Is physics non-local?

[sodium]

A recent topic in philosophy just veered off into this question, and I thought it might be a better idea to raise the question in this forum.

Anyway, this topic is about quantum physics and the EPR paradox, etc. The idea is this. Locality, for our purposes, means that you can't have event X cause event Y, unless it is possible that an object traveling the speed of light (or slower) could move from the position and time where X happens to the position and time where Y happens. Another way of saying this is that information cannot be transmitted faster than the speed of light.

Anyway, it is certainly true that we know of no way to usefully transmit information faster than the speed of light. However, some people argue that nonlocality does (and perhaps must) occur in order to explain the physical results we get in things like the Aspect experiment.

Anyone with an opinion? I'm not a physicist, BTW (in case you couldn't tell).

[Schneibster]

I suggest that all interpretations of quantum mechanics either explain the EPR and similar experiments as the consequences of non-local phenomena, or contain explicit non-local phenomena in their explanations of all behavior.

I am sure there will be many comments. This has been an interesting topic in the past.

[Anglican]

When two events are causally disconnected events for one event to 'cause' the other makes little sense (at least within relativty) as their time-ordering is frame dependent (i.e whether event A happens before, after or at the same time as event B depends on the refrence frame you choose).

Quantum mechanics is intrinsically non-local, 'local' theories cannot explain the evidence (i.e. experiments such as the Aspect experiment).

[PoodleLovinPessimist]

Doesn't Many Worlds preserve some form of locality?

[Schneibster]

Doesn't Many Worlds preserve some form of locality?You have to match the universes that have the correct correlation of second event to those that have the right first event, and this process is explicitly non-local.

[sodium]

Here's my take on the subject.

In experiments like Aspect, we see a correspondence between two measurements. Now, that in itself doesn't mean that there is nonlocality. In our everyday lives the fact that you can measure my existence at the pub means you won't find me existing at work at the same time, but no faster-than-light signaling is involved.

But in Aspect, this is combined with the lack of hidden variables. It isn't the case that the result has some value before it is measured (like in the example of me in the pub or at work). I think people conclude that a correspondence must be the result of some deterministic common cause.

My view is that the events are correlated but have no common cause in the past or present. However, the correlation itself is the result of a cause that lies enough in the past that it doesn't propagate faster than the speed of light. So locality isn't violated, but our intuition about why events may be correlated is violated.

So for example, the measurement of the z-spin of two different particles is correlated. This correlation is the result of an event in the past, for example the decay of a particle with neutral spin. This correlation doesn't propagate faster than light speed, because the particles involved don't move faster than light speed. This decay event doesn't, however, cause the specific spin measurements that later occur. The result of these measurements are random, and not determined by "hidden variables" that initiated with the decay.

In fact, if you measure a positive z-spin on one particle and a negative z-spin on the other, nothing caused either particle to have its specific spin. Only the relationship between the two was caused, by the initial decay.

[Protoctista]

You have to match the universes that have the correct correlation of second event to those that have the right first event, and this process is explicitly non-local.

Many-worlders don't agree with you that there are any explicitly non-local processes in many-worlds QM. Could you explain further how this "matching" happens?

Where do arguments like the following go wrong:

http://www.hedweb.com/manworld.htm#local
http://www.hedweb.com/manworld.htm#epr
http://arxiv.org/abs/quant-ph/0003146
http://xxx.lanl.gov/abs/quant-ph/9906007

?

Does your argument assume that the notion of a "world" and its splitting is explicitly written into the formalism, like collapse in the Copenhagen interpretation? (See Wallace (http://philsci-archive.pitt.edu/archive/00000208/)).

In my understanding, this is where many arguments against the MWI go wrong (including the Afshar experiment), because I don't know of any many-worlder who believes this. (I call it the "strawmany worlds interpretation" :p).

Many-worlds-according-to-its-adherents and many-worlds-according-to-its-opponents seem to be two different things.

[evil-religion]

In fact, if you measure a positive z-spin on one particle and a negative z-spin on the other, nothing caused either particle to have its specific spin. Only the relationship between the two was caused, by the initial decay.

Could be wrong here but....

I thought that if you looked at one set of particles with a device that detects X property you could statistically come up with 2/3 of the particles having that property so the other set of particles must have 1/3 with that property (if you measure a bivalent property that is) BUT it does not matter which set you measure you will still get the same results. SO how does one set affect the other at a distance? As you can't have both sets with 2/3 of the particles having mutually exclusive bivalent properties. So the act of measurement affects both sets of particles and this occurs after the decay and potentially at any distance.

My university physics is a little rusty (its been 8 years or so since I graduated) so please correct me if I missing something. I think I need to re read my final year disitation it was on the philosophy of quantum mechanics.

[Schneibster]

(I call it the "strawmany worlds interpretation" :p)LOL Good one.

Many-worlders don't agree with you that there are any explicitly non-local processes in many-worlds QM. Could you explain further how this "matching" happens? I know they don't. The problem is that for both of the following:
a) The measurements of the two photons in Aspect's EPR realization in the same polarization (spin) plane to be correlated, and
b) the measurements in different planes to violate Bell's Inequality
to be true, the alternatives where the measurements are in identical planes must be "matched" at the end to achieve correlation, and the alternatives where they are in different planes must also be "matched" in order to violate Bell's Inequality. This is because MWI violates Contrafactual Definiteness, or CFD. CFD violation means that unmeasured alternatives do not have a definite value; this assumption results in the violation of Bell's Inequality, a known behavior of the Aspect EPR realization. It has been shown that any interpretation of QM must either violate CFD or show explicitly non-local behavior in space.

I have had an extensive argument with Jesse in another thread regarding this, so let me describe the exact meaning I assign to "non-local in space" as opposed to "non-local in time," and then show why I assert that CFD requires behavior that is "non-local in time."

"Non-local in space" is the standard type of non-locality that is described in the Copenhagen Interpretation. That is, when either measurement of one of the two photons' polarization is made, the polarization of the other in that plane is determined; in other words, the collapse of the wave function occurs across all of space instantaneously in violation of the speed-of-light limit. Remember that the polarization (which is spin) of a particle in two planes is complementary, in terms of the Uncertainty Principle; if the spin in one plane is completely known, the spin in the other plane is completely indeterminate. Remember also that spin, unlike time/energy or position/momentum is an "all or nothing" type of thing; it is either up or down in one plane, and once it has been determined in that plane, it is indeterminate in any other plane. This can be understood to mean that if the spins are measured in different planes, the spin in the corresponding plane at the other particle has no value, and on the other hand if the spins are measured in the same planes, the spins in other planes also have no value. But for the spins in the same planes to have the same value, i.e. be correlated, as they always are, the spins must always come out the same and this cannot be accounted for by them having a previous value because of the violation of Bell's Inequality. Thus, the spin from one particle must have affected the other at the time of measurement and since the measurement areas for the two photons are spacelike separated, this must have occurred superluminally violating causality in a spacelike manner.

On the other hand, in Cramer's Transactional Interpretation or in Everett's Many Worlds Interpretation, the correlation of the measurements is achieved by behavior that is non-local in time. In the TI, this is achieved by the use of advanced wave solutions to the wave equations; these advanced wave solutions exhibit negative energy and negative frequency, and are therefore "...usually ignored as unphysical because they seem to have no counterpart in nature" according to Cramer's TI web site (http://mist.npl.washington.edu/npl/int_rep/tiqm/TI_30.html#3.1). The advanced wave travels backward in time, thus its negative energy and frequency; and to interpret the Aspect EPR realization, Cramer postulates that the advanced wave travels back in time from the measurement event to the event where the two photons were created, and sets the spin of the other photon to the correlated value in the plane of the first measurement. This propagation backward in time constitutes an interaction that is "non-local in time" as opposed to the instantaneous collapse of the wave function across all of space in the CI that is "non-local in space." In MWI, the mechanism is the association of the earlier alternative with its measurement with the later alternative's correlated measurement, in the case where the measurements are in the same plane. This also requires that an "influence" work backward in time to ensure that the correlated measurements occur in the same alternative; there is no alternative where the measurements in the same plane are not correlated, but there is no reason why in MWI there should not be such an alternative, and this is not a result of the operation of the interpretation but of its assumptions.

CFD violation postulates that the spins have no value until they are measured; but this requires that an influence work backward in time to ensure that the spins are correlated if the chosen planes of measurement happen to be the same, and also to ensure that they are not correlated if the measurements happen to be different. Only in this way can CFD work with both the correlation and the violation of Bell's Inequality.

One more definition of "non-local in space" vs. "non-local in time" can be made by referring to the concepts of "spacelike separation" and "timelike separation," more properly "negative timelike separation," from relativity theory. A process is non-local in space if two correlated events are spacelike separated from one another, yet the participants in these events can be shown not to have had their correlated values prior to the measurement. A process is non-local in time if two correlated events are timelike separated from one another, and the second is in the negative time direction from the first, yet the participants in these events can be shown not to have had their correlated values prior to the measurement. This second concept is a little harder to understand, because it is easy to confuse the formal measurements in EPR with the initial event where the measured particles originated; but when one sees that there are actually three events instead of just two, this second concept becomes much clearer.

The violation of CFD in MWI is postulated to be because the alternatives that do not "match" represent Hilbert space vectors that have a vanishing norm- in other words, in the frequentist interpretation of probability, their probability goes to the limit of zero as the number of trials goes to the limit of infinity. There is no process or mechanism proposed by which this happens; it is simply a result of the mathematics of the interpretation and need not be so. It is therefore a postulate that underlies the interpretation, and this postulate requires behavior that is explicitly non-local in time as opposed to the Copenhagen Interpretation or Consistent Histories Interpretation which postulate behavior that is non-local in space to explain the Aspect EPR realization.

Where do arguments like the following go wrong:

http://www.hedweb.com/manworld.htm#local
http://www.hedweb.com/manworld.htm#eprThe first leaves out that the requirement of locality in space plus CFD violation equates to non-locality in time. The second says, "It has been pointed out by both Albert [A] and Cramer [C] (who both support different interpretations of QM) that Bell and Eberhard had implicity assumed that every possible measurement - even if not performed - would have yielded a single definite result. This assumption is called contra-factual definiteness or CFD [S]. What Bell and Eberhard really proved was that every quantum theory must either violate locality or CFD. Many-worlds with its multiplicity of results in different worlds violates CFD, of course, and thus can be local." What they leave out is that it can be local in space, but not local in time, in order to ensure that the zero-probability events never occur. This is a consequence of CFD violation, which this excerpt explicitly shows is an assumption of MWI.

http://arxiv.org/abs/quant-ph/0003146
http://xxx.lanl.gov/abs/quant-ph/9906007

?As above, both of these assume that the correct alternatives will be matched at both ends; in terms of the first, when the "third measurement," that is, the comparison of the two measurements, is made, the correct correlation or lack of correlation will be observed. This assumption requires that the alternatives that match be associated in order to obtain this observation, and the method by which they are associated correctly requires that the alternative "know" what the other alternatives are and "choose" the correct one.

