The Impact of Weak Realism on Bell's Theorem and Bohmian Interpretations

[Dmitry67]

So then, in their seminal paper, EPR argued that if the value of a physical quantity can be predicted without measurement or disturbing the system in any way, the quantity corresponds to a real property of the underlying system.

So then, to refute EPR, there is only one option

Wait, wait, the very first assumption was wrong...

 

[Ilja]

In fact, they proved that non-commuting observables CAN NOT have simultaneously precise values and thus can not be simultaneously real. So it is mind boggling why Dr Chinese keeps insisting that the EPR definition of so called "realism" is "simultaneous values of non-commuting observables".

You must have read another EPR paper. This is not what EPR prove. This is what Bell was able to prove with his inequalities.

- Bell's inequality ONLY proved that the mathematical model of probability used by Bell to derive the inequalities, does not apply to the systems modeled by QM
- Aspect type experiments ONLY prove that
1) QM is a much better theory for modelling real systems than Bell's inequalities.
2) the mathematical model of probability used by Bell to derive the inequalities, does not apply to real systems

I disagree. The mathematical model of probability used by Bell is fine and compatible with observation. It has to be combined with locality to give the inequalities.

 

[mn4j]

You must have read another EPR paper. This is not what EPR prove. This is what Bell was able to prove with his inequalities.



I disagree. The mathematical model of probability used by Bell is fine and compatible with observation. It has to be combined with locality to give the inequalities.

I'm afraid you too have misunderstood EPR. EPR never admitted or suggested that it was possible for non-commuting observables to be simultaneously real. The only way you come to that point, is if YOU insist contrary to EPR that QM is complete.

Bell's prove was based on naive hidden variables not non-commuting ones. So if any thing, it only proves the non-validity of a theory based only on non-contextual variables.

Are you claiming that Bell's inequality are more accurate description of reality than experiments? Isn't it obvious that the violation of Bell's inequalities by real experiments implies Bell's model of reality used in deriving the inequalities must be wrong? Or do you consider the experiments unreal?

 

[DrChinese]

First, I can get Bell's result without assuming the simultaneous realism of a, b and c, because I can apply the EPR argument to prove the simultaneous realism of a, b and c, assuming only weak realism (which allows the values of a, b, c to be created in case of measurement and transferred to the other part, thus, does not assume the simultaneous realism of a, b and c) and locality (which closes this possibility).

What is called naive is to postulate the simultaneous realism of a, b and c. This is not what EPR or Bell do. They derive this, based on locality.

This is completely wrong. You might consider referencing materials or providing proofs rather than just making things up (as you have in prior posts as well).

First, EPR is about a and b (i.e. 2 operators, actually p and q), not a, b and c. There is a difference, as Bell needed a, b and c to obtain his result.

Second: EPR is about realism, not locality. They define and mention reality, elements of reality and the like perhaps a dozen times in their paper. Locality is mentioned once in passing. They also assume a particular result in their reasoning regarding the outcome of a hypothetical experiment. The question is whether you can measure, say p, on Alice and then obtain additional information about Alice from a measurement on Bob (which had previously interacted with Alice). However, we now know that their assumed result - more information than QM allows per the HUP - would not be born out. In other words, if Bell's Theorem had never been discovered we would still know that EPR was wrong. That is because entangled particles do not provide more information than the HUP permits.

Lastly, you cannot derive Bell without reference to a, b and c existing simultaneously. Instead of making unsubstantiated claims, do it and show us! Even Norsen has never answered that challenge and believe me, I have made it. All he ever does is say that separability is all you need (Bell's (2)). Yet: how many versions of Bell's Inequality have you seen with just 2 measurement settings? You need 3, which requires that you assume the simultaneous realism of something not observed! If you assume local hidden variables, you cannot construct result sets that are consistent with QM at all variations of 3 settings. But you can with just 2.

Now, if there was not the possibility that the measurement apparati might be able to communicate non-locally, then there would be no need to consider the locality question at all. But as it is, non-locality is a "loophole" if you will. As a result, Bell's Theorem must allow for either a non-realistic theory AND/OR a non-local theory. QED.

 

[SDetection]

I'm not getting it, why the position and momentum of a particle are necessarily non-commuting observables:
I mean for example, a photon's position in a laser beam is always observed/measured all the time by other particles (I meant photons), yet this doesn't alter its momentum.
Am I missing something?. And please be easy on me, I represent the public audience!.

 

[Dmitry67]

Other particles? If you are talking about the air/glass, then yes, light interacts with these particles. But what is important that such interaction does not leave any traces, it is absolutely reversible. So it is not considered a measurement (details may vary in different interpretations)

 

[SDetection]

Other particles? If you are talking about the air/glass, then yes, light interacts with these particles. But what is important that such interaction does not leave any traces, it is absolutely reversible. So it is not considered a measurement (details may vary in different interpretations)

But that means that the uncertainty is because the fault of the measurement process, while Heisenberg clearly stated that this is not the case. He said that it's just how the real world works.

 

[DrChinese]

This really doesn't have anything to do with this thread. This is a general question about the HUP. I really think this should be asked elsewhere...

