Living Opponents of the Copenhagen Interpretation

[bhobba]

Bill, thanks for the book order! I'll work on getting those! (I live in Nepal, so it will take time to figure out how to get them here…) But even if I LEARN it, am I wrong to assume that it would be very hard to get anyone to take anything I say seriously without a degree? Or would there be a way to learn it on my own first and then someone get a degree?!

You do have a degree - just not in physics.

Well surprise surprise neither do I - mine is in applied math. I taught myself QM - although I did study analysis and Hilbert Spaces as part of my degree which was a help.

Mate physicists generally listen to what people say - not what they put after their name.

Thanks
Bill

 

[bhobba]

An interference pattern is, for all intents and purposes, the visibility of non-localised information. It's visibility is embodied in the distribution of otherwise localised information (ie. distribution of particle detections).

Cant follow that.

Here is its explanation in terms of QM:
http://cds.cern.ch/record/1024152/files/0703126.pdf

Cant find anything about non locality there. Indeed on the issue of locality QM is pretty much agnostic - although it does depend on your definition of it in the first place. Mine is the cluster decomposition property - but it only applies to uncorrelated systems:
http://en.wikipedia.org/wiki/Cluster_decomposition_theorem

Anyway can't see how any of it applies to the double slit experiment.

Thanks
Bill

 

[harrylin]

[..] All interpretations of quantum mechanics must have the Copenhagen interpretation as emergent. [..]

I had the impression that the Copenhagen interpretation includes philosophy and claims about the unseen; but in view of your answer, that must be wrong, and also http://en.wikipedia.org/wiki/Copenhagen_interpretation must be wrong according to you. As it no doubt matters for the discussion, please give your definition of the Copenhagen interpretation (or a link to it) - thanks!

 

[carllooper]

Cant follow that.
Here is its explanation in terms of QM:
http://cds.cern.ch/record/1024152/files/0703126.pdf

I'm not suggesting anything controversial here - just probably (or obviously) not expressing it very well.

So we're talking about what is visible in a pattern of detections. Proposed was that non-locality is visible in such, ie. that it's not hidden. So the pattern is an image, or a signal. Given enough detections one can perform an FFT on this image and reconstruct a reasonable image of the two slits. The more detections the better. The pattern of detections makes visible the two slits. The pattern of detections therefore provides, or encodes, information about the slits, eg. if one or both are open. An FFT isn't necessary if one knows how to read the pattern of course. The FFT can be regarded as just a means to decode a more familiar image of the slits.

But this information is not localised in anyone detection, so it's "non-localised" in the sense that the information is distributed across the pattern of detections.

Arguably one might say that if only one slit is open the information is localised, but that only works if you already know only one slit is open. From the detections alone, until you accumulate enough detections to rule it in or out, you don't know if one slit is open or not. In other words the information is still not localised in anyone detection. The pattern for a single slit is still an interference pattern - just one which expresses that interference (in the usual sense) isn't happening.

C

 

[Fredrik]

I was wondering if anyone could tell me who are considered the leading opponents of the Copenhagen Interpretation who are alive today?--and even what would be the best way to get myself some formal physics education under such individuals?

What do you think the Copenhagen interpretation is? (If you can't define it, how can you have strong opinions about it?)

I avoid the term "Copenhagen interpretation" as much as I can, because it's very hard to find two people who mean the same thing by it. People who try to define it often include claims that look like extreme misunderstandings of Bohr's ideas.

it is really physicists (and not philosophers of science) who are seen as having the authority to pronounce on the proper interpretation of their field.
[...]
are philosophers of science considered authorities on the interpretation of physics?

Certainly not by physicists.

No one has the authority to label one interpretation as "proper". In fact, it would be very unscientific to do so.

 

[Fredrik]

Opponents of Copenhagen are either crackpots or Ballentine (and Ballentine is wrong).

Ballentine seems to think that his argument against a particular way of thinking about states proves that the way he's thinking about them is the right way. I would agree that he should have toned it down a bit, but you keep saying that "Ballentine" is wrong, as if his entire book can be dismissed. I strongly disagree with that view.

Here is a peer-reviewed paper showing that Ballentine is wrong: Zurek, W. H. 1981. Pointer Basis of Quantum Apparatus: Into What Mixture Does the Wave Packet Collapse? Phys. Rev. D 24: 1516.
Also one can just take a look at standard texts like Landau & Lifshitz, Cohen Tannoudji, Diu & Laloe, Nielsen and Chuang, or Weinberg.
One can also find a discussion of an error made by Ballentine in http://arxiv.org/abs/quant-ph/0312059 (Rev.Mod.Phys.76:1267-1305,2004) and in http://books.google.com/books/about/Exploring_the_Quantum.html?id=QY6YuU-Qi-AC.
This is not a matter of taste like having a favourite interpretation. Ballentine is wrong.

I read that first article years ago. I don't see how it could possibly be relevant. If Schlosshauer's article is anything like his book, that's also my comment about his article. You certainly can't just "take a look at standard texts" to see that Ballentine is wrong.

 

[Fredrik]

There are a few around.
The three I am most aware of are Ballentine:

Apparently you and atyy both consider Ballentine an opponent of Copenhagen. The view that Ballentine argues against is the view that the system can be literally identified with its state, i.e. that the state represents all the properties of the system. That view is the starting point of the MWI, so I find it odd to label it "Copenhagen".

I realize that people often include this view as a part of something they call "Copenhagen", and then they assume that there's a mysterious physical process called "collapse" that's supposed to eliminate the many worlds. To me that's just the MWI plus an absurd idea that probably makes the whole thing inconsistent. There's no evidence that Niels Bohr held beliefs like that, so I wouldn't call this "Copenhagen", even though I know that many people do.

 

[atyy]

Ballentine seems to think that his argument against a particular way of thinking about states proves that the way he's thinking about them is the right way. I would agree that he should have toned it down a bit, but you keep saying that "Ballentine" is wrong, as if his entire book can be dismissed. I strongly disagree with that view.

