When Quantum Mechanics is thrashed by non-physicists #1

[TrickyDicky]

Well, if nothing else what that paper reveals is that most interpretational discussions in QM are "non-discussions" and therefore essentially useless because they are mainly based on misconceptions about what the theory says, either due to wrong or confusely presented postulates or ignoring some basic part of the process or misinterpreting what the predictions
of QM entail in practice.

 

[stevendaryl]

From my perspective, this thread is done may posts ago.

Well, experience shows that any discussion of interpretations of QM is eventually shut down by the moderators. They should just go ahead and put it on the list of forbidden topics. I actually enjoy those discussions, but the goal of Physics Forums seems to be to steer away from anything where the answer isn't already well-established.

 

[bhobba]

But MWI can be falsified. For example, if collapse is found to be real :) Maybe something like in http://arxiv.org/abs/1410.0270.

Atyy - I think you mean MAY be falsifiable.

None of this stuff can be proved one way or the other right now. Which interpretation you prefer is purely a psychological choice that really has no value scientifically - it's like the choice of SR vs LET. Most people choose SR once they understand it's simply a theory of space-time geometry and its simpler than this unobservable aether - of course simplicity is also in the eye of the beholder.

Thanks
Bill

 

[vanhees71]

I think that's backwards. MWI doesn't "invoke" an assumption of infinitely many parallel universes. Rather, the assumption that there is only one value for macroscopic observables is an ADDITIONAL assumption, not strictly speaking required.

I understand that it is convenient to make the assumption, and I'm fine with just saying: Well, let's assume something along those lines so that we can along with doing science. But to me, it's ad hoc, and it's only a rule of thumb, and not to be taken seriously. "The mind collapses the wave function" or "Measurement collapses the wave function" or "Decoherence collapses the wave function", or (Bohm-DeBroglie), "the only thing that is real is position, so it's already collapsed at all times" are all essentially equivalent. They are just ad hoc rules to get past interpretational difficulties in QM. I think that's fine, but I don't think people should pretend that there is anything other than an ad hoc choice being made.

According to standard quantum theory the state of a quantum system can be such that a given observable either has a determined value or it has not a determined value. In the former case the statistical operator is of the form
$$\hat{\rho}=\sum_{\alpha} P_{\alpha} |o,\alpha \rangle \langle o,\alpha|.$$
Here, $|o,\alpha \rangle$ is a complete orthonormal set of eigenvectors of the self-adjoint operator $\hat{O}$, which represents the observable of interest $O$. The $P_{\alpha} \geq 0$ and $\sum_{\alpha} P_{\alpha}=1$.

In all other states the observable $O$ has no determined value, and the state only describes the probablities to find a certain value. That's it. There's no more content in the quantum-theoretical setup concerning observables. I don't need a collapse only because I read of a pointer observable at a macroscopic measurement apparatus or, more likely in our digitalized world, the reading gets saved into a computer file providing a data base for the physicst to evaluate the outcome of the measurement (on an ensemble of equally and independently prepared systems measuring the observable $O$). Does, in your opinion, the collapse occur when the hardware has performed the measurement or when the DAQ writes it to the storage system of the computer or only when Alice evaluates the data or when Bob reads Alice's paper about the result?

I hope, it's clear that I consider these questions as rhetorical.

 

[Demystifier]

Each copy has a complete memory of definite observations. I guess that could be considered a "preferred basis" assumption, in some sense, but to really investigate the extent to which this needs to be an assumption, I think you would have to look at how "memories" work. It's possible that the mechanism of memory only works in one basis. The basis in which the states consists of superpositions of "seeing heads" and "seeing tails" has nothing like a memory to record such things.

Even if it is true that the preferred basis can be derived from the working of memory, the point is that somehow the existence of a preferred basis can be derived. The paper we discuss in this thread can also be viewed as a derivation of its existence.

 

[Demystifier]

They should just go ahead and put it on the list of forbidden topics.

I strongly hope they will not do it!

 

[vanhees71]

I'd also not put it on the list of "forbidden" topics, although it's pretty fruitless, because obviously there's no chance to keep it scientific, because most people take it as a philosophical topic.

The reason is simply that on the one hand from a scientific point of view, there's no problem with quantum theory. It's a well-defined mathematical setup (if you restrict yourself to non-relativistic quantum theory for systems with fixed (and conserved) particle number), which has a very clear FAPP interpretation such that it can be used to describe (so far even all) observations of Nature. With this the case is indeed close from a purely scientific point of view, because that's all what science is about: To find an as comprehensive quantitative description of that part of the world that is objective.

