Is there a fundamental flaw in our understanding of space and conservation laws?

[ueit]

In as much as this *sounds* like a scientific question, it isn't actually one. The scientific question is always "will I observe that...".

OK, I'll reformulate:

"will I observe that the electron will go up or down in a Stern Gerlach device?"

MWI cannot answer that. (this is true about QM in general, but MWI, by accepting QM as complete, cannot be expected to give an answer, never).

Why are you the person you are, and aren't you George Bush ?

1. my parents' DNA
2. my father's name is not Bush
3. my parents did not put me the name George.

Why do you think this question is relevant?

This is in fact not MWI, but rather the Landscape in string theory.

Yeah, but the method is similar. It proves the theory logically consistent but it gives useless answers.

The only thing "useful" that comes out of it, is to have an ontology which fits perfectly with the formalism of quantum theory as we know it, without any desire to fiddle with it. It is a kind of tranquilizer which helps you come to peace with the quantum formalism - and to help you develop some intuition for it. That's what "interpretations" are for. Peace of mind.

The problem is that QM's formalism fails to explain some experimental results (like the exact moment of decay of an unstable particle). Therefore the theory must be amended or changed (by adding hidden variables, like in BM for example). Trying to save the theory by in spite of unexplained experimental data is not, IMHO, a good way to do science. We should not "come to peace" with the present situation but try to improve it.

We are absolutely NOT forced to accept MWI because of EPR-Bell! We are not forced AT ALL to accept any interpretation (and certainly not MWI). Only, EPR-Bell is there in the first place because of the quantum formalism. So why look for *another* explanation if we have a machinery already on paper which has made us put up with the situation in the first place!

Because the "machinery" we have does not give exact answers (like the spin of the particle detected at a specific detector. The obvious answer is that we have a statistical theory that should be amended.

The reason I mention this is different. If all this discussion were not there, and if, from the start, one was to have an MWI-kind of view on quantum theory, then one could easily see an EPR-experiment as a kind of "confirmation" of quantum interference with macroscopic bodies: you entangle two macroscopic systems with different spin decompositions of a pair of entangled particles, and you do this at spacelike separation in order to be sure that they don't decohere immediately with one another. Then you let them "interfere" (calculate the correlations), and you show that this is statistically not possible without quantum effects.

You need quantum effects only if you assume statistical independence between the choice of detection angle and the properties of the entangled particles. EPR proves this assumption wrong.

If, in Newtonian mechanics, you calculate the force on a planet, and find, from those calculations, that the orbit is then going to be an ellipse, then why would you want to look for *another* mechanism which could produce elliptic orbits ? The very formalism that gave you the orbit (the Newtonian force of gravity) can also serve as the explanation or the mechanism. So you imagine with your mind's eye that some "force" is pulling on a planet, and that as such, it follows the orbit it should. If you do Newtonian mechanics, it wouldn't come to your mind to think of that as a kind of "bending of spacetime" or "crystal ellipsoids on which the planets roll" or whatever: if you do Newtonian mechanics, you take "as real" the elements of the theory. You don't lie awake at night of what "mechanism" might be responsible for elliptical orbits which come out of the mathematical formalism.
Well, to me, MWI is just a similar kind of reasoning held wrt the unitary quantum formalism. That's all. It's just a picture to keep in mind, when working with a formalism. A kind of "image of a reality" that will do the trick. But the nice thing is that because it sticks closely to the formalism, that it helps you reason intuitively in quantum theory.

When Newtonian mechanics failed to explain observations it was replaced. Not so with QM. Here, it is assumed that the random character of the measurement results must be fundamental, therefore a question that doesn't comply with this assumption is deemed irrelevant. I maintain that such a view is unscientific. MWI goes along with it, saying that there is no reason whatsoever for the specific result you get in an experiment because there are other, "superposed", experimenters getting the other possible results. Once you assume MWI as true, it makes sense, but I think such an assumption goes against scientific progress.

MWI will learn you nothing more than what is already in the quantum formalism. To me, it is the kind of "minimalistic" ontological picture that one can have. It simply gives you the "ease of mind" not to have the desire to change the formalism. In exactly the same way as the usual view of Newtonian mechanics doesn't give you any desire to change it.

I agree with this, but I don't see any merit in this lack of desire to change QM.

 

[rewebster]

It's interesting to see a discussion between a person concerned in applied physics and another in theoretical physics.

 

[vanesch]

OK, I'll reformulate:

"will I observe that the electron will go up or down in a Stern Gerlach device?"

MWI cannot answer that. (this is true about QM in general, but MWI, by accepting QM as complete, cannot be expected to give an answer, never).

I think you fail to see the gist of what is "an interpretation" of a theory. It is not at all an extension, or a suggestion for *another* theory, but it is just that: an interpretation (a way of looking at/understanding/having some intuition for) a theory. Several "interpretations" of QM start with saying that QM cannot be completely correct and modify it. But that's, strictly speaking, not trying to give an interpretation of QM! That is: not accepting QM and trying to fiddle with it to make it suit your wishes.

As to the random character of QM, I don't see in which way that needs to be a problem. After all, there is no fundamental theory - and you don't seem to be bothered by it either - which tells you how things on the left side of the road should look if you know how things on the right side look. In the same way, there is no a priori reason why we should be able to have a theory which tells us how things "tomorrow" look when we know how they look "today". It is simply a property of nature that there is such a relationship, but the nature of that relationship doesn't, in any way, need to satisfy properties we made up for ourselves. In classical physics, there is indeed a form of determinism: if we know how things are today, then we know (in principle) how they are tomorrow. But that is a specific property of classical physics. And classical physics itself ALSO has "unknowable" things ab initio: you still DO need to say how things are today to know them tomorrow. In other words, classical physics doesn't tell you what initial conditions you should choose. So not everything is "known" in classical physics either.

Quantum theory gives us a relationship between observations, and that relationship is stochastic. From that fact alone, we cannot conclude that it is "incomplete": nature might be such that there is no way to know for sure. There doesn't *need* to be any unique and deterministic relationship.

Yeah, but the method is similar. It proves the theory logically consistent but it gives useless answers.

The answers are not useless. QM does give a lot of information, in the form of statistical distributions, which are observationally verified.

The problem is that QM's formalism fails to explain some experimental results (like the exact moment of decay of an unstable particle). Therefore the theory must be amended or changed (by adding hidden variables, like in BM for example).

There is no fundamental need that the information of such a moment is "present" in nature before the thing actually happens. Maybe, maybe not. Observational determinism doesn't need to be a principle of nature.

Trying to save the theory by in spite of unexplained experimental data is not, IMHO, a good way to do science. We should not "come to peace" with the present situation but try to improve it.

Well, if you CAN improve upon it, than you should do so. However, nor Bohmian mechanics nor anything else has ever come up with an *observational improvement* of quantum theory. So saying that the theory is internally deterministic, but fundamentally observationally random, is scientifically speaking, not any different from saying that the theory is random.

Because the "machinery" we have does not give exact answers (like the spin of the particle detected at a specific detector. The obvious answer is that we have a statistical theory that should be amended.

The obvious fact is that we have a working observationally stochastic theory. Maybe nature is such, or maybe it isn't. But it is not up to us to decide that, in an armchair. Imagine that nature has a truly stochastic element to it. You can look forever for your deterministic improvement, you will be looking in vain.

You need quantum effects only if you assume statistical independence between the choice of detection angle and the properties of the entangled particles. EPR proves this assumption wrong.

EPR doesn't prove this assumption wrong, because it is only one amongst many.
It is true that there is a conceptually very simple theory of nature which is deterministic and which doesn't give many interpretational problems: nature is a pre-filled bag of events. There is somewhere a huge list of events, which fill spacetime. There is no causality, no determinism, nothing of the kind, just a big list of everything that happened and will happen in the universe. There are no laws of nature apart from this list, which includes everything.
Now, that list can have funny properties, and one can start to think that there are causal relationships, because of the funny correlations that appear in that list, but all that is just by coincidence.
With such a view, every theory and its contrary are possible.

When Newtonian mechanics failed to explain observations it was replaced. Not so with QM.

No, it is not because Newtonian mechanics failed to predict the correct initial conditions (you need to put them in by hand!) that it was replaced. It was because it made predictions that were observed not to be so.
QM doesn't make predictions that are observed not to be true.

Here, it is assumed that the random character of the measurement results must be fundamental, therefore a question that doesn't comply with this assumption is deemed irrelevant. I maintain that such a view is unscientific.

I could say: "it is assumed that the need to give initial conditions must be fundamental, therefor a question that doesn't comply with this assumption is deemed irrelevant..."

MWI goes along with it, saying that there is no reason whatsoever for the specific result you get in an experiment because there are other, "superposed", experimenters getting the other possible results. Once you assume MWI as true, it makes sense, but I think such an assumption goes against scientific progress.

Again, an interpretation is not meant to go beyond the theory it interprets. It just wants to give a mental picture that "explains" the formalism. That's all. All the rest is not "interpretational" but rather "exploratory".

I agree with this, but I don't see any merit in this lack of desire to change QM.

The specific problem QM faces is that it seems difficult to set up a mental picture of it that doesn't require one to say that, deep down, something doesn't turn round as it should. Now, that's very nice, but there is not any satisfying replacement that keeps the power of the QM formalism (Bohmian mechanics comes close, but needs to kick out a very fundamental principle, namely, relativity, without anything to replace it with just as much predictive power). So MWI proposes a way to set up such a mental picture without that need. With MWI, you can develop an intuition for QM which goes with its formalism. When you USE the formalism, you can rely on an MWI view when you are a bit in doubt on exactly how to use it (as is sometimes the case in delayed quantum erasers and EPR situations). That's the only use of MWI!

It is not from interpretational issues that any progress will follow in any case. It is not from re-interpreting Newtonian mechanics that people came to relativity or anything. A new theory was build upon a new principle, and THEN people tried to find out how the link with Newtonian mechanics could be re-established. Personally, I don't think that much progress beyond QM will come from considerations in Bohmian mechanics. It is already conceptually difficult to get some correspondence with QFT, which is a working horse.

 

[rewebster]

For me, this discussion is what is unusual about laws, theories, interpretations, etc. in theoretical physics.

Debate is one of the most useful tools that I’ve found to expand an idea, or initiate a creative thought; whether it is with another person, or, if you want/can call it a ‘debate’, if the person is able, to examine and criticize one’s own thought patterns.

I personally use set theory, bell curve grading (how and where the thought fits into each idea and subset), logic (in the way of deductive and inductive reason, in using judgement and reason), and common sense (very personal as it relates differently to each person). And, yes, all of these can be personal and introspective when it comes to anyone’s (and my) philosophy; and, of course, each and every statement made (by anyone, including me, here) is a reflection of one’s personal belief systems as it relates to each and every area of a thought taken (transcribed) to words, and the function of time to present those thoughts (reflection time to present the thoughts in the most ‘logical’ and correct way).

Laws, theories, interpretations, etc. can be accepted seemingly TRUE, FALSE, or PARTIALLY TRUE (and, therefore, PARTIALLY FALSE); and ‘accepted’ could mean generally, personally, academically, scientifically, etc.;

---and, again, the ‘ideas’ of the interpretation of the ‘words’ and the ‘ideas' behind the words, ‘true’ and ‘false’ can be interpreted in different ways (the idea of interpretation of any of the ‘words’ used here will be a ‘given’ as such).

There could be another subset (one of many) under this of USEFUL, PARTIALLY USEFUL (and therefore, PARTIALLY USELESS) and USELESS. This could be viewed from the perspective of (some) applied physicists (technicians, researchers, engineers, some teachers, etc.).

