mean contextual?

To me, contextual means: the nature of the observation shapes the reality and influences the results. A contextual interpretation (or hypothetical mechanism) does not assert that there are real answers to questions which cannot be measured. So that takes us to a point at which "the moon is not there when we are not looking at it". Which is what Einstein disliked.

 

[DrChinese]

1. I suspect the basic source of this apparent inconsistency is the logic that allows us to imagine that reality must account for the answers to questions that are not actually being posed in that reality. The 0.56 calculation requires such a hypothetical aspect-- it is a logical analysis of certain observations that does not take into account in any explicit way whether or not those observations have actually occurred. In other words, we can no longer assume observations are purely hypothetical-- if they are going to be used in describing a certain reality, they have to be included in the reality we are describing.

2. This I believe is the crux of the issue. I'm not concerned with the problem of changing the settings mid-flight, because it doesn't matter what the settings were when no observation occurred, it only matters what the settings were when observations did occur. Only the latter is part of the reality that we are trying to describe, the reality we are wondering whether or not is deterministic. We never asked determinism to explain things that don't happen, only things that do. In other words, we can no longer think of an observation as a hypothetical interaction, we cannot say "because I can set the instrument, it doesn't matter if I actually execute the setting or not", we have to actually ask, "what observation are you actually going to execute, that is the only reality that I need to understand". If you take that approach, then you do not run into any internal inconsistencies.

The question you pose, then, is whether or not this can count as determinism. I believe it can, just a more subtle version where we do not say that any hypothetical result is determinable, but merely that the outcome of experiments that are actually executed have to be determinable. Specifically, that frees reality from needing to be able to logically combine any two observations without explicitly allowing the presence of either observation to alter the reality that is being described. No logical chains of hypothetical realities are quantum mechanically meaningful, that's the bottom line.

3. Maybe the future can affect the past in purely quantum systems, but all traces of that back-reaction must be wiped out when we couple them to the classical world.

1. I agree with this, I am using the example to show that there is an internal inconsistency in the realistic position.

2. I am good with there being a deterministic solution which is NOT realistic. But here is the rub: when is it determined? We measure Allice and Bob after changing their settings mid-flight. But suppose we measure Alice first? Then Bob's setting doesn't matter in a deterministic non-local solution. So the non-local solution (in this case) needs to consider Alice's setting and ignore Bob's. That's a pretty tall order. And if we are truly asserting it is "deterministic", I guess it should be supplying the actual measurement result as well.

3. Maybe, but also maybe the observed randomness is somehow a function of the future state - in which case all of the traces are not wiped out. Again, this is pure speculation.

 

[Ken G]

I am good with there being a deterministic solution which is NOT realistic. But here is the rub: when is it determined?

Perhaps what is determined in advance, when the system is still in local contact with itself, is not the answer to all hypothetical questions that can be posed simultaneously, it is the answer to all real questions that can actually be physically posed by a possible experimental setup, in its entirety. If we cannot think of a setup that tests for the spin along three axes, then there is no such thing as the spin along three axes, and reality is under no obligation to supply an answer to such hypothetical questions. In other words, it is not necessary for "realism" to assert that all questions have answers, merely that all questions that can actually be physically posed must have answers, whether or not they are actually posed to that particular systems. So although whether or not a question actually gets posed is a matter for trees falling in woods, what questions are possible to pose are a matter of fact.

We measure Allice and Bob after changing their settings mid-flight. But suppose we measure Alice first? Then Bob's setting doesn't matter in a deterministic non-local solution. So the non-local solution (in this case) needs to consider Alice's setting and ignore Bob's.

That depends on what the solution means, i.e., what the question is. If the question is just about what Alice measures by herself, then yes, Bob's setting will never matter, whether Bob's measurement has happened yet or not. If the question includes correlations with Bob's setting, then any changes to Bob's setting are going to be part of the physical posing of that question, which reality must answer and which we have to be able to determine in advance to call it a pre-determined reality. But we can. In other words, it is part of the physical posing of a question about correlations to include both settings, regardless of when they were last changed.

And if we are truly asserting it is "deterministic", I guess it should be supplying the actual measurement result as well.

This is the part we know we can't do, but the question in regard to "determinism" is what is the source of our failure. Is it that reality itself cannot determine that, for fundamental reasons, or is it just our lack of sufficient knowledge about reality that introduces this barrier? I think the idea behind preserving a kind of "possible but not practical" determinism is to say that we simply face overwhelming difficulties in establishing the full information that reality itself uses to determine the outcomes of the experiments.

3. Maybe, but also maybe the observed randomness is somehow a function of the future state - in which case all of the traces are not wiped out.

It is certainly possible that some of the information that we have no access to, but reality does, comes from the future. A tantalizing possibility, to be sure. But I think we already face significant problems in establishing the information that comes from the past (and is stored in untraceable noise modes and convoluted entanglements), such that it will not be possible to separate that from what we don't know could be coming from the future! If the latter is included, however, then we really have no hope of ever improving that situation (a likelihood I tend to accept anyway).