Commonsense local realism refutes Bell's theorem

Foundations of physics and/or philosophy of physics, and in particular, posts on unresolved or controversial issues

Re: Commonsense local realism refutes Bell's theorem

Postby gill1109 » Wed Jun 11, 2014 8:26 am

Ben6993 wrote:Hi Richard

Thanks for posting your Vaxjo slides and notes so quickly.

Hope you are not relying on sales of the t shirts to pay for any particular expenses!


I have no particular forseeable expenses for which I need to sell a lot of t-shirts.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Commonsense local realism refutes Bell's theorem

Postby Heinera » Wed Jun 11, 2014 8:48 am

gill1109 wrote:
Ben6993 wrote:Hi Richard

Thanks for posting your Vaxjo slides and notes so quickly.

Hope you are not relying on sales of the t shirts to pay for any particular expenses!


I have no particular forseeable expenses for which I need to sell a lot of t-shirts.

But what about the unforeseeable expenses?
Heinera
 
Posts: 917
Joined: Thu Feb 06, 2014 1:50 am

Re: Commonsense local realism refutes Bell's theorem

Postby Heinera » Wed Jun 11, 2014 11:27 am

PS: I want that t-shirt. Both, actually. :)
Heinera
 
Posts: 917
Joined: Thu Feb 06, 2014 1:50 am

Re: Commonsense local realism refutes Bell's theorem

Postby Xray » Wed Jun 11, 2014 3:23 pm

gill1109 wrote:
minkwe wrote:
gill1109 wrote:He just gave a talk at the Växjö conference, the slides are here: http://www.slideshare.net/gill1109/vaxjo-2014. The 10th slide is about Gordon Watson but he kept it anonymous.



Slide #20 is also funny, as it shows the lack of understanding that a cosine curve which plots the DIFFERENCE between Alice and Bob's angles vs the correlation, means in fact that Alice and Bob's angles have been varied, otherwise you would get a single point.

Slide #5 is my favorite. The whole thing looks to me like a classified ad, so you could easily ask what he is selling. Who knows, hopefully it was more interesting than is apparent from the slides alone.


I had better improve the wording of slide #20. Seems some readers are misunderstanding the point here.

A lot of people liked the talk a great deal! It led to a lot of good discussion with people who think that Bell got it completely wrong. There are a lot of such people at this conference.

I am already getting requests for t-shirts.

But this is all getting rather off-topic and Michel is getting rather personal. If anyone wants to start a new topic about my slides, go ahead. But they may get updated from time to time.


Gill,

I opened this thread and I am very happy for your Slide-show to be discussed here forever

thank you for bringing them here so quickly

all Admins please note my wishes

Xray
Xray
 
Posts: 44
Joined: Mon Apr 21, 2014 2:23 pm

Re: Commonsense local realism refutes Bell's theorem

Postby Xray » Wed Jun 11, 2014 3:37 pm

gill1109 wrote:He just gave a talk at the Växjö conference, the slides are here: http://www.slideshare.net/gill1109/vaxjo-2014. The 10th slide is about Gordon Watson but he kept it anonymous.


Gill

Please explain the meaning of Slide 10 for me please

can you use other words?

I am wanting to know why Gordon Watson kept it anonymous

Did he have other words that he wanted instead?

Thanks

Xray
Xray
 
Posts: 44
Joined: Mon Apr 21, 2014 2:23 pm

Re: Commonsense local realism refutes Bell's theorem

Postby Xray » Wed Jun 11, 2014 4:59 pm

Gill,

bringing slide 10 here to help us all

Gill slide #10

Logic is difficult

Bell proved a theorem that a certain inequality could not be violated

Bell was delighted that experiment had violated (or could be expected to violate) his inequality

XXX sees this as proof that Bell's theorem is false


I have fixed your typo [spelling is difficult] so in addition to explaining more carefully your point in this slide could you add something like

Bell was delighted because ….

The point c****pot miss is ….


Xray
Xray
 
Posts: 44
Joined: Mon Apr 21, 2014 2:23 pm

Re: Commonsense local realism refutes Bell's theorem

Postby gill1109 » Wed Jun 11, 2014 9:33 pm

According to CHSH "[Bell] showed that ... no local hidden-variable theory can reproduce all of the statistical predictions of quantum mechanics."

In Bertlmann's socks, Bell lists four (not exhaustive) possibilities which follow from his analysis. Everyone on this forum should know them. One of the possibilities is that quantum mechanics is wrong.

