Is wave function a real physical thing?

[atyy]

Not from unitary evolution alone. You always need coarse graining to derive the classical behavior of measurement/preparation devices.

Sure, introducing coarse graining as an additional postulate is equivalent to introducing a cut and collapse as postulates. Then the measurement problem is that the coarse grained theory makes sense, but the fine grained theory (without hidden variables or MWI) does not, whereas in classical physics both the fine-grained or more fundamental theory and the coarse-grained or emergent theory make sense. It is in this sense that I consider the cut and collapse essential: if you remove it, in a minimal interpretation you must reintroduce the measurement problem by introducing an additional FAPP postulate beyond unitary evolution.

I can live with any interpretation without collapse as a real process, becausevit violates causality.

In the minimal interpretation, the cut and collapse are not necessarily real, they are FAPP. So we have collapse or coarse graining, both of which are FAPP. So here are the questions: Is collapse ontic or epistemic? Is coarse graining ontic or epistemic? Is FAPP ontic or epistemic?

If collapse is not physical, then it is presumably at least partly epistemic. So my point against your argument that the wave function is ontic is that collapse is part of the time evolution of the wave function. Consequently, if one considers collapse to be epistemic, it isn't obvious how the wave function can be purely ontic.

 

[vanhees71]

Sure, introducing coarse graining as an additional postulate is equivalent to introducing a cut and collapse as postulates. Then the measurement problem is that the coarse grained theory makes sense, but the fine grained theory (without hidden variables or MWI) does not, whereas in classical physics both the fine-grained or more fundamental theory and the coarse-grained or emergent theory make sense. It is in this sense that I consider the cut and collapse essential: if you remove it, in a minimal interpretation you must reintroduce the measurement problem by introducing an additional FAPP postulate beyond unitary evolution.

NO! There's a big difference in this approach: It's showing that there are no instantaneous interactions at a distance as claimed with collapse postulates but only local interactions as postulated in all successful relativistic-QFT models (including the standard model).

See also the very nice paper, somebody brought up in one of our "interpretation discussions". I don't like some subtleties like using the word "collapse" and "wave functions" for photons, but the overall conclusion is right. Note that Fig. 1 does not provide the correct interpretation of measurements according to Sect. 3.

Of course, he misses the point somewhat by oversimplifying the math with the entangled states somewhat. I plead guilty for myself in this respect, when I discussed the Scully quantum eraser experiment in this forum. The oversimplification is in leaving out the spatial part of the two-photon state. Here a "wave-packet formulation" is mandatory to make the issue utmost clear, and this also leads to the correction of Fig. 1 described in words at the end of Sect. 3. Of course "wave packet" should not be understood as introducing a "wave function for photons". There cannot be such an object, because the photon hasn't even a position operator in the usual sense. You have locations of detection events, which are well-defined by the fact that photons are detected with devices consisting of massive particles and not because the asymptotic free photon states have a position.

 

[atyy]

NO! There's a big difference in this approach: It's showing that there are no instantaneous interactions at a distance as claimed with collapse postulates but only local interactions as postulated in all successful relativistic-QFT models (including the standard model)

Can you show that the coarse-graining is local and preserves relativistic causality? Peres talks about coarse-graining such that the Wigner function becomes entirely positive, which means that the theory resulting from the coarse-graining is a classical probabilistic theory and therefore realistic. If the coarse-graining is local, the resulting theory is presumably a local realistic theory. However, the Bell theorem forbids local realistic theories, so the theory that results from local coarse-graining presumably cannot explain violations of the Bell inequalities at spacelike separation.

See also the very nice paper, somebody brought up in one of our "interpretation discussions". I don't like some subtleties like using the word "collapse" and "wave functions" for photons, but the overall conclusion is right. Note that Fig. 1 does not provide the correct interpretation of measurements according to Sect. 3.

Of course, he misses the point somewhat by oversimplifying the math with the entangled states somewhat. I plead guilty for myself in this respect, when I discussed the Scully quantum eraser experiment in this forum. The oversimplification is in leaving out the spatial part of the two-photon state. Here a "wave-packet formulation" is mandatory to make the issue utmost clear, and this also leads to the correction of Fig. 1 described in words at the end of Sect. 3. Of course "wave packet" should not be understood as introducing a "wave function for photons". There cannot be such an object, because the photon hasn't even a position operator in the usual sense. You have locations of detection events, which are well-defined by the fact that photons are detected with devices consisting of massive particles and not because the asymptotic free photon states have a position.

I think the paper you are thinking about is Braam Gaasbeek's "Demystifying the Delayed Choice Experiments" http://arxiv.org/abs/1007.3977. I agree with this paper completely. Section 3 does not correct Figure 1. Section 3 says Figure 1 is correct, but that collapse is not necessarily physical (not an frame-invariant event). Quantum mechanics in the minimal interpretation is an FAPP theory, and the predictions of FAPP collapse are thus far completely successful and consistent with special relativity. So this paper does not support your point (unless we are agreeing, but using different language). Rather it supports my point that collapse is part of the standard postulates of quantum mechanics, and is not in conflict with relativity.

 

[DrChinese]

Of course "wave packet" should not be understood as introducing a "wave function for photons". There cannot be such an object, because the photon hasn't even a position operator in the usual sense. You have locations of detection events, which are well-defined by the fact that photons are detected with devices consisting of massive particles and not because the asymptotic free photon states have a position.

Are you in the school of thought that free photons don't exist because they are excitations of the EM field?

