Interpretations of Quantum Mechanics

[alexepascual]

Which interpretation of QM do you like?

 

[fleem]

Non-local hidden variables, but not Bohm/pilot waves. Bohm & de Broglie assumed the rules we've devised simply to describe the average, large-scale behavior of many countless little machines (particle-particle interactions), namely classical mechanics like space, time, continuums, etc., must also be the founding axioms of understanding the behavior of a single such machine. That assumption is not logical.

 

[Count Iblis]

The MWI, because it is a minimal interpretation. Tegmark in his most recent article has argued that the "many worlds" are actually a consequence of symmetry. Thgis is implemted by unitary time evolution, but even if you get rid of unitary time evolution, you don't necessarily get rid of the many worlds.


E.g., we have some system in some symmetrical but unstable state, like a pencil balanced perfectly on its tip at zero temperature. Then QM predicts that the pencil will fall to the ground. According to CI, the rotational symmetery around the z-axis must somehow be broken. But that's impossible if the laws of physics are exactly symmetrical and the initial state respects this symmetry, no matter how you modify the unitary time evolution law.

 

[humanino]

Rovelli's relational QM by far.

 

[humanino]

But that's impossible if the laws of physics are exactly symmetrical and the initial state respects this symmetry, no matter how you modify the unitary time evolution law.

That's called "spontaneous symmetry breaking" and I do not see why this should be impossible. It is used in the standard model of particle physics for instance.

 

[Count Iblis]

That's called "spontaneous symmetry breaking" and I do not see why this should be impossible. It is used in the standard model of particle physics for instance.

What is impossible is for any proposed physical mechanism for the "collapse" to respect the symmetry. In the CI this whole issue is sidestepped and in the MWI you have a global superposition which is symmetric. Of course, this does not stop anyone from doing physics in the usual way...

 

[RUTA]

I prefer Relational Blockworld, but I'm very biased

 

[Adrian59]

The MWI, because it is a minimal interpretation.

I have one problem with MWI, though I accept I may be looking at it too literally. The original MWI was devised to explain the double slit experiment conundrum. So the one photon goes through each slit but in different worlds. Not to upset causality or the conservation of mass or energy, each world only sees one photon – no problem. However now each photon is in a parallel world (universe), why does interference happen?

 

[Fra]

I do not confess cleanly to any of the major interpretations. There are fragments of reasoning in Rovelli's RQM paper that I think is outstanding, but others are not.

I guess I have my own intrinsic relational subjective information interpretation, without multiverses and most of all without the bird view realism that rovelli has.

Take RQM and add more undecidability and replace his realist symmetries with emergent frog-view symmetries and I am close.

/Fredrik

 

[humanino]

I just collapsed to preferring Fra's interpretation. Can I change my relational interpretation of my own self ?

 

[Fredrik]

QM is just an algorithm that tells us how to compute probabilities of possible results of experiments. It doesn't need an interpretation, and I doubt that it's even possible to find one that's both well-defined and logically consistent. I think the attempts to find a correct interpretation are fundamentally misguided. They are all trying to tell us what QM really describes, but a theory doesn't have to describe anything. All it has to do in order to be falsifiable is to make predictions about probailities of possible results of experiments.

 

[Fra]

QM is just an algorithm that tells us how to compute probabilities of possible results of experiments. It doesn't need an interpretation, and I doubt that it's even possible to find one that's both well-defined and logically consistent. I think the attempts to find a correct interpretation are fundamentally misguided. They are all trying to tell us what QM really describes, but a theory doesn't have to describe anything. All it has to do in order to be falsifiable is to make predictions about probailities of possible results of experiments.

If we are talking about interpretations only that have no ambition to improve or solve new problems that I can almost agree. But my "interpretation" actually implies that the formalism of quantum mechanics is only a special case and not the formalism we need for a intrinsic relational measurement theory. I do not only view the standar model as an effective theory, it's worse that that, I view the standard framework or QM and QFT as en effective framework! And my personal view is that the problems implicit in the open question in physics can not be solved or even properly posed within the current framework of proper QM etc.

And if you want to improve something, just knowing that it apparently works isn't enough I think. You need to ask different question to drive a car than to build a car, or to improve an existing car.

/Fredrik

 

[Fra]

But that's impossible if the laws of physics are exactly symmetrical and the initial state respects this symmetry, no matter how you modify the unitary time evolution law.

If you see no reason to treat physical law different than any other measurement, then inference of physical law from interaction experiences, are subject to the same physical constraints as is any physical measurement/interaction.

