EPR paradox revisited, again. hehehe

[alfredblase]

Simply put the EPR paradox argues that the fact QM predicts that a measurement at A affects reality at B instantaneously, means that QM is violating the speed of light limit.

I ask for a clear proof or explanation of how QM does not violate the speed of light limit in EPR type experiments as I have not been able find this anywhere.

Thanks. ;)

 

[vanesch]

In the post:

https://www.physicsforums.com/showpost.php?p=849418&postcount=12

I gave a proof of how the *results* at Alice's cannot be affected by the choice of measurement at Bob's, and as such, that the effect cannot be used to build a FTL telephone.

As such, there is no *direct* violation of locality.

As for Bell's theorem, it does indicate that, if the results at Bob and at Alice are to be caused by an underlying mechanism, then it seems to indicate that the only mechanism we can think of is indeed a non-local one, in that every LOCAL mechanism would satisfy Bell's theorem.

In standard quantum theory, the non-local mechanism is of course the collapse of the wavefunction, which collapses it everywhere instantaneously.

There is a way out, when you take on the many worlds interpretation (I am known here as one of its proponents, btw). In the MWI there is no collapse, and there is no definite result at Bob (from Alice's perspective) as long as Alice didn't learn from Bob. In other words, the "outcome" at Bob is in a quantum superposition (wrt Alice) as long as she didn't hear from it.
As such, there _was_ no definite result, and as such, Bell's theorem was not violated ; it is only violated once Alice has the two results together, but then they are not non-local anymore.

 

[alfredblase]

hmm I am very sorry, but i didnt find that very clear..

 

[jackle]

I ask for a clear proof or explanation of how QM does not violate the speed of light limit in EPR type experiments as I have not been able find this anywhere.

You won't find any proof because it is generally accepted that you CAN reasonably interpret QM to be violating the speed of light limit through an underlying mechansim.

You can find a clear proof that you can't use it to send information faster than light but you will have to learn the maths first. Any faster than light mechanism underlying QM is not possible to expose and can not be exploited to directly send information.

 

[Tez]

hmm I am very sorry, but i didnt find that very clear..

Well, here's an analogy from classical probability theory as to how it works. Quantum mechanically the result is very similar to this:

Imagine someone places two die in different boxes, and gives them to you and a friend. They promise you that the values are the same, but you don't know which they are. Let me label the 6 different values a dice can take as a,b,c,d,e,f instead of 1 through 6 for clarity.

So you and your friend have a joint state of die which you can describe as the "state" (in what is a figurative but I hope obvious notation)

1/6*(a,a)+1/6*(b,b)+1/6*(c,c)+1/6*(d,d)+1/6*(e,e)+1/6*(f,f)

I hope it is clear that if each of you was asked to write down how you would describe each of your die individually, i.e. without regard to its possible correlations with the other persons die, that you'd say:
1/6*(a)+1/6*(b)+1/6*(c)+1/6*(d)+1/6*(e)+1/6*(f)

Now, imagine you perform a measurement which only tells you whether the value of the dice in your box is 1,2,3 (a,c,e) or 4,5,6 (b,d,f), and you happen to get the outcome that it is 1 to 3. You will immediately "collapse" the probability distribution of the two die to

1/3*(a,a)+1/3*(b,b)+1/3*(c,c)

Now what is your friend going to say the state of his own die is? Well, he doesn't know what outcome you got! So while you might say he should think the state of his die is
1/3*(a)+1/3*(b)+1/3*(c),
he will continue to think it is
1/6*(a)+1/6*(b)+1/6*(c)+1/6*(d)+1/6*(e)+1/6*(f)

Now, what if you are trying to signal by such measurements? Well, he might know that you've definitely performed the measurement (because you agreed that you would try at a certain time say) if you are trying to signal a 1, else you do nothing if you're trying to signal a 0. So he knows that you will either think the state is
1/3*(a)+1/3*(b)+1/3*(c),
or you will think the state is
1/3*(d)+1/3*(e)+1/3*(f),
but since each of these has probability 1/2 from his perspective (he doesn't know which outcome you got remember)the final state he has to assign is
1/2*[1/3*(a)+1/3*(b)+1/3*(c)]+1/2*[1/3*(d)+1/3*(e)+1/3*(f)]
=1/6*(a)+1/6*(b)+1/6*(c)+1/6*(d)+1/6*(e)+1/6*(f)

That is, as you might expect intuitively, he has no change in what he describes his system by regardless of whether you have performed a measurement or not! Because of this, nothing he does will tell him whether you've measured your system or not, and thus you can't use this process to signal.

