# Against realism

Against "realism"

If you would like to discuss the pros and cons of Norsen's paper (and I would be happy to do so), we should start another thread for that.

Ok, no problem.

My point about Norsen's paper is that IF you accepted his basic argument - as you say you do - THEN you naturally wind up with a favorable perspective on Bohmian interpretations. That is not the only way to get to that point, but I think it is natural. Because then, you essentially view Bell's Theorem as a rejection of locality - as Norsen essentially does, since he thinks that "naive realism" (his words) is not so relevant to Bell's Theorem.

That's not a problem, once I indeed favour a Bohmian interpretation.

(By the way, I laughed when Norsen attached the word "naive" to a respectable viewpoint. Apparently, no one bothered to explain to Einstein that his cherished viewpoint was in fact naive.

What he describes as naive (out of memory, once we start a new thread about it, I would have to reread it) is naive if taken as an axiom. EPR argue that this special situation follows from locality. Such an argumentation is not naive at all.

Anyway, realism - robust, healthy and meaningful - is fully present in Bell's Theorem as an assumption and there is nothing weak about it. It is difficult to imagine how Bell's famous paper can be considered absent the arguments presented in the EPR paper - in which the famous phrase "elements of reality" is defined - given the title Bell chose. I think Bell could have labeled his realism assumption more clearly, but that does not change the argument or its impact.)

It is clearly important to recognize that it is the first part of Bell's theorem (and, essentially, the EPR argumentation) that from weak realism and locality follows what is appropriately named "naive realism".

Many presentations of the violation of BI fail to recognize this, and from point of view of these presentations it looks like only the "naive realism" (which is, indeed, quite unreasonable to postulate even in classical situations), fails.

To give up weak (meatphysical) realism is quite another category. It means to give up the search for realistic explanation with only one reason: One does not like the fact that all realistic explanations require a preferred frame.

Then, giving it up does not preserve the realistic versions of Einstein causality or relativistic symmetry: Without reality, they are about nothing, thus, become meaningless anyway.
Thus, they have to be given up anyway. No reason to give up metaphysical realism.

Without reality, they are about nothing, thus, become meaningless anyway.

The best approach to realism is MUH
But I dont know how to call it.
I like MUH, so I would rather ask you - MUH is it an ultimate realism or an ultimate denial of the one?

DrChinese
Gold Member

What he describes as naive (out of memory, once we start a new thread about it, I would have to reread it) is naive if taken as an axiom. EPR argue that this special situation follows from locality. Such an argumentation is not naive at all.

Norsen had an interesting (read: wrong) take on EPR. Locality is assumed in EPR, not really a factor per se. It is simply used as a technique to have particles interact and then separate. So the separability requirement is an experimental requirement so that we can go deeper into the central question of EPR: is there a "more complete" level of reality of a single particle.

1. The abstract of EPR reads: "In a complete theory, there is an element corresponding to each element of reality... In quantum mechanics in the case of tow physical quantities described by non-commuting operators, the knowledge of one precludes the knowledge of the other... One is led to conclude that the description of reality as given by a wave function is not complete." The paper concludes (before a final paragraph): "No reasonable definition of reality could be expected to permit this."

2. Nowhere does EPR discuss FTL influences, which they take to be sacrosanct. Of course, EPR's conclusion was wrong in the sense that they expected a "reasonable" definition of reality, one which is not observer independent. As Bell showed, that was not possible. We might get a clue as to what the paper was about from its title, "On the Einstein Podolsky Rosen Paradox." Clearly, not about locality. These papers are about the nature of reality. Is reality observer dependent, as QM seems to imply? Or is the moon round even when no one is looking, as Einstein believed?

In 1935, there was no question in the minds of EPR that their thought experiment - if performed - would demonstrate that the HUP could be beaten. On the other hand, most others (Bohr etc.) felt the opposite. But everyone could see one case which would not answer the question either way: the case of the so-called perfect correlations. This is essentially the kind of example EPR uses, where total momentum is known before 2 particles interact. EPR asked if you could learn more about a SINGLE particle (in this case) than the HUP allows. So that is the ultimate question: does an independent particle have well-defined values for non-commuting observables? If YES, well, that is realism, my friend. You can call it naive if you like but that is CERTAINLY what Einstein believed in.

3. So in the case of the perfect correlation, to switch to the entangled photon setup which was not experimentally feasible in 1935: A measurement on Alice allows you to predict, with perfect confidence, the result of an identical measurement on Bob. Now clearly, this was predicted by both QM and the local realist school led by Einstein. And Bell discussed this in his paper as well, showing in his (4) - (8) some examples in which these results could be derived by simple models (which themselves could be called naive or toy if you need to label them). Clearly, when Alice and Bob are measured at identical settings, the results should match perfectly. This is a requirement of any candidate mechanical theory, local or non-local. (In a way, it is simply a consequence of the greater requirement that any candidate theory reproduce the QM predictions.)

4. Bell introduces the requirement of realism after his (14), when he hypothesizes: "it follows that c is another unit vector". This is the leap. Previously, we had 2 particles and 2 measurement setting possibilities. Now he adds the "realistic" requirement that there must also be values at other measurement settings.

It is pretty obvious that if you don't have an a, b and c, then you can't get to Bell's inequality (15). Realism is NOT inserted as a requirement because it was previously assumed, as Norsen claims. Bell adds it here because it wasn't explicit previously, and needed to be added to obtain Bell's key result.

5. Now, you are free to disagree with my analysis, as Norsen does. But you will be hard pressed to make Norsen's argument by direct reference to EPR and Bell (not an absolute requirement, but you will see quickly that Norsen's reasoning depends on some very shaky inference). Norsen is completely wrong when he challenges:

"...anyone who claims that Bell’s Theorem is a theorem about 'local realist' theories ... needs to explain clearly what they mean by 'realism' [done, see 2. above] and show precisely where such 'realism' is assumed in the derivation of Bell’s inequalities. [done, see 4. above]"

I have refuted his argument by reference to the key articles themselves. Norsen's historical perspective is extremely skewed, and this was pointed out by Shimony himself in another one of Norsen's historical outings. I know Norsen is bright and knowledgeable, but that does not change my opinion of the interpretation of Bell's result (nor my conclusion that Norsen is wrong). And apparently, it hasn't swayed the opinions of his peers either. I could probably produce any number of recent quotes that state to the effect:

"No physical theory of local Hidden Variables [i.e. local realism] can ever reproduce all of the predictions of Quantum Mechanics."

