Perceived contradiction in non-locality principle

[entropy1]

Predictions of theory does not change from which measurement is considered 'first'. Why do you want unambiguous definition of order of measurement?
If measurements are time-like separated there is unambiguous order of measurements but still if you would calculate later measurement first and earlier measurement second result will not change. So from the perspective of theory there is no need to establish order of measurements. Both will give equally valid result.

I get that. But this result is phrased in terms of a resulting wavefunction. This resulting wavefunction is, in my eyes, not unambiguously defined and therefore not real and academic. I'd rather speak in terms of correlation between measurement results than suggesting a wavefunction that is not really there. It like saying something is blue, but if you look at it differently it is red. Then, is it in fact blue or red? You may assert it is blue, but I am equally right if I assert it is red, or neither.

 

[zonde]

But this result is phrased in terms of a resulting wavefunction.

Why do you say that? On the contrary, what I wrote was based on predicted probabilities and correlations for actual measurements.

 

[entropy1]

Why do you say that? On the contrary, what I wrote was based on predicted probabilities and correlations for actual measurements.

Sorry, I must have misread you. You were quoting my response to Morrobay's post concerning wavefunction collapse. It is the notion of wavefunction collapse (resulting in a wavefunction) I have a problem with. We seem to agree about that.

 

[zonde]

Sorry, I must have misread you. You were quoting my response to Morrobay's post concerning wavefunction collapse. It is the wavefunction collapse I have problems with. We seem to agree about that.

Hmm, so you are trying to argue against collapse because it requires unambiguous order of measurement?
But if you use wrong order of measurement the end result is still correct (even if middle stage is 'wrong'). So it does not seem that your argument works.

 

[entropy1]

Hmm, so you are trying to argue against collapse because it requires unambiguous order of measurement?
But if you use wrong order of measurement the end result is still correct (even if middle stage is 'wrong'). So it does not seem that your argument works.

No, the unambiquity of the interpretation of the collapse!

Maybe this will illuminate the matter a bit: if Alice's filter is at angle a, and Bob's filter at angle b, the 'collapse' could result in angle a or angle b (sorry for my poor terminology - I am a layman) depending on how you look on it. So, the interpretation (of the collapse) is not unambiquously defined! It is like saying x+y may be 2, but it may also be 3.

I'd like to interpret non-locality as a correlation between measurements only. There is no need to pose a mathematical construct that bears no existence in reality (collapse).

 

[zonde]

Maybe this will illuminate the matter a bit: if Alice's filter is at angle a, and Bob's filter at angle b, the 'collapse' could result in angle a or angle b (sorry for my poor terminology - I am a layman) depending on how you look on it. So, the interpretation (of the collapse) is not unambiquously defined!

Collapse is unambiguously defined if you introduce preferred frame simultaneity convention.
You can say that you don't like preferred frame simultaneity convention (say because it's not very convenient) but you can't build an argument about collapse not being 'real' based on your likes or dislikes.

 

[entropy1]

Let me ask then why I should be adopting a preferred frame of reference. Isn't that purely a matter of convenience?

 

[zonde]

You shouldn't. That is if you do not try to model something that does not respect relativity.
Simultaneity is just a convention within domain of applicability of relativity. If you step out of that domain simultaneity is not just a convention any more.

 

[entropy1]

So how do I step out of the domain of relativity if I am an experimenter in a spaceship observing the experiment on earth?

In other words: why should we leave relativity out of it?

 

[Nugatory]

here is no need to pose a mathematical construct that bears no existence in reality (collapse).

Right... Collapse is not part of quantum theory, it's something that some interpretations have added on. If it doesn't help - and it as I've said before it is seldom helpful in understanding spatially separated measurements of multi-particle systems - you don't use it.

Don't confuse collapse with definite outcomes; there's no doubt that measurements produce outcomes, but collapse is merely one of the ways of explaining this. Also, don't confuse state preparation with collapse; I can prepare a system in a mixed state that is not reachable by collapse alone. That's why in general we have to represent states as density operators.

 

[DrChinese]

So how do I step out of the domain of relativity if I am an experimenter in a spaceship observing the experiment on earth?