The quantum situation is very analogous...

 

[vanesch]

That is, as you might expect intuitively, he has no change in what he describes his system by regardless of whether you have performed a measurement or not! Because of this, nothing he does will tell him whether you've measured your system or not, and thus you can't use this process to signal.

The quantum situation is very analogous...

Except for the correlations, of course!

I remember a long time ago a great post on the subject... wait...

here it is:

https://www.physicsforums.com/showpost.php?p=439887&postcount=4

 

[vanesch]

hmm I am very sorry, but i didnt find that very clear..

In the following post
https://www.physicsforums.com/showpost.php?p=851269&postcount=29

I tried to outline the MWI view on the EPR situation. Now, you can agree, or not, with this view, that's not the point. What's however clear, is that this is an example where no non-local interaction took place, and nevertheless the EPR correlations are found.

I think this view is interesting (even if you don't like it for other reasons) as a counter example to an often-made claim, namely that EPR situations IMPLY some FTL (= non-local) effect: this is a view in which this is not needed. As such, it plays the same role of counter example as does Bohm's theory, which is a counter-example to the claim that no deterministic hidden variable theory can be empirically equivalent to quantum theory. Clearly Bohm's theory *is* such an example.

 

[alfredblase]

I don't like any of this unfortunately. Let's look at my view of the problem please and then discuss =)

Ok imagine Alice and Bob are 100000 metres apart. Now Alice and Bob are sent 2000 pairs of entangled electrons between them. They agree before hand that they will always measure in the x direction, and they agree before the experiment which particular direction this will be. Also they synchronise their measurements, so that while these are of course not simultaneous, the time difference between the measurement events is too small for a signal traveling at the speed of light to travel from Alice to Bob in such an interval. Further they synchronise their measurements so that Alice will always be the first one to make the measurement.

Then they carry out the 2000 measurements and compare their results. Now if Alice is not really affecting Bob's measurement and if therefore Bob still has a fifty fifty chance of measuring either up or down every time, then you would expect that Bob would NOT have the exact opposite results to Alice. But he will! So it seems quite clear that Alice's measurement events do really affect Bob's measurement events, before any signal would have had time to travel between them.

 

[leandros_p]

Ok imagine Alice and Bob are 100000 metres apart. Now Alice and Bob are sent 2000 pairs of entangled electrons between them. They agree before hand that they will always measure in the same direction and they synchronise their measurements, so that while these are of course NOT simultaneous, the time difference between the measurement events is too small for a signal traveling at the speed of light to travel from Alice to Bob in such an interval. Also they agree that Alice will ALWAYS be the first one to make the measurement.

Then they compare their results. Now if Alice is not really affecting Bob's measurement and if therefore Bob still has a fifty fifty chance of measuring either up or down every time, then you would expect that Bob would NOT have the exact opposite results to Alice. But he would! Right? So it seems quite clear that Alice's measurement events do really affect Bob's measurement events, before any signal would have had time to travel between them, unless QM is an incomplete theory...

You say first that they agree before hand that they will always measure in the same direction...they synchronise their measurements...they also agree that Alice will ALWAYS be the first one to make the measurement...

And yet you say then that the results are affected before any signal would have had time to travel between them !

How on Earth did they agree on the direction of measurement, on the synchronisation, and on who is going to do the measurement first without having "any signal to travel between them before the measurements" ?

Leandros

 

[alfredblase]

Leandros...

dear oh dear...

they of course agree BEFORE they carry out any measurements. It would be a carefully planned experiment...

What I mean is that there is no way a signal can get from Alice to Bob in the time interval between Alice measuring an electron and Bob measuring its oppositely spinned pair.

