Why is entanglement necessary for understanding quantum mechanics?

[Jabbu]

It might be worth going back to the beginning of this thread, reading through it again... It started with johan0001 asking why there was anything strange going on at all, how measuring the polarization of two entangled photons and finding a correlation is any different than putting two gloves in two boxes then finding that the presence of a left-handed glove in one box is correlated with finding a right-handed glove in the other.

I've actually read quite a bit about this stuff, but the crucial thing which eluded me is that photon polarization is supposed to be random, I thought it's constant. This is very subtle difference if you don't know it's implied. But what is fascinating I was reading through those experiments and it all actually made sense, it gives the same result, only there is no any mystery. Allegedly random photon polarization is a big news to me, it changes everything, and nothing makes sense anymore.

 

[billschnieder]

I've actually read quite a bit about this stuff, but the crucial thing which eluded me is that photon polarization is supposed to be random, I thought it's constant.

If you think about what "random" means you'll see there is some confusion in the way it is being used here. The only time it makes sense to talk of randomness with respect to the polarization of a single photon, has to do with inability to predict what it will be, but in that case it doesn't mean it won't be constant. For many different photons, you can say that their polarization directions are random, which simply means there is no discernible pattern from one photon to the other in the set. It does not mean each individual photon does not have a fixed constant polarization direction (although some people believe that). It is important to distinguish claims about individual photons, and claims about ensembles of photons.

 

[bhobba]

If you think about what "random" means you'll see there is some confusion in the way it is being used here.

Indeed there is.

But please bear in mind the polarisation is not random - its nothing at all - ie it doesn't even have the property of polarisation until observed. Its the act of observation that gives a random outcome. The weirdness is that it's always correlated with the other photon. But the other photon doesn't have the property of polarisation until observed either.

One may think some instantaneous communication went between the two photons, but what Bell's theorem shows is you don't have to view it that way. You can keep locality if you assume it doesn't have the property until observed - which is actually what the formalism says so really there is no issue. I personally go further and think there is some kind of communication, and the polarization don't exist until observed - which the theorem also allows. But that's just me - you don't have to do it.

Thanks
Bill

 

[Jabbu]

We have two photons, entangled so that if one of them passes a filter at a given angle, the other one definitely will not pass a filter at that angle.

You mean if they are orthogonally entangled? But if they are entangled with parallel polarization then whenever one goes through so must the other, like this:

theta_A = 0, theta_B = 0
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

Consider the first pair of photons from that sequence. Let's say they have 45 degrees polarization so they both have 50% chance to pass through their 0 degrees aligned polarizers. Photon A happens to go through, first, and informs his twin brother he must go through as well, or else! So then photon B finally meets with polarizer B, bribes it, and continues as planned. Even if the explanation is some instantaneous connection between the photons, it still doesn't explain why would second polarizer just go along with that deal.

 

[DrChinese]

Ok, all together it should now look like this:


theta_A = 0, theta_B = 0
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

sequence length = 20
matching pairs (00 or 11) = 20
mismatching pairs (01 or 10) = 0
Correlation = (match - mismatch) / sequence length = 20/20 = 100%


theta_A = 0, theta_B = 45
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 0 0 0 1 0 0 1 0 1 1 0 0 1 1 0 1 1

sequence length = 20
matching pairs (00 or 11) = 15
mismatching pairs (01 or 10) = 5
Correlation = (match - mismatch) / sequence length = 10/20 = 50%


theta_A = 0, theta_B = 90
A: 1 0 1 0 1 0 1 0 1 0 1 1 0 0 0 0 1 1 0 1
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

sequence length = 20
matching pairs (00 or 11) = 10
mismatching pairs (01 or 10) = 10
Correlation = (match - mismatch) / sequence length = 0/20 = 0%

I just saw this post. Unfortunately, you have mixed up the entangled match percentages and correlation somewhat such that the above is not correct.