Norsen is completely wrong when he challenges:

"...anyone who claims that Bell’s Theorem is a theorem about 'local realist' theories ... needs to explain clearly what they mean by 'realism' [done, see 2. above] and show precisely where such 'realism' is assumed in the derivation of Bell’s inequalities. [done, see 4. above]"]

I don't know much about this topic, but I would like to understand your point of view. It looks like you define realism in "2. above" as "naive realism". On the other hand, I believe Norsen argues against using this definition with respect to the set of local realist theories allegedly banned by the Bell theorem using the following reasoning: "it was already known, prior to Bell’s Theorem and prior to any experimental tests of Bell’s inequalities, that Non-Contextual HVTs (i.e., Naive Realist theories) are wrong, are not empirically viable." So let me ask you if you disagree with the quoted Norsen's statement or you just believe that, while the statement is correct, naive realism is still what the Bell theorem is about?

I have refuted his argument by reference to the key articles themselves. Norsen's historical perspective is extremely skewed, and this was pointed out by Shimony himself in another one of Norsen's historical outings. I know Norsen is bright and knowledgeable, but that does not change my opinion of the interpretation of Bell's result (nor my conclusion that Norsen is wrong). And apparently, it hasn't swayed the opinions of his peers either. I could probably produce any number of recent quotes that state to the effect:

"No physical theory of local Hidden Variables [i.e. local realism] can ever reproduce all of the predictions of Quantum Mechanics."

One can agree or disagree with the statement in bold font, but my problem with it (and I apologise if the following may look as a distraction for this thread) is that a much stronger statement may be equally true or wrong: "No physical theory free of mutually contradictory assumptions can ever reproduce all of the predictions of Quantum Mechanics." Indeed, it seems to me that standard quantum mechanics contains mutually contradictory assumptions: while unitary evolution does not allow irreversibility, the projection postulate requires irreversibility, thus unitary evolution and the projection postulate are mutually contradictory, and the Bell theorem requires the latter or something similar as an essential assumption.

DrChinese
Gold Member

1. I don't know much about this topic, but I would like to understand your point of view. It looks like you define realism in "2. above" as "naive realism". On the other hand, I believe Norsen argues against using this definition with respect to the set of local realist theories allegedly banned by the Bell theorem using the following reasoning: "it was already known, prior to Bell’s Theorem and prior to any experimental tests of Bell’s inequalities, that Non-Contextual HVTs (i.e., Naive Realist theories) are wrong, are not empirically viable." So let me ask you if you disagree with the quoted Norsen's statement or you just believe that, while the statement is correct, naive realism is still what the Bell theorem is about?

2. One can agree or disagree with the statement in bold font, but my problem with it (and I apologise if the following may look as a distraction for this thread) is that a much stronger statement may be equally true or wrong: "No physical theory free of mutually contradictory assumptions can ever reproduce all of the predictions of Quantum Mechanics." Indeed, it seems to me that standard quantum mechanics contains mutually contradictory assumptions: while unitary evolution does not allow irreversibility, the projection postulate requires irreversibility, thus unitary evolution and the projection postulate are mutually contradictory, and the Bell theorem requires the latter or something similar as an essential assumption.

1. I completely disagree with Norsen's assertions. Notice that I quote from EPR and Bell directly to come to my point. Norsen takes a tour of things that are tangential to the issues of EPR and Bell (such as "metaphysical realism"). EPR and Bell talk about realism (elements of reality) and I don't see any reason to call it "naive" in the circumstances. You would be hard pressed to convince science historians that Einstein had abandoned realism later in life after specifically saying:

"I think that a particle must have a separate reality independent of the measurements. That is: an electron has spin, location and so forth even when it is not being measured. I like to think that the moon is there even if I am not looking at it."

I really don't see how this definition is ambiguous in any way, nor do I see it as naive. Nor have I seen any historical evidence that the subject of "local realism" had simply become an issue of "locality" by the time of either the EPR or the Bell papers - as Norsen asserts. If it had, we'd all be Bohmians now (and obviously we aren't).

How can Norsen say, with a straight face, that "it was already known, prior to Bell’s Theorem and prior to any experimental tests of Bell’s inequalities, that Non-Contextual HVTs (i.e., Naive Realist theories) are wrong, are not empirically viable." I have read my share of historical papers, and that is as far from accurate as it gets. For that to be true, we'd need to believe that all scientists would have agreed that Bell tests (not yet invented, of course) would have expected the results per Aspect. Now, how exactly could that have been true, considering such experiments were not generally discussed until many years later? Yes, a few had heard of EPR-B, but not many until later.

Regarding the Kochen Specker theorem: it not only was not mainstream in 1965, it wasn't published yet. Please reference: Kochen, S. and Specker, E. (1967): "The Problem of Hidden Variables in Quantum Mechanics". Gleason's theorem was around, but I doubt you'd find too many who would call that a well-known result (even today).

Finally, there are plenty of scientists today who believe that nature is local non-realistic, in diametric opposition to Norsen's assertion. We would need to accept that scientists correctly rejected realism prior to Bell, but erroneously did not reject locality after Aspect. Show me anything that makes that sound like a mainstream history of the past 50 years.

No, Bell is about realism a la Einstein (who accepted it and thought QM was incomplete), a la Bohr (who rejected it because he saw QM as complete), a la the many other scientists who had no idea whether Einstein was right or Bohr was right. That is what Bell's Theorem is all about.

2. As to mutually contradictory assumptions: it is locality and realism that are the 2 on the table, and they are not mutually contradictory. Perhaps it is possible to tie them to ideas that become contradictions, but that would require additional debate.

1.
How can Norsen say, with a straight face, that "it was already known, prior to Bell’s Theorem and prior to any experimental tests of Bell’s inequalities, that Non-Contextual HVTs (i.e., Naive Realist theories) are wrong, are not empirically viable.".

Thank you, this is what I wanted to understand. So you just disagree with this statement. I am not sure I have my own opinion on this matter.

