What does a local non-realistic theory look like?

[greypilgrim]

Hi,

Bell says there is no local realistic theory that reproduces the (experimentally verified) predictions of QM. That leaves us with three possibilities:

1. Non-local non-realistic (e.g. Copenhagen interpretation)
2. Non-local realistic (e.g. Bohmian mechanics)
3. Local non-realistic

How would a theory of the 3rd kind look like? I can't see how to possibly explain entanglement in a local way.

 

[lugita15]

Well, a lot of people don't view Copenhagen as nonlocal. They don't view wave function collapse as a phenomenon that genuinely violates locality, so they'd classify it in category 3.

 

[Demystifier]

3. Local non-realistic

How would a theory of the 3rd kind look like? I can't see how to possibly explain entanglement in a local way.

See e.g.
http://lanl.arxiv.org/abs/1112.2034 [Int. J. Quantum Inf. 10 (2012) 1241016]
In this interpretation only the observers are real, while the observed objects are not.

 

[Maui]

Hi,

Bell says there is no local realistic theory that reproduces the (experimentally verified) predictions of QM. That leaves us with three possibilities:

1. Non-local non-realistic (e.g. Copenhagen interpretation)
2. Non-local realistic (e.g. Bohmian mechanics)
3. Local non-realistic

How would a theory of the 3rd kind look like? I can't see how to possibly explain entanglement in a local way.

Brains are information processors, not particle processors, hence cutting down assumptions leaves us with the only solid knowledge available - that nobody has seen a particle, but always information about particles. If you were able to bypass the brain's ability to represent whatever is 'out there' in terms of particles(where they are applicable), you'd certainly see a very different 'world'. As someone remarked - it must be stranger than we can imagine with our mundane concepts(admittedly mathematical physics has broadened that horizon substantially and some bizarre experiments appear to confirm it in very categorical way). There is no problem with locality and realism as long as you realize that reality is not completely classical and cannot be regarded as such anymore(this has been so for a century but change is slow).

 

[DrChinese]

Hi,

Bell says there is no local realistic theory that reproduces the (experimentally verified) predictions of QM. That leaves us with three possibilities:

1. Non-local non-realistic (e.g. Copenhagen interpretation)
2. Non-local realistic (e.g. Bohmian mechanics)
3. Local non-realistic

How would a theory of the 3rd kind look like? I can't see how to possibly explain entanglement in a local way.

Welcome to PhysicsForums, greypilgrim!

There are a number of local non-realistic interpretations of quantum mechanics. Not all classify themselves as such, but here are a few:

1. Time symmetric class includes the retro-causal, transactional, relational blockworld and a few others.

http://arxiv.org/abs/0908.4348
http://arxiv.org/abs/0706.1232
http://www.npl.washington.edu/npl/int_rep/gat_80/

In the above, there is never any influence that propagates in excess of c. Each constituent interaction is local. However, because time direction is allowed to go either way, the net effect can appear non-local.

2. Many Worlds (MWI), although they often consider themselves both local and realistic. But obviously they can't be.

 

[kaplan]

Hi,

Bell says there is no local realistic theory that reproduces the (experimentally verified) predictions of QM. That leaves us with three possibilities:

1. Non-local non-realistic (e.g. Copenhagen interpretation)
2. Non-local realistic (e.g. Bohmian mechanics)
3. Local non-realistic

How would a theory of the 3rd kind look like? I can't see how to possibly explain entanglement in a local way.

Many Worlds - which is really just the statement that measuring devices should be quantum mechanical too. If you take that view, collapse is impossible (because QM time evolution is linear and unitary, while collapse is neither), and interactions are local.

What happens in MW (or just in QM applied to detectors too) is that when a detector measures a particle that's in a superposition of eigenstates, the detector ends up in a superposition of states. So there's no unique "reality" - instead, Schrodinger's cat states are allowed.

 

[morrobay]

Well, a lot of people don't view Copenhagen as nonlocal. They don't view wave function collapse as a phenomenon that genuinely violates locality, so they'd classify it in category 3.

Then from the Copenhagen interpretation can there be a local - non realistic model ( besides Many worlds and retro causality) that agrees with QM (non-linear) correlation predictions for Bell Inequality violations. Where locality and realism are not conjoined. Where instead of classical EPR/ Bell realism , a less restrictive (QM) definition of realism. And this model applied to modern experiments with entangled photons and emphasis on the cos² (θ -∅) coincidence rate in the explanation.

