Why is superdeterminism not the universally accepted explanation of nonlocality?

[DrChinese]

Didn't Bell himself say the contrary in his famous superdeterminism quote? His theorem shows that no local realistic theory can reproduce all predictions of quantum mechanics under the assumption of free choice of the observables to be measured (please object if you think this is wrong).

Sure, he did. I don't take that very seriously though, and I doubt he did either (I think of it as a joke). As I have said, you can also say that God intervenes to make it look like Bell test results reflect reality. My point being that it takes a candidate theory to have anything meaningful to discuss. There aren't any.

And as to your comments about initial conditions: I repeat that this does NOT produce an answer for Bell test results. You need completely new physics of a type never seen before. Please do not wave your hands on this point. If you want to explain Bell test results outside QM, you need to explain the mechanism or otherwise give me something to bite my teeth into. So okay, what is it? (Because anything you throw I will falsify experimentally.)

 

[Demystifier]

In fact, I think there was an old thread where Demystifier set out to show that almost no major interpretation of quantum mechanics would actually interpret delayed choice as changing the past.

Yes, here it is for the case someone is interested:
https://www.physicsforums.com/showthread.php?t=402497
Pay particular attention to posts #1 and #21.

 

[Demystifier]

Thanks for the reply. But you're doing it also. I really don't know what you guys mean by superdeterminism.

OK, here is a simple example of superdeterminism which does not involve quantum mechanics.

Assume that John runs a lottery business. And assume that his son Bob wins the main prize every time.

How to explain that? Well, the most obvious explanation is that John cheats, in order to make his son pick up the big money. This is what most people believe, including police. Confronted with the police accusations, he argues with police as follows:

John: No, I don't cheat. It is not my guilt that my son Bob wins every time.
Police: Then how do you explain that your son wins every time?
John: I don't now, maybe he is just lucky.
Police: Come on, nobody can be that lucky to win every fu****g time.
John: Maybe God wants him to win every time. So it's God, it has nothing to do with me.
Police: God cannot have anything to do with it. That's because lottery is a deterministic process, i.e., the winning numbers are determined by the details of initial conditions. God cannot change the initial conditions.
John: OK, I admit that lottery is deterministic, but maybe it's not only that. Maybe lottery is not only deterministic, but superdeterministic.
Police: What do you mean by "superdeterministic"?
John: Well, God wants my son to win every time. And yet, He wants to obey his own deterministic laws. So what does He do? He chooses the initial conditions in a very special way to make sure that my son will win every time. That's superdeterminism; deterministic laws plus very special initial conditions chosen by God to create one additional rule that otherwise could not be explained by the deterministic laws alone. In this case, the additional rule is: my son will win every time.
Police: So, is that supposed to prove that you are innocent?
John: Of course. I certainly don't have this power to fine tune all the initial conditions in the Universe. Only God can do that.

---------

So, would you buy this superdeterminism argument and reject the idea that John is cheating? Likewise, would you buy superdeterminism as explanation of quantum correlations and reject the idea that nature is nonlocal?

 

[IttyBittyBit]

Demystifier, I don't see how your analogy is valid at all. No one mentioned anything about 'God'. How would superdeterminism be like rejecting the idea that John cheats? If anything, it means accepting that he cheats, and acknowledging that maybe the Universe cheats all the time and this is the natural order of things.

There are certain facts we have to accept. One fact is that the Universe certainly seems to be non-local. Thus, any deterministic theory that attempted to model the Universe accurately would have to be superdeterministic. This is one of the implications of Bell's theorem. That's it. Bell's theorem is not a proof that superdeterminism is false, any more than the EPR paradox is proof that hidden variables must exist, or the twin paradox is proof that relativity is impossible, or Schrodinger's cat is proof that wavefunction collapse does not apply to macroscale objects.

Everyone in this thread should acknowledge that there is, at this point in time, no scientific reason to reject superdeterminism. It's all just speculation and intuition, and we all know how well that goes. As I said in my first reply, first we need to develop a set of superdeterministic theories (there are currently none), then test them against experiment, and compare them with non-deterministic theories. Until then, any thoughts on the matter are premature.

 

[juanrga]

Everyone in this thread should acknowledge that there is, at this point in time, no scientific reason to reject superdeterminism. It's all just speculation and intuition, and we all know how well that goes. As I said in my first reply, first we need to develop a set of superdeterministic theories (there are currently none), then test them against experiment, and compare them with non-deterministic theories. Until then, any thoughts on the matter are premature.

Determinism and variants are clearly in contrast with scientific method. Let's leave determinism and variants for philosophers.

 

[Demystifier]

Demystifier, I don't see how your analogy is valid at all. No one mentioned anything about 'God'. How would superdeterminism be like rejecting the idea that John cheats? If anything, it means accepting that he cheats, and acknowledging that maybe the Universe cheats all the time and this is the natural order of things.

Thomas asked us to explain what do we mean by superdeterminism, and I have explained what do I (and not only I) mean by that. Obviously you mean something else by this term, so it would be nice if you could give your explanation as well. In particular, you should explain the difference between determinism and superdeterminism.

Of course, it was not necessary to use God in my explanation, but the concept of "God" in some very general sense seems suitable to talk about the cause of initial conditions. For what else might be the cause of initial conditions?

 

[lugita15]

Yes, here it is for the case someone is interested:
https://www.physicsforums.com/showthread.php?t=402497
Pay particular attention to posts #1 and #21.

Thanks, Demystifier. And while I have your attention, do you have any thoughts on the Zurek paper I referenced earlier in this thread? Here it is again:
http://arxiv.org/abs/quant-ph/0405161
If decoherence is sufficient to derive the Born rule, why can't this derivation be carried over into Bohmian mechanics?

 

[lugita15]

Of course, it was not necessary to use God in my explanation, but the concept of "God" in some very general sense seems suitable to talk about the cause of initial conditions. For what else might be the cause of initial conditions?