2. As to mutually contradictory assumptions: it is locality and realism that are the 2 on the table, and they are not mutually contradictory. Perhaps it is possible to tie them to ideas that become contradictions, but that would require additional debate.

What I was trying to say is that predictions of standard quantum mechanics are mutually contradictory, which makes it quite difficult to reproduce all of them anyway:-).

DrChinese
Gold Member

What I was trying to say is that predictions of standard quantum mechanics are mutually contradictory, which makes it quite difficult to reproduce all of them anyway:-).

There are certainly some strange elements; spin and entanglement predictions being some. Clearly, that sets a very high bar for theory development; we can't have a new candidate theory that makes predictions which are different than reproducible experiments...

EPR asked if you could learn more about a SINGLE particle (in this case) than the HUP allows. So that is the ultimate question: does an independent particle have well-defined values for non-commuting observables? If YES, well, that is realism, my friend. You can call it naive if you like but that is CERTAINLY what Einstein believed in.

From my understanding of the EPR paper, the above seems to be not quite accurate. The EPR paper says:

If, without in any was disturbing a system, we can predict with certainty (i.e., with probability equal to unity) the value of a physical quantity, then there exists an element of physical reality corresponding to this physical quantity. (pp 777)

It appears to me they are saying, position and momentum can not both be simultaneously real parameters of particle. In other words, not being able to simultaneously predict or measure both values with certainty indicates that they are not both simultaneously real quantities. This is subtly different from saying a real particle is one in which both position and momentum are simultaneously well defined.

Clearly they indicate this to be the issue because on the next page they say:

From this follow that either (1) the quantum-mechanical description of reality given by the wave function is not complete or (2) when the operators corresponding to two physical quantities do not commute the two quantities cannot have simultaneous reality.

Their definition of real quantities also did not indicate that the two observables be non-commuting. They go on to explain that no reasonable definition of reality should permit the reality of the position and momentum of one system to be dependent on the process of measurement carried out on a different system which does not disturb the first system in any way.

DrChinese
Gold Member

DrC said: "EPR asked if you could learn more about a SINGLE particle (in this case) than the HUP allows. So that is the ultimate question: does an independent particle have well-defined values for non-commuting observables? If YES, well, that is realism, my friend. You can call it naive if you like but that is CERTAINLY what Einstein believed in."

1. From my understanding of the EPR paper, the above seems to be not quite accurate. The EPR paper says:

"If, without in any was disturbing a system, we can predict with certainty (i.e., with probability equal to unity) the value of a physical quantity, then there exists an element of physical reality corresponding to this physical quantity. (pp 777)"

2. It appears to me they are saying, position and momentum can not both be simultaneously real parameters of particle. In other words, not being able to simultaneously predict or measure both values with certainty indicates that they are not both simultaneously real quantities. This is subtly different from saying a real particle is one in which both position and momentum are simultaneously well defined.

3. Clearly they indicate this to be the issue because on the next page they say:

"From this follow that either (1) the quantum-mechanical description of reality given by the wave function is not complete or (2) when the operators corresponding to two physical quantities do not commute the two quantities cannot have simultaneous reality."

Their definition of real quantities also did not indicate that the two observables be non-commuting. They go on to explain that no reasonable definition of reality should permit the reality of the position and momentum of one system to be dependent on the process of measurement carried out on a different system which does not disturb the first system in any way.

1. There is no difference in what I say and what you quote. Realism = elements of reality are simultaneous for non-commuting (or commuting, for that matter) observables. This exactly matches the words and intent of EPR. They flat out say that there are elements of reality to Alice and Bob BECAUSE they can be predicted in advance. And that is an excellent argument, one that makes perfect sense. But Bell demonstrated it was simply wrong - but not 'til later!

Let's use a specific example, going back to photons. According to EPR, Alice's photon is measured at any angle X. If Bob's photon is measured at X, the outcome can be predicted with certainty. Therefore, there is an element of reality associated with X. Now, we know if we measure Alice and Bob at Y instead where X<>Y, we can still predict the results. So if we do the same at every angle and circle 360 degrees, we will convince ourselves that there was an element of reality associated with every possible position. Because when Alice and Bob are measured at the same angle, the results are perfectly correlated.

So, EPR asserts that hidden variables (or realism, or elements of reality) make sense as an explanation for the phenomena. The problem was, the essential flaw in the EPR argument does not show up in the particular example used - which QM also predicts the same results. Their basic concept of reality was fine, but it took Bell to see the additional requirement that makes local realism untenable. We will come back to this in 3.

2. It is QM, not EPR, that says p and q are not simultaneously well-defined. EPR denies this because they are asserting realism. And it is not the particle's reality that is in question, it is the simultaneous realism of well-defined values for non-commuting observables of that particle.

3. EPR's conclusion, as listed here, is technically correct. The issue is that EPR thought they proved (1), when actually it is (2) that is true.

Repeating the question EPR asks (my words): "does an independent particle have well-defined values for non-commuting observables?" Translated (their words, which are identical in essence to mine): "when the operators corresponding to two physical quantities do not commute, do the two quantities have simultaneous reality? EPR's correct conclusion was "(2) when the operators corresponding to two physical quantities do not commute the two quantities cannot have simultaneous reality".

Now, why not (1) which is what EPR thought they had proved (that QM was incomplete)? Because EPR ruled out (2) because they felt it was "unreasonable". That is the only mistake in the paper, as Bell later showed that they did not use a sufficiently strong mathematical requirement for realism. If they had, they would have seen that it is their viewpoint that was not "complete", as it yields predictions at odds with experiment.

In their view of their paper's result, they "proved" that you could beat the HUP. But we know now that you cannot. I doubt you disagree with that point.

1. There is no difference in what I say and what you quote. Realism = elements of reality are simultaneous for non-commuting (or commuting, for that matter) observables.
This exactly matches the words and intent of EPR. They flat out say that there are elements of reality to Alice and Bob BECAUSE they can be predicted in advance.
If that were the case, a real system could be made unreal by describing it with non-commuting observables. That is why I said the difference is subtle. The reality in question was that of the physical quantities in the theory and not of the underlying system being studied.