Does your argument assume that the notion of a "world" and its splitting is explicitly written into the formalism, like collapse in the Copenhagen interpretation? (See Wallace (http://philsci-archive.pitt.edu/archive/00000208/)). No, it merely requires the assumption of CFD violation. See above.

In my understanding, this is where many arguments against the MWI go wrong (including the Afshar experiment), because I don't know of any many-worlder who believes this. (I call it the "strawmany worlds interpretation" :p).

Many-worlds-according-to-its-adherents and many-worlds-according-to-its-opponents seem to be two different things.Hee hee, they like to hide the non-locality in their assumptions and pretend it doesn't exist. I have an opinion that all interpretations of QM require non-locality either in time or space, either explicitly in the explanation of EPR or implicitly in their assumptions. I do not yet have sufficient evidence (and may never bother to gather it) to make this a theorem. But it certainly is borne out in every interpretation I have seen so far.

[Schneibster]

Here's my take on the subject.

In experiments like Aspect, we see a correspondence between two measurements. Now, that in itself doesn't mean that there is nonlocality. In our everyday lives the fact that you can measure my existence at the pub means you won't find me existing at work at the same time, but no faster-than-light signaling is involved.

But in Aspect, this is combined with the lack of hidden variables. It isn't the case that the result has some value before it is measured (like in the example of me in the pub or at work). I think people conclude that a correspondence must be the result of some deterministic common cause.

My view is that the events are correlated but have no common cause in the past or present. However, the correlation itself is the result of a cause that lies enough in the past that it doesn't propagate faster than the speed of light. So locality isn't violated, but our intuition about why events may be correlated is violated.You should examine Cramer's Transactional Interpretation (http://mist.npl.washington.edu/npl/int_rep/tiqm/TI_toc.html). He proposes an explicit mechanism by which this can occur.

So for example, the measurement of the z-spin of two different particles is correlated. This correlation is the result of an event in the past, for example the decay of a particle with neutral spin. This correlation doesn't propagate faster than light speed, because the particles involved don't move faster than light speed. This decay event doesn't, however, cause the specific spin measurements that later occur. The result of these measurements are random, and not determined by "hidden variables" that initiated with the decay.

In fact, if you measure a positive z-spin on one particle and a negative z-spin on the other, nothing caused either particle to have its specific spin. Only the relationship between the two was caused, by the initial decay.Right, but what causes the correlation? Remember, in the same plane, they are always correlated, and in different planes, they are always truly random and violate Bell's Inequality. You have to account for both of these conditions, and you have only accounted for the correlation, not the violation.

[sodium]

Right, but what causes the correlation? Remember, in the same plane, they are always correlated, and in different planes, they are always truly random and violate Bell's Inequality. You have to account for both of these conditions, and you have only accounted for the correlation, not the violation.
Here BTW for anyone trying to follow this, is a good simplified explanation of Bell's Inequality that I might refer to later,
http://www.berlinet.de/schmelzer/PG/Bell.html

I think the problem is that it is natural to try to devise a classical mechanism for giving the observed results. I'm not claiming that one exists. There is no classical mechanism through which the observed correlations are maintained and yet Bell's inequality is violated. I'm just admitting that this is what happens, and questioning whether this is really nonlocal.

The strictest test for non-locality would be to demand that we send information at faster than light speed. That we cannot do. I think, though, that you might say that the experimental results require nonlocality, or require that information is being sent at faster than light speed, even though we can't make use of it.

But I think there is a big assumption in such an analysis. What is meant by claiming that the observation requires nonlocality is that it requires nonlocality if we are imagining a classical mechanism. If we don't restrict ourselves in this way, but allow any natural mechanisms, then the argument becomes circular. Unless we already concede that the quantum effects under discussion are nonlocal, we have no reason to believe that they require a nonlocal mechanism to be implemented.

[Schneibster]

I think the problem is that it is natural to try to devise a classical mechanism for giving the observed results. I'm not claiming that one exists. There is no classical mechanism through which the observed correlations are maintained and yet Bell's inequality is violated. I'm just admitting that this is what happens, and questioning whether this is really nonlocal.If it is not, then explain the interaction by which this correlation occurs. The fact that you cannot explain it in terms of any other interaction, quantum or classical, or any pre-existing property of the system (remember that violation of Bell's Inequality?), when added to the fact that the two detection events are spacelike separated and by Special Relativity cannot be causally connected, is the reason that it is referred to as "non-local."

The strictest test for non-locality would be to demand that we send information at faster than light speed. That we cannot do. I agree we cannot; but I disagree that this is a necessary test. All that is necessary is to show that there can be causality that travels faster than light, not information that travels faster than light.


I think, though, that you might say that the experimental results require nonlocality, or require that information is being sent at faster than light speed, even though we can't make use of it.Not information; influence. I.e., causality. But never information.

But I think there is a big assumption in such an analysis. What is meant by claiming that the observation requires nonlocality is that it requires nonlocality if we are imagining a classical mechanism. Not so. The mechanism is explicitly quantum mechanical; "spin" for example (which is the correlated measured quantity in the Aspect EPR realization) has no meaning in terms of the classical physical properties of anything. It is merely an internal freedom of action of quanta that behaves mathematically like the macroscopic phenomenon we call "rotation." There is no evidence that it actually is rotational in any way at all. Thus, the relationships between spins cannot have a classical meaning, but only a quantum mechanical one. Yet, the nonlocality is still required to explain the observed behavior.

If we don't restrict ourselves in this way, but allow any natural mechanisms, then the argument becomes circular. Unless we already concede that the quantum effects under discussion are nonlocal, we have no reason to believe that they require a nonlocal mechanism to be implemented.They are what they are; we merely observe, and attempt to explain. I cannot conceive of the various interpretations of QM being "true;" none of them can be "true." The maximal description of the quantum mechanical model of microscopic behavior is the mathematical description of that model. Any attempt to reduce such behavior to language, or even to a conceptual framework that is congruent to our perceptions of reality, introduces problems in understanding. But even that does not sufficiently describe it; note that even the mathematics breaks down when trying to describe what happens to quanta! It is possible (and I am not the first to say it) that there not only is not, but cannot ever be, a completely consistent mathematical description of reality! Look at Godel's Incompleteness Theorem and tell me that I am wrong; if you understand its implications, you will see that it is a distinct possibility that our very conception of "number" contains limitations that are not obeyed by reality.

[sodium]

If it is not, then explain the interaction by which this correlation occurs. The fact that you cannot explain it in terms of any other interaction, quantum or classical, or any pre-existing property of the system (remember that violation of Bell's Inequality?), when added to the fact that the two detection events are spacelike separated and by Special Relativity cannot be causally connected, is the reason that it is referred to as "non-local."
First, I wouldn't want to explain it by another interaction because I don't think it's an interaction at all. Second, the EPR paradox can be seen in the behaviour of a number of different reactions. You could explain any of them with reference to one of the others if you wanted.

You could imagine tiny particles, much smaller than photons, that decay into two even tinier particles that have the same kind of measured correspondence as those we're discussing. Each particle in the result of a quantum decay carries away one of these particles, and uses it as a reference to get the result we measure. So the weird correspondence we can get is explained by a weird correspondence in these tiny reference particles.

Of course, I don't find that satisfying. But that's all we do with a lot of correspondences that are possible in classical physics. They're explained by correspondences in smaller particles, or just accepted.
Not so. The mechanism is explicitly quantum mechanical; "spin" for example (which is the correlated measured quantity in the Aspect EPR realization) has no meaning in terms of the classical physical properties of anything.
That isn't the problem. The problem is, that we have a situation where there is a correspondence between a measurement of two particles. Of course, we're used to correspondences. Classical physics expects correspondences in momentum, so that if we measure the momentum of two halves of an exploded object, we'll get a correspondence. This will occur even if the halves are light years apart when we measure them. We don't claim this means they are still interacting when we measure them.

My claim is that the two particles in the quantum experiment also have correspondence in their properties, although they do not interact at a distance. The tricky part is that the correspondence isn't of a kind that we can emulate using particles as described by classical physics. But it isn't self-contradictory or mathematically impossible. It's just that we'd like to imagine a model based on classical physics, not using any quantum effects, that could give distant particles with the same correspondence. And it can't be done.

But that's not that strange. Let's say I asked you to explain the behaviour of objects in special relativity, but only by constructing mechanisms using pre-Einstein Newtonian physics. Could you do it? And even if you could, what's the point?
They are what they are; we merely observe, and attempt to explain. I cannot conceive of the various interpretations of QM being "true;" none of them can be "true." The maximal description of the quantum mechanical model of microscopic behavior is the mathematical description of that model.
But then you shouldn't talk about particles "interacting" or nonlocal causation, or any of that. You should just say, "here are the results. No one understands it." When you say that its the result of nonlocal interaction you are attempting to interpret the result, even if you aren't explaining it in full detail.

[Schneibster]

First, I wouldn't want to explain it by another interaction because I don't think it's an interaction at all. Well, there is a problem: an interaction is generally defined as a type of action that occurs as two or more objects, systems, or substances have an effect on one another; in other words, as they participate in causal behavior with one another. And this is causal behavior; that is the meaning of the combination of the correlation and the violation of Bell's Inequality. So you can call it anything you like- but the fact remains that this behavior fits the definition of an interaction.

Second, the EPR paradox can be seen in the behaviour of a number of different reactions. You could explain any of them with reference to one of the others if you wanted.First, you just denied it is an interaction- and then you called it a reaction. Is that the term you want to use to describe it? Being that a reaction is one part of an interaction, I have to say that I am not very much likely to share your opinion on the terminology, since it does not appear that the dictionary does either.

Be that as it may, second, I'm not quite clear on the distinguishing characteristics of the class you are proposing. You need to be much more clear on exactly what "reactions" you are talking about, and on the exact nature of the explanatory references you propose to make.

You could imagine tiny particles, much smaller than photons, that decay into two even tinier particles that have the same kind of measured correspondence as those we're discussing. Each particle in the result of a quantum decay carries away one of these particles, and uses it as a reference to get the result we measure. So the weird correspondence we can get is explained by a weird correspondence in these tiny reference particles. This is called a "hidden variable" interpretation. Such interpretations are ruled out by the violation of Bell's Inequality if they explicitly define the measured characteristic. They can define other characteristics that result in the measured characteristic; but the problem is that there is only one unexplained behavior, and all such interpretations that have so far been proposed result in other behavior besides the desired one that is never observed. So it looks like all such interpretations have been ruled out. There is only one such "hidden variable" interpretation that has proved to be viable, and it is rather esoteric. It was proposed by David Bohm in the 1960s. You might want to look into it. But I don't think that it's going to get you out of the situation you are in with this unexplained interaction that results in correlation plus violation of Bell's Inequality; it is a specifically non-local interpretation, which is something you have been trying to avoid.

