Quantum theory - Nature Paper 18 Sept

• I
Earlier this morning I came across this article -- https://phys.org/news/2018-09-errors-quantum-world.html -- which is about an open for all article in Nature Communications -- https://www.nature.com/articles/s41467-018-05739-8

Quantum theory provides an extremely accurate description of fundamental processes in physics. It thus seems likely that the theory is applicable beyond the, mostly microscopic, domain in which it has been tested experimentally. Here, we propose a Gedankenexperiment to investigate the question whether quantum theory can, in principle, have universal validity. The idea is that, if the answer was yes, it must be possible to employ quantum theory to model complex systems that include agents who are themselves using quantum theory. Analysing the experiment under this presumption, we find that one agent, upon observing a particular measurement outcome, must conclude that another agent has predicted the opposite outcome with certainty. The agents’ conclusions, although all derived within quantum theory, are thus inconsistent. This indicates that quantum theory cannot be extrapolated to complex systems, at least not in a straightforward manner.
I've yet to fully work through the paper, but if this thread is appropriate I would love to read other people's views on it.

atyy

Related Quantum Interpretations and Foundations News on Phys.org
PeterDonis
Mentor
2020 Award
These authors have a number of previous papers on this general topic; here is a typical one from 2016:

https://arxiv.org/abs/1604.07422

What appears to me to be new about this latest paper is that the claimed inconsistency is not limited to single-world interpretations, as it was in previous papers like the one linked to above. The difference now is only that different interpretations have the inconsistency show up in different places (depending on which assumptions they make).

bhobba and vanhees71
the claimed inconsistency is not limited to single-world interpretations, as it was in previous papers like the one linked to above. The difference now is only that different interpretations have the inconsistency show up in different places (depending on which assumptions they make).
Where do you see that? In the Nature article the main result seems to be "Theorem 1. Any theory that satisfies assumptions (Q), (C), and (S) yields contradictory statements when applied to the Gedankenexperiment of Box 1."

(S) is the single-world assumption

Michael Price and Demystifier
DarMM
Gold Member
It seems to be their 2016 paper tidied up and with a discussion about the various interpretations modified to have some responses from the proponents of those interpretations taken into account. I'll have a more careful look later. I think most see it as an inconsistency in uninterpreted QM, i.e. you have to reject Q, C or S or else you have a inconsistent framework.

atyy, bhobba and Demystifier
Demystifier
Gold Member
The paper does not claim that QM is inconsistent. It claims that what is inconsistent is any interpretation of QM in which all 3 assumptions Q, C and S are satisfied. And then it explains that most known interpretations are consistent by violating at least one of those assumptions. The question that seems unanswered is the following: Is there any known interpretation of QM that satisfies all 3 assumptions and is therefore inconsistent? As far as I can see, the paper does not identify any such inconsistent interpretation.

bohm2, atyy, bhobba and 2 others
DarMM
Gold Member
The question that seems unanswered is the following: Is there any known interpretation of QM that satisfies all 3 assumptions and is therefore inconsistent? As far as I can see, the paper does not identify any such inconsistent interpretation.
There was some discussion about certain orthodox versions of Copenhagen (as opposed to unorthodox like QBism) trying to hold all three conditions. See Matt Leifer's talk here:
39:44

Earlier in the same lecture he breaks down the "types" of Copenhagen into Objective and Perspectival.

However see the paper by Bub that argues that Orthodox (Objective) Copenhagen is fine with this result:
https://arxiv.org/abs/1804.03267

Truecrimson, atyy, kith and 1 other person
DarMM
Gold Member
Okay, I've read it and compared it to the previous paper. Small differences:
1. They simplify the experiment (less steps), but this is a trivial detail
2. The discussion on interpretations is longer with more references to how the interpretation responds to the theorem
3. They include a nice discussion of the original Wigner's friend
4. The point out how their set up reduces to Wigner's friend and Deustch's modification of it
5. (Biggest difference) The original paper derived the result within a framework they called "The story framework", how physical theories define narratives about results. This Nature paper simply presents it from the perspective of normal quantum mechanics. This was the hardest part of the original paper in my opinion. The assumptions Q, C, S now have a shorter forms stated purely in terms of textbook QM.
Overall one should just read this version.