To be fair, a superdeterminist need not believe that the special initial conditions that give rise to conspiratorial behavior are due to supernatural agency or chance. Instead, you can have a theory where at some time in the past all the particles in the universe (or their ancestors) interacted with each other in such a way that they gave each other the right initial conditions. In other words, they literally met and "agreed" on the conspiracy. So then superdeterminists would have to invent some form of local particle interaction which gives the particles the special initial states necessary to produce nonlocal correlation later on. And that seems like a really hard task.

 

[lugita15]

Demystifier, I don't see how your analogy is valid at all. No one mentioned anything about 'God'. How would superdeterminism be like rejecting the idea that John cheats? If anything, it means accepting that he cheats, and acknowledging that maybe the Universe cheats all the time and this is the natural order of things.

Yes, "the Universe cheats" is actually a pretty good description of superdeterminism

There are certain facts we have to accept. One fact is that the Universe certainly seems to be non-local. Thus, any deterministic theory that attempted to model the Universe accurately would have to be superdeterministic. This is one of the implications of Bell's theorem. That's it. Bell's theorem is not a proof that superdeterminism is false, any more than the EPR paradox is proof that hidden variables must exist, or the twin paradox is proof that relativity is impossible, or Schrodinger's cat is proof that wavefunction collapse does not apply to macroscale objects.

Bell's theorem implies that any local deterministic theory which reproduces the predictions of quantum mechanics must be superdeterministic. But you can have a nonlocal deterministic theory like Bohmian mechanics, and that need not be superdeterministic. But you're right that Bell's theorem does not disprove superdeterminism.

Everyone in this thread should acknowledge that there is, at this point in time, no scientific reason to reject superdeterminism.

Science does not accept all theories until they are rejected by experiments. In science we do not accept claims unless we have good experimental reason to do so. Currently we have a nondeterministic theory, quantum mechanics, which is extremely accurate in its predictions. Thus the burden of proof is on the determinist to demonstrate his thesis, not on the scientific community to disprove it.

 

[IttyBittyBit]

Bell's theorem implies that any local deterministic theory which reproduces the predictions of quantum mechanics...

Yes that's why I said models our own Universe.

Science does not accept all theories until they are rejected by experiments.

Actually, science does not accept anything. Theories are never proven; only rejected based on evidence.

(Some might go further and even dispute that, saying that the best we can do is assign likelihoods to theories.)

Thus the burden of proof is on the determinist to demonstrate his thesis, not on the scientific community to disprove it.

This viewpoint is often thrown around and represents a confusion between proving the existence of something (which is the subject of logic and mathematics, not science - though logic and mathematics are tools that are often used in science), and evaluating the likeliness of a hypothesis (which is the subject of science).

Quite contrary to your statement, in science all theories (all falsifiable theories, that is) are given equal consideration until evidence comes along that disproves any number of them. This evidence might be direct or indirect. Science resembles a courtroom more than anything.

That being said, I repeat that superdeterminism by itself is not a scientific theory, because it is unfalsifiable. However, there are strong indications that it might be possible to construct a valid theory of the Universe that is superdeterministic in nature.

 

[kith]

If you want to explain Bell test results outside QM, you need to explain the mechanism or otherwise give me something to bite my teeth into.

You really have a biting reflex, when it comes to Bell's theorem, don't you? ;-) Just to be clear: I don't question its validity. We are simply talking about different aspects of the discussion. Amongst others, the following two arguments against superdeterminism occurred during the discussion.

First, that there is no local realistic superdeterministic (LRSD) model (yet). This is what you are talking about.

Second, people object to LRSD theories in general, because they have to be conspirational. That is they have to look nonlocal to (predeterminded) experimenters, in spite of their real laws beeing local. By choosing the wording "conspiracy", it is suggested that such theories are absurd or very unlikely to be true.

My point was to question this view by noting, that all non-conspirational LRSD theories would be incompatible with quantum mechanics, so the anthropic principle could be used to explain the unlikely "conspiracy" (see my last posts).

So my bottom line is: the "conspiracy" part alone is not an argument against LRSD theories, if one acknowledges the anthropic principle (as many people do in cosmology). This is indepent of the construction of actual superdeterministic theories, which of course has to be done in order to discuss superdeterminism in a more scientific way.

And as to your comments about initial conditions: I repeat that this does NOT produce an answer for Bell test results.

This part of your posts I still don't get. Do you agree with Demystifier's definition of superdeterminism as determinism+special initial conditions in post #9? And do you agree with Bell himself that his theorem relies on the assumption that the experimenters can choose the observables they measure? Given the deterministic laws, what else can determine the outcome of Bell tests if not the special initial conditions?

 

[lugita15]

Yes that's why I said models our own Universe.

What I was quibbling with was your assertion that any deterministic theory must be superdeterministic. But all Bell says is any LOCAL deterministic theory must be superdeterministic.

Actually, science does not accept anything. Theories are never proven; only rejected based on evidence.

(Some might go further and even dispute that, saying that the best we can do is assign likelihoods to theories.)



This viewpoint is often thrown around and represents a confusion between proving the existence of something (which is the subject of logic and mathematics, not science - though logic and mathematics are tools that are often used in science), and evaluating the likeliness of a hypothesis (which is the subject of science).

Quite contrary to your statement, in science all theories (all falsifiable theories, that is) are given equal consideration until evidence comes along that disproves any number of them. This evidence might be direct or indirect. Science resembles a courtroom more than anything.

OK, I guess we just have a difference of opinion about the philosophy of science.

 

[lugita15]

My point was to question this view by noting, that all non-conspirational LRSD theories would be incompatible with quantum mechanics, so the anthropic principle could be used to explain the unlikely "conspiracy" (see my last posts).

Let's get some terminology straight. By definition, a superdeterministic theory is a deterministic theory which is conspiratorial. So "non-conspiratorial LRSD" doesn't make any sense. I think what you mean to say is "all local deterministic theories which are not superdeterministic are would be incompatible wiht quantum mechanics".