(Nine possible subsets)

Theoretical physicists may have a subset of ACCEPTABLE, PARTIALLY ACCEPTABLE (and therefore, PARTIALLY UNACCEPTABLE), and UNACCEPTABLE.

(Nine possible subsets, or, if combined with both the above, twenty seven subsets)

“The sun rises in the east” could be considered by some to be TRUE and USEFUL and ACCEPTABLE. Others will automatically say it’s FALSE, and, maybe PARTIALLY USEFUL and UNACCEPTABLE.

If you (and you can) add ‘unknown’ or ‘not known’, ‘not yet known’ etc., it would add more subsets, but also indicate a level of doubt, lack of completeness, indecisiveness, etc.

The question I asked myself is ‘If a theory is only partially true to every circumstance (QM, relativity, MWI, etc.), as it is (they are); and, it is (they are) not (yet) THE total and complete “The Theory of Everything”, is it (are they), or can it be (could they be), therefore, totally false (a little lie is still a lie, for example)?'


I’m in the thinking that MWI is PARTIALLY TRUE, PARTIALLY USEFUL, and PARTIALLY ACCEPTABLE.

 

[ueit]

I think you fail to see the gist of what is "an interpretation" of a theory. It is not at all an extension, or a suggestion for *another* theory, but it is just that: an interpretation (a way of looking at/understanding/having some intuition for) a theory. Several "interpretations" of QM start with saying that QM cannot be completely correct and modify it. But that's, strictly speaking, not trying to give an interpretation of QM! That is: not accepting QM and trying to fiddle with it to make it suit your wishes.

Of course, it depends on how you define "interpretation". I don't think that it is necessary to assume QM as complete in order to give an interpretation. For example one can understand QM as a statistical theory, like thermodynamics.

As to the random character of QM, I don't see in which way that needs to be a problem. After all, there is no fundamental theory - and you don't seem to be bothered by it either - which tells you how things on the left side of the road should look if you know how things on the right side look. In the same way, there is no a priori reason why we should be able to have a theory which tells us how things "tomorrow" look when we know how they look "today". It is simply a property of nature that there is such a relationship, but the nature of that relationship doesn't, in any way, need to satisfy properties we made up for ourselves. In classical physics, there is indeed a form of determinism: if we know how things are today, then we know (in principle) how they are tomorrow. But that is a specific property of classical physics. And classical physics itself ALSO has "unknowable" things ab initio: you still DO need to say how things are today to know them tomorrow. In other words, classical physics doesn't tell you what initial conditions you should choose. So not everything is "known" in classical physics either.

To say that something is random is to say that it cannot have any explanation. A theory that proposes such a possition goes against the very idea of science. One cannot seriously accept "there is no answer" for an answer.

There can be a connection between the two sides of the road, but it is so complex that nobody can "reverse engineer" the causality chain.

Determinism is the most powerful idea behind science. Randomness is a dead end. Maybe it is true, but it is as usefull as the "brain in the vats" hypothesis.

Quantum theory gives us a relationship between observations, and that relationship is stochastic. From that fact alone, we cannot conclude that it is "incomplete": nature might be such that there is no way to know for sure. There doesn't *need* to be any unique and deterministic relationship.

I disagree. The fact that we cannot explain an experimental result is a strong indication that the theory is incomplete. Otherwise, no theory will ever be suspected as being incomplete. Just answer "it's random" to any question.

The answers are not useless. QM does give a lot of information, in the form of statistical distributions, which are observationally verified.

I'm not saying that QM is useless. Useless is an interpretation of it which takes "there is no answer" as a good answer.

There is no fundamental need that the information of such a moment is "present" in nature before the thing actually happens. Maybe, maybe not. Observational determinism doesn't need to be a principle of nature.

Sure, but one must assume it in order to do science. Otherwise, can you propose a scheme by which one can distinguish between pure random phenomena and deterministic but yet unexplained phenomena worthy of research?

Well, if you CAN improve upon it, than you should do so. However, nor Bohmian mechanics nor anything else has ever come up with an *observational improvement* of quantum theory. So saying that the theory is internally deterministic, but fundamentally observationally random, is scientifically speaking, not any different from saying that the theory is random.

There is a difference. We can be 100% sure that a theory based on fundamental randomness is a dead end. If something is random it cannot be explained. Other theories may (or may not) evolve.

The obvious fact is that we have a working observationally stochastic theory. Maybe nature is such, or maybe it isn't. But it is not up to us to decide that, in an armchair. Imagine that nature has a truly stochastic element to it. You can look forever for your deterministic improvement, you will be looking in vain.

This is true, but what else can we do? There is also the possibility that a proof against deterministic theories is found.

It is true that there is a conceptually very simple theory of nature which is deterministic and which doesn't give many interpretational problems: nature is a pre-filled bag of events. There is somewhere a huge list of events, which fill spacetime. There is no causality, no determinism, nothing of the kind, just a big list of everything that happened and will happen in the universe. There are no laws of nature apart from this list, which includes everything.
Now, that list can have funny properties, and one can start to think that there are causal relationships, because of the funny correlations that appear in that list, but all that is just by coincidence.
With such a view, every theory and its contrary are possible.

I do not hold such a view and I find it completely useless. My deterministic hypothesis concerning EPR is based on the assumption that the entangled particles can "read", locally, from the surrounding fields how they will be detected/measured in the future.

It should be possible to rewrite Bohm's theory in a local manner by describing each event as a function of an arbitrary state in the distant past. I wonder if such an attempt has ever been made.

QM doesn't make predictions that are observed not to be true.

That's why I think it is a good statistical theory.

I could say: "it is assumed that the need to give initial conditions must be fundamental, therefor a question that doesn't comply with this assumption is deemed irrelevant..."

But Newtonian mechanics can give you the initial conditions, granted that you measure the present conditions with an arbitrary high accuracy.

Again, an interpretation is not meant to go beyond the theory it interprets. It just wants to give a mental picture that "explains" the formalism. That's all. All the rest is not "interpretational" but rather "exploratory".

See my opinion above.

The specific problem QM faces is that it seems difficult to set up a mental picture of it that doesn't require one to say that, deep down, something doesn't turn round as it should. Now, that's very nice, but there is not any satisfying replacement that keeps the power of the QM formalism (Bohmian mechanics comes close, but needs to kick out a very fundamental principle, namely, relativity, without anything to replace it with just as much predictive power). So MWI proposes a way to set up such a mental picture without that need. With MWI, you can develop an intuition for QM which goes with its formalism. When you USE the formalism, you can rely on an MWI view when you are a bit in doubt on exactly how to use it (as is sometimes the case in delayed quantum erasers and EPR situations). That's the only use of MWI!

OK.

It is not from interpretational issues that any progress will follow in any case. It is not from re-interpreting Newtonian mechanics that people came to relativity or anything. A new theory was build upon a new principle, and THEN people tried to find out how the link with Newtonian mechanics could be re-established. Personally, I don't think that much progress beyond QM will come from considerations in Bohmian mechanics. It is already conceptually difficult to get some correspondence with QFT, which is a working horse.

I think QM is more like thermodynamics than Newtonian mechanics. Here, evidence for the atomic structure was found by expressing some laws of unknown origin as a statistical approximation of many particles' motion(the P-V-T relationship of gases for example).

 

[vanesch]

Of course, it depends on how you define "interpretation". I don't think that it is necessary to assume QM as complete in order to give an interpretation. For example one can understand QM as a statistical theory, like thermodynamics.

Well, that's not so easy! If you do so, you need to introduce a sample space over which a statistical uncertainty is defined ; in other words, you need to build a hidden-variable theory. The only known one is Bohmian mechanics, and you run in quite some troubles with relativity there. Moreover, in doing so, you've been fiddling with the formalism, just for sake of a principle which doesn't bring in anything new and predictive, but rather destroys a predictive principle (namely, relativity).

See, saying that a theory "is a statistical theory, like thermodynamics" already creates a very restrictive frame. MWI, on the other hand, says that quantum theory is indeed, something like "thermodynamics", but then the sample space is the *observer* sample space. And then, you do not need to introduce any new elements in the formalism, or to break the principle of relativity.
In other words, what's particular to QM (through MWI glasses), is to say that there is "statistical mechanics" on the observer side, and not on the system side.

To say that something is random is to say that it cannot have any explanation. A theory that proposes such a possition goes against the very idea of science. One cannot seriously accept "there is no answer" for an answer.

Again, random doesn't mean "just anything". Random is that strange concept in between "determined" and "totally unknown": namely: with statistical regularity, but individually unknown.

Determinism is the most powerful idea behind science. Randomness is a dead end.

Determinism is a principle like any other. It can hold, or it can't.

Otherwise, can you propose a scheme by which one can distinguish between pure random phenomena and deterministic but yet unexplained phenomena worthy of research?

Well, if a theory can give you the verifiable statistical distributions of phenomena, then that's already something quite worth while. Nothing should stop you to try to find anything that gives more information, but there is no guarantee that you can find such a thing. As I said, Bohmians didn't come up with anything that gives "finer" predictions than the statistical predictions of QM.

There is a difference. We can be 100% sure that a theory based on fundamental randomness is a dead end. If something is random it cannot be explained. Other theories may (or may not) evolve.

I'm pretty sure that, if ever there is a theory which does finer predictions than QM, it will be of a totally different nature. So the interpretations suggested by QM will not help in uncovering such a theory (in the same way as different interpretations of Newtonian mechanics didn't lead to relativity). Again, the interpretation of QM has not as a goal to go beyond it, just to give an mental picture of it.

This is true, but what else can we do? There is also the possibility that a proof against deterministic theories is found.

Cannot be, given that there *IS* a deterministic theory (the big list of events) that can describe all what happens. Only, we will need first to observe all events in the universe before we can write it down (and even then, the sheet of paper on which it is to be written doesn't fit in the visible universe).

I do not hold such a view and I find it completely useless. My deterministic hypothesis concerning EPR is based on the assumption that the entangled particles can "read", locally, from the surrounding fields how they will be detected/measured in the future.

And how useful is such a view when you do QM computations ? Don't confuse a wet dream of a would-be theory with something that gives you an actual picture that goes with actual calculations.

It should be possible to rewrite Bohm's theory in a local manner by describing each event as a function of an arbitrary state in the distant past. I wonder if such an attempt has ever been made.

See, that's a wet dream of a would-be theory (which has in fact not much to do with the quantum formalism).

But Newtonian mechanics can give you the initial conditions, granted that you measure the present conditions with an arbitrary high accuracy.

No, you didn't understand my objection. I said: one could also have the principle that a good theory doesn't need ANY INPUT, and can make predictions WITHOUT input (such as initial or final or boundary conditions).

You require a theory to be deterministic: meaning: if events on a spacelike surface are given, then you should determine uniquely what are the events on another spacelike surface. You do not accept that only statistical correlations of those events can be found. I give you an extreme alternative: NO correlations have to exist really and all those that we think we've found are spurious (the "big list" theory) ; on the other hand the "a single theory that gives you all events WITHOUT having to specify any initial conditions"

Is the "big list" theory (which doesn't need any input at all!):
- deterministic or
- totally random ?

From a PoV of principle, the "big list" theory is superdeterministic: you do not even need to specify initial conditions, it contains all events AS SUCH (and no statistical distributions or anything: you just look in the list what are the events in the year 3050, and you also read in the list how the dinosaurs became extinct).
But from a practical PoV, it is totally random, as there aren't any a priori regularities in the list of events. It cannot even predict statistical regularities. Given that we don't have the list, this theory is practically way worse than QM.