Last edited:
dextercioby, atyy, Deepblu and 3 others
stevendaryl
Staff Emeritus
The axiom that mainstream interpretations (variants of Copenhagen) must reject (even though they don't explicitly reject it) is C, which I would paraphrase as follows:

If Alice is 100% certain that Bob is 100% certain that X is true, then Alice should also be 100% certain that X is true.

where being 100% certain of something means that the Born probability for something is 1, or that it's derived from other 100% certain facts using the various assumptions.

I couldn't completely follow the argument of the thought experiment, but my rough impression is that under certain circumstances, you can have a situation where Alice, treating Bob as a quantum mechanical system, calculates that there is a 100% certainty that Bob is in a state of being 100% certain that X is true, although Alice knows that X is false.

atyy
vanhees71
Gold Member
The axiom that mainstream interpretations (variants of Copenhagen) must reject (even though they don't explicitly reject it) is C, which I would paraphrase as follows:

If Alice is 100% certain that Bob is 100% certain that X is true, then Alice should also be 100% certain that X is true.
This is not logically sound at all. If Alice knows that Bob is certain about X this doesn't imply that she can be certain about X too. That's often forgotten in discussions about the foundations.

Put it in physical terms. Say, we have our beloved two-photon Bell state again in the beginning, i.e., there are two photons in the polarization state
$$|\Psi \rangle=\frac{1}{\sqrt{2}} (|HV \rangle - |VH \rangle).$$
If now Bob measures the polarization of his photon and tells Alice that he measured it, Alice still describes her photon's state as
$$\hat{\rho}_{A}=\frac{1}{2} \hat{1}$$
She cannot be certain about Bob's outcome, before Bob has told her.

where being 100% certain of something means that the Born probability for something is 1, or that it's derived from other 100% certain facts using the various assumptions.

I couldn't completely follow the argument of the thought experiment, but my rough impression is that under certain circumstances, you can have a situation where Alice, treating Bob as a quantum mechanical system, calculates that there is a 100% certainty that Bob is in a state of being 100% certain that X is true, although Alice knows that X is false.
I've still to read the Nature Communications paper, though. Since it has been discussed in the community for a while, I'm sure, it's not as simple a (unfortunately quite) common misunderstanding of QT as suggested in #8.

bhobba and dextercioby
Andy Resnick
The paper does not claim that QM is inconsistent. It claims that what is inconsistent is any interpretation of QM in which all 3 assumptions Q, C and S are satisfied. And then it explains that most known interpretations are consistent by violating at least one of those assumptions. The question that seems unanswered is the following: Is there any known interpretation of QM that satisfies all 3 assumptions and is therefore inconsistent? As far as I can see, the paper does not identify any such inconsistent interpretation.
I admit that I don't fully understand the paper- the quantitative details of the gedankenexperiment are (IMO) cryptic and hard to parse. Tables 1 and 2 gave me a headache, for example....

However, the main point of their argument seems to concern a situation where multiple independent observers each have (partially overlapping) incomplete information about a pure state; when the observers compare their predictions about what the pure state is, they have irreconcilable differences. Is that a reasonable summary?

If that's a correct summary, then I'm not sure there's any paradox: Gibbs' paradox is a related problem with a well-understood solution: whether or not we agree that 2 populations of objects mix depends on if we agree that the groups of objects are the same or different.

That said, I appreciate their Table 4 and related discussion.

stevendaryl
Staff Emeritus
Put it in physical terms. Say, we have our beloved two-photon Bell state again in the beginning, i.e., there are two photons in the polarization state $$|\Psi \rangle=\frac{1}{\sqrt{2}} (|HV \rangle - |VH \rangle).$$ If now Bob measures the polarization of his photon and tells Alice that he measured it, Alice still describes her photon's state as $$\hat{\rho}_{A}=\frac{1}{2} \hat{1}$$

She cannot be certain about Bob's outcome, before Bob has told her.
I'm talking about the case in which Alice knows what it is that Bob is 100% certain of.