 

[vanhees71]

No, there are free-photon states within QED and thus they exist within this framework. However, photons cannot be interpreted as particles like massive quanta, because massless particles with spin $s \geq 1$ have no position observable (at least not one in the strict sense). See Arnold Neumaier's FAQ:

http://arnold-neumaier.at/physfaq/topics/position.html

 

[vanhees71]

Can you show that the coarse-graining is local and preserves relativistic causality? Peres talks about coarse-graining such that the Wigner function becomes entirely positive, which means that the theory resulting from the coarse-graining is a classical probabilistic theory and therefore realistic. If the coarse-graining is local, the resulting theory is presumably a local realistic theory. However, the Bell theorem forbids local realistic theories, so the theory that results from local coarse-graining presumably cannot explain violations of the Bell inequalities at spacelike separation.

I've the kinetic approach in mind. There the "coarse-graining" is done via gradient expansion. I've to think harder, whether this leads to a violation of Bell's theorem. Maybe this is a loophole in my argument!

I think the paper you are thinking about is Braam Gaasbeek's "Demystifying the Delayed Choice Experiments" http://arxiv.org/abs/1007.3977. I agree with this paper completely. Section 3 does not correct Figure 1. Section 3 says Figure 1 is correct, but that collapse is not necessarily physical (not an frame-invariant event). Quantum mechanics in the minimal interpretation is an FAPP theory, and the predictions of FAPP collapse are thus far completely successful and consistent with special relativity. So this paper does not support your point (unless we are agreeing, but using different language). Rather it supports my point that collapse is part of the standard postulates of quantum mechanics, and is not in conflict with relativity.

Yes, that's the paper. Sorry, I forgot to cite it properly again. Here's the crucial point (at the end of Sect. 3):

B. Gaasbek:
We can now solve the problem we started with in the
introduction. If the measurement is nothing but an iso-
lated event in space time, there is no point whatsoever in
trying to associate a spatial slice to it. So the horizontal
and tilted lines in Figure 2 actually have no meaning at
all! Nothing happens along these slices - the only place
where something physical happens is the place of the
measurement, and the implications on conditional prob-
abilities hold for other measurements throughout the en-
tire spacetime, present and past.

If "nothing happens along these slices" in Fig. 2 then the instantaneous collapse proposed in Fig. 1 doesn't happen, or did I get this wrong?

 

[atyy]

If "nothing happens along these slices" in Fig. 2 then the instantaneous collapse proposed in Fig. 1 doesn't happen, or did I get this wrong?

Yes, I agree with that. Let me use my flavour of Copenhagen. I will take a classical/quantum cut (don't take the term "classical" too literally), by which I mean that only things on the classical side are real or physical, while things on the quantum side are not (necessarily) real or physical. As shorthand, I will simply say that things on the quantum side are not real or not physical. So yes, in Copenhagen the instantaneous collapse is not physical and does not really happen. Only the measurement choice and the measurement outcome are real, and these are local classical events.

 

[DrChinese]

No, there are free-photon states within QED and thus they exist within this framework. However, photons cannot be interpreted as particles like massive quanta, because massless particles with spin $s \geq 1$ have no position observable (at least not one in the strict sense). See Arnold Neumaier's FAQ:

http://arnold-neumaier.at/physfaq/topics/position.html

Thanks, that's a treasure trove of his material. :-)

 

[vanhees71]

Yes, I agree with that. Let me use my flavour of Copenhagen. I will take a classical/quantum cut (don't take the term "classical" too literally), by which I mean that only things on the classical side are real or physical, while things on the quantum side are not (necessarily) real or physical. As shorthand, I will simply say that things on the quantum side are not real or not physical. So yes, in Copenhagen the instantaneous collapse is not physical and does not really happen. Only the measurement choice and the measurement outcome are real, and these are local classical events.

Ok, then that's just other wording for the minimal interpretation. I can live with this understanding of "collapse".

Then you would you also call the result of tossing a die and finding it showing "6" a collapse (before you'd have asserted $P("6")=1/6$, if you've known nothing about the specific die before)?

 

[TrickyDicky]

Then you would you also call the result of tossing a die and finding it showing "6" a collapse (before you'd have asserted $P("6")=1/6$, if you've known nothing about the specific die before)?

Well, if all the information about observables you had in classical physics was in this form, I bet you'd call it collapse too, or maybe would use like previously the fancier expression "coarse graining" ;), but it amounts to the same thing: sneaking irreversibility into the reversible Schrodinger equation picture, or into reversible Newtonian laws thru entropy(Boltzmann variety).

 

[atyy]

Ok, then that's just other wording for the minimal interpretation. I can live with this understanding of "collapse".

Then you would you also call the result of tossing a die and finding it showing "6" a collapse (before you'd have asserted $P("6")=1/6$, if you've known nothing about the specific die before)?

Within my flavour of Copenhagen, I cannot answer whether the wave function is ontic or epistemic. The wave function is not necessarily real or ontic because I do take a classical/quantum cut. I like the die analogy, but the analogy between collapse and Bayesian updating is not exact, so I don't know whether collapse is epistemic.

Here's the question I've been trying to ask you: Since you support the die analogy, it seems that collapse is epistemic in the minimal interpretation. However, earlier you argued that the wave function is ontic in the minimal interpretation. Collapse (measurement and selection of a sub-ensemble) is a method of preparing a wave function. If one method of preparing a wave function is epistemic, how can the wave function be ontic?

 

[stevendaryl]

Ok, then that's just other wording for the minimal interpretation. I can live with this understanding of "collapse".

Then you would you also call the result of tossing a die and finding it showing "6" a collapse (before you'd have asserted $P("6")=1/6$, if you've known nothing about the specific die before)?