A deterministic theory is an intermediate: GIVEN a slice of events in the list, we can calculate the whole list. So, you've shifted the whole list into a single slice + a reasonably short (?) algorithm. A theory such as QM is also an intermediate: given a slice of events, we can calculate statistical distributions of events of another slice. Here, the algorithm doesn't allow you to reconstruct the list, but just to find correlations.
In the "superdeterministic" list, no algorithm needs to exist at all, that gives you either a deterministic OR a stochastic link between two slices. So in as much you consider the list itself as a "trivial but huge" algorithm, you have a superdeterministic theory, and in as much you consider the list as "necessary initial data" (and an empty algorithm), you have a totally random "theory" without any predictive value.

 

[ueit]

Vanesch, our disagreement comes from the principles we consider important in doing science.
We have a set of experimental observations that cannot be explain/predicted by our current quantum theory (like the moment of a radioactive decay).

There are different things one can assume when confronting them:

1. Nothing. Reality is a “bag of events”
2. Nothing. They are fundamentally random (from our subjective experience), cannot be explained by any mathematical algorithm.
3. Try to find a mechanism by which they are “produced”
4. ….
5. ….

It is my opinion that “3” is the good way of doing science. In spite of its problems, BM provides a framework to explore this possibility. MWI goes with “2” and nothing interesting can be expected from it, just as in the case of “1”.
I repeat that MWI may very well be true, just like the “bag of events” or “universe as computer simulation” hypothesis. But, starting from such an assumption is not of any use for scientific progress. The theory is a dead end.

Now, because you assume “2”, you see nothing wrong/incomplete about QM formalism and see any deviation from it as harmful. On the other side, “3” requires the formalism to be completed. If by doing so, we have to change other principles like relativity, so be it, as long as there is no conflict with the experimental evidence (one can simulate relativistic effects on an absolute frame).

I'm pretty sure that, if ever there is a theory which does finer predictions than QM, it will be of a totally different nature.

Why?

So the interpretations suggested by QM will not help in uncovering such a theory (in the same way as different interpretations of Newtonian mechanics didn't lead to relativity). Again, the interpretation of QM has not as a goal to go beyond it, just to give an mental picture of it.

Again, I think Newtonian mechanics is a bad analogy because it is not a statistical theory. Thermodynamics is a better example. Because a new theory must reduce to QM, trying to reproduce QM’s formalism from a hidden variable approach might be useful.

Cannot be, given that there *IS* a deterministic theory (the big list of events) that can describe all what happens. Only, we will need first to observe all events in the universe before we can write it down (and even then, the sheet of paper on which it is to be written doesn't fit in the visible universe).

OK, but at least a proof against a certain class of deterministic theories, of the Newtonian type, could be found. I think “the big list of events” is not deterministic in the usual understanding of the term as the present does not determine the future. It can be rejected on Ockham’s razor alone.

And how useful is such a view when you do QM computations ? Don't confuse a wet dream of a would-be theory with something that gives you an actual picture that goes with actual calculations.

It is the obvious conclusion when confronting EPR from a deterministic point of view. I cannot myself develop a theory as I lack the necessary math skills and knowledge. I only wonder if such a view has been disproved. It strikes me as odd that nobody seems to think about it.

Time to go home, I’ll continue later.

 

[vanesch]

Vanesch, our disagreement comes from the principles we consider important in doing science.
We have a set of experimental observations that cannot be explain/predicted by our current quantum theory (like the moment of a radioactive decay).

Although the exact moment of a decay cannot be explained by QM, its statistical properties can. I agree with you that it could a priori be conceivable that there is a theory that can say so, however, in 80 years of trying, no such evidence has ever been found (in other words, no empirical relation has ever been put forward that gave a more precise relationship than the statistical relationship given by QM). As such, for all we know - today, and this can change tomorrow - we have no empirical indication at all that there is such a thing as determinism.

The "principles that are important in doing science" are basically that observations are what matters. Well, no observational correlations have ever been discovered that are more precise than QM predictions. This was, remember, different with Newtonian mechanics. Before the advent of Newtonian mechanics, there were Kepler's laws, for the motion of planets. They were empirically based laws, which could give a precise correlation between positions "yesterday" and positions "tomorrow", even though the theoretical frame was not known. But for QM phenomena, no such empirical relations have ever been devised. Not that there haven't been attempts! The early literature is full of "attempts to violate the Heisenberg uncertainty relationships". They all fail.

There are different things one can assume when confronting them:

1. Nothing. Reality is a “bag of events”
2. Nothing. They are fundamentally random (from our subjective experience), cannot be explained by any mathematical algorithm.
3. Try to find a mechanism by which they are “produced”
4. ….
5. ….

Well, before even having any hope of formulating a general deterministic theory, one should have some observations that suggest it.

It is my opinion that “3” is the good way of doing science. In spite of its problems, BM provides a framework to explore this possibility. MWI goes with “2” and nothing interesting can be expected from it, just as in the case of “1”.
I repeat that MWI may very well be true, just like the “bag of events” or “universe as computer simulation” hypothesis. But, starting from such an assumption is not of any use for scientific progress. The theory is a dead end.

I agree. An interpretation of a theory is not a way to make progress, it is a way to get a feel for the theory at hand. This is why I don't bother, for instance, to try to incorporate quantum gravity in any MWI picture. There is simply no physical theory as of today which gives a good formal description of quantum gravity phenomena, so I don't think that "interpretations" of an existing theory should do what the theory itself cannot.

But let's be clear about something: EPR-Bell phenomena are perfectly well described by the formalism of QM. So the "secret" already resides in the QM formalism. One mustn't "go beyond" QM to "discover" them, they are already included into QM. So it would be very strange that an *interpretation* of this perfectly well predicted phenomenon by QM would need elements that are NOT included in said formalism. Nevertheless, these (well-predicted) QM phenomena seem to be ununderstandable for some. MWI is a view which helps one understand how the CURRENT QM formalism works, and comes to these predictions.

Now, because you assume “2”, you see nothing wrong/incomplete about QM formalism and see any deviation from it as harmful.

No, not at all. I consider the exercise of *interpreting* an existing theory as a totally different exercise from the exercise to go beyond the given theory.

There is a difference between "inventing new, speculative theories" or rather having wet dreams about would-be new theories on one hand, and trying to understand what we already have in our hands. I consider MWI in the last category. It is not a tool to go beyond QM. It is a tool to understand contemporary QM, including its strange predictions such as EPR-Bell and quantum erasers and so on - of which I repeat that they are perfectly well described by the CURRENT formalism (so all the knowledge is already included in this formalism!).

On the other side, “3” requires the formalism to be completed. If by doing so, we have to change other principles like relativity, so be it, as long as there is no conflict with the experimental evidence (one can simulate relativistic effects on an absolute frame).

Yes, but in order to save a "desired" principle which, as of today, doesn't bring in any new formal predictive power, we've then sloshed down the drain another principle which DOES have quite some formal predictive power. That's where I think that we lose some.

Why?

Historically, things always happened that way. The view on the old theory was never a good starting point to find a new theory. In fact, the view on the old theory has always *hindered* the correct conceptualisation of the new one. This was so with electromagnetism (mechanical "ether"), with relativity (Lorentz ether), with quantum theory (Bohmian mechanics), ...
Sticking to the old view on things has never been very productive or suggestive in the new paradigm. That's why I think that if it happens again, all our considerations concerning QM are to be dropped also.

Again, I think Newtonian mechanics is a bad analogy because it is not a statistical theory. Thermodynamics is a better example. Because a new theory must reduce to QM, trying to reproduce QM’s formalism from a hidden variable approach might be useful.

It might. Then, it might not.

OK, but at least a proof against a certain class of deterministic theories, of the Newtonian type, could be found. I think “the big list of events” is not deterministic in the usual understanding of the term as the present does not determine the future. It can be rejected on Ockham’s razor alone.

What I wanted to illustrate is that the separation into "algorithm" (or "laws of nature") on one hand, and "external data" (boundary conditions) is, up to a certain point, arbitrary. A very complex algorithm is difficult to distinguish from external data. How nature is put together is not to be decided in an armchair, but by observation and deduction.
"deterministic" theories have a rather simple algorithm, and need as external data the "initial conditions" on a spacelike surface. They are just ONE possible configuration for nature, but that doesn't need to be so.
There could be a single, universal algorithm with NO external data needed (no "initial conditions" needed). Or there could just be statistical relations. Or there could be nothing at all in the form of an algorithm (big bag of events) - this is already probably not the case, given that we already FOUND laws of nature of a rather simple kind. But it is not up to us to decide in advance what should be the kind of structure that nature has. We can only make observations, and try to find relationships between them (deterministic, statistical, whatever). That's science.

For instance, superdeterminism is also a "bad" hypothesis to do science. It would mean that we cannot discover any laws of nature, because all of our observations are biased in unknown but strong ways. All the "correlations" we find between different events are actually DIFFERENT correlations due to superdeterminism. It would be the end of observational science, given that every observation is going to be the result of a big conspiracy.

 

[ueit]

It should be possible to rewrite Bohm's theory in a local manner by describing each event as a function of an arbitrary state in the distant past. I wonder if such an attempt has ever been made.

See, that's a wet dream of a would-be theory (which has in fact not much to do with the quantum formalism).

I disagree. I think it is easy to show that a deterministic theory, like BM, can be reformulated in a local manner. For example, one could predict each spin measurement in an EPR experiment, let's say, one hour before the said experiment is actually performed, only from the initial conditions. Any trick, like using a "delayed choice" device is irrelevant, because the "choice" already exists, hidden in those initial parameters. So, we can interpret the so-called non-local interactions in terms of a local mechanism. An example is the local description of gravity in GR versus the non-local one in Newtonian theory.
So, I think BM is only an approximation to the "true", local theory.

Although the exact moment of a decay cannot be explained by QM, its statistical properties can. I agree with you that it could a priori be conceivable that there is a theory that can say so, however, in 80 years of trying, no such evidence has ever been found (in other words, no empirical relation has ever been put forward that gave a more precise relationship than the statistical relationship given by QM). As such, for all we know - today, and this can change tomorrow - we have no empirical indication at all that there is such a thing as determinism.

The "principles that are important in doing science" are basically that observations are what matters. Well, no observational correlations have ever been discovered that are more precise than QM predictions. This was, remember, different with Newtonian mechanics. Before the advent of Newtonian mechanics, there were Kepler's laws, for the motion of planets. They were empirically based laws, which could give a precise correlation between positions "yesterday" and positions "tomorrow", even though the theoretical frame was not known. But for QM phenomena, no such empirical relations have ever been devised. Not that there haven't been attempts! The early literature is full of "attempts to violate the Heisenberg uncertainty relationships". They all fail.

First, I'm not so sure that evidence for patterns in the data was seriously looked for (I've seen complains about the lack of raw, "uncleaned" data from various experiments). Second, we simply don't have the same data about radioactive decay as we did for planetary motion. For example, we don't know the charge or mass distribution inside of an atomic nucleus that is ready to "blow up". We don't know the detailed structure of the electron cloud in the place where a tunneling has happened. So, it's not unexpected that a pattern was not observed.

AFAIK Heisenberg uncertainty applies only to predictions, not to past observations. For example we can determine with unlimited accuracy both the momentum and position of a particle simply by detecting it arbitrary far away from the source. We cannot use this result to predict its future behavior, but this has nothing to do to the random/determined character of its motion.

Well, before even having any hope of formulating a general deterministic theory, one should have some observations that suggest it.

We have two types of experiments:

1. we start with 100% identical systems (like two atoms produced from a diatomic molecule) and we get identical measurements. (this points to determinism)

2. we start with unknown initial parameters end get random results.

Do you think that "2" points to randomness at fundamental level? If so, why?