In other words: why should we leave relativity out of it?

Consideration of relativity does not improve the quality of your predictions.

 

[.Scott]

So, am I right when I say that if Alice measures before Bob does, that A collapses and B takes on A's polarisation, and vice-versa: if Bob measures before Alice does, that B collapses and A takes on B's polarisation? Then how can you establish who measures first (Alice or Bob), and which photon of the pair takes on the other's polarisation? Is it even defined?

As you have already noticed, whether it is Alice or Bob that measures can be a matter of reference frame.
It would be better to describe the situation this way: As a single event, the photons separate and their non-local environment, which includes both detectors, determines how it will be measured at A and B. In other words, the entire system is one measurement. The before and after don't matter.

 

[entropy1]

Consideration of relativity does not improve the quality of your predictions.

No, but it does seem to suggest that 'collapse' is just a mathematical notion, and is not rooted in reality (as, for instance, the result of a conscious observer).

 

[zonde]

In other words: why should we leave relativity out of it?

You can't have FTL time-like sequence in relativity. FTL motion, causality, communication or anything timelike is not defined in relativity.
Basically you are saying "but if we try to describe FTL influence in relativity we get nonsense". Yes, that's right. Don't do that.

 

[entropy1]

You can't have FTL time-like sequence in relativity. FTL motion, causality, communication or anything timelike is not defined in relativity.
Basically you are saying "but if we try to describe FTL influence in relativity we get nonsense". Yes, that's right. Don't do that.

Wavefunction-collapse seems to require a relativistic approach, and the approach won't hold, it seems to me. So you can throw out relativity, but why not throw out collapse?

On the other hand: non-locality seems to imply FTL influence. This is not, on first sight, compatible with relativity.

Then again, you could also endorse other interpretations, such as the pure statistical one, retrocausality, and the EPR alternatives, such as for instance (the lack of) counterfactual definiteness.

There seem to be many different angles to the whole thing.

 

[.Scott]

Let me go further. Because there is no result without a measurement, it doesn't make sense to talk about A causing B or B causing A - because you would be assuming that had only one measurement been made, it would have been the same as the measurement that was made.

Here's the extreme example that I gave in another thread - but now I'll work a little more with it. Entangled particles split. One is measure immediately, the other is stored in a box for a year. In all reference frames, measurement A is made before measurement B. While waiting for the year to go by, we can publish the result of measurement A - and it would be very tempting to say that measurement A is independent of whatever we will eventually do with B. But the Bell inequality shows us that would be an impossible interpretation to consistently maintain. If we don't take the A+B measurement as a single measurement in this case, we would be forced to make arbitrary choices in how we interpret situations where the time period (1 year) was so short that no before/after could be unambiguously determined.

Even more importantly, we could capture thousands of those particles, a thousand each at angle -45,0,+45, storing all of them for a year in boxes. And then at the end of the year, Bill could make measurement B of each one - randomly choosing angles among the -45,0,+45 degrees. Presuming that the Bell inequality was demonstrated, to explain the result in terms of A influenced B, we would have to presume that somehow the "A" measurement got into the "B" box, or influenced Bill's best effort to randomize his measurements.

 

[Weddgyr]

The fact that someone could fly overhead in a rocket at .3c and reverse the detection events is just academic.

Not entirely. Coincidence times in Bell tests have have the order of nano-second differences. Depending on how spacelike separated you can make your experiment, you don't need to be moving that fast to really shift the timing of events to make one detection before another in a different frame. At the low orbital velocity of the ISS (~7 km/s), a ground experiment with spacelike separation of 20km in total would give a time shift of about 1ns if I got it right (using SR only). Anyway, you'd probably be better off putting the experiment in orbit for maximum spacelike separation if you really wanted to do this, but the point is that there probably have been actual human satellites which would have disagreed on which photon in Earth based experiments was detected "first". This is an almost completely academic point yes, but not entirely academic.

Collapse is unambiguously defined if you introduce preferred frame simultaneity convention.

I would be interested in reading references which formalise this method if anyone has them. In particular if they deal with the problem of entanglement across multiple frames.