And yes, there are plenty of times that particles interact with other particles and fields and still act according to the HUP and "as if" they had not been measured/observed. Examples include: light through a lens, off a mirror, changing from one medium to another. In each of these cases, if the photon is an entangled photon, the entanglement can be maintained.

 

[mn4j]

First, EPR is about a and b (i.e. 2 operators, actually p and q), not a, b and c. There is a difference, as Bell needed a, b and c to obtain his result.

Actually EPR is about prediction of 2 different physical quantities ((q1, p2) or (p1, q2) ie 2 values from a set of pairs). Bell's analysis of course missed the mark because he treated pairs of values from a set of triples as though they were being selected from a set of pairs. In other words, he naively thought you could select pairs of values from a set of triplets and they would be equivalent to selecting pairs of values from a set of pairs. That is why it is not possible to derive the inequalities without introducing the error of using 3-stations. (See Hans De Raedt, Karl Hess, Kristel Michielsen (2009) Extended Boole-Bell inequalities applicable to quantum theory http://arxiv.org/abs/0901.2546)

Second: EPR is about realism, not locality.

Actually it is both, because their main conclusion is based on their claim that FTL is unreasonable. I take it you believe FTL is reasonable.

The question is whether you can measure, say p, on Alice and then obtain additional information about Alice from a measurement on Bob (which had previously interacted with Alice).

Actually, that was not exactly their issue. The issue was that according to QM, by choosing to measure p1, you can know q2 exactly but not p2, however for exactly the same system, you can choose to measure q1 and end up knowing p2 exactly but not q2. Therefore according to their definition of reality, q2 will be real in the first experiment, and unreal in the second experiment. However, particle two is so far away that interaction is not possible (read non-local, except FTL), therefore it is unreasonable to expect a theory, which claims to completely model reality, to permit the reality of a physical quantity of a very distant particle (so far away that no iteraction is possible) to be changed instantaneously based on which measurement we choose to perform on particle 1. The key is not about being able to obtain additional information but being able to change the reality of a quantity at particle 2 by deciding which measurement you are going to do on particle 1.

So clearly, both locality and reality are at issue:
1- If you contest their definition of reality, provide an alternative
2- If you accept their definition of reality, and also accept that QM is complete, then you must also admit that FTL is possible.
3- If the QM state vector is not complete then a more complete state vector must exist

However, we now know that their assumed result - more information than QM allows per the HUP - would
not be born out.

EPR was not about HUP. Just because the incomplete description of QM involves p and q with an uncertainty relationship does not mean the more complete description of the underlying system needs or relies on p and q or even other non-commuting observables. EPR did not make any suggestion that p and q must be part of this more complete state vector. The only requirement was that a more complete description existed, not that the specific description must include non-commuting observables.

Now, if there was not the possibility that the measurement apparati might be able to communicate non-locally, then there would be no need to consider the locality question at all. But as it is, non-locality is a "loophole" if you will. As a result, Bell's Theorem must allow for either a non-realistic theory AND/OR a non-local theory. QED.

Seems to me you are admitting that you believe superluminal communication or psychokinesis is reasonable.

 

[DrChinese]

1. Actually, that was not exactly their issue. The issue was that according to QM, by choosing to measure p1, you can know q2 exactly but not p2, however for exactly the same system, you can choose to measure q1 and end up knowing p2 exactly but not q2.

2. Seems to me you are admitting that you believe superluminal communication or psychokinesis is reasonable.

1. QM does NOT say that you can measure p1, and know q2 exactly. It says the exact opposite, that entangled p1 and q2 are non-commuting just like p1 and q1 are. This is supported by experiment.

(By the way, we have previously discussed that the papers of the De Raedts do not follow Bell. They obviously have a local realist agenda, as do you. If you want to debate Bell, this is not the thread. This thread is about Bell and realism.)

2. Per Bell. yes. However, I would stipulate that if superluminal, it is instantaneous.

 

[Ilja]

I'm afraid you too have misunderstood EPR. EPR never admitted or suggested that it was possible for non-commuting observables to be simultaneously real. The only way you come to that point, is if YOU insist contrary to EPR that QM is complete.

Bell's prove was based on naive hidden variables not non-commuting ones. So if any thing, it only proves the non-validity of a theory based only on non-contextual variables.

Are you claiming that Bell's inequality are more accurate description of reality than experiments? Isn't it obvious that the violation of Bell's inequalities by real experiments implies Bell's model of reality used in deriving the inequalities must be wrong? Or do you consider the experiments unreal?

First, I disagree with your reading of EPR. As I already said, you must have read another EPR paper. I even see no starting point to argue.

It appears you have also read a different Bell paper, because in the Bell paper I have read there was nothing naive in the assumptions he has used. Non-contextuality is certainly not among the assumptions used by Bell.

The important assumption used by Bell is Einstein causality.

Last but not least, I fully agree with the experiments, as well as I fully agree with the notion of realism used in Bell's theorem. There is no contradiction, because Bell has used in his theorem also another assumption, namely a realistic version of Einstein causality. Thus, the experiments simply prove that this version of Einstein causality is wrong.