There is a difference between "inventing new, speculative theories" or rather having wet dreams about would-be new theories on one hand, and trying to understand what we already have in our hands. I consider MWI in the last category. It is not a tool to go beyond QM. It is a tool to understand contemporary QM, including its strange predictions such as EPR-Bell and quantum erasers and so on - of which I repeat that they are perfectly well described by the CURRENT formalism (so all the knowledge is already included in this formalism!).

OK.

For instance, superdeterminism is also a "bad" hypothesis to do science. It would mean that we cannot discover any laws of nature, because all of our observations are biased in unknown but strong ways. All the "correlations" we find between different events are actually DIFFERENT correlations due to superdeterminism. It would be the end of observational science, given that every observation is going to be the result of a big conspiracy.

I think we have a different understanding of superdeterminism. IMO, any deterministic theory (classical physics, MWI, BM) is superdeterministic, that is, no free-will. In fact I see no trace of evidence, from QM or any other branch of science, that would give a foundation for free-will. You need a sort of mind/brain dualism for that.
I fail to see why a determined "choice" implies a "conspiracy", or biased observations. There is no reason to assume that the experimental data is not representative. We have to accept however that there might be experiments that cannot be done. For example, you cannot "fool" a particle in a EPR experiment. But you cannot "fool" gravity or conservation laws either.

 

[vanesch]

I disagree. I think it is easy to show that a deterministic theory, like BM, can be reformulated in a local manner.

No. It is easy to show that an unknown theory, for which there is a common cause of the settings of the analysers (in other words, in which there is a common cause to the settings of the analysers of Alice, Bob, and of the pair production) doesn't have a problem with Bell, because one of the assumptions of Bell's theorem is not satisfied.

But the theory is still unknown! Can you figure out a theory in which all the nitty-gritty details are explained on how it came about that Alice decided to set exactly THAT analyser setting which she had to set ?
I mean, it is not sufficient to say that, because in a deterministic theory, there is indeed no free will and whatever Alice will set WAS predetermined, that it is also THE CORRECT VALUE of the setting! You have to show that the *correct* settings at Alice and Bob have to come out! Did you ever hear of any such derivation which shows that the complicated (deterministic) process in the brain of Alice and so on, cannot but result in exactly that correct setting which will lead to the (illusion of) EPR correlations ?

So I stick to my claim of a wet dream of a non-existing would-be theory.

For example, one could predict each spin measurement in an EPR experiment, let's say, one hour before the said experiment is actually performed, only from the initial conditions. Any trick, like using a "delayed choice" device is irrelevant, because the "choice" already exists, hidden in those initial parameters. So, we can interpret the so-called non-local interactions in terms of a local mechanism. An example is the local description of gravity in GR versus the non-local one in Newtonian theory.
So, I think BM is only an approximation to the "true", local theory.

Yes, but there is a difference between saying that such a thing could in principle be possible, and propose an actual theory in which one can demonstrate that *THIS* is what will happen ! Do you realize the complexity of such a feat ? Irrespective of the specific physical setup, all small interactions will ALWAYS, in many different circumstances, conspire such that exactly that "choice" is made which makes us think that there are correlations!

 

[ueit]

But the theory is still unknown! Can you figure out a theory in which all the nitty-gritty details are explained on how it came about that Alice decided to set exactly THAT analyser setting which she had to set ?
I mean, it is not sufficient to say that, because in a deterministic theory, there is indeed no free will and whatever Alice will set WAS predetermined, that it is also THE CORRECT VALUE of the setting! You have to show that the *correct* settings at Alice and Bob have to come out! Did you ever hear of any such derivation which shows that the complicated (deterministic) process in the brain of Alice and so on, cannot but result in exactly that correct setting which will lead to the (illusion of) EPR correlations ?

First, you speak about the "illusion of EPR correlations". I don't think this is apropriate. The correlations are real and "enforced" by the deterministic law, like gravity. They are inevitable.

Yes, but there is a difference between saying that such a thing could in principle be possible, and propose an actual theory in which one can demonstrate that *THIS* is what will happen ! Do you realize the complexity of such a feat ? Irrespective of the specific physical setup, all small interactions will ALWAYS, in many different circumstances, conspire such that exactly that "choice" is made which makes us think that there are correlations!

I think there is another way to look at this problem, by focusing not on Alice's or Bob's brain but on the entangled particles.

Let's say that the total force acting on the particle source is represented by the function f(t) = d1(t) + d2(t), where d1 and d2 correspond to the detectors' state at time t (by "state" I mean any field comming from them, gravitational, EM, weak, whatever). I assume that no other object interferes with the experiment.
Because d1, and d2 evolve deterministicaly, their state at any future time can be known at source's position because of f(t).
We can now postulate that the the entangled particles are "produced" only when the measurement orientation, at the time of detection (which can be determined from f as well), is in accordance with QM's prediction.

 

[ttn]

I just skimmed the recent posts on this thread and wanted to make one comment.

ueit, you seem to miss (or unfairly minimize) the distinction between a merely deterministic theory, and a super-deterministic theory. The relevant point for this distinction is the idea of "external fields". An external field is a field which has causal influences on the beables of a theory, but which is not itself fixed by those beables. Most familiar physical theories have external fields: for example, the charge and current distributions can be external fields in classical E&M (though they can also be included dynamically in the theory), simple QM usually allows external electromagnetic potentials to be set, QFT allows coupling to external sources, etc. Indeed, it is hard to think of an example of a real theory (as opposed to a dream, like string theory... perhaps vanesch would call this a "dry dream" since it hasn't yielded any pleasurable results?!) which doesn't contain such external fields. And the reason for this is clear: scientific theories must be testable if they are going to be believed, and "external fields" are the way that theories leave room for experimenters. Without external fields, it is extremely difficult to imagine how we could test a theory -- or more precisely, it is extremely difficult to understand how the theory would permit something which we could validly describe as a test of the theory. (It'd be like a randomized drug trial in which the assignment of pills to patients was made by the pills themselves -- even if, in the end, the placebo patients did worse than the drug patients, it would be impossible to infer that the drug actually helped those patients, i.e., it would be impossible to believe that we had actually *tested* the drug.)

The point is: there are many extant examples of deterministic theories. But all of these leave room for "free human experimenters" in their external fields. So it is very misleading when you suggest that the EPR-Bell apparent statistical evidence for non-locality, can be explained away if we merely posit instead a deterministic theory. What you actually need to posit is instead a *superdeterministic* theory, i.e., a theory in which the kinds of things we normally (in science) think of as being "freely or randomly settable", actually *aren't* "freely or randomly settable". So when Weihs et al set up quantum random number generators to determine along which axis a given photon's polarization is measured, you have to believe that that random number generator is actually not random at all (with respect to the state of the incoming photon and whatever else it happens to be entangled with). Rather, by some gigantic cosmic conspiracy, whatever factors are determining the output of the random number generator, are also determining the state of the photon pair (or, I guess, since you want to preserve locality, the factors which immediately determine the output of the random number generator are determined/influenced by some other thing in the distant past which *also* turns out to the relevant thing for influencing the state of the photon pair). And as if that we're insane enough, you have to hold that this magical cosmic conspiracy is clever enough to magically work out exactly the same way, even if the random number generator is switched out in favor of some completely different process, such as a computer running a "random number" algorithm, a young child flipping coins, a human making last-minute "random free-will choices" (with scare quotes because, in superdeterminism, there can't actually be genuine freewill), a chimpanzee throwing either even or odd numbers of poop projectiles, etc. The point is, in order to take your idea seriously, you need to be able to construct a theory in which some causal influence from the distant past influences *all* of these sorts of things, and in just exactly the right way so as to correlate the apparatus settings (which are determined by one of these silly methods) and the state of the photon pair, so as to later fool us into believing that (a) a genuine empirical test of Bell's inequality has occurred and (b) that the outcomes violated the inequality. Or at very least, you have to be able to give at least the hint of some kind of plausibility argument that maybe all of these things *could*, in principle, be so influenced as to bring about the results in question. Without this, all you have is a conspiracy theory. And so it is no better than if you said, about some randomized drug trial which showed strong correlation between being given the actual drug and feeling better later, "I don't believe it because maybe the allegedly-random assignment of pills to patients wasn't random at all, but was somehow magically biased by some causal factors in the distant past which caused the assignment to somehow or other be correlated to the prior health of the patient." Yeah, "maybe." But in the absense of a single shred of evidence to suggest this, it is scientifically meaningless.

This is why, as vanesch keeps saying, it's not enough to just say (in the Bell case) "maybe the allegedly-random setting of apparatus orientations wasn't random at all..." There is strong empirical reason to think that the orientations are indeed random enough for the purposes at hand here. So the burden is on you to explain how -- all evidence accumulated so far to the contrary notwithstanding -- these things could actually have gotten correlated up in the necessary way. And doing that means providing an actual example of a theory which predicts this, a super-deterministic theory. And as vanesch has said, this is just a wet dream. Nobody has such a theory. Indeed, as I started out saying, nobody really has even any idea what such a theory might look like, and how such a thing, if it existed, could even be considered as a scientific theory (since it is highly questionable that it could even be rendered empirically testable). And so your idea for a local explanation of the relevant experiments here turns out to be a lot like string theory -- (paraphrasing Peter Woit) not even a wet dream.

 

[ueit]

ueit, you seem to miss (or unfairly minimize) the distinction between a merely deterministic theory, and a super-deterministic theory. The relevant point for this distinction is the idea of "external fields". An external field is a field which has causal influences on the beables of a theory, but which is not itself fixed by those beables.

Any "merely deterministic" becomes superdeterministic when applied to those "external fields", therefore there is no real difference between the two. If BM for example is true then everything in this universe is "fixed" since the big-bang. Free will (defined here as the possibility of multiple choices) is an illusion because there is really only one option, the one determined by the previous state. Assuming both determinism and free external fields is logically contradictory (at least when forces with infinite range, like gravity or EM are described). Sooner or later this contradiction will lead to "paradoxes". It is my opinion that these paradoxes should be dealt with not by abandoning locality (which is required by relativity) but by accepting that the "external field" assumption is false.

Indeed, it is hard to think of an example of a real theory (as opposed to a dream, like string theory... perhaps vanesch would call this a "dry dream" since it hasn't yielded any pleasurable results?!) which doesn't contain such external fields. And the reason for this is clear: scientific theories must be testable if they are going to be believed, and "external fields" are the way that theories leave room for experimenters. Without external fields, it is extremely difficult to imagine how we could test a theory -- or more precisely, it is extremely difficult to understand how the theory would permit something which we could validly describe as a test of the theory. (It'd be like a randomized drug trial in which the assignment of pills to patients was made by the pills themselves -- even if, in the end, the placebo patients did worse than the drug patients, it would be impossible to infer that the drug actually helped those patients, i.e., it would be impossible to believe that we had actually *tested* the drug.)

The GR theory did not require "external fields" to be tested by comparing its predictions relating the orbit of Mercury with what it was observed. Not all theories can be tested in a controlled environment but this doesn't make them useless.

The point is: there are many extant examples of deterministic theories. But all of these leave room for "free human experimenters" in their external fields.

They "leave room for free human experimenters" only with the price of their logical consistency. You cannot have both determinism and multiple choices at the same time.