In this context, it is Ballentine's opposition to Copenhagen that is wrong, not the whole book. If you don't wish to label the view that Ballentine opposes as "Copenhagen", I can agree to that, which is why I talked about "Copenhagen-style/instrumental/operational" interpretations, by which I would include "orthodox" or "shut-up-and-calculate". Bohr in fact never wrote down his interpretation in a paper, so strictly speaking, Copenhagen does not exist (just like the quantum world :) To give a concrete example of what I am calling Copenhagen, I would give the interpretation given in Landau and Lifshitz. It is also true that the view Ballentine opposes is a misrepresentation of Copenhagen, so in that sense, one could also say that he does not oppose any correct view, merely a caricature of it, so in that sense Ballentine is right. However, Ballentine does make the view he opposes seem to be "mainstream", so he is at least misleading. My claim is not that the entire book by Ballentine can be dismissed, but applies to Chapter 9.

I read that first article years ago. I don't see how it could possibly be relevant. If Schlosshauer's article is anything like his book, that's also my comment about his article. You certainly can't just "take a look at standard texts" to see that Ballentine is wrong.

The quotation of standard texts shows that what I am saying is "mainstream physics", in response to strangerep's suggestion that I am violating PF's rules. It is clear that Ballentine is not mainstream physics, and is in opposition to virtually every other textbook.

Zurek's article is cited because the Stern-Gerlach experiment there is analyzed correctly, whereas Ballentine gets it wrong. Here is another correct version of the Stern-Gerlach experiment http://arxiv.org/abs/quant-ph/0306072.

The Schlosshauer article and the book by Haroche and Raimond are cited to show that Ballentine is wrong in criticizing the collapse postulate, yet effectively using it or an equivalent postulate when taking improper mixtures to be proper mixtures.

 

[DrChinese]

Dr. Chinese, isn't it possible for a philosophy-laden theory to experimentally vindicated, but yet philosophically wrong? For example, I could have a theory that there is an invisible string connecting the moon and the Earth and thus the orbit. To refute that, you may bring in other observations (like that they are also spinning relative to one another), but at one point in the discussion (before the new observations were brought to bear) the philosophically-wrong theory matched at least some of the observations. That situation could be compared to the current situation with the failure to find a unifying theory.

You can't really say the "invisible string" theory is incorrect unless it makes incorrect predictions that another theory does not. Quantum interpretations can be judged by their utility, which for now is equal among competing versions. Besides, I don't know that I would call CI a "a philosophy-laden theory". :-)

 

[DrChinese]

Arguably one might say that if only one slit is open the information is localised, but that only works if you already know only one slit is open. From the detections alone, until you accumulate enough detections to rule it in or out, you don't know if one slit is open or not. In other words the information is still not localised in anyone detection.

To use your line of thinking: you can place polarizers over BOTH slits. The relative orientation (parallel or orthogonal) will then determine whether there is interference or not. So in that manner, the variable [relative orientation of polarizers] is always "non-local".

 

[atyy]

I had the impression that the Copenhagen interpretation includes philosophy and claims about the unseen; but in view of your answer, that must be wrong, and also http://en.wikipedia.org/wiki/Copenhagen_interpretation must be wrong according to you. As it no doubt matters for the discussion, please give your definition of the Copenhagen interpretation (or a link to it) - thanks!

There is no single well-defined Copenhagen interpretation, which is why I have referred to Copenhagen-style/instrumental/operational interpretations. One example of what I mean is the interpretation in Landau and Lifshitz, which is associated with Bohr by Bell in his article "Against 'measurement'". Another example is the Copenhagen-style interpretation mentioned by Weinberg in his quantum mechanics textbook.

It is true that historically there is a flavour of Copenhagen which has asserted that quantum mechanics requires a rejection of "naive realism". As I have mentioned twice in this thread (posts #2 and #29), this is not philosophical bias, but mathematical error, and this view has been discredited. In non-relativistic quantum mechanics, we in fact know that there is no such theorem, because of counter-examples, the first of which was discovered by Bohm. Modern Copenhagen-style interpretations do not claim that quantum mechanics requires a rejection of "naive realism".

The key elements of what I consider a Copenhagen-style interpretation are a classical/quantum cut, and agnosticism as to whether the wave function is real. The classical/quantum cut is basically assumed in the quantum information literature, where you will see channels with classical and/or quantum output. The classical/quantum cut taken with relativity provides one way to argue that the wave function is not necessarily real, because only the classical outcomes and their probabilities are Lorentz-invariant, but not wave function evolution if state reduction is taken into account. For example, state reduction is not an "event" in the same sense that simultaneity is not absolute in special relativity (we can say of course say that simultaneity is absolute relative to a family of observers).

One example of a Copenhagen-style interpretation whose authors also support "naive reality" is Leifer and Spekkens's http://arxiv.org/abs/1107.5849. Leifer and Spekkens do have classical/quantum cuts, and they also say "the picture we have in mind is of the quantum state for a region representing beliefs about the physical state of the region, even though we do not yet have a model to propose for the underlying physical states."

 

[Fredrik]

The key elements of what I consider a Copenhagen-style interpretation are a classical/quantum cut, and agnosticism as to whether the wave function is real.

Some people consider "a classical/quantum cut" to be the idea that QM simply doesn't apply to some systems, in particular measuring devices. I find that notion absurd, and I think you do too. If two systems have Hilbert spaces $H_1$ and $H_2$, the composite system has Hilbert space $H_1\otimes H_2$. This is an essential part of the QM framework, and it implies that any object that consists of quantum systems, is a quantum system.

In my opinion, the proper way to talk about these things is this: Science defines a "theory" (roughly) as a set of statements about the real world that makes testable predictions. A testable prediction is a claim about what the result of a measurement will be, or what the average result of a sequence of measurements performed on identical systems will be. A measurement* is an interaction between a measuring device and the system that the theory is about, which leaves the measuring device in one of its possible final states. The possible final states must be easily distinguishable by a human observer. If they're not, we wouldn't consider the object a measuring device.

In the context of QM, these ideas are sometimes very inaccurately described as "a classical/quantum cut". They shouldn't be described this way, because they are just basic requirements of science. That makes them part of every theory of physics, including QM. So they do not need to be added as part of an interpretation, and in fact, it would be wrong to do so.

