Doubts regarding my interpretation of Wigner’s Friend Scenario

[Marco211298]

TL;DR Summary

In this thread, i want to see if my understanding of the Wigner’s friend scenario can be correct. I also ask what kind of mainstream interpretation does suit to it the most

I have some doubts regarding my personal interpretation that i was contemplating about in the context of Wigner's friend experiment (also tested in the laboratory recently). Could it be that a system is always in a superposition, and when we perform a measurement, we obtain a definite value due to the interaction, but after it, the system returns to a superposition? For Wigner, who will check if superposition exists for him after his friend's measurement, he will find the system again in a superposition. If they were to measure at the same time, they would see the same definite result instead. Wigner and his friend might have different measurements, but this wouldn't imply a different reality, only their knowledge of it. What kind of “mainstream” intrepretation does this resemble the most? Please note that i am not a physicist and i do not intend to propose any theory or interpretation because i certantly don’t have the knowledge required. I am just expressing my curiosity.

 

[PeroK]

TL;DR Summary: In this thread, i want to see if my understanding of the Wigner’s friend scenario can be correct. I also ask what kind of mainstream interpretation does suit to it the most

I have some doubts regarding my personal interpretation that i was contemplating about in the context of Wigner's friend experiment (also tested in the laboratory recently). Could it be that a system is always in a superposition, and when we perform a measurement, we obtain a definite value due to the interaction, but after it, the system returns to a superposition? For Wigner, who will check if superposition exists for him after his friend's measurement, he will find the system again in a superposition. If they were to measure at the same time, they would see the same definite result instead. Wigner and his friend might have different measurements, but this wouldn't imply a different reality, only their knowledge of it. What kind of “mainstream” intrepretation does this resemble the most? Please note that i am not a physicist and i do not intend to propose any theory or interpretation because i certantly don’t have the knowledge required. I am just expressing my curiosity.

After a measurement, the state of a system is in the appropriate eigenstate of the observable that was measured. If that state is not an eigenstate of the Hamiltonian, then the state will immediately evolve into a superposition of eigenstates of the observable. This is generally the case in QM. Wigner's Friend, however, assumes that the measurement was recorded on some macroscopic measuring device. That device shows a definite number, and not a supersposition of two numbers!

To Wigner's friend, however, the measuring device and the original quantum system should be in a superposition of the various possible measurement outcomes. Whether the system that was measured has evolved is not the issue. The issue is what the macroscopic measuring device shows.

There's a discussion here:

https://en.wikipedia.org/wiki/Wigner's_friend

 

[PeterDonis]

Wigner's friend experiment (also tested in the laboratory recently)

Unfortunately, since the claim you describe has indeed been made, the recent experiments you refer to do not test anything like the actual Wigner's friend scenario. They test a scenario in which qubits are "measured" but then the "measurement" can be undone (which is why I put "measurement" in scare-quotes--actually what is being done is just garden variety quantum computing-type reversible operations). They do not test a scenario where a human being (the friend) makes a measurement but then another human being (Wigner) makes a measurement on the first human being that effectively undoes the first measurement. Human beings are not qubits. They are not even remotely close to being qubits. The experiments actually done do not show anything useful about quantum mechanics and human beings.

Could it be that a system is always in a superposition, and when we perform a measurement, we obtain a definite value due to the interaction, but after it, the system returns to a superposition?

No.

 

[vanhees71]

Which experiments are you referring to? It's too vague to seriously discuss your question in a scientific way.

 

[Marco211298]

Unfortunately, since the claim you describe has indeed been made, the recent experiments you refer to do not test anything like the actual Wigner's friend scenario. They test a scenario in which qubits are "measured" but then the "measurement" can be undone (which is why I put "measurement" in scare-quotes--actually what is being done is just garden variety quantum computing-type reversible operations). They do not test a scenario where a human being (the friend) makes a measurement but then another human being (Wigner) makes a measurement on the first human being that effectively undoes the first measurement. Human beings are not qubits. They are not even remotely close to being qubits. The experiments actually done do not show anything useful about quantum mechanics and human beings.No.