 

[Ilja]

This is completely wrong. You might consider referencing materials or providing proofs rather than just making things up (as you have in prior posts as well).

Just to clarify: This is not a scientific paper, but a simple posting in a discussion forum. To write a paper about it I see no reason because there would be nothing new in it - all this is well-known.

First, EPR is about a and b (i.e. 2 operators, actually p and q), not a, b and c. There is a difference, as Bell needed a, b and c to obtain his result.

That's nothing to object. I know that only Bohm improved the thought experiment by replacing p and q by spins.

Second: EPR is about realism, not locality. They define and mention reality, elements of reality and the like perhaps a dozen times in their paper. Locality is mentioned once in passing.

How often they mention realism in comparison with causality is completely irrelevant. Their argument requires locality. As well, it requires a weak form of realism (the EPR criterion of reality). And, combining these two, they obtain something which Norsen calls naive realism about the relevant observables.

They also assume a particular result in their reasoning regarding the outcome of a hypothetical experiment. The question is whether you can measure, say p, on Alice and then obtain additional information about Alice from a measurement on Bob (which had previously interacted with Alice). However, we now know that their assumed result - more information than QM allows per the HUP - would not be born out. In other words, if Bell's Theorem had never been discovered we would still know that EPR was wrong. That is because entangled particles do not provide more information than the HUP permits.

Their argument remains valid. From the EPR criterion of reality, Einstein causality, and the observable correlations it follows that the observables considered have predefined values. We know that they cannot have them, thus, one of the assumptions used in this argument should be wrong. The only reasonable candidate is Einstein causality.

Lastly, you cannot derive Bell without reference to a, b and c existing simultaneously. Instead of making unsubstantiated claims, do it and show us! Even Norsen has never answered that challenge and believe me, I have made it.

I observe that, whenever the same direction (a,b,resp,c) is measured at A and B, the results agree. I apply the EPR criterion of reality and Einstein causality. This gives that
local values for a,b, and c have to be elements of reality.

Thus, I don't have to assume them, once I can derive this from other assumptions.

All he ever does is say that separability is all you need (Bell's (2)). Yet: how many versions of Bell's Inequality have you seen with just 2 measurement settings? You need 3, which requires that you assume the simultaneous realism of something not observed!

Of course, as I acknowledge without any hesitation, I assume some weak form of realism. This form of realism is, of course, about what really exists, not about what is observed.

And I do not think that Norsen thinks he can do the proof without some assumption of realism. He discusses favourously two notions of realism - perceptual realism and metaphysical realism - more or less as prerequisites for doing science. Rejecting them you cannot do science at all. Thus, he has any right to assume them.

If you assume local hidden variables, you cannot construct result sets that are consistent with QM at all variations of 3 settings. But you can with just 2.
Now, if there was not the possibility that the measurement apparati might be able to communicate non-locally, then there would be no need to consider the locality question at all. But as it is, non-locality is a "loophole" if you will. As a result, Bell's Theorem must allow for either a non-realistic theory AND/OR a non-local theory. QED.

Your point being? I clearly favour the non-local version - pilot wave theory. There is absolutely no reason to give up the weak, metaphysical version of realism on needs in this proof. Therefore one can as well omit mentioning "realism", as one does omit mentioning "logic" and "mathematics" and "probability theory".

 

[Ilja]

1. QM does NOT say that you can measure p1, and know q2 exactly. It says the exact opposite, that entangled p1 and q2 are non-commuting just like p1 and q1 are. This is supported by experiment.

False. For standard QM, observables of different particles commute. Thus, p1 commutes with q2 and p2 commutes with q1.

[p_i, q^j] = -i hbar delta^j_i

 

[DrChinese]

False. For standard QM, observables of different particles commute. Thus, p1 commutes with q2 and p2 commutes with q1.

Not for entangled particles. Otherwise: you could measure p1 and q2, and q2 would tell you what q1 was to any desired precision. You would then know both p1 and q1. And we know that is wrong. You cannot learn more about the entangled system than the HUP permits. Now, please, don't confuse that with being able to measure p1 and q2 as that point is not in question. The question is whether p1 and q2 are non-commuting where particles 1 and 2 are entangled in p and q.

 

[p764rds]

Not for entangled particles. Otherwise: you could measure p1 and q2, and q2 would tell you what q1 was to any desired precision. You would then know both p1 and q1. And we know that is wrong. You cannot learn more about the entangled system than the HUP permits. Now, please, don't confuse that with being able to measure p1 and q2 as that point is not in question. The question is whether p1 and q2 are non-commuting where particles 1 and 2 are entangled in p and q.

Why would entanglement affect the non-commutability?

 

[mn4j]

From my reading of this thread, it is obvious that DrChinese clearly understands reality as defined by EPR to be what Norsen calls "naive realism". So my understanding is that both Ilja and DrChinese agree on this point that realism ala EPR is "naive realism".

Just to be clear, using the photon polarization example at three different angles (A, B and C), "naive realism" would mean the photon has simultaneously three different internal physical quantities corresponding to polarization at those angles, and that those quantities belong to the photon. So the claim then is that this is what EPR meant by realism.