The kind of "classical/quantum cut" mentioned in my opening sentence could be part of an interpretation, but in my opinion, this is a crackpot idea that started as an extreme misunderstanding of things that Niels Bohr said. I could be wrong about what I'm about to say, but I think Bohr's view was essentially what I just said, and that he was misunderstood in part because this is a difficult subject, and in part because he never found a good way to explain these things.

Because of this, I wouldn't include either of these types of "classical/quantum cuts" in a definition of "Copenhagen".
*) Some of the simplest measurements aren't accurately described by this. For example, if you measure the length of your phone with a ruler, interactions between the phone and the ruler are irrelevant, and the number that we consider "the result" isn't associated with the final state of the ruler; it's associated with a final state of your brain. I suppose I could try to modify my definition of "measurement" to include every conceivable type of process that we might want to call a measurement, but for my current purposes, I think it's better to just keep the definition simple. I think that my definition is good enough for all possible theories, because in situations where a different type of "measurement" is possible, we always have the option to do a measurement that satisfies my definition and is for all practical purposes equivalent to the other measurement.

 

[atyy]

Some people consider "a classical/quantum cut" to be the idea that QM simply doesn't apply to some systems, in particular measuring devices. I find that notion absurd, and I think you do too. If two systems have Hilbert spaces $H_1$ and $H_2$, the composite system has Hilbert space $H_1\otimes H_2$. This is an essential part of the QM framework, and it implies that any object that consists of quantum systems, is a quantum system.

In my opinion, the proper way to talk about these things is this: Science defines a "theory" (roughly) as a set of statements about the real world that makes testable predictions. A testable prediction is a claim about what the result of a measurement will be, or what the average result of a sequence of measurements performed on identical systems will be. A measurement* is an interaction between a measuring device and the system that the theory is about, which leaves the measuring device in one of its possible final states. The possible final states must be easily distinguishable by a human observer. If they're not, we wouldn't consider the object a measuring device.

Because of this, I wouldn't include either of these types of "classical/quantum cuts" in a definition of "Copenhagen".

I think this is a subtle issue - is there a real difference between "naive realism" and "instrumentalism"? One can argue that "reality" is just a tool to help us predict the results of experiments, so in that sense "reality" is not necessarily "real".

But at a less refined level, the classical/quantum cut in Copenhagen is distinctive, because it means that Copenhagen does not assign any meaning to the "wave function of the universe". Copenhagen does not assert that the whole universe is a quantum system. One can contrast the classical/quantum cut with Many-Worlds (or a very literal form of Bohmian Mechanics), which tries to make sense of the "wave function of the universe". One can also contrast the classical/quantum cut with general relativity, in which there is in principle no problem to including the measuring apparatus in the state of the system, and we do believe that general relativity can be a theory of the whole universe.

 

[carllooper]

To use your line of thinking: you can place polarizers over BOTH slits. The relative orientation (parallel or orthogonal) will then determine whether there is interference or not. So in that manner, the variable [relative orientation of polarizers] is always "non-local".

Yes that's right.

 

[aleazk]

I think this is a subtle issue - is there a real difference between "naive realism" and "instrumentalism"? One can argue that "reality" is just a tool to help us predict the results of experiments, so in that sense "reality" is not necessarily "real".

But at a less refined level, the classical/quantum cut in Copenhagen is distinctive, because it means that Copenhagen does not assign any meaning to the "wave function of the universe". Copenhagen does not assert that the whole universe is a quantum system. One can contrast the classical/quantum cut with Many-Worlds (or a very literal form of Bohmian Mechanics), which tries to make sense of the "wave function of the universe". One can also contrast the classical/quantum cut with general relativity, in which there is in principle no problem to including the measuring apparatus in the state of the system, and we do believe that general relativity can be a theory of the whole universe.

That's certainly a subtle issue. According to Isham, we need a more realist (i.e., non-instrumentalist) interpretation of QM if we want to build a QG theory, and QC (quantum cosmology):

http://arxiv.org/pdf/1004.3564.pdf

Apparently, his solution to all these problems (in particular, the clash between the continuum in many of our theories and the supposed non-continuum spacetime of QG) is to rewrite all physics in terms of topos theory rather than set theory.

 

[kith]

It is also true that the view Ballentine opposes is a misrepresentation of Copenhagen, so in that sense, one could also say that he does not oppose any correct view, merely a caricature of it, so in that sense Ballentine is right. However, Ballentine does make the view he opposes seem to be "mainstream", so he is at least misleading.

I think that's one of very few passages in your posts about Ballentine where a person who is not familiar with the book doesn't get a heavily distorted view of it. As Frederik remarked, many of your posts make it seem like Ballentine is a really bad book.

I agree that Ballentine makes it seem like he is criticizing a mainstream view while he is actually arguing against collapse as a physical process. But disregarding this, most of his arguments are valuable and not present in many other standard texts.

Cohen-Tannoudji -which was my first book on QM-, for example states that there's a problem with measurements but instead of discussing it, he only writes half a page of vague things like that the interaction between system and measurement apparatus is important. My second book was Sakurai whose only comment on the matter is a quote from Dirac's book reading "A measurement always causes the system to jump into an eigenstate". After having read these two standard textbooks and having attended 2-3 graduate level courses on QM, my impression of collapse was exactly one of the wrong notions which Ballentine disproofs.

So I don't think that Ballentine does a bad job here. Sure, he should have chosen a more modest wording and he should acknowledge that QM can be interpreted in different ways. But I don't think you have more misconceptions about interpretations after reading Ballentine than after reading the average standard textbook on QM. (And Ballentine's presentation of the physics is far above average.)

 

[carllooper]

Carllooper, I don't believe that physics CAN really function without philosophy. I know this will be a controversial and perhaps even offense statement, and I'm sorry for that!--but I don't mean by "philosophy" the stuff that is taught in the academy, but something like "first order principles that govern thought-towards truth" (sometimes called metaphysics, although widely rejected these days). I think when scientists appeal to the basic tools of thought they are simply appealing to these things. But if you get those principles wrong, the science will start to go astray, which is what seems to me to be happening. For example, if Einstein did not have a strong belief in the principle that nothing happens without a reason, would he have made the discoveries that he did? And if not, and given the widespread rejection of that very principle today….well, that's why I think these things are important.