On the other hand, the full understanding of quantum theory is subtle, because it describes a part of the objective world that is far from our everyday experience, because quantum effects (despite the very fact that we have stable "solid" matter around us) are rare in our macroscopic world, and we are thus not used to it.

Also still for many people, even 90 years after the formulation of modern quantum theory, it seems not to provide a satisfactory picture about Nature, although it's the best and most comprehensive picture we have. This thus is a subjective feeling rather than a scientific shortcoming of quantum theory and strictly speaking not on topic in a scientific forum. Nevertheless these discussions help to sharpen the understanding of the theory. As long as the discussion is as civilized as here, it's fine also for a scientific forum.

Personally, I get some insight into what's discussed in this corner of science, which is far from my everyday use of quantum theory in high-energy nuclear/particle physics. So it's interesting, to widen my horizon about what's going on on the border between science and philosophy.

What I'd recommend is to make an extra subforum, labeled "Interpretational issues of Quantum Theory" or something like that. Then people, who don't like to discuss those issues can stay out of these threads more easily, and the Quantum Mechanics forum can concentrate on the "hard science and mathematical" part of quantum theory.

 

[TrickyDicky]

According to standard quantum theory the state of a quantum system can be such that a given observable either has a determined value or it has not a determined value.

This sentence contains a semantic trap in the word "either" that is related to the thread's paper. Because in practice for the theory to have any applicability and empirical and predictive validity only the case with determined values(so in practice always mixed states) is actuallly considered. That is ultimately why the standard quantum theory state of the first postulate isn't physical.
I recall that recently Weinberg published an approach to QM in which mixed state matrices took the central place instead of the usual vector states.

 

[vanhees71]

In any case the statistical operators are the true representants of states. Also Weinberg's textbook is very on the typical "no-nonsense line" of his usual brialliant textbook style. The chapter on "interpretations" is one of the best I've ever read about this issue although I don't agree with his statement that there are undecided issues. The minimal interpretation without collapse is a fully satisfactory one from a scientific point of view. Everything beyond this is philosophy (see my previous posting).

I don't know, where there is a "semantic trap" here. It's simply the standard interpretation of quantum theory since Born's famous footnote in his scattering-theory paper.

 

[Demystifier]

because that's all what science is about: To find an as comprehensive quantitative description of that part of the world that is objective.

What is science all about is determined by scientists. Unfortunately, there is no complete agreement among scientists on that. In particular, many scientists would not agree that the above is all what science is about.

For example, even theoretical physicists who are not interested in philosophical issues are not satisfied by having only quantitative description without a conceptual understanding. In particular, many theoretical physicists feel that a numerical simulation on a computer without a simple (even if approximate) analytical solution does not provide a completely satisfying description. To quote the Nobel laureate Eugene Wigner "It is nice to know that the computer understands the problem, but I would like to understand it too".

What I'd recommend is to make an extra subforum, labeled "Interpretational issues of Quantum Theory" or something like that. Then people, who don't like to discuss those issues can stay out of these threads more easily, and the Quantum Mechanics forum can concentrate on the "hard science and mathematical" part of quantum theory.

I think that's a great idea.

 

[TrickyDicky]

The minimal interpretation without collapse is a fully satisfactory one from a scientific point of view. Everything beyond this is philosophy

Aside from interpretations, the "hard", FAPP quantum theory applied by experimentalists with so much success includes a form of collapse (that it is of course epistemic, nothing physically collapses instantly, no ftl implications whatsoever) whether one wants to acknowledge it explicitly or not, most people applying QM don't think about it or even care about it. They just set up experiments, measure and calculate. They don't need to be explicitly aware of the quantum/classical interface, they are actually implementing it by doing measurements, and any scientific theory in physics(an emprical discipline) must include measurements in their formulation, otherwise they are pure math and have no connection with nature. I think the first three sections of the Ronde and Maasri paper make this clear with different words.

 

[vanhees71]

What is science all about is determined by scientists. Unfortunately, there is no complete agreement among scientists on that. In particular, many scientists would not agree that the above is all what science is about.

For example, even theoretical physicists who are not interested in philosophical issues are not satisfied by having only quantitative description without a conceptual understanding. In particular, many theoretical physicists feel that a numerical simulation on a computer without a simple (even if approximate) analytical solution does not provide a completely satisfying description. To quote the Nobel laureate Eugene Wigner "It is nice to know that the computer understands the problem, but I would like to understand it too".