However the statistical predictions of quantum mechanics were indeed confirmed yet again by experiment, so Bell was kind of delighted that we could not "escape" from his "theorem" by just saying - oh well, QM works in some circumstances, but in an EPR-B experiment it actually breaks down.

It seems that XXX is confused as to what is Bell's theorem. If we take a narrow view, Bell's theorem says that if you have functions A, B and rho satisfying certain properties, then a certain quantity S defined in terms of A B and rho is less than or equal to 2.

The "S" which we see in an experiment is not defined that way. It is defined quite differently.

A local hidden variables theory says that if we would repeatedly measure pairs of particles in directions a and b and average the product of the outcomes, then in the long run we would see a value approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda

So Bell was excited that experiment was seeming to prove that a local hidden variables theory was not possible.

The experiment was saying that we had to discard one of the four possibilities. That left three. The field is narrowed down.

Unfortunately at the time the experiment which was actually done was not the experiment which was needed to be done. And even today they still haven't done the experiment which needed to be done and observed violation of Bell inequalities. Some people, Kristel Michielsen for instance, believes that this will never happen. Emilio Santos believes this will never happen because QM itself (through some kind of uncertainty relations) will make it impossible to engineer the initial conditions which are needed for the meaningful experiment. I call this Bell's fifth position.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Commonsense local realism refutes Bell's theorem

Postby minkwe » Thu Jun 12, 2014 7:28 am

gill1109 wrote:It seems that XXX is confused as to what is Bell's theorem.

Looking at slides #9 and #10, it seems that it is Gill that is confused about what is Bell's theorem. He says on one slide that there is no Bell's theorem, and on the very next slide he talks about how Bell proved a theorem, and then accuses xxx of not understanding what Bell's theorem is.

Maybe he will explain what he meant by Bell's theorem on page #9 and what he meant by Bell's theorem on page #10. They must be different, otherwise how did Bell prove a theorem that does not exist?

gill1109 wrote:A local hidden variables theory says that if we would repeatedly measure pairs of particles in directions a and b and average the product of the outcomes, then in the long run we would see a value approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda

For any theory whatsoever, local or non-local, if we would repeatedly measure pairs of particles in directions a and b, the average of the product of outcomes will be approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda. There is absolutely nothing in that expression that restricts it to LHV.

gill1109 wrote:Some people, Kristel Michielsen for instance, believes that this will never happen. Emilio Santos believes this will never happen because QM itself (through some kind of uncertainty relations) will make it impossible to engineer the initial conditions which are needed for the meaningful experiment. I call this Bell's fifth position.

According to Slide #5, you invented it. I wonder what Kristel and Emilio would think about that.
minkwe
 
Posts: 1441
Joined: Sat Feb 08, 2014 10:22 am

Re: Commonsense local realism refutes Bell's theorem

Postby Heinera » Thu Jun 12, 2014 10:17 am

minkwe wrote:
gill1109 wrote:
gill1109 wrote:A local hidden variables theory says that if we would repeatedly measure pairs of particles in directions a and b and average the product of the outcomes, then in the long run we would see a value approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda

For any theory whatsoever, local or non-local, if we would repeatedly measure pairs of particles in directions a and b, the average of the product of outcomes will be approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda. There is absolutely nothing in that expression that restricts it to LHV.


Wrong. For a non-local theory, the functions A(a, lambda) and B(b, lambda) are not even well defined. Only in an LHV model is it possible to talk about two functions like that with those arguments.
Heinera
 
Posts: 917
Joined: Thu Feb 06, 2014 1:50 am

Re: Commonsense local realism refutes Bell's theorem

Postby minkwe » Thu Jun 12, 2014 11:17 am

Heinera wrote:
minkwe wrote:
gill1109 wrote:A local hidden variables theory says that if we would repeatedly measure pairs of particles in directions a and b and average the product of the outcomes, then in the long run we would see a value approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda

For any theory whatsoever, local or non-local, if we would repeatedly measure pairs of particles in directions a and b, the average of the product of outcomes will be approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda. There is absolutely nothing in that expression that restricts it to LHV.

Wrong. For a non-local theory, the functions A(a, lambda) and B(b, lambda) are not even well defined. Only in an LHV model is it possible to talk about two functions like that with those arguments.

As usual, you don't know what you are talking about:

Let us imagine a non-local function in which our non-local hidden variable lambda is actually {a,b} ie, Each outcome uses information about the settings on both sides. Therefore, A(a, lambda) and B(b, lambda) are well defined. And the expectation value for the product of outcomes if we repeatedly measure pairs of particles in directions a, and b under such a theory will be exactly the same as integral A(a, lambda) B(b, lambda) rho(lambda) d lambda, contrary to your outburst above.

