Why do we think that collapse occurs at measurement?

[mmiguel]

Pardon me for my ignorance. This is a question of curiosity.
I look for a concise logical argument as to why the community believes the following hypothetical situation and explanation is wrong.

What justification is there for believing in waveform collapse at time of measurement?
Is it not possible that, say, entangled particles collapse into anti-correlated spins on their creation and that until we measure them, we model our uncertainty as a probability distribution?
Once measured, our uncertainty is gone, and we say the system has collapsed, when really only our uncertainty has vanished?
If we are unable to measure the system (without affecting it) before we measure it (affecting it), then how can we say anything at all about this system prior to "disturbing" it, which is beyond our ability to measure?

 

[Nugatory]

Is it not possible that, say, entangled particles collapse into anti-correlated spins on their creation and that until we measure them, we model our uncertainty as a probability distribution? Once measured, our uncertainty is gone, and we say the system has collapsed, when really only our uncertainty has vanished?

It is not possible - Google around for "Bell's Theorem" to find the argument for why. Basically, Bell's Theorem proves that no theory in which the entangled particles are created with their spin attributes already set can agree with the predictions of quantum mechanics in all situations - the interesting differences arise when we measure the spins at angles that are not exactly 180 degrees apart. These experiments have been done, and they agree with the quantum mechanical predictions.

If we are unable to measure the system (without affecting it) before we measure it (affecting it), then how can we say anything at all about this system prior to "disturbing" it, which is beyond our ability to measure?

We measure one member of the pair on one angle, and the other member of the pair on a different angle that is not exactly 180 degrees off of the first angle. Now we have two measurements on two different angles, and we know that for both angles, the unmeasured particle would have exactly the opposite spin if we had measured it. Study these results for a large enough number of particles and we find statistical results that cannot be reconciled with the assumption that both spins were set when the pair was created.
Google for "Bell's inequality experiment" to get the details of what are some of the most subtle and philosophically important experiments of the 20th century.

 

[mmiguel]

It is not possible - Google around for "Bell's Theorem" to find the argument for why. Basically, Bell's Theorem proves that no theory in which the entangled particles are created with their spin attributes already set can agree with the predictions of quantum mechanics in all situations - the interesting differences arise when we measure the spins at angles that are not exactly 180 degrees apart. These experiments have been done, and they agree with the quantum mechanical predictions.

We measure one member of the pair on one angle, and the other member of the pair on a different angle that is not exactly 180 degrees off of the first angle. Now we have two measurements on two different angles, and we know that for both angles, the unmeasured particle would have exactly the opposite spin if we had measured it. Study these results for a large enough number of particles and we find statistical results that cannot be reconciled with the assumption that both spins were set when the pair was created.
Google for "Bell's inequality experiment" to get the details of what are some of the most subtle and philosophically important experiments of the 20th century.

Thank you for your reply Nugatory.
Reading about Bell's experiment is exactly what brought me to this forum with my question.
My interpretation of Bell's experiment according to the source I was reading (Wikipedia) is the following:

1) Assume wavefunction collapse occurs at measurement
2) Show that two measurements of two entangled particles can be made space-like (i.e. events have a positive spacetime interval and therefore cannot be causally related without violating the speed of light)
3) Conclude that the measurement of the first particle measured collapse the common wavefunction of both particles, which in effect, produces instantaneous change over a positive spacetime interval (violates speed of light)
4) Conclude that local hidden variable theories are impossible

In my question about wavefunction collapse at measurement, I had thought I was asking about an assumption or precondition of the Bell experiment, and if that were so, the Bell experiment would not be helpful in answering my question. Is there another aspect of the Bell experiment that justifies the assumption that wavefunction collapse occurs exactly at the point of measurement?

I assume of course, this logical gap is due to my ignorance of the details of the subject. Because my question is very specific, I believe it is far more direct to ask a person who may know the answer than searching through texts that are trying to prove something other than what I am asking.