Well, many interesting things in theoretical physics, cannot be addressed by analytical means (yet?). An example is lattice QCD, which is the only way to study full QCD without making use of perturbation theory, and there are a lot of fundamental things, one cannot study perturbatively (confinement, the hadron spectrum, the QCD phase diagram of strongly interacting matter) but with lattice. Are you saying that's not a satisfactory method to test QCD against observations and doesn't this lead to further progress in our understanding of hadrons and the strong interaction? I'd disagree with this entirely!

 

[vanhees71]

Aside from interpretations, the "hard", FAPP quantum theory applied by experimentalists with so much success includes a form of collapse (that it is of course epistemic, nothing physically collapses instantly, no ftl implications whatsoever) whether one wants to acknowledge it explicitly or not, most people applying QM don't think about it or even care about it. They just set up experiments, measure and calculate. They don't need to be explicitly aware of the quantum/classical interface, they are actually implementing it by doing measurements, and any scientific theory in physics(an emprical discipline) must include measurements in their formulation, otherwise they are pure math and have no connection with nature. I think the first three sections of the Ronde and Maasri paper make this clear with different words.

I've never heard or read a convincing argument for the necessity of a collpase. The naive collapse is unphysical, and it doesn't help to call it epistemic. I have minimally interpreted quantum theory without a collapse, and that's all what's needed by experimentalists and theorists alike to describe quantitative observations with a high accuracy. So again: What do I need a collapse for?

 

[Demystifier]

Are you saying that's not a satisfactory method to test QCD against observations and doesn't this lead to further progress in our understanding of hadrons and the strong interaction?

No, I'm saying that many physicists (not necessarily including myself) think that QCD would be even better understood if could solve it analytically, without computer simulations.

 

[vanhees71]

No question about that. If one could get an analytic solution of QCD that would for sure be the best...

 

[Demystifier]

What do I need a collapse for?

To account for the information which is available to you after the measurement, i.e. to make measurable predictions for a measurement performed after a measurement.

 

[vanhees71]

But that entirely depends on the measurement apparatus. If I destroy, e.g., my photon in absorbing it in detecting it, I can't do further measurements on it. What you are referring to is not measurement but preparation. The most simple preparation procedure is to perform a filter measurement, but still, where do you need a collapse there?

Let's discuss the Stern-Gerlach experiment for a spin-1/2 particle again to have a concrete example. You shoot a beam of particles into an inhomogeneous magnetic field, which leads to the entanglement between the particles' position and it's spin component determined by the direction of the magnetic field (usually taken as the $z$ direction). This is entirely understood by quantum dynamics. You can even evaluate this (nearly) completely analytically.

Then you simply only look at one of the two beams. You just only consider particles in this beam, which then is prepared in a state with a determined $\sigma_z$. Then you can do experiments with these particles, e.g., measuring $\sigma_x$. Nowhere is a collapse needed to describe state preparation!

 

[Demystifier]

No question about that. If one could get an analytic solution of QCD that would for sure be the best...

Of course. But if both approaches eventually give the same numbers, then what is the advantage of the analytical approach? Many think that one advantage is that the analytical approach gives a conceptual understanding that the purely numerical approach lacks. My point is, the mere fact that some physicists think so is a proof that physics is not only about numbers.

And if we can agree that theoretical physics is not only about getting the correct numbers but also about understanding the origin of those numbers, then we should not be very far from agreeing that physics may also be about philosophy of physics.

 

[Demystifier]

If I destroy, e.g., my photon in absorbing it in detecting it, I can't do further measurements on it.

You cannot do further measurements with the photon, but you can do further measurements with the atom which absorbed the photon. The claim that your new state is an atom in an excited state can also be expressed in terms of a collapse.

... which then is prepared in a state with a determined $\sigma_z$. ... Nowhere is a collapse needed to describe state preparation!

It is needed. You can prepare the state to be $\sigma_z=+1/2$, or you can prepare the state to be $\sigma_z=-1/2$. In the first case you can say that the state collapsed to $|+1/2>$, while in the second case you can say that the state collapsed to $|-1/2>$. If you do not use the word "collapse", then what would you say, how did the state come to one of those two states? Certainly not by unitary evolution governed by the Schrodinger equation!

 

[vanhees71]

Of course. But if both approaches eventually give the same numbers, then what is the advantage of the analytical approach? Many think that one advantage is that the analytical approach gives a conceptual understanding which the purely numerical approach lacks. My point is, the mere fact that some physicists think so is a proof that physics is not only about numbers.

And if can agree that theoretical physics is not only about getting the correct numbers but also about understanding the origin of those numbers, then we should not be very far from agreeing that physics may also be about philosophy of physics.

