EPR paradox revisited, again. hehehe

[ttn]

Ok. Bob and Alice adhere to the speed of light limit. They used to believe QM was an adequate description of reality. However since QM clearly violates the speed of light limit they discard QM and (in answer to your question) they begin to look for an adequate description of reality that does not violate the speed of light limit.

Now you've got it. That's exactly what Einstein thought. This is exactly the EPR argument (or what the EPR argument was supposed to be, before Podolsky botched the writing of the paper...). Orthodox QM provides a nonlocal explanation of these correlations. If you want to have a local theory, you have to look elsewhere -- specifically, you have to posit a certain kind of hidden variables that determine the outcomes on each side (independent of what's going on on the other side). That is just exactly the point of EPR.

...which means you're now ready for the *second* part of the proof that nature is nonlocal: Bell's Theorem. Bell's Theorem proves that the kind of theory just posited in the last paragraph (posited, remember, on pain of violating locality!) cannot reproduce the full slew of QM predictions. Thus, *no* theory whatsoever (respecting a certain locality condition, Bell Locality) can be consistent with the full slew of QM predictions. And since we know from experiment that those predictions are *correct*, it follows that nature ain't Bell Local.

 

[Sherlock]

... I restate my original appeal: I ask for a clear proof that OQM does not violate the speed of light limit.

There can't be a proof that OQM doesn't violate locality for the same reason that there can't be a proof that OQM does violate locality. OQM isn't a causal theory, it's a probabilistic one. There are correlations, that's all. The underlying physical reasons for the correlations are a mystery.

The assumption that the clearly deterministic or causal components of the mathematical structure of quantum theory are necessarily a 1-1 mapping of an underlying quantum world is just that --- an assumption. It's not an assumption that's part of OQM. OQM is the probabilistic interpretation.

Given the assumption that formal qm is a 1-1 mapping of an underlying quantum world, then maybe it's possible to infer that quantum theory (and, hence, underlying reality --- since it's being assumed that the theory is a description of underlying reality) is or isn't nonlocal. But, given that assumption, you're not dealing with OQM --- and, given that assumption, you might, as well, infer the existence of the alternative realities of MWI.

So, the proof you seek depends on what is assumed about the relationship of the formalism of quantum theory to an underlying quantum world. Since we have no sensory apprehension of this underlying reality, the wisest course seems to me to make no assumption about this relationship given the extant data. This is what OQM does, at least wrt its pedagogical presentation. It treats this consideration as essentially meaningless --- even though it's probably a good bet that at least some of the physicists who apply and develop the theory DO have their own ideas and intuitions about such a relationship.

Wrt OQM then, Alice and Bob are not affecting each other. The rate of detection, the probability of detection, at one end is not altered by events at the other end.

Make an assumption about the relationship of the theory to an underlying reality and it's a different story. The catch is that there's no definitive way to ascertain whether or not such an assumption is true.

 

[vanesch]

Ok, since you believe I misunderstand his "question" (in fact there are 3 question marks in his post, indicating three questions) will you rephrase his post and state the _question_ (singular) which you think I have not answered please?

The question I asked (and which Doc Al clearly understood) is: do you conclude a faster-than-light emission FROM THE DATA or do you conclude it from the formalism and its interpretation ?

I had the impression, in your post #8, that you concluded it FROM THE DATA.

And as your data (for the example that you presented in post #8) are identical with data that you can obtain with a set of colored balls, I was wondering how you could do so.

There's only one question mark here, so there should be no ambiguity as to what the question is

 

[alfredblase]

Leandros,

if anyone else can make sense of your objections and put them forward in a clear manner I shall try and answer them. I'm sorry but I do not have a clue what you are talking about.

 

[alfredblase]

I think what you mean is: the state of Bob's particle changes (when Alice makes her measurement). It goes from something "fuzzy" in regard to the value of a certain attribute, to something "definite" in regard to that attribute. That's all true.

we agree on this then.

And I think we agree that that involves a kind of nonlocality (a kind which can be made precise using a precise def'n of locality such as Bell Locality).

From here on you mention locality many times, (something I haven't mentioned at all). Please give the definition of locality given by Bell Locality.

Thanks.

 

[alfredblase]

sherlock you also refer to locality as being the violated principle. Is your definition of locality that given by Bell Locality or do you use another one? If another please give this definition.

Thanks.

 

[ttn]

From here on you mention locality many times, (something I haven't mentioned at all). Please give the definition of locality given by Bell Locality.

Thanks.

Bell discusses this in detail in several essays, including "La Nouvelle Cuisine" and "The Theory of Local Beables." Both essays appear in the 2nd edition of "Speakable and Unspeakable". If you don't have easy access to either of those, there's a nice discussion in section 2 of this paper

http://www.arxiv.org/abs/quant-ph/0601205

which is basically a rigorous version of the EPR argument formulated in terms of Bell Locality.

 

[alfredblase]

The question I asked (and which Doc Al clearly understood) is: do you conclude a faster-than-light emission FROM THE DATA or do you conclude it from the formalism and its interpretation?

The answer is that I never conclude a faster than light emission.

But I think you also mean to test my understanding of what is in question...

I suspect you also mean to ask me if I think there is a difference between an experiment involving classical objects such as pairs of oppositely coloured balls, and one involving objects that exhibit QM properties such as pairs of electrons with opposite spin angular momentum, right?

QM tells us that spin is a non-commuting observable. This tells us that until we measure the spin of one of the entangled electrons, in a particular direction, we are violating local realism. Meaning that we are violating the assumption that all properties are defined in reality as being something in particular (up or down in this case), even if a particular properety has never been measured.

Coloured balls on the other hand... well there are some who would argue that... well anyway we need not concern ourselves with balls because there is no QM wavefunction defining the colour of billiard sized balls, and so since we are performing this experiment to test the validity of QM we need not discuss balls.

Thanks.

 

[alfredblase]

The following is Bell’s description of the locality con-
dition, along with his accompanying figure:

“A theory will be said to be locally causal
[i.e., what we are calling Bell Local] if the
probabilities attached to values of local be-
ables in a space-time region 1 are unaltered
by specification of values of local beables in
a space-like separated region 2, when what
happens in the backward light cone of 1 is al-
ready sufficiently specified, for example by a
full specification of local beables in a space-
time region 3.”

(from that paper ttn linked in post 57, oh and I didn't include the figure)

what is a beable?

oh and concerning my experiment, what do you consider region 1 to be? region 2? and region 3?

what is considered a "full specification"?

 

[ttn]

what on Earth is a beable?

It's a kind of semi-joke of Bell's. Standard quantum mechanics talks a lot about observ-ables. Bell thought that any real theory ought, in addition, to specify that which is *real* -- not just what can be observed, but what *is*. Hence "be-able" as an alternative to "observ-able". As has been pointed out many times, this neologism is unfortunate in a way, since it implies that "beables" are only potentially real, whereas Bell actually meant to refer not to a mere potential but to what is really actually real.

An example he gives somewhere: in Maxwellian electrodynamics, the E and B fields (and functionals of them) are beables, while the potentials V and A are not. That is why nobody thinks locality is violated when, in Coulomb gauge, the potential V at some point changes instantaneously when a distant charge is moved. By contrast, if the fields E or B were to change instantaneously, that *would* be a problematic kind of nonlocality since those fields are supposed to represent "beables".

Does that clarify? The point is basically: every theory has to say that something or other is real; it has to be *about* *something*. And then Bell Locality is the requirement that those real somethings shouldn't be causally affected by stuff outside of the past light cone.