So it is very misleading when you suggest that the EPR-Bell apparent statistical evidence for non-locality, can be explained away if we merely posit instead a deterministic theory. What you actually need to posit is instead a *superdeterministic* theory, i.e., a theory in which the kinds of things we normally (in science) think of as being "freely or randomly settable", actually *aren't* "freely or randomly settable". So when Weihs et al set up quantum random number generators to determine along which axis a given photon's polarization is measured, you have to believe that that random number generator is actually not random at all (with respect to the state of the incoming photon and whatever else it happens to be entangled with). Rather, by some gigantic cosmic conspiracy, whatever factors are determining the output of the random number generator, are also determining the state of the photon pair (or, I guess, since you want to preserve locality, the factors which immediately determine the output of the random number generator are determined/influenced by some other thing in the distant past which *also* turns out to the relevant thing for influencing the state of the photon pair). And as if that we're insane enough, you have to hold that this magical cosmic conspiracy is clever enough to magically work out exactly the same way, even if the random number generator is switched out in favor of some completely different process, such as a computer running a "random number" algorithm, a young child flipping coins, a human making last-minute "random free-will choices" (with scare quotes because, in superdeterminism, there can't actually be genuine freewill), a chimpanzee throwing either even or odd numbers of poop projectiles, etc. The point is, in order to take your idea seriously, you need to be able to construct a theory in which some causal influence from the distant past influences *all* of these sorts of things, and in just exactly the right way so as to correlate the apparatus settings (which are determined by one of these silly methods) and the state of the photon pair, so as to later fool us into believing that (a) a genuine empirical test of Bell's inequality has occurred and (b) that the outcomes violated the inequality. Or at very least, you have to be able to give at least the hint of some kind of plausibility argument that maybe all of these things *could*, in principle, be so influenced as to bring about the results in question. Without this, all you have is a conspiracy theory.

In my previous post I've provided a sketch of a hypothesis that overcomes the difficulties you've raised. I only need the followings to be true:

1. The detectors (with everything that influences them like computers, humans, monkeys, whatever) are governed by a deterministic law (their future is fixed by their present state). This is true for every deterministic theory.

2. The state of the detectors can be inferred at the source by means of the known fields. (this is true about gravity, not sure about EM field though)

3. Particle pairs are produced only when the detectors have a "favorable" state, allowing the measurements to comply with QM's prediction.

I have to insist here that I do not claim that EPR correlations are false but we are predetermined to see a biased sample (a conspiracy theory). I think they are true, and they are enforced by the way the entangled particles are produced at the source.

This is why, as vanesch keeps saying, it's not enough to just say (in the Bell case) "maybe the allegedly-random setting of apparatus orientations wasn't random at all..." There is strong empirical reason to think that the orientations are indeed random enough for the purposes at hand here. So the burden is on you to explain how -- all evidence accumulated so far to the contrary notwithstanding -- these things could actually have gotten correlated up in the necessary way. And doing that means providing an actual example of a theory which predicts this, a super-deterministic theory. And as vanesch has said, this is just a wet dream. Nobody has such a theory. Indeed, as I started out saying, nobody really has even any idea what such a theory might look like, and how such a thing, if it existed, could even be considered as a scientific theory (since it is highly questionable that it could even be rendered empirically testable). And so your idea for a local explanation of the relevant experiments here turns out to be a lot like string theory -- (paraphrasing Peter Woit) not even a wet dream.

See my idea about such a theory above.

 

[ttn]

2. The state of the detectors can be inferred at the source by means of the known fields. (this is true about gravity, not sure about EM field though)

Suppose the state of the detectors is set by the even-ness or odd-ness in the millionth digit of the energy of some randomly selected CMBR photon. Of course, this photon was emitted in some specific process 15 billions years ago 15 billion light years away. And the past light cone of that emission event does (in the very remote past) overlap with the past light cone of the event at which the particle pair (to be measured) is emitted by the source. So think about what you are saying here: it's not just that "the state of the dectors can be inferred at the source by means of the known fields", but that some truly incomprehensible set of events 10s of billions of years ago has magically set things up so that the particle-pair-emission event and the CMBR photon emission event (events which happen 15 billion light years apart from each other) are correlated -- and not just correlated in any old way, but correlated so that the millionth decimal place in the energy of the one has some connection with the state of the emitted pair.

So you'll have to let us know when you actually produce a theory which explains how this could come about.

3. Particle pairs are produced only when the detectors have a "favorable" state, allowing the measurements to comply with QM's prediction.

Again, you state it misleadingly. It's not just that "particle pairs are produced only" in some favorable circumstances, but that events in the remote remote past have been contrived to produce (by mechanisms which you cannot name and merely envision in the wee hours of the night) correlations between the emitted particle pairs and pretty much everything else in the universe (incoming CMBR photons, the poop-throwing behavior of monkeys, etc). But seriously. There was some lucky bunching of free quarks in the goop 20 billion years ago, which caused a monkey to throw an even number of poops at 3:00 PM on a Tuesday at Fermilab? Give me a break. It's not science. It's a conspiracy theory, except that it's not even a theory but a wet dream.

But that's the last time I'll try to argue that here. The burden is clearly on anyone who advocates this kind of idea to put up or shut up.

 

[ueit]

Suppose the state of the detectors is set by the even-ness or odd-ness in the millionth digit of the energy of some randomly selected CMBR photon. Of course, this photon was emitted in some specific process 15 billions years ago 15 billion light years away. And the past light cone of that emission event does (in the very remote past) overlap with the past light cone of the event at which the particle pair (to be measured) is emitted by the source. So think about what you are saying here: it's not just that "the state of the dectors can be inferred at the source by means of the known fields", but that some truly incomprehensible set of events 10s of billions of years ago has magically set things up so that the particle-pair-emission event and the CMBR photon emission event (events which happen 15 billion light years apart from each other) are correlated -- and not just correlated in any old way, but correlated so that the millionth decimal place in the energy of the one has some connection with the state of the emitted pair. So you'll have to let us know when you actually produce a theory which explains how this could come about.

I’ll make an analogy with gravity. We have two theories:

1. Newtonian gravity (non-local, the force is assumed to propagate instantaneously)
2. GR (local, the changes in space curvature propagate at c)

An analogous experiment would be set up this way:

The “source” is an asteroid pair. The “detectors” are two galaxies, each 1 bln. parsecs away from the “source”. Now, in order for the two asteroids to leave their orbit towards the “detectors”, it is necessary that the stars in those galaxies to be in a certain relative position (otherwise, the resultant force is too weak).

From a Newtonian point of view it looks like the force between “detectors” and “source” acts instantaneously. Once the galaxies have a certain internal structure the asteroids start flying towards them. However, from a GR point of view, the two asteroids follow the space curvature, “carved” by the two galaxies 1 bln years ago, WHEN NONE OF THEM HAD THE REQUIRED STRUCTURE.
This example is quite similar with my hypothesis and it shows that the later is not such absurd as you might think.

Again, you state it misleadingly. It's not just that "particle pairs are produced only" in some favorable circumstances, but that events in the remote remote past have been contrived to produce (by mechanisms which you cannot name and merely envision in the wee hours of the night) correlations between the emitted particle pairs and pretty much everything else in the universe (incoming CMBR photons, the poop-throwing behavior of monkeys, etc). But seriously. There was some lucky bunching of free quarks in the goop 20 billion years ago, which caused a monkey to throw an even number of poops at 3:00 PM on a Tuesday at Fermilab? Give me a break. It's not science. It's a conspiracy theory, except that it's not even a theory but a wet dream.

While I agree that my hypothesis is more like a “wet dream” than a theory, your criticism is not, IMHO, well founded. You assume that physics is somehow scale-dependant as you think that by using more distant or complex detectors would change anything. I think this assumption is wrong. If BM (or other deterministic theory) is right it is irrelevant how far or how old is a system. You simply replace the value of time or coordinate parameter.

 

[reilly]

With all due respect, much of the discussion here, has little to do with physics. Rather it's about philosophy, in which there is seldom a right answer, and thus everyone has great freedom to pontificate, and argue about the maybe.

What so many people, at least in this forum, fail to realize is that physics is phenomena driven -- Nature provides puzzles, scientists try to solve them, ideally with simple math and elegant experiments -- and now with experiments run as if they were the province of small corporations.

With all due respect to the geniuses of String Theory, they are still in business 'cause there's no data to prove or disprove their arguments. With such a good deal, a new priestthood has been born, with hopes of eventual redemption. But now, they are free to do virtually anything, free from the difficult constraints of lot's of data.

Quantum theory, at every level, was born to explain, to provide understanding of very strange, classically impossible phenomena. Now we even have such gems as MWI, and Bohm's gallent efforts to turn back the clock on QM, and who knows what else. The problem with these efforts is that they seemed to be developed without recourse to any data, except that already used in standard QM. What have these alternatives really added?

They don't explain new phenomena, and they give obviously bright people like David Deutsh the chance to come up with some of the silliest arguments physics has ever seen. In his case, I refer particularly to the jive and con of "shadow photons", one of his better nonsensical ideas. (In some sense, we've gone back to pre-Galileo times in which authority was the primary criteria for vetting ideas. Data, while nice, was really not that important.)

A bottom line is: 1, Nature is not always classical.That is, God knows why, Nature is not a 19th century construct; it's far more complex than Newton or Maxwell ever imagined. To suppose that 20th century Nature can and should be explained by 19th century ideas and intellectual structures is, shall we say, pushing research in the wrong direction.

2. Nature is the boss, whether it confoms to our current strictures or not. That QM is, at times, bizarre simply means that to explain the bizarre requires new ways of thinking. To expect that these new ways of thinking should provide a complete, or just extensive world view is, I think, wrong, absurd, and as arrogant as the English physicists at the turn of the century, the TOE folks today, and certainly the weird notions of the anthropic priniciple.(Isn't amazing that these anthropic souls have no more empirical clue than the rest of us about life, and its role and place in the Universe.) As people are won't to say these days, Hello.

Hey, humans are far from the end-all, be-all; we are lucky to be here, and luckier still to have the capabilities to understand Nature. With all due respect, this anthropic stuff, this MWI stuff , belongs in the Monty Python world. These theories have absolutely no empirical foundation, and, in my opinion, are acknowleged only beause a few heavy hitters like John Wheeler like such notions. Back to authority)

3. I believe that when in conflict, human ideas are often trumped by Nature. The best current example is QM. For various understandable reasons, more than a few physicists have problems, some slight, some major, with QM. More often than not, these problems can be associated with 19th century ideas, which are very much at the center of our-day-to-day life. That is, many are unwilling to take the position that it's a new ball game, that many of our old ideas -- causality, notions of the external world, the role of observers, closed theories -- do not necessarily apply to QM as they did to classical 19th century ideas. and, so it seems that many refuse to, or are simply comfortble with the old ideas, and simply can't' accept the sparse quantum world.

The 2-slit experiments drives some to distraction. But, theory and experiment seem to do well together.The strange interference is there, like it or not. What's good about QM in this case is that the basics of 2-slit interference are generalizable, and serve nicely in, for example, scattering experiments. No amount of slick exposition can take the triumph from QM. This means to me, the real issue is to find a framework in which QM is celebrated as an extraordinary human acheivement, and to keep pushing toward a deeper understanding of QM and Nature -- requires experiments and theories that give numbers. Worked before.
Regards
Reilly

 

[vanesch]

Quantum theory, at every level, was born to explain, to provide understanding of very strange, classically impossible phenomena. Now we even have such gems as MWI, and Bohm's gallent efforts to turn back the clock on QM, and who knows what else. The problem with these efforts is that they seemed to be developed without recourse to any data, except that already used in standard QM. What have these alternatives really added?