That's a tricky business. Of course, theoretical physics should provide mathematical models which are as simple as possible and at the same time as comprehensive as possible. That's why Newtonian mechanics is considered a "better theory" than the epicycle theory of ancient scientists.

On the other hand, a theory which "doesn't get the numbers out" to be compared to experiments, is no science either. At best it's interesting mathematics, which might lead to physics as soon as it is so far understood to make testable predictions. A nice example is string theory, which is an interesting piece of mathematics so far but no physics yet although it's study for sure is motivated from physics. This example also shows that from a scientific point of view it's more fruitful to try to attack the real open problems of contemporary physics than to ponder "cold cases" like the "collapse" or the "measurement problem".

 

[vanhees71]

You cannot do further measurements with the photon, but you can do further measurements with the atom which absorbed the photon. The claim that your new state is an atom in an excited state can also be expressed in terms of a collapse.Of course there is. You can prepare the state to be $\sigma_z=+1/2$, or you can prepare the state to be $\sigma_z=-1/2$. In the first case you can say that the state collapsed to $|+1/2>$, while in the second case you can say that the state collapsed to $|-1/2>$. If you do not use the word "collapse", then what would you say, how did the state come to one of these two states? Certainly not by unitary evolution governed by the Schrodinger equation!

Nothing has collapsed in my SG example. There was just unitary time evolution of one particle beam into two spatially separated particle beams with (FAPP) determined spin-$z$ components (one has $\sigma_z=1/2$, the other $\sigma_z=-1/2$). Nowhere did I envoke non-quantum dynamics, let alone an instantaneous action at a distance, to explain that I have well-separated beams with determined spin-$z$ components, and that's very important, because a collapse would be an unphysical process envoked as an "explanation" for a physical process, and that's not physics but esoterics!

The same with the photon part. Here you have a single atom as "detector" and a photon to begin with. Then you look at the case where a photon gets absorbed by the atom which is excited after that. Great, but no collapse either. The absorption of the photon can be well described with quantum dynamics (at least perturbatively).

 

[Demystifier]

This example also shows that from a scientific point of view it's more fruitful to try to attack the real open problems of contemporary physics than to ponder "cold cases" like the "collapse" or the "measurement problem".

How do you distinguish the "real" open problems from the "cold" ones? Most lists of the most important open problems in physics, such as this one
http://en.wikipedia.org/wiki/List_of_unsolved_problems_in_physics (see 1.9 Other problems),
include the collapse and measurement problem. Even Weinberg in his recent book admits that this is one of the most important open problems in physics.

 

[Demystifier]

There was just unitary time evolution of one particle beam into two spatially separated particle beams with (FAPP) determined spin-zz components (one has σz=1/2\sigma_z=1/2, the other σz=−1/2\sigma_z=-1/2).

If your initial beam contains only one particle, the experiments show that there are no two separated beams, but only one! How can you explain that, by unitary means without a collapse?

 

[zonde]

The same with the photon part. Here you have a single atom as "detector" and a photon to begin with. Then you look at the case where a photon gets absorbed by the atom which is excited after that. Great, but no collapse either. The absorption of the photon can be well described with quantum dynamics (at least perturbatively).

Wait, but what about (irreversible) amplification of signal? Single excited atom won't work as detector.

 

[stevendaryl]

The reason is simply that on the one hand from a scientific point of view, there's no problem with quantum theory

Yes, that's part of what is confusing and frustrating about QM, and intractable about ending the interpretation debates: There is no problem to be solved, in the sense of having an experiment whose result we are unable to predict (at least probabilistically). On the one hand, conceptually, there is lots of things that we don't understand about quantum mechanics, but on the other hand, there's no empirical guidance from unexplained results. It seems to me that if there is a solution to the interpretation problem for QM, we already have enough information to solve it--we're just not smart enough, maybe.

 

[stevendaryl]

Even if it is true that the preferred basis can be derived from the working of memory, the point is that somehow the existence of a preferred basis can be derived. The paper we discuss in this thread can also be viewed as a derivation of its existence.

Yes, I think that a preferred basis might very well be derivable. My suggestion, which I don't know whether anyone has looked into it, is that the preferred basis isn't inherent in the physics of particles and fields, but instead comes from the human need to be able to remember past observations. The physics doesn't care about one basis versus another, but WE do.

 

[Demystifier]

It seems to me that if there is a solution to the interpretation problem for QM, we already have enough information to solve it--we're just not smart enough, maybe.

Yes, that's why many physicists try to solve it without making new experiments or even without making new measurable predictions.

 

[Demystifier]

Yes, I think that a preferred basis might very well be derivable. My suggestion, which I don't know whether anyone has looked into it, is that the preferred basis isn't inherent in the physics of particles and fields, but instead comes from the human need to be able to remember past observations. The physics doesn't care about one basis versus another, but WE do.

