Entangled "information" sent faster than light (....not)

[Micheth]

Hello^^

It's always interesting to ponder the various ways that quantum entanglement *appears* to allow transmission of faster-than-light information, but on closer examination, actually doesn't.
One that occurred to me today was the situation of two persons who receive each of a pair of entangled photons.
Just to make it more "interesting", suppose they live on opposite sides of a galaxy with a black hole in the middle, and since they cannot travel directly toward each other, their task is to each set out in a particular direction in order to meet each other.
With the photon they each receive (by conventional means), they are given instructions to set out at a 45deg angle from the imaginary line connecting them, the direction to take being to their right if their photon's measured spin is up, or to their left if their photon's measured spin is down.
Obviously if they take "opposite" paths (one to his right, the other to his left), they will meet, whereas if they both go to their respective right, they will miss each other by the diameter of the galaxy.

So, the sticky part is of course that before they measure their photons, they have no idea which way to go, and neither do the photons know which spin they have, presumably. (in other words, they as yet have no definite spin).
(I believe all of the above is correct, maybe the phrasing is somewhat off)

But once one person measures his photon (up, let's say), he knows to take off in a rightward direction, but according to QM, at that moment the other photon now knows to have a down-spin, and when the other persons measures it, he'll now know to take off in his leftward direction, which will intersect with the first person.

I don't really have an definite way of explaining why it's not faster-than-light communication, except to propose that the info was really already there in the instructions, which were transmitted conventionally.

But there still is this nagging sense that "something" was transmitted when one person measured their photon, because otherwise, had the other person not also performed his measurement and discovered which direction to embark upon, the two might not have ever been able to meet, whereas *with* that crucial information, they *were* able to meet. Two very different outcomes.

It's just another thought experiment, but any comments would be interesting to read.
Cheers,

 

[mfb]

except to propose that the info was really already there in the instructions, which were transmitted conventionally.

That is the point.
You can do the same experiment with written letters and a purely classical rocket delivering them.

 

[Micheth]

That is the point.
You can do the same experiment with written letters and a purely classical rocket delivering them.

The difference with that case being that the written letters in the rocket are determined beforehand, right?
Whereas the spin of the photon on the other side of the galaxy hasn't yet been determined, so that information hasn't yet been "transmitted".

 

[Derek Potter]

Hello^^
I don't really have an definite way of explaining why it's not faster-than-light communication, except to propose that the info was really already there in the instructions, which were transmitted conventionally.
But there still is this nagging sense that "something" was transmitted when one person measured their photon, because otherwise, had the other person not also performed his measurement and discovered which direction to embark upon, the two might not have ever been able to meet, whereas *with* that crucial information, they *were* able to meet. Two very different outcomes.

You are quite correct. Information is being transmitted faster than light. However it is not transmitted from one person to the other, it is transmitted from one photon to the other (various caveats apply). This is why entangled particles are said to "break local causality". There is no escaping it in a "realistic" model i.e. one which assumes that there are definite outcomes e.g. Bob and Alice do actually find out which way to go.
edit - note that the "information" is not information in the usual sense, it is !"whatever allows the two photons to make co-ordinated decision" or some such.

That is the point.
You can do the same experiment with written letters and a purely classical rocket delivering them.

Sure, but the scenario is not intended to show Bell inequality violation. It is intended to show FTL is inescapable if the quantum description of entanglement is correct.

Edit:

The difference with that case being that the written letters in the rocket are determined beforehand, right?
Whereas the spin of the photon on the other side of the galaxy hasn't yet been determined, so that information hasn't yet been "transmitted".

This is a big assumption. Entanglement can be seen in those terms and it's rather fashionable to talk about observation actually creating reality - a determined state. But one can equally say that there is always a determined state, it's just that the state may be superposition of more classical states. In those terms, the information is spread over both particles so detection at one place is always detection of the joint property. That is non-local rather than involving FTL, though FTL is required if the detection also gets rid of the superposition (aka "collapses the wavefunction"). One of the attractions of the Many Worlds Interpretation is that it doesn't have wavefunction collapse and therefore FTL is not required either.

 

[mfb]

Sure, but the scenario is not intended to show Bell inequality violation. It is intended to show FTL is inescapable if the quantum description of entanglement is correct.