Any philosophical principles that Einstein may have used may very well have played an important, if not key role in his contribution to physics. Indeed, personally, I'm sure of it. But how can we conclude that if we "get those principles wrong" science will go astray? Bohr was equally inspired by philosophical principles, but a somewhat different set of principles, yet was perfectly able to make a contribution to physics.

The way in which physics generally works is that ideas are put to a physical test. If there's agreement between the physical test and the ideas that conceive it, then the idea is considered provisionally correct. Or useful. In other words, the idea could be philosophically right or wrong, but in terms of the physics it points out, (to the extent that it does) it wouldn't actually matter. What matters is whether it's physically so, (ie. physically wrong or provisionally correct). Not whether it's philosophically so.

Of course, in practice, it can get a lot more complicated than this.

C

 

[atyy]

Apparently you and atyy both consider Ballentine an opponent of Copenhagen.

I agree that Ballentine makes it seem like he is criticizing a mainstream view while he is actually arguing against collapse as a physical process. But disregarding this, most of his arguments are valuable and not present in many other standard texts.

Ballentine's book does not mention Copenhagen. However, Ballentine's earlier review is explicitly anti-Copenhagen, and the view opposed in his book is the strawman version of Copenhagen in his review.

 

[bhobba]

I agree that Ballentine makes it seem like he is criticizing a mainstream view while he is actually arguing against collapse as a physical process. But disregarding this, most of his arguments are valuable and not present in many other standard texts.

That's exactly what's going on.

I don't know why there has been this big discussion about it.

Thanks
Bill

 

[atyy]

That's exactly what's going on.

I don't know why there has been this big discussion about it.

To be honest, I have never imagined that Ballentine was arguing against collapse as a physical process till now - possibly because the view opposed in his book and review are the same, and the review labels it "Copenhagen". So if one is working in Copenhagen, and has already taken the classical quantum cut, and the wave function is taken as a tool to calculate the probabilities of experimental outcomes, then one can interpret the plain English statement that a pure state provides a "complete and exhaustive description of an individual system" as correct within Copenhagen. A pure state is "complete and exhaustive" in the sense that it is an extremal state - even very modern versions of Copenhagen such as Hardy's "Five Reasonable Axioms" identify pure states with extremal states - the difference being that the classical space is a simplex, but the quantum space is not. There is also no problem in Copenhagen if one takes a pure state to label an individual system, as long as one adds that the theory only predicts probabilities. It is completely a matter of taste whether the pure state labels an individual system or an ensemble - either way, there are no differences in predictions, and no deviations from current experiments.

Furthermore, the view Ballentine ends up with, if it is correct, is essentially some flavour of Copenhagen renamed. However, it is not a very coherent presentation of any flavour of Copenhagen. If Ballentine is proposing that the wave function is not necessarily real, then he has to take a classical/quantum cut to get observable results. If he does not take the classical/quantum cut, then he will get a wave function of the universe and all the problems associated with it. So let's say he takes a classical quantum cut, and the wave function is not necessarily real - in that case there is no problem with state reduction - yet he argues against state reduction, and claims to be able to derive it from unitary evolution alone. Ballentine's derivation of "effective" state reduction is flawed, because he has implicitly assumed that proper and improper mixtures are equivalent, which is implicitly assuming the very postulate he has rejected. If he had derived effective collapse from unitary evolution alone, he would have made decoherence (without additional assumptions) a plausible solution to the measurement problem! So in the end, it is completely unclear what interpretation Ballentine wants. In his review, he secretly wanted hidden variables (nothing wrong with that, but one should say so explicitly if that is the case). In his book that seems to be corrected, and it seems he wants some flavour of Copenhagen, since the wave function is not necessarily real. But he also secretly wants Many-Worlds, since he objects to state reduction and wants unitary evolution only.

 

[vanhees71]

Opponents of Copenhagen are either crackpots or Ballentine (and Ballentine is wrong).

I still don't get, why the minimal statistical interpretation is wrong, because it's a subset of Copenhagen. One should say which flavor of Copenhagen anyway; e.g., I can well agree with all flavors without a collapse. An example is Bohr, who stressed that QT is a description about our objective (possible) knowledge about a quantum system and that the preparation procedures and measurement apparati select what we observe. There is no clear explicit statement concerning the collapse question (anyway, Bohr is usually not very explicit and clear, but that's another story). What I disagree with is the statement that there is a separate quantum dynamics and classical dynamics and there is a "cut" between these to realms. So far quantum theory has been seen as the most comprehense model, and the classical behavior of macroscopic systems is rather well understood from quantum many-body theory, where one derives transport equations, hydrodynamics, etc. from the full quantum Kadanoff-Baym equations via some coarse-graining formalism like the gradient expansion.

What I also don't buy is the collapse hypothesis, which only makes problems rather than explaining anything. So what's left as a physical theory is the usual postulates about quantum kinematics and dynamics + the Born postulate, and that's the minimal statistical interpretation. A quantum state refers to a single system being operationally defined as a (equivalence class) of preparation procedures but leads only to probabilistic knowledge about the outcome of further measurements which can empirically validated only using a large enough ensemble and statistical analysis. That's all what's need to use QT as the most successful physical theory, i.e., to map the formalism to observable objective facts about Nature, and that's all what physics is about.

Ontological or other philosophical questions beyond this is the problem of philosophers not of physicists! I don't see in which respect Ballentine's ensemble point of view is wrong (perhaps there are details he got wrong in his book, but I still think it's the best book on interpretations I've seen yet; also the book by Peres and Weinberg's new QM book are very good too).

