Insights Why the Quantum | A Response to Wheeler's 1986 Paper - Comments

[vanhees71]

Newton's equation is a statement of the form "position of the particle is such and such". It is not a statement of the form "When position of the particle is measured, then position of the particle is such and such".

But quantum mechanics is different. QM does not state that "probability of the position of the particle is such and such". It states that "When position of the particle is measured, then probability of the position of the particle is such and such".

The statement of QT is indeed not "When position of the particle is measured, then position of the particle is such and such" but "When the position of the particle is measured the probability distribution for the outcome of this measurement is given by Born's rule". Within QT there's nothing else known about the position than these probabilities. It seems as if we agree in fact about this. The difference only is that in classical mechanics it's postulated that all observables always have a determined value and of course that means that if you measure these observables you observe these determined values, while QT tells us that not all observables can take determined values, no matter in which state the system may be prepared, and thus the only sensible meaning of the quantum state are the probabilities given by Born's Rule.

 

[vanhees71]

I guess your interpretation is not an "interpretation" in traditional quantum terminology since it retains the cut and doesn't attempt to solve the measurement problem.

Especially since you frame it with Wheeler's question, which was not about solving the measurement problem, I guess your programme is more like trying to provide alternative axioms for quantum theory, like the odl quantum logic thinking of von Neumann, Birkhoff, Mackey, etc, and the more recent ones of Lucien Hardy https://arxiv.org/abs/quant-ph/0101012 or of Chiribella and colleagues https://arxiv.org/abs/1011.6451 ?

Well, I think there is no measurement problem for the simple reason that QT works extraordinary well to describe what's observed in Nature and measured with high precision in the lab.

Wow, it always amazes me how many human-IQ-hours have been invested trying to find a way to reinvent QM so as to rid it of the measurement problem (see Schlosshauer quote in #135). Accepting QM as supplying spatiotemporal constraints on the distribution of quantum events, rather than dynamical laws for the behavior of quantum systems, automatically rids us of the MP. Then, QM is seen as complete by simply accepting quantum-classical contextuality. There is nothing in Nature that demands we recover classical reality from a quantum reality in toto. Certainly not with any empirical consequences. That's just a reductive bias. If Weinberg tried and failed, it's certainly above my pay grade! But, it looks to be entertaining lots of brilliant mathematical minds, so by all means, enjoy :-)

I couldn't agree more!

 

[stevendaryl]

There is no physical content of any theory without measurements

I don't think that's true. Newton's laws have the same physical content even in the absence of human beings and measuring devices.

if not you can't compare it to experiment so its not testable ie its not science.

I agree with that. Science is about how we go about finding out what is true about the world---developing theories and testing them, etc. But that doesn't mean that the physical content of the theory is about measurements.

The issue with QM is defining, using nothing but QM itself, what a measurement its.

I don't think that is the issue. A measurement device can be said to measure a physical property of a system being measured if the interaction between measuring device and system leads to a macroscopic difference in the device such that different values of the physical property lead to persistent, observable differences in the state of the device. Or something like that. That is not the issue for quantum mechanics.

People sometimes talk about things like "the irreversible interaction between the system and an environment" as a kind of measurement. But it certainly isn't. Not in the sense of the rhetoric that "science is about measurement".

The issue for quantum mechanics is understanding how probabilities arise and how alternative possibilities become real.

It has long been my 'feeling' that some issues in QM people worry about are also present in other theories like probability and classical mechanics, however they are generally not worried about in those areas - except maybe by philosophers.

In classical physics, probability is not fundamental but a result of trying to deal with imperfect information. So it's very different from quantum mechanics in that respect.

 

[stevendaryl]

I thought that QT can make predictions without any cut. Say, if you smash these particles, then the probability to get those is so and so. No cut and a very spesific prediction. Or something along the lines a black hole will radiate and loose energy, no cut. Or is the cut somewhere implicit.

Quantum mechanics gives the amplitudes for particle interactions. To actually get probabilities out of them, you have to assume that somewhere down the road, someone is going to be detecting those particles.

 

[martinbn]

Quantum mechanics gives the amplitudes for particle interactions. To actually get probabilities out of them, you have to assume that somewhere down the road, someone is going to be detecting those particles.