I see, but what bothers me is that everything is quantum, including ourselves. Heisenberg did mention that measurement is essentially interaction. So, why isn't a qubit a satisfactory model for an observer since it measures through interaction? I am aware that it doesn't collapse the wave function, and I recall reading that the collapse was introduced somewhat ad hoc to the mathematics to make predictions. There are interpretations that do not postulate a wave function collapse.

 

[Marco211298]

Which experiments are you referring to? It's too vague to seriously discuss your question in a scientific way.

Sorry, this is the experiment i was referring to https://www.nature.com/articles/s41567-020-0990-x

 

[PeterDonis]

what bothers me is that everything is quantum, including ourselves.

We don't actually know this; we have no direct evidence of quantum behavior of macroscopic objects. For example, nobody has ever done a double slit experiment with humans, or even with rocks. (IIRC the largest objects that have had a double slit experiment done are buckyballs, C₆₀. Other experiments, such as with the Josephson effect, have shown quantum effects with larger numbers of atoms, about a trillion IIRC, but that is still at least 12 or 13 orders of magnitude smaller than, say, a 1 kg rock.)

why isn't a qubit a satisfactory model for an observer since it measures through interaction?

No, a qubit does not "measure through interaction". The actual measurements in quantum experiments are done by macroscopic measuring devices, not qubits. Qubits are the measured systems, but that doesn't mean the qubits themselves are measuring devices. They aren't.

I am aware that it doesn't collapse the wave function, and I recall reading that the collapse was introduced somewhat ad hoc to the mathematics to make predictions. There are interpretations that do not postulate a wave function collapse.

Yes, there are, and collapse is irrelevant to the issues you are asking about here. Decoherence is a sufficient condition for "measurement" for this discussion.

 

[Nugatory]

So, why isn't a qubit a satisfactory model for an observer since it measures through interaction?

It's the difference between a single isolated qubit (naturally described as a pure state in a two-dimensional Hilbert space, arguably the simplest non-trivial case of all) and an incoherent collection of $10^{25}$ or so particles each with more degrees of freedom than a qubit and continuously interacting with one another.

There's an analogy from classical physics: An ideal gas is made up of a very large number of particles, each one of which is satisfactorily modeled as a point mass moving under the influence of Newton's laws. But put a few moles of these particles together in a container and we need a completely different model involving concepts of pressure, volume, and temperature. The two models are related, but the relationship is seriously non-trivial; it takes a semester of college-level statistical mechanics to see how the bulk behavior of gases can emerge from the simple Newtonian behavior of the individual particles.

The analogous bridge between single-qbit behavior and the behavior of macroscopic multi-particle systems is decoherence (google for "quantum decoherence") - and it had not been discovered back when Wigner and Schrodinger proposed their paradoxes. If they had known about it they would have been far less troubled by the discrepancy between quantum behavior and macroscopic. (There's still plenty of other stuff to be troubled by, which is why we need a Foundations/Interpretations subforum).

A good layman-friendly introduction to the subject is David Lindley's book "Where does the weirdness go?"

 

[vanhees71]

We don't actually know this; we have no direct evidence of quantum behavior of macroscopic objects. For example, nobody has ever done a double slit experiment with humans, or even with rocks. (IIRC the largest objects that have had a double slit experiment done are buckyballs, C₆₀. Other experiments, such as with the Josephson effect, have shown quantum effects with larger numbers of atoms, about a trillion IIRC, but that is still at least 12 or 13 orders of magnitude smaller than, say, a 1 kg rock.)

I'd put it in the other way: We have no evidence that QT has any limit concerning the system size. On the other hand there's no classical model consistent with the atomistic structure of matter and its stability. That's resolved only by QT.