If this is your claim, then both of you could not be further from the truth for the following reason. Imagine that rather than having three different internal physical quantities the photon only had one quantity that could be said to belong to the photon. Now consider that on interacting with the polarizer, the interaction reveals a result based on the angle of the polarizer or other device settings such that the observed polarizations are just three different manifestations of the same underlying quantity. In other words, it is not so much a property of the photon but of the combined "photon+polarizer" system. Clearly this scenario is NOT "naive realism" as defined by Norsen. However, this scenario is clearly real as defined by EPR. This is why I keep insisting that both of you have misunderstood EPR.

The EPR definition of reality did not mandate that the reality of one quantity be independent from the reality of another quantity. If you disagree, point to the sentence in the paper where they made the requirement. Within their definition, it is perfectly acceptable to have two quantities describing the same underlying reality so long as the two descriptions are not contradictory. Non-commuting observables get in the picture because they can not be simultaneously real and thus lead to contradictory descriptions (ie, being able to instantaneously change the reality of a remote quantity by deciding which measurement to perform locally).

Therefore it is not true that EPR realism is "naive realism".

 

[mn4j]

Yet: how many versions of Bell's Inequality have you seen with just 2 measurement settings? You need 3, which requires that you assume the simultaneous realism of something not observed!

This is why I explained that you can not expect correlations of pairs of values from a system consisting of 2 photons and three settings be equivalent to those from a system consisting of 2 photons and 2 measurement settings. Its apples and oranges! The only way you can be consistent to the EPR definition is if you consider the "system" as composed of particles PLUS measuring device, and the quantities as being properties of this "system". But then you run into the delimma that it is an unrealizable and thus unverifiable experiment. It is impossible to measure two photons using three devices. No experiment has ever been or ever will be performed using three stations to measure 2 particles. How then can anyone ever say with a straight face that Bell's inequalities have been verified experimentally?

The only thing that has happened is that the QM correlations have been confirmed by experiment, without saying anything about the validity of Bell.

 

[DrChinese]

Why would entanglement affect the non-commutability?

Because non-entangled particles have separate wave functions, and you could obtain a good p1 and q2 in that case.

 

[DrChinese]

1. From my reading of this thread, it is obvious that DrChinese clearly understands reality as defined by EPR to be what Norsen calls "naive realism". So my understanding is that both Ilja and DrChinese agree on this point that realism ala EPR is "naive realism".

Just to be clear, using the photon polarization example at three different angles (A, B and C), "naive realism" would mean the photon has simultaneously three different internal physical quantities corresponding to polarization at those angles, and that those quantities belong to the photon. So the claim then is that this is what EPR meant by realism.

Pretty much, this is the standard hidden variable (a/k/a realistic) interpretation. The requirement stems from the fact that we have "perfect" correlations at matching angle settings for entangled pairs. I separated your definition from the rest of your post so we could see that I do agree with this idea as being the basis for realism, either naive or otherwise.

I do not agree with the portion that requires there to be underlying attributes for every possible angle setting. I.e. if you divided a circle into 360 degrees and made each one an atrribute, there would be 360 attributes to describe spin. I do not believe realism asserts that there are 360 (or 360 million, or infinity, etc.) such attributes. Realism is merely the expression of the idea that there is a definite real outcome to a measurement of a specific observation, regardless of whether or not such observation is made. The mapping to an underlying physical attribute, operator, etc. is optional.

 

[DrChinese]

If this is your claim, then both of you could not be further from the truth for the following reason. Imagine that rather than having three different internal physical quantities the photon only had one quantity that could be said to belong to the photon. Now consider that on interacting with the polarizer, the interaction reveals a result based on the angle of the polarizer or other device settings such that the observed polarizations are just three different manifestations of the same underlying quantity. In other words, it is not so much a property of the photon but of the combined "photon+polarizer" system. Clearly this scenario is NOT "naive realism" as defined by Norsen. However, this scenario is clearly real as defined by EPR. This is why I keep insisting that both of you have misunderstood EPR.

The EPR definition of reality did not mandate that the reality of one quantity be independent from the reality of another quantity...

It really doesn't matter to Bell whether or not this version of realism is somehow different than the previously agreed version (per above post). And I don't think it matters to EPR either. But let's review the facts before we go much further. What DO we need to have? It is not 1 attribute for photon spin, it is at least 2!

The reason is that across 360 degrees, there are completely independent results at 0 and 45 degrees (and all multiple thereof) that have no value overlaps. But the strange thing is, it doesn't matter how you orient the 0 degrees direction, there are still at least 2. There could be more, but maybe not. So I certainly do not assert that realism requires there be some large number of independent attributes - I only think there are 2 to begin with (using spin of linear polarized photons as the example). But we need to allow for the facts, which definitely point to at least 2.

I believe the QM description is essentially correct or complete or whatever you want to call it. So, it would not make sense for me to think the debate about realism - which might be needed in a theory or might not - would not fit my notion of what there is to be "real" in the first place. After Bell, the question is whether a correct QM is non-local or non-realistic (or possibly both).