 

[zonde]

we would be forced to make arbitrary choices

What so bad about arbitrary choices? You mean that we could make the wrong choice and that is unacceptable?
But in science we only care about "does not work" and "works for now". And from that perspective we can make arbitrary choice (even the 'wrong' one) as long as it works.

 

[zonde]

Wavefunction-collapse seems to require a relativistic approach, and the approach won't hold, it seems to me. So you can throw out relativity, but why not throw out collapse?

Don't ask me too much about wavefunction-collapse. I just saw that you tried to discard it based on flawed argument.
Why you can't follow Nugatory's suggestion and ignore collapse idea? Is it because you want to sort everything into 'correct' or 'flawed'?

 

[entropy1]

Why you can't follow Nugatory's suggestion and ignore collapse idea?

I may not ignore it, but I am challenging it, for I gather that some people treat it as real, whereas I don't.

 

[zonde]

I may not ignore it, but I am challenging it, for I gather that some people treat it as real, whereas I don't.

You can challenge the idea on the grounds that:
- it's predictions do not agree with experiments;
- it gives ambiguous predictions (it is not self consistent).
As I understand you are skeptical that collapse approach gives unambiguous predictions, right?

 

[entropy1]

As I understand you are skeptical that collapse approach gives unambiguous predictions, right?

No no. The expression "the wavefunction collapses" suggests that there is something in the real world 'collapsing'. But perhaps I am not familiar enough with the math. I still have to start with my first book.

 

[Demystifier]

On the other hand: non-locality seems to imply FTL influence. This is not, on first sight, compatible with relativity.

But on second sight, it might be:
http://lanl.arxiv.org/abs/1002.3226 [Int. J. Quantum Inf. 9 (2011) 367-377]

 

[.Scott]

What so bad about arbitrary choices? You mean that we could make the wrong choice and that is unacceptable?
But in science we only care about "does not work" and "works for now". And from that perspective we can make arbitrary choice (even the 'wrong' one) as long as it works.

No. I mean that we try to say measurement A causes measurement B or vice versa, we would have to describe very similar situations in opposite ways. And when the measurements we space-like separated, we would be making an arbitrary choice on how to interpret the measurements. That lack of time-based cause and effect is the core reason for taking the entire measurement process as one fully-integrated event. And, of course, the fact that the math that supports the modelling of the measurements also treats the measurements as a single operation.

 

[Demystifier]

So how do I step out of the domain of relativity if I am an experimenter in a spaceship observing the experiment on earth?

Relativity, just like any other theory in physics, has its domain of established validity. Certainly there are domains which have not yet been investigated, which leaves a possibility that relativity might be violated in some of those domains. One frequently studied possibility is that relativity is violated at very small distances (say of the order of Planck length). At such small distances, the correct theory (still unknown to us) might contain an explicit non-local mechanism lying behind the known phenomenological laws of quantum mechanics (QM). The fact that we don't see any such mechanism in the standard laws of QM is then a consequence of the fact that those standard laws do not describe the "true reality" at such small distances.

So, if you are an experimenter in a spaceship observing the experiment on earth, either you are able to measure such small distances or you are not. If you are not (which corresponds to current technology), then you don't see those explicit non-local mechanisms and all what you see looks in accordance with relativity. If you are able to measure them, then you are able to see the violation of relativity at those small distances.

 

[zonde]

That lack of time-based cause and effect is the core reason for taking the entire measurement process as one fully-integrated event.

Lets say you measure one particle from entangled pair and then depending from the measurement result you decide to perform or not perform entanglement swapping of second particle with a particle from another entangled pair.
Can you take it as one fully-integrated event?

 

[.Scott]

Lets say you measure one particle from entangled pair and then depending from the measurement result you decide to perform or not perform entanglement swapping of second particle with a particle from another entangled pair.
Can you take it as one fully-integrated event?

Well, you can. But that makes things too easy. Normally you want to keep one measurement from having an obvious influence on the other so that there is no "hidden value" solution.
Perhaps more importantly, the measurement of just one particle is "random", he equivalent of a coin flip. It is only when you have a population of A measurements to compare to a corresponding population of B measurements that you see a pattern.