Of course, I don't find that satisfying. But that's all we do with a lot of correspondences that are possible in classical physics. They're explained by correspondences in smaller particles, or just accepted.I think based on this statement that you are unclear on exactly what "classical physics" is. I admit that there is a certain fuzziness to it; but there are also some very well-accepted guidelines, and I will tell you about them.

One of the great dividing lines between "classical physics" and "modern physics" is taken to be the quantum theory. Max Planck proposed the quantum theory in 1900; and this is often taken as another of the dividing lines. A third dividing line involves the difference between "classical mechanics" (which is a slightly different concept than "classical physics") and "quantum mechanics." A "mechanics" is the set of equations that are used to describe motion and interaction between objects, systems, or substances.

The reason that I say that you are apparently not clear on the difference is because you are referring to particle physics as classical. Particle physics are completely non-classical; quantum mechanics is the accepted description of particle physics. Quantum mechanics proposes that there are some types of behavior of particles that have no corresponding behavior in macroscopic objects. This is one of the great differences between classical and modern physics; and among those different behaviors are the Uncertainty Principle, and the type of non-locality that you are arguing against.

That isn't the problem. The problem is, that we have a situation where there is a correspondence between a measurement of two particles. Of course, we're used to correspondences. Classical physics expects correspondences in momentum, so that if we measure the momentum of two halves of an exploded object, we'll get a correspondence. This will occur even if the halves are light years apart when we measure them. We don't claim this means they are still interacting when we measure them.However, we also can prove that the two pieces had that correspondence in momentum at every moment, and in every position, between the original interaction and the point where we decide to measure them. But because Bell's Inequality is violated by the particles, we know absolutely for certain that the particles did not have the corresponding values until we measured them! And this is not a theory- it is an observable, provable physical fact! Any theory we come up with must explain why this is. Any theory that fails to explain it also fails to describe reality.

This is a primary difference between classical and quantum mechanics. Quantum mechanics says that things happen that have absolutely no correspondence to any observed physical behavior of a macroscopic object. Position and momentum are not merely beyond our ability to measure simultaneously with unlimited accuracy, they in fact do not exist with unlimited accuracy! The uncertainty in the particle's position and momentum is not merely a measurement problem- it actually does not have a precise position and/or momentum!

My claim is that the two particles in the quantum experiment also have correspondence in their properties, although they do not interact at a distance. The tricky part is that the correspondence isn't of a kind that we can emulate using particles as described by classical physics. But it isn't self-contradictory or mathematically impossible. It's just that we'd like to imagine a model based on classical physics, not using any quantum effects, that could give distant particles with the same correspondence. And it can't be done.I think you aren't quite clear here on exactly what you envision these properties to be. They can be mathematically described; it is a tenet of physics that all properties can be mathematically described, and any property that cannot is non-existent. They can also be measured; it is another tenet that all properties can be measured, and any that cannot is non-existent. So here we have a measureable, mathematically describable property, but we have no explanation for it. But wait! We do have an explanation for it! It is a non-local interaction, subject to the probability laws of quantum mechanics, and perfectly describable; furthermore, we can measure this interaction in the Aspect realization of the EPR experiment.

After writing this, I'm not entirely certain that what you just said isn't exactly the same as what I just said, with one exception: you are maintaining that there is no action at a distance, but the Aspect experiment says you are wrong and there is. I really think that you should very carefully consider what you said, and what I just said, and see whether you might not be wrong to be maintaining this.

But that's not that strange. Let's say I asked you to explain the behaviour of objects in special relativity, but only by constructing mechanisms using pre-Einstein Newtonian physics. Could you do it? And even if you could, what's the point?I don't think you're getting that the quantum mechanical description of these interactions is that they are non-local. The most widely accepted interpretation of quantum mechanics, the Copenhagen Interpretation, explicitly describes the interaction between the two particles in EPR as non-local. All other interpretations either do the same, or contain non-local assumptions as part of their framework. Which is what I have been saying all along.

So, your premise, which is that the perceived non-locality is the result of using classical concepts to describe quantum phenomena, is proven incorrect; because the Copenhagen Interpretation is in no way classical physics. It is quantum mechanics. And that means that (to extend your analogy) we aren't trying to describe relativity using Newtonian physics; we are describing it using relativity.

But then you shouldn't talk about particles "interacting" or nonlocal causation, or any of that. Why not? The most widely accepted interpretation of quantum mechanics does exactly that. Do you suggest it be abandoned? What to you propose to replace it with? Why is it better?

You should just say, "here are the results. No one understands it." When you say that its the result of nonlocal interaction you are attempting to interpret the result, even if you aren't explaining it in full detail.Well, they've been trying to sweep it under that particular corner of the carpet now for about eighty years, and have yet to do so successfully. In fact, the sweeping was interrupted most ungraciously when the results of the Aspect experiment were published. It is now entirely and unambiguously clear that influences occur between events that are spacelike, or negative timelike, separated from one another. We can call them non-local interactions; or we can call them some other terminology; but they happen, and it is experimentally provable, and I don't really care whether you say potay-toe or potah-toe, it grows in the ground and you boil it and eat it.

[sodium]

And this is causal behavior; that is the meaning of the combination of the correlation and the violation of Bell's Inequality.
You interpret it as causal behaviour, because when you think about a non-causal correlation, you are limiting yourself to ones that can be achieved in classical physics.
First, you just denied it is an interaction- and then you called it a reaction
By reaction, I meant the start of the experiment where the particles initially fly apart. Now that I think about it, that's probably not necessarily a reaction. But I'm certainly not claiming that the particles are "reacting" at a distance. That would be a nonlocal interaction!
This is called a "hidden variable" interpretation.
No. My facetious theory involving tiny sub-particles shows results that can't be explained by hidden variables any more than can the Aspect result. Of course, the trick is that my explanation assumed that quantum weirdness was just a way particles could behave, and we could explain Aspect without reducing the weirdness to being the result of some more intuitive behaviour.
It was proposed by David Bohm in the 1960s. You might want to look into it. But I don't think that it's going to get you out of the situation you are in with this unexplained interaction that results in correlation plus violation of Bell's Inequality; it is a specifically non-local interpretation, which is something you have been trying to avoid.
Exactly.
The reason that I say that you are apparently not clear on the difference is because you are referring to particle physics as classical.
I think that if you imagine particles as behaving as Newton would have described, or Newton + General Relativity, then that's thinking in terms of classical physics. Of course, you personally aren't thinking entirely in terms of classical physics, because you allow properties like spin. But the kind of explanations you accept are still limited by the expectations of classical physics.
But because Bell's Inequality is violated by the particles, we know absolutely for certain that the particles did not have the corresponding values until we measured them! And this is not a theory- it is an observable, provable physical fact!
If you are talking about measured values, like spin, then I don't think the particles had these values before you measured them, corresponding or not. That's certainly the Copenhagen Interpretation.
Any theory we come up with must explain why this is. Any theory that fails to explain it also fails to describe reality.
Finally we may have reached the crux of the argument. What constitutes an explanation, and when, if ever, are we allowed to say that there is simply no explanation? More on this below.
The uncertainty in the particle's position and momentum is not merely a measurement problem- it actually does not have a precise position and/or momentum!
Agreed.
After writing this, I'm not entirely certain that what you just said isn't exactly the same as what I just said, with one exception: you are maintaining that there is no action at a distance, but the Aspect experiment says you are wrong and there is. I really think that you should very carefully consider what you said, and what I just said, and see whether you might not be wrong to be maintaining this.
I really believe that I understand what you are saying, and why your saying it. I just think you're making an assumption that isn't warranted, and that you can't believe that I'm not making the same assumption because it seems so obviously true. And perhaps it is true, and perhaps I'm wrong. But we haven't really started event to debate that issue.
I don't think you're getting that the quantum mechanical description of these interactions is that they are non-local. The most widely accepted interpretation of quantum mechanics, the Copenhagen Interpretation, explicitly describes the interaction between the two particles in EPR as non-local.
Well, I think there's some ambiguity. It might help to consider the two-slit experiment. One kind of interpretation imagines that a wave drifts out and breaks against the opposing wall. Then, a collapse occurs, where one position is chosen, with probabilities depending on the results of the wave, and finally the particle is placed at the chosen end point. Meanwhile, the wave itself vanishes into the nothingness from which it came.

You've probably guessed that I find this silly, and from my reading it isn't what the Copenhagen Interpretation means. Or at least it isn't a necessary way to resolve the ambiguity of the Copenhagen Interpretation. Instead, I think that the wave isn't real, it's just a mathematical description of the probability of the particle arriving in various positions. But then, I'm sure you would ask, what is my explanation for the pattern of probabilities given by the double slit experiment?

And to this, I say that there isn't any. Or at least, there isn't any explanation that resolves this weirdness by using everyday classical mechanisms, like physical waves. There is also no pre-relativistic explanations involving instant transfers of information. If there is any explanation, it's in terms of more basic objects that display the same kind of weird behaviour.

I think that the only reason this seems wrong to so many people is that in our everyday experience things just don't behave in this way. There's a temptation to assume that everyday behaviour describes how thing must be, at least at the most basic level. And that any true explanation must reduce these kind of weird quantum correlations to the kinds of correlations we're used to. Even if that means describing things in a way that makes no sense under relativity.

Of course, I think the EPR paradox is just the same kind of situation as the double slit experiment. There is no explanation of the kind you want. But there's also no real wave or instantaneous transmission. There's only a weird correlation that we can't explain with reference to the kinds of things we see in everyday life.

[Schneibster]

You interpret it as causal behaviour, because when you think about a non-causal correlation, you are limiting yourself to ones that can be achieved in classical physics.Hee hee, this is the really interesting part of the conversation. I have been having an argument with Jesse about whether causality is defined for quantum mechanics or not. It has been an extremely long and convoluted conversation, and unfortunately it is at this time unfinished.

I have some ideas about causality, and I am planning some serious study of it as well (I have a couple of books in mind). Based on the definitions I have found so far, it appears that causality has meaning in both the classical and the quantum realms; however, in the quantum realm, certain types of causality violation are allowed under certain circumstances. The EPR experiment and its brethren show such violations, and create such circumstances. The mechanisms by which these violations occur are unknown at this time, but the circumstances seem (so far) to be well constrained.

The very idea of "non-local" interactions is an explicit violation of causality which is not permitted in classical physics. I am not at all sure of what you mean when you say, "when you think about a non-causal correlation, you are limiting yourself to ones that can be achieved in classical physics," when the phenomena under consideration are explicitly non-classical. I'm not entirely convinced you are either. Please consider this carefully.