Yes, some variants on MWI already look into that direction.

 

[vanhees71]

If your initial beam contains only one particle, the experiments show that there are no two separated beams, but only one! How can you explain that, by unitary means without a collapse?

The "beam" can consist of single particles. By assumption the location of the particle at the 2nd Stern-Gerlach apparatus is enough to be sure that there enters a particle with definite $\sigma_z$.

Of course, to test the predictions of QT you always need "beams" in the sense that you have to perform the experiment with single particles often enough to collect "enought statistics" to have a large enough confidence level to confirm disconfirm the probabilistic prediction. That's why this point of view is usually called "(Minimal) Statistical Interpretation" or "Ensemble Interpretation".

 

[Demystifier]

The "beam" can consist of single particles. By assumption the location of the particle at the 2nd Stern-Gerlach apparatus is enough to be sure that there enters a particle with definite $\sigma_z$.

Of course, to test the predictions of QT you always need "beams" in the sense that you have to perform the experiment with single particles often enough to collect "enought statistics" to have a large enough confidence level to confirm disconfirm the probabilistic prediction. That's why this point of view is usually called "(Minimal) Statistical Interpretation" or "Ensemble Interpretation".

Ah, now I think I finally see what interpretation of QM do you have in mind. In the ensemble interpretation, the wave function does not represent individual measured objects, but only statistical properties of the ensemble. With that view, you don't need collapse.

That's fine, but then, instead of a collapse, you need something else. Since the individual measured objects are not wave functions, it follows that there is something which is not a wave function. But then what it is? Are you (like Ballentine) agnostic about that question, or do you believe in some sort of hidden variables such as the Bohmian ones?

 

[vanhees71]

I'm agnostic about this question. As long as there's no clear evidence against standard QT, there seems not to be a chance to find a more comprehensive theory.

 

[Ilja]

I've never heard or read a convincing argument for the necessity of a collpase. The naive collapse is unphysical, and it doesn't help to call it epistemic. I have minimally interpreted quantum theory without a collapse, and that's all what's needed by experimentalists and theorists alike to describe quantitative observations with a high accuracy. So again: What do I need a collapse for?

Strange. I have never heard about a minimal version without collapse.

The point is very simple: Measurements play a double role. First, they measure A of something, which is in a state with wave function $\psi=\sum a_i\psi_i$. The minimal interpretation tells us that with probability $a_i^2$ the result is the $\lambda_i$ which is defined by $A \psi_i=\lambda_i\psi_i$. Fine, this part we can manage without introducing a collapse. But then there is the second part, state preparation. If we prepare a state in $\psi_i$, what do we have to do? To make a measurement, and wait until the measurement result is $\lambda_i$. This prepares the state as being $\psi_i$. You cannot make a minimal interpretation without such a rule for state preparation. And Nature does not know if the measurement is made for getting a result or for preparing a state. Thus, this double role is sufficient to require that the measurement transforms the initial state $\psi=\sum a_i\psi_i$ with probability $a_i^2$ into the final state $\psi_i$. But this is the collapse.

 

[Demystifier]

I'm agnostic about this question. As long as there's no clear evidence against standard QT, there seems not to be a chance to find a more comprehensive theory.

But you do admit that probably there is something which is not the wave function (otherwise you would need either the collapse or many-worlds), even if you don't care what exactly that something might be. Am I right?

 

[Ilja]

It seems to me that if there is a solution to the interpretation problem for QM, we already have enough information to solve it--we're just not smart enough, maybe.

I think "smart" is not the problem. Most are not ready to accept the hypothesis that relativistic symmetry is not fundamental, not even as a working hypothesis, which one accepts for the sake of the argument, to have a look at how it fails, no, it is rejected out of hand, without any discussion. Like here, where even to discuss the Lorentz ether interpretation vs. Minkowski interpretation is forbidden (BTW with a completely wrong argument that there is no difference in the predictions, in a situation where one can derive Bell's inequalities only in the Minkowski interpretation, but not in the Lorentz interpretation, which allows hidden causal influences, SCNR).

And this prejudice, of course, severely damages all realistic interpretations of QT, because they all require (because of Bell's inequality) a preferred frame.

 

[Ilja]

With such a more common terminology, their theorems say that the physical content of QM is not basis independent. But this claim is not new at all. This is nothing but a restatement of the preferred basis problem appearing in one way or another in all interpretations of QM.

Where do you see the preferred basis problem in the de Broglie-Bohm interpretation?