Quantum gravity and the 'measurement problem'

[JustinLevy]

There has at times been expressions of hope that a theory of quantum gravity may help resolve issues such as the 'measurement problem' in quantum mechanics (Penrose, and Smolin I think, and probably others).

Have any of the recent 'advances'/'new lines of study' towards a quantum theory of gravity provide any indication that they would indeed help?

 

[Fra]

There has at times been expressions of hope that a theory of quantum gravity may help resolve issues such as the 'measurement problem' in quantum mechanics (Penrose, and Smolin I think, and probably others).

Have any of the recent 'advances'/'new lines of study' towards a quantum theory of gravity provide any indication that they would indeed help?

I for one think there is a relation, but I am not yet aware of any published satisfactory connection.

I personally see many good reasons to suspect a connection, but it's clear that people have quite differing ideas on this too, exactly what the connection is and how to best exploit it.

Penrose vision seems to be to exploit gravity, to restore some kind of objective wave function collapse of a superposition, by somehow fiddling with the energy differences. I personally do not find his reasoning attractive and I even think his doing it backwards. But I think his intutition of a connection is justified.

Another way, for those that doesn't see the collapse as a "problem" in the same sense as those who seek to restor seom level of realism, and isntead simply see it as a subjective information update, one can exploit this possible connection in another way - to use a proper reconstruction of the measurement theory itself, and post-dict gravity. In this way, you would get a measurement perspective on gravity from start and thus probably solve QG.

Thus, a proper understanding of measurement theory may solve QG and post-dicts gravity, rather than they other way around.

This is the lineout I believe in, but I think there is very little, if anything published on this that I'm aware of at least :( So if you want to see such a paper chances are you might have to write it yourself ;)

On the penrose angle, there is I think at least more published. But I don't share that perspective so I haven't searched that much down that road.

/Fredrik

 

[RUTA]

There has at times been expressions of hope that a theory of quantum gravity may help resolve issues such as the 'measurement problem' in quantum mechanics (Penrose, and Smolin I think, and probably others).

Have any of the recent 'advances'/'new lines of study' towards a quantum theory of gravity provide any indication that they would indeed help?

I would say any path integral approach to QG works de facto, since a definite final state is part of that approach. Thus, the measurement problem is a nonstarter in QG theories like quantum Regge calculus and causal dynamical triangulation.

 

[Finbar]

I think that solving one of these problems will naturally involve solving the other. Ultimately what is needed is a proper understanding of QFT and its extension to curved space. Already at the semi-classical level there seem to be fundamental issues that need to be resolved.

But before we even get started we need a good definition of the measurement problem.

This paper is not on QG but i think its very interesting in terms of the measurement problem.

http://arxiv.org/abs/0906.4919

 

[marcus]

This paper is not on QG but i think its very interesting in terms of the measurement problem.

http://arxiv.org/abs/0906.4919

Christof Wetterich! I just started a thread about an Asymptotic Safety paper by him and Shaposhnikov. It was Wetterich's insight that opened a door for Martin Reuter to restart the A.S. program. Some people have a combination of creativity and luck. Not only do they come up with ideas but then the ideas sometimes work. I think he may be one of those.

 

[Fra]

This paper is not on QG but i think its very interesting in terms of the measurement problem.

http://arxiv.org/abs/0906.4919

Thanks for the link, 33 pages it too long for me to skim today, but I'll try to skim at a later time. Looks interesting.

My first impression from the first page is that I like the idea that he seems QM as an emergent statistics for isolated subsystems and that the unitary evolution is merely a special case. This is just how I see it as well. I am curious to see what his "general case" is.

But my main second thoughts just from the first page is his starting point of a classical statistics with infinitely many states. Although in a certain sense, I think this is correct (since I think QM somehow is a limiting case also in terms of complexity), it is still missing the physics of the non-limit, and I'm not sure if it's satisfactory to resolve the issues I see. To me a system with infinitely many states is somehow "non-physical", except in FAPP sense, but FAPP sense isn't enough for a fundamental rethinking I thikn. I would expect the scaling of this complexity, to yield new physics insights. I'm not sure if he says anything about that later. I personally think this complexity scaling is where gravity enters the picture, as the degrees of freedom that is distinguishable to a given observer may increase or decrease in my envisioned general case.