 

[atyy]

The Bell inequalities and the observed correlations in the Bell experiments can rule out local hidden variables regardless of whether collapse occurs at the measurement or not, and do not even depend on quantum mechanics being true.

Within the Copenhagen interpretation, the predictions of quantum mechanics for the Bell experiment do depend on wave function collapse occurring at measurement. However, there may be interpretations of quantum mechanics such as many-worlds, in which collapse does not occur. If many-worlds works (I'm not sure there is consensus on whether it works), it is not necesssary for quantum mechanics to depend to wave function collapse. However, within the postulated many-worlds interpretation, the Bell inequalities and Bell experiments do not rule out local hidden variables, because many-worlds violates an assumption in the derivation of the inequalities that each measurement produces a unique outcome.

 

[DrChinese]

Is it not possible that, say, entangled particles collapse into anti-correlated spins on their creation and that until we measure them, we model our uncertainty as a probability distribution?

atyy's comments are good. I would like to add the following relative to your comment above. The local realistic viewpoint is essentially as you say, with the added condition that the nature of a measurement here cannot affect the outcome of a measurement "there".

Given that, Bell's Theorem shows that no such local realistic theory can agree with a number of quantum mechanical predications that have been well tested. (QM could still be wrong, but so are all local realistic theories.)

Have you taken the time to work through the Bell logic? It is difficult to explain further if you haven't.

 

[mmiguel]

The Bell inequalities and the observed correlations in the Bell experiments can rule out local hidden variables regardless of whether collapse occurs at the measurement or not, and do not even depend on quantum mechanics being true.

Within the Copenhagen interpretation, the predictions of quantum mechanics for the Bell experiment do depend on wave function collapse occurring at measurement. However, there may be interpretations of quantum mechanics such as many-worlds, in which collapse does not occur. If many-worlds works (I'm not sure there is consensus on whether it works), it is not necesssary for quantum mechanics to depend to wave function collapse. However, within the postulated many-worlds interpretation, the Bell inequalities and Bell experiments do not rule out local hidden variables, because many-worlds violates an assumption in the derivation of the inequalities that each measurement produces a unique outcome.

Thank you for your response atty.
To make sure I correctly understand what you are saying, I will try to summarize your post - please correct any mistakes in my expression of your meaning:
1) Bell experiments rule out local hidden variables as long as we assume that wavefunction collapses occur, regardless of whether they occur at the point of measurement or not.
2) Under other interpretations, such as many-worlds, in which wavefunction collapses do not occur, the assertion of Bell that local hidden variables cannot explain all physical phenomena is not valid because certain assumptions in the Bell experiment are not met (unique outcomes, wavefunction collapse, both?).

My reaction to this (if my interpretation of your words is correct) would be to change my question to be, what justification do we have for believing in wavefunction collapse at all (or for that matter each of the assumptions e.g. unique outcomes)? Was it part of a historical first cut hypothesis that has been unchallenged for majority consensus due to satisfactory statistical prediction of quantum phenomena? In order to believe the Bell conclusion, one must first believe all the assumptions. It sounds like there are multiple perspectives that are each consistent with all the experimental evidence, some of which support Bell and some of which do not - however most things that I read about this assert that the Bell conclusion is fact. My curiosity is regarding the confidence with which these assertions are made and whether there is evidence other than credibility/ethos that points us towards one of these interpretations over the other. From my perspective and rudimentary knowledge, the Bell experiment does not do this because of some of the assumptions it makes.

 

[mmiguel]

atyy's comments are good. I would like to add the following relative to your comment above. The local realistic viewpoint is essentially as you say, with the added condition that the nature of a measurement here cannot affect the outcome of a measurement "there".

Given that, Bell's Theorem shows that no such local realistic theory can agree with a number of quantum mechanical predications that have been well tested. (QM could still be wrong, but so are all local realistic theories.)

Have you taken the time to work through the Bell logic? It is difficult to explain further if you haven't.

Thank you DrChinese. Do you know/have a recommended online source that concisely states the fundamental logic behind these experiments?