One way of reading orthodox QM is to take wave functions as beables. If we do that, the theory violates Bell Locality. On the other hand, if we don't do that -- and hence say that the theory has no beables at all, then the question of the theory's locality becomes meaningless because, really, it isn't even a theory unless it asserts *something* about the way the world works.

 

[alfredblase]

thanks ttn, I also added more questions afterwards, sorry I tend to post and keep retouching them until I'm happy.

 

[leandros_p]

Leandros,

if anyone else can make sense of your objections and put them forward in a clear manner I shall try and answer them. I'm sorry but I do not have a clue what you are talking about.

OK.

Thank you for trying to understand my post.

Leandros

 

[ttn]

oh and concerning my experiment, what do you consider region 1 to be? region 2? and region 3?

1 and 2 are the past light cones of the two measurement events.

what is considered a "full specification"?

That's a very important question. The answer is: how the heck should anybody know, a priori? But the point is a *theory* provides some definite candidate for what a "full specification" is supposed to consist of. For example, if we're talking about Maxwell electrodynamics, a full specification means: the positions and velocities of all charged particles plus the E and B field configurations. To posit a theory is to posit some definite proposal for what a full specification of beables would mean. And once you do that, you can ask: does this theory respect Bell Locality? This is why the criterion of Bell Locality applies primarily to theories. If all you have are, say, the results of some experiments, there is no way to answer the question "Was Bell Locality respected?" But, using the 2-part argument that Bell gave, it is, in the end, possible to conclude that no Bell Local theory can agree with what we know about the correlations from experiment. So therefore nature doesn't respect Bell Locality (which of course leaves open exactly what the right theory is -- we know the right theory will not be Bell Local, but that's about it).

 

[alfredblase]

you havn't specified region 3 :P =)

 

[ttn]

you havn't specified region 3 :P =)

Look at the other figure in the paper. The slice across spacetime (after the particles are emitted, but before the measurements are made) plays the role of "region 3" for *both* detection events. So the point is, you need (i.e., "lambda" is) a complete specification of the state of the particles on this slice, i.e., at this time. According to OQM that is just the wave function. According to some kind of hidden variable theory, it would be something else or something more.

I have class in 5 seconds, so to be continued later if you're still confused...

 

[alfredblase]

reply to ttn post 60:

An example he gives somewhere: in Maxwellian electrodynamics, the E and B fields (and functionals of them) are beables, while the potentials V and A are not. That is why nobody thinks locality is violated when, in Coulomb gauge, the potential V at some point changes instantaneously when a distant charge is moved. By contrast, if the fields E or B were to change instantaneously, that *would* be a problematic kind of nonlocality since those fields are supposed to represent "beables".

ok so a beable is a description of a variable that does not need a field from a source different to its own in order to be defined. do you agree?

 

[vanesch]

QM tells us that spin is a non-commuting observable. This tells us that until we measure the spin of one of the entangled electrons, in a particular direction, we are violating local realism.

Ok, so you were NOT talking, finally, about the experimental results (which were identical for the balls and the electrons). You are talking about the inner gears and wheels of the FORMALISM of quantum theory, and you are looking at how the inner wheels and gears of the formalism run, and from their appearance, you conclude about a non-locality.

But then you have to know that how exactly these inner wheels and gears are seen to correspond to the real world, is exactly the interpretation of the theory!

Now, I tell you, that in the MWI interpretational frame, there is NO communication (as a FTL interaction) between the different parts of the photon pair upon measurement.

I also agree that in the "standard" (von Neuman - Copenhagen) view, there IS a clear FTL mechanism, which is the projection postulate, which is applied when the measurement is performed. However, as the wavefunction is not seen as representing something physical in this picture, it is not really clear if this means that there is an FTL ACTION in nature. I would like to point out that the MWI view does away with this projection postulate as a physical operation.

Finally, Bohm's theory (which is empirically equivalent to QM) gives you a clear, explicit FTL interaction, by the quantum potential.

Meaning that we are violating the assumption that all properties are defined in reality as being something in particular (up or down in this case), even if a particular properety has never been measured.

But these are statements which only make sense within the Copenhagen interpretation of QM, and NOT in the MWI view.

well anyway we need not concern ourselves with balls because there is no QM wavefunction defining the colour of billiard sized balls, and so since we are performing this experiment to test the validity of QM we need not discuss balls.

Again, in the MWI view, billiard size balls have just as well a wavefunction description as electrons. In the Copenhagen view, on the other hand, the macroscopic world does NOT have a quantum description.

So, you see, when looking at the inner gears and wheels of the quantum formalism, it depends on HOW you look upon it. All you've been arguing was by implicitly taking on the Copenhagen view. There, of course, you cannot "prove" that the machinery is not doing anything FTL, because the basic operation of projection after measurement is an FTL operation!
And in MWI, on the other hand, there is NEVER an FTL operation. Nevertheless, they both share the same unitary quantum theory.

 

[ttn]

ok so a beable is a description of a variable that does not need a field from a source different to its own in order to be defined. do you agree?

The beables of a theory are those variables (in it) which we are supposed to take as descriptions of really-existing entities. Or, for short: the beables are whatever some theory says is *real*.

For Newtonian mechanics, massive particles and the forces they exert on each other are the "beables." For Maxwell's theory, the electric and magnetic fields are the "beables." For Bohmian quantum theory, the wave function and the particle positions are the "beables." For orthodox QM, the wave function is the only "beable." Maybe that helps?

If you can manage to get ahold of a copy of Bell's book ("Speakable and Unspeakable") you really really really really should. He is a breath of fresh air, and if you have any interest in quantum physics at all, there is literally no other author which it is so mandatory to have read! :!) In honor of Valentine's Day I'll state it here publicly: I love Bell!

 

[alfredblase]

hmmm so all wavefunctions in QM are beables?

 

[alfredblase]

replyto vanesch post 67:

and from their appearance, you conclude about a non-locality.

I have at no point talked about locality in any of my arguments. What is your definition of locality please?

 

[ttn]

hmmm so all wavefunctions in QM are beables?

That's the only way I know of to make sense of Bohr's claim that QM was "complete." I'm sure there are some people who would deny that, for OQM, the wf is a beable. They'd say: no, it's just some meaningless abstract thing in our heads that we use to calculate probabilities. The problem is, they then don't have a theory at all -- just some meaningless abstract formal rules in their heads that refer *only* to measurement outcomes. And then any questions like "Is the theory complete?" or "Is the theory local?" become meaningless. If your "theory" doesn't actually say anything about anything, there's no answer to these questions because, really, you don't have a theory.

So, in my opinion, yes, wavefunctions in QM are beables for QM. And that is precisely why OQM violates Bell Locality: what you decide to do over here can affect what is real (the wf) over there, and the effect is instantaneous. Nonlocal action at a distance. This repulsed Einstein, who thus argued that we ought to reject the completeness doctrine and construct a hidden variable theory to replace/supplement OQM.

 

[ttn]

I have at no point talked about locality in any of my arguments. What is your definition of locality please?

Maybe vanesch wants to answer, but... you have too talked about locality! You talked all about violating (or not) the "speed of light limit." Well, that's all "locality" means. A local theory is one that doesn't violate relativity's speed of light limit. The problem is: what exactly is this limit a limit *on*? Do we require no particles moving faster than light? No information transfer faster than light? No signalling faster than light? No causal influences faster than light? Or what?

