Written in the skies: why quantum mechanics might be wrong

[SF]

What's this guy saying?

The question of whether quantum mechanics is correct could soon be settled by observing the sky — and there are already tantalizing hints that the theory could be wrong.

Antony Valentini, a physicist at Imperial College, London, wanted to devise a test that could separate quantum mechanics from one of its closest rivals — a theory called bohmian mechanics.

[..]

So far it’s been impossible to pick apart quantum mechanics from bohmian mechanics — both predict the same outcomes for experiments with quantum particles in the lab.

But Valentini thinks that the stalemate could be broken by analysing the cosmic microwave background — the relic radiation left behind after the Big Bang. The cosmic microwave background contains hot and cold temperature spots that were generated by quantum fluctuations in the early Universe and then amplified when the Universe expanded.

Using the principles of quantum mechanics, cosmologists have calculated how these spots should be distributed.

However, Valentini’s calculations show that the hidden-variables theory might give a different answer. “Any violation of quantum mechanics in the early Universe would have a knock-on effect that we could see today,” says Valentini.

Almost all measurements of the cosmic microwave background seem to fit well with the predictions of quantum mechanics, says Valentini. But intriguingly, a distortion that fits one of Valentini’s proposed signatures for a failure of quantum mechanics was recently detected by Amit Yadav and Ben Wandelt at the University of Illinois at Urbana-Champaign (see 'Deflating inflation?'). That result has yet to be confirmed by independent analyses, but it is tantalizing, Valentini adds.

“It’s far too early to say that this is definite evidence of a breakdown in quantum mechanics — but it is a possibility,” he says.

http://www.nature.com/news/2008/080515/full/news.2008.829.html

 

[malawi_glenn]

What is that you don't understand?

It is easier to answer those kind of questions.

By observing the distribution of CMB, one can distinguish from a QM description and a Bohmian mechanical description of the CMB.

Bohmian quantum mechanics is a different formulation of the "ordinary" quantum mechanics:
http://plato.stanford.edu/entries/qm-bohm/

 

[SF]

Didn't Bell show that there canot be a quantum theory with hidden variables?

 

[Phycho]

No, he did not, and I can't believe in 2008 people still have this complete misunderstanding.
It's like saying Charles Darwin believed in creationism.

John S Bell showed no hidden variable theory could have both realism and nonlocality.
Proving nonlocality is a fact of nature (atleast as he himself believed) because he denied to throw out realism (which is the foundation of science).

As a matter of fact:
John S Bell himself was the strongest Bohm interpretation (yes the HV) proponent 'til his own death.

So what he would say himself is that he showed HV is correct, just not local ones.
In other words, he showed nonlocality is a fact of reality and reality doesn't have to be abounden to satisfy a interpretation of QM.

 

[vanesch]

This sounds like yet another erroneous interpretation of quantum theory ; amazing that you can get this published in Nature now. As far as standard quantum mechanics and Bohmian mechanics are concerned, both are *identical* in their empirical predictions. It is only when one makes a mistake in using standard quantum mechanics (by, for instance, "projecting out too early") that one might arrive at any difference. It is only in potential *extensions* of both that one might see any difference, so I'm really curious at what this claim could mean.

 

[Phycho]

Well you know, they love to print everything.
Whether it's right or wrong.

Remembre that article in NewScientist a few months back titled "parallel universes proof - study" ?
Has got to be one of the worst cases of journalism in recorded history.

However I think Antony Valentini got a little different Bohm interpretation than the standar or "Casual/Ontological interpretation".

 

[Demystifier]

This sounds like yet another erroneous interpretation of quantum theory ; amazing that you can get this published in Nature now. As far as standard quantum mechanics and Bohmian mechanics are concerned, both are *identical* in their empirical predictions. It is only when one makes a mistake in using standard quantum mechanics (by, for instance, "projecting out too early") that one might arrive at any difference. It is only in potential *extensions* of both that one might see any difference, so I'm really curious at what this claim could mean.

You are wrong. BM and standard QM are equivalent ONLY if the quantum equilibrium hypothesis is assumed to be valid. The idea of Valentini is that the Universe is in the equilibrium now (which is why physics now can be effectively described by standard QM), but that it might not have been in equilibrium at the beginning.

 

[Icosahedron]

John S Bell showed no hidden variable theory could have both realism and nonlocality.

You mean realism and locality.

 

[Fredrik]

You mean realism and locality.

Isn't it objectivity and locality? (That's the way I've heard it a bunch of times, but I'm guessing objectivity and realism are two words for the same thing).

What is objectivity/realism anyway? I read an old paper by Clauser and Horne once, where they defined "objective local theories" as theories that satisfy a condition that looked a lot like a locality condition to me, but I never figured out why they included the word "objective".

 

[Phycho]

Icosahedron, yeah sorry writing mistake there :)

Objectivity means real, independant of meassurement.
as in "out there" independant of our existence.