Please do not confuse the statement that there is a causal relationship for the statement that there is or is not a causal violation in that relationship; they are separate, though related, concepts.

The fact that there is a causal interaction is proven by the fact that the correlation occurs. Correlation is a minimal definition of causality. The fact that this correlation cannot be a matter of a pre-existing characteristic or parameter value is proven by the violation of Bell's Inequality. If the mensurable were a pre-existing characteristic, Bell's Inequality would not be violated. The fact that this causal interaction actually violates standard causality is proven by the location of the interacting events at a spacelike distance from one another. If the causal interaction did not violate causality, then it would not manifest across spacelike distances.

A correlation is an effect. Detection of spin is an action; so is creation of a photon pair. Therefore, since we have two actions connected via a third action that have an effect on one another, we have an interaction by definition.

By reaction, I meant the start of the experiment where the particles initially fly apart. Now that I think about it, that's probably not necessarily a reaction. But I'm certainly not claiming that the particles are "reacting" at a distance. That would be a nonlocal interaction!I'm really looking to match up terminology here, so that we both know what the other is talking about as well as ourselves!

Do you have some word you prefer to the fairly neutral "interaction" to describe these phenomena? Please read my previous post again and respond to it to communicate, rather than to argue. We need to communicate effectively before we can argue to any conclusion.

This is called a "hidden variable" interpretation.No. My facetious theory involving tiny sub-particles shows results that can't be explained by hidden variables any more than can the Aspect result. Of course, the trick is that my explanation assumed that quantum weirdness was just a way particles could behave, and we could explain Aspect without reducing the weirdness to being the result of some more intuitive behaviour.Actually, I prefer "hypothetical" to facetious, but I understand what you mean: it was an example, not a serious proposal. I understood that before; but what you have not realized is that because the photons were carrying these other hypothetical particles along inseparably, the parameters of these hypothetical particles constituted hidden variables.

The reason that I say that you are apparently not clear on the difference is because you are referring to particle physics as classical.I think that if you imagine particles as behaving as Newton would have described, or Newton + General Relativity, then that's thinking in terms of classical physics. But particles do not behave that way, and any deductions you might make about their behavior using such a model would be wrong. What is the point?

Of course, you personally aren't thinking entirely in terms of classical physics, because you allow properties like spin. But the kind of explanations you accept are still limited by the expectations of classical physics.I disagree. You have failed to comprehend that the kind of explanations I will accept; but even that is not the problem. The problem is that you have failed to comprehend that local means specifically that the interactions must all be classical! So what you are saying when you say that quantum physics does not contain non-locality is that it is all classical in terms of causality. You are essentially not allowing there to be any non-classical causality, of the type that is clearly demonstrated by the Aspect EPR realization. That is what I am arguing against. What I am saying is that there can be causality that is explicitly not a matter of an interaction involving any field we are familiar with, nor the exchange of any quanta that we are familiar with. Unless or until such a field or interaction is found, the phenomena described as "non-local" will have no classical explanation.

Did we cross wires somewhere? I'm beginning to think we might have. Are you entirely certain you understand what non-local interaction means? I am not at all certain you do, since you seem to be arguing that the existence of non-local interaction is proof that non-local interaction does not exist!

But because Bell's Inequality is violated by the particles, we know absolutely for certain that the particles did not have the corresponding values until we measured them! And this is not a theory- it is an observable, provable physical fact!If you are talking about measured values, like spin, then I don't think the particles had these values before you measured them, corresponding or not. That's certainly the Copenhagen Interpretation.So how did they wind up with correlated values? That is the point. And the answer is, "through a non-local interaction." Why? Because without an interaction of some kind, there can be no correlation! Remember, correlation itself is minimal proof of causality, and anti-correlation of the negative (i.e., proof that there is no other correlated potential cause- something that we also have here) is further proof.

Any theory we come up with must explain why this is. Any theory that fails to explain it also fails to describe reality.
Finally we may have reached the crux of the argument. What constitutes an explanation, and when, if ever, are we allowed to say that there is simply no explanation? More on this below.OK, we'll discuss it there.

After writing this, I'm not entirely certain that what you just said isn't exactly the same as what I just said, with one exception: you are maintaining that there is no action at a distance, but the Aspect experiment says you are wrong and there is. I really think that you should very carefully consider what you said, and what I just said, and see whether you might not be wrong to be maintaining this.I really believe that I understand what you are saying, and why your saying it. I just think you're making an assumption that isn't warranted, and that you can't believe that I'm not making the same assumption because it seems so obviously true. And perhaps it is true, and perhaps I'm wrong. But we haven't really started event to debate that issue.Well, I have identified what I believe is the assumption I was making that was wrong: I assumed you knew what locality was, and therefore what non-local means!

Here is a statement of the Principle of Locality from Wikipedia (http://en.wikipedia.org/wiki/Principle_of_locality): "In physics, the principle of locality is that distant objects cannot have direct influence on one another: an object is influenced directly only by its immediate surroundings." In other words, if we show that physics describes influences that operate between distant objects, specifically spacelike separated objects, and especially if we can do experiments that show such influences, then we have explicitly proven that physics is non-local.

I really think that you need to consider that definition of locality carefully. I also think that there is confusing further material in that article; local realism is discussed, along with the fact that despite the fact that an influence can be transmitted superluminally, information cannot. Because of this, quantum field theory is often defined as a local theory; but actually, by the above definition of locality, it is non-local!

Here's some more supporting information: classical physics represents everything except electromagnetic and gravitic interactions as direct contact forces, i.e. local. And when amplified by Maxwell's Equations, or the quantum theory, electromagnetic interactions are rendered local either by the field equations or by exchange of quanta. Finally, relativity renders gravity local by describing it as curvature of spacetime. And the strong and weak forces are rendered local because they also involve quantum exchanges. So every force and every action is local. Either quantum exchange, or fields, or spacetime curvature, or physical contact, mediates each interaction and makes it local.

But these other interactions we are discussing from the EPR experiment and its cousins, despite the fact that they meet other criteria for causality, i.e. correlation, negative anti-correlation/dependence, and consistency, do not meet the requirements for locality. They also do not necessarily meet other causality requirements, i.e. sequentiality and an action-based explanation (which may be the same as locality- I am still thinking about this). Thus, we must conclude that there is some influence that we cannot describe in terms of any existing force or field or contact that nevertheless exerts a causal influence between these events. This influence is a "non-local interaction."

I don't think you're getting that the quantum mechanical description of these interactions is that they are non-local. The most widely accepted interpretation of quantum mechanics, the Copenhagen Interpretation, explicitly describes the interaction between the two particles in EPR as non-local.Well, I think there's some ambiguity. It might help to consider the two-slit experiment. One kind of interpretation imagines that a wave drifts out and breaks against the opposing wall. Then, a collapse occurs, where one position is chosen, with probabilities depending on the results of the wave, and finally the particle is placed at the chosen end point. Meanwhile, the wave itself vanishes into the nothingness from which it came.

You've probably guessed that I find this silly, and from my reading it isn't what the Copenhagen Interpretation means. Or at least it isn't a necessary way to resolve the ambiguity of the Copenhagen Interpretation. Instead, I think that the wave isn't real, it's just a mathematical description of the probability of the particle arriving in various positions. But then, I'm sure you would ask, what is my explanation for the pattern of probabilities given by the double slit experiment?

And to this, I say that there isn't any. Or at least, there isn't any explanation that resolves this weirdness by using everyday classical mechanisms, like physical waves. Precisely- the explanation is non-local!

There is also no pre-relativistic explanations involving instant transfers of information. If there is any explanation, it's in terms of more basic objects that display the same kind of weird behaviour.I partly disagree. We have yet to find a mechanism that mediates this interaction; but we have positive proof that the interaction exists, and we have many parameters that constrain the possible explanations. Whether this will turn out to be mediated by more basic objects, or by some other as-yet unimagined effect, remains to be seen. But until we have some explanation, I believe we need to keep looking for one.

I think that the only reason this seems wrong to so many people is that in our everyday experience things just don't behave in this way. There's a temptation to assume that everyday behaviour describes how thing must be, at least at the most basic level. And that any true explanation must reduce these kind of weird quantum correlations to the kinds of correlations we're used to. Even if that means describing things in a way that makes no sense under relativity.I agree with you there. Most people have not developed any kind of sense of the rules of quantum behavior. I am still working on mine- but I am not sure that there is any proof that there is not a local explanation of these non-local effects, that we just don't understand yet. Nor am I convinced that there must be an explanation of that type; what I am convinced of is that there must be some explanation of these effects, and that there is none that fully meets the requirement (although Many Worlds comes close- if I could just get past that "alternative matching").

Of course, I think the EPR paradox is just the same kind of situation as the double slit experiment. There is no explanation of the kind you want. I remain unconvinced that you know what kind I "want." Or even that I "want" one- at least in the sense I think you mean this.

But there's also no real wave or instantaneous transmission. There's only a weird correlation that we can't explain with reference to the kinds of things we see in everyday life.Gravity is a weird correlation we can't explain with reference to the kinds of things we see in everyday life- take a look at Newton's Law of Universal Gravitation, which makes absolutely no statements about what gravity is, only statements about how it behaves. But, along comes relativity, and now we understand all about gravity. Or so we think until we try to construct a quantum theory of it!

Now, how about the weak force? There's a really weird one. It doesn't even result in an attraction or a repulsion- it just changes quark flavors into one another, with lepton pair leftovers. What does this have to do with anything that we can see? Very little, and only under special conditions. Yet, there remains a correlation- and a bunch of conservation laws to go with it. There is nothing we see in macroscopic physics that works anything like the weak force. Yet, it has a complete explanation.

So, you see, weird correlations and explanations for them are already part of physics. Nothing out of the ordinary here. Why should this new effect be any different?

[sodium]

In general, we think of objects as having properties. These properties are based on what we get from measurements. For example, when I say something is green, I mean, it has whatever property it takes so that I recognize it as green. That's what it means for something to be green.

Of course, there's a recognition here that a test can fail. Green objects fail the green test in the dark. So, our assignment of properties is based on ideal conditions. But we still believe that if we can perfectly carry out the test, and the conditions are in our favour, we can get the correct result.

In the quantum world, however, there is a certain amount of randomness that can't be avoided. This causes a problem for assigning properties. For example, it would be nice to be able to say that a particle has the property of having positive x spin, but there is no test that corresponds to that alleged property, because there is no test that we can use to reliably test for it. However, we might say that there are some properties of the particle, some possible states for the particle to be in, that the x spin measurement is more likely. We would have to list the possible states of the particle, and how they correspond to the probability of each resulting measurement. But we could no longer identify the state with the measurement. Or at least it would be arbitrary to do so.