I already noted/explained Bell Locality as one particular answer here. Bell Locality means "no causal influences faster than light." What more are you looking for from poor old vanesch?

 

[alfredblase]

reply to ttn post 71:

this is what i suspected. It seems that beable is not a clearly and well defined word... And that therefore neither is Bell Locality. Indeed I have read this elsewhere; I have even read that no physical significance is attached to Bell Locality...

that is why I have avoided using the word "locality" because it was never clear.

So it seems you cannot definitely and inequivocably (not sure if i spelt that right :P, I am a bit tipsy at the mo :P ) define what you mean by locality. Therefore I will not consider that any post so far has demonstrated that QM doesn't violate something that really shouldn't be violated. I hesitate to say the speed of light limit now since bizzarely this doesn't seem to be a problem for most of you, even when arguably reality is changing; since I'm tipsy and reckless right now I'm going to put my chips on the table, (not literarly [i'm sure i spelt THAT wrong]; I am actually eating an omellete) and state that if one adopts a QM description of reality in EPR type experiments, causality is violated... (i'm a bit scared of the responses I might get after this post... :s heheh :P )

 

[vanesch]

this is what i suspected. It seems that beable is not a clearly and well defined word...

It is a clearly defined concept, but is dependent on the particular interpretation. The interpretation is exactly that: saying which elements of a formalism of a physical theory have ontological existence (= are beables).

And that therefore neither is Bell Locality. Indeed I have read this elsewhere; I have even read that no physical significance is attached to Bell Locality...

Bell locality is a clear concept, because it deals with OUTCOMES of experiment (or the empirical predictions of the outcomes of experiment of a physical theory). As such, quantum theory, and all other theories that are empirically equivalent to it, are Bell-non-local. They violate the conditions which define Bell locality. Bell locality is independent of any interpretation of the formalism, because it deals only with experimental outcomes.
Even without any theory, a list of observations can be judged to be Bell local or not.

So it seems you cannot definitely and inequivocably (not sure if i spelt that right :P, I am a bit tipsy at the mo :P ) define what you mean by locality. Therefore I will not consider that any post so far has demonstrated that QM doesn't violate something that really shouldn't be violated.

It's difficult to give a proof of a statement of which you yourself claim that it cannot be defined correctly

What's the relationship between locality and speed of light ? Locality means, essentially, that "things happening at an event (x,y,z,t)" should only depend directly on all beables that are related to the event (x,y,z,t), and not to any other event (x',y',z',t) (same t). As such, locality is "beable-dependent" - it is dependent on the interpretation.
However, if there is no upper limit to the speed of anything, then it doesn't make sense to say that the event at (x,y,z,t) did depend on the a beable at (x',y',z',t) because there might be a small error on the last t, and with high enough speed, this can arrive at (x,y,z,t). So the concept of locality would depend upon an infinite precision of the time variable.
However, with a finite speed limit, it DOES make sense (even with finite measurement errors on x,y,z,t) to say that something happening on a beable at (x,y,z,t) should only depend upon other beables in its neighbourhood, and hence should NOT depend upon the beables at (x',y',z',t) if (x',y',z') is spatially remote enough from (x,y,z).

So it is thanks to the speed of light limit for beables, that the locality concept has ueberhaupt a meaning.

But you see that it also depends on what is taken to be a beable (= what is taken to be ontologically there). If you assign "beable" status to measurement results, then locality implies Bell locality. If you DON'T assign ontological status to measurement results (such as MWI does), then Bell locality has nothing to say about locality (beable locality). It is of course objectionable to NOT assign beable-status to measurement outcomes - this is only possible if outcomes are an effect of the relationship between observer and ontology. Many people object to this, understandably, and hence do not consider MWI.

Now, if your ontology has to have any sense what so ever, then SIGNALS should have some or other beable status. So a theory that does not satisfy SIGNAL LOCALITY will have a hard time having "beable"-locality. Signal non-locality leads to paradoxes in relativity.
Signal locality is ANOTHER condition on experimental outcomes (less severe than Bell locality). Quantum theory (and empirically equivalent theories) are signal-local (that was my proof with the reduced density matrix).
As such, the gate is still OPEN for (beable) locality.

 

[ttn]

Bell locality is a clear concept, because it deals with OUTCOMES of experiment (or the empirical predictions of the outcomes of experiment of a physical theory). As such, quantum theory, and all other theories that are empirically equivalent to it, are Bell-non-local. They violate the conditions which define Bell locality. Bell locality is independent of any interpretation of the formalism, because it deals only with experimental outcomes.
Even without any theory, a list of observations can be judged to be Bell local or not.

vanesch, perhaps you are also tipsy? This is completely wrong.

Bell Locality does not pertain exclusively to experimental outcomes. It is a statement about the probabilities for such outcomes *as assigned by some particular theory*. You *can't* just look at some outcomes and say yes/no Bell Locality was/wasn't violated.

Here's a simple example. Say there's a game where you put two balls into two little boxes so you can't see the balls' colors. Then Alice and Bob each carry a box with a ball to some distant location. Then they simultaneously open them and observe the color. And say that they always see opposite colors: whenever Alice sees green Bob sees red, and vice versa. OK? That's what we know from observation. Is Bell Locality respected? It depends:

Theory1: Balls are neither green nor red until someone looks at them; they're, say, grey. Whoever (Alice or Bob) opens their box first (relative to some preferred frame) *causes* their own ball to switch from grey to one or the other of the observed colors, with 50/50 probability each. And this same act of observation also causes the *distant* ball to pick a definite color which is always opposite to her own. So: alice looks in her box, which causes her initially grey particle to turn green and also causes bob's distant as-yet-unobserved particle to turn red. Or maybe the other way round. Anyway, clearly this theory can account for the observations described above, yes?

Theory2: Balls are either green or red regardless of whether anyone has looked. Either a red ball goes into Alice's box and a green into Bob's, or vice versa, and nothing funny is going on at all when they carry the boxes apart and then look in them. Alice sees green if/when her particle has been green all along, etc. So the correlations are explained by the fact that there is always one green and one red put into the boxes. OK? So this theory too can explain the observations perfectly well.

Here's the rub: Theory1 violates Bell Locality, while Theory2 respects it. So it is just wrong to say that you can tell from the predictions alone whether Bell Locality is violated. Bell Locality is not about the outcomes, it's about the theories which predict those outcomes. It's particular theories which are or are not Bell Local.

That said, I agree with you that any theory making the same prediction as QM will not be Bell Local. But I think it's very misleading to argue for this claim the way you did.

 

[ttn]

Now, if your ontology has to have any sense what so ever, then SIGNALS should have some or other beable status. So a theory that does not satisfy SIGNAL LOCALITY will have a hard time having "beable"-locality. Signal non-locality leads to paradoxes in relativity.
Signal locality is ANOTHER condition on experimental outcomes (less severe than Bell locality). Quantum theory (and empirically equivalent theories) are signal-local (that was my proof with the reduced density matrix).
As such, the gate is still OPEN for (beable) locality.

This is a cheap shot that I don't intend all that seriously, but...

Does MWI respect signal locality? Or rather, does the claim that MWI is signal local have any meaning? The reason I wonder is that, according to MWI, all of the other "people" in the universe are actually mindless hulks. So if I transmit a signal to them, it is never really consciously received, i.e., it wasn't really a signal. So not only is superluminal signalling impossible, all signalling is impossible, and the idea of signal locality has no meaning.

If a tree falls in the forest and crushes a mindless hulk, ...?