How exactly in these examples? A photon will not decay to a proton. What detection is needed to make such a prediction? Just for the prediction, not to actually test it. Or black holes radiate, where is the detection here? I am not saying that it isn't there, it is just not obvious to me.

 

[stevendaryl]

How exactly in these examples? A photon will not decay to a proton. What detection is needed to make such a prediction? Just for the prediction, not to actually test it. Or black holes radiate, where is the detection here? I am not saying that it isn't there, it is just not obvious to me.

Okay. My remark is about the non-exceptional probabilities, that are neither 0 nor 1, but somewhere in between. I guess I would agree that if QM predicts that the probability amplitude for something is 0 or 1, then you don't need to know anything more than that.

 

[RUTA]

I guess your interpretation is not an "interpretation" in traditional quantum terminology since it retains the cut and doesn't attempt to solve the measurement problem.

Especially since you frame it with Wheeler's question, which was not about solving the measurement problem, I guess your programme is more like trying to provide alternative axioms for quantum theory, like the odl quantum logic thinking of von Neumann, Birkhoff, Mackey, etc, and the more recent ones of Lucien Hardy https://arxiv.org/abs/quant-ph/0101012 or of Chiribella and colleagues https://arxiv.org/abs/1011.6451 ?

No, no, our interpretation totally solves the MP and keeps QM intact. The MP obtains because physicists want a model of physical reality that is dynamical and reductive, but Nature is under no obligation to be the way we want it to be.

 

[atyy]

Okay. My remark is about the non-exceptional probabilities, that are neither 0 nor 1, but somewhere in between. I guess I would agree that if QM predicts that the probability amplitude for something is 0 or 1, then you don't need to know anything more than that.

One always needs a cut, even for probabilities of 0 or 1. If the observer is included in the wave function, there are no measurement outcomes and no probabilities, which come only when the Born rule is applied.

 

[atyy]

No, no, our interpretation totally solves the MP and keeps QM intact. The MP obtains because physicists want a model of physical reality that is dynamical and reductive, but Nature is under no obligation to be the way we want it to be.

But I thought you retain the classical-quantum cut?

 

[stevendaryl]

No, no, our interpretation totally solves the MP and keeps QM intact. The MP obtains because physicists want a model of physical reality that is dynamical and reductive, but Nature is under no obligation to be the way we want it to be.

To me, the measurement problem should really be called the probability problem. The equations of quantum mechanics describe probability amplitudes. To get a probability out of that, you have to pick a basis (or a projection operator or more generally, a positive-operator valued measure). Picking a basis or an operator is roughly speaking what the "cut" is about.

 

[RUTA]

But I thought you retain the classical-quantum cut?

Quantum-classical contextuality means there is a classical context for every quantum exchange of momentum. However, there is no “cut” as regards size. As Gerry et al. showed, you can reach the Tsirelson bound with large angular momentum if properly screened off. In other words, you’ll get an elephant interference pattern if you screen off elephants in the twin-slit experiment (as already done with 60-atom molecules). QM states don’t care about the size of the objects involved.

Maximal violations of a Bell inequality by entangled spin-coherent states
author = {Gerry, Christopher C. and Benmoussa, Adil and Hach, Edwin E. and Albert, Jaroslav},
journal = {Phys. Rev. A},
volume = {79},
issue = {2},
pages = {022111},
numpages = {4},
year = {2009},
month = {Feb},
publisher = {American Physical Society},
doi = {10.1103/PhysRevA.79.022111},
note = {\url{https://link.aps.org/doi/10.1103/PhysRevA.79.022111}}

 

[RUTA]

Keep in mind I’m not a “research physicist,” I just teach physics at a 4-year college. Therefore, I’m just trying to figure out a way to make sense of the physics we already have. The interpretation I’m presenting allows me to do that without instrumentalism, i.e., I’m providing a model of physical reality as Becker argued. It took me 24 years and now that I have it I must admit per Wheeler, “how could I have been so stupid for so long?” It’s actually pretty simple once you see it. But, I understand research physicists have a different agenda, so I don’t want to rain on your parade :-)

 

[Fra]

I don't think that's true. Newton's laws have the same physical content even in the absence of human beings and measuring devices.