Let's use a specific example, going back to photons. According to EPR, Alice's photon is measured at any angle X. If Bob's photon is measured at X, the outcome can be predicted with certainty. Therefore, there is an element of reality associated with X. Now, we know if we measure Alice and Bob at Y instead where X<>Y, we can still predict the results. So if we do the same at every angle and circle 360 degrees, we will convince ourselves that there was an element of reality associated with every possible position. Because when Alice and Bob are measured at the same angle, the results are perfectly correlated.
Again you are confusing the reality of the quantities in the theory with that of the system.

The problem was, the essential flaw in the EPR argument does not show up in the particular example used - which QM also predicts the same results.
I think maybe you misunderstood their example.

2. It is QM, not EPR, that says p and q are not simultaneously well-defined. EPR denies this because they are asserting realism. And it is not the particle's reality that is in question, it is the simultaneous realism of well-defined values for non-commuting observables of that particle.
This is not accurate. No where in that article does EPR deny the fact that p and q are not simultaneously well-defined. In fact it is a central aspect of their proof that QM must not be complete.

3. EPR's conclusion, as listed here, is technically correct. The issue is that EPR thought they proved (1), when actually it is (2) that is true.

Repeating the question EPR asks (my words): "does an independent particle have well-defined values for non-commuting observables?" Translated (their words, which are identical in essence to mine): "when the operators corresponding to two physical quantities do not commute, do the two quantities have simultaneous reality? EPR's correct conclusion was "(2) when the operators corresponding to two physical quantities do not commute the two quantities cannot have simultaneous reality".

Now, why not (1) which is what EPR thought they had proved (that QM was incomplete)? Because EPR ruled out (2) because they felt it was "unreasonable". That is the only mistake in the paper, as Bell later showed that they did not use a sufficiently strong mathematical requirement for realism. If they had, they would have seen that it is their viewpoint that was not "complete", as it yields predictions at odds with experiment.

In their view of their paper's result, they "proved" that you could beat the HUP. But we know now that you cannot. I doubt you disagree with that point.

I think maybe you misunderstood their conclusion. EPR proved that there were two alternatives:
(1) The QM description of reality given by the wave function is not complete
(2) When the operators corresponding to two physical quantities do not commute, they cannot have simultaneous reality.

This derivation of both alternatives IS the main conclusion of the EPR paper.

They then go on to show that if you ASSUME (1) to be false (as Bohr was arguing then), you end up contradicting (2) as well which they have already proven is the only other valid alternative to (1).

In essence they were forcing the hand of those who claimed QM was complete by showing that if they believed that, they must also accept that position and momentum are both simultaneously real physical quantities (see their defintion of complete earlier in the paper). But if they accept that, then they find themselves in a situation in which the reality of the position AND momentum of the SECOND particle which has not been disturbed in any way was dependent on a measurement carried out on the FIRST particle with which the SECOND particle is not interacting.

The issue was not an attempt to defeat HUP for a single particle as you seem to be saying, but to show that if QM is complete, a reasonable definition of reality should not permit what the QM wave function permitted.

DrChinese
Gold Member

1. If that were the case, a real system could be made unreal by describing it with non-commuting observables. That is why I said the difference is subtle. The reality in question was that of the physical quantities in the theory and not of the underlying system being studied... Again you are confusing the reality of the quantities in the theory with that of the system...I think maybe you misunderstood their example.

2. This is not accurate. No where in that article does EPR deny the fact that p and q are not simultaneously well-defined. In fact it is a central aspect of their proof that QM must not be complete.

3. I think maybe you misunderstood their conclusion. EPR proved that there were two alternatives:
(1) The QM description of reality given by the wave function is not complete
(2) When the operators corresponding to two physical quantities do not commute, they cannot have simultaneous reality.

This derivation of both alternatives IS the main conclusion of the EPR paper.

They then go on to show that if you ASSUME (1) to be false (as Bohr was arguing then), you end up contradicting (2) as well which they have already proven is the only other valid alternative to (1).

In essence they were forcing the hand of those who claimed QM was complete by showing that if they believed that, they must also accept that position and momentum are both simultaneously real physical quantities (see their defintion of complete earlier in the paper). But if they accept that, then they find themselves in a situation in which the reality of the position AND momentum of the SECOND particle which has not been disturbed in any way was dependent on a measurement carried out on the FIRST particle with which the SECOND particle is not interacting.

The issue was not an attempt to defeat HUP for a single particle as you seem to be saying, but to show that if QM is complete, a reasonable definition of reality should not permit what the QM wave function permitted.

I guess we can argue all day as to who doesn't understand and who does. So I will skip that part.

1. I do not assert that the system is unreal or real, etc. What we are discussing is whether there are well-defined values for non-commuting observables at points in time where those observables are not being measured. So it is the values of p and q we are discussing. Not sure what subtlety you refer to. Could you reference a specific relevant point if this is still an issue? Otherwise we can move on.

2. Again, it is QM that asserts that p and q are not simultaneously well-defined. Obviously, if you read the final sentence, EPR doubt that is so: "We, however, believe such a [more complete] theory is possible." Do you really question their viewpoint on the HUP at that point?

3. I completely agree with your assessment of what their proper conclusion is, which is that either (1) or (2) is true. So we agree here.

I agree that Bohr would argue that (1) is false, ergo (2) is true. So we agree here too.

EPR assumes (2) is false - which makes (1) true by their logic - because of 2 different reasoning components:

a) In their words, "no reasonable definition of reality could be expected to permit this." They had previously included their "reasonable definition of reality" in their argument that leads to the negation of (2) above. Thus, the negation of (2) that they claimed is dependent on their assumptions.

b) And what were those assumptions? They were: p is an element of reality, and q is an element of reality, and they could both be measured to any degree of accuracy, thereby beating the HUP. Of course, we now know this assumption has been experimentally falsified via Aspect et al. So here is the quote from EPR which backs up my description of their assumption, from the first paragraph after their (18): "In the first case we must consider the quantity P as being an element of reality, in the second case the quantity Q is an element of reality." They then say that this conclusion could be objected to if you insist on simultaneous requirement of measurement or prediction (which we would in an experiment).

So what we are left we is that their deduction of either (1) or (2) being true is a good and proper conclusion. But their ruling out (2) was erroneous, because they based their contradiction on an unwarranted assumption. That assumption being the result of a thought experiment (given their view of a reasonable definition of reality) when, in fact, the true result of that experiment was the opposite of their assumption when it was actually performed nearly 50 years later.