 

[alfredblase]

I'm sorry but I can't test more than one theory at a time. I thought QM was a widely accepted theory and that there was only one QM theory! I thought it was ironclad; the most proven theory in the history of science... please tell me which theory I was thinking of so that I can test it... please? But don't tell me there are a few of them and so making my test impossible.

 

[Sherlock]

sherlock you also refer to locality as being the violated principle. Is your definition of locality that given by Bell Locality or do you use another one? If another please give this definition.

I think I said something like that you can infer that quantum theory and nature are nonlocal depending on how you interpret the theory.

Bell Locality is a straightforward mathematical criterion for evaluating whether or not a theory is locally causal. Bell Locality requires that the sort of nonseparable quantum states that are used in Bell tests be factorizable. But in quantum theory these states are not factorizable, so the theory is evaluated as not being a locally causal theory according to the Bell Locality test.

This doesn't automatically mean that quantum theory is a nonlocally causal theory. Depending on how you interpret the formalism of the theory, it might be said that it's not a causal theory --- in which case it's also not a nonlocally causal theory.

There are parts of the qm algorithm that are deterministic, but their relationship to (an underlying) reality isn't quite clear.

If the whole of quantum theory is interpreted as being an acausal theory (which is the standard interpretation), then an evaluation of some part(s) of its formal structure wrt the Bell Locality criterion is meaningless.

One way of reading orthodox QM is to take wave functions as beables. If we do that, the theory violates Bell Locality. On the other hand, if we don't do that -- and hence say that the theory has no beables at all, then the question of the theory's locality becomes meaningless because, really, it isn't even a theory unless it asserts *something* about the way the world works.

The only thing that the theory is asserting unambiguously about reality is the calculation of probability distributions of instrumental output. That's what it does, that's what it's for. It's a theory about quantum phenomena. Quantum phenomena are instrumental phenomena. Even if that's all the formalism is about, it's nevertheless telling us *something* about the way the world works.

I like to think that it's telling us more than that --- something about an underlying reality. But exactly what it's telling us about an underlying reality is still a matter of speculation.

 

[alfredblase]

A theory must be testable on all counts.

It seems you all agree that whether or not QM violates causality depends upon interpretation.

But you must understand that an interpratation that does not violate causality is essential to QM, and that it must be found before we can accept QM as a a physical theory.

 

[vanesch]

vanesch, perhaps you are also tipsy? This is completely wrong.

Bell Locality does not pertain exclusively to experimental outcomes. It is a statement about the probabilities for such outcomes *as assigned by some particular theory*. You *can't* just look at some outcomes and say yes/no Bell Locality was/wasn't violated.

I thought that Bell locality came down to requiring that all observed (or empirically predicted) correlations respected all thinkable Bell inequalities. In other words, that they CAN be produced by a theory that you call "Bell local".

Your theory 1 is a theory that is Bell local, but is not beable local (and as such a very strange theory!), that is, the outcomes do not violate the Bell inequalities (and hence CAN be generated by a theory that is, according to your wordings, Bell local). However, the inner gears and workings of the theory do involve non-local actions, but which are such, that this compensates entirely any potential signal or Bell non-locality (my definition).

If Bell locality were a property of a theoretical construction, one could not test it in a lab! I think you're mixing up Bell locality and "beable locality".
(and then, it is maybe just a matter of semantics, but I prefer to keep Bell locality for that typical requirement of respecting Bell inequalities, something that is entirely independent of any theory behind it).

 

[vanesch]

But you must understand that an interpratation that does not violate causality is essential to QM, and that it must be found before we can accept QM as a a physical theory.

Hence my insistance upon MWI as the preferred view on QM...

 

[vanesch]

The reason I wonder is that, according to MWI, all of the other "people" in the universe are actually mindless hulks. So if I transmit a signal to them, it is never really consciously received, i.e., it wasn't really a signal.

The "mindless hulk" has to come to you with his notebook. If you notice a correlation that depends upon your CHOICE of experiment (and not on your OUTCOME) then you know you signalled something.
It is like an EPR experiment, except that you now look for the correlation between your SETTING, and his results, and not your RESULTS and his RESULTS.

 

[alfredblase]

you forgot to quote my first sentence vanesch...

"a theory must be testable on all counts"

since MWI predicts many worlds and since these many worlds can never be observed, MWQM is not a physical theory either.

 

[ttn]

I thought that Bell locality came down to requiring that all observed (or empirically predicted) correlations respected all thinkable Bell inequalities.

No, Bell Locality doesn't just mean "respects Bell inequalities." See the discussion in section 2 of quant-ph/0601205 for more info about Bell Locality.

The problem is, to derive the inequality you have to assume that there exist "hidden variables" of a certain type which determine the outcomes. So what kind of theory is required to respect those inequalities? Only that type of hidden variable theory. So it wouldn't even make sense to say something like "OQM is nonlocal because it violates a Bell inequality." It does violate the inequality, yes, but that doesn't prove squat about whether it's local or not, because it isn't the kind of theory (namely, the kind of hidden variable theory) to which the inequalities are supposed to apply.

Your theory 1 is a theory that is Bell local, but is not beable local (and as such a very strange theory!), that is, the outcomes do not violate the Bell inequalities (and hence CAN be generated by a theory that is, according to your wordings, Bell local). However, the inner gears and workings of the theory do involve non-local actions, but which are such, that this compensates entirely any potential signal or Bell non-locality (my definition).

Well maybe we're just using words differently. I have no idea what you mean by "beable local." But what *I* mean by "Bell Local" is what Bell meant, as explained in several of his papers and in the paper I mentioned above. And my theory1 from that previous post is definitely not Bell Local -- even though, as you point out, the theory doesn't predict any violation of bell inequalities.

Your talk of the "inner gears and workings" is more along the lines of what I (and Bell) mean by Bell Locality.

By the way, my toy theory1 is, in all relevant respects, exactly like orthodox QM. Theory1 and OQM violate Bell Locality for exactly the same reasons (and are signal local for exactly the same reasons too).

If Bell locality were a property of a theoretical construction, one could not test it in a lab! I think you're mixing up Bell locality and "beable locality". (and then, it is maybe just a matter of semantics, but I prefer to keep Bell locality for that typical requirement of respecting Bell inequalities, something that is entirely independent of any theory behind it).

Can you define "beable locality"?

BTW, you're absolutely right that you can't *directly* test Bell Locality in a lab. That's why Bell's two part argument is so important. The first part shows that the only way to Bell-Locally explain a certain set of the observed correlations is for certain kinds of hidden variables to exist. Then the second part (the derivation of the inequality) shows that that kind of hidden variable theory can't account for some of the other observed correlations. So it's only at the end of that whole chain of reasoning that one is entitled to conclude that Bell Locality fails (in the sense that no Bell Local theory can be consistent with the observed facts).

 

[vanesch]

you forgot to quote my first sentence vanesch...

"a theory must be testable on all counts"

since MWI predicts many worlds and since these many worlds can never be observed, MWQM is not a physical theory either.

The effects of the "worlds" can be observed, in principle, by quantum interference experiments. Now, of course, from a certain entanglement on, the experiment is not feasible anymore. I could devise an "in principle" experiment for each statement where one says that a "parallel world" (a term in the wavefunction) has disappeared. Of course it would be technically totally unfeasable: it would be quantum erasure experiments on the scale of macroscopic items.