If the wave function simply reflects the knowledge (or lack thereof) of the observer, then there is nothing weird about a "collapse". When you observe something, you're just discovering a pre-existing value. But that interpretation is difficult to maintain for quantum mechanics, because in some circumstances (for example, EPR), there is no way to interpret the results of a measurement in terms of pre-existing values.

 

[bohm2]

EPR stated such requirement was unreasonable (“No reasonable definition of reality could be expected to permit this.”).

Here's the exact quote from the EPR paper:

This makes the reality of P and Q depend upon the process of measurement carried out on the first system, which does, not disturb the second system in any way. No reasonable definition of reality could be expected to permit this.

Can Quantum-Mechanical Description of Physical Reality Be Considered Complete' ?
http://journals.aps.org/pr/pdf/10.1103/PhysRev.47.777

I take this as just point 2. ("Nonlocality is unreasonable..") mentioned above; that is, the "unreasonableness" that elements of reality could depend on nonlocal effects. Blaylock and Fine agree on this point:

Regarding the possibility that elements of reality could depend on nonlocal effects, they concluded: “No reasonable definition of reality could be expected to permit this.”

The EPR paradox, Bell’s inequality, and the question of locality
http://www.stat.physik.uni-potsdam.de/~pikovsky/teaching/stud_seminar/Bell_EPR-1.pdf

The unreasonableness to which EPR allude in making “the reality [on the second system] depend upon the process of measurement carried out on the first system, which does not in any way disturb the second system” is just the unreasonableness that would be involved in renouncing locality understood as above.

The Einstein-Podolsky-Rosen Argument in Quantum Theory
http://plato.stanford.edu/entries/qt-epr/

 

[DrChinese]

1. Here's the exact quote from the EPR paper:...

2. I take this as just point 2. ("Nonlocality is unreasonable..") mentioned above; that is, the "unreasonableness" that elements of reality could depend on nonlocal effects. Blaylock and Fine agree on this point:

1. Yes, I included that specific quote in my post.

2. Certainly, you must be able to see that going from "This makes the reality of P and Q depend upon the process of measurement carried out on the first system, which does, not disturb the second system in any way. No reasonable definition of reality could be expected to permit this" to "non-locality is unreasonable" is a completely circular argument. EPR says one thing, and the conclusion you reach requires making a substitution that EPR would never agree with. Whether that conclusion is warranted or not is in some ways irrelevant. It wouldn't be a conclusion of EPR, but a conclusion of someone else.

Please note that although Fine's otherwise excellent Plato article skips it, I presented the quote from EPR where realism is assumed/defined explicitly. They flat out said: if you require elements of reality to be simultaneously predictable, you won't reach our conclusion. Ergo, they assume it. It doesn't get much more explicit than that. If you then further ASSUME that non-locality is the ONLY way to have non-realism, then you get your result (as Fine does).

 

[Ross D]

Why? For example EM can be written in complex form:
file:///C:/Users/Administrator/Downloads/Complex%20Maxwell%2527s%20equations.pdf

Physics is basically a mathematical model - all sorts of things can be used to model it.

Thanks
Bill

What I mean by that statement is not that you can't use imaginary numbers in modeling, but that they are not what is actually happening in the real world. For example, you cannot have an imaginary amount of ducks, but it is possible to create a complex function that gives you the amount of ducks. The wave function itself is imaginary and as such is not a physical object, but the probability distribution that it creates is very real and carries a finite amount of energy.

 

[Ross D]

How is the wave function measured?

The wave function is not measured directly, but instead it is defined in terms of three things that can be measured directly: energy, probability, and frequency. The wave function has defined mathematical relationships to all of these, so you can use measurements of these to determine the wave function. You usually have to have some constraints to simplify the process, though.

 

[TrickyDicky]

If the wave function simply reflects the knowledge (or lack thereof) of the observer, then there is nothing weird about a "collapse". When you observe something, you're just discovering a pre-existing value. But that interpretation is difficult to maintain for quantum mechanics, because in some circumstances (for example, EPR), there is no way to interpret the results of a measurement in terms of pre-existing values.

There is no way to make the interpretation within the theory, but the whole point of the EPR paper was that the theory was incomplete, that is to come up with an example that showed clearly the incompleteness. The only possible answer to the paper from within the theory is, either in Bell's form or any other, to get immersed ever deeper in the QM nonlocal flavor of antirealism. Only some interpretations of QM like many-worlds or BM attempt to do that. QFT's locality is in some way an example of the incompleteness of QM if one considers it as realist .

 

[bhobba]

but that they are not what is actually happening in the real world.

The map is not the territory. None of our models tell us what's happening in the 'real' world - whatever that is - philosophers can't seem to agree on that one - it only describes it. Any model using anything is just as valid as any other model - the only criteria is - does it agree with experiment.

you cannot have an imaginary amount of ducks,

You can't have a negative amount of ducks either - but if you owe someone some ducks its a perfectly good way to model it.

Thanks
Bill

 

[bhobba]

The wave function is not measured directly, but instead it is defined in terms of three things that can be measured directly: energy, probability, and frequency.

Energy has nothing directly to do with measuring a wave-function. To measure it you need a large ensemble and have a positive on each observation to determine if its in that state. The only invocation of probability and frequency is determining if the ensemble is large enough so that a positive on every member means the chance of it not being in that state is for all practical purposes zero.

Thanks
Bill

 

[stevendaryl]

QFT's locality is in some way an example of the incompleteness of QM if one considers it as realist .

I'm not sure what you mean by that. QFT is nonlocal, in the same sense that QM is, and I don't see that the nonlocality shows anything about completeness.