Now, what does this say about Norsen?

I. Clearly, the definition of elements of reality are not naive or trivial and I hope we can drop usage of these words.
II. Clearly, locality is not the main issue either. For if it were, EPR would have needed to add a third part to their main conclusion of (1) and (2), and that would be (3) that the process of measurement on one system disturbs the second system in some way. Gosh, how did they forget that if EPR was all about locality and realism had already been shown to be false? So Norsen is completely wrong about his point that EPR and Bell assume realism is false. Realism is the main point.
III. Clearly, the HUP is an issue and one could derive the result, as EPR did, that belief in the HUP leads to the assertion that there cannot be simultaneous reality for P and Q. We certainly know now, after Aspect, that EPR realism is false...
IV. Unless, of course, the process of measurement on one system disturbs the second system in some way. I.e. non-locality is true. Which is exactly what every book says about Bell, that either realism is false or non-locality is true. In complete contradiction to Norsen's conclusion.

Last edited:
mysearch
Gold Member

As a newcomer to QM, I am simply trying to get a handle on some of the key issues and do not want to get in the way of a very interesting discussion. However, I have posted some questions about HUP in the following thread and was interested in comments linked to this topic: https://www.physicsforums.com/showthread.php?t=306829

It appears to me they are saying, position and momentum can not both be simultaneously real parameters of particle. In other words, not being able to simultaneously predict or measure both values with certainty indicates that they are not both simultaneously real quantities.

2. Again, it is QM that asserts that p and q are not simultaneously well-defined. Obviously, if you read the final sentence, EPR doubt that is so: "We, however, believe such a [more complete] theory is possible."

In the thread above, I was trying to get a clarification of the QM position that appears to state that uncertainty is intrinsic, at least, at the quantum level and not just as issue of measurement. However, my question in this thread is whether the extent of realism you are discussing has to be qualified by the scale of the particles in question, i.e. is there a point where [p] and [q] cease being simultaneously ambiguous and classical reality returns?

1. I do not assert that the system is unreal or real, etc. What we are discussing is whether there are well-defined values for non-commuting observables at points in time where those observables are not being measured. So it is the values of p and q we are discussing. Not sure what subtlety you refer to. Could you reference a specific relevant point if this is still an issue? Otherwise we can move on.
The non-existence of well-defined values for non-commuting observables at points in time where those observables are not being measured is not at issue. EPR proved this aspect from QM. They never question it. Your statement seems to imply that they questioned it. They didn't.

2. Again, it is QM that asserts that p and q are not simultaneously well-defined.
Agreed. This is never in question in the EPR paper. (2) in their conclusion is derived from that fact.

Obviously, if you read the final sentence, EPR doubt that is so: "We, however, believe such a [more complete] theory is possible." Do you really question their viewpoint on the HUP at that point?
That is not how I understand the final sentence. They are not here saying, they believe it is possible to have a more complete theory in which non-commuting observables will be simultaneously real. The bold part is the crucial part, they have already previously proven that to be impossible! I understand them to be saying that it is possible to have a more complete theory in which the physical quantities describing the system each have simultaneous reality (note the absence of non-commuting observables in this sentence). Your characterization of their final sentence implies they claim it it is possible to have a a theory in which non-commuting observables will have simultaneous reality. They already proved that this is not possible (alternative (2)). Again the crucial phrase is the one in bold.

I agree that Bohr would argue that (1) is false, ergo (2) is true. So we agree here too.
Not quite. EPR showed that according to if you assume QM to be complete, you arrive at the conclusion that non-commuting observables can have simultaneous reality which as they proved in the paper is not possible since (1) and (2) can not both be false. Therefore, your ergo above leads to a contradiction. Therefore the assumption that QM is complete must be false.

EPR assumes (2) is false - which makes (1) true by their logic - because of 2 different reasoning
components:

a) In their words, "no reasonable definition of reality could be expected to permit this." They had previously included their "reasonable definition of reality" in their argument that leads to the negation of (2) above. Thus, the negation of (2) that they claimed is dependent on their assumptions.

b) And what were those assumptions? They were: p is an element of reality, and q is an element of reality, and they could both be measured to any degree of accuracy, thereby beating the HUP. Of course, we now know this assumption has been experimentally falsified via Aspect et al. So here is the quote from EPR which backs up my description of their assumption, from the first paragraph after their (18): "In the first case we must consider the quantity P as being an element of reality, in the second case the quantity Q is an element of reality." They then say that this conclusion could be objected to if you insist on simultaneous requirement of measurement or prediction (which we would in an experiment).

Note that the assertions were either:
(1) QM is not complete, OR
(2) non-commuting observables do not have simultaneous reality

So it is not clear to me how you can say they claim (2) is false. What negates (2) is the when the assumption that QM is complete is taken to it's logical conclusion. This is what they set out to show:

"In QM it is usually assumed that the wave function does contain a complete description of the physical reality of the system in the state to which it corresponds. At first sight this assumption is entirely reasonable, for the information obtainable from a wave function seems to correspond exactly to what can be measured without altering the state of the system. We shall show, however, that this assumption, together with the criterion of reality given above, leads to a contradiction"

So what we are left we is that their deduction of either (1) or (2) being true is a good and proper conclusion. But their ruling out (2) was erroneous, because they based their contradiction on an unwarranted assumption.

They never ruled out (2). Rather, they showed that assuming that QM as complete leads naturally to the elimination of (2) which is the only other possible alternative, thus a contradiction -- therefore the assumption that QM is complete is faulty. So clearly this is affirming (2) rather than ruling it out.

As to the relevance of their thought experiment, they explained that no reasonable definition of reality should permit the reality of an observable at a distant non-interacting system to be dependent on a measurement done on a different distant system which is not interacting with the first system or disturbed it in any way. If you take issue with this, suggest a reasonable definition of reality which permits this.

I believe I have shown elsewhere on this forum the faults of Bell's theorem and interpretations of Aspect-type experiments. So I would not rehash those arguments here.

Last edited:
DrChinese
Gold Member

1. The non-existence of well-defined values for non-commuting observables at points in time where those observables are not being measured is not at issue. EPR proved this aspect from QM. They never question it. Your statement seems to imply that they questioned it. They didn't.