Also the extreme empiricist idea that a theory must be testable on all counts would simply imply that every theory which is more than a simple catalog of past observations would not satisfy the requirement. Try to account testability of the concept of "force" in Newtonian physics...

 

[vanesch]

So it wouldn't even make sense to say something like "OQM is nonlocal because it violates a Bell inequality." It does violate the inequality, yes, but that doesn't prove squat about whether it's local or not, because it isn't the kind of theory (namely, the kind of hidden variable theory) to which the inequalities are supposed to apply.

Right. In the Copenhagen view, where the wavefunction does not represent any physical quantity (where it is even left open as to whether nature exists on the microscopic level ) it would then be silly to say that the theory is not "Bell local" according to this definition.

Well maybe we're just using words differently. I have no idea what you mean by "beable local." But what *I* mean by "Bell Local" is what Bell meant, as explained in several of his papers and in the paper I mentioned above.

I admit not really knowing what Bell meant with the word (I even think he simply called it "local", no ?), and I have to say that historical prerogatives are not my strongest point. The essence of Bell's work is, I'd say, the derivation of his inequalities, and the particularity of quantum theory is that it violates _IN ITS PREDICTIONS OF OBSERVABLE MEASUREMENTS_ these inequalities. So I'd say that THIS property is what captures most what Bell meant with his concept of "locality".

And my theory1 from that previous post is definitely not Bell Local -- even though, as you point out, the theory doesn't predict any violation of bell inequalities.

Your talk of the "inner gears and workings" is more along the lines of what I (and Bell) mean by Bell Locality.

By the way, my toy theory1 is, in all relevant respects, exactly like orthodox QM. Theory1 and OQM violate Bell Locality for exactly the same reasons (and are signal local for exactly the same reasons too).

"Orthodox" quantum theory does not even assign *ANY* inner gears and workings to its formalism (according to the Copenhagen view - maybe less according to the von Neumann view), so saying that there is a violation of "Bell locality" according to gears and wheels would mean nothing in that respect.

Can you define "beable locality"?

I think it is exactly what you mean by "Bell locality": that the beables (the inner parts of the formalism that are supposed to correspond to something real out there) have only interactions (changes in their nature dictated by) with things they are in local spatial contact with.

BTW, you're absolutely right that you can't *directly* test Bell Locality in a lab. That's why Bell's two part argument is so important. The first part shows that the only way to Bell-Locally explain a certain set of the observed correlations is for certain kinds of hidden variables to exist. Then the second part (the derivation of the inequality) shows that that kind of hidden variable theory can't account for some of the other observed correlations. So it's only at the end of that whole chain of reasoning that one is entitled to conclude that Bell Locality fails (in the sense that no Bell Local theory can be consistent with the observed facts).

That's why I found it more logical to call THIS aspect, Bell locality. You may be right in the historical definition, I don't really know.

 

[ttn]

Right. In the Copenhagen view, where the wavefunction does not represent any physical quantity (where it is even left open as to whether nature exists on the microscopic level ) it would then be silly to say that the theory is not "Bell local" according to this definition.

Well, it's true that Bohr once said "there is no quantum world" or whatever. But he (and generations of followers) also insisted that the wave function alone provides a *complete* description of... [something]. I can only assume that something is the relevant aspect of the quantum world. What else could the completeness doctrine mean? The whole anti-hidden-variables attitude of the orthodoxy is precisely against the idea of *supplementing* the wf's description of the quantum world with something else.

So, I guess I think you shouldn't accept so easily something that is often said but is not actually accepted in practice. Plus, as I've said before, if there is some interpretation in which the wf does not refer to anything actually real (any gears and wheels) then that interpretation is not a theory, and there is therefore no way to apply terms like complete/incomplete/local/nonlocal to it. Those terms refer to gears and wheels, period. So if the copenhagen/orthodox people don't believe their theory provides any gears and wheels, what the heck are they talking about when they keep on insisting decade after decade that OQM is both complete and local?

The essence of Bell's work is, I'd say, the derivation of his inequalities, and the particularity of quantum theory is that it violates _IN ITS PREDICTIONS OF OBSERVABLE MEASUREMENTS_ these inequalities. So I'd say that THIS property is what captures most what Bell meant with his concept of "locality".

Yes, Bell's main achievement is indeed the derivation of his inequalities. But he also understood very clearly that the inequality is only the second part of a two part argument. See, for example, the section "QM is not locally causal" in la nouvelle cuisine (I think), where he notes that EPR pointed out years ago that QM (if taken as complete) is not local. Bell considered the EPR argument to be the first half of the argument. OQM is not local and (as EPR suggested) we need a certain kind of hv's to reinstate locality. Then enter Bell's inequality, which shows that no hv theory of that type can agree with experiment. Conclusion: the locality criterion (on which was based our belief in the kind of hv theory that the inequality further constrains) cannot be maintained in the face of experiment.

So... the point is, the first half of this argument is *crucial*. Without it, your left with the muddle-headed view that is so widely held today: Bell proved not that there is any problem with QM itself, but only with attempts to add hidden variables -- i.e., Bell proved that Bohr was right and Einstein was wrong. This view is complete BS.

And it is based, in part, on the confusion between Bell Locality (which is a basic requirement for theories) and the Bell Inequalities (which is merely a consequence for a certain class of Bell Local theories).

 

[selfAdjoint]

But he (and generations of followers) also insisted that the wave function alone provides a *complete* description of... [something]. I can only assume that something is the relevant aspect of the quantum world.

No in Copenhagen it's a complete description of anything we might find if we did an experiment. So it's not ontological, at least in the traditional sense, but kind of meta-epistomological. It's not "knowledge" exactly, because it's complete, and knowledge for Bohr can only be knowledge of the familiar macroworld; it's the prior necessity for knowledge.

If this sounds like pop Kant, you're right. All that generation of German-influenced physicists studied Kant as teenagers; this was the peak of the German educational tradition, before the deluge. It informed their thinking.

 

[ttn]

No in Copenhagen it's a complete description of anything we might find if we did an experiment. So it's not ontological, at least in the traditional sense, but kind of meta-epistomological. It's not "knowledge" exactly, because it's complete, and knowledge for Bohr can only be knowledge of the familiar macroworld; it's the prior necessity for knowledge.

Perhaps you are right about what Bohr or some other particular person actually intended. But to me it's clear that this philosophy is a muddled and contradictory hash. It doesn't actually form a coherent position. So I am more concerned with finding some definite (if wrong) position that at least captures some aspect of what Bohr was aiming for.

Now, specifically, what you say in the first sentence doesn't make sense. The wave function in QM is *not* a description of measurement outcomes. It just isn't. That isn't the role it plays in the theory. Rather, you *use* the wave function to compute the possible measurement outcomes. So a statement like "the wave function is a complete description of possible measurement outcomes (or their probabilities or whatever)" is incoherent. Or, if coherent, it has nothing to do with how orthodox QM actually works as a theory. So I'm unwilling to accept that position as what Copenhagen really means.

Don't get me wrong. I don't necessarily insist that, according to Copenhagen, the wf is ontological. My claim is weaker: *if* Copenhagen is interpreted as positing any ontology at all, it can *only* be in terms of the wave function. Or put it this way: forgetting about traditional terminology, there exists a reasonably coherent interpretation which takes the wave function as a literally true and complete description of the actual state of quantum systems, i.e., which takes the wf as a "beable". And that interpretation (whatever you want to call it) violates Bell Locality.