To me, that's the lesson of Bell's theorem--the apparent nonlocality of quantum mechanics is not something that is likely to be addressed by a more complete theory, because there isn't a completion that lacks the nonlocality.

 

[stevendaryl]

Certainly, you must be able to see that going from "This makes the reality of P and Q depend upon the process of measurement carried out on the first system, which does, not disturb the second system in any way. No reasonable definition of reality could be expected to permit this" to "non-locality is unreasonable" is a completely circular argument. EPR says one thing, and the conclusion you reach requires making a substitution that EPR would never agree with. Whether that conclusion is warranted or not is in some ways irrelevant. It wouldn't be a conclusion of EPR, but a conclusion of someone else.

I don't understand the quote. If you allow for nonlocal (FTL) interactions, then you CAN'T conclude that a measurement on the first system does not disturb the second system. It's only if you assume Einstein locality that allows you to conclude that distant measurements can't affect each other.

 

[TrickyDicky]

I'm not sure what you mean by that. QFT is nonlocal, in the same sense that QM is, and I don't see that the nonlocality shows anything about completeness.

I was referring to locality of QFT in the sense vanhees was commenting above, my point was
that QFT improves upon the NRQM approximation but I agree I expressed it badly in that sentence.

To me, that's the lesson of Bell's theorem--the apparent nonlocality of quantum mechanics is not something that is likely to be addressed by a more complete theory,
because there isn't a completion that lacks the nonlocality.

Well yes but as you know there are loopholes that exploit avoiding the assumptions of the theorem. QM adheres to those assumptions but a different theory could dodge some of them.

 

[stevendaryl]

Well yes but as you know there are loopholes that exploit avoiding the assumptions of the theorem. QM adheres to those assumptions but a different theory could dodge some of them.

I'm not sure what loopholes you are talking about. To me, the loopholes are variants of nonlocality: If there are FTL influences, then that could explain violations of Bell's inequality. Also superdeterminism and backwards-causality could explain it. (But there is a sense in which both superdeterminism and backwards-causality are ways of messing with locality, or at least with causality, which locality is really about).

 

[stevendaryl]

The map is not the territory. None of our models tell us what's happening in the 'real' world - whatever that is - philosophers can't seem to agree on that one - it only describes it. Any model using anything is just as valid as any other model - the only criteria is - does it agree with experiment.

I've heard people say this, and in one sense, it is clearly true: All we have is models, and we can never know that a model reflects reality, the best we can know is whether it makes predictions that agree with experiment.

However, there is a big difference, in my opinion, between a theory whose entities are intended to be real-world entities, and theories containing entities that are purely for calculation purposes. Newton's theory with particles moving through 3-D space under the influence of forces may have been wrong, or an oversimplification, or an idealization, but it was supposed to be describing the real world. In contrast, in probability theory (at least in many applications), probabilities are not assumed to correspond to anything objective in the world, but to our knowledge about the world. (I guess our knowledge is part of the real world, but that's a more complex correspondence between entities and the real world than in something like Newtonian mechanics).

Quantum mechanics is a little unusual, in that the status of the entities (such as wave functions) are unclear.

 

[bhobba]

However, there is a big difference, in my opinion, between a theory whose entities are intended to be real-world entities, and theories containing entities that are purely for calculation purposes.

Sure - theories often contain stuff that maps out there and stuff that doesn't - it all part of modelling. The number of ducks in a pen maps rather easily - the number you owe to someone doesn't - it's a much more nebulous thing. But just because it maps out there doesn't make it real. Think for example of a point - its supposed to have position but no size - they don't exist out there. But in applying Euclidean geometry we map it to things like pegs surveyors use etc etc that do. We use the human faculty of abstraction to do that. The stuff in our theories are conceptualisations - whether they map easily or not.

The question here is the quantum state real (I dislike wave-function because its enshrines the position basis). I take 'real' to mean - to map to something out there like the ducks in a pen. I take not real to be something like the ducks you owe to someone. Of course a philosopher would likely argue about that - and that's bread and butter for what they do. But I think physicists in general eschew that sort of sophistry and rely on simple common-sense notions - which I believe my view is - but really I can only speak for me.

Based on that, for me, the answer for a state is it depends on your interpretation - so talking about if a state is real or not is meaningless without first specifying what interpretation you are using.

I hold to the ignorance ensemble interpretation which is basically the ensemble interpretation applied to improper mixed states. The observation selects from a conceptual ensemble of possible outcomes. The ensemble resides purely in the head of the theorest as a conceptualisation - it doesn't exist out there like a point - its like the negative ducks - rather nebulous.

Oh and I need to add the objects we model with, say numbers, are ambivalent to existing out there or not. Positive numbers when applied to ducks exist that way - negative numbers in the sense of owing someone do not. But they both exist as points on a line - it purely depends on context.

With regard to the argument since the state is complex it can't be real its exactly the same as with numbers - it exists out there in terms of the points in a plane viewed as an argand diagram - but if it models a quantum state - that purely depends on how you view the quantum state.

Isn't applied math marvellous. Its a conceptualisation - but what it conceptualises varies enormously.

Quantum mechanics is a little unusual, in that the status of the entities (such as wave functions) are unclear.

Most definitely.

Thanks
Bill

 

[DrChinese]

I don't understand the quote. If you allow for nonlocal (FTL) interactions, then you CAN'T conclude that a measurement on the first system does not disturb the second system. It's only if you assume Einstein locality that allows you to conclude that distant measurements can't affect each other.

If anyone thinks this portion of the thread needs to be discussed in a separate thread, just say so...