Note that the assertions were either:
(1) QM is not complete, OR
(2) non-commuting observables do not have simultaneous reality

So it is not clear to me how you can say they claim (2) is false. What negates (2) is the when the assumption that QM is complete is taken to it's logical conclusion... They never ruled out (2). Rather, they showed that assuming that QM as complete leads naturally to the elimination of (2) which is the only other possible alternative, thus a contradiction -- therefore the assumption that QM is complete is faulty. So clearly this is affirming (2) rather than ruling it out.

2. As to the relevance of their thought experiment, they explained that no reasonable definition of reality should permit the reality of an observable at a distant non-interacting system to be dependent on a measurement done on a different distant system which has not interacted with the first system or disturbed the first system in any way. If you take issue with this, suggest a reasonable definition of reality which permits this.

3. I believe I have shown elsewhere on this forum the faults of Bell's theorem and interpretations of Aspect-type experiments. So I would not rehash those arguments here.

To keep things straight:

(1) QM is not complete.
(2) non-commuting observables do NOT have simultaneous reality.
(3) either (1) is true, or (2) is true, this is an uncontested conclusion of EPR.
(4) according to EPR: if QM is complete, i.e. (1) is false, then non-commuting observables DO have simultaneous reality.
(5) since (4) negates (2), then by (3) we must accept (1).
(6) conclusion (4) was contested by Bohr and many others QM proponents at the time, and they rejected (1) and (4) and of course (5).

Hopefully we can agree on the above. If you don't accept these, there is really no point because EPR was NOT generally accepted at that time as proving QM was not complete.

1. EPR says that QM implies (2). Then they prove (4), which negates (2) making (1) true. So of course they reject (2). Not sure how anyone can walk away with any other conclusion. They say that QM is incomplete! So obviously they deny (2). They repeat this at the end and I quoted it.

2. Of course I take issue with this, and so does virtually the entire physics community!! That is because Bell's Theorem is generally accepted. Reality does NOT need to match the definition provided by EPR, and it certainly doesn't need to be reasonable. That is the whole point!!!

3. I am just now realizing that you deny Bell, which means that you also disagree with Norsen. So what stake do you have in this thread? I thought we were discussing Norsen's paper, and now I realize you have ignored Norsen's paper altogether. We can talk about Bell on another thread.

Not sure how anyone can walk away with any other conclusion. They say that QM is incomplete! So obviously they deny (2). They repeat this at the end and I quoted it.
The fact that both (1) and (2) can not both be simultaneously false, does not mean both (1) and (2) can not both be simultaneously true. That is your error in interpreting their conclusion.

2. Of course I take issue with this, and so does virtually the entire physics community!! That is because Bell's Theorem is generally accepted. Reality does NOT need to match the definition provided by EPR, and it certainly doesn't need to be reasonable. That is the whole point!!!
So then if you disagree with Norsen, what defintion of reality would you provide? In as much as I disagree with Norsen about Bell, I do agree with him that most of the time when people talk of "local reality" they themselves do not know what reality means.

Also Norsen's definition of "naive realism" does not apply to EPR. Unless you misunderstand EPR to be claiming a real system must have simultaneous values for "non-commuting observables". However, that is not what EPR claimed at all.

DrChinese
Gold Member

1. The fact that both (1) and (2) can not both be simultaneously false, does not mean both (1) and (2) can not both be simultaneously true. That is your error in interpreting their conclusion.

2. So then if you disagree with Norsen, what defintion of reality would you provide? In as much as I disagree with Norsen about Bell, I do agree with him that most of the time when people talk of "local reality" they themselves do not know what reality means.

Also Norsen's definition of "naive realism" does not apply to EPR. Unless you misunderstand EPR to be claiming a real system must have simultaneous values for "non-commuting observables". However, that is not what EPR claimed at all.

1. Now you are really going off the deep end with rabbit trails. The question at hand is whether realism, simultaneous well-defined values for non-commuting operators, is a subject of discussion in EPR and Bell. The question is NOT whether realism is supported or not by these papers, or even if both (1) and (2) can both be true (which of course EPR argues that they cannot). The fact is that EPR's (1) is akin to saying QM makes correct predictions. EPR's (2) is akin to saying realism does not hold. Essentially, EPR says that is QM is correct, then realism does not hold. It also says, by contranegative, that if reality holds then QM is not correct. See where this is going? It's about realism, i.e. FOR 'REALISM'.

2. Definition has been provided above already, of course is the same one used by EPR and Bell. Glad to hear that you and Norsen have a lock on your own private definition of local realism and the rest of us don't have access to same. For what purpose is that? Use the language the rest of us use.

If you want to discuss the role of realism and locality in Bell's Theorem, as Norsen actually does, let's continue. Otherwise I would prefer to discuss with those who are interested in this topic.

The question at hand is whether realism, simultaneous well-defined values for non-commuting operators, is a subject of discussion in EPR and Bell.
That of course seems to be a private definition of realism. It is certainly contrary to the EPR definition of realism and that is what I have been trying to show you. Norsen does not mention that definition anywhere. So how can that be the question?

2. Definition has been provided above already, of course is the same one used by EPR and Bell.
As I have alreay pointed out to you, the definition you keep repeating is totally contrary to the definition from EPR. If by "local realism" you are understand realism to mean what you defined above, then such "local realism" will have nothing to do with EPR as I've already explained and can explain again if you wish.

Glad to hear that you and Norsen have a lock on your own private definition of local realism and the rest of us don't have access to same. For what purpose is that? Use the language the rest of us use.
The definitions used by Norsen in his paper are not new. He did not create them. The have been known and discussed in philosophical circles for centuries. You can not voluntarily ban the use of the term "Naive realism" just because you don't like it. It has a specific widely accepted meaning.

DrChinese
Gold Member

As I have alreay pointed out to you, the definition you keep repeating is totally contrary to the definition from EPR. If by "local realism" you are understand realism to mean what you defined above, then such "local realism" will have nothing to do with EPR as I've already explained and can explain again if you wish.

Definition of realism (actually non-realism) per EPR: "...when the operators corresponding to two physical quantities do not commute the two quantities cannot have simultaneous reality."