I think a scientific theory can never be detached from its historical inference; because the explanation lies in its history. You can certainly do that mathematically, but then you loose contact to the experiment, and it even misses out the important theoretical aspect and meaning of interactions and evolution of the theory.Even Newtons laws are indeed inferred from actual interactions (measurements)! In this sense there is actually not much fo a difference.

The difference is that in classical mechanics we can afford to "trivialize" measurements in a way, that makes us think they arent part of interactions. But I think this is a deep mistake.

The typical rebuttal to this argument is that i am here confusing the laws of nature with the human knowledge of them (and human theories) but its not that simple unfortunately. There is a much deeper meaning of this.

/Fredrik

 

[Fra]

I meant the theory cannot make any predictions without a cut. If the observer is included in the wave function and all we have is the unitarily evolving quantum state, the theory makes no predictions.

I agree. And to include the observer, we need ANOTHER observer. And then what we get are the predictions of this OTHER observers measurements on the composite system.

But if we repeat this, we realize that at some point the predictions are ONLY at the infinite boundary of the composite system. Which is the same situation as saying that all we can predict are the S-matrix, we can not make predictions of the interior, we can only predict the interactions at the observer boundary. All this is fine for lab work, but becomes completely meaningless for cosmological theories. This is for me the real motivation for this discussion IMO.

/Fredrik

 

[Demystifier]

I meant the theory cannot make any predictions without a cut. If the observer is included in the wave function and all we have is the unitarily evolving quantum state, the theory makes no predictions.

So is the cut necessary to predict that electric charge will be conserved?

 

[Demystifier]

The difference only is that in classical mechanics it's postulated that all observables always have a determined value and of course that means that if you measure these observables you observe these determined values, while QT tells us that not all observables can take determined values, no matter in which state the system may be prepared, and thus the only sensible meaning of the quantum state are the probabilities given by Born's Rule.

Consider a classical stochastic process, e.g. a random walk. The observables are not deterministic, but given by a probabilistic law. Yet, observables have definite values at each time, irrespective of whether you measure them or not. I think the key to understand QM is to explain what exactly is the difference between QM and classical stochastic processes.

 

[atyy]

So is the cut necessary to predict that electric charge will be conserved?

Yes, because there is no physical interpretation to the wave function alone.

 

[vanhees71]

Consider a classical stochastic process, e.g. a random walk. The observables are not deterministic, but given by a probabilistic law. Yet, observables have definite values at each time, irrespective of whether you measure them or not. I think the key to understand QM is to explain what exactly is the difference between QM and classical stochastic processes.

Classical stochastic processes (as described, e.g., by Langevin equations) only occur, because we are not able to describe the detailed equations of motion of the classical system (e.g., the Brownian particle in a fluid consisting of very many classical particles, with which this Brownian particle interacts) and instead describe the interaction of a subsystem (e.g., the Brownian particles) in terms of a friction force and a randomly fluctuating force in the sense of classical statistical physics. The system as a whole is still deterministic, i.e., here we use probabilistic arguments to mimic the interactions which we cannot fully describe due to the complexity of the situation.

The probabilities in QT given by Born's rule are different: If you take QT as complete, then Nature is intrinsically random and not only by ingnoring too complicated deterministic dynamics, i.e., a particle doesn't take determined positions and momenta no matter how well I try to prepare the particle to have determined positions and momenta. Even worse, the Heisenberg-Robertson uncertainty relation tells us that if we attempt to make the position pretty well determined we must buy this at the prize that necessarily the momentum gets only very badly determined (and vice versa). Of course, I don't claim that QT is necessarily complete. We can't know, whether one day someone finds a deterministic non-local theory from which QT can be derived as a stochastic effective theory as with the above discussed example of a Brownian particle, where a stochastic theory is used to describe the effect of many unresolved degrees of freedom.