I agree with you that EPR Locality is assumed. I said that in my post #97. In the EPR paper, realism is first defined. Their definition includes an assumption, should be clear as follows:

Comment: the Bohr view of the EPR example was: there is one system consisting of 2 particles, not 2 systems of one particle as EPR suppose. Therefore, saying “the two systems no longer interact” is not accurate at some unspecified level (in that view).

To bring the EPR and Bohr views into a common frame, you must assume that the *choice* of measurement basis on one part of an entangled system (say by Alice) cannot affect the outcome of a measurement on another part of the system (say by Bob). I would say that the assumption of Einstein/EPR locality would include this. Note that the Bohr view - that there is 1 entangled system and not 2 independent systems - applies in QM even if the 2 entangled particles have never been in a common light cone, or have never even co-existed. That is a strange definition of a single physical system indeed!

You could therefore say that EPR's definition of realism does not hold water if you don't also assume locality (and so defining local realism). That is not the only possible definition of realism, but as they say, it is a reasonable one and should be sufficient. Bell was comfortable with it, since he used a version of it. I.e. Bell had 3 simultaneous elements of reality (a, b and c) compared to EPR's 2 (P and Q).

 

[Ilja]

Well, I don't think that any physical theory can be proven to be complete. So I don't bother about this question very much. So far we have no hint that quantum theory is incomplete, but that doesn't imply that it is complete.

EPR wanted to show that it is incomplete. It used Einstein causality (locality is a misnomer, a theory with 100000 c as a maximal speed would still be local but not Einstein causal). So if one assumes Einstein causality, one can prove that it is incomplete.

By the way, I would sayto prove completeness is also easy: A deterministic theory, if it is correct, is also complete.

Now, the most comprehensive QT we have is relativistic quantum field theory (let's ignore the substantial mathematical problems in its foundations and let's take the physicist's practical point of view to define it in a perturbative sense). By construction the interactions within this theory are strictly local. Nobody could construct a consistent QFT with non-local interactions so far.

Sorry, but QFT is the same old quantum theory, nothing changed, it violates Bell's inequalities, thus, cannot be Einstein causal except in the minimal, weak sense of not allowing FTL phones.

This weak sense forbids only correlations, but refuses to say anything about causation, thus, to name this "Einstein causality" is also misleading. Something like "Einstein correlationality" would be more appropriate.

but these do not violate the relativistic causality structure, as long as you don't consider the collapse as a real process and stick, e.g., to the minimal statistical interpretation (some time ago we had a debate along these lines when discussing the quantum-eraser experiment by Scully et al). Thus, I think EPR rightfully criticized the Kopenhagen collapse doctrine rather than quantum theory itself.

I disagree. If you use "causality" in a meaning which has something to do with causality (instead of being a pure description of correlations without any speculation about how to explain these correlations by causal influences) then a violation of Bell's inequality requires a violation of Einstein causality.

This puzzle was finally solved by Einstein in his famous 1905 paper about what we call Special Relativity Theory today. Then one could have thought that relativistic mechanics + electrodynamics is complete. This idea hold for at most 2 years, when Einstein discovered that he couldn't make easy sense of gravity, which lead to the development of the General Relativity Theory, which was finished by Einstein (and at the same time also Hilbert) in 1915 (big anniversary next year .

Historically incorrect. The idea that one needs a relativistic theory of gravity was clear and well-known already 1905, and the first proposal for a relativistic theory of gravity is part of Poincare's 1905 paper. This was simply scalar gravity distributed by the relativistic wave equation, thus, a wrong solution, but a clear sign that the problem was already known (and that one needs no Minkowski spacetime interpretation to search for relativistic equations for other fields).

 

[DrChinese]

...but it was supposed to be describing the real world. In contrast, in probability theory (at least in many applications), probabilities are not assumed to correspond to anything objective in the world, but to our knowledge about the world.

Isn't that the point of the PBR line of reasoning? I.e. that the probabilities themselves correspond to something that is real/ontic (of course by their specific definition of same)?

I would assert that the wave function - relating to probablities of an outcome - can be manipulated as a real physical thing. For example, I could take a PBS and split an entangled wavefunction in two and later recombine it to restore the entangled state*. Of course, that wave function is not local realistic. If I can move it from place to place, doesn't it have some real attributes?
*J.H. Eberly, Bell inequalities and quantum mechanics (2001).

 

[atyy]

Probability itself has several interpretations. The interpretation that is most beautiful (non-technical term) and coherent (technical term) is the impractical Bayesian interpretation. Although the interpretation of the Kolmogorov axioms for probability can be debated, the algorithms of the "competing" Frequentist interpretation can be derived from the Bayesian interpretation via exchangeability and the de Finetti representation theorem.

If probability itself is open to interpretation, does that complicate the question about whether quantum mechanics is ontic or epistemic? One formal way out of this problem is to ask instead: is quantum mechanics a generalization of classical probability, or is it a special case of classical probability? By asking the question this way, we use the Kolmogorov axioms and don't need to choose between Bayesian and Frequentist interpretations of the Kolmogorov axioms. Here is a tentative classification of some options.

1) QM is a generalization of classical probability (Piron, Hardy http://arxiv.org/abs/quant-ph/0101012, Leifer and Spekkens http://arxiv.org/abs/1107.5849)

2) QM is an ontic special case of classical probability (Bohmian Mechanics, PBR http://arxiv.org/abs/1111.3328)

3) QM is an epistemic special case of classical probability (LJBR http://arxiv.org/abs/1111.3328, ABCL http://arxiv.org/abs/1303.2834)

 

[atyy]

The question I am asking in this post uses the Harrigan and Spekkens definition of ontic and epistemic, in line with the PBR paper of the OP.