 

[kith]

Let's get some terminology straight. By definition, a superdeterministic theory is a deterministic theory which is conspiratorial. So "non-conspiratorial LRSD" doesn't make any sense. I think what you mean to say is "all local deterministic theories which are not superdeterministic are would be incompatible wiht quantum mechanics".

Yes, you can also put it like that.

Personally, I prefer a broader definition which includes the classical case. There has already been much philosophical debate before quantum mechanics and Bell's position (as expressed in the BBC quote) is hard to distinguish from Laplacian determinism.

 

[IttyBittyBit]

OK, I guess we just have a difference of opinion about the philosophy of science.

Then you have a difference of opinion with what is usually taken to be the definition of science.

Take a good long read of this page: http://undsci.berkeley.edu/teaching/misconceptions.php#b14

This is just basic stuff really.

(In particular, on that page see 'MISCONCEPTION: Science proves ideas' and 'MISCONCEPTION: The job of a scientist is to find support for his or her hypotheses'. In regard to: 'MISCONCEPTION: Science can only disprove ideas' that is what I was referring to by stating that some would say that the best we can do is assign likelihoods to hypotheses.)

 

[DrChinese]

... Bell's position (as expressed in the BBC quote) is hard to distinguish from Laplacian determinism.

And just to be clear, Laplacian determinism is NOTHING like superdeterminism.

 

[DrChinese]

1. You really have a biting reflex, when it comes to Bell's theorem, don't you? ;-)

2. First, that there is no local realistic superdeterministic (LRSD) model (yet). This is what you are talking about.

3.Do you agree with Demystifier's definition of superdeterminism as determinism+special initial conditions in post #9? And do you agree with Bell himself that his theorem relies on the assumption that the experimenters can choose the observables they measure? Given the deterministic laws, what else can determine the outcome of Bell tests if not the special initial conditions?

1. You got me.

2. Yes, I agree.

3. No, and I am not necessarily disagreeing with Demystifier on this. Superdeterminism requires both specific initial conditions AND new laws of physics. This is necessary for the kind of cheating that Demystifier mentions. The reason is that there are no known initial conditions that can lead to Bell test results. As I have said many times, there just isn't enough information carried locally to support this. So there must be something additional, and it must be of a type of physics not currently contemplated.

Just stating that initial conditions does it is not sufficient by a long shot. (You may as well assert that the true speed of light is 1 kph too, but initial conditions cause every test made to yield c as the answer.)

 

[kith]

And just to be clear, Laplacian determinism is NOTHING like superdeterminism.

On a conceptual level, Laplacian determinism (LD) and conspirational superdeterminism (CSD) are not very different.

The philosophical question which steered the debate about determinism is the question of free will. If the laws are deterministic, does the experimenter still have the freedom to choose initial conditions in an experiment or are they predetermined by the past? I find it only reasonable to have an umbrella term for theories, which have deterministic laws and deny this freedom. Since these theories are "more" deterministic than theories which leave the question open, the wording "super"-deterministic seems adequate. Bell's BBC quote suggests that this is what led him to coin the wording.

So in this sense, both LD and CSD are superdeterministic. And another theory which needs to fit under the umbrella is Bohmian mechanics. So instead of calling these theories "Laplacian determinism", "superdeterminism" and "Bohmian mechanics without free choice of initial conditions" and not having an umbrella term, I think it is more logical (and also in the spirit of Bell) to call the umbrella "superdeterminism" and refer to local realistic hidden variable theories by something like "conspirational superdeterminism".

The main difference between LD and CSD is of course, that CSD is only a hypothetical theory and no one knows how its laws would look like. With your second post, I mostly agree. The question is whether one wants to discuss general philosophical implications or hard science. Since there is no CSD theory available, the second one is quite futile.

 

[lugita15]

The philosophical question which steered the debate about determinism is the question of free will. If the laws are deterministic, does the experimenter still have the freedom to choose initial conditions in an experiment or are they predetermined by the past? I find it only reasonable to have an umbrella term for theories, which have deterministic laws and deny this freedom. Since these theories are "more" deterministic than theories which leave the question open, the wording "super"-deterministic seems adequate. Bell's BBC quote suggests that this is what led him to coin the wording.

What you call by the term "superdeterminism" is what everybody else just calls determinism. If you think the experimenter has free will, then obviously you're not a determinist. As I think you already know, what you call "conspiratorial superdeterminism" everyone else just calls superdeterminism. It's important for everyone to use the same terms; otherwise you have situations like ThomasT thinking there is no difference between determinism and superdeterminism as they are conventionally used.

 

[lugita15]

The reason is that there are no known initial conditions that can lead to Bell test results. As I have said many times, there just isn't enough information carried locally to support this. So there must be something additional, and it must be of a type of physics not currently contemplated.

I think this assertion's a bit too strong. You can obviously have a conspiracy where all the particles in the universe carry literally all information about all other particles in the universe, and use this information to do exactly what they need to do to make Bell's inequality appear violated, and never use this information except when dealing with quantum entanglement. That's why it's called ad hoc.

 

[kith]

What you call by the term "superdeterminism" is what everybody else just calls determinism. If you think the experimenter has free will, then obviously you're not a determinist. As I think you already know, what you call "conspiratorial superdeterminism" everyone else just calls superdeterminism. It's important for everyone to use the same terms; otherwise you have situations like ThomasT thinking there is no difference between determinism and superdeterminism as they are conventionally used.

Yes, you are probably right. But I think a big part of the confusion comes from inadequately naming things.

 

[ThomasT]

Ok, thanks for the replies, I understand what's meant by superdeterminism now.

Paraphrasing lugita15 (from post #81), the assumption of superdeterminism can be used to construct a scenario in which violation of BI's makes it seem "as if local determinism is false even though it is really true".

However, imho, the reason that BIs are violated has to do with constraints encoded in clearly explicitly local realistic models of quantum entanglement. These constraints, necessary for a model to be identified as local or local realistic, have the net effect wrt simplistic models (such as that described by Herbert) of requiring light to behave in an uncharacteristic way. That is, instead of expecting the correlation between angular difference and coincidental detection to be linear, the correlation would more reasonably be expected to approximate the nonlinear correlation that QM predicts.