/Fredrik

 

[meopemuk]

There has at times been expressions of hope that a theory of quantum gravity may help resolve issues such as the 'measurement problem' in quantum mechanics

I don't think that 'measurement problem' is really a problem. The only problem that I can see there is a widespread misunderstanding of the true meaning of quantum mechanics.

I also don't see any fundamental difficulty in constructing a quantum theory of gravity. In the simplest approach such a theory can be constructed by taking usual Newton's Hamiltonian and replacing dynamical variables there with quantum operators. This simplistic approach already describes the red shift and time dilation in a perfect agreement with experiment. Mercury's perihelion shift can be accommodated by adding a few extra velocity-dependent terms (the Einstein-Infeld-Hoffmann Hamiltonian). A simple modification of this Hamiltonian takes care of the light bending and spin effects (de Sitter and Lense-Thirring effects). The only difficulty is that there is no regular way to derive higher-order terms in this Hamiltonian. But in my opinion this difficulty is more technical than fundamental.

Eugene.

 

[hamster143]

I would say any path integral approach to QG works de facto, since a definite final state is part of that approach. Thus, the measurement problem is a nonstarter in QG theories like quantum Regge calculus and causal dynamical triangulation.

But all possible intermediate states are also part of the path integral approach.

 

[George Jones]

A post I made almost four years summarizes some of the tension that Penrose sees between quantum theory, the measurement problem, and gravity.

https://www.physicsforums.com/showthread.php?p=910951#post910951

 

[Finbar]

Thanks for the link, 33 pages it too long for me to skim today, but I'll try to skim at a later time. Looks interesting.

My first impression from the first page is that I like the idea that he seems QM as an emergent statistics for isolated subsystems and that the unitary evolution is merely a special case. This is just how I see it as well. I am curious to see what his "general case" is.

But my main second thoughts just from the first page is his starting point of a classical statistics with infinitely many states. Although in a certain sense, I think this is correct (since I think QM somehow is a limiting case also in terms of complexity), it is still missing the physics of the non-limit, and I'm not sure if it's satisfactory to resolve the issues I see. To me a system with infinitely many states is somehow "non-physical", except in FAPP sense, but FAPP sense isn't enough for a fundamental rethinking I thikn. I would expect the scaling of this complexity, to yield new physics insights. I'm not sure if he says anything about that later. I personally think this complexity scaling is where gravity enters the picture, as the degrees of freedom that is distinguishable to a given observer may increase or decrease in my envisioned general case.

/Fredrik

At some point he mentions the "coarse graining" of the infinitely many states. I think this is very much like the RG so that at the scale of observation there only seems to be a finite number of states. In a way it is like "hidden variables" except the variables are not hidden they are just coarse grained such that they appear to be absent in the QM description at some scale. So if what you mean by "non-physical" is "non-measureable" then I agree.

Your last sentence is very nice. Are you thinking along the lines of how the definition of the vacuum is dependent on the gravitational field?

Because in the paper he mentions that the infinitely many states should be associated to the vacuum(say of an atom). Then in gravity the degrees of freedom that are observable for each observer becomes dependent of the gravitational field.

I think that reality is scale dependent. In the sense that if I preform measurements at a certain scale and therefore only care about measuring lengths and energies to the nearest meter or 1/meter say then all by observables will be unaffected by the uncertainty principle. So I have a set of observable {A} that are defined at the scale of a meter which are unaffected by QM and hence seem to form a complete description of reality. On a completely different scale say 1000000th of a meter I can also define a different set of observables {B} which I measure only up to 1000000th of a meter or 1/(1000000th of a meter) then these too can form a complete description of reality. Its only when we try and form a description of reality at multiple scales that the description is incomplete and reality breaks down.

Perhaps quantum gravity will provide a link between scales to form a scale independent reality.