On the other hand classicality also follows via quantum theory. In my opinion this was beautifully already clarified very early in Mott's famous paper on the question, how $\alpha$ particles from a radioactive nucleus can leave classical straight tracks in a cloud chamber. Indeed, you can view this at many science exhibitions: An $\alpha$ particle's track can indeed be easily followed by eye, looking as if a classical particle moves with (almost constant) velocity in a straight line. Indeed, this behavior is nice explained in Mott's paper: The issue is that (a) we do not look too closely, i.e., we are satisfied with the finite resolution of the track in the cloud chamber and the corresponding finite resolution for the velocity (or momentum) of the particle you can get by measuring the speed by following the head of the track and (b) that the straight-line trajectory is the vastly most probable trajectory of the particle. Only the initial momentum of the $\alpha$ particle is of course completely random, i.e., the direction of the track is indeed random, as one can see by watching more and more nuclei emit their $\alpha$ particles. After a while you see a quite isotropic distribution of tracks. So the single $\alpha$ particle indeed seems to look classical when you take the track's head as the momentaneous position of the particle, but that's a very much coarse-grained macroscopic quantity, which seems to behave classically. In this way one sees that there is no contradiction in an apparently classical behavior of a particle when (constantly) "observed" through it's interaction with a macroscopic "measurement apparatus", which here is the vapour in the cloud chamber, and the observation of the track by a human being is completely irrelevant. You can as well film the whole thing and watch the movie much later, i.e., the track is indeed there, no matter whether a conscious being is watching it or not. So there's indeed no necessity for any solipsistic collapse arguments of some flavors of Copenhagen interpretations (like the Princeton interpretation).

 

[martinbn]

Yes, because there is no physical interpretation to the wave function alone.

Where would you put the cut in these examples?

 

[Boing3000]

Then please precisely explain to me what you mean when you say "measurements are special" (within quantum mechanics)

I mean precisely what is said in the minimal interpretation of QM. See post #127 or #142, and for that matter nearly ever other post in this thread.

I have no clue what that should mean if you admit that measurement devices are usual "stuff" and thus behaves according to the generally valid physical laws.

This is the salient point of your misunderstanding. My belief, or your belief, are inconsequential here. We are not discussing what you call philosophy (some taste based word salad). We are asserting the coherence of some statement, only on the merit of what is written, nothing more.
QM that you claim to appreciate for its coherence and simplicity (which is fine) does make the distinction between interaction and measurement. Why does it have to do that distinction, while the universe is obviously only make of the same "stuff" ? That's the measurement problem.

I'm arguing against this claim of the Copenhagen-like interpretation all the time.

OK, but then you argumentation must be based on a version off QM that does not contains the Born rule. Yours does.

All of physics is about phenomenology. Theory aims at ever more comprehensive and ever more precise description of phenomena that are objectively observable in Nature.

That is insufficient, the theory must also predict new/unknown phenomena, and this is very important. I suppose you'll agree that the future of physics is not to find zillion of equivalent abstract phenomenology (like string theory, ...)
Also maybe you don't think of quantum "field" to be "ontological" field. So far so good. But then by saying that QM is complete, you've just give up on positivism, by asserting that you cannot find anything better (no even bettering QM itself) without any shred of evidence.
Also to assert that nothing can be gained by trying to connect a phenomenology to some ontology (like strings) is also an anti-positivist claim. There is no evidence for that.

Again: QT is causal but not deterministic.

OK, given the time you have taken to explain, I'll use that word like you do. Although that I don't think it's the correct word to encompass the uncertainty relations, because I have always read that the Schrodinger equation is all there is, and even if every knowledge cannot be known "at once", all is continuous and unitary (and thus determined)

What do you mean by that? I don't use Born's rule inside the Schrödinger equation.

Really, all I wanted you is to admit the following and ...

For me Born's rule is an independent postulate, necessary to give an interpretation to the wave function (in this very special case of systems, where a wave function is a sufficient description of the (pure) quantum states of the system) usable in the lab.
{...}
So far the Born rule seems to be an independent postulate, necessary to give a minimal interpretation needed to apply the QT formalism to real-world observations.

...that this independent postulate is only for the lab, which need a special interpretation. There is no such thing in classical mechanics, which treat all the stuff in the universe equally.

Which "distinction"?

The same you've made above. We are on the same page now.

I have no clue, what you want to tell by this statements. The same units are used in QT as in classical physics.

No. That's not even the same field. And going to real by using a modulo is one thing, but the squaring implies that the unit of the Hilbert space is "square root" of probabilities...

Probabilities are of course numbers between 0 and 1. I've no clue, why you think probabilities might be complex numbers.