Now, if you are a realist, and believe there is a reasonable alternative and physical explanatory mechanism that fits the facts: please come forward with a potential model. You will see quickly that you must still postulate real definite results for n>2 measurement settings. Clearly, you could postulate n<=2 and that would match the QM view of internal degrees of freedom. (Of course, it doesn't explain the HUP but that is a separate issue.) That is what separated Bell from everything else.

Bell gave us the leap to n=3 and the realization that consistent resultsets could not be constructed by a (local) realist. There is not internal consistency because the results at settings between 0 and 45 degrees are a mixture of 0 and 45 degrees, but the mix is not a linear combination as it would need to be fit the realistic view. Of course, this description is for entangled linear polarized photons, but the analogy is the same for electrons (used in Bell). If you would like me to further explain this logic, I can provide specific examples.

 

[mn4j]

Pretty much, this is the standard hidden variable (a/k/a realistic) interpretation. The requirement stems from the fact that we have "perfect" correlations at matching angle settings for entangled pairs. I separated your definition from the rest of your post so we could see that I do agree with this idea as being the basis for realism, either naive or otherwise.

If this is the definition of realism at issue in Bell, then it is obvious why Bell and his proponents have missed the mark. Using my previous analogy of the sun and it's "color", you would be saying realism means the sun has a color when nobody is looking. You can see this by having two people wear different colored goggles. If you believe the realism means the sun has a color when nobody is looking, then you must also admit that the realism of the sun can be changed instantaneously by changing the color of goggles you put on (ie FTL). It is obvious to see how such a person would conclude that local hidden variable theorems are forbidden -- because in fact no local hidden variable theorem can explain being able to change the color of the sun superluminally. But as I hope is getting obvious already, the error was made by attributing to the sun only, properties that belong to the combined "sun + observer system". Correctly recognizing that the appropriate system must include the observer, you then realize that it becomes possible to use a hidden variable theorem to explain the results. The local hidden variable theorem in this case being along the lines that the sun produces a spectrum of electromagnetic radiation which when combined with the goggles results in absorption of some regions and not others resulting in different perceived colors.

I do not agree with the portion that requires there to be underlying attributes for every possible angle setting. I.e. if you divided a circle into 360 degrees and made each one an atrribute, there would be 360 attributes to describe spin. I do not believe realism asserts that there are 360 (or 360 million, or infinity, etc.) such attributes. Realism is merely the expression of the idea that there is a definite real outcome to a measurement of a specific observation, regardless of whether or not such observation is made. The mapping to an underlying physical attribute, operator, etc. is optional.

It may seem reasonable until you realize that the concept of "polarization" includes the measuring device in this particular case. You can't have it both ways. So then it is inconsistent to talk of the photon having a polarization even when no measurement has been made.

Note that to say the sun does not have a color when nobody is looking is not the same thing as saying the moon does not exist when nobody is looking. We are merely saying "color" by definition includes the object, observer and medium in the system.
Therefore
1) it doesnot even make sense to talk of "measuring" the color of the sun, because the sun does not have a color
2) it does not make sense to talk of the "sun+goggles+observer" system having a color when no measurement has been made because, the "color" of the "sun+goggles+observer" system is by definition, the result of a measurement.

Therefore I would modify your definition of realism to: Realism is merely the expression of the idea that there are objective facts about a system, whether or not observers know them (or can know them) and Local causality is therefore merely the expression of the idea that FTL or psychokinesis is unreasonable. This is the only definition of "realist local causality" that is exhaustive and consistent with EPR.

Naive realism is limited to the concept that all observations are passive revelations of properties owned by the objects being observed. But as we have seen, even if the objects own objective properties, their revelation need not be passive, in which case the revealed outcomes will be properties of the combined object/observer system. So naive realism is unreasonable. Any attempt to limit the EPR definition of reality to naive realism for the purpose of arguing against it is a straw-man argument.

In order to argue against all local realist theories therefore, you MUST prove that either:
1) No objective facts can exist about a system OR
2) argue why FTL or psychokinesis is reasonable.

Arguing against (1) is self defeating because the validity of any argument against (1) depends on the validity of (1). In other words, in order for the statement "No objective facts can exist about a system" to be valid, it must be an objective fact about all systems. So then we are left with (2). Has (2) been proven?

 

[mn4j]

It really doesn't matter to Bell whether or not this version of realism is somehow different than the previously agreed version (per above post). And I don't think it matters to EPR either. But let's review the facts before we go much further. What DO we need to have? It is not 1 attribute for photon spin, it is at least 2!

You contradict yourself, Why MUST you have at least 2. Clearly that means Bell's treatment does not apply to situations in which you have only one. Why can it not be just one. You also have admitted that Bell's inequalities can not be derived from less than 3. So it seems you are saying there must be at least 3 not 2.

The reason is that across 360 degrees, there are completely independent results at 0 and 45 degrees (and all multiple thereof) that have no value overlaps. But the strange thing is, it doesn't matter how you orient the 0 degrees direction, there are still at least 2. There could be more, but maybe not. So I certainly do not assert that realism requires there be some large number of independent attributes - I only think there are 2 to begin with (using spin of linear polarized photons as the example). But we need to allow for the facts, which definitely point to at least 2.