 

[Fra]

In a way it is like "hidden variables" except the variables are not hidden they are just coarse grained such that they appear to be absent in the QM description at some scale. So if what you mean by "non-physical" is "non-measureable" then I agree.

Yes, non-measureable, non-observable, or non-inferrable are synonyms to what I mean. However usually something beeing observable has a specific meaning in the context of normal QM, this is why I avoid that word. Since we are speculating in about a generalistion of QM here, the concept of observables are also generalised. As we already know the exact meaning of "what is an observable" is indeed one of the key problems.

But I think we mean the same thing.

Rovelli's partial/complete observables paper illustrates part of the problem, but he does not resolve it. The problem is that observables are unavoidably observer dependent, at the same time there is the vision to find an oberver invariant description of physical law.

IMO, this "problem" is exactly why there is no static observer invariant description of physical law. This is where the evolutionary perspective enter in my view. Note the - somewhat weak, but still - analogy here to Einsteins resistance to a nonstatic universe.

Your last sentence is very nice. Are you thinking along the lines of how the definition of the vacuum is dependent on the gravitational field?

Because in the paper he mentions that the infinitely many states should be associated to the vacuum(say of an atom). Then in gravity the degrees of freedom that are observable for each observer becomes dependent of the gravitational field.

I haven't had time to read th epaper yet, need to do some xmas preps this weekend, but I certainly use different metaphors that than, but my first impresison is that the paper is more to my liking than the average paper! this is why I wil lread it when get around to.

My thinking in shorts starts from the inside perspective, where each observer has what one can call an observable horizon, which I see as an "index", the size of this index is always bounded by the total complexity of the observer. This index can via certain processes grow or shrink. Now gravity (in my view) is the phenomenon by which an thirs observer sees two systems get attracted because their "probing into their environment", will yield a very slowly by steadily attraction. This attraction is in my view described to take place in an information space, where the distance measure is a kind of information divergence. So in my view, both space and gravity emerge as a result of two observers sharing the same environment are doing randow "moves"/"walks" and that the feedback from the environment brings them together. But the exact formalism from this is not yet mature. But the general vision is reasonably clear.

I think that reality is scale dependent.
...
Perhaps quantum gravity will provide a link between scales to form a scale independent reality.

I fully agree that the inferrable or measurable reality is scale dependent, where in my view "scaling" refers to the complexity(or mass) of the observer.

However my bet is that QG will provide the link to the scaling not in a independent way as in terms of realist (like I think Rovellis thinks), but rather in terms of an evolving description (more like SMolins evolving law). But for practical purposes as compare to present models, scale independent reality might be close.

I hope to read it during the wekend.

/Fredrik

 

[Fra]

where each observer has what one can call an observable horizon, which I see as an "index", the size of this index is always bounded by the total complexity of the observer. This index can via certain processes grow or shrink. Now gravity (in my view) is the phenomenon by which an thirs observer sees two systems get attracted because their "probing into their environment", will yield a very slowly by steadily attraction. This attraction is in my view described to take place in an information space, where the distance measure is a kind of information divergence. So in my view, both space and gravity emerge as a result of two observers sharing the same environment are doing randow "moves"/"walks" and that the feedback from the environment brings them together. But the exact formalism from this is not yet mature. But the general vision is reasonably clear.

I guess what I tried to say is that I think that the process responsible for the origin and accumulation of mass, is the same as the one responsible for gravitatinal acttraction. And this is moreover related to intertia - connected to "inerta of revising your opinon in the light of new evidence" when you see space as a special kind of information space.

/Fredrik

 

[Finbar]

My thinking in shorts starts from the inside perspective, where each observer has what one can call an observable horizon, which I see as an "index", the size of this index is always bounded by the total complexity of the observer. This index can via certain processes grow or shrink. Now gravity (in my view) is the phenomenon by which an thirs observer sees two systems get attracted because their "probing into their environment", will yield a very slowly by steadily attraction. This attraction is in my view described to take place in an information space, where the distance measure is a kind of information divergence. So in my view, both space and gravity emerge as a result of two observers sharing the same environment are doing randow "moves"/"walks" and that the feedback from the environment brings them together. But the exact formalism from this is not yet mature. But the general vision is reasonably clear.