I don't, and that's exactly what I write ==> "As far as I known, probabilities are not complex numbers... even (0,0)"

 

[Fra]

Then please precisely explain to me what you mean when you say "measurements are special" (within quantum mechanics). I have no clue what that should mean if you admit that measurement devices are usual "stuff" and thus behaves according to the generally valid physical laws. Indeed, measuring a force with a balance invokes the very laws the concept of force is based on within the theory (necessarily Newtonian mechanics, because the force concept only makes sense within Newtonian mechanics). I'm not doubting that, but you do, if I understand the statement "measurements are special". I'm arguing against this claim of the Copenhagen-like interpretation all the time.

I think i repeat myself, but again I want to note that there is a circularity or chicken/egg situtuation here. And this is a key observation this is why i emphasise it.
The circulatory is also not of the circular reasoning kind, its of the evolutionary kind.

I agree with vanhees that there exists and equivalence between interactions and observations. But its also clear that current QM formalism, does not manifest this equivalence, except for the special case where the class of observers are only classical. Because the statistics are "objective" only (or at least at best) in the classical realm. But even there is non-trivial if we include the classical observers obeying also GR. If we stick to SR and particle physics in a lab if we ignore the problem of unification of forces other than GR which is still lacking.

Any well defined measurement sort of DEPENDS on the theory. It involves (depending on how you frame this) preparation of measurement devices, signal processing of the let's say "raw data" coming off an actual detector. All these things are the "baggage" that are essentially put in by hand as constraints in QM. None of it is "explained". In here lies the reference to the classical measurement device, it implicitly includes all these things. Without this "background" one can not define any definite expectations (probabilities) due to undefined references in the conditional probability.

But of equal importance, the origin of the current state of laws as we know them, are de facto a result of hundreds of years of human scientific work. So all the stuff we "put in by hand" are not as ad hoc as it seems. But if we take Vanhees equivalence of measurement and interactions seriously (and i do as well) then this must hold even for complex systems, such as physicists. So the observers abduction of laws in its own environment, must be a physical process and most probably a survival trait of any system. In here lies also the key to understand how symmetries are emergent as a result of evolving interacting systems. Symmetries are most liekely NOT god given constraints.

As i see it can be no other way. But our understanding of this process, ie the PHYSICAL process by which one system infers and encodes predictive rules about another system, in which this inference process, is the key to understand the interactions as well, is undeveloped.

/Fredrik

 

[vanhees71]

This is the salient point of your misunderstanding. My belief, or your belief, are inconsequential here. We are not discussing what you call philosophy (some taste based word salad). We are asserting the coherence of some statement, only on the merit of what is written, nothing more.
QM that you claim to appreciate for its coherence and simplicity (which is fine) does make the distinction between interaction and measurement. Why does it have to do that distinction, while the universe is obviously only make of the same "stuff" ? That's the measurement problem.

Again you just claim this, but that's not what standard QT claims. According to standard QT the functioning of measurement apparati are completely consistent with the laws of physics valid for all matter observed yet (usually one assumes also that that's how all matter in the universe behaves, but that's of course an extrapolation which never can be tested empircially). My point is that there is no measurement problem at all. What's called a measurement problem is usually a metaphysical quest for an ontological interpretation of the notion of quantum states, which is however not subject of science but only of philosophy.

 

[atyy]

Again you just claim this, but that's not what standard QT claims. According to standard QT the functioning of measurement apparati are completely consistent with the laws of physics valid for all matter observed yet (usually one assumes also that that's how all matter in the universe behaves, but that's of course an extrapolation which never can be tested empircially). My point is that there is no measurement problem at all. What's called a measurement problem is usually a metaphysical quest for an ontological interpretation of the notion of quantum states, which is however not subject of science but only of philosophy.

You are completely wrong - see Landau and Lifshitz.

 

[vanhees71]

Where in Landau Lifshitz is the claim you are not willing to state explicitly in this thread? It's strange that you always make these claims about measurement devices being outside of the standard laws of QT but never explaining in which sense you mean it and then point to textbooks and not giving the clear statement nor where to find it in these books. I've never heard such a statement nor read it in any serious textbook about QT, and LL for sure is one very serious textbook.