The issue is not merely the number of attributes but the mechanism of revelation of the outcome. Passive or not.

I believe the QM description is essentially correct or complete or whatever you want to call it.

I believe the QM description is correct but incomplete, and so did EPR. The correctness of QM has never been in question.

After Bell, the question is whether a correct QM is non-local or non-realistic (or possibly both).

QM can not be correct and non-realistic because QM is a theory about real systems. As I have shown in my previous post, the argument against realism is self-defeating. It is like saying "reality" is not realistic. I would say the real issue is what the nature of "reality" is. In otherwords, is reality local or non-local, or in yet other words, is FTL reasonable or not. You believe it is reasonable so maybe you can convince me that it is.

Now, if you are a realist, and believe there is a reasonable alternative and physical explanatory mechanism that fits the facts: please come forward with a potential model.

I don't need to have such a model in order to appreciate that FTL is unreasonable because to believe FTL will require that I simultaneously hold contradictory beliefs which would be intellectually dishonest.

In any case, an alternative explanation has been provided. See

Event-by-Event Simulation of Einstein-Podolsky-Rosen-Bohm Experiments
Shuang Zhao · Hans De Raedt · Kristel Michielsen
Found Phys (2008) 38: 322–347​

Read it and we can discuss it in a separate thread.
All that this boils down to is the question: Is a deterministic local hidden variable explanation for Malus Law possible? This question has been answered in the affirmative.

 

[DrChinese]

1. If this is the definition of realism at issue in Bell, then it is obvious why Bell and his proponents have missed the mark. ... But as I hope is getting obvious already, the error was made by attributing to the sun only, properties that belong to the combined "sun + observer system".

2. It may seem reasonable until you realize that the concept of "polarization" includes the measuring device in this particular case. You can't have it both ways. So then it is inconsistent to talk of the photon having a polarization even when no measurement has been made.

3. Naive realism is limited to the concept that all observations are passive revelations of properties owned by the objects being observed. But as we have seen, even if the objects own objective properties, their revelation need not be passive, in which case the revealed outcomes will be properties of the combined object/observer system. ...

You are arguing pro or con realism here, yet the question is whether realism is an assumption of EPR/Bell. And NOT whether it is valid or not.

1. The issue is whether the observer determines reality. In this case, there are 2 observers, Alice and Bob. Does Alice determine Bob's reality? EPR argues NO.

As to whether the choice of pink or blue goggles determines anything: no one is arguing against your point. But a local realist requires the result for Bob to be independent of the choice of goggles made by Alice. Further, what if Alice instead chooses yellow goggles? Bell showed that the results were incompatible with your idea.

2. Of course, once Alice has made a measurement, Bob does have a definite polarization.

3. Essentially, it is passive but only in the sense that the interaction of the particle and the apparatus reveals a value that, due to whatever rules were in place, would always be revealed by a measurement of that type (per EPR). Now, if there were elements being introduced from the observer, and that is your idea, then it is incumbent that the observer's choice be communicated from Alice to Bob (per Bell).

 

[DrChinese]

1. You contradict yourself, Why MUST you have at least 2. Clearly that means Bell's treatment does not apply to situations in which you have only one. Why can it not be just one. You also have admitted that Bell's inequalities can not be derived from less than 3. So it seems you are saying there must be at least 3 not 2.

2. I believe the QM description is correct but incomplete, and so did EPR. The correctness of QM has never been in question.

QM can not be correct and non-realistic because QM is a theory about real systems. As I have shown in my previous post, the argument against realism is self-defeating. It is like saying "reality" is not realistic. I would say the real issue is what the nature of "reality" is. In otherwords, is reality local or non-local, or in yet other words, is FTL reasonable or not. You believe it is reasonable so maybe you can convince me that it is.

3. In any case, an alternative explanation has been provided. See

Event-by-Event Simulation of Einstein-Podolsky-Rosen-Bohm Experiments
Shuang Zhao · Hans De Raedt · Kristel Michielsen
Found Phys (2008) 38: 322–347​

Read it and we can discuss it in a separate thread.
All that this boils down to is the question: Is a deterministic local hidden variable explanation for Malus Law possible? This question has been answered in the affirmative.

1. Why must you have at least 2? Because you can do an experiment on 2 (although the experiment does not prove conclusively that local realism cannot be correct). Why can't you have 3, as local realism requires? Because Bell shows that 3 are incompatible with the predictions of QM.

2. EPR did vary from QM. They were wrong in their assumption that p1 and q2 could be determined, violating the HUP. We now know that for entangled particles, you cannot learn anything additional about Alice by measuring Bob.

3. I won't discuss local realistic papers as they are a complete waste of my time. Of course, you cannot have a local realistic theory that postulates Malus and gets past Bell. So go ahead and start the thread, I just won't be there.