 

[Nugatory]

Is there another aspect of the Bell experiment that justifies the assumption that wavefunction collapse occurs exactly at the point of measurement?

Wavefunction collapse is not a formal part of quantum mechanics, it's what's called an "interpretation" - a mental model that we can choose to use when we're trying to make sense of what the mathematical formalism is telling us. Thus, it's neither an assumption nor a result of the Bell theorem.

If you choose to use the collapse interpretation, then you will interpret the experiments as involving a collapsing wave function. But that's not necessary, and indeed the Bell inequalities can be derived without using any quantum mechanics at all: The theorem just tells us that the assumption that the result of a measurement of one particle in the pair can be determined from properties it acquired at creation time (your original question) leads to a prediction (the inequality) for the correlations. QM only comes into the picture because this prediction is inconsistent with the prediction that QM makes, and therefore allows us to falsify one or the other.

 

[atyy]

Thank you for your response atty.
To make sure I correctly understand what you are saying, I will try to summarize your post - please correct any mistakes in my expression of your meaning:
1) Bell experiments rule out local hidden variables as long as we assume that wavefunction collapses occur, regardless of whether they occur at the point of measurement or not.

No. Bell experiments rule out local hidden variables without assuming quantum mechanics or wave function collapse.

2) Under other interpretations, such as many-worlds, in which wavefunction collapses do not occur, the assertion of Bell that local hidden variables cannot explain all physical phenomena is not valid because certain assumptions in the Bell experiment are not met (unique outcomes, wavefunction collapse, both?).

Yes. Unique outcomes from a measurement is an assumption used in the derivation of the logic that enables bell experiments to rule out local hidden variables.

My reaction to this (if my interpretation of your words is correct) would be to change my question to be, what justification do we have for believing in wavefunction collapse at all (or for that matter each of the assumptions e.g. unique outcomes)? Was it part of a historical first cut hypothesis that has been unchallenged for majority consensus due to satisfactory statistical prediction of quantum phenomena? In order to believe the Bell conclusion, one must first believe all the assumptions. It sounds like there are multiple perspectives that are each consistent with all the experimental evidence, some of which support Bell and some of which do not - however most things that I read about this assert that the Bell conclusion is fact. My curiosity is regarding the confidence with which these assertions are made and whether there is evidence other than credibility/ethos that points us towards one of these interpretations over the other. From my perspective and rudimentary knowledge, the Bell experiment does not do this because of some of the assumptions it makes.

If one uses the Copenhagen interpretation of quantum mechanics, it comes along with wave function collapse. The rationale for believing in the Copenhagen interpretation of quantum mechanics is that it has passed all experimental tests to date.

Yes, of course the Bell experiments rule out local hidden variables only if one grants the assumptions in the derivation. So if experiments do not give a unique outcome, then the Bell experiments do not rule out local hidden variables. If the many-worlds interpretation of quantum mechanics is possible and in fact true, then there will be no empirical way of choosing between it and the Copenhagen interpretation.

 

[bhobba]

Wavefunction collapse is not a formal part of quantum mechanics, it's what's called an "interpretation"

Indeed it isn't.

And indeed even in interpretations that have it, such as most versions of Copenhagen, since the wave function is simply an aid to calculating outcomes in those interpretations, that it instantaneously changes is of zero concern, anymore than probabilities change from one sixth for all sides to one for one side when a dice is thrown.

You may be interested in what I wrote in another thread:
https://www.physicsforums.com/showthread.php?t=755406

Thanks
Bill

 

[DrChinese]

Thank you DrChinese. Do you know/have a recommended online source that concisely states the fundamental logic behind these experiments?

Sure, see below. But there is little reason to review the experiments until you see WHY Bell's Theorem works.

http://arxiv.org/abs/quant-ph/9810080

http://arxiv.org/abs/quant-ph/0205171

Please keep in mind that in Bell experiments, a Bell inequality is derived and then found to be violated in the experiment. There are hundreds of different such inequalities, and they are not limited to photons. Each is a rejection of local realism and support for QM.