Suppose we have an interpretation that is ontic in a particular aether frame and reproduces the Lorentz invariant predictions of relativistic quantum mechanics. Does this interpretation remain ontic in all Lorentz inertial frames, or can the notion of ontic versus epistemic be frame dependent?

 

[bhobba]

Probability itself has several interpretations.

True.

But to the best of my knowledge none are real in a physical sense.

We have frequentest - which to avoid circularity is based on the abstract Kolmogorov axioms, and Bayesian, where it's some kind of degree of confidence a rational agent has.

Thanks
Bill

 

[LachyP]

I think that wave functions can be both, they are a very helpful mathematical tool, especially when dealing with the quantum realm, but I also think that they are real. I wouldn't go as far to say "physical" but i definitely believe that wave functions are real. A nice way of looking at the reality of wave functions is the "double slit experiment". This experiment has been executed not only for study of the quantum nature and wave-particle duality of light but also to study the nature of wave-particle duality and probability waves of electrons. I hope this helps..

 

[vanhees71]

According to quantum theory the probabilistic nature of observables is "ontic" in the sense that "really" not all observables of a quantum mechanical system have determined values in any possible (pure) quantum state. That's it. What's the problem with this idea? Why should Nature be deterministic? The more I read our discussions and the suggested (partially more metaphysical than physics papers) I come to the conclusion that "realistic" is just another word for "deterministic". As long as there is not a deterministic (then necessarily nonlocal) theory that is as comprehensive as quantum theory, I tend to believe that quantum theory is correct, and then nature simply isn't deterministic and the probabilities of quantum theory are the realistic description of nature. That's it, isn't it?

 

[LachyP]

According to quantum theory the probabilistic nature of observables is "ontic" in the sense that "really" not all observables of a quantum mechanical system have determined values in any possible (pure) quantum state. That's it. What's the problem with this idea? Why should Nature be deterministic? The more I read our discussions and the suggested (partially more metaphysical than physics papers) I come to the conclusion that "realistic" is just another word for "deterministic". As long as there is not a deterministic (then necessarily nonlocal) theory that is as comprehensive as quantum theory, I tend to believe that quantum theory is correct, and then nature simply isn't deterministic and the probabilities of quantum theory are the realistic description of nature. That's it, isn't it?

I have been thinking deeply of this, and I think that your statement has said exactly what I have been thinking, but have been unable to express in words.. so I would like to thank you for that haha. PS - Hopefully I interpreted your reply correctly..

 

[atyy]

According to quantum theory the probabilistic nature of observables is "ontic" in the sense that "really" not all observables of a quantum mechanical system have determined values in any possible (pure) quantum state. That's it. What's the problem with this idea? Why should Nature be deterministic? The more I read our discussions and the suggested (partially more metaphysical than physics papers) I come to the conclusion that "realistic" is just another word for "deterministic". As long as there is not a deterministic (then necessarily nonlocal) theory that is as comprehensive as quantum theory, I tend to believe that quantum theory is correct, and then nature simply isn't deterministic and the probabilities of quantum theory are the realistic description of nature. That's it, isn't it?

But what is this "nature" that you talk about? Does "nature" exist after all physicists are dead? If the answer is "no", then you are correct. But if so, I never wish to hear again that "nature does not care what we like" :)

 

[LachyP]

But what is this "nature" that you talk about? Does "nature" exist after all physicists are dead? If the answer is "no", then you are correct. But if so, I never wish to hear again that "nature does not care what we like" :)

Sorry, that may have been a poor choice of words, what I meant by "nature" was "the way it behaves". By knowing that, does it change the meaning of my above statement? Sorry about that.

 

[vanhees71]

I'm not a solipsist. Of course, nature exists well without us. How do you come to the conclusion that the validity of quantum mechanics implies that nature doesn't exist without us humans?

 

[LachyP]

I'm not a solipsist. Of course, nature exists well without us. How do you come to the conclusion that the validity of quantum mechanics implies that nature doesn't exist without us humans?

There are many philosophical theories that suggest that nothing is real, and that it is all inside of our minds. For instance, some would suggest that you are real in my mind but may not really exist in anyone else's mind in the entire world. I despise the concept of it. Some, brought about with the "help" of Niels Bohr's Copenhagen Interpretation of Quantum Mechanics, suggests the possibility that nothing is real, ever, unless observed, or that it is not there when nobody is observing it. I don't like those theories at all. I just think it is horrible that some philosophers took his brilliant interpretation of quantum mechanics and turned it into such a preposterous accusation. Don't you?

 

[stevendaryl]

According to quantum theory the probabilistic nature of observables is "ontic" in the sense that "really" not all observables of a quantum mechanical system have determined values in any possible (pure) quantum state. That's it. What's the problem with this idea? Why should Nature be deterministic? The more I read our discussions and the suggested (partially more metaphysical than physics papers) I come to the conclusion that "realistic" is just another word for "deterministic". As long as there is not a deterministic (then necessarily nonlocal) theory that is as comprehensive as quantum theory, I tend to believe that quantum theory is correct, and then nature simply isn't deterministic and the probabilities of quantum theory are the realistic description of nature. That's it, isn't it?

Actually, I don't agree. There can be nondeterministic theories that I would consider realistic. Stochastic processes are "realistic" in a sense, it's just that the dynamics is nondeterministic.

To me, a "realistic" theory describes the world in terms of objective entities existing in space, with states that evolve over time. The evolution could be nondeterministic. (Okay, I guess this characterization has a nonrelativistic bias).