 

[DrChinese]

In the words of Zeilinger (1999), on EPR and Bell:

In 1935 Einstein, Podolsky, and Rosen (EPR) studied entangled states of the general type used in the twophoton experiment discussed above. They realized that in many such states, when measuring either linear momentum or position of one of the two particles, one can infer precisely either momentum or position of the other. As the two particles might be widely separated, it is natural to assume validity of the locality condition suggested by EPR: ‘‘Since at the time of measurement the two systems no longer interact, no real change can take place in the second system in consequence of anything that may be done to the first system.’’ Then, whether or not momentum or position can be assigned to particle (system) 2 must be independent of what measurement is performed on particle 1 or even whether any measurement is performed on it at all. The question therefore arises whether the specific results obtained for either particle can be understood without reference to which measurement is actually performed on the other particle. Such a picture would imply a theory, underlying quantum physics, which provides a more detailed account of individual measurements.

... A most important development was due to John Bell (1964) who continued the EPR line of reasoning and demonstrated that a contradiction arises between the EPR assumptions and quantum physics. The most essential assumptions are realism and locality. This contradiction is called Bell’s theorem.


As can be seen from the above (and there are no shortage of similar comments from others), the general view matches what I am saying: Norsen is ignoring the arguments that have been made, and addressed experimentally, with his presumed objective being to recast the local realistic debate in a different form than was made originally (i.e. to make it about locality alone). OK, fine, but we are missing the support for this point of view! Clearly, there is substantial support for the traditional view.

If Norsen - or any reader - thinks realism is superfluous to the debate about Bell, there is a simple solution: write it up without realism and see if it is accepted by your peers. I do not see that as possible, but maybe someone smarter that me could do it.

On the other hand: there are quite a number of arguments to the effect that realism is not only central to Bell, but that locality itself is not an issue. Examples: Kochen Specker, GHZ, etc. These demonstrate that realism must be rejected in ANY physical theory. I am not personally arguing that position, but there is certainly strong support for it. Again, from Zeilinger:

...It was initially argued by Specker (1960) for Hibbert spaces of dimension larger than two that quantum mechanics cannot be supplemented by additional variables. Later it was shown by Kochen and Specker (1967) and by Bell (1966; for a review see Mermin, 1993), that for the specific case of a spin-1 particle, it is not possible to assign in a consistent way measurement values to the squares of any three orthogonal spin projections, despite the fact that the three measurements commute with each other. This is a purely geometric argument which only makes use of some very basic geometric considerations. The conclusion here is very important. The quantum system cannot be assigned properties independent of the context of the complete experimental arrangement. This is just in the spirit of Bohr’s interpretation. This so-called contextuality of quantum physics is another central and subtle feature of quantum mechanics.

(Contextuality being essentially equivalent to the non-realistic position.)

 

[p764rds]

Non-commuting observables get in the picture because they can not be simultaneously real and thus lead to contradictory descriptions (ie, being able to instantaneously change the reality of a remote quantity by deciding which measurement to perform locally).

When a measurement is performed locally -for two entangled particles - nothing changes remotely apart from the local and remote states that are no longer correlated.
So, in my view there is no physical-like communication needed between the two particles. Its just that what was once correlated is no longer correlated (after local observation). Its a 'negative' logic-wise which is different from a positive logic-wise.

I can't see how a Bohmian pilot wave comes in here - its not needed - and its different from a particle arriving at its destination and reporting back (through time) to its initial position (which I see might possibly be the case) in the normal preparation and observation of one particle. Entanglement correlations is different and IMO requires that both particles refer to the same data point in information space wherever they are in spacetime - after local observation (or remote) that data point is no longer flagged as belonging to the particles.

 

[Ilja]

Not for entangled particles. Otherwise: you could measure p1 and q2, and q2 would tell you what q1 was to any desired precision. You would then know both p1 and q1. And we know that is wrong. You cannot learn more about the entangled system than the HUP permits. Now, please, don't confuse that with being able to measure p1 and q2 as that point is not in question. The question is whether p1 and q2 are non-commuting where particles 1 and 2 are entangled in p and q.

Yep. And if the operators p1 and q2 commute or not depends only on the operators, but in no way on the particular state (however entangled or not). And the operators q2 and -i\hbar\partial_1 commute.

What we, by measuring q2, obtain a previously existing value q1 is a consequence of the EPR argument, and needs locality. Quantum theory is nonlocal, therefore this is not essential.

 

[Ilja]

From my reading of this thread, it is obvious that DrChinese clearly understands reality as defined by EPR to be what Norsen calls "naive realism". So my understanding is that both Ilja and DrChinese agree on this point that realism ala EPR is "naive realism".

There are two different meanings of "realism a la EPR": What EPR use as their assumptions (the EPR criterion of reality), and what EPR obtain as a consequence of this assumption, locality, and the QM correlations. It is only the last which, in case one starts with it, could be (and has been by Norsen) named "naive realism".

Just to be clear, using the photon polarization example at three different angles (A, B and C), "naive realism" would mean the photon has simultaneously three different internal physical quantities corresponding to polarization at those angles, and that those quantities belong to the photon. So the claim then is that this is what EPR meant by realism.