So, in this view (which isn't mine) a particle has a state. That state may be randomly assigned at some point, like when a particle decays. The particle moves along, and then eventually has some interaction where the state gets measured. This measurement doesn't necessarily reveal the state, because there may be some randomness involved. The outcome is, however, based on the state.

This is a model of reality that works very well for our day to day lives. In fact, even the indeterminancy is intuitive, because a lot of the interactions we observe are unpredictable to us, even if they aren't truly random if we had perfect knowledge.

But in a case like the EPR paradox, and even the double slit experiment, this model doesn't work very well. In order to maintain it, we either need instant communication across distances, or we need particles to know what is going to happen to them in the future. I doubt you share my view of these solutions as ugly patches for a bad model, but that's the position I've taken.

So, what was that model again? It's such an intuitive view, that it may not seem like any assumptions are being made at all, but what I'm describing is a view in which it makes sense to claim that all objects have internal states, and that all future interactions must be based on the internal state of an object. What I'm claiming is that we have to give up not only the idea that particles have spin before they are measured, but that they have any internal states or properties that correspond to the propensity for giving particular measurements. Instead, all we have is a description of the natural world that says that when particles are created or influenced in various ways, that later measurements will reveal a particular pattern. When this happens, we say that the earlier interaction caused the later measurement. Note that the difference in time and space between cause and effect always obeys the principle of locality.

Consider the double slit experiment. There's no wave, in my view, that moves from where the particle is emitted to the far wall where it is measured. There's only a pattern between the cause (the inital emission) and the effect, the "measurement" when the particle hits the wall. I'm using the word "measurement" in a broad sense, and don't mean to imply any conscious interaction.

The same is true of EPR. The particle has no spin before it is measured. Not only does it have no spin, but it has no internal properties that correspond to a propensity for particular measurements of spin. There is only a known pattern between the initial cause (the particle decay), and the eventual effects, of measured spin. This cause and these effects obey locality.

But isn't there a correlation between the spin of the two particles, and isn't correlation a minimal kind of interaction? No. If I discover a pattern in which a baseball that lands fair never lands foul, I don't claim that its landing fair caused in to not land foul. I don't imagine an instantaneous message being transmitted from where it landed to everywhere it didn't land. I just say that there is a correlation, and that likely means that there is some common cause to both observations. And of course, there is a common cause for both observed spins, and that common cause is the explanation for the correlation, and that common cause obeys locality.

[sodium]

I am not at all sure of what you mean when you say, "when you think about a non-causal correlation, you are limiting yourself to ones that can be achieved in classical physics," when the phenomena under consideration are explicitly non-classical. I'm not entirely convinced you are either. Please consider this carefully.
I hope I've explained this in the post I just made. I'm not really implying that your worldview is entirely classical. Just that some classical assumptions are affecting your interpretation. I'm sure this is true of me too, of course.
The fact that this correlation cannot be a matter of a pre-existing characteristic or parameter value is proven by the violation of Bell's Inequality.
Right. EPR is incompatible with the state/property based view that I mention in the post I just made.
The fact that this causal interaction actually violates standard causality is proven by the location of the interacting events at a spacelike distance from one another. If the causal interaction did not violate causality, then it would not manifest across spacelike distances.
I think this confuses the results with one possible explanation of the results.
I'm really looking to match up terminology here, so that we both know what the other is talking about as well as ourselves!
I think we basically agree on how we use the word "interaction", but you can't believe I really want to say what I'm saying.

Let's say I have a strange habit. Every year, I get one red ball, and one blue ball, and place each in its own box. I mail these off to you and a friend. Let's say you get sent a blue ball, and your friend gets sent a red ball. My sending the balls causes you to get them. That is an interaction, or a series of interactions. As soon as you open your box, you know what coloured ball your friend gets. That is not an interaction, however. Your discovery of the red ball doesn't cause your friend to get a blue ball. This is a correlation, but not an interaction. Now I know this isn't exactly what EPR is like, and I understand the reasons why people think this isn't merely a correlation, but a nonlocal interaction. But I'm claiming, against this view, that it really is a correlation and not an interaction after all.
Actually, I prefer "hypothetical" to facetious, but I understand what you mean: it was an example, not a serious proposal. I understood that before; but what you have not realized is that because the photons were carrying these other hypothetical particles along inseparably, the parameters of these hypothetical particles constituted hidden variables.
The particles have no hidden variables that explain the phenomenon. I think this tangent isn't really working out.
The problem is that you have failed to comprehend that local means specifically that the interactions must all be classical!
Not really. If we take Relativistic classical physics, then this implies locality. But locality doesn't imply relativistic classical physics. It could be non-classical for other reasons.
I am not at all certain you do, since you seem to be arguing that the existence of non-local interaction is proof that non-local interaction does not exist!
Uh, no. I don't believe in nonlocal interaction, although it's a conclusion you draw from a premise I do accept. I haven't tried to prove that nonlocal interaction doesn't occur in nature, only that my interpretation is consistent with the view that nonlocal interaction doesn't occur. Of course, I believe that locality holds, but there's no point in trying to argue for my position when I still haven't succeeded in explaining it.
So how did they wind up with correlated values? That is the point. And the answer is, "through a non-local interaction." Why? Because without an interaction of some kind, there can be no correlation!
There is an interaction in the sense that both spin measurements are partly the result of the initial decay. So we don't have correlation without some common causation. But we do have a correlation that isn't explained, in the sense that we can't tell a story of the kind I mention in my last post, which explains this in terms of objects moving around with states.
what I am convinced of is that there must be some explanation of these effects
So is it explanations all the way down, then? I mean, every time we explain something by reducing it to something else, which needs further explanation. What sort of bedrock explanation-that-needs-no-explanation are you hoping for?

[Schneibster]

In general, we think of objects as having properties. These properties are based on what we get from measurements. For example, when I say something is green, I mean, it has whatever property it takes so that I recognize it as green. That's what it means for something to be green.We can be more specific about what it means to be green: it means that the object reflects green light, of about 500 to 565 nm wavelengths, with greater efficiency than other colors.

Of course, there's a recognition here that a test can fail. Green objects fail the green test in the dark. So, our assignment of properties is based on ideal conditions. But we still believe that if we can perfectly carry out the test, and the conditions are in our favour, we can get the correct result.In other words, that a green object will still reflect green light in the dark. This is a logical error; do you see the problem? There is no light; therefore, the object cannot be green, because the definition of green is that it reflects green light more than any other wavelength.

In the quantum world, however, there is a certain amount of randomness that can't be avoided. This causes a problem for assigning properties. For example, it would be nice to be able to say that a particle has the property of having positive x spin, but there is no test that corresponds to that alleged property, because there is no test that we can use to reliably test for it. Yes, there is; measurement. However, we also know that the particle has a probability of the spin being positive at angle x; since there are only two available spins, that probability is 0.5. Which means that if we measure it right now, and then measure a bunch more quanta under exactly the same conditions (like at this same point, generated under the same conditions, everything else the same but a later time) we will find that the spin is positive in x 50% of the time and negative in x 50% of the time. So quanta don't have definite properties; they have probabilities of having certain properties. But according to the violation of Bell's Inequality, that is all they have; they do not have the actual properties until they are measured. In some interpretations, measurement is not necessary; any interaction in which the property is important serves as a measurement, i.e. forces the probability to convert to a definite value.

However, we might say that there are some properties of the particle, some possible states for the particle to be in, that the x spin measurement is more likely. We would have to list the possible states of the particle, and how they correspond to the probability of each resulting measurement. But we could no longer identify the state with the measurement. Or at least it would be arbitrary to do so.I presume you meant, "that the positive spin measurement at angle x is more likely."

This listing of alternatives is an implicit part of quantum mechanics. The properties of a particle cannot be known without measurement; all that can be known about the particle without measurement is the probability that it will have a particular value when it is measured.

The observable variables of a particle's state are called "eigenvectors" or "eigenstates." Potential values of the eigenstate that have a probability greater than zero are called "eigenvalues." Thus, the particle has a eigenstate for every observable value, like spin, velocity, etc., and for each eigenstate there is a set of eigenvalues, corresponding only to the possible (not the impossible) values of that observable. When an observation occurs, the eigenstate is reduced to a state, in which the former eigenvalues are converted from their original probability to a probability of zero, except one eigenvalue which has a probability of one. That is the measured value of that observable for that particle at that time and place. So, as you can see, the state is implicitly identified with the measurement; there is no state until the measurement is taken. Only an eigenstate, and its eigenvalues, exist in the absence of a measurement.

So, in this view (which isn't mine) a particle has a state. Then this view is not that of quantum physics, which says that a particle has only an eigenstate unless it is being measured, in which case it has a state.

That state may be randomly assigned at some point, like when a particle decays. The particle moves along, and then eventually has some interaction where the state gets measured. This measurement doesn't necessarily reveal the state, because there may be some randomness involved. The outcome is, however, based on the state. No, the outcome is based upon the eigenstate, and the probabilities of its eigenvalues. The particle does not have any state while it is moving along unmeasured.

This is a model of reality that works very well for our day to day lives. In fact, even the indeterminancy is intuitive, because a lot of the interactions we observe are unpredictable to us, even if they aren't truly random if we had perfect knowledge.It works in our day to day lives only for objects that are not particles; for particles, it never works. That is not a correct view of how particles behave, and we know for certain it is not because Bell's Inequality is violated in a situation where correlation occurs.

But in a case like the EPR paradox, and even the double slit experiment, this model doesn't work very well. That's because it doesn't describe reality.

In order to maintain it, we either need instant communication across distances, or we need particles to know what is going to happen to them in the future. I doubt you share my view of these solutions as ugly patches for a bad model, but that's the position I've taken.What you have failed to understand is that we need these same things in order to account for the correlation, even in this unfamiliar world of eigenstates and eigenvalues (not the one you were propounding above, but the real world); and that we need them only and specifically because of the violation of Bell's Inequality, because otherwise we could assume that the particles all had values the whole time, but we just didn't know it, and they showed those values when we measured them. Using this model, all interactions (including EPR and the dual-slit experiment) would be local, because there would be no need for a correlation to exist between spacelike separated points that was not the result of a pre-existing value. Unfortunately, as I say, Bell's Inequality is violated, and we cannot account for the correlation by a pre-existing value. Thus, we must account for it by some other means; and whatever that means is, it must act across a spacelike distance, and thus it must not be local.

So, what was that model again? It's such an intuitive view, that it may not seem like any assumptions are being made at all, but what I'm describing is a view in which it makes sense to claim that all objects have internal states, and that all future interactions must be based on the internal state of an object. What I'm claiming is that we have to give up not only the idea that particles have spin before they are measured, but that they have any internal states or properties that correspond to the propensity for giving particular measurements. Instead, all we have is a description of the natural world that says that when particles are created or influenced in various ways, that later measurements will reveal a particular pattern. When this happens, we say that the earlier interaction caused the later measurement. Note that the difference in time and space between cause and effect always obeys the principle of locality. But, you see, it does not always obey the principle of locality; that is because we have observed an effect that is connected to its apparent cause by a means that we have no description of. Until we can describe that means, we cannot in good conscience state that there actually is a means; and if there is no means, then the effect cannot be local.