Correct is:
Theta=0 degrees, match=100%, correlation=1
Theta=0 degrees, match=50%, correlation=0
Theta=90 degrees, match=0%, correlation=-1

So the 2nd and 3rd patterns are inaccurate.

 

[PeterDonis]

Consider the first pair of photons from that sequence. Let's say they have 45 degrees polarization

Here is where you are going wrong: you are assuming that the photons have some definite polarization (45 degrees or anything else) before their polarization is measured. But that would mean that both photons started out in a polarization eigenstate; and that is a very restrictive assumption. There are *lots* of states of the photons which are *not* polarization eigenstates (and in the actual experiments that are done to test this, the photons are not in polarization eigenstates). When people say the polarization is "random", or that the photon does not have a definite polarization until it's observed, what they mean is that the photon's state at the start is not a polarization eigenstate.

 

[DrChinese]

You mean if they are orthogonally entangled? But if they are entangled with parallel polarization then whenever one goes through so must the other, like this:

theta_A = 0, theta_B = 0
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

...

There are 2 types of PDC producing entangled photon pairs: Type I and Type II. Type I produces pairs that are parallel, Type II produces pairs that are orthogonal. I prefer to discuss Type I because the examples are much easier to describe in posts. The fundamental principle is the same either way. In fact you can rotate either Type to act like the other Type.

So sometimes one poster is referring to one type when another poster is referring to the other. I do my best to label as Type I PDC and entangled so as to be clear.

 

[DrChinese]

Consider the first pair of photons from that sequence. Let's say they have 45 degrees polarization so they both have 50% chance to pass through their 0 degrees aligned polarizers. Photon A happens to go through, first, and informs his twin brother he must go through as well, or else! So then photon B finally meets with polarizer B, bribes it, and continues as planned. Even if the explanation is some instantaneous connection between the photons, it still doesn't explain why would second polarizer just go along with that deal.

As PeterDonis indicates, you must be careful when you assume that there is a pre-existing polarization as Bell's Theorem places restrictions on such.

However, your model WILL more or less work as described above. It is called a non-local model. Under such model, there is instantaneous communication from A to B (assuming A is measured first). In this hypothetical case, B conforms to A. The statistics work out fine and Bell's Theorem is not a problem. Further, you don't need to answer "WHY" it happens that way any more than you ask "why" the speed of light is c. It just is.

The problem with this interpretation (model) is that there is no other evidence of the signalling mechanism other than in entanglement, and it is not supported by the so-called "standard model". Accordingly, it requires a new force currently undiscovered - or other significant changes to our understanding of physics. There are several non-local interpretations that explicitly solve these issues. Bohmian Mechanics is one such, you can google that.

 

[DrChinese]

The premise must be wrong, surely?

Yes, but which premise?

That is what Bell helps us to understand. Usually the premises are:

1. QM is correct (this is strongly confirmed by experiment).
2. There is no action at a distance.
3. Quantum particles such as photons have well-defined properties independent of the act of observation.
4. There is causality (the future does not affect the past).

At least one of the above is incorrect, according to Bell's Theorem.

 

[Jabbu]

If you think about what "random" means you'll see there is some confusion in the way it is being used here. The only time it makes sense to talk of randomness with respect to the polarization of a single photon, has to do with inability to predict what it will be, but in that case it doesn't mean it won't be constant. For many different photons, you can say that their polarization directions are random, which simply means there is no discernible pattern from one photon to the other in the set. It does not mean each individual photon does not have a fixed constant polarization direction (although some people believe that). It is important to distinguish claims about individual photons, and claims about ensembles of photons.

That's why I'm asking about how actual data streams look like.

theta_A = 0, theta_B = 0
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

When there is always 50% ones and 50% zeroes in each individual sequence then according to Malus's law photon polarization is either random or constant at 45 degrees relative to polarizers. I thought photon polarization was constant and centered relative to polarizers, so that data stream for theta_A = 0, theta_B = 0 looks like this:

A: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
B: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

It turns out 100% correlation as well. Only this time it's not weird why sequence A coincides with sequence B, they simply both have 100% chance to go through. It's very subtle difference to distinguish if you don't look at some actual experimental data, which I can't find anywhere on the internet.