Wrt more sophisticated LR models (still clearly local or local realistic models -- ie., Bell-type formulations), which predict a nonlinear correlation, the QM-predicted angular dependency is more closely approximated, but the range predicted by those LR models is reduced, and the data points skewed.

From this, and consideration of the way light would reasonably be expected to behave in a local deterministic universe, it seems to me that there might be a simpler solution to the problem than to invoke superdeterminism.

That is, I don't believe that either locality or nonlocality have been definitively shown to be false, and I think it's quite possible that violation of BIs might be making it seem "as if local determinism is false" even if it's true, and that this is due to constraints encoded in the LR formalism, and that violations of inequalities based on those constraints don't inform as to whether locality or nonlocality is the case.

Entanglement stats result from measuring, via global measurement parameters, relationships between entangled entities. Experimental preparations are designed to produce these relationships via local transmissions and interactions. So, on the face of it, there's nothing particularly mysterious about entanglement correlations except for the fact that explicit Bell-type LR models are a bit off. Hence, imo, no need for superdeterministic theories.

And I agree with what DrC (et al.) said about the problem(s) with superdeterminism.

 

[lugita15]

ThomasT, let me just tell you that Bell's theorem rigorously proves that it is IMPOSSIBLE for any local deterministic theory to reproduce the predictions of quantum mechanics without being superdeterministic, so you are badly mistaken if you think a more "sophisticated" local realist model which is not superdeterministic can succeed. It is not an arbitrary "assumption" or "constraint" that correlations must be linear, in the sense that the mismatch at 60 cannot be greater than twice the mismatch at 30. It is a conclusion of an argument, not an assumption, and that argument is given in Herbert's article. Here is my detailed outline of the argument from another thread:

1. If you have an unpolarized photon, and you put it through a detector, it will have a 50-50 chance of going through, regardless of the angle it's oriented at.
2. A local realist would say that the photon doesn't just randomly go through or not go through the detector oriented at an angle θ; he would say that each unpolarized photon has its own function P(θ) which is guiding it's behavior: it goes through if P(θ)=1 and it doesn't go through it P(θ)=0.
3. Unfortunately, for any given unpolarized photon we can only find out one value of P(θ), because after we send it through a detector and it successfully goes through, it will now be polarized in the direction of the detector and it will "forget" the function P(θ).
4. If you have a pair of entangled photons and you put one of them through a detector, it will have a 50-50 chance of going through, regardless of the angle it's oriented at, just like an unpolarized photon.
5. Just as above, the local realist would say that the photon is acting according to some function P(θ) which tells it what to do.
6. If you have a pair of entangled photons and you put both of them through detectors that are turned to the same angle, then they will either both go through or both not go through.
7. Since the local realist does not believe that the two photons can coordinate their behavior by communicating instantaneously, he concludes the reason they're doing the same thing at the same angle is that they're both using the same function P(θ).
8. He is in a better position than he was before, because now he can find out the values of the function P(θ) at two different angles, by putting one photon through one angle and the other photon through a different angle.
9. If the entangled photons are put through detectors 30° apart, they have 25% chance of not matching.
10. The local realist concludes that for any angle θ, the probability that P(θ±30°)≠P(θ) is 25%, or to put it another way the probability that P(θ±30°)=P(θ) is 75%.
11. So 75% of the time, P(-30)=P(0), and 75% of the time P(0)=P(30), so there's no way that P(-30)≠P(30) 75% of the time.
12. Yet when the entangled photons are put through detector 60°, they have a 75% chance of not matching, so the local realist is very confused.

The only way around this, and it's a somewhat far-fetched way, is to assume a conspiracy involving the particles under observation and the (unrelated) particles controlling the measurement decisions, i.e. superdeterminism.

 

[ThomasT]

ThomasT, let me just tell you that Bell's theorem rigorously proves that it is IMPOSSIBLE for any local deterministic theory to reproduce the predictions of quantum mechanics without being superdeterministic, so you are badly mistaken if you think a more "sophisticated" local realist model which is not superdeterministic can succeed.

I agree with you. In fact, I think I said as much. I don't think I said anything indicating that I believe that a clearly LR (ie., Bell-type) model of quantum entanglement could ever be viable. So I don't know why you might think that I might think that any clearly LR model which is not superdeterministic might succeed.

It is not an arbitrary "assumption" or "constraint" that correlations must be linear, in the sense that the mismatch at 60 cannot be greater than twice the mismatch at 30.

I didn't say the constaints encoded in LR models are arbitrary. It seems clear enough to me that Bell put a lot of thought into his work on this. He made the reasons for his formulation quite clear. And that formulation stands unrefuted as far as I know.

But it also seems clear to me that, given what's known about the behavior of light, it's not reasonable to expect the correlation between angular difference and coincidental detection to be a linear one.

It is a conclusion of an argument, not an assumption, and that argument is given in Herbert's article.

I didn't say it (a linear correlation) was an assumption. Just that, even assuming a local deterministic world, we wouldn't expect the correlation to be linear.

So, the problem for the local realist's was to either come up with Bell-type (ie., clearly local realistic) formulations that described a nonlinear correlation (which has been done, but with apparently insurmountable limitations on the range and distribution of data point predictions wrt those models), or to come up with non-Bell-type formulations (which has also been done, but which are generally disregarded because they're just not clearly local realistic).

The only way around this, and it's a somewhat far-fetched way, is to assume a conspiracy involving the particles under observation and the (unrelated) particles controlling the measurement decisions, i.e. superdeterminism.

I don't see this as a solution. In fact, I think it detracts from what I think should be the main focus which has to do with what makes an LR formulation, and associated inequality, clearly local realistic ... and why, given those constraints, a viable LR model of quantum entanglement is impossible -- even if our universe is local deterministic.

 

[lugita15]

But it also seems clear to me that, given what's known about the behavior of light, it's not reasonable to expect the correlation between angular difference and coincidental detection to be a linear one.