Consider the double slit experiment. There's no wave, in my view, that moves from where the particle is emitted to the far wall where it is measured. Then what causes the observed interference? Interference is a signature effect of waves of all kinds; it is produced by both transverse and longitudinal waves; and it is produced by all waves. There is no wave that does not produce interference when it interacts with another wave of the same kind; and there is no kind of interference that produces interference fringes that we have ever seen that is not caused by waves. If you see interference, then you are looking at waves, period. If you wish, you may attempt to produce an example of interference (and I am speaking specifically here of interference fringes, the specific effect produced in the dual-slit experiment) that is not a result of waves; you will find there is no such effect. All interference is produced by waves. To observe interference is sufficient proof that waves are present. Is there any other way I need to say it so that you get the point?

There's only a pattern between the cause (the inital emission) and the effect, the "measurement" when the particle hits the wall. I'm using the word "measurement" in a broad sense, and don't mean to imply any conscious interaction.That pattern is an interference pattern, and is produced solely and only by waves, period. There is no other method of producing an interference pattern. What is most amazing is that the interference pattern can be produced by the pattern of single impacts of particles- thus, we know that the wave has physical existence because there is interference, but we can clearly see the individual particle impacts.

Feynman said that you can see all of physics in this experiment. I tend to agree with him.

The same is true of EPR. The particle has no spin before it is measured. Not only does it have no spin, but it has no internal properties that correspond to a propensity for particular measurements of spin. You are correct, as proven by Bell's Inequality.

There is only a known pattern between the initial cause (the particle decay), and the eventual effects, of measured spin. This cause and these effects obey locality.You are wrong. You cannot show a mechanism by which this cause accomplishes these effects; that is required to maintain the presumption of locality. I know you cannot show this, because no one else can either, and if you could, you would be receiving a call from a certain committee in Stockholm regarding a prize that they give out annually for discoveries like that.

That's why they call it "non-local."

But isn't there a correlation between the spin of the two particles, and isn't correlation a minimal kind of interaction? No. So, you claim to have discovered the one and only 100% correlation (because this one is not a probability- the probability is 1) that satisfies all the tests for a statistical correlation, including negative correlation of alternatives, and repeatability, and all the rest, that does not represent a causal relationship? Again, I sincerely doubt it. And if the causal relation is with the producing event, by what means is the relation propagated to the two detection events? This is no more explainable than the presumption of causation between the two events! There is no pre-existing value, so there is no means by which the "causative" event could produce either result.

I keep telling you, you aren't getting this. You now try to push the causation back to the origin event; but there still is no mechanism by which it can cause the effect we see, and that is guaranteed by the violation of Bell's Inequality. The purpose of physics is to explain things; and this is unexplained.

Let me take another type of example to put this in perspective. Why is the value of the fine structure constant 1/137.03599911(42)? Why not 1/136, or 1/5, or 10^35? This question has no answer. We assume that it is not that there is no answer possible; it is that we do not know the answer. Similarly, there is no answer to why the spin values of the two photons in the Aspect experiment should be correlated. Again, we assume that it is not that there is no answer possible; it is that we do not know the answer. What we do know is that the answer (if there is one) is not that the photons had those spin values all the time. Remember: there is no guarantee that there is an answer!

The definition of an effect for which there is no explanation is that it is "non-local." Locality assumes that there is a known cause that creates the effect, and that known cause is the result of a local interaction, and potentially the movement of the consequents of the local interaction to widely separated points. But in this case, there is no possibility of moving the consequences; a photon has only a very few characteristics, and none of them can be responsible for this measurement. Therefore, the effect is "non-local" until we can determine a mechanism. Just as gravity was in Newton's Theory of Universal Gravitation; just as electricity was until Franklin proposed the "electric fluid" (and by the way got the signs wrong), and just as magnetism was until Faraday described the "lines of force."

If I discover a pattern in which a baseball that lands fair never lands foul, I don't claim that its landing fair caused in to not land foul. I don't imagine an instantaneous message being transmitted from where it landed to everywhere it didn't land. I just say that there is a correlation, and that likely means that there is some common cause to both observations. And of course, there is a common cause for both observed spins, and that common cause is the explanation for the correlation, and that common cause obeys locality.But there is no connection between the cause and the effect. How precisely does this cause operate to create these effects? Until you can explain that, these types of events will continue to be called non-local.

Personally, I believe you may be correct; there may be some sort of local thing that happens that creates the observed effects from the apparent cause. But the problem is that we don't know that now, and it remains possible that there is no answer to this question; and until we know there is an answer, by discovering what that answer is, we cannot assume what form it will take, and must therefore continue to describe this effect as "non-local."

[Schneibster]

I think that we really did have our wires crossed; we were saying essentially the same thing, but we got hosed around because we were using different terminology. I don't think you can prove what you are maintaining; but I believe it too. "No non-local interactions" has been a rallying cry ever since Newton penned the Theory of Universal Gravitation, and in that case it eventually resulted in relativity.

So is it explanations all the way down, then? I mean, every time we explain something by reducing it to something else, which needs further explanation. What sort of bedrock explanation-that-needs-no-explanation are you hoping for?Well, when the universe was created, it had no dimensions; it was just a point. Then, (something we don't yet know the explanation for occurred), and it resulted in inflation. During inflation, the three large space and one time dimensions were blown up to be big; this was because (something else we don't know the explanation for).

You're following me here, I'm sure. What I'm looking for is

A calcium atom is excited until it is in the proper state to emit two photons in opposite directions, of opposite spin polarization. Because the state of the atom is such that the total angular momentum of the electron shell that emits the photons is zero before and after emission, we know that the two photons must have orthogonal spins. However, because of the violation of Bell's Inequality, we also know that the spins cannot have correlated values until they are measured. The photons travel to a pair of points with spacelike separation. We measure them, and due to (something we don't know the explanation for), we measure correlated spins.

What I want is the something we don't know the explanation for. And until I get it, I see no choice but to call it "non-local."

[sodium]

You are wrong. You cannot show a mechanism by which this cause accomplishes these effects; that is required to maintain the presumption of locality.
But what if there is no mechanism? This is what I was trying to get at with my comment about what would constitute an acceptable end-point to the search for explanations. At some point, don't we expect to reach a level of understanding of the universe where all we have are raw facts with no underlying mechanism? That's certainly what the lowest level would look like, and we have to accept that one day we may find it.

I don't think we necessarily have the lowest level now, but I'm claiming that it may well be that even if we go lower and lower, this weird effect may not disappear. It may be fundamental, with no underlying mechanism. And it seems to me that the reason you think this must have an underlying mechanism is that we don't have these kinds of effects in everyday life, and you want the lowest level of explanation to be intuitive. If it isn't intuitive, then that suggests there is something we'd want to have explained, which would require a lower level. But I don't see that as a real problem. I mean, do we expect that at a lower level the strange Einstein relativity effects will disappear and we'll be left with good old-fashioned Newtonian common sense?

[Schneibster]

But what if there is no mechanism? I'll believe it when it is a theoretical consideration, at minimum. And if that turns out to be the case, I will continue to call it "non-local!"

At some point, don't we expect to reach a level of understanding of the universe where all we have are raw facts with no underlying mechanism? That's certainly what the lowest level would look like, and we have to accept that one day we may find it.Actually, string theory seems to bring up the possibility that we may be able to do it with only a very few free variables. But that's still a ways off. Certainly, physicists hope to reduce the free parameters to well below the current 19 (or 29, including the neutrino masses, if they're right about that).

I don't think we necessarily have the lowest level now, but I'm claiming that it may well be that even if we go lower and lower, this weird effect may not disappear. It may be fundamental, with no underlying mechanism. Sure; its entirely possible. But I have to point out that if it doesn't, then I will continue to call it non-local.

And it seems to me that the reason you think this must have an underlying mechanism is that we don't have these kinds of effects in everyday life, and you want the lowest level of explanation to be intuitive. If it isn't intuitive, then that suggests there is something we'd want to have explained, which would require a lower level. If it turns out this way, it will be the first truly non-local interaction we have ever found. We have found several others we thought were non-local, but none of them actually turned out to be.

Interestingly, this is my position; you are the one who chose to say, "it's local." If it's local, then it has an underlying mechanism, a lower-level explanation; only if its non-local does it not require one.

But I don't see that as a real problem. I mean, do we expect that at a lower level the strange Einstein relativity effects will disappear and we'll be left with good old-fashioned Newtonian common sense?Nope- but none of the weird relativity stuff is non-local either. I say again, it would be the only one.

Personally, I'm in favor of five forces. I think this may be the lever that prys open the box of goodies. We'll have to see what people come up with.

[sodium]

Sure; its entirely possible. But I have to point out that if it doesn't, then I will continue to call it non-local.
So, even if there is no internal nonlocal mechanism for the result, you would continue to call this behaviour nonlocal. I hope we're agreed that a nonlocal result would involve some kind of superluminal influence or causation.

Given that we are considering the hypothetical situation where there is no internal, hidden superluminal influence, you must believe that the interaction of the EPR experiment is itself an example of superluminal influence or causation. Note that you are not claiming, I hope, that it is merely evidence for such a thing. You aren't saying just that you need superluminal influence or causation to explain the result, but that the result itself is superluminal influence or causation.

For that, I suppose, we might want to try to define causation (or influence). I'm not really pushing for any particular definition, and I hope you'll suggest your own, but here's an example. If whenever an event of type X exists somewhere in space-time we know that an event of type Y exists somewhere in space-time, and we are able (at least sometimes) to voluntarily cause X to occur, then we say that X causes Y.

For example, I might know that if a ball gets hit with a bat (under some conditions), the ball goes flying off immediately afterwards. And I know I can voluntarily hit the ball with the bat, so I say that hitting the ball with the bat causes the ball to go flying off.

This is why the transactional interpretation is said to have causality back in time. It states that certain states of the particle correspond to certain later measurements. Since we can control the measurements, we describe this as causation working backwards.

Anyway, if you use this definition of causation, then there is no causation between the measurements in EPR, only causation between the decay and the measurements. So there is no nonlocality.

Do you have some other definition of causality/influence under which it makes sense to say that locality is violated under my interpretation of what happens in EPR?

[Schneibster]

So, even if there is no internal nonlocal mechanism for the result, you would continue to call this behaviour nonlocal. I hope we're agreed that a nonlocal result would involve some kind of superluminal influence or causation.Errrmmm, "internal nonlocal mechanism?" What is that? You mean like, when these two photons get emitted by this calcium atom, and they have no state or other information about them that makes them correlated, and you can check them with Bell's Inequality and they are not correlated, but then when you measure them in the same plane, they are correlated?