 

[Jabbu]

One may think some instantaneous communication went between the two photons, but what Bell's theorem shows is you don't have to view it that way. You can keep locality if you assume it doesn't have the property until observed - which is actually what the formalism says so really there is no issue. I personally go further and think there is some kind of communication, and the polarization don't exist until observed - which the theorem also allows. But that's just me - you don't have to do it.

Is there mathematically even any difference between "random" and "not existing - then existing"? I can conceive instantaneous interaction and stuff disappearing and reappearing in different locations, if I must. But that there is no Moon when I don't look at it, that's too much.

 

[Jabbu]

Here is where you are going wrong: you are assuming that the photons have some definite polarization (45 degrees or anything else) before their polarization is measured. But that would mean that both photons started out in a polarization eigenstate; and that is a very restrictive assumption. There are *lots* of states of the photons which are *not* polarization eigenstates (and in the actual experiments that are done to test this, the photons are not in polarization eigenstates). When people say the polarization is "random", or that the photon does not have a definite polarization until it's observed, what they mean is that the photon's state at the start is not a polarization eigenstate.

I don't think that changes the conclusion. Photon A interacts with polarizer A first, and the decision whether photon B will go through polarizer B is made right there and then. Ok? But even if the explanation is some instantaneous connection between the photons, it still doesn't explain why would second polarizer just go along with that deal.

 

[Jabbu]

As PeterDonis indicates, you must be careful when you assume that there is a pre-existing polarization as Bell's Theorem places restrictions on such.

However, your model WILL more or less work as described above. It is called a non-local model. Under such model, there is instantaneous communication from A to B (assuming A is measured first). In this hypothetical case, B conforms to A. The statistics work out fine and Bell's Theorem is not a problem. Further, you don't need to answer "WHY" it happens that way any more than you ask "why" the speed of light is c. It just is.

The problem with this interpretation (model) is that there is no other evidence of the signalling mechanism other than in entanglement, and it is not supported by the so-called "standard model". Accordingly, it requires a new force currently undiscovered - or other significant changes to our understanding of physics. There are several non-local interpretations that explicitly solve these issues. Bohmian Mechanics is one such, you can google that.

I thought instantaneous signaling is "standard model". If not that, what's the explanation then?

 

[DrChinese]

But that there is no Moon when I don't look at it, that's too much.

Type I entangled photons

A=0 degrees:
A: 1 0 1 0 1 0 0 0 1 0 1 1 0 1 0 0 1 1 0 1

B=120 degrees:
B: 0 1 0 1 0 0 1 1 0 1 1 0 1 0 0 1 1 1 1 0

Match rate above is 25%. If the moon is there when not observed, what is C below (had it been measured)?

C=240 degrees:
C: ? ? ? ...

 

[DrChinese]

I thought instantaneous signaling is "standard model". If not that, what's the explanation then?

Not really. The "standard model" is actually silent on the mechanism. There are also models where there is no instantaneous signalling.

The term you will sometimes hear is: "quantum non-locality". That means it sort of appears non-local, but strictly in a "quantum" manner - such that it does not conflict with relativity. Fully non-local models violate relativity.

 

[DrChinese]

PS regarding my post #164:

A-B match rate is 25%
A-C match rate should be 25%
B-C match rate should also be 25%

After all, theta is the same for all 3.

Good luck!

 

[PeterDonis]

Photon A interacts with polarizer A first, and the decision whether photon B will go through polarizer B is made right there and then.

But this way of looking at it is frame-dependent; there will be another frame in which photon B interacts first, and the decision whether photon A will go through is made right there and then. There is no invariant fact of the matter about which photon interacts first. That's why DrChinese said that fully non-local models violate relativity.