 

[ttn]

In the following post
https://www.physicsforums.com/showpost.php?p=851269&postcount=29

I tried to outline the MWI view on the EPR situation. Now, you can agree, or not, with this view, that's not the point. What's however clear, is that this is an example where no non-local interaction took place, and nevertheless the EPR correlations are found.

Ah yes, that was nice to look at again. Let me ask a couple of follow-up questions on this, though. At several points you describe some state or other as from Alice's POV or Bob's POV. Mathematically, you just seem to be using these phrases to mean: think of the big expressions as linear combinations of definite belief states for Alice. Fine. But this signals a deeper worry that (despite your many eloquent attempts to answer) I still can't make any sense of. This whole MWI explanation of the EPR correlations makes sense (and is local) *either* "from Alice's POV" *or* "from Bob's POV" but not in any objective POV sense, if that makes sense. What I mean is: if you're Alice, you do the experiment, you find yourself in some definite belief state afterwards (b/c in Patrick's MWI, the Born rule is just applied by hand in such situations, yes?), and still believe yourself to be in a universe where Bob is in a superposition. Then, later, when Bob wanders over, the local interactions permit the appropriate kinds of coupling-up between states of definite belief for Bob and the already-real definite belief state of you, Alice. Yes? The problem is, this is (as a proposed ontology) inconsistent with how things appear "from Bob's POV". I mean, he too performed an experiment and found himself in a definite belief state afterwards, right? If we "look at things from Bob's POV" then you, Patrick, will apply the Born rule ("for Bob") after he makes his measurement, etc... all in parallel to the story we told for Alice, but actually inconsistent with it. For example, is there a fact of the matter about whether or not Bob is in a definite belief state while he's walking over to Alice's house?

Saying the same thing another way and using (Patrick) your idea of a "token" (representing, in your MWI theory, a person's conscious belief states) that follows along one or the other of the branches (with probability given by Born's rule), some of the time the Alice-token and the Bob-token will end up in different branches. Right? So the conversation at the coffee shop between the two sentient humans Alice and Bob, in fact, never happens in this theory.

Now, I realize you can get around this by bringing in David Albert's "mindless hulks" -- so the Bob that shows up to exchange notes with Alice is not the *real* (conscious) Bob (who exists now in a parallel universe) but is instead a mindless hulk who *acts* like the real Bob (except he reports a different outcome than the real Bob might have). Or if you prefer to be Bob, it's the same thing in reverse: with some fair odds, the Alice that shows up to exchange notes with you at the coffee shop is a mere mindless hulk. The point here really is this: MWI is able to provide a local explanation of the EPR correlations only by going solipsistic -- that is, only by rejecting any pretension of telling a real story about what's going on in the real world, and instead talking merely about how things appear to some one particular observer (but then from this POV there is no making sense of the apparent sentience of other people, since those appearances are overwhelmingly likely to be *wrong*). It is, in effect, a theory of a single consciousness "processing" a constant stream of influxing information. Yes? And if the assignment was to give a local explanation for how (sentient beings) Alice and Bob can have a certain conversation in the coffee shop after their experiments, this means MWI has not suceeded. It *doesn't* provide a local explanation of *that*. Instead, it tells a (bizarre) story according to which there *aren't* two sentient beings having a conversation at all.

I think this view is interesting (even if you don't like it for other reasons) as a counter example to an often-made claim, namely that EPR situations IMPLY some FTL (= non-local) effect: this is a view in which this is not needed. As such, it plays the same role of counter example as does Bohm's theory, which is a counter-example to the claim that no deterministic hidden variable theory can be empirically equivalent to quantum theory. Clearly Bohm's theory *is* such an example.

Now you're moving in the right direction.

 

[abszero]

Leandros...

dear oh dear...

they of course agree BEFORE they carry out any measurements. It would be a carefully planned experiment...

What I mean is that there is no way a signal can get from Alice to Bob in the time interval between Alice measuring an electron and Bob measuring its oppositely spinned pair.

Which means that useful information did not get transmitted faster than light. You can only transmit information faster than light in a way that violates relativity if they can know each other's experimental results without any other communication between them aside from the entangled units. Therefore, their knowledge of what orientation to measure and such does not violate the whole speed of light thing.

Nature is very good at preserving causality, and this is another example of that.