In my view, there is no such thing as an exact certain inference of perfect symmetry. Therefore the symmetry arguments commonly used, are not really strict. They represent special cases.

I share Lee Smolin critique of the notion of timeless law, and by consequence also the notion of timeless universal symmetry. This doesnt' mean symmetry arguments are useless, it just means that we might need a new undertand of what a symmetry is, and it's general origin. A system can act as if the symmetry is exact, but this doesn't mean the symmetry will stay exact forever, and neither is there a contradiction.

I think a lot of contradiction comes because an expectation that the universe as a whole is more decidable and predictable than it really is. Probabilistic determinism isn't enough. The notion of probability is even a key suspect as the physical basis of statistics is one of the things that isn't properly abstracted in current formalisms.

Rovelli's also avoids the "physical meaning of probability" s in his RQM papper. If he didn't, a good reasoning could I think have been even better.

/Fredrik

 

[ajw1]

... but a theory doesn't have to describe anything. All it has to do in order to be falsifiable is to make predictions about probailities of possible results of experiments.

I have seen these kinds of arguments before, but for me this is a really strange attitude towards physics:
- you can make a fitting simulation model for any problem without knowing any law of physics.
- it would mean that physicist only needs to start thinking when an experiment contradicts their current models.

Furthermore it seems this view on physics is mostly used in relation to QM and its interpretaions.

 

[sokrates]

QM is just an algorithm that tells us how to compute probabilities of possible results of experiments. It doesn't need an interpretation, and I doubt that it's even possible to find one that's both well-defined and logically consistent. I think the attempts to find a correct interpretation are fundamentally misguided. They are all trying to tell us what QM really describes, but a theory doesn't have to describe anything. All it has to do in order to be falsifiable is to make predictions about probailities of possible results of experiments.

Don't youthink this is a little too humble? How do you reach the conclusion that it's "just an algorithm"?

You are confusing the scientific method with intrinsic facts about nature. These are two independent things. Quantum Physics is at a level low enough to answer those very fundamental questions. It is definitely not "just an algorithm". We don't even know how it works yet...

 

[Count Iblis]

Yes, I think it is fair to say that if you think QM is "just an algorithm" you must also be similarly open minded about the very existence of particles like electrons, quarks etc. We can't directly detect them. All that we can directly observe are the states of macroscopic measurement devices.

So, if I play the Devil's advocate, I could claim that Nature is described by a fundamental theory that one can turn into an effective theory yielding statistical predictions using a path integral formalism involving integrating over fictitious field configuarations.

The formalism is then the same as the Standard Model, except for the fact that only the sources and sinks of particles are real, not the particles themselves. So, when you compute amplitudes for process using Feynman diagrams, the external lines always start end end at detectors/sources. The amplitude is associated to the state of the detectors, not to particle states.

 

[fleem]

These debates are always really debates about reductionism, Occam's razor, that sort of thing, and why Occam's razor is true. Epicycles worked pretty well and there were those that said, "Therefore we understand the movement of planets pretty well". But embracing epicycles certainly slowed progress. Here's why Occam's razor is true: The simpler the explanation, the more things it will describe. Thus we replaced epicycles with gravity and were able to predict both planetary motion and apples falling from trees with one theory. Right now QM, QFT, QED, all seem a little contrived, especially considering they really don't predict individual experiments (only statistically). So someone using the phrase "interpretation of QM" typically means, "Why is it only statistical?" and "Why does it look like classical laws like conservation of energy are disobeyed while a system is closed (not interacting with the rest of the universe) until it once again interacts with us?".

 

[Fredrik]

I have seen these kinds of arguments before, but for me this is a really strange attitude towards physics:
- you can make a fitting simulation model for any problem without knowing any law of physics.
- it would mean that physicist only needs to start thinking when an experiment contradicts their current models.

Those "fitting simultion models" would be pretty bad theories because of their limited predictive power. We would always be trying to find better theories, e.g. by unifying several different theories into a new one, like we did with electricity and magnetism.

Furthermore it seems this view on physics is mostly used in relation to QM and its interpretaions.

This is because before QM we could always claim that our theories at the very least are descriptions of "fictional" universes that are similar to our own. It's not clear that such a claim about QM even makes sense. I still haven't seen a well-defined and logically consistent many-worlds formulation of QM. (A "Copenhagenish" formulation of QM looks very much like it can't be interpreted as a description of the universe, since you can't include yourself in the physical system that you're trying to describe).

Don't youthink this is a little too humble? How do you reach the conclusion that it's "just an algorithm"?