 

[jk22]

Imo a local nonrealistic theory would be a non deterministic one because it does not know the real outcome but only a probability.

For example Chsh operator is AB+AB'+...

In a local deterministic theory like Bell the knowledge of the angle and the variable induce the result, hence the above quantity would have as result

ab+ab'+...

Note that the result for the A operator leads to twice the same whereas a nondeterministic variable would give

ab+\-ab'+\-...

Such a theory can violate the Bell inequality but there are orher arguments that show that such one does not reproduce quantum predictions so that quantum is probably both nonlocal and nonrealistic.

 

[bohm2]

Welcome to PhysicsForums, greypilgrim!

There are a number of local non-realistic interpretations of quantum mechanics. Not all classify themselves as such, but here are a few:

1. Time symmetric class includes the retro-causal, transactional, relational blockworld and a few others.

http://arxiv.org/abs/0908.4348
http://arxiv.org/abs/0706.1232
http://www.npl.washington.edu/npl/int_rep/gat_80/

In the above, there is never any influence that propagates in excess of c. Each constituent interaction is local. However, because time direction is allowed to go either way, the net effect can appear non-local.

The authors of the relational blockworld in a very recent article considered their model realistic but Ψ-epistemic:Relational Blockworld: Providing a Realist Psi-Epistemic Account of Quantum Mechanics
http://www.ijqf.org/archives/2087

 

[bhobba]

Many Worlds - which is really just the statement that measuring devices should be quantum mechanical too.

That's not many worlds - there are many interpretations that treat measuring devices quantum mechanically eg Consistent Histories.

If you take that view, collapse is impossible (because QM time evolution is linear and unitary, while collapse is neither), and interactions are local.

Collapse is not part of the QM formalism - its an interpretational thing - some interpretations have it (eg GRW models) and some do not (eg BM). The formalism is however ambivalent.

What happens in MW (or just in QM applied to detectors too) is that when a detector measures a particle that's in a superposition of eigenstates, the detector ends up in a superposition of states.

That is incorrect. In Many Worlds what happens is, via decoherencre, a superposition is transformed into a mixed state ∑pi |bi><bi| where each |bi><bi| is interpreted as a world. Because its a mixed state it, by definition of what a mixed state is, is NOT in superposition. You may be thinking of what happens to the system as a whole - yes that remains in superposition, but the interpretation specifically states its not in superposition as far as 'world splitting' is concerned.

Thanks
Bill

 

[bhobba]

Then from the Copenhagen interpretation can there be a local - non realistic model

The problem with Copenhagen is it comes in quite a few variants. Here is the detail of the usual variant:
http://motls.blogspot.com.au/2011/05/copenhagen-interpretation-of-quantum.html

Note since the state in that interpretation represents subjective knowledge, even though it has collapse, it is of no concern since it simply occurs in the head of a theorist. It is local and non realistic.

Thanks
Bill

 

[bhobba]

Many Worlds (MWI), although they often consider themselves both local and realistic. But obviously they can't be.

That's true. It's resolution lies in the subtle difference between realism and counterfactual definiteness:
http://en.wikipedia.org/wiki/Principle_of_locality

Personally I believe such subtle differences aren't really worth much, being more philosophy than physics, and MW rejects realism but is local.

Thanks
Bill

 

[atyy]

There cannot be a local non-realistic theory, because realism is a precondition for locality - if we classify Bohmian Mechanics as nonlocal realistic - since that means we are using the notion of "local causalty" or "local explainability" when we say "local".

Also, although it is ok to say Copenhagen is nonlocal nonrealistic, it is better to say that it is nonrealistic, and that is why it fails to be local, since realism is a precondition for locality.

I'm not sure MWI is local, but it is not necessarily nonlocal by Bell's theorem, since by assuming all outcomes occur, a condition for the theorem to apply is violated.

As for Demystifier's soplisistic hidden variables, which I like very much, I wouldn't classify it local nonrealistic in the sense implied in the OP, where it seems that the options are those allowed by Bell's theorem after accepting that a Bell inequality has been violated at spacelike separation. I would say that in Demystifier's scenario, the Bell inequalities are not violated at spacelike separation, so the question of nonlocality due to a Bell inequality violation does not arise. A similar situation can arise in Copenhagen, because Copenhagen allows us to consider the other observer "not real", in which case there is also no violation of a Bell inequality at spacelike separation, and the issue of nonlocality simply does not arise.