What the paper means by rejecting C is the possibility of the following sort of situation:

Alice says: "I'm 100% certain that Bob is 100% certain that his photon is vertically polarized. But his photon is actually horizontally polarized."

Last edited:
vanhees71
Gold Member
I still have to read the paper...

stevendaryl
Staff Emeritus
Okay, so I think I finally understand the argument. It's not presented in the clearest way, in my opinion.

So there are two different systems that are being treated as quantum-mechanical systems.

1. System ##\overline{F}##, has states ##|\overline{h}\rangle## and ##|\overline{t}\rangle##. For the thought experiment, we consider a few special superpositions:
• ##|\overline{ok}\rangle = \frac{1}{\sqrt{2}} (|\overline{h}\rangle - |\overline{t}\rangle)##
• ##|\overline{fail}\rangle = \frac{1}{\sqrt{2}} (|\overline{h}\rangle + |\overline{t}\rangle)##
• ##|init\rangle = \frac{1}{\sqrt{3}} (|\overline{h}\rangle + \sqrt{2} |\overline{t}\rangle)##
2. System ##F##, which has states ##|\frac{+1}{2}\rangle## and ##|\frac{-1}{2}\rangle##. For the thought experiment, we consider the alternative basis:
• ##|ok\rangle = \frac{1}{\sqrt{2}} (|\frac{+1}{2}\rangle - |\frac{-1}{2} \rangle)##
• ##|fail\rangle = \frac{1}{\sqrt{2}} (|\frac{+1}{2}\rangle + |\frac{-1}{2} \rangle)##
The interaction between systems is such that:
• If System ##\overline{F}## is in state ##|\overline{h}\rangle##, then it puts System ##F## into the state ##|\frac{-1}{2}\rangle##
• If System ##\overline{F}## is in state ##|\overline{t}\rangle##, then it puts System ##F## into the state ##|fail\rangle##
We set up ##\overline{F}## so that it is initially in state ##|init\rangle##.

The system evolves according to the interaction rule above, and the linearity of the evolution equations into the state:

##|final\rangle = \frac{1}{\sqrt{3}} |\overline{h}\rangle |\frac{-1}{2}\rangle + \frac{1}{\sqrt{3}} |\overline{t}\rangle |\frac{+1}{2}\rangle + \frac{1}{\sqrt{3}} |\overline{t}\rangle |\frac{-1}{2}\rangle##

For the sake of the reasoning that follows, I'm going to write this in three different ways:
1. ##|final\rangle = \sqrt{\frac{1}{3}} |\overline{h}\rangle |\frac{-1}{2}\rangle + \sqrt{\frac{2}{3}} |\overline{t}\rangle |fail\rangle##
2. ##|final\rangle = \sqrt{\frac{2}{3}} |\overline{fail}\rangle |\frac{-1}{2}\rangle + \sqrt{\frac{1}{3}} |\overline{t}\rangle |\frac{+1}{2}\rangle##
3. ##|final\rangle = \sqrt{\frac{3}{4}} |\overline{fail}\rangle |fail\rangle - \sqrt{\frac{1}{12}} |\overline{fail}\rangle |ok\rangle - \sqrt{\frac{1}{12}} |\overline{ok}\rangle |fail\rangle - \sqrt{\frac{1}{12}} |\overline{ok}\rangle |ok\rangle##
The twist in this thought experiment is that systems ##\overline{F}## and ##F## contain observers who can also reason using quantum-mechanics. So we have four observers: ##\overline{W}, \overline{F}, W, F## (I hope it's not confusing to use the same name for system ##F## and observer ##F## and the same name for system ##\overline{F}## and observer ##\overline{F}##). They are measuring different things:
• ##F## is measuring whether his system is in state ##|\frac{+1}{2}\rangle## or ##|\frac{-1}{2}\rangle##
• ##\overline{F}## is measuring whether his system is in state ##|\overline{h}\rangle## or ##|\overline{t}\rangle##
• ##W## is measuring whether system ##F## is in state ##|ok\rangle## or ##|fail\rangle##
• ##\overline{W}## is measuring whether system ##\overline{F}## is in state ##|\overline{ok}\rangle## or ##|\overline{fail}\rangle##
Let's write down what each observer can reason about the others, based on their observations:
1. If ##F## measures +1/2, then it means that it is impossible that ##\overline{F}## got result ##\overline{h}##. That's because there is no overlap between the final state and ##|\overline{h}\rangle |\frac{+1}{2}\rangle##. So ##F## concludes that if he got +1/2, ##\overline{F}## must have gotten ##\overline{t}##
2. If ##\overline{W}## measures ##\overline{ok}##, then it means that is impossible that ##F## got -1/2. That's because there is no overlap between the final state and ##|\overline{ok}\rangle |\frac{-1}{2}\rangle##. So ##\overline{W}## concludes that if he got ##\overline{ok}## then ##F## must have gotten +1/2.
3. If ##\overline{F}## gets ##\overline{t}##, then it is impossible that ##W## got ##ok##. That's because there is no overlap between the final state and ##|\overline{t}\rangle |ok\rangle##. So ##\overline{F}## concludes that if he got ##\overline{t}##, then ##W## got ##fail##.
Now, what happens if ##\overline{W}## gets ##\overline{ok}##?