I can't prove it conclusively, but arguments such as the one from section 9.3 in Ballentine's book look pretty strong to me. (I think that's the right section number, but it's from memory). Ballentine's view on these matters appears to be the same as mine, but he says it in a different way. His way is to say that a wavefunction isn't a representation of the properties of a physical system, but a representation of the properties of en ensemble of identically prepared systems.

You are confusing the scientific method with intrinsic facts about nature.

Uhm, no. What would it even mean to do that?

I don't doubt that there are "intrinsic facts about nature", but the only knowable facts in science are statements about how accurate a theory's predictions are.

These are two independent things. Quantum Physics is at a level low enough to answer those very fundamental questions. It is definitely not "just an algorithm".

Sounds like you think that's a fact. If you're sure about that, then you should be able to justify it.

Yes, I think it is fair to say that if you think QM is "just an algorithm" you must also be similarly open minded about the very existence of particles like electrons, quarks etc. We can't directly detect them. All that we can directly observe are the states of macroscopic measurement devices.

Yes. I think the right approach to these matters is to re-interpret questions like "do electrons exist?" to something like "does the theory that defines what an electron is make accurate predictions?".

The other day I answered the question "is gravity a force?" by saying that it's a force in Newton's theory and not a force in Einstein's theory. It just isn't possible to say that gravity "is a force" or "isn't a force" in some absolute, theory-independent, way. I think of electrons the same way.

So, if I play the Devil's advocate, I could claim that Nature is described by a fundamental theory that one can turn into an effective theory yielding statistical predictions using a path integral formalism involving integrating over fictitious field configuarations.

The formalism is then the same as the Standard Model, except for the fact that only the sources and sinks of particles are real, not the particles themselves. So, when you compute amplitudes for process using Feynman diagrams, the external lines always start end end at detectors/sources. The amplitude is associated to the state of the detectors, not to particle states.

I wouldn't have any problem with that, so I guess I'm the

 

[Dmitry67]

Interesting that nobody voted for Copenhagen :) Looks like it is finally declared dead :)

 

[humanino]

The problem with this kind of thread is that it systematically ends up in an endless gathering of poorly informed opinions, such as

We don't even know how it works yet...

You do not know how it works maybe, however since (1932) Von Neumann wrote a book where he clearly explained why it is unnecessary to let go mathematical rigor (as Dirac did), as a matter of fact we have a pretty robust definition of every single step in the quantum algorithm. The book is elementary and pretty much requires no prior knowledge. There is nothing left unproven, to such an extent that the couple of flawed demonstration have been corrected by other authors.

If you do not realize that there is no mathematical difficulty with QM, you may very well have misunderstood what drags people towards various interpretations. So for the (N+1)th time, where N is an arbitrarily large number, when we say "shut up and calculate" we do not mean there is nothing to understand or interpret. We mean that there is nothing convincing enough to unite the community available on the market.

when you compute amplitudes for process using Feynman diagrams, the external lines always start end end at detectors/sources. The amplitude is associated to the state of the detectors, not to particle states.

I do not know what "playing the devil advocate" has anything to do with the above statement. Any interpretation in disagreement with the above statement could not be falsified !

 

[RUTA]

Yes, I think it is fair to say that if you think QM is "just an algorithm" you must also be similarly open minded about the very existence of particles like electrons, quarks etc. We can't directly detect them. All that we can directly observe are the states of macroscopic measurement devices.

So, if I play the Devil's advocate, I could claim that Nature is described by a fundamental theory that one can turn into an effective theory yielding statistical predictions using a path integral formalism involving integrating over fictitious field configuarations.

The formalism is then the same as the Standard Model, except for the fact that only the sources and sinks of particles are real, not the particles themselves. So, when you compute amplitudes for process using Feynman diagrams, the external lines always start end end at detectors/sources. The amplitude is associated to the state of the detectors, not to particle states.

Are you calling me the Devil?

But seriously, do you have a citation for this interpretation -- proponents, opponents, discussion, etc?

 

[sokrates]

I can't prove it conclusively, but arguments such as the one from section 9.3 in Ballentine's book look pretty strong to me. (I think that's the right section number, but it's from memory). Ballentine's view on these matters appears to be the same as mine, but he says it in a different way. His way is to say that a wavefunction isn't a representation of the properties of a physical system, but a representation of the properties of en ensemble of identically prepared systems.

Ballentine is emphasizing the statistical nature of the wavefunction, he's not making general epistemological arguments about what a theory could tell us, and what it means. I don't think anyone would object to the argument that
" a wavefunction isn't a representation of the properties of a physical system, but a representation of the properties of en ensemble of identically prepared systems."