Well, honestly, it is from this perspective that I see MWI as one of the mental pictures that can make the standard quantum formalism "intuitively understandable". If you have no problems with QM, whatever your view is, then keep it that way! But you often see that one asks for "what is the action at a distance" in EPR situations or "how can this be?" in delayed choice quantum erasers and so on. Personally, I was also bugged by these apparently "nonsensical" (but empirically confirmed) situations, and taking on as a mental picture "MWI" (which is nothing else but telling yourself that the mathematical objects you handle in standard QM are "genuine" in the same way as "position of a particle in space" is genuine in classical mechanics), has helped me to build a "mental picture of what happens" in these EPR experiments and so on (in the same way as - I assume - you have some mental picture of "what happens" when you do Newtonian mechanics).
Now, whether this is "really true" or not doesn't matter for me: I just imagine that what the standard QM formalism tells me, is somehow "what really happens physically" in my mind's eye. As such, I don't have to fiddle with the formalism to get my picture up and running, within that picture, I don't wonder how there is "action at a distance" in EPR, or if there is some backward causal magic in delayed choice quantum erasers.

If this doesn't work for others, then they will have to devise their own way of getting - if they desire so - a different way of understanding quantum theory. I just want to be of some help for those who struggle with things I also struggled with (and many do), and for which I found "ease of mind" with MWI. As I repeated often, to me, MWI is just a way of getting a "feel" for quantum theory as it is now. It is not some credo of some weird religion.

To suppose that 20th century Nature can and should be explained by 19th century ideas and intellectual structures is, shall we say, pushing research in the wrong direction.

I agree with that. That's why I insist on letting the formalism (which has been build upon observation) guide the interpretation, and not vice versa.

 

[vanesch]

Suppose the state of the detectors is set by the even-ness or odd-ness in the millionth digit of the energy of some randomly selected CMBR photon. Of course, this photon was emitted in some specific process 15 billions years ago 15 billion light years away. And the past light cone of that emission event does (in the very remote past) overlap with the past light cone of the event at which the particle pair (to be measured) is emitted by the source. So think about what you are saying here: it's not just that "the state of the dectors can be inferred at the source by means of the known fields", but that some truly incomprehensible set of events 10s of billions of years ago has magically set things up so that the particle-pair-emission event and the CMBR photon emission event (events which happen 15 billion light years apart from each other) are correlated -- and not just correlated in any old way, but correlated so that the millionth decimal place in the energy of the one has some connection with the state of the emitted pair.

I wholeheartedly agree. (nice to agree some times too )

There is a difference between saying: you don't have a completely closed argument to forbid such a theory on one hand, and to propose a reasonable formulation of it which works out in such a spectacular way on the other.

 

[Demystifier]

Now we even have ... Bohm's gallent efforts to turn back the clock on QM ... What have these alternatives really added?

This for example:
http://arxiv.org/abs/quant-ph/0406173
Or this:
http://arxiv.org/abs/hep-th/0407228
http://arxiv.org/abs/hep-th/0601027
Or this:
http://arxiv.org/abs/gr-qc/0611037

 

[ueit]

Suppose the state of the detectors is set by the even-ness or odd-ness in the millionth digit of the energy of some randomly selected CMBR photon. Of course, this photon was emitted in some specific process 15 billions years ago 15 billion light years away. And the past light cone of that emission event does (in the very remote past) overlap with the past light cone of the event at which the particle pair (to be measured) is emitted by the source. So think about what you are saying here: it's not just that "the state of the dectors can be inferred at the source by means of the known fields", but that some truly incomprehensible set of events 10s of billions of years ago has magically set things up so that the particle-pair-emission event and the CMBR photon emission event (events which happen 15 billion light years apart from each other) are correlated -- and not just correlated in any old way, but correlated so that the millionth decimal place in the energy of the one has some connection with the state of the emitted pair.

I wholeheartedly agree. (nice to agree some times too )

The “argument” put forward by ttn is not even an argument but an appeal to emotion (“10s of billions”, “15 billion light years apart”, “millionth decimal place”, “magically”) . I’m sad that you’ve been fooled by it.

Any deterministic theory requires, in order for a good prediction, a good knowledge of the initial parameters. If you choose to introduce as a parameter the properties of a photon generated at the big-bang then you find yourself forced to introduce the big-bang in calculations. The absurdity comes not from the theory but from the physicist who decided to use a practically unknowable parameter in his experiment.
In order to describe Aspect’s experiments I only need information regarding the state of the two detectors, few fractions of a second before detection and nothing else.

My analogy with the “gravitational EPR” gedankenexperiment shows how lame ttn’s “argument” is.

There is a difference between saying: you don't have a completely closed argument to forbid such a theory on one hand, and to propose a reasonable formulation of it which works out in such a spectacular way on the other.

I agree with this but there is no need to misrepresent or exaggerate in order to close that argument.

 

[ttn]

I know I said I wasn't going to comment any more on this thread (god, I think I might have actually said that twice already ) but I can't pass this up because it makes everything so clear:

In order to describe Aspect’s experiments I only need information regarding the state of the two detectors, few fractions of a second before detection and nothing else.

That is simply not true. If you think that, you haven't sufficiently understood Bell's theorem and/or the superdeterminism you are proposing.

If you consider as "given" (relative to ascribing probabilities for the possible outcomes of the nearby measurement) only (a) the state of the particle pair to be measured and (b) information regarding the state of the nearby detector, then you cannot account for Bell-inequality-violating correlations. That's the theorem. On the other hand, if, as I think your statement above is meant to imply, you consider as "given" the above (a) and (b) and in addition "information regarding" the state of the other, distant detector, then your account of the correlations is not local. *Obviously* you can "describe Aspect's experiments (i.e., account for the outcomes) if you allow each outcome to depend on the particle pair state and "information regarding the state of the two detectors, few fractions of a second before detection". But this would not be a *local* explanation of the outcomes, which is supposed to be the whole point, right?

So it is not an appeal to emotion, or some pointless rhetorical distraction, that I bring up all the stuff about billions of years ago and billions of light years away. This is actually what your theory *requires* us to consider if we are going to convert what you say above into an actual *local* explanation of the correlations. It is not good enough to say that the current settings of the apparatus "just happen" to be so as to fool us into thinking we have performed a non-biased measurement of the correlations. You must shoulder the burden of explaining how this (apparent) conspiracy could have come about, and the *only* way to do this *locally* is to search in the distant past for events which have *caused* (*locally*) the two settings and the particle state to get appropriately correlated. So, good luck with that.

My good friend and philosophical partner-in-arms vanesch will have to take up any further discussion of this on my behalf because this time I'm serious when I say I won't comment anymore! (Unless I change my mind...)

 

[vanesch]

Any deterministic theory requires, in order for a good prediction, a good knowledge of the initial parameters. If you choose to introduce as a parameter the properties of a photon generated at the big-bang then you find yourself forced to introduce the big-bang in calculations. The absurdity comes not from the theory but from the physicist who decided to use a practically unknowable parameter in his experiment

The point is that it is not very difficult to set into action an experiment that uses almost ANY kind of measurement to decide upon the settings of the polarizers. This can range from cosmic microwave background noise, to stuff happening in the brain of an experimenter, to sampling a song of the 1930ies... you can pick MIRIADS of ways to select the polariser settings, based upon totally different kinds of physical systems, and ALL of them have to agree upon the RIGHT polarizer settings in order to generate the right EPR correlations. That is, if we use samples of a song of Vera Lynn to decide upon the polarizer settings, then these samples have to be in agreement with the polarisations of the sent pairs of light pulses so as to generate EPR correlations. But we could decide in the middle of the experiment, to switch to CMB radiation noise as a decider for the polarizations, and this should simply continue. Etc...
In other words, all these totally different physical quantities have to be in agreement to generate the correct polarizer settings in agreement with the pairs sent out.

In order to describe Aspect’s experiments I only need information regarding the state of the two detectors, few fractions of a second before detection and nothing else.

The point is that that "information" can be Vera Lynn's song, or the CMB radiation or whatever, and this all has to come out correctly.

My analogy with the “gravitational EPR” gedankenexperiment shows how lame ttn’s “argument” is.

Not really. In the gravitational EPR experiment you propose, everything is a simple gravitational interaction. Your example doesn't work anymore if there is one single electromagnetic phenomenon happening, which "helps align" the stars. It is clear that if a gravitational EPR experimenter had the impression of having some freedom to align or not, the stars in the galaxy, that the argument wouldn't hold anymore - even if this freedom is only apparent. It simply means that the experimenter has some "way of deciding" to his disposal to align, or not, the stars in the galaxy, which means essentially that the experiment can be set up in such a way that alignment or not of the galaxies can be correlated with just any other phenomenon.


I agree with this but there is no need to misrepresent or exaggerate in order to close that argument.[/QUOTE]

 

[ueit]

I know I said I wasn't going to comment any more on this thread (god, I think I might have actually said that twice already ) but I can't pass this up because it makes everything so clear:

In order to describe Aspect’s experiments I only need information regarding the state of the two detectors, few fractions of a second before detection and nothing else.

That is simply not true. If you think that, you haven't sufficiently understood Bell's theorem and/or the superdeterminism you are proposing.

If you consider as "given" (relative to ascribing probabilities for the possible outcomes of the nearby measurement) only (a) the state of the particle pair to be measured and (b) information regarding the state of the nearby detector, then you cannot account for Bell-inequality-violating correlations. That's the theorem. On the other hand, if, as I think your statement above is meant to imply, you consider as "given" the above (a) and (b) and in addition "information regarding" the state of the other, distant detector, then your account of the correlations is not local. *Obviously* you can "describe Aspect's experiments (i.e., account for the outcomes) if you allow each outcome to depend on the particle pair state and "information regarding the state of the two detectors, few fractions of a second before detection". But this would not be a *local* explanation of the outcomes, which is supposed to be the whole point, right?

Bell's theorem proves that the properties of entangled particles and detectors' orientation are not statistically independent parameters. The theorem is silent about the mechanism by which the correlations are realized. It doesn't require it to be non-local.
Aspect's experiment used a fast-switching device to set the detector while the particles were in flight. Does the experimental result rule out a local mechanism behind the correlations? Not at all.

As I said, few fractions of a second before detection (the time it takes the photon to travel from the source to the detectors), just before the entangled photons are released, the information regarding the detector's state arrives at the source. Because I propose a local mechanism, that information is not "up to date" but this is not a problem. We've agreed that the detectors are deterministic systems, therefore knowing their state at a certain time is enough to predict their state in the future. So, when the particles leave the source they "know" how they will be detected even if they only have old information.

To better illustrate my point, let's say some aliens set up an EPR-type experiment. The source is placed on Earth and the detectors about 1 parsec away.
But the humans found out about the experiment and wanted to make fun of the aliens by sending them a message encoded in the experimental data. But, they don't know how the aliens will perform the measurements so what to do? Well someone comes with the idea of building a telescope so powerful that it can see the electrons and quarks the detectors are made of. He then puts the data into a computer and calculates the detectors' state one year into the future, when the measurement will be performed. It doesn't matter how the aliens will do to "chose" the detectors' orientation. A quick "scan" of their brain will inform humans about the so-called choice.

In conclusion, as I said before, there is no need for a non-local mechanism if the detectors are predictable and if their state (even if an old one, it doesn't matter) can reach the source via EM or other signals traveling at or bellow the speed of light.

Of course, finding the exact mechanism could be a difficult task, perhaps more dificult than going from Newtonian gravity to GR, but the fact that such an mechanism is allowed by Bell's theorem and all experiments performed to date is quite important.

 

[ueit]

The point is that it is not very difficult to set into action an experiment that uses almost ANY kind of measurement to decide upon the settings of the polarizers. This can range from cosmic microwave background noise, to stuff happening in the brain of an experimenter, to sampling a song of the 1930ies... you can pick MIRIADS of ways to select the polariser settings, based upon totally different kinds of physical systems, and ALL of them have to agree upon the RIGHT polarizer settings in order to generate the right EPR correlations. That is, if we use samples of a song of Vera Lynn to decide upon the polarizer settings, then these samples have to be in agreement with the polarisations of the sent pairs of light pulses so as to generate EPR correlations. But we could decide in the middle of the experiment, to switch to CMB radiation noise as a decider for the polarizations, and this should simply continue. Etc...
In other words, all these totally different physical quantities have to be in agreement to generate the correct polarizer settings in agreement with the pairs sent out.