For what it's worth, quantum field theory has a somewhat different take on this, at least as I understand it. In QFT, "causality" does not mean that spacelike separated measurements can't be correlated, even to a degree that violates the Bell inequalities; it only means that spacelike separated measurements must commute, i.e., the results must not depend on which measurement happens first. The photon measurements satisfy this criterion, so QFT simply says "no problem".

 

[Jabbu]

I just saw this post. Unfortunately, you have mixed up the entangled match percentages and correlation somewhat such that the above is not correct.

Correct is:
Theta=0 degrees, match=100%, correlation=1
Theta=0 degrees, match=50%, correlation=0
Theta=90 degrees, match=0%, correlation=-1

So the 2nd and 3rd patterns are inaccurate.

After PeterDonis' post #132 I adjusted it to match the new formula:

correlation = (match - mismatch) / sequence length

...where "match" is number of 00 + 11 pairs, and "mismatch" is number of 01 + 10 pairs.


Originally this seemed to be the formula:

correlation = (match_0 + match_1) / sequence length

...where "match_0" is number 00 pairs, and "match_1" is number of 11 pairs.



What is your "match" formula?

 

[DrChinese]

After PeterDonis' post #132 I adjusted it to match the new formula:

correlation = (match - mismatch) / sequence length

...where "match" is number of 00 + 11 pairs, and "mismatch" is number of 01 + 10 pairs.Originally this seemed to be the formula:

correlation = (match_0 + match_1) / sequence length

...where "match_0" is number 00 pairs, and "match_1" is number of 11 pairs.
What is your "match" formula?

Type I Match: cos^2(theta)
Type I Correlation: cos^2(theta)-sin^(theta)

I did warn you in post #108, it is best to discuss matches and for Type I.

 

[Jabbu]

Type I entangled photons

A=0 degrees:
A: 1 0 1 0 1 0 0 0 1 0 1 1 0 1 0 0 1 1 0 1

B=120 degrees:
B: 0 1 0 1 0 0 1 1 0 1 1 0 1 0 0 1 1 1 1 0

Match rate above is 25%. If the moon is there when not observed, what is C below (had it been measured)?

C=240 degrees:
C: ? ? ? ...

True range is only from 0 to 90 degrees relative, from parallel to orthogonal. So both 120 and 240 are really just the same 60 degrees relative angle. Anyway, what I thought is that photon polarization is actually constant and deliberately centred between the two polarizers, so in the case of master_theta = 60 or 120 or 240:

theta_A = -30, theta_B = +30
A: 1 1 0 1 1 0 0 1 1 0 1 0 1 1 1 1 1 1 1 1
B: 1 1 1 1 0 1 0 1 1 1 1 1 0 1 1 0 0 1 1 1

Both photons now according to Malus have cos^2(30) = 75% chance to pass through, and therefore there is 75% ones and 25% zeros in each stream. So what is the chance of getting 11 or 00 matching pairs vs chance of getting 10 or 01 mismatching pairs?

chance of match: (0.75 * 0.75) + (0.25 * 0.25) = 0.625
chance of mismatch: (0.25 * 0.75) + (0.75 * 0.25) = 0.375
correlation: 0.625 - 0.375 = 25%

 

[Jabbu]

Type I Match: cos^2(theta)
Type I Correlation: cos^2(theta)-sin^(theta)

I did warn you in post #108, it is best to discuss matches and for Type I.

I'm not talking about any other experiment but what you call "Type I". cos^2(theta) is theoretical prediction equation. Correlation is calculated differently from experimental data, in terms of matching and mismatching pairs, along these lines:

correlation = (match - mismatch) / sequence length
...where "match" is number of 00 + 11 pairs, and "mismatch" is number of 01 + 10 pairs

correlation = (match_0 + match_1) / sequence length
...where "match_0" is number 00 pairs, and "match_1" is number of 11 pairs


The formula defines how actual data streams are supposed to look like, it makes all the difference, but the difference is very subtle, so it is important to get this formula straight.