I fully agree that the inferrable or measurable reality is scale dependent, where in my view "scaling" refers to the complexity(or mass) of the observer.

I hope to read it during the wekend.
/Fredrik

I like that you think the correct arena to describe gravity is information space. Your aware of this derivation of gravity theories from a maximal entropy? Papers by Padmanabhan? If you read "Entropy=information" this might gel with your ideas.


Let's be a bit more explicit by what we mean by "observer". Let's say if I have a sub system of the universe that is suitably complex such that if it interacts with some other energy/matter its entropy will increase. Then such an increase in entropy can be considered a "measurement". This may seem like a broad definition but I'm assuming by observer your not only referring to humans or animals but also say a particle detector or a photomultiplier but would want to rule out say a helium atom from being an observer?
So if we say that for something to be an observer we have to be able to assign it a finite entropy this may also be a good starting point for an "index" of its complexity.


Fra, out of interest are you an active researcher? Can you devote a large enough amount of time to the "exact formalism" of your ideas?

 

[Fra]

I like that you think the correct arena to describe gravity is information space. Your aware of this derivation of gravity theories from a maximal entropy? Papers by Padmanabhan? If you read "Entropy=information" this might gel with your ideas.

Yes, I'm aware of similar attempts. Padmanabhan has written some I think nice papers. Ariel Caticha also wants to infer GR from the rules of inference, mainly using max ent methods.

All this is close to this discussion! but ingredients are missing. The main objection to max ent methods are the ambigous construction of the entropy measure.

Let's be a bit more explicit by what we mean by "observer".
Let's say if I have a sub system of the universe that is suitably complex such that if it interacts with some other energy/matter its entropy will increase. Then such an increase in entropy can be considered a "measurement". This may seem like a broad definition but I'm assuming by observer your not only referring to humans or animals but also say a particle detector or a photomultiplier but would want to rule out say a helium atom from being an observer?

You'r right I'm not referring to just humans or biological systems. I do not rule out a helium atom as an observer. In my view, every subsystem qualifies as an observer of it's environment.

So observer for me, does not mean just "classical observer" a general quantum system will do.

Fra, out of interest are you an active researcher? Can you devote a large enough amount of time to the "exact formalism" of your ideas?

I'm not a professional researcher, no. I've got a fulltime job, that is quite remote from this.

So I guess I do not have enough time, but then does anyone ever? :)

I will let it take whatever amount of time it takes.

/Fredrik

 

[Finbar]

Yes, I'm aware of similar attempts. Padmanabhan has written some I think nice papers. Ariel Caticha also wants to infer GR from the rules of inference, mainly using max ent methods.

All this is close to this discussion! but ingredients are missing. The main objection to max ent methods are the ambigous construction of the entropy measure.



You'r right I'm not referring to just humans or biological systems. I do not rule out a helium atom as an observer. In my view, every subsystem qualifies as an observer of it's environment.

So observer for me, does not mean just "classical observer" a general quantum system will do.


/Fredrik

I disagree I don't think a helium atom can be considered an observer because it has no entropy. It has no way to absorb information about events. If you include all general systems as "observers" then there's no meaning to the word anymore.

 

[Fra]

I disagree I don't think a helium atom can be considered an observer because it has no entropy. It has no way to absorb information about events. If you include all general systems as "observers" then there's no meaning to the word anymore.

We my obviously be thinking differently here, and we probalby won't settle that in this thread :)

First of all as I see it, entropy is not an observer independent unique measure as I see it. Each observing system encodes it's own measure of the missing information about it's own environment, relative to itself - this is something completely different than some other observer evaluating an measure of entropy of this first system, relative to this second observer.

So the choice of this measure, is related to the choice of an observer.

I rather think that ther mere fact that there exists is more or less stable helium atom in a given environment, does contain non-trivial information about this environment. A helium atom is not stable in an arbitrary environment. Just increase the energy a bit and there are no atoms at all. An helium atom in my view clearly contains information, lots of it.