The point is that that "information" can be Vera Lynn's song, or the CMB radiation or whatever, and this all has to come out correctly.

The key issue here is that those "totally different physical quantities" are not as different as you claim they are. If, at the bottom, all particles follow deterministic trajectories and if they can "communicate" their state via a local mechanism, similar to gravity in GR, it doesn't matter how complex or big a system is or what internal distribution it has. A button-pressing monkey or a computer analyzing "Vera Lynn's song" are, after all both quark-electron systems and I see no reason to look for different physical laws to describe each of them.

Not really. In the gravitational EPR experiment you propose, everything is a simple gravitational interaction. Your example doesn't work anymore if there is one single electromagnetic phenomenon happening, which "helps align" the stars. It is clear that if a gravitational EPR experimenter had the impression of having some freedom to align or not, the stars in the galaxy, that the argument wouldn't hold anymore - even if this freedom is only apparent. It simply means that the experimenter has some "way of deciding" to his disposal to align, or not, the stars in the galaxy, which means essentially that the experiment can be set up in such a way that alignment or not of the galaxies can be correlated with just any other phenomenon.

Sure, in my example, you can appeal to EM to "fool" gravity. But how can you propose to "fool" an electron or a quark if it follows a deterministic trajectory? You should use something else that is not "covered" by QM, a different kind of force. I don't have any idea of what that may be.

So, if you are limited to what QM (or better QFT) describes, and if the theory is deterministic you are like an astronaut, without any propulsion system, orbiting some planet. He might believe that it goes round and round because of his "free will" but he can do nothing but "choose" to follow that trajectory.

 

[reilly]

This for example:
http://arxiv.org/abs/quant-ph/0406173
Or this:
http://arxiv.org/abs/hep-th/0407228
http://arxiv.org/abs/hep-th/0601027
Or this:
http://arxiv.org/abs/gr-qc/0611037

Interesting work indeed. But where does it go? Theory is nice, particularly when it is close to experiment. I first enountered Bohm's work some 50 years ago. It's still outside the mainstream, beause it has contributed little or nothing to further any branch of physics that's connected to experiments. My sense is that the arguments about Bohm have not changed much over the last 50 years.

Regards,
Reilly Atkinson

 

[reilly]

Now, whether this is "really true" or not doesn't matter for me: I just imagine that what the standard QM formalism tells me, is somehow "what really happens physically" in my mind's eye. As such, I don't have to fiddle with the formalism to get my picture up and running, within that picture, I don't wonder how there is "action at a distance" in EPR, or if there is some backward causal magic in delayed choice quantum erasers.

If this doesn't work for others, then they will have to devise their own way of getting - if they desire so - a different way of understanding quantum theory. I just want to be of some help for those who struggle with things I also struggled with (and many do), and for which I found "ease of mind" with MWI. As I repeated often, to me, MWI is just a way of getting a "feel" for quantum theory as it is now. It is not some credo of some weird religion.



I agree with that. That's why I insist on letting the formalism (which has been build upon observation) guide the interpretation, and not vice versa.

Sounds good to me. I find MWI confusing, and take my Born interpretation with a grain of salt.

I think about an imaginary QM world based largely on a three-point interactions, transitions if you will. I think sometimes pictures of such are called Feynman Diagrams. Like many,I think the standard diagrams are a God send. But, it would be a great stretch to say they represent real processes -- but then many don't

Regards,
reilly

 

[vanesch]

The key issue here is that those "totally different physical quantities" are not as different as you claim they are. If, at the bottom, all particles follow deterministic trajectories and if they can "communicate" their state via a local mechanism, similar to gravity in GR, it doesn't matter how complex or big a system is or what internal distribution it has. A button-pressing monkey or a computer analyzing "Vera Lynn's song" are, after all both quark-electron systems and I see no reason to look for different physical laws to describe each of them.

Yes, but look at the following. We have the impression that we can "decide" about the settings of the polarizers. This means that we can essentially correlate this with almost any other phenomenon. Now, I don't want to drift in a discussion here about "free will" - I can accept that free will is an illusion in a deterministic universe, but nevertheless, it has a meaning. The meaning is that the choices can be based upon "arbitrary" correlations (human brain states, photons coming from the back end of the universe, samples of songs sung long ago...). We have learned somehow, throughout the entire evolution of science (and even before that), that when things seem to depend upon "free will" that they are statistically independent. In fact, the whole endeveour of science is based upon that idea! "Controlled experiments" are nothing else but exactly that: we change "arbitrarily" some parameters while keeping others constant, and we deduce "causal relations" from that. You push arbitrarily the button, and you see the light go on. You do that 50 times, and you conclude that there is some causal relationship between you pushing the button as cause, and the light going on as consequence... maybe due to the current that will flow in the wire, or maybe for another reason. In fact, THE WHOLE OF SCIENCE (and all of our common sense knowledge) has been deduced directly or indirectly that way: what we can change "arbitrarily" must be the "cause" and what is "observed" is the consequence. Even the whole idea of a deterministic time evolution comes from those observations (and their theoretical extrapolations) in the first place.
Now, whether we "really" had some choice, or whether we were just passive observers of spurious correlations between disparate events (such as me suddenly having a desire to push the button which is the result of some brain activity) doesn't matter. What matters is that we know when things are "arbitrary choices" we (think we can) set, and which find their origin if not in free will, then in some processes that have "nothing to do" with the phenomenon that is to be observed.

Now, if we are going to claim, that the light went on just by some thermodynamical coincidence (suddenly, the random motion of the atoms in the wire increased a lot as a statistical fluctuation), and that, exactly at that time, in my brain, there was a process that gave me some desire to push the button, and that this coincidence repeats itself several times, then indeed, I will be fooled into thinking that there is a causal relation between me pushing the button, and the light going on, but in fact there is none. But then, ALL WE KNOW is based upon erroneous deductions of causal relationships which were in fact nothing else but funny coincidences.

This is different with your GR "experiment". In GR, if we don't use electromagnetism, we wouldn't "have the impression that we had the freedom to choose when the galaxies aligned". We would just be passively observe the galaxies evolve, see them align, and then see that our asteroids got pulled apart. We wouldn't have that same impression of 'being able to decide when we push the switch". Indeed, in a passively evolving system, where we cannot "arbitrarily decide", we don't know whether correlations represent cause-effect relationships. At no point, we would be fooled into thinking that we had a choice.

And finally, my argument against "unknown physics/conspiracies/..." to explain EPR correlations is this: quantum mechanics predicts them correctly. That means that all the physical knowledge of the mechanism is already included in one way or another in quantum theory. Now, we might not understand this totally, or quantum theory might be an "effective formalism" of a deeper theory of course. But it would be very strange that a superdeterministic theory of which we don't even have a clue how it might be put together (in such a way that samples of a song of Vera Lynn are correlated with the photon pairs emitted in a lab 60 years later) produces as only effect, exactly those effects that are nicely described by the current quantum formalism. A quantum formalism that was deduced by looking at observations that were often made by using "free choices" of settings to derive "causal relationships".

Again, I don't say that superdeterminism is impossible. But it stretches imagination (much more so, imo, than the extravagance of MWI!).

I have outlined several arguments:

1) The argument that was used against MWI in the beginning of this thread (namely, that MWI supposes a difference between observation and "reality", while that MWI was derived by looking at QM, which was a formalism that was derived by looking at observations, probably by people who thought that observations corresponded to reality), can be, in a much stronger form, used against superdeterminism:
If superdeterminism is true, then all "causal relationships" we've ever derived were the result of funny correlations without the slightest bit of causality. This includes all we think we know, because all our knowledge is based upon the assumption that free choices are uncorrelated with what is going to be observed.

2) There is not the slightest hint of an existing theory that does this, even though it is not impossible.

3) All that trouble for finding, finally, something that we know already: the EPR correlations are correctly predicted by quantum mechanics as we know it today.

He might believe that it goes round and round because of his "free will" but he can do nothing but "choose" to follow that trajectory.

He wouldn't. If he had the impression of free will, he would find correlations with several distinct phenomena. He could say: now, the planet will have the position given by a 'random number generator'. Next, he could say: now the planet will have the position given by samples of my favorite song. Next, he could say...
Of course, this is maybe not truly free will, but the point is that the correlations could be with just anything. If this happens, THEN he will have the impression of "being able to decide where the planet is". And if, as a function of that position of the planet, he observes certain phenomena (like tides on another planet or something), he might start to think that there is a causal relationship. But he wouldn't have that impression from just the planet following passively an orbit on which he "cannot decide" how to "change" it.

I'm not fooled into thinking that the sun's position is causally determined by the hands on my watch. If however, I can "change" the hands of my watch "arbitrarily" and the sun would "follow" instantaneously, then I would be quite puzzled. But for that, I have to be able to get the impression that I can freely change the position of the hands of my watch.

 

[tehno]

I first enountered Bohm's work some 50 years ago. It's still outside the mainstream, beause it has contributed little or nothing to further any branch of physics that's connected to experiments. My sense is that the arguments about Bohm have not changed much over the last 50 years.

To me,Bohm's work is an interesting approach how to interpret some aspects of QM.I don't see how it can contribute to any branch of experimental physics or predict something new what can't be predicted by an ortodox school of QM.Therefore ,as such,it's just a try to "see" what is in background of QM and its' equations. What would Einstein say about Bohm's interpretations?

 

[vanesch]

To me,Bohm's work is an interesting approach how to interpret some aspects of QM.I don't see how it can contribute to any branch of experimental physics or predict something new what can't be predicted by an ortodox school of QM.Therefore ,as such,it's just a try to "see" what is in background of QM and its' equations.

Right. That's also how I look upon Bohm (and MWI, btw).

What would Einstein say about Bohm's interpretations?

Now, that's an interesting question

He would have liked the "realist" aspect, and disliked the fact that relativity goes out of the window, I guess (but then, who am I to say what Einstein would have thought :shy: ...)

 

[Anonym]

Quote:

What would Einstein say about Bohm's interpretations?

Now, that's an interesting question.

He would have liked the "realist" aspect, and disliked the fact that relativity goes out of the window, I guess (but then, who am I to say what Einstein would have thought)

For your pleasure

A. Einstein to M.Born (12.5.1952):

”Did you see as Bohm (as by the way de Broglie 25 years ago) believes that the quantum theory may be interpreted deterministically otherwise? It, in my view, cheap consideration, however, your judgment is better indeed.”

Regards, Dany.

 

[ttn]

Now, that's an interesting question

He would have liked the "realist" aspect, and disliked the fact that relativity goes out of the window, I guess (but then, who am I to say what Einstein would have thought :shy: ...)

Actually, Einstein lived long enough to know about Bohm's theory. He famously remarked that it was "too cheap" and basically dismissed it. Vanesch is probably right that the main reason for this is that relativity goes out the window. It's also relevant that Einstein had more or less invented Bohm's theory himself about 40 years earlier, and had rejected it way back then precisely because it seemed to conflict with relativity. Einstein never got fully behind de Broglie's discovery of the same theory either, no doubt for about the same reason. And then, between the 20s (when de Broglie proposed this theory) and the 50s (when Bohm rediscovered it), Einstein had been working on his "unified field theory" idea, which was supposed to be some kind of massive replacement for QM which unified E&M and General Relativity and sort of "got underneath" the need for anything like the by-then-standard QM formalism. So I think that is basically what Einstein meant when he said Bohm's theory was "too cheap" -- it kept too much of the standard QM formalism, and just added something on top of it, something which *did not fix* what was for Einstein *the fundamental* problem with orthodox QM, namely the conflict with relativity. So while Einstein probably appreciated the conceptual clarity of Bohm's theory (relative to copenhagen) he didn't see Bohm as something that solved the fundamental problem, and simultaneously saw it as something that wasn't nearly radical enough in its break with orthodoxy. Hence, "too cheap".