From 2 above, ##\overline{W}## concludes that ##F## got +1/2.
From 1 above, it follows that ##F## concludes that ##\overline{F}## got ##\overline{t}##
From 3 above, it follows that ##\overline{F}## concludes that ##W## got ##fail##

So in this case, ##\overline{W}## is certain that ##F## is certain that ##\overline{F}## is certain that ##W## got ##fail##. So if we adopt the inference rule:

Rule C: If agent A is certain (according to the rules of quantum mechanics) that agent B is certain (according to the rules of quantum mechanics) that fact X is true, then agent A should be certain that X is true

then it follows:

If ##\overline{W}## gets ##\overline{ok}##, then he should be certain that ##W## gets ##fail##.

But actually, there is a 1/12 chance that ##\overline{W}## gets ##\overline{ok}## and that ##W## gets ##ok##. This just follows from the fact that the overlap between ##|final\rangle## and ##|\overline{ok}\rangle |ok\rangle## is ##- \frac{1}{\sqrt{12}}## leading to a probability of 1/12.

Last edited:
msumm21, eloheim, dextercioby and 5 others
atyy
There was some discussion about certain orthodox versions of Copenhagen (as opposed to unorthodox like QBism) trying to hold all three conditions. See Matt Leifer's talk here:
It's an interesting to talk, but it has not much to do with most "orthodox" Copenhagen-type interpretations, since he defines the "objective" Copenhagen interpretation not to be instrumentalism (ie. does not regard measurement as fundamental), and also to assume that hidden variables don't exist. In contrast, the Copenhagen-type interpretation in Landau and Lifshitz clearly treats measurement as fundamental, and Messiah goes with Copenhagen while not ruling out hidden variables. Also, one just has to say in his words that "Alice's outcomes exist for Alice (but not for Bob), and Bob's outcomes exist for Bob (but not for Alice)", which is already part of the orthodox Copenhagen-type interpretation, eg. Wiseman's and Cavalcanti say 'strict operationalists do indeed operate "by denying independent real situations as such", as Einstein subsequently allowed.'

Last edited:
atyy
Is the Frauchiger and Renner result related to an earlier result by Hays and Peres about conditions needed for quantum and classical descriptions of a measuring apparatus to agree?

https://arxiv.org/abs/quant-ph/9712044
Quantum and classical descriptions of a measuring apparatus
Ori Hay, Asher Peres
A measuring apparatus is described by quantum mechanics while it interacts with the quantum system under observation, and then it must be given a classical description so that the result of the measurement appears as objective reality. Alternatively, the apparatus may always be treated by quantum mechanics, and be measured by a second apparatus which has such a dual description. This article examines whether these two different descriptions are mutually consistent. It is shown that if the dynamical variable used in the first apparatus is represented by an operator of the Weyl-Wigner type (for example, if it is a linear coordinate), then the conversion from quantum to classical terminology does not affect the final result. However, if the first apparatus encodes the measurement in a different type of operator (e.g., the phase operator), the two methods of calculation may give different results.