 

[atyy]

I still don't get, why the minimal statistical interpretation is wrong, because it's a subset of Copenhagen. One should say which flavor of Copenhagen anyway; e.g., I can well agree with all flavors without a collapse. An example is Bohr, who stressed that QT is a description about our objective (possible) knowledge about a quantum system and that the preparation procedures and measurement apparati select what we observe. There is no clear explicit statement concerning the collapse question (anyway, Bohr is usually not very explicit and clear, but that's another story).

Yes, the minimal statistical interpretation is at best a flavour of Copenhagen. What I don't like about Ballentine's work is that he explicitly opposes Copenhagen in his review article, but at best the Copenhagen he opposes is such a caricature and misunderstanding of Copenhagen. In his book he doesn't identify his opponent as Copenhagen, but the interpretation he opposes is the one he identifies as Copenhagen in his review.

What I disagree with is the statement that there is a separate quantum dynamics and classical dynamics and there is a "cut" between these to realms. So far quantum theory has been seen as the most comprehense model, and the classical behavior of macroscopic systems is rather well understood from quantum many-body theory, where one derives transport equations, hydrodynamics, etc. from the full quantum Kadanoff-Baym equations via some coarse-graining formalism like the gradient expansion.

Copenhagen does not assign specific dynamics to the classical realm. The term "classical" refers to the the fact that we get a particular or definite experimental outcome on any single run of an experiment. This terminology goes back at least to the English translation of Landau and Lifshitz, and is standard in the literature. For example, http://www.quantiki.org/wiki/Channel_(CP_map): "Any device taking classical or quantum systems of a certain type as input and (possibly different) classical or quantum systems as output is a channel.", "Measurements are channels with classical range", "Preparations are channels with classical domain".

What I also don't buy is the collapse hypothesis, which only makes problems rather than explaining anything. So what's left as a physical theory is the usual postulates about quantum kinematics and dynamics + the Born postulate, and that's the minimal statistical interpretation. A quantum state refers to a single system being operationally defined as a (equivalence class) of preparation procedures but leads only to probabilistic knowledge about the outcome of further measurements which can empirically validated only using a large enough ensemble and statistical analysis. That's all what's need to use QT as the most successful physical theory, i.e., to map the formalism to observable objective facts about Nature, and that's all what physics is about.

In quantum mechanics one can use measurement as a method of state preparation. Collapse in quantum mechanics is a tool to describe the relationship between preparation and the preceding measurement used as a preparation procedure.

Ontological or other philosophical questions beyond this is the problem of philosophers not of physicists! I don't see in which respect Ballentine's ensemble point of view is wrong (perhaps there are details he got wrong in his book, but I still think it's the best book on interpretations I've seen yet; also the book by Peres and Weinberg's new QM book are very good too).

Ballentine is wrong in the following:
1. Opposing Copenhagen or characterizing Copenhagen with a caricature.
2. Claiming that pure states are not in some sense the "complete" information about an individual system. Within Copenhagen, after a classical/quantum cut is taken, pure states are the most "complete" information in the sense that they are extremal points of the space of density operators. One can identify pure states with individual systems or ensembles, and no mistake is made as long as one adds that the theory only predicts probabilities.
3. Claiming that Copenhagen cannot predict the results of Stern-Gerlach experiments correctly. In Copenhagen, if a measurement is made, there is a classical apparatus producing a definite result, or at least a quantum ancilla on which a measurement is later made. Ballentine is mssing the ancilla in his caricature of Copenhagen's version of Stern-Gerlach. So Copenhagen gets it right, Ballentine's caricature of Copenhagen gets it wrong. A correct version of the Stern-Gerlach with ancilla is shown in Zurek's http://arxiv.org/abs/quant-ph/0306072.
4. Claiming that one can do away with state reduction, because effective state reduction can be derived from decoherence without any additional assumptions. Here he has made the error of assuming that proper and improper mixtures are equivalent which is equivalent to assuming state reduction as Haroche and Raimond explain in https://www.amazon.com/dp/0198509146/?tag=pfamazon01-20.
5. If he is presenting some version of Copenhagen, then the wave function is already just a tool, and there is no problem with state reduction. His objection to state reduction makes more sense if he is considering the wave function as real, as in Many-Worlds.

 

[vanhees71]

Only because something is in Landau Lifshitz (whose volume 3 is, however, excellent) it needs not to be true, indeed. An there is no physically valid definition or any empirical hint for a cut, and as far as I know, it was Bohr, who emphasized the necessity for classical dynamics of a measurement apparatus. So that's part of the Copenhagen interpretation. Anyway, that's just semantics.

Further the collapse postulate is almost never fulfilled in practice. Usually a quantum system gets destroyed when it's measured but not prepared in a new state. Let's discuss a concrete example, where both aspects of measurements, gaining information about a system and destroying it and using the measurement for the preparation of a system for further measurements.

Let's consider a photon from an entangled state gets absorbed when it is registered by a photo plate or CCD or however you measure it (call the experimentalist Alice). If you have let Alice's photon go through an ideal polarizer, then the 2nd photon (measured by Bob before or after Alice's measurement) is in a definite polarization state (if Alice's photon was measured to be horizontally polarized in the direction determined by the polarizer, then Bob's is usually vertically polarized when the original entangled biphoton was in the singlet state). You can take this as a preparation procedure for a definite polarization state of Bob's single photon through measurement of the other photon in an entangled pair, if Bob does his polarization measurement after Alice did hers, but already at this point you come into trouble with causality if Bob does his measurement before Alice. This is, what I call a preparation procedure (however, a very tricky one). No classical dynamics is needed. Going throuh a polarization filter and detction of the one photon is perhaps difficult to describe in quantum-theoretical detail, but there's no hint that you need any classical dynamics to describe the meausrement/preparation procedure. So you need no cut.

What is, in my opinion, really unsolved with regard to ontology within quantum dynamics, is how Bob's photon gets into the pure polarization state. Here comes the place, where I think one has to invoke something like a collapse, when insisting on an ontological interpretation of quantum states, and this is highly problematic, because then Bob's photon must instantly go somehow into the definite pure polarization state, at Bob's place, when I measure the one photon due to the entanglement the other photon's state which may be registered at the same or a previous time (or at any space or timelike distance in spacetime). Note however, that the photon is not well localized. So the collapse, if it's a real process in an ontological interpretation of quantum states, place over long distances instantaneously, which violates relativistic causality.