 

[DrChinese]

The formula defines how actual data streams are supposed to look like, it makes all the difference, but the difference is very subtle, so it is important to get this formula straight.

And I keep telling you that the formula for matches and the formula for correlation are completely different. Forget correlation, you will simply make things more complicated than needed.

Matches at 60 degrees = matches at 120 degrees = 25%.

Before you start modeling things, you would find it beneficial to understand fully what happens experimentally. And many of your ideas about that are incorrect, especially most everything in post 170. For example, 1's are not more likely than 0's for entangled streams. They are equally likely. The correlation for 60 degrees is not 25%, it is -.5.

 

[bhobba]

Is there mathematically even any difference between "random" and "not existing - then existing"?

You are viewing it incorrectly.

When we say something exists we usually have the view of what's called naive reality - you can look up exactly what it is. I will not get into a philosophical discussion about what existing means except to say most assume some kind of naive reality.

Now QM is a theory about observations. It is silent on what's going on when not observed. Naive reality may apply - or it may not - the theory says nothing one way or the other - on the surface that is.

We can't say anything about the photons polarisation when it's not being observed. That is the key difference between your analogy, Bertlmanns socks vs QM. Both, being classical objects, obey naive reality.

What Bells theorem shows is QM is not only silent on the issue, but in fact rules out naive reality. That's the striking and interesting thing about it. The Bertlmann socks analogy is not correct.

Thanks
Bill

 

[bhobba]

it still doesn't explain why would second polarizer just go along with that deal.

That's the definition of correlated.

What Bell shows is classical correlations like Bertlmann's socks are different to quantum ones.

Why are they correlated?

Apply the Born Rule to 1/root 2 (|a>|b> + |b>|a>).

Its basic QM.

Thanks
Bill

 

[Jabbu]

Before you start modeling things, you would find it beneficial to understand fully what happens experimentally.

Can you point any document on the internet where we can see samples of actual experimental data?

 

[Jabbu]

We can't say anything about the photons polarisation when it's not being observed. That is the key difference between your analogy, Bertlmanns socks vs QM. Both, being classical objects, obey naive reality.

I'm talking about photons after they have been observed. If 50% photons pass through regardless of polarizer rotation, does that not mean those photons had random polarization?

 

[microsansfil]

Hello

Perhaps look at Kochen–Specker theorem, which shows that the result of any individual measurement of spin was not fixed independently of the choice of measurements. An outcome is "not determined" by prior conditions.

Patrick

 

[Jabbu]

Hello

Perhaps look at Kochen–Specker theorem, which shows that the result of any individual measurement of spin was not fixed independently of the choice of measurements. An outcome is "not determined" by prior conditions.

Patrick

All the equation needs is two polarizer angles. I don't see how can that be anything but "prior condition".

 

[microsansfil]

I don't see how can that be anything but "prior condition".

The result cannot be determined before make the measurements.

http://arxiv.org/abs/quant-ph/0604079

Conway and Kochen do not prove that free will does exist. The definition of "free will" used in the proof of this theorem is simply that an outcome is "not determined" by prior conditions.

Patrick
PS
http://www.researchgate.net/post/How_do_you_define_unpolarized_light_What_is_the_difference_between_polarized_light_and_unpolarized_light

 

[Goodies]

The error with your analogy is that, after they are entangled, if you replace one glove with its opposite without actually measuring it (for example, we have an automatic mold machine within each box that creates the opposite of the glove inside of one box and the original was destroyed), the other glove would have to change automatically. In this case, we can assume that the "entangled" gloves right and left can be considered the particles' spin. When we change one, the other MUST change as a result. There's just fundamental differences between classical and quantum mechanics.

 

[DrChinese]

All the equation needs is two polarizer angles. I don't see how can that be anything but "prior condition".

As I indicated in an early post in this thread, photons can be entangled AFTER they are measured. Clearly prior condition is not possible in this light.

 

[Jabbu]

That's the definition of correlated.