Today we have evidence for quantum behavior of really macroscopic objects, e.g., the LIGO mirrors, where thanks to the advanced quantum technology (squeezed and most recently also frequency-dependent squeezing) the quantum fluctuations of the corresponding harmonic oscillators can be resolved. The mirrors weigh around 40kg!

No, a qubit does not "measure through interaction". The actual measurements in quantum experiments are done by macroscopic measuring devices, not qubits. Qubits are the measured systems, but that doesn't mean the qubits themselves are measuring devices. They aren't.Yes, there are, and collapse is irrelevant to the issues you are asking about here. Decoherence is a sufficient condition for "measurement" for this discussion.

Whether or not the collapse postulate, as a FAPP description, holds, depends on the details of the measurement apparatus. It's usually not possible to realize such ideal von Neumann filter measurements. It's of course true that a measurement always involves decoherence and irreversibility, i.e., an irreversible storage of the measurement result. This part of Bohr's philosophy is indeed correct.

 

[vanhees71]

It's the difference between a single isolated qubit (naturally described as a pure state in a two-dimensional Hilbert space, arguably the simplest non-trivial case of all) and an incoherent collection of $10^{25}$ or so particles each with more degrees of freedom than a qubit and continuously interacting with one another.

Of course, a single qubit does not need to be necessarily prepared in a pure state. E.g., if you have two maximally entangled qubits (preparation in a Bell state) the single qubits in the pair are maximally indetermined, i.e., in the maximum-entropy mixed state $\hat{\rho}=\hat{1}/2$.

There's an analogy from classical physics: An ideal gas is made up of a very large number of particles, each one of which is satisfactorily modeled as a point mass moving under the influence of Newton's laws. But put a few moles of these particles together in a container and we need a completely different model involving concepts of pressure, volume, and temperature. The two models are related, but the relationship is seriously non-trivial; it takes a semester of college-level statistical mechanics to see how the bulk behavior of gases can emerge from the simple Newtonian behavior of the individual particles.

That's classical statistical physics. Of course the same arguments hold true within QT, and there you get to the more comprehensive description by quantum statistics.

The analogous bridge between single-qbit behavior and the behavior of macroscopic multi-particle systems is decoherence (google for "quantum decoherence") - and it had not been discovered back when Wigner and Schrodinger proposed their paradoxes. If they had known about it they would have been far less troubled by the discrepancy between quantum behavior and macroscopic. (There's still plenty of other stuff to be troubled by, which is why we need a Foundations/Interpretations subforum).

I'm not so sure. I think Schrödinger was well aware about quantum statistical mechanics, but still skeptical concerning the probabilistic interpretation on a fundamental level. I don't know what happened with Wigner in his later years. Some of his writeups border on esoterics. He seems to have followed a kind of "Princeton interpretation" a la von Neumann, according to which a conscious being is necessary to define the quantum-classical cut a la Heisenberg... That's of course far in the realm of "interpretation" and thus speculation.

A good layman-friendly introduction to the subject is David Lindley's book "Where does the weirdness go?"

There's only "weirdness" if you insist on the validity of classical concepts, which are incompatible with QT, particularly if you don't accept that there's "objective randomness" in Nature's behavior.

 

[PeterDonis]

I'd put it in the other way: We have no evidence that QT has any limit concerning the system size.

But that's the wrong way to put it, because we do have evidence that macroscopic objects behave classically and do not display effects like quantum interference, at least not with any experiments we can do now or in the foreseeable future. So the burden of proof is not on us to show why QT does have a limit; we already have evidence that suggest that. If QT wants to claim to be a complete theory valid on all scales, then QT needs to show positive evidence that it does apply to macroscopic objects.

 

[PeterDonis]

Today we have evidence for quantum behavior of really macroscopic objects, e.g., the LIGO mirrors, where thanks to the advanced quantum technology (squeezed and most recently also frequency-dependent squeezing) the quantum fluctuations of the corresponding harmonic oscillators can be resolved. The mirrors weigh around 40kg!