One more point about Copenhagen. Although the wave function is not real, it is real FAPP, and collapse is real FAPP. So althouh Copenhagen is nonrealistic (without FAPP), it is also nonlocal realistic FAPP.

 

[zonde]

There cannot be a local non-realistic theory, because realism is a precondition for locality - if we classify Bohmian Mechanics as nonlocal realistic - since that means we are using the notion of "local causalty" or "local explainability" when we say "local".

I think the same - that there can't be local non-realistic theory, but I would like to add my opinion that there can't be non-local realistic theory just as well.
People imagine non-locality as FTL communication and that of course does not conflict with realism (while it conflicts with Special relativity). But the problem is that the way non-locality is described and how it is viewed in QM is not FTL communication but fundamental absence of distance in case of entangled particles. But such non-locality is completely new philosophical concept that is mutually exclusive with locality i.e. two concepts can't be fundamental at the same time. And without concept of distance (if non-locality is fundamental) there is no realism.

 

[morrobay]

There cannot be a local non-realistic theory, because realism is a precondition for locality

Bell inequalities are not violated at spacelike separation, so the question of nonlocality due to a Bell inequality violation does not arise.

Why is realism a precondition for locality ? Why can't they be independent ?
And in general why no Bell inequality violations at spacelike separation ? I thought that was an example of non locality ?

 

[zonde]

Why is realism a precondition for locality ? Why can't they be independent ?

Locality is described in terms of physical things that exist and interact independently from our ideas.

 

[stevendaryl]

I think both the words "local" and "realism" are ambiguous in their connotations. (Even if they are given unambiguous definitions, there are cases that are intuitively at odds with those definitions) There are at least two senses of "local": (1) Signal locality: No effects can propagate faster than the speed of light. (2) I don't know what the technical word is, but something like "local factorizability". Roughly speaking, the "state" of the universe is completely described by the states of little neighborhoods, together with information about how the neighborhoods fit together. The first is amenable to direct falsification through experiment, and so far seems completely solid. The second is more difficult to test experimentally (because it's much less specific in its observable consequences), but Bell's analysis shows that it can have testable consequences, nevertheless.

I think "realism" means several different things as well. I don't quite understand Dr. Chinese' claim that retrocausality (back-in-time influences) violate realism. But the two meanings that occur to me are: (1) The mathematical objects in the theory assumed to describe the state of an objective, observer-independent world. This is in contrast to a subjectivist theory, where the mathematical objects might represent an observer's information about the world (subjective probabilities, for instance), but don't directly represent anything objective. (2) Observations or experiments are understood as revealing something objective about the world. A many-worlds type theory can be understood as non-realistic in this sense, because an observation such as looking to see the result of a coin flip is not actually revealing anything objective about the world. You know before you look at the coin that the result will be heads in some branches of the multiverse and tails in some other branches. So observation is not revealing anything about the world (or collection of worlds), it's only "self-locating"; it tells which of the branches you find yourself in.

QM, in which the wave function is viewed purely epistemologically, as a description of an observer's knowledge, then that's not a realistic model according to sense (1), but it can be realistic according to sense (2). QM is local according to sense (1), but not according to sense (2) (because of entangled states).

MWI is realistic, according to sense (1), because it views the wave function as real, but it is not realistic according to sense (2). It is local according to sense (1), but not according to sense (2).

 

[atyy]

Why is realism a precondition for locality ? Why can't they be independent ?

By "locality", I meant "local causality" or "local explainability". A cause or an explanation must be real, by definition.

By using "local explainability" as a synonym for "local causality", I am using similar terminology as
http://arxiv.org/abs/0909.0015
http://arxiv.org/abs/1311.6852
http://arxiv.org/abs/1503.06413

And in general why no Bell inequality violations at spacelike separation ? I thought that was an example of non locality ?

Yes, if the Bell inequalities are violated at spacelike separation, then local causality is ruled out.

However, if one denies that the Bell inequalities are violated at spacelike separation, then Bell's theorem does not apply, and local causality is not ruled out.

So nonrealism can be used to prevent the violation of the Bell inequalities and thus prevent ruling out local causality by Bell's theorem. But if the Bell inequalities are violated at spacelike separation, then nonrealism cannot save local causality, since realism is a prerequisite for local causality.