And what is the conclusion to draw from that? It is that if the universe is not superdeterministic, it cannot be local deterministic. This is what was rigorously proved by Bell. Do you disagree with this conclusion?

I didn't say it (a linear correlation) was an assumption. Just that, even assuming a local deterministic world, we wouldn't expect the correlation to be linear.

Bell proved that in a local deterministic world which is not superdeterministic, the correlation MUST be linear, in the sense that the mismatch at 2θ cannot be greater than the mismatch at θ.

So, the problem for the local realist's was to either come up with Bell-type (ie., clearly local realistic) formulations that described a nonlinear correlation (which has been done, but with apparently insurmountable limitations on the range and distribution of data point predictions wrt those models), or to come up with non-Bell-type formulations (which has also been done, but which are generally disregarded because they're just not clearly local realistic).

I have no idea what you mean by Bell-type formulation and non-Bell-type formulation.

I don't see this as a solution. In fact, I think it detracts from what I think should be the main focus which has to do with what makes an LR formulation, and associated inequality, clearly local realistic ... and why, given those constraints, a viable LR model of quantum entanglement is impossible -- even if our universe is local deterministic.

If our universe was local deterministic and not superdeterministic, then Bell's inequality would be satisfied, a viable LR model would be possible, Bell's inequality would be confirmed by experiment, and quantum mechanics would be disproven. All of that follows from the argument I quoted.

 

[Demystifier]

Thanks, Demystifier. And while I have your attention, do you have any thoughts on the Zurek paper I referenced earlier in this thread? Here it is again:
http://arxiv.org/abs/quant-ph/0405161
If decoherence is sufficient to derive the Born rule, why can't this derivation be carried over into Bohmian mechanics?

Well, I wouldn't say decoherence is sufficient. The actual derivation in this paper contains some additional assumptions. Besides, even though this paper is well known in the decoherence community, it is not generally accepted in this community that this paper definitely solves the problem of the origin of Born rule.

 

[Demystifier]

Bell's theorem implies that any local deterministic theory which reproduces the predictions of quantum mechanics must be superdeterministic.

I wouldn't say so, because superdeterminism is not the only way to make the quantum world local:
https://www.physicsforums.com/blog.php?b=3622

 

[ThomasT]

And what is the conclusion to draw from that?

That maybe the standard LR formalism and BI violations don't inform wrt the reality underlying instrumental behavior. That maybe there is something about the Bell-type LR formal constraints that makes viable Bell-type LR models of entanglement impossible -- even if nature is local deterministic.

That is, there's a couple of hundred years of optics detailing the characteristic behavior of light in experimental situations similar to Bell test setups (eg., polariscopic setups), and what the simplest realization of Bell's theorem (eg., Herbert's popularization) suggests is that, in a local deterministic universe, light should be expected to behave contrary to that.

This makes sense if you assume that Bell-type LR constraints and BI violations inform wrt the reality underlying instrumental behavior. But, afaik, there's no reason to assume that.

It is that if the universe is not superdeterministic, it cannot be local deterministic.

Both are just assumptions that, afaik, can't be falsified. There are, afaik, some good reasons to assume that the universe is evolving determinstically, but no good reasons to assume superdeterminism.

This is what was rigorously proved by Bell. Do you disagree with this conclusion?

Yes, I disagree with that conclusion. What, imo, was rigorously proved by Bell was that, wrt quantum entanglement setups, the supplementation of QM formalism with certain "local realistic hidden variable" constraints produced expectation values which were incompatible with the predictions of standard QM, and as was eventually shown, incompatible with experimental results. No more, no less.

Bell proved that in a local deterministic world which is not superdeterministic, the correlation MUST be linear, in the sense that the mismatch at 2θ cannot be greater than the mismatch at θ.

No, he didn't prove that. He proved that, wrt a certain entanglement setup, the predictions of standard QM are incompatible with the predictions of a certain local realistic hidden variable supplementation of standard QM.

As mentioned before, the fact that the correlation between angular difference and coincidental detection is predicted by the achetypal LR model to be linear raises a red flag wrt the applicability of the model to the experimental situation -- unless one assumes that the model is informing, precisely, wrt the underlying reality. But that's a, maybe, unwarranted assumption wrt which there's no definitive test.

And, if there's any possibility of a simpler explanation for the violation of BIs via the formalism and the experimental setup, then that's the preferred approach, imho -- not the over the top assumption of superdeterminism.

I have no idea what you mean by Bell-type formulation and non-Bell-type formulation.

The salient characteristic of an essentially Bell-type formulation, and associated BI, is the encoding of an explicit locality condition. If a formulation is clearly, explicitly local, then it's also, necessarily, realistic.

Examples of non-Bell-type, purported LR models of quantum entanglement have been proposed by Christian* and Unikrishnan among others. But these have been generally disregarded because, although they reproduce the standard QM predictions, they're not clearly LR models.

If our universe was local deterministic and not superdeterministic, then Bell's inequality would be satisfied, a viable LR model would be possible, Bell's inequality would be confirmed by experiment, and quantum mechanics would be disproven. All of that follows from the argument I quoted.

And if the Queen had ***** she'd be the King. The superdeterministic argument isn't compelling, imho. And I think it's the wrong approach to take to understanding why BIs are violated.

Whether nature is local or nonlocal is still an open question, imho.
-----------------------

* Wrt Christian, I recall that a while back, before I understood some stuff which I've subsequently learned, I made the comment to DrChinese that he just didn't understand Christian's LR offerings. On getting more into it I found that I really didn't understand them either, and, it seems to be the case that most people don't understand them. So, I apologize to DrC for that comment -- and, by the way, DrC has demonstrated a much deeper knowledge of this stuff than I previously gave him credit for. So, I must defer to whatever he (and also Demystifier, who has demystified some stuff for me) have to say about it.