However, unfortunately, Einstein did not live long enough to witness the next major development in this story, namely Bell's discovery (motivated by Bohm's theory, by the way... so much for the idea that Bohm's theory never led to any important new discoveries) that the non-locality (the anti-relativistic character) of Bohm's theory was *unavoidable* in any theory making (what we now know are) the empirically correct predictions for a certain class of experiments. So it wasn't, as Einstein thought, a matter of working to find a theory that would solve the fundamental problem with orthodox QM (its nonlocality), but rather of coming to grips with the fact that something is wrong or incomplete about relativity (or we've misunderstood it if we think it outlaws nonlocality). And once this is fully grasped (most still don't grasp it today), it removes completely the one otherwise-valid objection to Bohm's theory, namely that relativity goes out the window. If Einstein had lived long enough to see Bell's Theorem, he probably would have understood this and would have had a very very different reaction to Bohm's theory.

Sadly, though, that has to remain speculation.

 

[tehno]

However, unfortunately, Einstein did not live long enough to witness the next major development in this story, namely Bell's discovery (motivated by Bohm's theory, by the way... so much for the idea that Bohm's theory never led to any important new discoveries) that the non-locality (the anti-relativistic character) of Bohm's theory was *unavoidable* in any theory making

But wasn't Bell's discovery, you are reffering to,and associated experiments motivated also by Einstein's so called *spooky action at distance* thought experiment?
Famous "Quantum entanglement" ,and that our Universe is quantumly coupled (with or without relativity) are consequence of ortodox QM theory.
Or am I mistaken?

 

[Anonym]

Actually, Einstein lived long enough to know about Bohm's theory... If Einstein had lived long enough to see Bell's Theorem, he probably would have understood this and would have had a very very different reaction to Bohm's theory.

I read your essay- post #101. I have two questions:

1. Do you have professional education in physics?
2. If yes, who was your teacher?

In addition, I am not familiar with “what was for Einstein *the fundamental* problem with orthodox QM, namely the conflict with relativity”. A. Einstein never talks about nonexistent problems; the problem was formulation of consistent theory of measurements.

Regards, Dany.

 

[ueit]

Yes, but look at the following. We have the impression that we can "decide" about the settings of the polarizers. This means that we can essentially correlate this with almost any other phenomenon.

OK, but keep in mind that this "impression that we can "decide" doesn't mean much, especially when dealing with unintuitive experiments like EPR. We have also the impression that we don't exist in a superposition of states, that we live in a local universe, and so on. However, QM forces us to accept that at least some of these impressions are false, so why not drop the "free will" one?

And because all matter have the same basic components (mainly electrons and quarks) any "phenomenon" you choose as the source of "random" settings brings nothing new to the experiment. It is like trying to "fool" energy conservation by using all kinds of devices (chemical, mechanical, nuclear and so on) and wondering each time about the failure to do so. The explanation is the same. Energy conservation is enforced at the fundamental particle level and no matter how you play with different devices you cannot brake the law.

Now, I don't want to drift in a discussion here about "free will" - I can accept that free will is an illusion in a deterministic universe, but nevertheless, it has a meaning. The meaning is that the choices can be based upon "arbitrary" correlations (human brain states, photons coming from the back end of the universe, samples of songs sung long ago...).

In fact free will is not the problem here, and neither randomness (understood as a lack of pattern, not as a lack of cause), but only predictability. And I think we agree that in a deterministic world everything is predictable.

We have learned somehow, throughout the entire evolution of science (and even before that), that when things seem to depend upon "free will" that they are statistically independent. In fact, the whole endeveour of science is based upon that idea! "Controlled experiments" are nothing else but exactly that: we change "arbitrarily" some parameters while keeping others constant, and we deduce "causal relations" from that. You push arbitrarily the button, and you see the light go on. You do that 50 times, and you conclude that there is some causal relationship between you pushing the button as cause, and the light going on as consequence... maybe due to the current that will flow in the wire, or maybe for another reason. In fact, THE WHOLE OF SCIENCE (and all of our common sense knowledge) has been deduced directly or indirectly that way: what we can change "arbitrarily" must be the "cause" and what is "observed" is the consequence. Even the whole idea of a deterministic time evolution comes from those observations (and their theoretical extrapolations) in the first place.
Now, whether we "really" had some choice, or whether we were just passive observers of spurious correlations between disparate events (such as me suddenly having a desire to push the button which is the result of some brain activity) doesn't matter. What matters is that we know when things are "arbitrary choices" we (think we can) set, and which find their origin if not in free will, then in some processes that have "nothing to do" with the phenomenon that is to be observed.

1. I think you overestimate the importance of free choice here. There are branches of science where we are forced in the position of passive observers, like galaxy formation or paleontology.

2. Even in your example I don't see any problem. The causality remains, we still observe that whenever the button is pressed the light is on.

3. The device that sets the detector in an EPR experiment generates an EM field by its own existence. I don't see any good reason to assume that this field has "nothing to do" with the experiment.

Now, if we are going to claim, that the light went on just by some thermodynamical coincidence (suddenly, the random motion of the atoms in the wire increased a lot as a statistical fluctuation), and that, exactly at that time, in my brain, there was a process that gave me some desire to push the button, and that this coincidence repeats itself several times, then indeed, I will be fooled into thinking that there is a causal relation between me pushing the button, and the light going on, but in fact there is none. But then, ALL WE KNOW is based upon erroneous deductions of causal relationships which were in fact nothing else but funny coincidences.

You seem to be building a case against the "bag of events" hypothesis where we have coincidences without a causal, deterministic mechanism (the random fluctuation in the wire). I reject such conspiracy theories, just like you and I see no similarity between them and (super)determinism. On the contrary, I propose a causal connection between the generation of entangled particles and the surrounding fields. Nothing happens at random and there are no "erroneous deductions". The correlations are not an illusion, they reflect the underlying deterministic mechanism.

This is different with your GR "experiment". In GR, if we don't use electromagnetism, we wouldn't "have the impression that we had the freedom to choose when the galaxies aligned". We would just be passively observe the galaxies evolve, see them align, and then see that our asteroids got pulled apart. We wouldn't have that same impression of 'being able to decide when we push the switch". Indeed, in a passively evolving system, where we cannot "arbitrarily decide", we don't know whether correlations represent cause-effect relationships. At no point, we would be fooled into thinking that we had a choice.

The GR was meant to be an analogy for how QM interactions take place. So, you have to "replace" GR for EM-interactions wherever these appear, including the experimenter's brain. So, if QM is deterministic we are "passively evolving systems" just like the other parts of the experiment. So, the experimenter is a galaxy, the detector another one, a computer, if present, another one and so on. Regardless of what each "galaxy" does or believes, the asteroids (our entangled particles) "know" exactly the detector configuration at the detection time.

And finally, my argument against "unknown physics/conspiracies/..." to explain EPR correlations is this: quantum mechanics predicts them correctly. That means that all the physical knowledge of the mechanism is already included in one way or another in quantum theory.

Just like all chemistry "is already included in one way or another" in thermodynamics.

Now, we might not understand this totally, or quantum theory might be an "effective formalism" of a deeper theory of course. But it would be very strange that a superdeterministic theory of which we don't even have a clue how it might be put together (in such a way that samples of a song of Vera Lynn are correlated with the photon pairs emitted in a lab 60 years later) produces as only effect, exactly those effects that are nicely described by the current quantum formalism. A quantum formalism that was deduced by looking at observations that were often made by using "free choices" of settings to derive "causal relationships".

I find the alternatives no less strange and the possibility (even if remote) to recover the nice classical ontology in QM world very attractive. Again, you make it sound more unlikely that it is. After all a computer and a physicist will generate the same amount of energy per each Kg, they fall at the same speed, etc. Why not behave similarly in an EPR experiment too?
I agree that here it's not the place to discuss "free choice" but I think there are experiments proving that someone can be determined to "freely choose" what to do. I'll try to find some links and post them here and close this argument.

Again, I don't say that superdeterminism is impossible. But it stretches imagination (much more so, imo, than the extravagance of MWI!).

I have outlined several arguments:

1) The argument that was used against MWI in the beginning of this thread (namely, that MWI supposes a difference between observation and "reality", while that MWI was derived by looking at QM, which was a formalism that was derived by looking at observations, probably by people who thought that observations corresponded to reality), can be, in a much stronger form, used against superdeterminism:
If superdeterminism is true, then all "causal relationships" we've ever derived were the result of funny correlations without the slightest bit of causality. This includes all we think we know, because all our knowledge is based upon the assumption that free choices are uncorrelated with what is going to be observed.

No, superdeterminism is just the logically consistent determinism, and is completely different from "bag of events" conspiracy theories. Our "free choices" are certainly correlated with observation.

2) There is not the slightest hint of an existing theory that does this, even though it is not impossible.

Yeah, that's bad.

3) All that trouble for finding, finally, something that we know already: the EPR correlations are correctly predicted by quantum mechanics as we know it today.

I don't think it's useless to find a mechanism by which the statistical predictions of QM are accomplished.

He wouldn't. If he had the impression of free will, he would find correlations with several distinct phenomena. He could say: now, the planet will have the position given by a 'random number generator'. Next, he could say: now the planet will have the position given by samples of my favorite song. Next, he could say...
Of course, this is maybe not truly free will, but the point is that the correlations could be with just anything. If this happens, THEN he will have the impression of "being able to decide where the planet is". And if, as a function of that position of the planet, he observes certain phenomena (like tides on another planet or something), he might start to think that there is a causal relationship. But he wouldn't have that impression from just the planet following passively an orbit on which he "cannot decide" how to "change" it.

I'm not fooled into thinking that the sun's position is causally determined by the hands on my watch. If however, I can "change" the hands of my watch "arbitrarily" and the sun would "follow" instantaneously, then I would be quite puzzled. But for that, I have to be able to get the impression that I can freely change the position of the hands of my watch.

See above my clarification regarding the correct understanding of my analogy

 

[ttn]

But wasn't Bell's discovery, you are reffering to,and associated experiments motivated also by Einstein's so called *spooky action at distance* thought experiment?

You mean the EPR argument? That was an argument that a local explanation of certain correlations predicted by QM would require deterministic local hidden variables. (And secondarily that, as long as one accepts locality, orthodox QM is therefore incomplete since it doesn't contain those variables.) Or put it this way: EPR proved that *the only hope for locality* is a certain specific type of hidden variable model.

Bell's theorem is then that this can't work, that the one kind of model that EPR showed *might* be able to explain things locally, can't. So *nothing* can explain things locally. Non-locality is a real fact of nature.

Famous "Quantum entanglement" ,and that our Universe is quantumly coupled (with or without relativity) are consequence of ortodox QM theory.
Or am I mistaken?

I don't really understand what you said. Orthodox QM is not a local theory. That's the whole point of EPR. What bothered Einstein was the spooky action at a distance that OQM posited. He hoped to get rid of that by introducing something like the hidden variables I mentioned above. But Bell proved that that won't work -- you can't get rid of the non-locality (and keep the right empirical predictions, *even* by adding hidden variables). So non-locality is a real fact of nature.