Norsen makes this point in http://arxiv.org/abs/quant-ph/0607057, which is basically in line with the references from Hall, Cavalcanti and Lal, and Wiseman and Cavalcanti given above. (Norsen is entertaining, maybe too entertaining, and for "serious" physics perhaps I should only have given the other references, but Norsen is where I first read the point, and his argument is acknowledged by others, eg. http://arxiv.org/abs/0706.2661.)

 

[Jilang]

Norsen sure has no time for MWI, does he?
"The bottom line is the impossibility of any scientific basis for any (allegedly scientific) theory requiring the rejection of Perceptual Realism. MWI requires such a rejection, and hence cannot be taken seriously as a scientific theory"
Pulls no punches there!

 

[atyy]

Norsen sure has no time for MWI, does he?
"The bottom line is the impossibility of any scientific basis for any (allegedly scientific) theory requiring the rejection of Perceptual Realism. MWI requires such a rejection, and hence cannot be taken seriously as a scientific theory"
Pulls no punches there!

He does like some aspects of MWI: "Despite the harsh criticisms of MWI in the previous section, there is one aspect of the theory which I fully support: it accepts the existence of a single, objective, external world “out there” whose existence and identity is independent of anyone’s awareness (or, in the case of MWI, non-awareness) of it."

 

[zonde]

I think "realism" means several different things as well. I don't quite understand Dr. Chinese' claim that retrocausality (back-in-time influences) violate realism. But the two meanings that occur to me are: (1) The mathematical objects in the theory assumed to describe the state of an objective, observer-independent world. This is in contrast to a subjectivist theory, where the mathematical objects might represent an observer's information about the world (subjective probabilities, for instance), but don't directly represent anything objective. (2) Observations or experiments are understood as revealing something objective about the world. A many-worlds type theory can be understood as non-realistic in this sense, because an observation such as looking to see the result of a coin flip is not actually revealing anything objective about the world. You know before you look at the coin that the result will be heads in some branches of the multiverse and tails in some other branches. So observation is not revealing anything about the world (or collection of worlds), it's only "self-locating"; it tells which of the branches you find yourself in.

Realism means that there is objective reality independent of our models of reality. One important consequence that we can draw from this statement is that if we have two mutually exclusive descriptions of reality then they both can't be right.
Your option (1) is rather fundamental vs phenomenological description and about option (2) - realism can't say much about how objective are observations but if two are contradictory than at least one is wrong (not ojective).

 

[stevendaryl]

Realism means that there is objective reality independent of our models of reality.

I don't think that's a very useful definition. Are there any scientists who don't believe in realism in that sense?

 

[zonde]

I don't think that's a very useful definition. Are there any scientists who don't believe in realism in that sense?

I would say it's very good philosophical statement if significant majority agrees with it because in philosophy there are not too many ways how to convince those thinking otherwise.
But statement might not be very useful if we can't draw useful conclusions from it, say like "reality is consistent". And then maybe we can draw such conclusion because "inconsistency" is specific property of our models of reality as a side effect of flexibility of our thinking process.

 

[ddd123]

I don't think that's a very useful definition. Are there any scientists who don't believe in realism in that sense?

Try this: https://en.wikipedia.org/wiki/Constructivist_epistemology

 

[stevendaryl]

I would say it's very good philosophical statement if significant majority agrees with it because in philosophy there are not too many ways how to convince those thinking otherwise.
But statement might not be very useful if we can't draw useful conclusions from it, say like "reality is consistent". And then maybe we can draw such conclusion because "inconsistency" is specific property of our models of reality as a side effect of flexibility of our thinking process.

I don't think that "reality is consistent" is a very meaningful thing to say. "Consistency" is a property of statements, not of the world. Specifically, a collection of statements is said to be inconsistent if you can deduce a contradiction from them. A collection of statements is said to be consistent if it's impossible to derive a contradiction from them. This is a property of theories, not a property of objects, real or otherwise.

You can certainly say of a theory that if it's inconsistent, then it's wrong. Or you can say of a pair of theories that if they are inconsistent in an unresolvable way, then one or the other (or both) is wrong. But to say that two theories are inconsistent in an unresolvable way is pretty complicated, it seems to me. The problem is that the same sentence might mean different things in different theories, and so the theories would be superficially inconsistent, but they would become consistent by making distinctions. For example, you can prove in real analysis that there is no square-root of -1, but in complex analysis, there is a square-root of -1. Those two theories are not inconsistent, you can view real analysis as a subset of complex analysis. There is no real square root of -1, but there is an imaginary square root.