I just don't see how it gets any simpler than this. I will pass on further debate with you on the matter, as it seems to have no purpose. If anyone else wants to discuss the Norsen paper, I would be happy to oblige.

The best approach to realism is MUH
But I dont know how to call it.
I like MUH, so I would rather ask you - MUH is it an ultimate realism or an ultimate denial of the one?

I don't take is seriously. It's more the ultimate denial of realism, by making the notion meaningless.

Norsen had an interesting (read: wrong) take on EPR. Locality is assumed in EPR, not really a factor per se. It is simply used as a technique to have particles interact and then separate. So the separability requirement is an experimental requirement so that we can go deeper into the central question of EPR: is there a "more complete" level of reality of a single particle.

Of course locality is assumed in EPR. That is Norson's reading as well: EPR is an argument from pure locality to deterministic (naive) realism about the spin components.

2. Nowhere does EPR discuss FTL influences, which they take to be sacrosanct. Of course, EPR's conclusion was wrong in the sense that they expected a "reasonable" definition of reality, one which is not observer independent. As Bell showed, that was not possible. We might get a clue as to what the paper was about from its title, "On the Einstein Podolsky Rosen Paradox." Clearly, not about locality. These papers are about the nature of reality. Is reality observer dependent, as QM seems to imply? Or is the moon round even when no one is looking, as Einstein believed?

It does not discuss FTL influences and does not have to, once it is an argument from locality to naive realism about the spin components. To conclude that the EPR conclusion (QM incomplete) is wrong is one thing, that their argument (locality -> naive realism) is wrong another one. Pilot wave theory proves that they may be wrong about the reality of the spin components (assuming for simplicity Bohms version of the experiment) without being wrong in their argument (because it is locality that fails).

Once their conclusion was that QM is incomplete, I see no reason to say that they were wrong in this point.

In 1935, there was no question in the minds of EPR that their thought experiment - if performed - would demonstrate that the HUP could be beaten. On the other hand, most others (Bohr etc.) felt the opposite.

Let's not speculate about states of mind. We have facts. One fact is that EPR give an argument from locality to naive realism. Not? If yes, the next question is if the argument is valid. (To be distinguished from the question if the conclusion is valid or the assumption false.)

So that is the ultimate question: does an independent particle have well-defined values for non-commuting observables? If YES, well, that is realism, my friend. You can call it naive if you like but that is CERTAINLY what Einstein believed in.

I CERTAINLY do not have to care, in this question, about what Einstein believed in. (What I believe is that Einstein, knowing about Bell's theorem, would have claimed that this theorem gives the ultimate test between QM and relativity: Or QM, or relativity should be simply wrong. This is what I believe as well. And, given the results of the BI tests, the conclusion is that relativity is wrong. (I resist to speculate about the question if the founder of relativity would have accepted this or tried to save his child in some way. The latter would have been an excusable human weakness.)

4. Bell introduces the requirement of realism after his (14), when he hypothesizes: "it follows that c is another unit vector". This is the leap. Previously, we had 2 particles and 2 measurement setting possibilities. Now he adds the "realistic" requirement that there must also be values at other measurement settings.

Without having Bell at hand, I remember that all we need from realism is already applied in his formula (2). If we hypothesize something, and we do not try to cheat about this, we do not use an expression of type "it follows".

It is pretty obvious that if you don't have an a, b and c, then you can't get to Bell's inequality (15). Realism is NOT inserted as a requirement because it was previously assumed, as Norsen claims. Bell adds it here because it wasn't explicit previously, and needed to be added to obtain Bell's key result.

Correct. But that we have a, b, and c follows from locality (and the weak, non-naive version of realism used explicitly in formula (2)).

5. Now, you are free to disagree with my analysis, as Norsen does. But you will be hard pressed to make Norsen's argument by direct reference to EPR and Bell (not an absolute requirement, but you will see quickly that Norsen's reasoning depends on some very shaky inference).

The inference is an inference following from the EPR argument. It seems, you have not understood the crucial difference between the two notions of realism involved here:

1. The naive realism that any measurement reveals some predefined property of the measured object, say, a spin component.

2. The much weaker metaphysical realism, which allows as well that "measurement results" are results of complex interactions between "measurement device" and "measured object", which makes the use of measurement terminology in these cases inadequate.

Norsen is completely wrong when he challenges:

"...anyone who claims that Bell’s Theorem is a theorem about 'local realist' theories ... needs to explain clearly what they mean by 'realism' [done, see 2. above] and show precisely where such 'realism' is assumed in the derivation of Bell’s inequalities. [done, see 4. above]"

I have refuted his argument by reference to the key articles themselves.

Which I have refuted ;-)

Norsen's historical perspective is extremely skewed, and this was pointed out by Shimony himself in another one of Norsen's historical outings. I know Norsen is bright and knowledgeable, but that does not change my opinion of the interpretation of Bell's result (nor my conclusion that Norsen is wrong). And apparently, it hasn't swayed the opinions of his peers either.

I prefer not to argue about majorities (of peers or whatever group you like to care). Even in politics I'm not a democrat.

I could probably produce any number of recent quotes that state to the effect:

"No physical theory of local Hidden Variables [i.e. local realism] can ever reproduce all of the predictions of Quantum Mechanics."

I can easily write something similar without contradicting Norsen: I have fulfilled his requirement to define what I mean with realism (http://ilja-schmelzer.de/realism/definition.php" [Broken]) and where this definition is used (formula (2) of Bell). So you have to try harder. As well as Bell was not against measurements, Norsen is not against realism, as long as if it is a well-defined notion.

Last edited by a moderator:
DrChinese
Gold Member

1. Of course locality is assumed in EPR. That is Norson's reading as well: EPR is an argument from pure locality to deterministic (naive) realism about the spin components.

2. Without having Bell at hand, I remember that all we need from realism is already applied in his formula (2). If we hypothesize something, and we do not try to cheat about this, we do not use an expression of type "it follows".

Correct. But that we have a, b, and c follows from locality (and the weak, non-naive version of realism used explicitly in formula (2)).

The inference is an inference following from the EPR argument. It seems, you have not understood the crucial difference between the two notions of realism involved here:

3. The naive realism that any measurement reveals some predefined property of the measured object, say, a spin component...