After photon A goes through polarizer A, how can photon correlation make photon B go through polarizer B with 100% chance? What does photon correlation have to do with polarizers?

Why is photon entanglement not part of the equation?

 

[Jabbu]

The result cannot be determined before make the measurements.

Set polarizer A to -30, polarizer B to +30 degrees, and by doing so you will predetermine the outcome to be 25% correlation, every time. What result and what measurements are you talking about?

 

[Jabbu]

As I indicated in an early post in this thread, photons can be entangled AFTER they are measured. Clearly prior condition is not possible in this light.

Photons can be measured without being absorbed/destroyed? Where did you read that?

 

[microsansfil]

Set polarizer A to -30, polarizer B to +30 degrees, and by doing so you will predetermine the outcome to be 25% correlation, every time. What result and what measurements are you talking about?

Made a bet on your salary with this way of thinking, in this quantum context, and you will understand what I'm talking about.

Patrick

 

[stevendaryl]

After photon A goes through polarizer A, how can photon correlation make photon B go through polarizer B with 100% chance? What does photon correlation have to do with polarizers?

Why is photon entanglement not part of the equation?

Even though polarizing filters in QM work similarly to the way they work classically, the details seem very different.

Classically, if you have an electromagnetic field falling on a filter, you can think of what happens like this: Let \vec{E} be the electromagnetic field of the light. You can write this as a superposition of two different fields: \vec{E}_{||} which is parallel to the filter, and \vec{E}_{\bot}, which is perpendicular to the filter. The perpendicular component is absorbed, so the part that passes through is just \vec{E}_{||}

Now, if you drop the intensity low enough that you start seeing individual photons, then things start looking very different. A photon is not partly absorbed. It's either absorbed completely, or it passes through unchanged. So it seems as if every photon is either polarized perpendicular to the filter, or is polarized parallel to the filter.

What's special about entanglement is that the twins photons have the same polarization state. So if Alice's photon passes through her filter, which is at angle 30°, say, then it's as if the photon was always polarized at angle 30°. And Bob's corresponding photon acts as if it were always polarized at angle 30°. So if Bob's filter is at 30°, his photon will also pass.

I say "as if", because the photons did not have a definite polarization state before they were detected.

 

[atyy]

Can you point any document on the internet where we can see samples of actual experimental data?

Here's some actual data:

http://arxiv.org/abs/quant-ph/9810080
Violation of Bell's inequality under strict Einstein locality conditions
Gregor Weihs, Thomas Jennewein, Christoph Simon, Harald Weinfurter, Anton Zeilinger

 

[Jabbu]

What's special about entanglement is that the twins photons have the same polarization state. So if Alice's photon passes through her filter, which is at angle 30°, say, then it's as if the photon was always polarized at angle 30°. And Bob's corresponding photon acts as if it were always polarized at angle 30°. So if Bob's filter is at 30°, his photon will also pass.

What if photon A doesn't pass, what is then preventing photon B to go through polarizer B?

 

[Nugatory]

After photon A goes through polarizer A, how can photon correlation make photon B go through polarizer B with 100% chance?

it cannot and it does not. What the correlation says is that if photon A goes through polarizer A, then we know something about the probability that photon B will or will not go through the polarizer.

We can directly measure this, by watching what happens to a large number of entangled photon pairs. That's an experiment, and it's been done by Alain Aspect and many others.

Or we can calculate what should happen in such an experiment. We can do this using the Rules of Quantum Mechanics (in which the entanglement appears explicitly in the wave function for the system) and get one prediction for the correlation, or we can get a different prediction using classical rules in which there is no entanglement. One calculation matches the experimental result, and the other does not, so we know that which one is right.

But after all of this, we still don't know how this correlation came to be. We just know that quantum mechanics predicted it correctly and classical mechanics did not. Bell's theorem further tells us that any calculation that does not assume that the setting at A can influence the measurement at B cannot make the correct prediction - but it says nothing about the nature and mechanism of the influence.