In fact, guess who presented such a non-local hidden variable theory not too long ago in response to a challenge: (viewtopic.php?f=6&t=53).

Think before you shout, "Wrong!".
minkwe
 
Posts: 1441
Joined: Sat Feb 08, 2014 10:22 am

Re: Commonsense local realism refutes Bell's theorem

Postby Heinera » Thu Jun 12, 2014 12:21 pm

minkwe wrote:Let us imagine a non-local function in which our non-local hidden variable lambda is actually {a,b} ie, Each outcome uses information about the settings on both sides. Therefore, A(a, lambda) and B(b, lambda) are well defined. And the expectation value for the product of outcomes if we repeatedly measure pairs of particles in directions a, and b under such a theory will be exactly the same as integral A(a, lambda) B(b, lambda) rho(lambda) d lambda, contrary to your outburst above.

Nope, that won't work. Your hidden variable lambda = (a,b) actually depends on the first argument to A(a, lambda). If you check up the definition of a multivariate function, you will see that the arguments are supposed to be independent. So your definition of A and B only applies to a very restricted subset of the arguments. And the subset is too restricted to make sense:

If we go back to the original integral A(a, lambda) B(b, lambda) rho(lambda) d lambda, you will see that the integral is over all possible values of lambda, while still keeping a and b fixed. This is not well defined with your definition of lambda as equal to (a,b).
Heinera
 
Posts: 917
Joined: Thu Feb 06, 2014 1:50 am

Re: Commonsense local realism refutes Bell's theorem

Postby minkwe » Thu Jun 12, 2014 1:35 pm

Heinera wrote:Nope, that won't work. Your hidden variable lambda = (a,b) actually depends on the first argument to A(a, lambda).

Sorry, that works just fine. If lambda is (a,b) then A(a, lambda) is A(a, b), and integral A(a, lambda) B(b, lambda) rho(lambda) d lambda, is just integral of A(a, b) B(a,b) rho(a,b) d (a,b), you have a problem with that?

So your definition of A and B only applies to a very restricted subset of the arguments. .. And the subset is too restricted to make sense:

So what? It is a non-local model, it is already nonsensical. The outcome at Alice is A(a, b) (well defined), the outcome at Bob is B(a,b) (well defined), the product of the paired outcomes is

A(a, b)B(a,b), well defined

All well defined, if you repeatedly measure two particle pairs at Alice and Bob, the average of the product is well defined, the expectation value of that product will indeed be well defined and will be the integral, contrary to your earlier outburst. There is nothing about the expectation value of a paired product that is specific to LHV. What exactly are you arguing against here?

If we go back to the original integral A(a, lambda) B(b, lambda) rho(lambda) d lambda, you will see that the integral is over all possible values of lambda, while still keeping a and b fixed. This is not well defined with your definition of lambda as equal to (a,b).

Wrong. lambda is (a,b), you can argue that there is only one value of lambda if a and b are fixed and lambda = (a,b) but what relevance has that got to do with the fact that the expression is exactly the same well defined integral? There is nothing in your argument. Zilch.
minkwe
 
Posts: 1441
Joined: Sat Feb 08, 2014 10:22 am

Re: Commonsense local realism refutes Bell's theorem

Postby Heinera » Thu Jun 12, 2014 1:51 pm

So you mean we can make substitutions A(a, lambda) = A(a, b) and B(b, lambda) = B(a, b) and lambda = (a, b) so that the integral becomes

integral A(a, b) B(a, b) rho(a, b) d (a,b) ?
Heinera
 
Posts: 917
Joined: Thu Feb 06, 2014 1:50 am

Re: Commonsense local realism refutes Bell's theorem

Postby gill1109 » Thu Jun 12, 2014 8:20 pm

minkwe wrote:
gill1109 wrote:It seems that XXX is confused as to what is Bell's theorem.

Looking at slides #9 and #10, it seems that it is Gill that is confused about what is Bell's theorem. He says on one slide that there is no Bell's theorem, and on the very next slide he talks about how Bell proved a theorem, and then accuses xxx of not understanding what Bell's theorem is.

Maybe he will explain what he meant by Bell's theorem on page #9 and what he meant by Bell's theorem on page #10. They must be different, otherwise how did Bell prove a theorem that does not exist?

gill1109 wrote:A local hidden variables theory says that if we would repeatedly measure pairs of particles in directions a and b and average the product of the outcomes, then in the long run we would see a value approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda

For any theory whatsoever, local or non-local, if we would repeatedly measure pairs of particles in directions a and b, the average of the product of outcomes will be approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda. There is absolutely nothing in that expression that restricts it to LHV.

gill1109 wrote:Some people, Kristel Michielsen for instance, believes that this will never happen. Emilio Santos believes this will never happen because QM itself (through some kind of uncertainty relations) will make it impossible to engineer the initial conditions which are needed for the meaningful experiment. I call this Bell's fifth position.