In my opinion, this issue can only be solved by giving up an ontological interpretation and thus to give up the collapse postulate. Then I'm right at the ensemble interpretation, which does not associate an ontological meaning to states but taking them as the precise description of our knowledge about the quantum under investigation, which is probabilistic, i.e., takes an epistemical interpretation of quantum states. Then the association of the pure state after the measurement of the first photon is just due to the gain of our knowledge about the polarization of the measured photon and the knowledge that both photons were previously prepared in that entangled Bell state. This knowledge, however I can only gain when I take notice about the polarization state of the first photon. For this I need to communicate the measured polarization state of Alice's photon to Bob to take it as a preparation process for Bob's photon, before he has measured its polarization. This communication of information can take place at most with the speed of light, and thus Bob's measurement to validate the polarization state can only occur after (i.e., at a time like distance) to Alice's measurement. No problems with causality occur. That the correlation of polarizations is 100% for the photons in the entangled state is due to the very first preparation process, when the pair was created. Thus the epistemological interpretation of the ensemble interpretation saves the causality structure of quantum theory, which of course is at the heart of constructing all the most successful relativistic quantum theories in terms of microcausal local quantum field theories.

However, many people insist on an ontological interpretation of the notion of states in physics, and then you are left with this unsolved problem of preparations of states. For physics, which is just about the description of nature not about finding out ontological meaning of our knowledge, this is quite irrelevant, and thus it's a metaphysical problem of philophers :-).

 

[atyy]

Only because something is in Landau Lifshitz (whose volume 3 is, however, excellent) it needs not to be true, indeed. An there is no physically valid definition or any empirical hint for a cut, and as far as I know, it was Bohr, who emphasized the necessity for classical dynamics of a measurement apparatus. So that's part of the Copenhagen interpretation. Anyway, that's just semantics.

Just to be clear, my criticism of Ballentine is that he is wrong when he says (1) Copenhagen is wrong (2) textbook quantum mechanics with state reduction is wrong. Yes, Copenhagen-type/instrumental/operational interpretations have a cut, Landau and Lifshitz have a cut, and Weinberg's description of Copenhagen has a cut. Copenhagen-type interpretations do have a problem, but Ballentine has neither identified it not solved it.

Further the collapse postulate is almost never fulfilled in practice. Usually a quantum system gets destroyed when it's measured but not prepared in a new state. Let's discuss a concrete example, where both aspects of measurements, gaining information about a system and destroying it and using the measurement for the preparation of a system for further measurements.

Let's consider a photon from an entangled state gets absorbed when it is registered by a photo plate or CCD or however you measure it (call the experimentalist Alice). If you have let Alice's photon go through an ideal polarizer, then the 2nd photon (measured by Bob before or after Alice's measurement) is in a definite polarization state (if Alice's photon was measured to be horizontally polarized in the direction determined by the polarizer, then Bob's is usually vertically polarized when the original entangled biphoton was in the singlet state). You can take this as a preparation procedure for a definite polarization state of Bob's single photon through measurement of the other photon in an entangled pair, if Bob does his polarization measurement after Alice did hers, but already at this point you come into trouble with causality if Bob does his measurement before Alice. This is, what I call a preparation procedure (however, a very tricky one). No classical dynamics is needed. Going throuh a polarization filter and detction of the one photon is perhaps difficult to describe in quantum-theoretical detail, but there's no hint that you need any classical dynamics to describe the meausrement/preparation procedure. So you need no cut.

What you are saying is we can take the measurement apparatus to be quantum, and use only unitary evolution, and no additional assumptions, and recover definite outcomes. If you can show this, you would have shown that decoherence solves the measurement problem. Decoherence does not solve the measurement problem, unless additional assumptions are introduced.

What is, in my opinion, really unsolved with regard to ontology within quantum dynamics, is how Bob's photon gets into the pure polarization state. Here comes the place, where I think one has to invoke something like a collapse, when insisting on an ontological interpretation of quantum states, and this is highly problematic, because then Bob's photon must instantly go somehow into the definite pure polarization state, at Bob's place, when I measure the one photon due to the entanglement the other photon's state which may be registered at the same or a previous time (or at any space or timelike distance in spacetime). Note however, that the photon is not well localized. So the collapse, if it's a real process in an ontological interpretation of quantum states, place over long distances instantaneously, which violates relativistic causality.

In my opinion, this issue can only be solved by giving up an ontological interpretation and thus to give up the collapse postulate. Then I'm right at the ensemble interpretation, which does not associate an ontological meaning to states but taking them as the precise description of our knowledge about the quantum under investigation, which is probabilistic, i.e., takes an epistemical interpretation of quantum states. Then the association of the pure state after the measurement of the first photon is just due to the gain of our knowledge about the polarization of the measured photon and the knowledge that both photons were previously prepared in that entangled Bell state. This knowledge, however I can only gain when I take notice about the polarization state of the first photon. For this I need to communicate the measured polarization state of Alice's photon to Bob to take it as a preparation process for Bob's photon, before he has measured its polarization. This communication of information can take place at most with the speed of light, and thus Bob's measurement to validate the polarization state can only occur after (i.e., at a time like distance) to Alice's measurement. No problems with causality occur. That the correlation of polarizations is 100% for the photons in the entangled state is due to the very first preparation process, when the pair was created. Thus the epistemological interpretation of the ensemble interpretation saves the causality structure of quantum theory, which of course is at the heart of constructing all the most successful relativistic quantum theories in terms of microcausal local quantum field theories.

However, many people insist on an ontological interpretation of the notion of states in physics, and then you are left with this unsolved problem of preparations of states. For physics, which is just about the description of nature not about finding out ontological meaning of our knowledge, this is quite irrelevant, and thus it's a metaphysical problem of philophers :).

It's actually the other way round. If states are not ontological or "physically real and existing", in comparison to measurement outcomes, then we can have collapse. Major approaches that treat the wave function as real such as Bohmian Mechanics or Many-Worlds do not have collapse. Also, if you now treat the wave function as not ontological, you have introduced some form of Heisenberg cut, because the wave function is not "out there" but "in your head" so to speak.