I don't think you thought this through.

Given that we are considering the hypothetical situation where there is no internal, hidden superluminal influence, you must believe that the interaction of the EPR experiment is itself an example of superluminal influence or causation. Not necessarily- but given that we know for certain that there can be no hidden internal state, because of the violation of Bell's Inequality, it's hard to see how it can be anything else. Clearly, the most likely thing is that there is some extra information available as a parameter of the photon (or electron, depending on what version of the experiment you run) that expresses itself only when it is necessary that there be correlation, but I'm damned if I can see how that works. That's the only way I can see that this can be a local interaction. But we've never seen anything that wasn't a local interaction, when we got done looking it over. It's quite a puzzle.

Note that you are not claiming, I hope, that it is merely evidence for such a thing. You aren't saying just that you need superluminal influence or causation to explain the result, but that the result itself is superluminal influence or causation.At this point, I can't see any other way to interpret it, except that it is evidence for some kind of influence that we don't know about. That influence may be superluminal if it involves the eigenstates of the two photons collapsing into correlated states without prior correlation; or it may be that there is some extra hidden parameter, not spin, that causes it so that it is local. Certainly, however, it is evidence of one kind or another of influence that we do not understand. And equally certainly, there is associated evidence that it is not a preexisting spin value.

For that, I suppose, we might want to try to define causation (or influence). I'm not really pushing for any particular definition, and I hope you'll suggest your own, but here's an example. I have a definition running on two other threads- correlation is not causation, and the delayed choice quantum eraser- that is worth your time to take a look at, I think. I don't like yours- it's too sparse. We can use it to prove that things we know are not causal are.

This is why the transactional interpretation is said to have causality back in time. It states that certain states of the particle correspond to certain later measurements. Since we can control the measurements, we describe this as causation working backwards.

Anyway, if you use this definition of causation, then there is no causation between the measurements in EPR, only causation between the decay and the measurements. So there is no nonlocality.Yes, because of the advanced wave. I am familiar with Jack Cramer's TI.

Do you have some other definition of causality/influence under which it makes sense to say that locality is violated under my interpretation of what happens in EPR?Unless you can show a means by which the initial correlation (caused by the fact that there is no momentum change in the electron shell of the calcium atom) can result in the correlation at measurement, CI (and friends) and Feynman's Many Paths both show it. MWI does not; and TI does not; but MWI has correlation of the alternatives, which is non-local, and TI is non-local in time (backward causality, aka the advanced wave). Bohm shows it as non-local, but Bohm shows everything as non-local.

[sodium]

I have a definition running on two other threads- correlation is not causation, and the delayed choice quantum eraser- that is worth your time to take a look at, I think. I don't like yours- it's too sparse. We can use it to prove that things we know are not causal are.
I've checked out your definition here (http://www.iidb.org/vbb/newreply.php?do=newreply&p=2090997)

Let's say a meteoroid explodes into two pieces out in space. If I measure the momentum of one, I know the momentum of the other, since momentum must be conserved. This holds even if I wait until the pieces are at a great distance from each other.

So, in your opinion, does the momentum of the one piece cause the momentum of the other? Well, there certainly is a correlation. It isn't local, but you are willing to accept nonlocal causation in some cases. Of course, there may be another possible correlation involving the initial explosion. But I don't think you given any reason to prefer one correlation to the other in terms of identifying causation.

Consider the double-slit experiment. When the photon is absorbed in one position, we know that it isn't absorbed anywhere else. Would you claim that the absorption caused the lack of absorption elsewhere? If not, why not?

[Schneibster]

I've checked out your definition here (http://www.iidb.org/vbb/newreply.php?do=newreply&p=2090997)So what didja think???? :p

Let's say a meteoroid explodes into two pieces out in space. If I measure the momentum of one, I know the momentum of the other, since momentum must be conserved. This holds even if I wait until the pieces are at a great distance from each other.

So, in your opinion, does the momentum of the one piece cause the momentum of the other? Well, there certainly is a correlation. It isn't local, but you are willing to accept nonlocal causation in some cases. Of course, there may be another possible correlation involving the initial explosion. But I don't think you given any reason to prefer one correlation to the other in terms of identifying causation.But this is not the same as Aspect- there is state information carried along with the two pieces. It can be measured at any time, and Bell's Inequality will not be violated.

Now, in terms of causality, as I have defined it; first, the relations of the two momenta to one another:
1. There is correlation.
2. There is not negative correlation of alternatives! What about the correlation between the force of the explosion (there must be force, or the two pieces will just sit there) and the momentum of either piece?
3. There is repeatability.
4. There may or may not be sequentiality; it depends on whether you measure one first, the other first, or both at the same time. Because sequentiality is not necessarily met, and this is a macroscopic experiment, we conclude at this point that there is a causal connection between these, but it is not that one is the cause and the other an effect. These must both be effects of a common cause. But if we did not stop here, then
5. There is not locality. There is no local means by which the measurement of one could affect the measurement of the other, and this is not a microscopic experiment, so that absolutely rules out these being cause and effect. However, the correlation shows that they must be effects of a common cause.

Now, with the explosion (which provided a force) as the cause, and the momentum of each piece as an effect:
1. There is correlation.
2. There is negative correlation of alternatives; the only other correlated datum for either momentum is the other momentum, and we have already shown that the two momenta cannot be causes for one another, because of the failure to meet either the locality or the sequentiality criteria.
3. There is repeatability.
4. There is sequentiality; first there is the explosion, then there are the two pieces with momenta correlated to the force of the explosion.
5. There is locality; the explosion is in direct physical contact with both pieces.

Thus, we conclude that the momenta of the two pieces are both effects, and the explosion is the cause.

Note that we got a clue from the negative correlation of alternatives. B_Sharp turned me on to this, and he was dead-bang damn spankin right.

Consider the double-slit experiment. When the photon is absorbed in one position, we know that it isn't absorbed anywhere else. Would you claim that the absorption caused the lack of absorption elsewhere? If not, why not?
1. There is correlation.
2. There is not negative correlation of alternatives! Anything that affects the probability distribution of the photon affects its precise impact point, and the particular choice of impact point is not the only factor that affects that. The position of the slit, the distance between the slits, the precise wavelength of the photon's probability function, the distance between the slits and the detection screen... there are many possibilities. Note that this is our first hint that the fact that the photon was detected in one place, and the fact that it was not detected elsewhere, may be effects of a common cause, just as it was for the meteor experiment above.
3. There is quantum-style repeatability; many photons sent throught the experiment will show the expected probability distribution.
4. There is not sequentiality; the impact at one point is simultaneous with the non-impact at any other, at least from a viewpoint that sees the entire potential probability distribution of the photon as equidistant. However, this criterion can be violated in some quantum interactions, so we do not consider this conclusive.
5. There is not locality, because the impact is spacelike from other points within the probability distribution where the impact did not occur. However, again, this criterion can be violated without conclusive presumption of non-cause-to-effect phenomena.

We review the criteria, and determine that there is no conclusive evidence either way; however, the fact that the second criterion was not met combined with the fourth and fifth indicates that there is a high probability that we are not looking at cause and effect, but at effects of a common cause. If we can find a cause that meets 2, 4, or 5, we will have conclusive evidence that that is the cause and these are effects. Let's see, are there any candidates?

[sodium]

My problem with your attempt at defining causation is that it doesn't seem to be a definition at all. I can't tell whether you are attempting to describe criteria that explain how you will use the term (a definition), or you are trying to describe things that we may or may not say are experientially true of causation.

For example, you have criteria 4 of sequentiality. But surely you don't mean that you would actually claim that something was not causation, by definition, because it violated sequentiality. In fact, you don't do this. Instead, this is just something that is generally true of causation, so that when something is non-sequential, we have more reason to believe it isn't causation.

If you want to define causation, you have to describe what the term means. A description of what behaviour is evidence for causation is not the same thing, although a description can be part of a definition.

I think this creates a certain amount of confusion when considering a particular model for what goes on in an experiment. One argument I'm trying to make is that under one internally-consistent interpretation of the results, locality is preserved. This is not true if the model in fact contains something that is equivalent to causation, even if I call it something different. But you don't have a definition of causation; what you have is a bunch of tools for recognizing causation. So when you declare that causation occurs, you aren't saying that causation has occurred by definition, only that something has occurred, and you would tend to think that it most likely causation.

[sodium]

You mentioned the Transactional Interpretation previously, so I thought I might suggest a problem with it.

First, I'll attempt to review what the transactional interpretation is according to my understanding. The idea is that matter gives off special waves, both forward and backward in time. In something like the EPR experiment, the initial state of the particles is in part determined by the measurement that will be done on them. This is possible, because the measurements send waves back in time, and influence the initial particle decay.

Now, let's consider EPR, using particle spin. Let's say I take two arbitrary choices of measurement, such as x and y, or y and y, or any of the nine possible combinations. These measurement, I hope you will agree, will, on average, give an equal number of positive and negative results. So, for example, it is possible that my x and y measurements will both yield positive, or both negative, but if I do the experiment repeatedly, I will get about equal positives and negatives, for whichever measurements I choose.

But it is possible to set up an experiment along the basis of EPR, but which is biased toward positive readings. You measure one particle on the x axis. Now, you send a signal (for example by light beam) to where the other particle is going to be measured, only if you get a positive spin. The second measurement then uses the x axis if it gets no signal, and the z axis otherwise. If the first measurement is negative, no signal is sent, and the x axis is used again, guaranteeing a positive result. If the first measurement is positive, a signal is sent, and the z axis is used. Then, both positive and negative results are possible for the second measurement. So you'll never get two negative measurements, but you'll sometimes get two positives. As a result, this apparatus will on average measure more positive spins than negative.

Now, let's use my modified EPR apparatus, and consider what will happen under the transactional interpretation. At the time of particle decay, the decaying particle receives waves telling it what the eventual measurements will be. It then uses these waves to decide what each measurement should give. But as I mentioned above, once you have chosen an arbitrary pair of spin measurements, the result is unbiased in that it on average should produce as many positive as negative spin results. So that's what should happen here under the transactional interpretation. But as I also explained, what we actually expect to happen is that there should be a bias toward positive results. And after enough repetitions of the experiment, it should be obvious that this is what happens.

So where am I going wrong? I find it hard to believe that I'm the first person to notice this. Have I just got the transactional interpretation confused, or is there some way to wriggle out of this problem?

[Peter Soderqvist]

Einstein's local reality point of view is no longer sustainable, either locality or reality has to go! Locality means that causality (first cause, then effect) is obeyed in the sense that a cause of fission happens in the sun takes 8 minutes to travel at light speed to the earth and so make our earth an hot effect! Nonlocal in this example means that light travel faster than its light speed limits and so make our earth hotter in the past and so causality is violated! David Bohm maintain Einstein's reality of unmeasured quantum objects to have real values before measurement, by let this classical particle be informed instantaneously, thus non-locally by his pilot wave how it shall move about in space, and so the particle behave classically, meanwhile let his absolute light speed limits theory be modified into a statistical theory!