The many-particle wave function isn't realistic, because it's a function on configuration space, rather than physical space. I guess it could be realistic if we think of configuration space as "real", maybe. In any case, the lack of realism is not about nondeterminism.

 

[vanhees71]

In quantum theory the objective entities are the states, and they evolve even deterministic over time. Why is the many-particle wave function of non-relativistic physics not realistic, only because it's a function on configuration space? It tells you the position probality density for $N$ particles and thus is a function of the $3N$ position coordinates (and $N$ spin components if you have particles with spin). For a classical $N$-body system you need $6N$ phase-space coordinates (e.g., position and canonical momentum coordinates). Is this then also not realistic (despite the fact that it's classical and thus only an approximation, but within classical physics, I'd consider this as a realistic description)?

 

[Ken G]

Some work has been done in this thread attempting to supply a scientific meaning to "physically real." That is of course essential if we wish to give a scientific answer to "is the wave function physically real." But I think the actual problem here is that there is only one scientific meaning to "physically real" that holds up to the process of science itself, and that is essentially the meaning "whatever we say is physically real." In other words, "reality" is our word, and we are the ones who say what it shall mean to us, it is not a word picked out of some pre-existing language that we only dimly understand because that is just not what language is.

As such, it is simply the wrong way to ask the question to frame it is as "which of the devices and concepts invoked by science are the ones that are actually physically real, and which ones are not", the question must be posed as "what whall we mean by physically real such that the devices and concepts we wish to manipulate may be regarded as having that property." In short, the question is not if the quantum state is physically real, it is, given the observed attributes that quantum states have (ability to be nonlocally entangled, inability to provide simultaneous answers to all hypothetical questions that could be put to the state) do we want the quantum state to be physically real, or don't we? You must answer that question first, or you are spinning your wheels. Because the simple truth is, some do want that state to be physically real, because they want to be able to talk about it that way, and others do not, because they want to be able to talk about physically real things as having properties that quantum states do not.

Framed like this, we put the question in its proper context: it is not a physics question, and it is not even a philosophy question-- it is a sociology question. We don't get to decide what a quantum state is, that is constrained by how it is found to behave, but we do get to decide what we want "physically real" to mean, in our modern scientific society. And that meaning is not handed to us in advance because nothing is handed to us in advance in science, we have to explore every nook and cranny ourselves, and it is not common sense because science is constantly challenging the idea that common sense is likely to be correct. We have to be scientists first, and then we have to ask what is physically real, and when we do that, the answer always comes out something like "what we mean by physically real is nothing but our current understanding of the situation, whatever that is." So I would say that the quantum state is indeed physically real, but not because it has attributes that conform to what is physically real-- instead, what is physically real must conform to the attributes of the quantum state, because why would we want our most fundamental understanding of nature to be not what we mean by what is physically real?

 

[vanhees71]

I strongly disagree. Real is what can be objectively observed and quantitatively measured. It's not a social construct or something like this. In history of science there have been independent dicoveries of the same thing (e.g., renormalization of QED has been discovered independently by Tomonaga, Feynman, and Schwinger). The results of meausrements are independent of who performs them and independent from religious, political, or philosophical believes. Also the observable consequences of quantum theory are independent of the interpretation the one or the other physicist follows. In some sense, this huge thread and all our discussions are pretty unscientific but rather philosophical or (at best) metaphysical!

 

[atyy]

I'm not a solipsist. Of course, nature exists well without us. How do you come to the conclusion that the validity of quantum mechanics implies that nature doesn't exist without us humans?

The validity of quantum mechanics does not imply that nature does not exist without us. However, I think your proposal that nature is irreducibly stochastic causes problems for solving the measurement problem.

If there is an observer-independent reality that is describable by physical laws, then there should be hidden variables in order to solve the measurement problem. Such a solution would probably use classical probability (Kolmogorov's axioms), which is always compatible with determinism. This is why I agree with the idea that reality always seems to have something to do with determinism. So if one rejects any possibility of determinism, and insists on irreducible stochasticity, it seems that one is rejecting solutions to the measurement problem, and therefore rejecting that there is a law-like observer-independent reality.

Or can there be solutions to the measurement problem which reject classical probability (Kolmogorov's axioms)?

 

[bohm2]

Some work has been done in this thread attempting to supply a scientific meaning to "physically real." That is of course essential if we wish to give a scientific answer to "is the wave function physically real." But I think the actual problem here is that there is only one scientific meaning to "physically real" that holds up to the process of science itself, and that is essentially the meaning "whatever we say is physically real." In other words, "reality" is our word, and we are the ones who say what it shall mean to us, it is not a word picked out of some pre-existing language that we only dimly understand because that is just not what language is.

Do you believe that some clear definitions of "realism" is even required to draw EPR or Bell's conclusions? As an aside, Gisin recently tried to give some definition of realism:

A theory is realistic if and only if, according to the mathematical structure of this theory, the collection of all physical quantities written in the system unambiguously determines the probabilities of all possible measurement outcomes.

A possible definition of a Realistic Physics Theory
http://arxiv.org/abs/1401.0419

Given this definition both classical mechanics and QM are "realistic". That is the problem with defining realism. It can mean different things to different people.

 

[Ken G]

Do you believe that some clear definitions of "realism" is even required to draw EPR or Bell's conclusions?