Now there is of course the alternative of treating the wave function as a mere calculation instrument that has nothing whatever to do with ontology. Well, the "completeness" claim is then totally meaningless as I've said before, so I don't know why anyone would want to identify this interpretation with Copenhagen. But whatever; leaving aside all the mere terminology issues, let's just look at this interpretation. Well, since there is no candidate for "beable status" *other* than the wf, it is clear that this interpretation provides no ontology at all. It says literally nothing about any reality "behind" measurement outcomes. It is exclusively a formal recipe for calculating probabilities of certain outcomes. There are then several points to make about this. First, it is *meaningless* to say that this "theory" is local or nonlocal. Those terms denote certain features of the ontology posited by a theory, namely, whether or not it includes faster-than-light causation; but this "theory" posits no ontology, so there is simply no way to apply such terms to it. Second, I think it is a stretch of terminology to even call this a theory. Blind formalisms for predicting measurement outcomes are precisely what one resorts to in the *absense* of a theory, when one has no idea whatsoever what's going on "behind" the measurement processes to produce outcomes. Of course, advocates of this approach will resist the claim that their approach isn't a theory, because it is usually part of their approach that we shouldn't look for a theory (in the traditional sense). But that simply unveils how philosophical and stupid this view is. It is one thing to play it safe and not commit to any definite ontology when there is not yet suffiicient evidence; it is another entirely to enshrine that normally-temporary state of ignorance and insist not only that we don't have a clear physical picture of what's going on, but that (paraphrasing Bell) "it is immoral to look for one." This is made even more ridiculous because we're talking after all about physicists! Imagine it! *Physicists* saying, in effect, we don't now have -- and *never should look for* -- a coherent physical picture to go along with the calculation formalism. That is so preposterous it shouldn't be taken seriously by anyone who calls themself a physicist!

...and it is precisely why I prefer to be generous and interpret Copenhagen as defining the wf as a complete ontology (for, at least, quantum systems).

If this sounds like pop Kant, you're right. All that generation of German-influenced physicists studied Kant as teenagers; this was the peak of the German educational tradition, before the deluge. It informed their thinking.

I agree 100%. But I would add that Kant is anti-scientific trash. To say that orthodox QM is based on Kant (and his stupid phenomenal/noumenal distinction, etc...) is to confess that the theory is completely arbitrary (from a *physics* standpoint) and we shouldn't bat an eye if we're going to reject it -- or, as I prefer, give it as much benefit of the doubt as possible by tweaking it into a coherent position, and then rejecting that for sound scientific reasons (such as that the resulting theory suffers from the measurement problem, unlike alternatives such as Bohmian Mechanics).

 

[selfAdjoint]

Now, specifically, what you say in the first sentence doesn't make sense. The wave function in QM is *not* a description of measurement outcomes. It just isn't. That isn't the role it plays in the theory. Rather, you *use* the wave function to compute the possible measurement outcomes. So a statement like "the wave function is a complete description of possible measurement outcomes (or their probabilities or whatever)" is incoherent. Or, if coherent, it has nothing to do with how orthodox QM actually works as a theory. So I'm unwilling to accept that position as what Copenhagen really means.

You keep saying "not coherent" but you don't justify it. Consider the giant's line in "Jack and the Beanstalk": "Fee Fi Fo Fum! I smell the blood of an Englishman! Be he alive or be he dead, I'll grind his bones to make my bread!". That is a complete description of the hypothetical life-states of a hypothetical Englishman (cf. cat).

The wavefunction's eigenvalues when acted on by the operator representing a particular experiment give a complete description of the possible outcomes of the experiment. Complete in the sense that if you actualize the experiment correctly, you WILL observe one of the indicated outcomes. The wave function it self is even more complete in that it contains the partial information suitable to determine the possible outcomes of any hypothetical (properly set up) experiment.

That seems coherent enough to me. I may or may not agree with it, but coherent? Yes.

 

[ttn]

You keep saying "not coherent" but you don't justify it. Consider the giant's line in "Jack and the Beanstalk": "Fee Fi Fo Fum! I smell the blood of an Englishman! Be he alive or be he dead, I'll grind his bones to make my bread!". That is a complete description of the hypothetical life-states of a hypothetical Englishman (cf. cat).

You mean: he's either alive or dead? This makes me think you don't understand QM very well. Take a nice 2-state quantum analogue: a measurement of the z-axis-spin of some electron will either result in "up" or "down." So "up" and "down" are the only two possible states? Not according to QM! "up" and "down" merely form a *basis* for a whole infinity of possible states, all of which are surely supposed to be in some sense *different* according to the completeness doctrine, yes? What you're saying (if I understand correctly) makes it sound like the completeness doctrine (combined with a purely epistemic attitude toward the wf) implies the old "ignorance interpretation" -- namely, what it means to be in a superposition is, really, to be in one or the other of the states but we're not sure which. But to say that is precisely to confess that the wave function is *not* a complete description of the real state!

The wavefunction's eigenvalues when acted on by the operator representing a particular experiment give a complete description of the possible outcomes of the experiment. Complete in the sense that if you actualize the experiment correctly, you WILL observe one of the indicated outcomes. The wave function it self is even more complete in that it contains the partial information suitable to determine the possible outcomes of any hypothetical (properly set up) experiment.

The wf doesn't have eigenvalues; the operator does.

This is actually an important point. A list of possible measurement outcomes can be produced without specifying the wf. So if that's what you mean by a "complete description" then you don't even need to specify the wave function to have a complete description. Maybe you want to be able to specify not only the possible outcomes, but also the probabilities for each outcome? But then |+x> and |+y> become "the same state" so long as you're about to measure the z-spin. And that again seems to conflict with any rational meaning of completeness.

But let's come to the fundamental: you say that the wf "contains the ... information suitable to determine the possible outcomes..." Look at the word "information". What do you mean by this? What is this "information" information *about*? Is it information about the really-existing quantum system? If so, then either that information is or isn't complete (in the usual ontological sense) and we just have to argue about whether or not there's some good reason to add additional variables. (I will argue that there is a good reason -- namely, to solve the measurement problem.) But if the "information" you speak of is information about something else, you'd better tell me what the something else is.

That seems coherent enough to me. I may or may not agree with it, but coherent? Yes.

Maybe we've just misunderstood each other. I didn't say that the purely epistemic interpretation wasn't coherent. It is. I said that this interpretation rendered the completeness doctrine (as well as claims about the locality of the "theory") incoherent. One is free to deny that one's calculation recipe is telling us anything about the gears and wheels. But then one cannot go on to claim that one's description of the gears and wheels is complete, nor that the gears and wheels don't affect each other superluminally. That's the point.

 

[selfAdjoint]

"up" and "down" merely form a *basis* for a whole infinity of possible states, all of which are surely supposed to be in some sense *different* according to the completeness doctrine, yes?

Yes. And from the point of view of the giant, the hypothetical englishman is in a superposition of the states alive or dead. But the coefficient field in his case (he is a stupid, classical giant) is Z_2 not C.

The wf doesn't have eigenvalues; the operator does.

True, I should have phrased it differently. But the operator's eigenvalues don't have any issue in reality unless it acts on the wavefunction. The Copenhagen view is that the whole operator-wavefunction apparatus is just a formalism for predicting outcomes; the wave function is like a database of hypothetical conditions, and the operator is like a program that reads the database and instantiates them. Neither amounts to anything without the other, but it is meaningful to say that the database contains what any well set-up program will need to instantiate outcomes of a well-prepared experiment.

If you don't like the term information for what the wave function comprises, and want to avoid the weasel word state, I suggest quantum hypothetical.