Copenhagen Interpretation maintain Einstein's locality by remove his reality from the picture, since the quantum particle is unreal until measured, and photon A and B's interaction in Alain Aspects experiment is nonlocal in the sense that it has no local, it simply doesn't exist somewhere, it is equally weird as if someone talking with another in the telephone without telephone signals or sound traveling between telephone A and B, if A is measured "real up", then B adjust instantaneously himself correspondingly to "real down", and nobody knows how it can be like that! :huh:

[Schneibster]

My problem with your attempt at defining causation is that it doesn't seem to be a definition at all. I can't tell whether you are attempting to describe criteria that explain how you will use the term (a definition), or you are trying to describe things that we may or may not say are experientially true of causation.Well, actually, I guess you missed where I said, "The minimal definition of causality in physics is shown by the use of correlation, negative correlation of alternatives, and repeatability in physics papers. I contend that these three requirements are the minimal definition of causality. Locality and sequentiality may also be present, or may not; they are always present (so far as we have been able to prove by experiment) in all experiments except for those of microscopic physics."

For example, you have criteria 4 of sequentiality. But surely you don't mean that you would actually claim that something was not causation, by definition, because it violated sequentiality. That depends on the circumstances. If it were a macroscopic experiment, certainly I would. Macroscopic violation of sequentiality is a causality violation. See above. Note also that the fact that it is possible to interpret the measurements in your meteor experiment as a causality violation would permit the conclusion that the two measurements cannot be cause and effect. I went on to analyze locality as well, and showed that it also is violated- but sequentiality was sufficient.

In fact, you don't do this. Instead, this is just something that is generally true of causation, so that when something is non-sequential, we have more reason to believe it isn't causation.In fact, I do precisely this, as shown above, if you read it.

If you want to define causation, you have to describe what the term means. A description of what behaviour is evidence for causation is not the same thing, although a description can be part of a definition.I described three behaviors out of those five that can be taken as a minimal definition of causality; they work in all cases in both the quantum and classical realms. Locality and sequentiality work in some cases in the quantum realm, and in some cases they do not- which of course refutes your premise that physics is not non-local. In the classical realm, locality and sequentiality always are obeyed in cause-to-effect relationships.

I think this creates a certain amount of confusion when considering a particular model for what goes on in an experiment. One argument I'm trying to make is that under one internally-consistent interpretation of the results, locality is preserved. Obviously, it is not. Your own dual-slit example proves it.

This is not true if the model in fact contains something that is equivalent to causation, even if I call it something different. But you don't have a definition of causation; what you have is a bunch of tools for recognizing causation. So when you declare that causation occurs, you aren't saying that causation has occurred by definition, only that something has occurred, and you would tend to think that it most likely causation.Defining the minimal characteristics of a phenomenon is a definition of the phenomenon. This is a tautology in physics. If you want to talk about philosophy, say so; but this is physics, and that's the title of the OP: "Is physics non-local." The definition is right there, if you will read it.

[Schneibster]

You mentioned the Transactional Interpretation previously, so I thought I might suggest a problem with it.

First, I'll attempt to review what the transactional interpretation is according to my understanding. The idea is that matter gives off special waves, both forward and backward in time. Ummmm, let's be very, very clear. According to QED, the most successfully predictive theory in the history of science, there are two types of solutions which lead to advanced and retarded waves in the papers published by Wheeler and Feynman on the wave equations of QED. The advanced solutions are generally ignored as non-physical, but there are significant problems in doing this, as Cramer shows in his TIQM (http://mist.npl.washington.edu/npl/int_rep/tiqm/TI_30.html#3.0) web site. The "special waves" are the electromagnetic wave equation, which as previously stated is part of the most successfully predictive theory in the history of science (not merely physics but all science).

In something like the EPR experiment, the initial state of the particles is in part determined by the measurement that will be done on them. This is possible, because the measurements send waves back in time, and influence the initial particle decay.This is incorrect. Please refer to Cramer's discussions of the application of TI to Wheeler's delayed-choice experiment (http://mist.npl.washington.edu/npl/int_rep/tiqm/TI_40.html#4.4), and photon transmission through polarizers (http://mist.npl.washington.edu/npl/int_rep/tiqm/TI_40.html#4.4).

Now, let's consider EPR, using particle spin. Let's say I take two arbitrary choices of measurement, such as x and y, or y and y, or any of the nine possible combinations. These measurement, I hope you will agree, will, on average, give an equal number of positive and negative results. So, for example, it is possible that my x and y measurements will both yield positive, or both negative, but if I do the experiment repeatedly, I will get about equal positives and negatives, for whichever measurements I choose.Note that whenever the measurements are in the same plane, you will always get correlated results. This skews the measurements if both same-plane and different-plane measurements are made, by the amount that is the percentage of measurements that are in the same plane.

But it is possible to set up an experiment along the basis of EPR, but which is biased toward positive readings. You measure one particle on the x axis. Now, you send a signal (for example by light beam) to where the other particle is going to be measured, only if you get a positive spin. The second measurement then uses the x axis if it gets no signal, and the z axis otherwise. If the first measurement is negative, no signal is sent, and the x axis is used again, guaranteeing a positive result. If the first measurement is positive, a signal is sent, and the z axis is used. Then, both positive and negative results are possible for the second measurement. So you'll never get two negative measurements, but you'll sometimes get two positives. As a result, this apparatus will on average measure more positive spins than negative.Because of the correlation, you will always measure a correlated value if you get a negative spin at the first measurement; you may or may not measure a correlated value if you get a positive spin at the first measurement.

Now, let's use my modified EPR apparatus, and consider what will happen under the transactional interpretation. At the time of particle decay, the decaying particle receives waves telling it what the eventual measurements will be. It then uses these waves to decide what each measurement should give. But as I mentioned above, once you have chosen an arbitrary pair of spin measurements, the result is unbiased in that it on average should produce as many positive as negative spin results. Incorrect; the results are biased as noted above.

So that's what should happen here under the transactional interpretation. But as I also explained, what we actually expect to happen is that there should be a bias toward positive results. And after enough repetitions of the experiment, it should be obvious that this is what happens.

So where am I going wrong? I find it hard to believe that I'm the first person to notice this. Have I just got the transactional interpretation confused, or is there some way to wriggle out of this problem?Where you went wrong is noted above.

[sodium]

Because of the correlation, you will always measure a correlated value if you get a negative spin at the first measurement; you may or may not measure a correlated value if you get a positive spin at the first measurement.
Is an experimental apparatus going to measure on average more negative than positive spins, more positive than negative spins, or about equal? What I mean by "bias" is bias toward getting more negative spins, or bias toward getting more negative spins.

I think you would agree that the apparatus I suggest is biased toward positive spins. I think you would also agree that for any given pair of measurements, there is no bias (in the sense that I'm using the term). I'm not using bias to mean the same thing as correlation.

In order to get the correct result for my apparatus, it isn't enough to know what the measurements are/will be. You can't get the correct result based solely on the set up of the source and the eventual measurements, because based only on that you wouldn't think the experiment had a bias toward reading positive spins. You need something else.

The TI describes information going back and forth between the source and measurements. But that doesn't indicate the relationship between the measurements. How does the TI get this information?

[sodium]

Einstein's local reality point of view is no longer sustainable, either locality or reality has to go!
Right. That's really what I've been arguing. We can have either locality or reality, and I choose locality.

Of course, I don't think denying reality, in this context, is as bad as it sounds. Denying reality in this sense doesn't mean embracing epiphenomenalism or solipsism or anything like that. It just means that certain questions that we think intuitively should have an answer, like, "What is the particles spin before it is measured?" and "Where precisely is that photon between the emitter and the wall?" and even "How is it like that?" don't have any answer.

It isn't the whole of reality that has to go, but at least some things that we thing ought to be real are in fact not real.

[Schneibster]

Is an experimental apparatus going to measure on average more negative than positive spins, more positive than negative spins, or about equal? What I mean by "bias" is bias toward getting more negative spins, or bias toward getting more negative spins.I didn't agree with your initial premise that it would get an equal number of positive and negative results if the measurements were random; in the face of that, it doesn't matter what I or you believe about what happens if the plane of the second measurement were controlled by the outcome of the first measurement. The degree of correlation depends upon the precise angles at which the random measurements are made; it varies for different angles. As it happens, the maximum variation is at 22.5 degrees or its mirror, 67.5 degrees.

I think you would agree that the apparatus I suggest is biased toward positive spins. I think you would also agree that for any given pair of measurements, there is no bias (in the sense that I'm using the term). I'm not using bias to mean the same thing as correlation.I don't know whether the apparatus you suggest is biased toward positive spins; it depends on the exact angles you choose. I do not agree that for any given pair of measurements, there is no bias because some measurements are correlated to some degree, which biases them.

[Schneibster]

Right. That's really what I've been arguing. We can have either locality or reality, and I choose locality.

Of course, I don't think denying reality, in this context, is as bad as it sounds. Denying reality in this sense doesn't mean embracing epiphenomenalism or solipsism or anything like that. It just means that certain questions that we think intuitively should have an answer, like, "What is the particles spin before it is measured?" and "Where precisely is that photon between the emitter and the wall?" and even "How is it like that?" don't have any answer.

It isn't the whole of reality that has to go, but at least some things that we thing ought to be real are in fact not real.You've gotten sideways on this issue. For Aspect to be local, the particle must have a spin before it is measured. If it does not, then the interaction is non-local, unless you think the backward wave in TI is "local."

[sodium]

You've gotten sideways on this issue. For Aspect to be local, the particle must have a spin before it is measured. If it does not, then the interaction is non-local, unless you think the backward wave in TI is "local."
I believe I mentioned a definition of causation for which TI is nonlocal, although Aspect is potentially local. Of course, it is possible to define locality so that it is not something that exists. I think that such a definition will be based on the kind of naive realism that we should reject.

[sodium]

I didn't agree with your initial premise that it would get an equal number of positive and negative results if the measurements were random;
Can you give a counter-example?

[Peter Soderqvist]

Right. That's really what I've been arguing. We can have either locality or reality, and I choose locality.

Of course, I don't think denying reality, in this context, is as bad as it sounds. Denying reality in this sense doesn't mean embracing epiphenomenalism or solipsism or anything like that. It just means that certain questions that we think intuitively should have an answer, like, "What is the particles spin before it is measured?" and "Where precisely is that photon between the emitter and the wall?" and even "How is it like that?" don't have any answer.

It isn't the whole of reality that has to go, but at least some things that we thing ought to be real are in fact not real.

I pick loca