Finally, just because you don't understand a statement and proof of Bell's inequality doesn't mean it isn't "simple". Some of the most worthwhile results in physics and math are pretty complicated, even to the experts.

 

[alfredblase]

You can only transmit information faster than light in a way that violates relativity

How irritating..
any transmition of information faster than light is a violation of special relativity. grrrr

people please stop posting silly threads or threads not pertaining to what I want to discuss or ill have to start a new one.

Finally, just because you don't understand a statement and proof of Bell's inequality doesn't mean it isn't "simple". Some of the most worthwhile results in physics and math are pretty complicated, even to the experts.

What on Earth are you talking about! FOr one thing you contradicted yourself in that paragraph and for another Bell's inequalities prove that hidden variable theories are not viable, but this is not a discussion about hidden variable theories!

stop posting silly threads!

 

[Doc Al]

Enough with the attitude and the insults, alfredblase.

 

[abszero]

How irritating..

What on Earth are you talking about! FOr one thing you contradicted yourself in that paragraph and for another Bell's inequalities prove that hidden variable theories are not viable, but this is not a discussion about hidden variable theories!

Bell's inequality is intimately related to EPR type things, since it declares that quantum mechanics is locally complete. This is clearly related to your "instantaneous transmission of information" idea.

 

[vanesch]

For example, is there a fact of the matter about whether or not Bob is in a definite belief state while he's walking over to Alice's house?

This is always the same misunderstanding. There is ONE Bob's body, but there are, if you want, TWO Bob-experiences, one of which is the continuation of the original (and whether, or not, the other is "mindless" or "came into existence as a copy" doesn't really matter - but if you don't like the mindlessness, just consider it to be a "copy that came into existence"). It is the one that is the continuation of the original that follows the Born rule. Now, if you prefer that the other copies also are conscious (it that eases your mind), you can postulate for them what you want ; for instance that they will ALSO follow subsequently the Born rule. But all those considerations have no implications what so ever on any observable thing - so it is completely open to your personal preferences.

Saying the same thing another way and using (Patrick) your idea of a "token" (representing, in your MWI theory, a person's conscious belief states) that follows along one or the other of the branches (with probability given by Born's rule), some of the time the Alice-token and the Bob-token will end up in different branches. Right? So the conversation at the coffee shop between the two sentient humans Alice and Bob, in fact, never happens in this theory.

As I said, if this disturbes you, then "create" new sentient minds during the split, it doesn't matter, as long as the *original* mind follows the Born rule - in fact, as long as YOUR mind follows the Born rule.

Now, I realize you can get around this by bringing in David Albert's "mindless hulks" -- so the Bob that shows up to exchange notes with Alice is not the *real* (conscious) Bob (who exists now in a parallel universe) but is instead a mindless hulk who *acts* like the real Bob (except he reports a different outcome than the real Bob might have). Or if you prefer to be Bob, it's the same thing in reverse: with some fair odds, the Alice that shows up to exchange notes with you at the coffee shop is a mere mindless hulk. The point here really is this: MWI is able to provide a local explanation of the EPR correlations only by going solipsistic -- that is, only by rejecting any pretension of telling a real story about what's going on in the real world, and instead talking merely about how things appear to some one particular observer (but then from this POV there is no making sense of the apparent sentience of other people, since those appearances are overwhelmingly likely to be *wrong*).

Again, induce them with NEW consciousnesses if you like. I don't find that necessary as I don't know the difference between a conscious person and a mindless hulk which acts exactly as a conscious person. I don't know if my cat is a conscious being, for instance. I take it that my family are conscious beings, but that's more like social convention :-) What I want to say here is simply this: assigning, or not, consciousness to another being _is a matter of convention or choice_ as there's no way to find out. You even have that problem in classical physics (cfr all the "paradoxes" that people working in artificial intelligence come up with).

What counts is how the world APPEARS wrt a particular observer, and that's good enough, no ?

It is, in effect, a theory of a single consciousness "processing" a constant stream of influxing information. Yes? And if the assignment was to give a local explanation for how (sentient beings) Alice and Bob can have a certain conversation in the coffee shop after their experiments, this means MWI has not suceeded. It *doesn't* provide a local explanation of *that*. Instead, it tells a (bizarre) story according to which there *aren't* two sentient beings having a conversation at all.