Do you have a reference for this? I was not aware that anyone was claming that we have detected quantum fluctuations of the LIGO mirrors.

 

[Marco211298]

Is Schrödinger's cat a problem of how we define identity? If we consider that a cat is composed of numerous atomic particles, defining particles in superposition presents no issue. A cat is a human construct to represent a grouping of atoms, and notions of life or death emerge from atomic behavior. . For instance, a single-celled bacterium is alive, but if you separate the macromolecules that combined to create the bacterium, these units are not alive. Since life is an emergent property, there's no contradiction in the cat being alive or not. Everything, including ourselves, is quantum. Consider a house made of bricks—when organized, it's a house, but they remain bricks. This underscores that observers, cats, and what we perceive as "classical" are emergent properties of an underlying quantum and elementary world. We label a certain group of atoms as living, as in the case of a live cat. If we consider everything as quantum, it seems that paradoxes like Schrodinger’s cat and Wigner's friend dissolve.

 

[PeterDonis]

If we consider everything as quantum, it seems that paradoxes like Schrodinger’s cat and Wigner's friend dissolve.

No, you have it backwards. The whole point of the Schrodinger's cat and Wigner's friend scenarios is that trying to consider everything as quantum appears to lead to absurdities.

With Schrodinger's cat, the issue is not superposition by itself, but an entangled superposition of the cat and the rest of the apparatus, which includes both dead and alive states of the cat. Yes, "dead" and "alive" are, on the "everything is quantum" view, just emergent properties of the underlying quantum states (or subspaces each containing a large number of quantum states), but they're emergent properties of different orthogonal states (or subspaces), and they are macroscopically distinguishable, so having a superposition of them makes no sense.

With Wigner's friend, the issue is that if "everything is quantum", measurements that supposedly irreversibly record a result, actually can't--there will be some quantum operation that can in principle be performed that undoes the measurement and "un-records" the result. But the whole framework of QM is built on the assumption that measurement results are irreversibly recorded. So "un-recording" a measurement result makes no sense.

In other words, taking the view that macroscopic properties are emergent from the underlying quantum state does not solve these problems--it creates them.

 

[Marco211298]

What if superposition is never truly destroyed even in macroscopic objects but is simply undetectable? How can we be certain that the alive and dead states are macroscopically distinguishable?

 

[PeterDonis]

What if superposition is never truly destroyed even in macroscopic objects

Who said it was "destroyed"? What interpretation are you using?

but is simply undetectable?

If it's undetectable, it might as well not exist.

How can we be certain that the alive and dead states are macroscopically distinguishable?

Um, because we define "alive" and "dead" that way?

 

[vanhees71]

But that's the wrong way to put it, because we do have evidence that macroscopic objects behave classically and do not display effects like quantum interference, at least not with any experiments we can do now or in the foreseeable future. So the burden of proof is not on us to show why QT does have a limit; we already have evidence that suggest that. If QT wants to claim to be a complete theory valid on all scales, then QT needs to show positive evidence that it does apply to macroscopic objects.

The "classical behavior" of macroscopic objects refers to coarse-grained macroscopic observables and can be derived from quantum many-body theory. As stressed before there is quantum behavior of macroscopic objects observed, e.g., the LIGO mirrors (these are around 40 kg objects, which I'd clearly accept as "macroscopic"). There is no objective classical-quantum cut in theory and no observation hinting at a failure of quantum theory. It's not said that in the future one may find something like this, but so far there's no hint. That's all I'm saying.

 

[gentzen]

There is no objective classical-quantum cut in theory and no observation hinting at a failure of quantum theory. It's not said that in the future one may find something like this, but so far there's no hint. That's all I'm saying.

But there is a subjective observation-quantum cut, which is part of the Copenhagen interpretation of quantum theory.