But in this quest it's not enough to use an entropy definition that has the current framwork as baggage, say von neumann entropy.

Instead, if we take one step back, and look what the purpose of entropy is, and how it's usually constructed. It's not a far stretch to picture that the introduction of this "unique" entropy measure is context dependent.

So about the comparasion to hidden variable you mentioned, I agree that it's not variable with hidden values, it's rather more like hidden complexions. Ie. hidden microstructures, rather than microstructures with hidden microstates. This is why the anzats of a probability distribution that is used in the bell-type arguments for hidden variable theories are invalid, becaseu the distribution itself is hidden.

I think of it so that the implication that two observers, that can't properly agree about the structure of the communication channel (could be space for example), is physical interactions, that deforms both views of the channel, ideally until there is a new equilibrium.

/Fredrik

 

[Finbar]

We my obviously be thinking differently here, and we probalby won't settle that in this thread :)

First of all as I see it, entropy is not an observer independent unique measure as I see it. Each observing system encodes it's own measure of the missing information about it's own environment, relative to itself - this is something completely different than some other observer evaluating an measure of entropy of this first system, relative to this second observer.

So the choice of this measure, is related to the choice of an observer.

I rather think that ther mere fact that there exists is more or less stable helium atom in a given environment, does contain non-trivial information about this environment. A helium atom is not stable in an arbitrary environment. Just increase the energy a bit and there are no atoms at all. An helium atom in my view clearly contains information, lots of it.

But in this quest it's not enough to use an entropy definition that has the current framwork as baggage, say von neumann entropy.

Instead, if we take one step back, and look what the purpose of entropy is, and how it's usually constructed. It's not a far stretch to picture that the introduction of this "unique" entropy measure is context dependent.

So about the comparasion to hidden variable you mentioned, I agree that it's not variable with hidden values, it's rather more like hidden complexions. Ie. hidden microstructures, rather than microstructures with hidden microstates. This is why the anzats of a probability distribution that is used in the bell-type arguments for hidden variable theories are invalid, becaseu the distribution itself is hidden.

I think of it so that the implication that two observers, that can't properly agree about the structure of the communication channel (could be space for example), is physical interactions, that deforms both views of the channel, ideally until there is a new equilibrium.

/Fredrik

Actually I think we agree. What you mean by a "helium atom" is the physical atom in a specific environment whereas I was talking about the helium atom that is described by QM and measured by an observer on a much lower energy scale.

So my point would be that we could define an observer at the scale of the helium atom but this system could not be the isolated helium atom but some combination of the atom and its environment. Then this system would have a finite entropy defined at the scale of a helium atom. So as I see it this is a scale/observer dependant entropy.

 

[Fra]

I was going to write that i haven't been able to read this yet, I've read the first two pages several nights in a row and then fell asleep :) But then I suddenly realized that this has been discussed already here:

"Wetterich's derivation of QM from classical "
-- https://www.physicsforums.com/showthread.php?t=322346

It is a thread by Marcus and it seems I even skimmed the paper back then and already commented.

Anyway, I hope to read it more properly, and in particutlar I'll focus on their generalisations.

I think their classification as emergent for a "isolated" subsystem in an infinite environment is pretty accurate and in line with how I see it. But this in itself indicates that QM is a special case. In my view, the environment per see, can only be read by another observer. So the missing ingredient (which I don't know if they solved since I keep falling asleep after page 2 every night!) is that the complexity of the observable environment, must be constrained by the complexity of the second observer.

This is why their initial characterisation to me, says that QM is emergent when you consider an (relatively speaking) infinitely massive observer that described a small subsystem.

This is why deformations are expected when the observer no longer can be considered to monitor ALL of the environment. As I see it, the implictaions of the deviation here is new and evolving interactions which is the main motivations for me holding this view. IT has the power to incorporate all the features of unification and symmetry breaking into the same framework if we only can describe this complexity scaling. This is pretty much exactly what I am personally working on, thought of course progress is slow.

If not sooner, I hope that xmas offer some time to get past page 2 on that darn paper.

/Fredrik