But you can have a perfectly normal conversation with a mindless hulk, or with a COPY of a sentient being, no ? BY DEFINITION, there is no difference in their behaviour. They will answer in the same way. How do you KNOW that your wife is not "a perfect copy of your wife" ?
With or without consciousness ?

 

[alfredblase]

Abszero,

Allow me to put Bell's theorem aside for the moment. I put to you that ttn's and vanesch's "proofs" (which they have admitted are essentially equivalent) are demonstrated to be inadequate by the variation on the EPR experiment and argument I have put forward in post 8 of this thread. If anyone has a decent, well thought out argument against my claim I would be very interested to hear it.

Abszero I have read your comments undermining my post 8 argument over and over: it is clear to me that you are confused. Perhaps I did not explain my idea in enough detail or perhaps not clearly enough. It seemed obvious to me at the time that my proposition was more than simple enough to be immediately understood by all. But then I do have a strange way of communicating sometimes. If people still appear not to understand what I mean then I will make a new thread/post with a more careful and detailed explanation of my idea.



Coming back to Bell's theorem: it seems that it has proven that local realism is violated... however no-one has adressed my arguments from post 8 yet... I will not dwell on Bell's inequalities until some one proves that the claim i make in post 8 of this thread is void.

Thanks

P.S. Vanesch, ttn would you mind discussing whatever it is you are discussing in a separate thread? You seem to go off into your own little big discussions everytime there is an EPR related thread; perhaps you should make a thread called "the vanesch and ttn thread about EPR related stuff".. heheh; although perhaps the title i used for the thread is now misleading

 

[vanesch]

Then they carry out the 2000 measurements and compare their results. Now if Alice is not really affecting Bob's measurement and if therefore Bob still has a fifty fifty chance of measuring either up or down every time, then you would expect that Bob would NOT have the exact opposite results to Alice. But he will! So it seems quite clear that Alice's measurement events do really affect Bob's measurement events, before any signal would have had time to travel between them.

But according to that reasoning, the case where we put balls of identical color in the two bags for Alice and Bob would make you come to the same conclusion, no ?

If we put each time 2 red balls, one for Alice, and one for Bob, or 2 green balls, one for Alice, and one for Bob, with a random choice 50-50 to take red or green, Alice has 50% chance of having red or green, and Bob has 50% chance of having red or green, but of course, when they compare notes, each time Alice had red ,Bob had red too, and each time Alice had green, Bob had green too. According to your reasoning, Alice seeing the color of the ball influenced remotely the color of Bob's ball ??

The problem with your setup is that you keep the directions of measurement fixed. No EPR effect can be seen that way.

 

[vanesch]

P.S. Vanesch, ttn would you mind discussing whatever it is you are discussing in a separate thread? You seem to go off into your own little big discussions everytime there is an EPR related thread; perhaps you should make a thread called "the vanesch and ttn thread about EPR related stuff".. heheh; although perhaps the title i used for the thread is now misleading

The problem is that you ask for a proof of non-violating locality by EPR. But that depends on what are the boundary conditions of such a proof (iow, the pre-admitted postulates). Now, the barest proof that non-locality is NOT strictly needed is simply that you cannot have FTL signalling - that was my first contribution. The second "proof" is the MWI view. But that MWI view needs an a priori absense of the postulate that measurements are "definite", and that's the argument between ttn and me. So I'd figure that the discussion is highly relevant to your topic.
In other words, I provided you with a view where non-locality (in interaction) is NOT strictly necessary, but that needs a denial of a very obvious postulate, namely that Bob and Alice HAD definite outcomes.

I will grant it to you (and to ttn - but we know that already) that if you do NOT allow for leaving aside this postulate that measurements are unique and definitive, that you need a non-local mechanism if you want to find such a mechanism.

 

[leandros_p]

Leandros...

dear oh dear...

they of course agree BEFORE they carry out any measurements. It would be a carefully planned experiment...

What I mean is that there is no way a signal can get from Alice to Bob in the time interval between Alice measuring an electron and Bob measuring its oppositely spinned pair.