I don't know what happened with Wigner in his later years. Some of his writeups border on esoterics. He seems to have followed a kind of "Princeton interpretation" a la von Neumann, according to which a conscious being is necessary to define the quantum-classical cut a la Heisenberg... That's of course far in the realm of "interpretation" and thus speculation.

I doubt that von Neumann assigned a special role to conscious beings. Why? Because all references for that claim ultimately reference back to his book "Mathematische Grundlagen der Quantenmechanik" (1932), and the word conscious simply does not appear in his book. The words "the so-called principle of the psycho-physical parallelism" ("Prinzip vom psychophysikalischen Parallelismus") are the closest to the concept of consciousness to be found in his book. (It is possible that Henry Stapp is the one responsible for this and other myths regarding the role of consciousness in QM and Copenhagen, at least if I extrapolate a bit from my research.) So even Luboš Motl was slightly wrong when he wrote:

Consciousness is just a "spiritually sounding" synonym of the ability of an observer to be aware of a measured value of an observable.

These views were associated with Wigner and von Neumann and their usage of the word "consciousness" made their comments provocative and famous in the broader scientific public.

But if you read an English translation of von Neumann's words, you see that Motl was not far-off (especially if you ignore the derogatory part):

The difference between these two processes $U \to U'$ is a very fundamental one: aside from the different behaviors in regard to the principle of causality, they are also different in that the former is (thermodynamically) reversible, while the latter is not (cf. V.3.).
Let us now compare these circumstances with those which actually exist in nature or in its observation. First, it is inherently entirely correct that the measurement or the related process of the subjective perception is a new entity relative to the physical environment and is not reducible to the latter. Indeed, subjective perception leads us into the intellectual inner life of the individual, which is extra-observational by its very nature (since it must be taken for granted by any conceivable observation or experiment). (Cf. the discussion above.) Nevertheless, it is a fundamental requirement of the scientific viewpoint -- the so-called principle of the psycho-physical parallelism -- that it must be possible so to describe the extra-physical process of the subjective perception as if it were in reality in the physical world -- i.e., to assign to its parts equivalent physical processes in the objective environment, in ordinary space. (Of course, in this correlating procedure there arises the frequent necessity of localizing some of these processes at points which lie within the portion of space occupied by our own bodies. But this does not alter the fact of their belonging to the "world about us," the objective environment referred to above.)

N. Bohr, Naturwiss. 17 (1929), was the first to point out that the dual description which is necessitated by the formalism of the quantum mechanical description of nature is fully justified by the physical nature of things, and that it may be connected with the principle of the psycho-physical parallelism.

There's only "weirdness" if you insist on the validity of classical concepts, which are incompatible with QT, particularly if you don't accept that there's "objective randomness" in Nature's behavior.

The interesting question is which part of the classical concepts you have to keep, which ones you have to discard completely, and which ones you have to modify or reinterpret. And making "phenomenological" sense of "objective randomness" would also be a good idea.

 

[PeterDonis]

As stressed before there is quantum behavior of macroscopic objects observed, e.g., the LIGO mirrors

Again, can you give a reference?

 

[renormalize]

Again, can you give a reference?

Perhaps @vanhees71 is referring to this? https://arxiv.org/pdf/2002.01519.pdf

 

[vanhees71]

Yes, the published article is here

https://doi.org/10.1038/s41586-020-2420-8

 

[gentzen]

And making "phenomenological" sense of "objective randomness" would also be a good idea.

Objective Probability​

Similar considerations apply to the interpretation of probability. The physical property in this case is the objective tendency or propensity of an individual system to do something at any given moment (even when it’s not interacting with any “measurement apparatus”).

Even so aleazk made some convincing points earlier against operational definitions for the interpretation of physical theories, somehow this propensity interpretation of probability seems to be missing precisely such operational clarifications regarding which properties are expected from such a propensity.