 

[Jabbu]

it cannot and it does not. What the correlation says is that if photon A goes through polarizer A, then we know something about the probability that photon B will or will not go through the polarizer.

I think stevendaryl provided very sensible answer, as far as QM answers go anyway, but it's a little bit incomplete. We are talking about the case where polarizers are aligned at the same angle, like this:

theta_A = +30, theta_B = +30
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

...so if distance to polarizer A is shorter than to polarizer B, then photon B must always do the same thing photon A did.

 

[billschnieder]

Experiments are far less spectacular than Alice and Bob adventures. I don't see why make up stories when we can describe actual experiments. In the experiment there is a photon A and polarizer A on one side, and on the other side there is a photon B and polarizer B. Photon A will try to pass through polarizer A, and photon B will try to pass through polarizer B. If both manage to pass or if both fail we record '1', it's a match, and if one goes through but not the other we record '0', it's a mismatch. This is repeated with 10,000 more photons, the number of matches and mismatches are compared and then somehow interpreted to imply all kinds of crazy stuff.

I'm not impressed. The result is so very indirect and only vaguely related to what is being inferred from it. There is Malus's law in classical physics which can calculate probability for a photon to pass through a polarizer. Can it be demonstrated the outcome of the experiment in not predetermined by the angles set on the polarizers and Malus's law before the experiment even begins?

I should highlight that in actual experiments it is impossible to know that a photon did not go through, let alone that both of them did not go through. The typical setup is usually quite different than the (0,1) values being discussed. Typically you have a beam-splitter with two arms at each station. One of them labelled +1 and other labelled -1. a (+1,+1) or (-1,-1) result is a match and a (+1, -1) or (-1, +1) result for each pair is a mismatch. The way correlatiosn (actually expectation values) are calculated in the experiment is also quite different from the equations being discussed. They are calculated as <AB>, ie the average of a product of results on both sides for the given pair of angular settings. It is this <AB> value that matches the QM expectation value.

But it is even worse than that. Experiment do not give you pairs. Rather, you have a random series of time-tagged +1/-1 results at Alice, and another random series of time-tagged +1/-1 results at Bob corresponding to when the various detectors clicked. Then after the experiment you try to find pairs of clicks close enough in time which you *assume* belong to the same pair. Any unpaired value is discarded. There are no one sided (or does not pass through) values in the calculation.

 

[Nugatory]

I think stevendaryl provided very sensible answer, as far as QM answers go anyway, but it's a little bit incomplete. We are talking about the case where polarizers are aligned at the same angle, like this:

theta_A = +30, theta_B = +30
A: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0
B: 1 0 0 1 0 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0

...so if distance to polarizer A is shorter than to polarizer B, then photon B must always do the same thing photon A did.

Yes, and that's consistent with what I said. If the photon goes through polarizer A, then we know something about the probability that photon B will go through its polarizer. In this particular case, we know that the probability is 100%.

 

[Jabbu]

Yes, and that's consistent with what I said. If the photon goes through polarizer A, then we know something about the probability that photon B will go through its polarizer. In this particular case, we know that the probability is 100%.

In that case the same question for you: what if photon A doesn't pass, what is then preventing photon B to go through polarizer B? And if those photons are really unpolarized or "undefined" upon emission, then why don't we simply make them polarized first and then see what is really going on and how it actually works?

 

[DrChinese]

Here's some actual data:

http://arxiv.org/abs/quant-ph/9810080
Violation of Bell's inequality under strict Einstein locality conditions
Gregor Weihs, Thomas Jennewein, Christoph Simon, Harald Weinfurter, Anton Zeilinger

I gave that reference already, but I don't believe Jabbu is reading those. Also gave the Dehlinger reference on Bell tests which is one of the best tutorials.

http://arxiv.org/abs/quant-ph/0205171

 

[Jabbu]

Here's some actual data:

http://arxiv.org/abs/quant-ph/9810080
Violation of Bell's inequality under strict Einstein locality conditions
Gregor Weihs, Thomas Jennewein, Christoph Simon, Harald Weinfurter, Anton Zeilinger

Are you referring to the graph? I'm talking about raw binary streams data. Apparently there used to be some raw data from the Innsbruck experiment of 1998, available at:

http://www.quantum.univie.ac.at/research/bellexp/data.html

... but it's not there anymore, and that's the best I could find. I don't understand why are all those papers published without actual data obtained, I thought that's supposed to be obligatory.