1. I agree that locality is assumed. "...no real change can take place in the second system as a consequence of anything that may be done to the first system." But that hardly makes this an argument of locality to determinism. Reality is defined and mentioned any number of times, and is obviously the prime point EPR makes. They never even question that the result of an actual experiment will be that there is no apparently non-local effect.

2. You have trouble with this idea: the simultaneous realism of a and b is discussed in EPR. In Bell, the simultaneous realism of a, b and c is discussed. If you do not assume the simultaneous realism of a, b and c, you cannot get Bell's result. Assuming the separability requirement (Bell's (2) is not enough). Realism is NOT assumed within (2). Because (2) discussed just a and b. QM does not assert that there is a simultaneous a, b and c because it is complete as is.

3. ...is called an "element of reality" according to EPR. Call it naive if you want, that is what EPR and Bell are about.

EPR and Bell talk about realism (elements of reality) and I don't see any reason to call it "naive" in the circumstances....
I really don't see how this definition is ambiguous in any way, nor do I see it as naive.

If I (following Norsen) use the term "naive realism", I do not mean as well that Einstein has been a naive realist.

His criterion of reality is clearly not one of naive realism. What is classified as "naive realism" is only the conclusion derived by EPR (and reused by Bell) from non-naive realism and locality to naive realism about spin directions (or, in the original, about momentum together with coordinates)

Nor have I seen any historical evidence that the subject of "local realism" had simply become an issue of "locality" by the time of either the EPR or the Bell papers - as Norsen asserts. If it had, we'd all be Bohmians now (and obviously we aren't).

That's your problem of not appreciating the arguments. The facts are one thing, the beliefs of people another one. Even if you prefer one of the popular religions, you should see that most of the religious people on Earth believe into the wrong deities. In a similar way, even if Norsen is right, it does not follow that everybody has to become a Bohmian.

How can Norsen say, with a straight face, that "it was already known, prior to Bell’s Theorem and prior to any experimental tests of Bell’s inequalities, that Non-Contextual HVTs (i.e., Naive Realist theories) are wrong, are not empirically viable."

Very simple, by comparing the dates of impossibility theorems for non-contextual HVT's with Bell's paper. I do not care much about history and have not done it, but at least von Neumann's theorem was well-known. Moreover, "it is known" in science always means only "known by specialists in this domain", and not "known by democratic majorities" in whatever sense.

I have read my share of historical papers, and that is as far from accurate as it gets. For that to be true, we'd need to believe that all scientists would have agreed that Bell tests (not yet invented, of course) would have expected the results per Aspect.

There have been lot's of experiments before Aspect. What was new in Aspect's experiment was how the speed of light was involved. (But I'm not firm enough in history, so I would prefer in my papers a weaker formulation which makes the same point, say, by replacing "empirically not viable" with "in conflict with quantum predictions".)

Now, how exactly could that have been true, considering such experiments were not generally discussed until many years later? Yes, a few had heard of EPR-B, but not many until later.

Very simple "is known" in science does not mean "is known by some majority". All those who have not cared about the particular question do not count.

Finally, there are plenty of scientists today who believe that nature is local non-realistic, in diametric opposition to Norsen's assertion. We would need to accept that scientists correctly rejected realism prior to Bell, but erroneously did not reject locality after Aspect. Show me anything that makes that sound like a mainstream history of the past 50 years.

Please a quote of this diametric opposition. As I see it, Norsen's conclusions are not mainstream, but this clearly does not make them wrong. Again, correctness has nothing to do with majorities. And why it should have been correct to reject realism before Aspect is beyond my understanding.

No, Bell is about realism a la Einstein (who accepted it and thought QM was incomplete), a la Bohr (who rejected it because he saw QM as complete), a la the many other scientists who had no idea whether Einstein was right or Bohr was right. That is what Bell's Theorem is all about.

I agree. And Norsen is about realism a la Einstein. Even if his paper is named "against realism".

2. As to mutually contradictory assumptions: it is locality and realism that are the 2 on the table, and they are not mutually contradictory. Perhaps it is possible to tie them to ideas that become contradictions, but that would require additional debate.

I disagree. They have to be specified, of course (Norsen's requirement). But for the most reasonable specifications they are contradictory.

1. There is no difference in what I say and what you quote. Realism = elements of reality are simultaneous for non-commuting (or commuting, for that matter) observables. This exactly matches the words and intent of EPR. They flat out say that there are elements of reality to Alice and Bob BECAUSE they can be predicted in advance. And that is an excellent argument, one that makes perfect sense. But Bell demonstrated it was simply wrong - but not 'til later!

Certainly not. The argument (which differs from the conclusion) remains valid. Bell proves that the conclusion is wrong, thus, the assumption is wrong. The assumption is localitiy
(and an extremely weak version of realism named "metaphysical realism" by Norsen).

2. It is QM, not EPR, that says p and q are not simultaneously well-defined. EPR denies this because they are asserting realism. And it is not the particle's reality that is in question, it is the simultaneous realism of well-defined values for non-commuting observables of that particle.

Realism and locality.

That is the only mistake in the paper, as Bell later showed that they did not use a sufficiently strong mathematical requirement for realism. If they had, they would have seen that it is their viewpoint that was not "complete", as it yields predictions at odds with experiment.

This has nothing to do with the difference between EPR and Bell. The requirements of EPR and Bell regarding realism as well as locality are essentially identical, and the first part of Bell's proof is (see Norsen) simply a repetition of the EPR argument.

In their view of their paper's result, they "proved" that you could beat the HUP. But we know now that you cannot. I doubt you disagree with that point.

There is no need. EPR clearly stated their assumptions. In particular, locality. Locality is wrong, as we know today. Thus, the EPR argument from locality remains valid and correct.

Definition of realism (actually non-realism) per EPR: "...when the operators corresponding to two physical quantities do not commute the two quantities cannot have simultaneous reality."

This is why I like MUH. Words are too slippery and fuzzy.
What quantity is physical? How can we tell physical quantity from non-physical one? Well, physical quantity is real :)

So, I can rewrite the sentence above as

"...when the operators corresponding to two physical quantities do not commute the two quantities cannot be simultaneously physical"

or, if you prefer it this way:

"...when the operators corresponding to two real quantities do not commute the two quantities cannot be simultaneously real"