 

[ttn]

The Copenhagen view is that the whole operator-wavefunction apparatus is just a formalism for predicting outcomes; the wave function is like a database of hypothetical conditions, and the operator is like a program that reads the database and instantiates them. Neither amounts to anything without the other, but it is meaningful to say that the database contains what any well set-up program will need to instantiate outcomes of a well-prepared experiment.

If you don't like the term information for what the wave function comprises, and want to avoid the weasel word state, I suggest quantum hypothetical.

To avoid historical confusion over terminology, let's call the above view the self-adjoint-interpretation. Then let me ask you: according to this interpretation, does the wave function provide a complete description? And then I hope you can clarify: a complete description *of what*? And also this: does the theory respect relativity's prohibition on superluminal causation?

I still don't see how you can address either of these questions unless you accept that the wave function is a description *of* *something*, i.e., unless you accept that the wf is supposed to be a beable. But since it's you who is apparently making this claim, I'll let you clarify things (i.e., discharge the burden of proof).

BTW, here's why the burden of proof is on you: if you interpret the wf as a beable (and as the only beable, i.e., as providing a complete specification of the real state of things) then there is an absolutely clear meaning to the "completeness" and "locality" claims (though, as proved by EPR, both claims can't be simultaneously true!). I don't see how you can say the wf provides a complete description, and also deny that it describes anything. Same sort of problem with the locality question. But prove me wrong if you can!

 

[alfredblase]

the extreme empiricist idea that a theory must be testable on all counts would simply imply that every theory which is more than a simple catalog of past observations would not satisfy the requirement. Try to account testability of the concept of "force" in Newtonian physics...

[the above quote was from vanesch]

Newton's second law states that:

The rate of change of momentum of a body is equal to the resultant force acting on the body and is in the same direction.

You prove this statement every day vanesch [within well defined limits of course]. Newtonian physics makes no other definition of force.

Next point:

The effects of the "worlds" can be observed, in principle, by quantum interference experiments.

No no no. "in principle" is certainly not good enough. You must first define the term "world" and all other terms involved in this definition. Then you must prove inequivocably that there are "many" of them. Either that or I denounce you as a crackpot physicist for claiming that MWQM [meaning "Many Worlds Quantum Theory" ] is an undeniable physical theory.

Ok; I'm going to going to present my argument fully as I can:

1. Causality must hold in all physical theories [I can provide arguments for this if needed]

2. In view of point 1: QM must have a provable physical interpretation that ensures causality is not violated in order to be accepted as a physical theory.

I ask for a brief description [if there is one] of such an interpretation please

 

[DrChinese]

So... the point is, the first half of this argument is *crucial*. Without it, your left with the muddle-headed view that is so widely held today: Bell proved not that there is any problem with QM itself, but only with attempts to add hidden variables -- i.e., Bell proved that Bohr was right and Einstein was wrong. This view is complete BS.

Well, you have an opinion on that... but the facts are a bit short and hinge on your interpretation of EPR and Bell (that not all of us agree upon).

First, let's agree that there is nothing in particular "wrong" with the current oQM formalism.

Second, Einstein - EPR - *was* wrong - at least in some ways. EPR absolutely felt that experiments would show that the Heisenberg Uncertainty Relations could be beaten. They never knew about Bell or Aspect. They contended that if oQM were complete, that there could not be simultaneous reality to non-commuting observables - a position they considered unreasonable. Of course, they too recognized that if locality were violated, this would provide an escape route. But that too was considered at least as unreasonable.

Third, I would cast doubt that Bohr's position that oQM is complete has not been successfully defended. And of course, by completeness I mean that the WF is complete.

I would agree with you that there is a sense in which oQM is non-local, that being the collapse of the WF. (And I don't mean to step on MWI in that statement because that is not my intention.) However, I do not agree that causality is violated by such non-local collapse; and we already agree there is no non-local signal mechanisms. So the only real disagreement is whether we now call oQM non-local. I don't (because local causality is not violated in the sense that Alice's choice of setting does not affect Bob's result); you do (because WF collapse is FTL).

 

[DrChinese]

1. Causality must hold in all physical theories [I can provide arguments for this if needed]

2. In view of point 1: QM must have a provable physical interpretation that ensures causality is not violated in order to be accepted as a physical theory.

I ask for a brief description [if there is one] of such an interpretation please

1. Causality is not a requirement of all physical theories. QM is a counter-example to that idea. See for instance a paper I wrote: Determinism Refuted[/url].

2. QM *is* generally accepted, and subsequent to Bell I doubt it is considered causal universally.

A theory could consist of voodoo if it worked - and by "worked" I mean: it is useful. Please do not confuse theories with "the truth".

 

[vanesch]

[the above quote was from vanesch]

Newton's second law states that:

The rate of change of momentum of a body is equal to the resultant force acting on the body and is in the same direction.

You prove this statement every day vanesch [within well defined limits of course]. Newtonian physics makes no other definition of force.

The concepts of "momentum", "force" and so on are only helpful quantities in order to EXPLAIN, conceptually, observations (which are usually visual impressions of pointers on a dial, or spots on a photograph or whatever tool you decide to use as experimental apparatus). You cannot observe DIRECTLY a force, you can only observe its pretended consequences. As such it is an organizing principle of your observations.

The "many worlds" (in other words, the wavefunction!) is exactly that too.

No no no. "in principle" is certainly not good enough. You must first define the term "world" and all other terms involved in this definition.

world = term in the wavefunction, when written in a particular basis (usually the one that corresponds to the Schmidt-decomposition between the Hilbert space of the observer and the rest of the world).

I protest against the rejection of "in principle": it is the essence of any theory, to be able to say what would happen in principle, without limitation by the state of experimental technology (as long as that limitation is also not a matter of principle of course, that's the danger...).

Then you must prove inequivocably that there are "many" of them. Either that or I denounce you as a crackpot physicist for claiming that MWQM [meaning "Many Worlds Quantum Theory" ] is an undeniable physical theory.

You can say the same about Newton, then. Or any other person who has set up a physical theory. You can NEVER PROVE the existence of all the theoretical concepts that appear in the theory, you can only argue about its empirical validity or not. Because if what you claim is right, then it would be sufficient to reformulate a theory in an empirically equivalent one to show the "crackpottishness" of both. Newtonian physics (with forces) can be reformulated as a stationarity principle (Lagrange, ...). So both are clearly crackpottish theories according to your criterium (because in the Lagrangian formulation, no concept of "force" appears explicitly).

 

[vanesch]

I can only assume that something is the relevant aspect of the quantum world. What else could the completeness doctrine mean? The whole anti-hidden-variables attitude of the orthodoxy is precisely against the idea of *supplementing* the wf's description of the quantum world with something else.

I think that what's meant is that the WF contains ALL POSSIBLE information that one could ever extract from the system by any conceivable experiment (in other words, that the statistical predictions by quantum theory of the outcomes of experiment contain already the MAXIMUM amount of information (in the information-theoretic sense) about these outcomes, and that no theoretical refinement ever is going to do any better.

Now, of course I share your problems with this view which vehemently refuses to consider the ontology of the microworld and nevertheless claims to know all about it that can be known, but I think that this IS the view that is proposed in the Copenhagen interpretation.

So, I guess I think you shouldn't accept so easily something that is often said but is not actually accepted in practice. Plus, as I've said before, if there is some interpretation in which the wf does not refer to anything actually real (any gears and wheels) then that interpretation is not a theory, and there is therefore no way to apply terms like complete/incomplete/local/nonlocal to it.