 

[vanhees71]

Just to be clear, my criticism of Ballentine is that he is wrong when he says (1) Copenhagen is wrong (2) textbook quantum mechanics with state reduction is wrong. Yes, Copenhagen-type/instrumental/operational interpretations have a cut, Landau and Lifshitz have a cut, and Weinberg's description of Copenhagen has a cut. Copenhagen-type interpretations do have a problem, but Ballentine has neither identified it not solved it.

I've to read Ballentine's book again, but I don't think he says that Copenhagen is wrong. Anyway, I think if you take the collapse as a real thing happening in the physical world it must contradict Einstein causality, and there is no hint of this being true whatsoever. To the contrary, the successful relativistic QFTs are based on the very assumption that this is not the case: It's local and microcausal. So there is no instantaneous interaction at a distance by construction. Of course, the caveat is that so far nobody could prove the mathematical consistency of this scheme for physically relevant theories.

What you are saying is we can take the measurement apparatus to be quantum, and use only unitary evolution, and no additional assumptions, and recover definite outcomes. If you can show this, you would have shown that decoherence solves the measurement problem. Decoherence does not solve the measurement problem, unless additional assumptions are introduced.

I can't show this, but I don't see, where one finds it necessary to introduce a non-quantum dynamics, because the classical behavior of macroscopic objects, including measurement apparati are well-understood as approximations of quantum dynamics.

It's actually the other way round. If states are not ontological or "physically real and existing", in comparison to measurement outcomes, then we can have collapse. Major approaches that treat the wave function as real such as Bohmian Mechanics or Many-Worlds do not have collapse. Also, if you now treat the wave function as not ontological, you have introduced the classical/quantum cut, because the wave function is not "out there" but "in your head" so to speak.

I don't know, whether Bohm is consistent with relativity yet. Within non-relativistic theory there is no problem with a collapse, because instantaneous interactions are no contradiction with basic principles. So there is no need for non-local dynamics on top of quantum theory, as long as it doesn't provide additional observational consequences, and this is not the case as far as I know. So this additional Bohmian orbits are simply superfluous and can be cut away with Occam's razor. The same is true for unobservable parallel universes in the many-world interpretation. It simply doesn't help to solve the problems with quantum theory.

The minimal interpretation just states the facts and makes the necessary connections between observations in the real world and the formalism provided by quantum theory. I don't claim that it solves the measurement problem but it's a consistent scheme to use quantum theory to describe the outcome of real experiments.

 

[Fredrik]

Yes, the minimal statistical interpretation is at best a flavour of Copenhagen. What I don't like about Ballentine's work is that he explicitly opposes Copenhagen in his review article, but at best the Copenhagen he opposes is such a caricature and misunderstanding of Copenhagen. In his book he doesn't identify his opponent as Copenhagen, but the interpretation he opposes is the one he identifies as Copenhagen in his review.

I'm puzzled by this comment. You're implying that Ballentine is opposing a bad idea that's based on misunderstandings. So what's the problem? Is it really that he used the term "Copenhagen" to describe this bad idea 40 years earlier? Surely that's not a problem with his book?

It's interesting that you're labeling Ballentine's views as Copenhagen, and the views he's opposing as "not Copenhagen". I would do that too. The view he's opposing is essentially the MWI with the many worlds removed by magic. The view he's advocating is essentially just the idea that the state can be used to assign probabilities to possible results of experiments. I don't even consider that an interpretation. It's just QM. But if I had to call it something, I'd call it "Copenhagen". The term "the minimal statistical interpretation" has gained some popularity at PF, but I haven't heard it elsewhere. I don't want to use a term that includes "statistical interpretation", because that term should refer to the 1970 version, in which particles are assumed to have well-defined positions even when their wavefunctions are spread out. Is it the removal of that assumption that makes it "minimal"? Who defined it that way?

Oh, and I don't really want to call it "Copenhagen" either, because that term has completely lost its meaning.

Ballentine is wrong in the following:
1. Opposing Copenhagen or characterizing Copenhagen with a caricature.

So he's using the wrong terminology...in an article that came out more than 40 years ago. And he's not using that terminology in the book. I don't see the problem. Also, there's no right way to use the term "Copenhagen" anymore, because its meaning has been changed by people who misunderstood Bohr in different ways.

2. Claiming that pure states are not in some sense the "complete" information about an individual system. Within Copenhagen, after a classical/quantum cut is taken, pure states are the most "complete" information in the sense that they are extremal points of the space of density operators. One can identify pure states with individual systems or ensembles, and no mistake is made as long as one adds that the theory only predicts probabilities.

This argument doesn't make sense. What Ballentine is opposing is the view that a state represents is all the "actual properties" of the system. You can't argue that he's wrong by saying that pure states are extreme points in a convex set. That isn't just a weak argument. The argument isn't even related to the thing you're applying it to.

3. Claiming that Copenhagen cannot predict the results of Stern-Gerlach experiments correctly.

I haven't read this recently, so I can't comment at this time.

4. Claiming that one can do away with state reduction, because effective state reduction can be derived from decoherence without any additional assumptions. Here he has made the error of assuming that proper and improper mixtures are equivalent which is equivalent to assuming state reduction as Haroche and Raimond explain in https://www.amazon.com/dp/0198509146/?tag=pfamazon01-20.

The table of contents offers no clues about where in the book they have done this. That's OK though. Not sure I would want to take the time to read it. I expect that if they prove a claim like that, then their version of "reduction" or "collapse" isn't something that I would describe in those terms.

5. If he is presenting some version of Copenhagen, then the wave function is already just a tool, and there is no problem with state reduction. His objection to state reduction makes more sense if he is considering the wave function as real, as in Many-Worlds.

I don't remember what he said about this. Aren't the negative comments about state reduction part of his criticism of the view that the wave function is "real"?

 

[Nugatory]

There's no reason to not to let this thread run as long as the current participants are enjoying it... but I do think that we may have left OP and his question somewhere behind us?