Honestly, I find this irrelevant to your previous remark on Quantum Mechanics being "just an algorithm". Ballentine isn't implying that. Or is he?

Uhm, no. What would it even mean to do that?

I don't doubt that there are "intrinsic facts about nature", but the only knowable facts in science are statements about how accurate a theory's predictions are.

The process of developing a theory and the measure of its success are well-defined by the scientific method. But overrating it, and blindly sticking to its "formulated" restrictions, I think, is unimaginative and rather boring. The facts are out there, and we know that in some examples (as Count Iblis aptly pointed out) different "formulations" could explain the same facts - and WE get to decide which "algorithm" is better than the other and so forth. So this is like choosing a basis to explain an experiment. It is irrelevant.

My take is that quantum mechanics is much more than just an algorithm and chances are that the theory is trying to lead us the underlying facts by creating hard problems (measurement, interpretation issues, and so on). Maybe when we understand it better some day, we'll understand these issues in retrospect.

Sounds like you think that's a fact. If you're sure about that, then you should be able to justify it.

I can't prove it conclusively either. But in science, you know, the burden of proof is on the person asserting the positive claim, not on the skeptics to refute it.

 

[sokrates]

You do not know how it works maybe, however since (1932) Von Neumann wrote a book where he clearly explained why it is unnecessary to let go mathematical rigor (as Dirac did), as a matter of fact we have a pretty robust definition of every single step in the quantum algorithm. There is nothing left unproven, to such an extent that the couple of flawed demonstration have been corrected by other authors.
[...] So for the (N+1)th time, where N is an arbitrarily large number, when we say "shut up and calculate" we do not mean there is nothing to understand or interpret. We mean that there is nothing convincing enough to unite the community available on the market.

When I said, "we don't even know how it works", I was referring to the existing problems of QM, i.e, non-locality, measurement problem, how different ways of thinking about it could possibly lead to different conclusions and so forth. As Feynman put it: We don't know the machinery was what I was implying.

I very well know how the "algorithm" works, and you don't know me, so please control yourself and refrain from pointless personal attacks. Believe me it pollutes the forum more than the interpretation debates. I think I told you this a number of times in a number of other threads, but you constantly stalk my posts and seize every opportunity to launch an assault :) Please don't hijack this one too, with your personal issues with me. You could PM me all day for personal comments. I don't mind.

You happily claim that "there's nothing left unproven" in Quantum Mechanics, so could you enlighten us on how the measurement problem is solved mathematically? Because to your dismay, it still is an unresolved problem.

There are lots of things "on the market"(whatever you mean by that) that are evolving to make us understand QM better, they are just not in 'textbooks' yet. Some people are working on this, and this is an active field.

 

[humanino]

When I said, "we don't even know how it works", I was referring to the existing problems of QM, i.e, non-locality, measurement problem, how different ways of thinking about it could possibly lead to different conclusions and so forth. As Feynman put it: We don't know the machinery was what I was implying.

I'm sorry I can not read your mind when you are trying to repeat what you read Feynman said 50 years ago.

I mention the measurement problem to you in another thread when you were complaining about angular momentum.

 

[freelance4lyf]

I vote for Copenhagen! Then again I haven't actually taken QM yet, only intro to modern phys. I recently read Atomic Physics and Human Knowledge by Bohr (a collection of essays and lectures) and he has some very interesting points about the Copenhagen Interpretation. Again, I'm barely a 3rd year undergrad but this is an interesting debate.

 

[Fredrik]

Ballentine is emphasizing the statistical nature of the wavefunction, he's not making general epistemological arguments about what a theory could tell us, and what it means. I don't think anyone would object to the argument that
" a wavefunction isn't a representation of the properties of a physical system, but a representation of the properties of en ensemble of identically prepared systems."

Honestly, I find this irrelevant to your previous remark on Quantum Mechanics being "just an algorithm". Ballentine isn't implying that. Or is he?

Actually yes. It doesn't make much sense to say that a theory "describes" something unless it contains a mathematical representation of the properties of the system it's supposed to be describing (or at least an approximate version of it). I would say that that's the minimum requirement that must be satisfied. What else can we mean by a "description" or a "model"?

But overrating it, and blindly sticking to its "formulated" restrictions, I think, is unimaginative and rather boring.

The scientific method is unimaginative and boring?! That's a pretty controversial claim in a science forum.