No, this is only what follows, after the EPR argument (including locality) from the EPR criterion, which does not make such an assumption.

If this is your claim, then both of you could not be further from the truth for the following reason.

It has never been my claim.

 

[Ilja]

In the words of Zeilinger (1999), on EPR and Bell:

... A most important development was due to John Bell (1964) who continued the EPR line of reasoning and demonstrated that a contradiction arises between the EPR assumptions and quantum physics. The most essential assumptions are realism and locality. This contradiction is called Bell’s theorem.


As can be seen from the above (and there are no shortage of similar comments from others), the general view matches what I am saying: Norsen is ignoring the arguments that have been made, and addressed experimentally, with his presumed objective being to recast the local realistic debate in a different form than was made originally (i.e. to make it about locality alone). OK, fine, but we are missing the support for this point of view! Clearly, there is substantial support for the traditional view.

First, referencing Zeilinger as an authority is not very impressive. Second, Norsen quotes AFAIR himself a lot of such references to locality and realism as the assumptions, thus, you prove nothing against him with this quote. He has provided arguments against this mainstream position. If ignoring his arguments and quoting the questioned mainstream position as the mainstream position is the new way of solving scientific problems, we have
finally reached the state of democratic science, where majority decides about truth without
hearing any arguments. That's the end of science as science.

If Norsen - or any reader - thinks realism is superfluous to the debate about Bell, there is a simple solution: write it up without realism and see if it is accepted by your peers. I do not see that as possible, but maybe someone smarter that me could do it.

How often I have to repeat that realism is not superfluous, but has to be precisely defined. And once this is done, it appears that the used version of realism is so weak that to reject it becomes nonsensical, and to mention it as a nontrivial assumption unnecessary.

On the other hand: there are quite a number of arguments to the effect that realism is not only central to Bell, but that locality itself is not an issue. Examples: Kochen Specker, GHZ, etc. These demonstrate that realism must be rejected in ANY physical theory.

Not at all. These demonstrate that the rather nonsensical restriction of non-contextuality is incompatible with quantum predictions. This has nothing to do with realism.

(Contextuality being essentially equivalent to the non-realistic position.)

LOL. Contextuality has something to do with naive realism, but not with the metaphysical notion of realism Norsen and I defend. In particular, pilot wave theory is a quite nice, realistic theory, but contextual. Thus, in your words pilot wave defenders would be antirealists. Which is complete nonsense.

 

[DrChinese]

1. First, referencing Zeilinger as an authority is not very impressive. Second, Norsen quotes AFAIR himself a lot of such references to locality and realism as the assumptions, thus, you prove nothing against him with this quote. He has provided arguments against this mainstream position. If ignoring his arguments and quoting the questioned mainstream position as the mainstream position is the new way of solving scientific problems, we have
finally reached the state of democratic science, where majority decides about truth without
hearing any arguments. That's the end of science as science.

2. LOL. Contextuality has something to do with naive realism, but not with the metaphysical notion of realism Norsen and I defend. In particular, pilot wave theory is a quite nice, realistic theory, but contextual. Thus, in your words pilot wave defenders would be antirealists. Which is complete nonsense.

1. Don't get this, you are essentially saying Norsen should be considered more of an authority than Zeilinger? Zeilinger is one of the pre-eminient authorities in the area, and entanglement is his specialty. Certainly this is relevant. Norsen is primarily a theorist as best I know, although I thought he was doing some experimental work more recently. I would certainly like to see something from him that delved into some new areas more at the forefront of the field, as Zeilinger is doing. Just last week, Zeilinger was co-author on a Bell experiment with entanglement on 2 degrees of freedom (finding a Bell Inequality violated by 24 SD).

Now anyone can be wrong, and anyone can be right, regardless of their prior background or reputation. But what we are discussing is not the outcome of an experiment, but rather the relevance of that outcome. So certainly, we *should* consider the opinion of the scientific community in this case.

Also: Travis frequently interprets sections of standard texts and comes up with his own views on their meaning and context - against standard opinion. OK, sometimes that can be good too. But he often ignores any text or argument which goes against his opinion. Example: Norsen says EPR is about locality and not realism. Yet EPR references locality only ONCE, versus over a dozen times for realism. He never addresses this obvious flaw in his reasoning. In addition, EPR provides the standard definition of realism ("elements of reality"). Yet Norsen seeks his own definition, changing the character of the paper in the process. It has been my hope to learn if there is any "meat" to Travis' argument that I am missing in his work. Thus, I was hoping you could enlighten me on this in some fashion. If the best anyone can do is quote Norsen over EPR, quote Norsen over Bell, and quote Norsen over Zeilinger, given the huge gaps I have just mentioned... well, that fails on every level.

2. I'm laughing... There are plenty who believe that a Bohmian interpretation is deterministic and therefore realistic. And additionally ruled out by contextual no-go proofs. I don't happen to fall in that camp. But clearly, there is something odd about asserting that a theory has hidden variables, is deterministic, and yet is also contextual. Essentially, that viewpoint agrees with the EPR conclusion that a more complete specification of the system IS possible. And I reject that.