Anyway, Christian has a slew of papers at arxiv.org purportedly refuting Bell's theorem and offering a viable LR model. The problem with Christian's LR thing is that his offerings aren't clearly LR models, and he has never explained them in a way that sufficiently clarifies his claim. But if you're a mathematician with some interest in natural philosophy, you might find them interesting.

 

[f95toli]

That is, there's a couple of hundred years of optics detailing the characteristic behavior of light in experimental situations similar to Bell test setups (eg., polariscopic setups), and what the simplest realization of Bell's theorem (eg., Herbert's popularization) suggests is that, in a local deterministic universe, light should be expected to behave contrary to that.

I don't normally get involved in these discussions, because ultimately I don't think they are very interesting.
However, I thought I'd add my (usual) comment about experimental QM.

Whereas nearly all "conceptual" (and many practical) experiments are done using light, there are lots of examples of QM experiments that do NOT involve light, angular momentum, polarization, photon detectors etc.

People have performed experiments that are formally exactly analogues to the early (optical) tests of Bell's inequalities. We will probably see the first demonstrations in my own field pretty soon (solid-state QIP, which has nothing to do with optics but the QM formalism is obviously the same)
Hence, any attempt to explain away Bell type experiments by saying that that results are due to the fact we do not understand a specific technical detail of Aspect's original experiments is ultimately futile.

Moreover, note also that we are nowadays often -in practical- terms more concerned about OTHER inequalities that for one reason or another are better to test experimentally. A good example are tests of Legget-type inequalities which can used to test whether or not QM is local.

 

[ThomasT]

I don't normally get involved in these discussions, because ultimately I don't think they are very interesting.
However, I thought I'd add my (usual) comment about experimental QM.

Whereas nearly all "conceptual" (and many practical) experiments are done using light, there are lots of examples of QM experiments that do NOT involve light, angular momentum, polarization, photon detectors etc.

People have performed experiments that are formally exactly analogues to the early (optical) tests of Bell's inequalities. We will probably see the first demonstrations in my own field pretty soon (solid-state QIP, which has nothing to do with optics but the QM formalism is obviously the same)
Hence, any attempt to explain away Bell type experiments by saying that that results are due to the fact we do not understand a specific technical detail of Aspect's original experiments is ultimately futile.

Moreover, note also that we are nowadays often -in practical- terms more concerned about OTHER inequalities that for one reason or another are better to test experimentally. A good example are tests of Legget-type inequalities which can used to test whether or not QM is local.

What's the more reasonable expectation, based on historical observations of the characteristic behavior of light, wrt archetypal optical Bell tests, a linear or a nonlinear correlation between angular difference and coincidental detection?

 

[DrChinese]

You can obviously have a conspiracy where all the particles in the universe carry literally all information about all other particles in the universe, and use this information to do exactly what they need to do to make Bell's inequality appear violated, and never use this information except when dealing with quantum entanglement. That's why it's called ad hoc.

Oh, I quite agree. My point was that within the known properties of particles, there is not enough information. Further, there are no known mechanisms for transferring that sufficient information to make Bell tests work out. So simply making a statement that "all particles have previously been in causal contact" (were that true) would not be enough to have the conditions for superdeterminism. Ditto for the statement that the experimenter does not have free will to choose. That doesn't get you to Bell test results either.

I think we both agree, you need something radically more - something, well, super.

I would say that superdeterminism is as meaningful for explaining Bell tests as it is for explaining special relativity. Or general relativity. Or the charge of the electron. Etc. (Which is to say, not meaningful at all, sorry that this conclusion seems to offend people. But the original question was a "why" question and this is in fact the answer.)

 

[DrChinese]

* Wrt Christian, I recall that a while back, before I understood some stuff which I've subsequently learned, I made the comment to DrChinese that he just didn't understand Christian's LR offerings. On getting more into it I found that I really didn't understand them either, and, it seems to be the case that most people don't understand them. So, I apologize to DrC for that comment -- and, by the way, DrC has demonstrated a much deeper knowledge of this stuff than I previously gave him credit for. So, I must defer to whatever he (and also Demystifier, who has demystified some stuff for me) have to say about it.

Anyway, Christian has a slew of papers at arxiv.org purportedly refuting Bell's theorem and offering a viable LR model. The problem with Christian's LR thing is that his offerings aren't clearly LR models, and he has never explained them in a way that sufficiently clarifies his claim. But if you're a mathematician, you might find them interesting.

I don't follow Christian, it's true, I think this is common to his work.

I instead push the "DrChinese challenge" when it comes to a candidate local realistic theory: give me a dataset with values for simultaneous polarization outcomes at 0, 120 and 240 degrees. Then show me how your model gets to a 25% correlation rate from that. If someone cannot do that, then I conclude their model is worthless and there is no reason to try to find their error. Saves me a lot of effort in trying to demonstrate that the model is not realistic, which is the usual flaw.

 

[billschnieder]

I don't follow Christian, it's true, I think this is common to his work.
I instead push the "DrChinese challenge" when it comes to a candidate local realistic theory: give me a dataset with values for simultaneous polarization outcomes at 0, 120 and 240 degrees. Then show me how your model gets to a 25% correlation rate from that. If someone cannot do that, then I conclude their model is worthless and there is no reason to try to find their error. Saves me a lot of effort in trying to demonstrate that the model is not realistic, which is the usual flaw.

Bah! I had hoped you will have seen problem the so called "DrChinese challenge" by now. You keep saying:

- give me a dataset with values for simultaneous polarization outcomes at 0, 120 and 240 degrees.

Don't you yet understand that "dataset with simultaneous outcomes" implies an experiment is being performed. Previously I asked you to describe the experiment and I will give you the dataset but you never described the experiment because you can not and nobody can because THERE CAN NEVER BE AN EXPERIMENT WHICH SIMULTANEOUSLY MEASURES TWO PHOTONS AT 3 ANGLES (yes I'm shouting this time).

Therefore failure of anybody to provide your purported dataset is not due to anything other than the fact that the request is nonsensical.