 

[atyy]

Where in Landau Lifshitz is the claim you are not willing to state explicitly in this thread? It's strange that you always make these claims about measurement devices being outside of the standard laws of QT but never explaining in which sense you mean it and then point to textbooks and not giving the clear statement nor where to find it in these books. I've never heard such a statement nor read it in any serious textbook about QT, and LL for sure is one very serious textbook.

page 3 in the 1991 reprint of the 1958 English translation

 

[vanhees71]

This must be the very first chapter, where exactly the contrary is stated to the claim you make. They say that classical behavior of macroscopic devices is understood as an approximative limit of QT. They also bring the very example of the cloud-chamber traces for an electron that I also stated in this thread. In other words, what I read in LL is completely consistent with what I stated. They are very careful and only use a very weak version of the collapse postulate, more careful than most other books following the Copenhagen doctrine. I guess that's because Landau is following more Bohr's than Heisenberg's opinion.

 

[stevendaryl]

As was said in a previous comment, a measurement device plays two different roles in the minimal interpretation:

1. It is a physical system, and so it is described by the Schrodinger equation.
2. It determines a "preferred basis" for computing probabilities.

In role #1, there is no distinction between measurement devices and any other physical system. In role #2, there is a big distinction. Probabilities only appear in quantum mechanics if you have measurements, not for any other interactions.

I don't understand why, when the issue is #2, people keep bringing up that #1 doesn't distinguish measurements from other interactions. That's true, but it's only half the story.

 

[atyy]

This must be the very first chapter, where exactly the contrary is stated to the claim you make. They say that classical behavior of macroscopic devices is understood as an approximative limit of QT. They also bring the very example of the cloud-chamber traces for an electron that I also stated in this thread. In other words, what I read in LL is completely consistent with what I stated. They are very careful and only use a very weak version of the collapse postulate, more careful than most other books following the Copenhagen doctrine. I guess that's because Landau is following more Bohr's than Heisenberg's opinion.

Of course classical mechanics is a limit of quantum mechanics. One can see this in the saddle point approximation to the path integral.

However, what you are not understanding and which Landau and Lifshitz state clearly, is that quantum mechanics cannot be formulated without "classical concepts" also in its assumptions. It is not possible to derive classical physics from "purely quantum" assumptions.

One can use different language to state this assumption, but they are all essentially equivalent - measurement has a different status than the interactions described in the Hamiltonian.

 

[Fra]

As was said in a previous comment, a measurement device plays two different roles in the minimal interpretation:

1. It is a physical system, and so it is described by the Schrodinger equation.
2. It determines a "preferred basis" for computing probabilities.

In role #1, there is no distinction between measurement devices and any other physical system. In role #2, there is a big distinction. Probabilities only appear in quantum mechanics if you have measurements, not for any other interactions.

I don't understand why, when the issue is #2, people keep bringing up that #1 doesn't distinguish measurements from other interactions.

I see a reason to bring #1 into discussing #2:

By observer equivalence i expect that the non probabilistical hamiltonian of a complex system is "explainable" terms of the transformed views from the inside views which would contain probabilistic components due to internal measurements. So the deterministic evolution of the state vector should have a probabiliatic explanation that removes the classical baggage. Then we would have a complete duality between interaction and measurement.

But it is not yet known of course. But this to me suggests revision on qm.

/Fredrik

 

[vanhees71]

As was said in a previous comment, a measurement device plays two different roles in the minimal interpretation:

1. It is a physical system, and so it is described by the Schrodinger equation.
2. It determines a "preferred basis" for computing probabilities.

In role #1, there is no distinction between measurement devices and any other physical system. In role #2, there is a big distinction. Probabilities only appear in quantum mechanics if you have measurements, not for any other interactions.

I don't understand why, when the issue is #2, people keep bringing up that #1 doesn't distinguish measurements from other interactions. That's true, but it's only half the story.

But also #2 doesn't distinguish measurements from other interactions. The "preferred basis" is just the observer's choice which observable s/he likes to measure. That's the only meaning in which a basis has to be chosen to evaluate probabilities: To get the probabilities for the outcome of measurements you need the eigenbasis of the self-adjoint operator representing this observable. It's not more special then the observer's choice in classical physics too. If I measure the position of a particle I need another device than when I measure its momentum.