The agreements you are making between Alice and Bob ARE NOT about the preparation of the experiments. They are about performing the experiments. The agreements are not just careful plans, they affect each experiment's observations as it happens and they preordain the behavior of each observer during the experiment by a deterministic interdependent/correlative pattern.

it seems that it has proven that local realism is violated... however no-one has adressed my arguments from post 8 yet... I will not dwell on Bell's inequalities until some one proves that the claim i make in post 8 of this thread is void.

The agreements are possible only if Alice and Bob share information about making a measurement (time of measurement and measurement direction). If this information is shared in advance, it means that Alice and Bob indeed exchanged information of *performing* the experiment in advance. Alice and Bob are obliged to perform their experiments having a specific knowledge of what is going on on the "other side" at every instant of the experiment, according to the agreements, and each one must comply to other's agreed directives.

This is not communication faster than the speed of light. This is communication in advance.

Leandros

 

[Tez]

I don't like any of this unfortunately. Let's look at my view of the problem please and then discuss =)

Ok imagine Alice and Bob are 100000 metres apart. Now Alice and Bob are sent 2000 pairs of entangled electrons between them. They agree before hand that they will always measure in the x direction, [snip]

...well then this is silly.

Re-read the dice example I gave above. There Alice and Bob will also always get the same outcome (or you can make them always opposite by having Bob relabel his outcomes). Now do you really think there's anything nonlocal going on in that example?

Not all correlations are necessarily nonlocal. You need to choose between at least two different non-orthogonal measurements on an entangled system in order to get a locality-violating correlation.

Even then - not any two measurements will do. E.g. if Alice and Bob only chose between x and y measurements, for example, everything could be mimiced in the dice example by having each of them either choosing 123 versus 456 (equivalent to x) or 246 versus 135 (equivalent to y).

 

[ttn]

Ok imagine Alice and Bob are 100000 metres apart. Now Alice and Bob are sent 2000 pairs of entangled electrons between them. They agree before hand that they will always measure in the x direction, and they agree before the experiment which particular direction this will be. Also they synchronise their measurements, so that while these are of course not simultaneous, the time difference between the measurement events is too small for a signal traveling at the speed of light to travel from Alice to Bob in such an interval. Further they synchronise their measurements so that Alice will always be the first one to make the measurement.

Then they carry out the 2000 measurements and compare their results. Now if Alice is not really affecting Bob's measurement and if therefore Bob still has a fifty fifty chance of measuring either up or down every time, then you would expect that Bob would NOT have the exact opposite results to Alice. But he will! So it seems quite clear that Alice's measurement events do really affect Bob's measurement events, before any signal would have had time to travel between them.

Oops, sorry for hijacking the thread. I didn't see your note about that until after I posted another reply to patrick... Will you forgive me if I answer this question from post #8?

You're confusing two different points here: (1) does orthodox QM provide a local explanation for the correlations in this suggested experiment? and (2) is a local explanation for the correlations possible?

You have to separate those two questions, because they have different answers. The answer to (1) is: no. The answer to (2) is: yes. Let me try to briefly explain that.

OQM says the wave function description of the entangled state is complete. No "hidden variables." It also says that the probability for Alice to measure +1 on a given particle is 50% (based on the singlet state for the pair). But that is only true until/unless Bob has made a measurement. As soon as Bob makes a measurement, the 50% probability for Alice's +1 changes to either 0% or 100% -- the wave function collapses. Now this is clearly a nonlocal mechanism. (Or if that's not clear, it can be made rigorous with a definition of locality such as Bell Locality. But it seemed from your post like it was obvious to you that this was nonlocal.) So that's why the answer to (1) is "no". The collapse of the wf means that whoever makes the first measurement thereby affects the state of the distant particle (in a way that, according to the theory, affects the subsequent probabilities for the various outcomes).