I agree with your statement concerning locality ; however, "completeness" in the above sense would make sense.

I also agree with your claim about the schizophreny of its practicians: when they do physics with the wavefunction (when they write out interaction terms and so on, and say they can neglect certain contributions and so on) I have a hard time imagining that they do not give it some kind of ontological status (I don't know how you devellop an intuition for something to which you assign no ontological status at all).

And it is based, in part, on the confusion between Bell Locality (which is a basic requirement for theories) and the Bell Inequalities (which is merely a consequence for a certain class of Bell Local theories).

Probably, but given that everybody already confuses Bell locality with Bell inequalities, why don't we just tag the word "Bell locality" to just that, and we tag the word beable locality to the "locality of interaction by the beables of the theory".

Or otherwise we call it "Bell inequalities induced locality"...

Now we clearly have that, when measurement outcomes are seen as beables (which they are NOT in the MWI view!), then beable locality is equivalent to Bell locality (and that was in fact Bell's reasoning, right ?). So probably because Bell took this statement as so very obvious (that observations are "real", hence, beables) that he didn't even gave it further thought, he could reason the way he did.

 

[ttn]

I think that what's meant is that the WF contains ALL POSSIBLE information that one could ever extract from the system by any conceivable experiment (in other words, that the statistical predictions by quantum theory of the outcomes of experiment contain already the MAXIMUM amount of information (in the information-theoretic sense) about these outcomes, and that no theoretical refinement ever is going to do any better.

But (to repeat an earlier question) what is this "information" information *about*? There is no such thing as free-floating information that isn't information about something. The very concept "information" is literally meaningless without some object (like other concepts such as "awareness").

As I've said, it is possible to deny any micro-ontology and regard the whole QM formalism as simply being about measurement outcomes and nothing else. But then, as I keep arguing, all talk about "completeness" or "locality" becomes meaningless.

Now, of course I share your problems with this view which vehemently refuses to consider the ontology of the microworld and nevertheless claims to know all about it that can be known, but I think that this IS the view that is proposed in the Copenhagen interpretation.

It's one of the views, but the advocates aren't consistent. They go back and forth between the common-sense ontological interpretation of wf-as-complete, and the completely epistemic version. When they want to rail against Bohmian mechanics, they deride the hidden variables as cumbersome metaphysics (or whatever) and insist that the wf alone provides a complete description of quantum states. Then when they want to avoid the charge that their theory (like Bohm's) is nonlocal, they switch to the epistemic version. Well, we shouldn't let them so easily have it both ways. There *are* two ways to think about it, but they're not the same. Each has a virtue and a vice, and it's just not reasonable to let people fuzz up the issue and pick and choose the virtues from mutually inconsistent theories as it suits them.

I agree with your statement concerning locality ; however, "completeness" in the above sense would make sense.

Only if what the "complete" description is a complete description *of* is measurement outcomes. But (a) this claim doesn't really make any sense and (b) it is not at all the same thing that is claimed or denied in the context of debates about "hidden variables".

I also agree with your claim about the schizophreny of its practicians: when they do physics with the wavefunction (when they write out interaction terms and so on, and say they can neglect certain contributions and so on) I have a hard time imagining that they do not give it some kind of ontological status (I don't know how you devellop an intuition for something to which you assign no ontological status at all).

Right, I totally agree.

Probably, but given that everybody already confuses Bell locality with Bell inequalities, why don't we just tag the word "Bell locality" to just that, and we tag the word beable locality to the "locality of interaction by the beables of the theory".

I personally think this confusion over terminology reflects a much deeper and more important/fundamental confusion over what Bell's theorem proves in the first place. So I think it's worth fighting to clarify this terminology, rather than just accepting the confusion and introducing new terminology.

Now we clearly have that, when measurement outcomes are seen as beables (which they are NOT in the MWI view!),

Aren't they as much beables as anything else in MWI? I mean, the measuring appratuses are made out of electrons and whatnot, and hence described by wave functions. It's just that, usually, the apparatus isn't in a definite pointer state. But the formal entities that MWI uses to refer to the apparatuses (namely, wave functions!) are not only beables -- they're the only (kind of) beables. Well, except for those pesky "consciousness tokens"...

then beable locality is equivalent to Bell locality (and that was in fact Bell's reasoning, right ?). So probably because Bell took this statement as so very obvious (that observations are "real", hence, beables) that he didn't even gave it further thought, he could reason the way he did.

That's right. But I object to your making it sound like it was some kind of dubious, uncareful "leap" to just assume (without "giving it further thought") that pointers actually point. I mean, if you can't believe what you see, how the heck are you going to believe anything? Even your precious quantum formalism is ultimately -- historically -- based on putting together a whole bunch of things that a whole bunch of people literally saw!

 

[Sherlock]

Well, it's true that Bohr once said "there is no quantum world" or whatever. But he (and generations of followers) also insisted that the wave function alone provides a *complete* description of... [something]. I can only assume that something is the relevant aspect of the quantum world. What else could the completeness doctrine mean?

... as I've said before, if there is some interpretation in which the wf does not refer to anything actually real (any gears and wheels) then that interpretation is not a theory, and there is therefore no way to apply terms like complete/incomplete/local/nonlocal to it. Those terms refer to gears and wheels, period. So if the copenhagen/orthodox people don't believe their theory provides any gears and wheels, what the heck are they talking about when they keep on insisting decade after decade that OQM is both complete and local?

If assumptions about quantum theory's relationship with an underlying quantum world are avoided, then the expansion theorem-postulate doesn't say anything about nonlocality. The theory is then interpreted as being acausal, and as such makes no statement about the existence, or not, of nonlocal causality in nature. In this view it isn't a locally causal theory either. So, it wouldn't, strictly speaking, be correct to call it a local theory. If OQM is thought of that way (as a local theory), then I would guess that it just has to do with it not violating the principle of local causality (which it doesn't as long as it's not being taken as mirroring an underlying quantum world).

What they are talking about wrt completeness is that the wavefunction is regarded as a complete description of what can be quantitatively determined about a quantum experimental preparation --- that the instrumental output will correspond to the probabilities assigned by the wavefunction for the setup.

The orthodox interpretation is about what the theory is, not what it might be. The theory is a mathematical scheme that assigns probabilities to qualitative instrumental behavior. Attributing some speculative significance (in terms of a correspondence to an underlying quantum world) to the qm algorithm or any part thereof is beyond the scope of the theory itself (and apparently beyond the scope of physics, at least for the foreseeable future).

So, the only interpretation of quantum theory that is clearly meaningful is the orthodox, probabilistic (or Copenhagen) interpretation. Specifying what quantum theory is known to be about (assigning probabilities to experimental results), while avoiding speculation about the theory's relationship to an underlying reality, doesn't make it any less a physical theory. It just can't necessarily be taken as a description of an underlying reality --- and this is maybe the most confounding way in which quantum theory differs from its classical predecessors.

Your two-part argument for nonlocality in nature, ttn, seems solid enough given the assumption that the mathematical gears and wheels of quantum theory are a 1-1 mapping, or at least in very close approximation to, the relevant (to the experimental results) qualitative aspects of an underlying quantum world. However, it seems just as reasonable to assume that they aren't, but rather are just charting the evolution of the instrumental probabilities. Seen from the latter point of view, the gaps in the quantum theoretical picture aren't surprising and don't imply nonlocal causality.