On Joy Christian, he is working on a book chapter http://lanl.arxiv.org/abs/1201.0775 FQXi which goes into details about his theory. He also recently posted a rebuttal of some of the recent claims about his theory http://lanl.arxiv.org/abs/1110.5876

To the opening poster, my answer as to why superdeterminism is not universally accepted as an explanation for nonlocality will simply be that nonlocality is not universally accepted as true. It's like asking why the flying pig is not universally accepted as an explanation for the unicorn.

 

[lugita15]

Don't you yet understand that "dataset with simultaneous outcomes" implies an experiment is being performed. Previously I asked you to describe the experiment and I will give you the dataset but you never described the experiment because you can not and nobody can because THERE CAN NEVER BE AN EXPERIMENT WHICH SIMULTANEOUSLY MEASURES TWO PHOTONS AT 3 ANGLES (yes I'm shouting this time).

Your bolded statement is certainly true, but even if we cannot measure the polarizations at all three angles, presumably if you're a local realist you still believe that definite polarizations at all three angles EXIST. Thus you should be able to give a possible set of values the polarizations at the three angles can have.

 

[ThomasT]

Bah! I had hoped you will have seen problem the so called "DrChinese challenge" by now. You keep saying:

- give me a dataset with values for simultaneous polarization outcomes at 0, 120 and 240 degrees.

Don't you yet understand that "dataset with simultaneous outcomes" implies an experiment is being performed. Previously I asked you to describe the experiment and I will give you the dataset but you never described the experiment because you can not and nobody can because THERE CAN NEVER BE AN EXPERIMENT WHICH SIMULTANEOUSLY MEASURES TWO PHOTONS AT 3 ANGLES (yes I'm shouting this time).

billschnieder ... nice to see you're still around and thinking about the Bell stuff. Not that I agree that there's anything wrong with Bell's stuff. But as you know I do disagree with interpretations of Bell's stuff that say it has anything to do with nature.

Therefore failure of anybody to provide your purported dataset is not due to anything other than the fact that the request is nonsensical.

Let's parse this objectively. Apparently DrC is assuming that Bell-type LR formulations and associated inequalities are general. Not an unreasonable assumption, imho. So, wrt that assumption, then his 'DrC Challenge' seems to me to be a very sensible and reasonable way to cut through the BS.

Of course, that assumption might not be correct. I think it probably is, but I don't know. But apparently you don't think that Bell's formulation and associated BIs are general. Is that the case?

On Joy Christian, he is working on a book chapter http://lanl.arxiv.org/abs/1201.0775 FQXi which goes into details about his theory. He also recently posted a rebuttal of some of the recent claims about his theory http://lanl.arxiv.org/abs/1110.5876

Thanks for the info. It might be interesting to see how this plays out.

To the opening poster, my answer as to why superdeterminism is not universally accepted as an explanation for nonlocality will simply be that nonlocality is not universally accepted as true. It's like asking why the flying pig is not universally accepted as an explanation for the unicorn.

Interesting, and perhaps provocative, way of putting it. Anyway, the first sentence is certainly true, I think.

 

[billschnieder]

Your bolded statement is certainly true, but even if we cannot measure the polarizations at all three angles, presumably if you're a local realist you still believe that definite polarizations at all three angles EXIST. Thus you should be able to give a possible set of values the polarizations at the three angles can have.

A polarization is an outcome of a physical measurement. It can not EXIST when the measurement has not been made. Realists, do not believe the outcomes of measurements exist when no measurement has been made. Three simultaneous polarization values are impossible so no realist believes it makes sense to ever contemplate three possible simultaneous polarization values, even hypothetically. In simple terms you can not measure an impossibility and it makes no sense to hypothesize an impossibility either.

see: https://www.physicsforums.com/showpost.php?p=3344159&postcount=193

 

[DrChinese]

Your bolded statement is certainly true, but even if we cannot measure the polarizations at all three angles, presumably if you're a local realist you still believe that definite polarizations at all three angles EXIST. Thus you should be able to give a possible set of values the polarizations at the three angles can have.

Ditto to the above. bill's comments are typical of a new breed of local realist who conveniently skirt the issue, claiming this is some kind of "naive realism" as if that means it can be dismissed. Guess it's great to assert something ("realism") that by their definition actually has no specific meaning.

Anyway, my apologies to everyone for responding to ThomasT's comment and accidently steering the discussion away from the topic at hand. If we need to continue, we can start a new thread.

 

[billschnieder]

billschnieder ... nice to see you're still around and thinking about the Bell stuff. Not that I agree that there's anything wrong with Bell's stuff. But as you know I do disagree with interpretations of Bell's stuff that say it has anything to do with nature.

Thanks TT, nice to see you around too. To me there is not just one way of characterizing "Bell stuff" and it is definitely possible to carve out some aspects of it that are valid and correct. However, some may say the interpretation is also "Bell stuff" and criticize everything together, which may come off as implying they think those more narrow aspects are individually wrong, which they don't -- get my drift?

Let's parse this objectively. Apparently DrC is assuming that Bell-type LR formulations and associated inequalities are general. Not an unreasonable assumption, imho. So, wrt that assumption, then his 'DrC Challenge' seems to me to be a very sensible and reasonable way to cut through the BS.

Not sure I understand what it means for Bell-type LR to be "general", all I can see is a logical contradiction in the "DrC Challenge" which I've pointed out many times. It is simply the fact that it is impossible to produce a dataset from an impossible experiment. And failure to do so says nothing about the validity or lack there-of of any Bell-type or realist arguments. Now this is not an unreasonable critique of the DrC challenge, is it?

 

[ThomasT]

Thanks TT, nice to see you around too. To me there is not just one way of characterizing "Bell stuff" and it is definitely possible to carve out some aspects of it that are valid and correct. However, some may say the interpretation is also "Bell stuff" and criticize everything together, which may come off as implying they think those more narrow aspects are individually wrong, which they don't -- get my drift? Not sure I understand what it means for Bell-type LR to be "general", all I can see is a logical contradiction in the "DrC Challenge" which I've pointed out many times. It is simply the fact that it is impossible to produce a dataset from an impossible experiment. And failure to do so says nothing about the validity or lack there-of of any Bell-type or realist arguments. Now this is not an unreasonable critique of the DrC challenge, is it?