If you are saying that "realism" is not the crucial issue in EPR/Bell, I agree-- the crucial issue is what the experiments are doing that are not what we might have expected. EPR and Bell are just means to help us understand the significance of the observational data, as must always be true in science. Philosophical implications are also important, because we want to gain a sense that there is some lesson here, and we need a philosophical framework to decide what that lesson is. But philosophical frameworks are erected by us, not by nature. So I think we must distinguish the EPR/Bell conclusions as what nature is telling us, from how we choose to frame the significance of the data. It is the latter that depends on our goals, whereas the former is much less subjective.

Given this definition both classical mechanics and QM are "realistic". That is the problem with defining realism. It can mean different things to different people.

Yes, we should not regard the meaning of "realism" as handed to us, it is very much a project of our own to decide what we wish that word to mean. Gisin seems to be taking the scientific approach that I also am advocating-- let's choose a meaning that fits our scientific understanding, rather than reverse engineer our understanding to fit our preconceptions. Of course, scientific understanding is a moving target, so if we adopt this approach, it means that what we regard as "real" will also be a moving target. This elicits discomfort from many, but scientists shouldn't recoil from that, discomfort is like a signpost to where we need to go rather than what we need to avoid.

 

[microsansfil]

Hi Ken G,

Some work has been done in this thread attempting to supply a scientific meaning to "physically real." That is of course essential if we wish to give a scientific answer to "is the wave function physically real."
...

You give an interresting epismetic analysis of the use of the sentence "physically real", that coming out of the metaphysical assertions based on a-priori of scientific realism. “This is how humans are: We question all our beliefs, except for the ones that we really believe in, and those we never think to question.”

however, what is the usefulness for science to seek to build a response to this metaphysics question : "is the wave function physically real." ?

Patrick

 

[vanhees71]

The validity of quantum mechanics does not imply that nature does not exist without us. However, I think your proposal that nature is irreducibly stochastic causes problems for solving the measurement problem.

If there is an observer-independent reality that is describable by physical laws, then there should be hidden variables in order to solve the measurement problem. Such a solution would probably use classical probability (Kolmogorov's axioms), which is always compatible with determinism. This is why I agree with the idea that reality always seems to have something to do with determinism. So if one rejects any possibility of determinism, and insists on irreducible stochasticity, it seems that one is rejecting solutions to the measurement problem, and therefore rejecting that there is a law-like observer-independent reality.

Or can there be solutions to the measurement problem which reject classical probability (Kolmogorov's axioms)?

Again, I don't understand your conclusions. The probabilities of quantum theory obey Kolmogorov's axioms. Why shouldn't they? Then, if Nature is intrinsically stochastic why do you need hidden variables to make our description realistic? If it's stochastic then some or even any observables can be indetermined. Only the probability for a measurement outcome then is "real". So what's then "the measurement problem"?

I don't say that there aren't hidden variables or that there isn't a non-local deterministic (and causal!) theory. I only say that so far we haven't found it, and so far I even don't see any necessity to look for one, since quantum theory works so far very well. Perhaps some motivation might be to find a consistent quantum theory of gravitation?

 

[atyy]

Again, I don't understand your conclusions. The probabilities of quantum theory obey Kolmogorov's axioms. Why shouldn't they? Then, if Nature is intrinsically stochastic why do you need hidden variables to make our description realistic? If it's stochastic then some or even any observables can be indetermined. Only the probability for a measurement outcome then is "real". So what's then "the measurement problem"?

The measurement problem is that within quantum mechanics, if there is no observer, there are no measurement outcomes, and thus no observer-independent reality.

I don't say that there aren't hidden variables or that there isn't a non-local deterministic (and causal!) theory. I only say that so far we haven't found it, and so far I even don't see any necessity to look for one, since quantum theory works so far very well. Perhaps some motivation might be to find a consistent quantum theory of gravitation?

But if you don't reject hidden variables or the possibility of a non-local deterministic theory, how can you come to the conclusion that nature is intrinsically stochastic?

 

[stevendaryl]

In quantum theory the objective entities are the states, and they evolve even deterministic over time. Why is the many-particle wave function of non-relativistic physics not realistic, only because it's a function on configuration space? It tells you the position probality density for $N$ particles and thus is a function of the $3N$ position coordinates (and $N$ spin components if you have particles with spin). For a classical $N$-body system you need $6N$ phase-space coordinates (e.g., position and canonical momentum coordinates). Is this then also not realistic (despite the fact that it's classical and thus only an approximation, but within classical physics, I'd consider this as a realistic description)?

Well, classically, a phase-space description is equivalent to particle density and current density as a function of 3-space.

 

[Ken G]

however, what is the usefulness for science to seek to build a response to this metaphysics question : "is the wave function physically real." ?

One can certainly go the road that science is separate from metaphysics, and just avoid a question like that on the basis that it doesn't help you solve an equation or predict an experiment. That is often called the "shut up and calculate" approach to science. Yet, I must say that I've never actually met someone who only does that, even among those who claim to be of that philosophical bent. It's just not what draws us to science-- we wish to understand, we wish to learn the lessons. The learning process often feels like a sense of discomfort and confusion, followed by a sense of a light bulb coming on, a veil being lifted. That is a description of a metaphysical experience, you just can't get a "Eureka" moment by shutting up and calculating. But this presents a problem-- is not that "Eureka" moment something personal, so how can that mesh with the objectivity of science? This is a fundamental conundrum that is at the heart of questions like "is the wave function real", I don't know how to make that conundrum go away. In fact, I don't think we should make that conundrum go away, we are supposed to wrestle with it-- what do we want these words to mean, such that we get a sense of understanding and learning the lessons of science, more so than just being able to predict the outcomes? What are the lessons of the quantum state that are telling us something different from what we thought was true about reality? That's the question of importance, reality is something we are supposed to figure out, not already know.