DarMM
Gold Member
It's an interesting to talk, but it has not much to do with most "orthodox" Copenhagen-type interpretations, since he defines the "objective" Copenhagen interpretation not to be instrumentalism (ie. does not regard measurement as fundamental), and also to assume that hidden variables don't exist.
I think it's very hard to know what Copenhagen is at this point. Many interpretations call themselves Copenhagen despite being quite unlike Bohr's views. He's using the classification given by Adán Cabello in his paper here:
https://arxiv.org/abs/1509.04711

It seems to be the typical view in Quantum Foundations currently.

From my own understanding, I thought Bohr didn't viewed measurements as fundamental ontologically, just that they must necessarily be treated classically due to human nature, Heisenberg was similar.

Bohr thought that the "stuff" under QM did exist but was ineffable/indescribable and thus there were no hidden variables, as there was no mathematical description. Again, this is my reading of his papers on the subject.

Last edited:
dextercioby
stevendaryl
Staff Emeritus
There is one part of quantum orthodoxy that I think is actually wrong, although it doesn't usually get anyone into trouble. The "soft" contradiction described in this paper exploits this part of the orthodoxy, and I think that the argument falls through without it.

The orthodox interpretation of the Born rules is that any Hermitian operator corresponds to an observable, and measuring any observable results in an eigenvalue of the corresponding operator. In practical terms, though, many Hermitian operators are not actually measurable.

I can illustrate the absurdity of "any operator is measurable" using Schrodinger's dead cat. Suppose you have a dead cat, and you want to bring it back to life. Then let me define two cat states: ##|A\rangle = \frac{1}{\sqrt{2}} |Dead\rangle + \frac{1}{\sqrt{2}} |Alive\rangle## and ##|B\rangle = \frac{1}{\sqrt{2}} |Dead\rangle - \frac{1}{\sqrt{2}} |Alive\rangle##. Let the AB-ness operator be some operator ##AB## such that ##AB |A\rangle = +|A\rangle## and ##AB |B\rangle = - |B\rangle##. If I perform a measurement of the AB-ness of a dead cat, I'll either get +1 or -1 with 50/50 probability. Afterward, the cat will be in the state ##|A\rangle## or ##|B\rangle## depending on whether I got ##\pm 1##. Regardless of what I got, I can now check whether the cat is alive or dead, and I'll find it alive (the state ##|Alive\rangle##) with 50% probability. If I do this procedure enough time, I will eventually get a live cat.

What's wrong with this story (well, there are many things wrong with it, but...) is that you can't actually measure the AB-ness of a cat.

kith and mfb
Demystifier
Gold Member
There is one part of quantum orthodoxy that I think is actually wrong, although it doesn't usually get anyone into trouble. The "soft" contradiction described in this paper exploits this part of the orthodoxy, and I think that the argument falls through without it.

The orthodox interpretation of the Born rules is that any Hermitian operator corresponds to an observable, and measuring any observable results in an eigenvalue of the corresponding operator. In practical terms, though, many Hermitian operators are not actually measurable.

I can illustrate the absurdity of "any operator is measurable" using Schrodinger's dead cat. Suppose you have a dead cat, and you want to bring it back to life. Then let me define two cat states: ##|A\rangle = \frac{1}{\sqrt{2}} |Dead\rangle + \frac{1}{\sqrt{2}} |Alive\rangle## and ##|B\rangle = \frac{1}{\sqrt{2}} |Dead\rangle - \frac{1}{\sqrt{2}} |Alive\rangle##. Let the AB-ness operator be some operator ##AB## such that ##AB |A\rangle = +|A\rangle## and ##AB |B\rangle = - |B\rangle##. If I perform a measurement of the AB-ness of a dead cat, I'll either get +1 or -1 with 50/50 probability. Afterward, the cat will be in the state ##|A\rangle## or ##|B\rangle## depending on whether I got ##\pm 1##. Regardless of what I got, I can now check whether the cat is alive or dead, and I'll find it alive (the state ##|Alive\rangle##) with 50% probability. If I do this procedure enough time, I will eventually get a live cat.