Following DrC's suggestion, it might be better if this were introduced as a new thread in the philosophy forum if you want to pursue it. Including any and all posts that motivated you to post in this thread. Or maybe the QM forum, where you might get some more knowledgeable contributors, but I think it's more of a philosophical consideration.

This thread is about superdeterminsism, and why it isn't universally accepted as an explanation for nonlocality?
Any thoughts on that?

 

[lugita15]

That maybe the standard LR formalism and BI violations don't inform wrt the reality underlying instrumental behavior.

I'm not sure what this means.

\No, he didn't prove that. He proved that, wrt a certain entanglement setup, the predictions of standard QM are incompatible with the predictions of a certain local realistic hidden variable supplementation of standard QM.

No, Bell did not just prove that a particular local realist model failed to match the predictions of QM. He proved that any possible local deterministic universe which is not superdeterministic must satisfy the Bell inequality. If you disagree with this, look at the 12-step outline of Bell's argument I gave in a previous post and tell me what step does not apply to ALL local deterministic universes which are not superdeterministic.

 

[lugita15]

A polarization is an outcome of a physical measurement. It can not EXIST when the measurement has not been made. Realists, do not believe the outcomes of measurements exist when no measurement has been made. Three simultaneous polarization values are impossible so no realist believes it makes sense to ever contemplate three possible simultaneous polarization values, even hypothetically. In simple terms you can not measure an impossibility and it makes no sense to hypothesize an impossibility either.

see: https://www.physicsforums.com/showpost.php?p=3344159&postcount=193

billschnieder, we clearly have different definitions of what realism means. As DrChinese and ThomasT said you should start a new thread if you want to discuss your definition.

 

[ThomasT]

I'm not sure what this means.

It means that BI violations don't tell you anything about the reality underlying the instrumental behavior.

No, Bell did not just prove that a particular local realist model failed to match the predictions of QM. He proved that any possible local deterministic universe which is not superdeterministic must satisfy the Bell inequality.

Well, that's just an unwarranted interpretation of the meaning of Bell's theorem, imho.

If you disagree with this, look at the 12-step outline of Bell's argument I gave in a previous post and tell me what step does not apply to ALL local deterministic universes which are not superdeterministic.

I've looked at it. Your conclusion is unreasonable, imho.

If you really think you're on to something, then make a superdeterministic model of quantum entanglement.

 

[lugita15]

I've looked at it. Your conclusion is unreasonable, imho.

So then tell me which of the 12 steps need not hold in all (non-superdeterministic) local deterministic theories.

If you really think you're on to something, then make a superdeterministic model of quantum entanglement.

Um, I'm not sure what you're talking about. When did I say I'm on to a superdeterministic model?

 

[ThomasT]

I'm not sure what you're talking about. When did I say I'm on to a superdeterministic model?

You didn't, afaik. But you seem to be defending the notion that superdeterminism can explain BI violations. So, I'm just suggesting, make a superdeterministic model that can be experimentally tested. Otherwise, what are you talking about?

As for your 12-step recounting of Bell's argument, the conclusion that the correlation between the angular difference of the crossed polarizers and the rate of coincidental detection should be linear given the assumption of local determinism is wrong.

 

[lugita15]

You didn't, afaik. But you seem to be defending the notion that superdeterminism can explain BI violations.

All I'm saying is that Bell's theorem as such does not necessarily rule out superdeterminism, but it definitely rules out all other forms of local realism.

As for your 12-step recounting of Bell's argument, the conclusion that the correlation between the angular difference of the crossed polarizers and the rate of coincidental detection should be linear given the assumption of local determinism is wrong.

If my conclusion is wrong, one of my steps must be wrong. Which one is it?

 

[ThomasT]

All I'm saying is that Bell's theorem as such does not necessarily rule out superdeterminism, but it definitely rules out all other forms of local realism.

I basically agree with this ... with the qualifier that what Bell's theorem definitively rules out are Bell-type LR models of quantum entanglement. Whether non-Bell-type LR models can be definitively ruled out is still an open question, afaik. (Because, it's still an open question as to whether certain non-Bell-type LR models are actually LR models.)

If my conclusion is wrong, one of my steps must be wrong. Which one is it?

You're saying that the local realist would expect a linear correlation between θ and coincidental detection. Why would the local realist expect that?

My point is that if the local realist is aware of the historically documented characteristic behavior of light, then he wouldn't expect, in a local deterministic world, a linear correlation between the angular difference of the crossed polarizers and the rate of coincidental detection. He would, rather, expect a nonlinear correlation ... something approximating cos²θ.

The fact that Bell inequalities are, more or less, based on formal constraints which require the light in Bell tests to behave in an uncharacteristic way suggests that there's something in those constraints which is not corresponding to Bell test preparation and associated data processing. But not necessarily that nature is nonlocal.

 

[lugita15]

You're saying that the local realist would expect a linear correlation between θ and coincidental detection. Why would the local realist expect that?

To repeat, the fact that local realism implies a linear correlation is NOT some arbitrary assumption or constraint we place on local realist theories. It is the conclusion of a careful argument, and if you're claiming the conclusion is wrong then there must be something wrong with this argument.

 

[ThomasT]

To repeat, the fact that local realism implies a linear correlation is NOT some arbitrary assumption or constraint we place on local realist theories. It is the conclusion of a careful argument, and if you're claiming the conclusion is wrong then there must be something wrong with this argument.

Why does the argument assume that the correlation between θ and rate of coincidental detection should be linear?

 

[lugita15]

Why does the argument assume that the correlation between θ and rate of coincidental detection should be linear?

Argh! ThomasT, I told you, the argument does not assume it, it proves it.

 

[ThomasT]

Argh! ThomasT, I told you, the argument does not assume it, it proves it.

I didn't get that. How does it prove it?