According to Slide #5, you invented it. I wonder what Kristel and Emilio would think about that.

These questions again beautifully illustrate name = value problems, which was also one of the themes of my talk. Unfortunately, though you see the slides, you don't hear the talk, so you miss even more contradictions. Another very serious theme of the talk was humour in science. If anything, one should read Bell's book from cover to cover in order to enjoy the jokes.

Talking about humour, and to be honest, I am still feeling devastated that Rik Mayall just died (two days before I gave the talk).

It was nice talking to Kristel. She was happy about my talk and admitted to holding to Bell's fifth position.

I named it thus. Possibly, Emilio Santos was the first to put if forward, seriously. Bell never mentioned it in his writings but he admitted in private correspondence with Santos that it was a logical possibility.

So there are many things *called* Bell's theorem. Before we discuss it, we ought to say which thing we meant. On different slides I mean different things. Deliberately.

I hope everyone noticed the picture of the book "The name of the rose" by Umberto Eco. And is reminded of Romeo's statement to Juliet: "A rose by any other name would smell as sweet".
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Commonsense local realism refutes Bell's theorem

Postby gill1109 » Thu Jun 12, 2014 8:26 pm

minkwe wrote:
Heinera wrote:
minkwe wrote:"gill1109 said "A local hidden variables theory says that if we would repeatedly measure pairs of particles in directions a and b and average the product of the outcomes, then in the long run we would see a value approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda"
For any theory whatsoever, local or non-local, if we would repeatedly measure pairs of particles in directions a and b, the average of the product of outcomes will be approximately equal to integral A(a, lambda) B(b, lambda) rho(lambda) d lambda. There is absolutely nothing in that expression that restricts it to LHV.

Wrong. For a non-local theory, the functions A(a, lambda) and B(b, lambda) are not even well defined. Only in an LHV model is it possible to talk about two functions like that with those arguments.

As usual, you don't know what you are talking about:

Let us imagine a non-local function in which our non-local hidden variable lambda is actually {a,b} ie, Each outcome uses information about the settings on both sides. Therefore, A(a, lambda) and B(b, lambda) are well defined. And the expectation value for the product of outcomes if we repeatedly measure pairs of particles in directions a, and b under such a theory will be exactly the same as integral A(a, lambda) B(b, lambda) rho(lambda) d lambda, contrary to your outburst above.

In fact, guess who presented such a non-local hidden variable theory not too long ago in response to a challenge: (viewtopic.php?f=6&t=53).

Think before you shout, "Wrong!".

A hidden variable theory in which the hidden variable includes a, b is called a conspiracy theory. Example: Christian's simulation model based on the Pearle model. Pairs of particles such that either particle fails a test never ever existed, in reality. Thus a and b are in some sense included in lambda. Another way to say this is that the probability distribution of the hidden variable is different for each a, b pair. The set of possible values is the same but the distribution is different. rho(lambda) is actually rho(lambda, a, b).

To say this a different way, we define Lambda = (lambda, a, b); we define d Lambda = d lambda; we define A(a, Lambda) = A(a, lambda) and B(b, Lambda) = B(b, lambda; we renormalize rho(lambda) to obtain rho(Lambda); and now we have

E(a, b) = integral A(a, Lambda) B(b, Lambda) rho(Lambda) d Lambda

First you see it, now you don't. A conjuring trick. This is how Accardi's model, Hess and Phillip's model, and Christian's adaptation of Pearle's model all work.

It is easy to convert a detection loophole model into a conspiracy loophole model. It is just a question of replacing the unconditional distribution of lambda with the conditional distribution given detection. Physicists who do not know much probability and statistics very easily make the mistake of confusing these things.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Commonsense local realism refutes Bell's theorem

Postby minkwe » Thu Jun 12, 2014 9:46 pm

gill1109 wrote:To say this a different way, we define Lambda = (lambda, a, b); we define d Lambda = d lambda; we define A(a, Lambda) = A(a, lambda) and B(b, Lambda) = B(b, lambda; we renormalize rho(lambda) to obtain rho(Lambda); and now we have

E(a, b) = integral A(a, Lambda) B(b, Lambda) rho(Lambda) d Lambda

Thank you for confirming that there is nothing specific about LHV in the above expression. You can change the meaning of the terms from a LHV to non-local HV to conspiracy and the expectation value of the paired product will still have the same function form. You have an outcome on one side, and an outcome on the other side, you multiply them together and take the pro ability weighted average as N tends to infinity. The procedure is exactly the same no matter what theory is assumed to be producing the results. There is nothing LHV -specific in the above expression.