But then there is question (2): is a local explanation for the outcomes possible? Sure, here's one: suppose the particles carry hidden variables which instruct them how to behave when their spins are measured. And suppose what we call a "singlet state preparation procedure" produces pairs which have opposite values for this hidden variable: so, with 50% probability a given pair is of the (+1,-1) variety, and also with 50% probability it is instead of the (-1,+1) variety. Then when the particles in the pair get to Alice and Bob and they make their measurements, they get Alice = +1 and Bob = -1 whenever the pair happened to be (+1,-1) and vice versa for the other type of pair. So the net effect is that they always disagree about the outcomes, but each sees a random string of +1 and -1 from pair to pair. So you *can* reproduce the actual (observed, or predicted-by-QM) correlations in a local way. (But only for this trivial kind of experiment; Bell proved that this kind of model *can't* work for a more complicated version where Alice and Bob get to choose from among several different axes along which to measure the spins.)

Does that make sense? Yes, you can explain the correlations locally, but orthodox QM provides a non-local explanation.

This was of course *precisely* the point that Einstein was after (but which got a bit obscured by Podolsky). If you accept the completeness doctrine, the collapse postulate means that the state of distant things changes because of something you do here. In other words, OQM (QM with the completeness doctrine) is a nonlocal theory. But Einstein saw that, probably (he thought), you could build a completely local theory that would agree with QM's predictions by dropping the completeness assumption and adding hidden variables. And despite what you have maybe heard elsewhere, there is nothing whatsoever wrong with this argument. It's absolutely true (and remains absolutely true after Bell) that OQM is a nonlocal theory and that the only possible way of constructing a local theory is to add hidden variables. (What Bell's theorem adds is the point that this "only possible way" doesn't work! So we're *stuck* with a nonlocal theory, period, and we just have to *choose*, using some criteria other than locality, whether to accept the completeness doctrine or not.)

 

[alfredblase]

Ok thank you all for the interesting replies. I will address each in turn and see what comes out.

From Vanesch 18:

But according to that reasoning, the case where we put balls of identical color in the two bags for Alice and Bob would make you come to the same conclusion, no ?

If we put each time 2 red balls, one for Alice, and one for Bob, or 2 green balls, one for Alice, and one for Bob, with a random choice 50-50 to take red or green, Alice has 50% chance of having red or green, and Bob has 50% chance of having red or green, but of course, when they compare notes, each time Alice had red ,Bob had red too, and each time Alice had green, Bob had green too. According to your reasoning, Alice seeing the color of the ball influenced remotely the color of Bob's ball ??

Indeed, I completely agree.

The problem with your setup is that you keep the directions of measurement fixed. No EPR effect can be seen that way.

Why is that a problem? I think that arguing that if we look at things the way QM would have us do, Alice does really affect Bob's measurements, is the whole point of the EPR paradox.

 

[alfredblase]

From vanesch post 19:

Now, the barest proof that non-locality is NOT strictly needed is simply that you cannot have FTL signalling - that was my first contribution.

But I reapeat once again that such a proof is shown to be inadequate when we consider the experimental setup i made in post 8...

The second "proof" is the MWI view

The Many Worlds Interpretation is an interpratation, nothing more, and as such is proof of nothing.

will grant it to you (and to ttn - but we know that already) that if you do NOT allow for leaving aside this postulate

I have at no point stated that local realism is inviolate... If you disagree with the argument I make in post 8 please give me good reasons why.

 

[vanesch]

Why is that a problem? I think that arguing that if we look at things the way QM would have us do, Alice does really affect Bob's measurements, is the whole point of the EPR paradox.

In the standard version, yes, because a collapse affects ALL the systems described instantaneously. So any measurement will affect the quantum state. If you take the quantum state to be "real" then yes.

That's why I think that collapse is ugly
Hence my preference for MWI, and then you DON'T have this situation.
Others will tell you that the wavefunction is "not really out there" but a kind of help to calculate probabilities (leaving in the middle, then, WHAT is out there).

 

[alfredblase]

From Leandros post 20:

The agreements you are making between Alice and Bob ARE NOT about the preparation of the experiments.

Yes they are... they are made before any measurements take place and are synchronised carefully. I repeat that you are confused.

They are about performing the experiments.

Well of course! All preparations for an experiment have a hell of a lot a lot to do with how the experiment is performed, don't be silly.

The agreements are not just careful plans, they affect each experiment's observations as it happens

Complete rubbish. You have not given a single reason why this would be the case.

and they preordain the behavior of each observer during the experiment by a deterministic interdependent/correlative pattern.