The facts are out there, and we know that in some examples (as Count Iblis aptly pointed out) different "formulations" could explain the same facts - and WE get to decide which "algorithm" is better than the other and so forth. So this is like choosing a basis to explain an experiment. It is irrelevant.

Huh. I said these things myself. Why are you repeating them to me?

But in science, you know, the burden of proof is on the person asserting the positive claim, not on the skeptics to refute it.

This is a joke, right? The only thing that everyone agrees about is that QM is an algorithm that tells us the probabilities of possible results of experiments. The disagreement is about whether it's more than that. You're claiming it is. I'm claiming it isn't. How on Earth did you come to the conclusion that the burden of proof is on me?

 

[Fredrik]

Interesting that nobody voted for Copenhagen :) Looks like it is finally declared dead :)

On the contrary. My view (which I described above) is a version of the Copenhagen interpretation. I have really just removed the (in my opinion absurd) claim that "wavefunction collapse" is exact, and clarified a few points.

Actually, I strongly doubt that Bohr and Heisenberg really meant that the collapse is exact, or that measuring devices are exactly classical. They were too smart for that. It seems much more likely to me that they meant that for practical purposes, it's OK to use the approximation that wavefunction collapse is exact, and that measuring devices are exactly classical. So I would say that the CI is alive and well, and that the only problem with it is that people today are defining it in a way that the originators would have hated.

By the way (this is unrelated to the content of this post), I don't see why people like the "relational interpretation" so much. I know that the basic idea, that systems don't have properties in an objective sense, but rather with respect to some other system, is pretty appealing. But to me this sounds a lot more like an ingredient of a many-worlds interpretation than like a standalone interpretation. Maybe it's possible to take Everett's MWI, which really only defines the "bird's view" properly, and complete it by adding something like the "relational interpretation" to define the "frog's view".

 

[humanino]

Actually, I strongly doubt that Bohr and Heisenberg really meant that the collapse is exact, or that measuring devices are exactly classical. They were too smart for that. It seems much more likely to me that they meant that for practical purposes, it's OK to use the approximation that wavefunction collapse is exact, and that measuring devices are exactly classical. So I would say that the CI is alive and well, and that the only problem with it is that people today are defining it in a way that the originators would have hated.

I must say it feels good to hear reasonable voices. So I wanted to thank you.

 

[sokrates]

Actually yes. It doesn't make much sense to say that a theory "describes" something unless it contains a mathematical representation of the properties of the system it's supposed to be describing (or at least an approximate version of it). I would say that that's the minimum requirement that must be satisfied. What else can we mean by a "description" or a "model"?

This doesn't mean that the theory cannot penetrate its own shell and lead to us new theories. In fact, this is what ALL scientific theories have been doing, evolving to better theories by methods of falsification. SO let me get back to the REAL issue we were discussing:

Whether QM is JUST AN ALGORITHM - or is it more than that is what I was objecting, you skillfully avoid this point but this is what we are debating. If it was just an algorithm, it would have no potential for a future theory that would replace it. It wouldn't be able to do predictions and so forth... To tell you the truth, it's the first time I am using Quantum and Algorithm in the same sentence.

I am sorry but Ballentine isn't even remotely implying what you are saying here.

The scientific method is unimaginative and boring?! That's a pretty controversial claim in a science forum.

Calm down. I didn't say that. Maybe you could try to understand what I wrote by reading my earlier post. I said you are overrating it by misrepresenting quantum theory as an "algorithm".

Huh. I said these things myself. Why are you repeating them to me?

No; you didn't. Try reading carefully. And what's the point of such a question?

This is a joke, right? The only thing that everyone agrees about is that QM is an algorithm that tells us the probabilities of possible results of experiments. The disagreement is about whether it's more than that. You're claiming it is. I'm claiming it isn't. How on Earth did you come to the conclusion that the burden of proof is on me?

Do you even know what the word "algorithm" means and implies? Does an algorithm make predictions?? You are obviously using the word out of its context and definition. Maybe you should look it up and clarify your claims around this new definition of QM.

 

[Fredrik]

If it was just an algorithm, it would have NO POTENTIAL for a future theory that will replace it.

Yes it would. This is just crazy talk.

I am sorry but Ballentine isn't even remotely implying what you are saying here.

Yes he does. But you seem to have misunderstood what I'm saying. I hope you have. I can't think of any other explanation for the bizarre claim of yours that I quoted first in this post.

No; you didn't. And what's the point of such a question?

Yes I did, but thank you for proving that it's pointless to continue this discussion. I won't waste any more time on you.