Maybe Heine will believe it if it comes from you.
minkwe
 
Posts: 1441
Joined: Sat Feb 08, 2014 10:22 am

Re: Commonsense local realism refutes Bell's theorem

Postby minkwe » Thu Jun 12, 2014 9:54 pm

gill1109 wrote:A hidden variable theory in which the hidden variable includes a, b is called a conspiracy theory.

Strictly speaking, the above is misleading. It is only a conspiracy theory if Alice and Bob are aware of each others settings. But that doesn't change the fact that it can also be non-local.
minkwe
 
Posts: 1441
Joined: Sat Feb 08, 2014 10:22 am

Re: Commonsense local realism refutes Bell's theorem

Postby Joy Christian » Thu Jun 12, 2014 9:58 pm

gill1109 wrote:A hidden variable theory in which the hidden variable includes a, b is called a conspiracy theory. Example: Christian's simulation model based on the Pearle model...

First you see it, now you don't. A conjuring trick. This is how ... Christian's ... model ... work.

Yeah, right, in the world of morons.

Everything works as a conspiracy theory in the world of morons.

On the other hand, anyone with a brain can see that my model has nothing to do with detection loophole or conspiracy theory: http://arxiv.org/abs/1405.2355.
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Commonsense local realism refutes Bell's theorem

Postby gill1109 » Fri Jun 13, 2014 1:05 am

minkwe wrote:
gill1109 wrote:A hidden variable theory in which the hidden variable includes a, b is called a conspiracy theory.

Strictly speaking, the above is misleading. It is only a conspiracy theory if Alice and Bob are aware of each others settings. But that doesn't change the fact that it can also be non-local.

Unfortunately, people use words in different ways. One person uses the word "conspiracy", another uses the word "non-local". What's in a name? Fortunately the mathematics is crystal clear.

Nowadays people like to separate out various ways to violate Bell's inequality.

According to Boris Tsirelson, "the theorem" could be summarized as "quantum theory is incompatible with locality+realism+no-conspiracy". Hence, if we believe that experiment has proven that Nature chooses the side of quantum theory, then we must reject either locality or realism or no-conspiracy.

Bell's later papers, especially the wonderful "Bertlmann's socks" paper, are quite clear about the logic of all this. He very explicitly goes through at least *five* options:

QM is false
Locality is false
Realism is false
No-conspiracy is false
Experiment is false

There is of course yet another option,

Bell's logic is false

Michel, did you read the later papers of John Bell yet? Maybe you could learn something from them. I have put two very important ones in the drop box which I have shared already with Gordon Watson. Anyone else who'ld like to join, please just send me an ordinary email, or a private message using the forum's facility for this.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Commonsense local realism refutes Bell's theorem

Postby Heinera » Fri Jun 13, 2014 1:16 am

minkwe wrote:
gill1109 wrote:To say this a different way, we define Lambda = (lambda, a, b); we define d Lambda = d lambda; we define A(a, Lambda) = A(a, lambda) and B(b, Lambda) = B(b, lambda; we renormalize rho(lambda) to obtain rho(Lambda); and now we have

E(a, b) = integral A(a, Lambda) B(b, Lambda) rho(Lambda) d Lambda

Thank you for confirming that there is nothing specific about LHV in the above expression. You can change the meaning of the terms from a LHV to non-local HV to conspiracy and the expectation value of the paired product will still have the same function form. You have an outcome on one side, and an outcome on the other side, you multiply them together and take the pro ability weighted average as N tends to infinity. The procedure is exactly the same no matter what theory is assumed to be producing the results. There is nothing LHV -specific in the above expression.

Maybe Heine will believe it if it comes from you.

The expression above can also be written as

E(a, b) = integral A(a, lambda) B(b, lambda) rho(a, b, lambda) d lambda

Now it is easier to see that it is LHV-specific, albeit with a conspiracy term rho(a, b, lambda).
Heinera
 
Posts: 917
Joined: Thu Feb 06, 2014 1:50 am

PreviousNext

Return to Sci.Physics.Foundations

Who is online

Users browsing this forum: No registered users and 79 guests

cron
CodeCogs - An Open Source Scientific Library