 

[Jabbu]

I gave that reference already, but I don't believe Jabbu is reading those. Also gave the Dehlinger reference on Bell tests which is one of the best tutorials.

http://arxiv.org/abs/quant-ph/0205171

Does any of those papers contain or link to actual raw data samples?

 

[DrChinese]

Photons can be measured without being absorbed/destroyed? Where did you read that?

I don't think you followed my statement. They can be entangled AFTER they are observed. They no longer exist. That is because time ordering is very different than you might expect. I gave this reference previously as well.

http://arxiv.org/abs/quant-ph/0201134

And in fact photons can be entangled that have never co-existed or even been in one another's time cone:

http://arxiv.org/abs/0809.3991
http://arxiv.org/abs/1209.4191

It is more correct to consider the entire context of an experiment, from setup to detection, when determining predictions. Ie future setting are relevant to the statistical prediction. This violates normal everyday views on things, but is fully in keeping with QM.

 

[DrChinese]

Does any of those papers contain or link to actual raw data samples?

No, the raw data is very difficult to decipher and is not in a format that can be readily analyzed. It would almost be easier to do the experiment yourself. Of course, there is no real reason to see the raw data unless you simply don't believe the results. The experiments are described in plenty of detail for those who are interested.

Please keep in mind that these experiments involve a lot of elements you will need to understand first. For once, billschnieder has said something correct when he states that "typical setup is usually quite different than the (0,1) values being discussed".

However, I would strongly urge you to ignore ANY other comment he makes UNTIL you understand Bell tests better. He has a very non-standard agenda and will lead you to a place where it will be impossible for anyone to assist you.

 

[DrChinese]

I don't understand why are all those papers published without actual data obtained, I thought that's supposed to be obligatory.

I have never seen raw data published and I have probably read 1000+ papers.

 

[stevendaryl]

I should highlight that in actual experiments it is impossible to know that a photon did not go through, let alone that both of them did not go through. The typical setup is usually quite different than the (0,1) values being discussed. Typically you have a beam-splitter with two arms at each station. One of them labelled +1 and other labelled -1. a (+1,+1) or (-1,-1) result is a match and a (+1, -1) or (-1, +1) result for each pair is a mismatch. The way correlatiosn (actually expectation values) are calculated in the experiment is also quite different from the equations being discussed. They are calculated as <AB>, ie the average of a product of results on both sides for the given pair of angular settings. It is this <AB> value that matches the QM expectation value.

But it is even worse than that. Experiment do not give you pairs. Rather, you have a random series of time-tagged +1/-1 results at Alice, and another random series of time-tagged +1/-1 results at Bob corresponding to when the various detectors clicked. Then after the experiment you try to find pairs of clicks close enough in time which you *assume* belong to the same pair. Any unpaired value is discarded. There are no one sided (or does not pass through) values in the calculation.

That's a very good point. I think there have been attempts to explain EPR using local means by exploiting the differences between actual experiments and the idealization presented in most theoretical discussions of Bell's Inequality. For example, I think that someone named "Dereiter" or something like that? The idea is that Alice and Bob don't necessarily always measure the corresponding photons. If you assume that the likelihood of getting a mismatch is correlated with their filter settings, then maybe it's possible to reproduce the QM predictions for EPR.

I'm not sure if all such loopholes have been closed by experiments, but it's a little puzzling to think that errors in interpreting data would just happen to reproduce the predictions of QM.