Objective reality and realism is same thing, yeah.

So obviously, realism is something that is to be kept, if you cut it out, there's no point of doing science and you can lay back with LSD trip your balls off and just believe that's real for the rest of your life.

So obviously nonlocality is a fact of nature.
Which is why what Valentini does is certainly very important.

 

[vanesch]

You are wrong. BM and standard QM are equivalent ONLY if the quantum equilibrium hypothesis is assumed to be valid. The idea of Valentini is that the Universe is in the equilibrium now (which is why physics now can be effectively described by standard QM), but that it might not have been in equilibrium at the beginning.

BM without the quantum equilibrium hypothesis is like QM without the Born rule. The very reason of existence of BM is the quantum equilibrium hypothesis which allows it to make definite predictions, just as the Born rule does for QM. When you relax this, you can predict about anything and its converse - in the same way as QM without the Born rule can predict about anything and its converse.

In other words, I consider that the quantum equilibrium hypothesis is part of BM. If you take it away, and replace it by something else, you simply have an extension of BM that isn't the original BM. And depending on how exactly you change it, you have a different alternative extension. This is BTW the same in QM: there are MWI extensions which also allow for a deviation from the Born rule (Hanson works in that direction). That also is an extension of QM.

And then you will have to prove that such an extended BM doesn't find an equivalent extended QM with a modified Born rule... because both are very very similar ! The reason is that their common "core" is the unitary evolution of the wavefunction, in both cases, and that "particle dynamics + the quantum equilibrium hypothesis" in the case of BM, or the "Born rule" in the case of QM (MWI version) is just an added axiom that let's us crank out probability distributions for observations starting from the unitary wave function evolution, in both cases. The standard approach in both (q.e.h. for BM, and Born rule for QM) gives equivalent such rules. If you modify this in one approach, there are chances you will find an equivalent modification in the other that gives you again equivalent outcomes.

 

[Demystifier]

Vanesch, now you are right.
However, there is a good motivation for such a generalization of QM. In 1992 Valentini found that the quantum equilibrium does not need to be postulated, but can be derived from the Bohmian equation of motion of particle trajectories. (By contrast, in orthodox QM, the Born rule must be simply postulated.) More precisely, he has found that even if the system starts evolution out of the equilibrium, it will end out in the equilibrium. With this fact, it is then quite natural to study the observable consequences of the possibility that the early Universe have not been in the equilibrium.

 

[vanesch]

Vanesch, now you are right.
However, there is a good motivation for such a generalization of QM. In 1992 Valentini found that the quantum equilibrium does not need to be postulated, but can be derived from the Bohmian equation of motion of particle trajectories. (By contrast, in orthodox QM, the Born rule must be simply postulated.) More precisely, he has found that even if the system starts evolution out of the equilibrium, it will end out in the equilibrium.

I'm not sure that in that derivation, there is not an extra (reasonable) hypothesis, like is also the case in statistical mechanics. You see, in quantum theory, especially in MWI views on it, there have been also several "derivations" of the Born rule - which, at least to me, also make an extra, reasonable hypothesis.
The different schemes to derive the Born rule in QM that come to mind are:
- Dewitt's approach, proving that the hilbert norm of worlds which do not experience a Born rule, will tend to 0 in the limit of an infinity of observations (this is like the frequentist interpretation of probability).
- there is Gleason's theorem, which shows us that the Born rule is the only possible statistical function one can define over Hilbert space which respects non-contextuality
- Deutsch's approach, which shows that the Born rule is the rule that a rational decider will assume as (Bayesian) probability for the future, under a set of axioms of what a rational decider is, and a set of equivalent situations.
- there is Hanson's approach, which shows (a bit like in BM) that starting from equiprobable fundamental states, only stable worlds can emerge which follow microdistributions which imply *roughly* a Born rule.

As I said, all of these approaches make at least a few extra hypotheses (and most of them seem to ignore this), but they nevertheless arrive also at the Born rule. I wouldn't be surprised that a detailled analysis of the derivation of the q.e.c. in BM also uses some hidden assumption (which nevertheless looks fairly acceptable).

 

[Demystifier]

I'm not sure that in that derivation, there is not an extra (reasonable) hypothesis, like is also the case in statistical mechanics. I wouldn't be surprised that a detailled analysis of the derivation of the q.e.c. in BM also uses some hidden assumption (which nevertheless looks fairly acceptable).

Well, the derivation is analogous to the classical Boltzmann H-theorem that proves that entropy cannot decrease with time. Some hidden assumptions are certainly there, but I would say that they are fairly acceptable.

The advantage of the Valentini proof is that one can also test it by a "numerical experiment". One starts with some different distribution, numerically evolves the particle trajectories with time, and really finds that they settle down in the quantum equilibrium distribution. There are no abstract hidden assumptions in such a numerical approach.