The behaviours of observers are determined before any experiment! arggghhhhh!

There is nothing wrong with communitcating before an experiment takes place!

 

[alfredblase]

from tez:

Not all correlations are necessarily nonlocal. You need to choose between at least two different non-orthogonal measurements on an entangled system in order to get a locality-violating correlation.

Firstly, this is one thing that is usually left unclear in most explanations of EPR. You are saying that locality must hold but local realism need not? Second emmm... what is your point exactly?

 

[alfredblase]

ttn post 23:

I have read elsewhere: http://en.wikipedia.org/wiki/Principle_of_locality that there is a crucial difference between locality and local realism. Bells theorem (this is also subject to doubt e.g. loopholes and such), shows that local realism is violated.

My argument is that post 8 demonstrates that QM tells us ALice actually affects Bob's reality.

 

[alfredblase]

reply to vanesch post 25:

In the standard version, yes, because a collapse affects ALL the systems described instantaneously. So any measurement will affect the quantum state. If you take the quantum state to be "real" then yes.

By quantum state what do you mean, do you mean the state often referred to as a superposition of states, or the state of spin up (also quantum) or both?

 

[jackle]

The behaviours of observers are determined before any experiment! arggghhhhh!

Can you prove this?

 

[ttn]

ttn post 23:

I have read elsewhere: http://en.wikipedia.org/wiki/Principle_of_locality that there is a crucial difference between locality and local realism. Bells theorem (this is also subject to doubt e.g. loopholes and such), shows that local realism is violated.

My argument is that post 8 demonstrates that QM tells us ALice actually affects Bob's reality.

Don't believe everything you read in encyclopedias -- especially ones where the articles are written by random members of the public -- and especially on subjects where even the experts (and hence also the public) are extremely muddle-headed!

Your post 8 demonstrates exactly what I said before: that orthodox QM's explanation for these correlations is not local. According to *that particular theory*, yes, Alice actually affects Bob's reality. This is inherent in the collapse postulate once the completeness doctrine is accepted. But you can't forget that this is just a statement about one particular theory (orthodox QM). The fact that that theory is nonlocal only proves that nature is nonlocal if you already accept that theory as correct. If you're willing to just accept that, then, yes, sure, it follows that nature is nonlocal -- but then you better have a good argument for just accepting the truth of OQM! The fact is, if you only consider the kind of experiment you described in post 8, it is *easy* to postulate a local explanation. Yes, this local explanation isn't the one given by OQM. But if your goal is to establish that nature (as opposed to just OQM) is nonlocal, you better have some way of rejecting this alternative local explanation.

 

[alfredblase]

reply to jackle post 30:

its self evident that an experiment must have an observer. Now in order to carry out a certain measurement the observer must follow clear instructions on how to carry out such a measurement. Surely you learned this in primary school science lessons?

 

[alfredblase]

reply to ttn post 31:

Your post 8 demonstrates exactly what I said before: that orthodox QM's explanation for these correlations is not local. According to *that particular theory*, yes, Alice actually affects Bob's reality.

Finally we can agree. However by agreeing with me on this point you also agree that your dice "proof" is inadequate as in that post you argued that:

he has no change in what he describes his system by

We now agree that this is a false statement. According to QM Bob goes from a state of less definate reality into a state of more definate reality. QM offers a description of Alice's and Bob's reality. Everytime Bob makes a measurement, he is aquiring more and more statistical evidence that the QM description of reality violates the speed of light limit.

Bobs results when compared to Alice make no sense to Bob (a QM advocate) who was expecting an infinately more random set of results.

So since the proofs/explanations that have been given so far are inadequate, I restate my original appeal: I ask for a clear proof that OQM does not violate the speed of light limit.

 

[jackle]

reply to jackle post 30:

its self evident that an experiment must have an observer. Now in order to carry out a certain measurement the observer must follow clear instructions on how to carry out such a measurement. Surely you learned this in primary school science lessons?

A set of clear instructions before an experiment doesn't prove determinism is correct. Try again.

 

[alfredblase]

jackle no one is talking about determinism here... you are confused. I recommend the philosophy forum.