What's wrong with this story (well, there are many things wrong with it, but...) is that you can't actually measure the AB-ness of a cat.
Would there be anything wrong with that if "dead" and "alive" were replaced by "spin up" and "spin down"? I guess not. So one could still say that any Hermitian operator of a microscopic system corresponds to an observable. But then there is no precise border between microscopic and macroscopic, so one should conclude that your example is, in principle, possible even with the macroscopic cat. And it should not be any more strange than the fact that physical laws do not, in principle, forbid to transform a dead cat into an alive one. Of course, it's impossible in practice due to the statistical 2nd law of thermodynamics, but in principle it's possible.

eloheim and dextercioby
stevendaryl
Staff Emeritus
Would there be anything wrong with that if "dead" and "alive" were replaced by "spin up" and "spin down"?
I actually think that spin isn't measurable, either. The way that you measure spin is that you set things up so that a particle with spin-up will go to the left, and a particle with spin-down will go to the right. You still haven't measured the spin. But now, you crash the electron into a photographic plate, and it makes a dark spot. Now, you either see a dark spot on the left, or on the right, and you deduce that the particle must have been spin-up or spin-down. But what you're actually measuring isn't spin, but the presence/absence of dark spots.

So it's not really Hermitian operators that are measurable. It's macroscopic facts about the world. You can try to be clever and set up an interaction between systems so that this or that value of the operator leads to this or that macroscopic fact about your measurement device. But there is no guarantee that you'll be able to do that for all possible operators.

I guess not. So one could still say that any Hermitian operator of a microscopic system corresponds to an observable. But then there is no precise border between microscopic and macroscopic, so one should conclude that your example is, in principle, possible even with the macroscopic cat. And it should not be any more strange than the fact that physical laws do not, in principle, forbid to transform a dead cat into an alive one. Of course, it's impossible in practice due to the statistical 2nd law of thermodynamics, but in principle it's possible.
I sort of agree. It's a FAPP (for all practical purposes) impossibility. So that's what makes the contradiction described in this paper a "soft" contradiction. You can't actually perform an experiment to resolve the contradiction.

Some paradoxes in physics suggest experiments to resolve them. If two different lines of reasoning lead to two different predicted outcomes, then you can set up an experiment to see which one is right. But if the contradiction would require a FAPP impossible experiment, then it stays in the realm of philosophy, rather than physics.

dextercioby, mfb and Mentz114
I sort of agree. It's a FAPP (for all practical purposes) impossibility. So that's what makes the contradiction described in this paper a "soft" contradiction. You can't actually perform an experiment to resolve the contradiction.
In the paper what they propose is to run the experiment on a quantum computer, with the experimenters implemented as agents that are simple enough to simulate. If you have a large enough quantum computer then the experiment should be practical, so it's only FAPP impossible if we never build large quantum computers.

Demystifier
Gold Member
But if the contradiction would require a FAPP impossible experiment, then it stays in the realm of philosophy, rather than physics.
I disagree. Even if you cannot do the experiment with true cats, you can do experiments with larger and larger objects. For instance, interference experiments have been performed not only with electrons and photons, but even with large molecules containing several hundreds of atoms. And there are even serious proposals that it could be done with viruses. So, if you cannot resolve a contradiction by doing experiment with cats, you can do an analogous experiment with something smaller. Whatever results of such experiments might be, it is scientifically justified to extrapolate conclusions and argue that essentially the same can be said about cats.

For an example of that kind of reasoning, see my https://lanl.arxiv.org/abs/1406.3221

mfb
Mentor
Only if you keep the objects coherent. Anything that would remotely resemble an intelligent system will lead to decoherence.

Demystifier