For two theories of physics, I don't know how you would ever show that they are unresolvably mutually inconsistent, unless they refer to observations. If one theory says that such and such will never be observed, and the other theory says that it will, then they are inconsistent. But that doesn't seem to be a matter of realism.

I don't want to get argumentative about it, but I do think that your definition of "realism" is pretty useless for physics.

 

[stevendaryl]

Try this: https://en.wikipedia.org/wiki/Constructivist_epistemology

I disagree. To quote from that article:
"According to constructivists the world is independent of human minds, but knowledge of the world is always a human and social construction."

So that is not a "non-realistic" theory in the sense of zonde. It's non-realistic in my sense.

 

[ddd123]

I think that's what zonde is saying. Philosophers tend to read "objective" as "corresponding to representation", which comes from the etymology, rather than the vague meaning it has in laymen's terms as "really existing" or whatever. Which means, "objective" has to do with statements about the world and not just the world in general, and statements implicitly carry models with them so...

 

[bhobba]

Philosophers tend to read "objective" as "corresponding to representation"

I think philosophers argue about such basic terms a lot, and reach all sorts of positions, so much so there is no 'tend' to anything in what they semantically dissect. But physics divorced itself from philosophy yonks ago and its not really part of science, nor is it really what we discuss on this forum.

For what objective means simply take what a dictionary says:
'Existing independent of or external to the mind; actual or real'

For the purposes of physics nothing hard about it.

For Bells theorem its fairly easy what is meant - the subtlety comes with the concept of counter-factual definiteness and reality - you can do an internet search to understand it. But I suspect most, including me, would likely say its not worth worrying about.

Thanks
Bill

 

[jk22]

In some sense quantum is a prediction of reality but itself is not realistic like a pre-reality.

But some elements are real : the eigenvalues are measurement results and others don't : the probabilities are 'ideal' ie after an infinite number of measurement it should converge to them, which is not feasible in reality ?

 

[bhobba]

In some sense quantum is a prediction of reality but itself is not realistic like a pre-reality.

We have all sorts of interpretations - real ones, many minds, many worlds, many interacting worlds, I have read about all sorts. QM is pretty ambivalent to whatever view you want to take.

Thanks
Bill

 

[stevendaryl]

We have all sorts of interpretations - real ones, many minds, many worlds, many interacting worlds, I have read about all sorts. QM is pretty ambivalent to whatever view you want to take

Except that the most straightforward, intuitive view--that (1) the universe is at each moment in some state or other, and (2) one part of the universe is only affected by events in the backward light cone, and (3) measurements just reveal local information about the pre-existing state of the universe--is apparently ruled out by QM.

 

[jk22]

Except that the most straightforward, intuitive view--that (1) the universe is at each moment in some state or other, and (2) one part of the universe is only affected by events in the backward light cone, and (3) measurements just reveal local information about the pre-existing state of the universe--is apparently ruled out by QM.

For 3 we can escape with the fact that the local values has to interact with the apparatus hence when the knowledge of the preexisting element lambda and the measurement setting theta does Not determine the outcome. Indeed a probabilistic theory would violate Bell's inequality but the problem is that quantum mechanics does not allow all values that such a theory would give and hence there are global conditions

 

[stevendaryl]

Indeed a probabilistic theory would violate Bell's inequality but the problem is that quantum mechanics does not allow all values that such a theory would give and hence there are global conditions

What do you mean by saying a probabilistic theory would violate Bell's inequality? If by "probabilistic theory" you mean a local, stochastic theory, then I don't believe that's true.

 

[zonde]

Except that the most straightforward, intuitive view--that (1) the universe is at each moment in some state or other, and (2) one part of the universe is only affected by events in the backward light cone, and (3) measurements just reveal local information about the pre-existing state of the universe--is apparently ruled out by QM.

And here you contradict realism. QM can't rule out some particular view of reality. Only experiments and observations can do that.
Probably you wanted to say that experiments almost rule out these options.

 

[stevendaryl]

And here you contradict realism. QM can't rule out some particular view of reality. Only experiments and observations can do that. Probably you wanted to say that experiments almost rule out these options.

No, I don't think so. QM is a theory, and that theory is inconsistent with a particular view of the universe.