 

[atyy]

I've to read Ballentine's book again, but I don't think he says that Copenhagen is wrong. Anyway, I think if you take the collapse as a real thing happening in the physical world it must contradict Einstein causality, and there is no hint of this being true whatsoever. To the contrary, the successful relativistic QFTs are based on the very assumption that this is not the case: It's local and microcausal. So there is no instantaneous interaction at a distance by construction. Of course, the caveat is that so far nobody could prove the mathematical consistency of this scheme for physically relevant theories.

In Ballentine's book he doesn't explicitly say that Copenhagen is wrong. However, in his 1970 review he does oppose Copenhagen, and the caricature of Copenhagen opposed in his review is the same view he opposes in his book. I have no problems with an Ensemble/minimal statistical interpretation if one says it is a flavour of Copenhagen, and Copenhagen is basically right.

I can't show this, but I don't see, where one finds it necessary to introduce a non-quantum dynamics, because the classical behavior of macroscopic objects, including measurement apparati are well-understood as approximations of quantum dynamics.

There are two properties of the classical world: (1) definite measurement outcomes (2) specific classical dynamics. The classical limit of quantum mechanics can get (2), so that is not the problem. But quantum mechanics itself seems not to be able to get (1), which is why there is the widely acknowledged measurement problem. Within Copenhagen, the subjective cut is introduced to get (1), in Bohmian mechanics hidden variables are introduced to get (1), and in MWI all possible outcomes are introduced to get (1).

I don't know, whether Bohm is consistent with relativity yet. Within non-relativistic theory there is no problem with a collapse, because instantaneous interactions are no contradiction with basic principles. So there is no need for non-local dynamics on top of quantum theory, as long as it doesn't provide additional observational consequences, and this is not the case as far as I know. So this additional Bohmian orbits are simply superfluous and can be cut away with Occam's razor. The same is true for unobservable parallel universes in the many-world interpretation. It simply doesn't help to solve the problems with quantum theory.

The minimal interpretation just states the facts and makes the necessary connections between observations in the real world and the formalism provided by quantum theory. I don't claim that it solves the measurement problem but it's a consistent scheme to use quantum theory to describe the outcome of real experiments.

Yes, it is not really collapse that Bohmian mechanics and other interpretations are introduced to solve. It is the problem of definite measurement outcomes. I agree that it is still a matter for research whether there is a satisfactory Bohmian mechanics for relativistic quantum field theory. However, if the problem of lattice chiral fermions interacting with non-abelian gauge fields can be solved, then there will be a lattice standard model (let's take Hamiltonian lattice theory), which for any finite lattice spacing will not be relativistic, in which case there should probably be a Bohmian standard model also.

Anyway, as I have said, I have no problems with the minimal interpretation (with a cut, and with collapse) taken be a flavour of Copenhagen. Basically once the wave function is not necessarily ontological, there is a subjective cut, and there is no problem with collapse since that is (at least partly) updating our knowledge.

 

[carllooper]

If asked how a wave function becomes a particle detection one might very well say: "I don't know, but one way to describe it is that the wave function collapses".

I don't think Copenhagen introduces the concept as anything much more than this. It certainly doesn't point back to any physical mechanism that anyone had elaborated in QM at the time of Copenhagen

And Einstein will get this. He'll understand by this that QM is incomplete, rather than incorrect.

For Bohr it's obvious that if there is any reality behind the measurements it's not going to be one that fits any classical conception of reality. Einstein (and others) will pursue a somewhat different course. If Bohr introduces the cut it's because about the only thing that remains classically real are the particle detections. One can assign them an unambiguous location in space (and the same would apply if one used relativistic space-time for this as well).

Bohr cleverly anchors the whole thing in terms of that which is the least problematic. Remember it's at a time when relativity is still quite new, and QM even newer. The audience for QM is classically trained. In other words Bohr anchors QM in terms which everyone can, at the very least, agree on. And that's really smart.

But what other way can it be done anyway?

If one starts with the concept of a wave function and tries to derive a particle detection, it doesn't quite work. But does it need to work that way anyway? To the extent that any proposed mechanism might yield new physics that would be just fine. It would be good for physics. But until then it's only good for science fiction (ie. philosophy). Of course nothing wrong with that.

C

 

[atyy]

I don't remember what he said about this. Aren't the negative comments about state reduction part of his criticism of the view that the wave function is "real"?

I read the rest of your post as well, but basically the problem is we not only have multiple interpretations of QM, we have multiple interpretations of Ballentine! I have never read Ballentine as opposing the wave function as "physically real" until this thread where that interpretation of Ballentine is mentioned by bhobba, kith and you. I have always read Ballentine as opposing a Copenhagen-style interpretation which one can also call instrumental/operational/orthodox/textbook/shut-up-and-calculate, which is certainly the impression conveyed by his review. I understood the distinctive elements he is opposing to be the assignment within these Copenhagen-style interpretations of a pure state as the most complete possible description of a single system, and state reduction. Although the book does not call what he is opposing "Copenhagen", it is the same view he earlier called "Copenhagen", and in both the review and the book he makes it seem that the view he is opposing is a mainstream view. The only mainstream view with state reduction I know is the Copenhagen-style interpretation, while two major approaches that treat the wave function as "physically real" are Bohmian Mechanics and MWI, neither of which have state reduction as fundamental.

The comment that assuming proper and improper mixtures to be equivalent is as good as assuming state reduction is found in section 2.5.4 "Decoherence Models versus the Copenhagen Interpretation" in Haroche and Raimond's https://www.amazon.com/dp/0198509146/?tag=pfamazon01-20. As far as I can tell, Ballentine believes he can get effective state reduction from unitary evolution alone. Ballentine also opposed the explanation of the quantum Zeno effect based on state reduction, ie. he opposed Sudarshan and Misra, and Wineland and colleagues. I think it is fair to assume that when Sudarshan and Misra and Wineland and colleagues use state reduction, they are using a Copenhagen-style interpretation, not some non-mainstream "the wave function is "real" and "really" collapses" view of quantum mechanics. So I basically don't think Ballentine is teaching "standard quantum mechanics".