That is exactly the point. You need Bell-type measurements to get clearly non-classical effects. If you just say "okay, we have two photons, we both measure their horizontal/vertical polarization" then the measurement result could be pre-determined and still in agreement with experiments. It is not pre-determined (in most interpretations), but this experiment is not sufficient to show it.

 

[Swamp Thing]

One of the attractions of the Many Worlds Interpretation is that it doesn't have wavefunction collapse and therefore FTL is not required either.

MWI would still have to explain how the very worlds where Alice got "heads" are exactly the ones in which Bob got "tails" -- even when they are space like apart.

 

[mfb]

There is nothing special to explain, it is a direct result of the evolution of the wave function and what we call "worlds" in that interpretation.

 

[stevendaryl]

That is exactly the point. You need Bell-type measurements to get clearly non-classical effects. If you just say "okay, we have two photons, we both measure their horizontal/vertical polarization" then the measurement result could be pre-determined and still in agreement with experiments. It is not pre-determined (in most interpretations), but this experiment is not sufficient to show it.

I had a similar disagreement a couple of weeks ago, and I'll make the same point again. I don't see why it is relevant that there exists a classical explanation for a specific experiment, if we already know that that explanation is not true (because of other experiments, such as EPR).

 

[Derek Potter]

I had a similar disagreement a couple of weeks ago, and I'll make the same point again. I don't see why it is relevant that there exists a classical explanation for a specific experiment, if we already know that that explanation is not true (because of other experiments, such as EPR).

Yes, and your argument was quite right and it still is. We could explain away these special cases with a classical mechanism. But as you pointed out, it doesn't make sense to say the system switches from entanglement to predetermination just for the special cases. Besides which, BI violation is proven for small angles so the OP's scenario could be modified to make use of some such small angle with the only difference being that Bob and Alice will then not be absolutely certain they will meet up and have to manage with 99.99% certainty or some such.

 

[mfb]

I had a similar disagreement a couple of weeks ago, and I'll make the same point again. I don't see why it is relevant that there exists a classical explanation for a specific experiment, if we already know that that explanation is not true (because of other experiments, such as EPR).

If you want to use an experiment to show a nonclassical property, then this experiment should not have an equivalent in classical physics.

 

[Derek Potter]

MWI would still have to explain how the very worlds where Alice got "heads" are exactly the ones in which Bob got "tails" -- even when they are space like apart.

Because the particles have interacted in a way that conserves spin. Hence, at the moment of interaction, there are two eigenstates, |up,down> and |down, up>, and the wave function is a superposition of these. In MWI the superposition is interpreted as two worlds, one relating to |up,down> and the other to |down, up>. So one world has Alice observing heads and Bob observing tails. And the other world, vice versa.

 

[stevendaryl]

If you want to use an experiment to show a nonclassical property, then this experiment should not have an equivalent in classical physics.

I don't see why that should be the case. Like I said, I don't see the relevance to pointing out that something COULD be explained by X if we know that X is not, in fact, the explanation.

 

[stevendaryl]

Here's a possibly useful, or possibly misleading analogy to the information transferred by entanglement in EPR-type experiments.

Suppose Alice has a device that works like this:

1. Alice enters a bit-string [itex]X[/itex], that is, a sequence of 0s and 1s, to represent a message.
2. The device generates another random bit-string [itex]Y[/itex].
3. The two strings are combined using the operator [itex]\otimes[/itex] to get a string [itex]Z[/itex]. The operator [itex]\otimes[/itex] acts on the corresponding bits according to:
   • [itex]1 \otimes 1 = 0[/itex]
   • [itex]1 \otimes 0 = 1[/itex]
   • [itex]0 \otimes 0 = 0[/itex]
   • [itex]0 \otimes 1 = 1[/itex].
4. [itex]Z[/itex] is instantaneously transmitted to Bob.

I think it's clear that if Alice has no way to predict the random string [itex]Y[/itex], then it is impossible to use this device to transmit FTL messages. Alice instantaneously knows what Bob will see, but she has no way to influence it.

To me, the EPR twin-pair experiment seems a lot like this. Alice makes a choice--the orientation of her detector. "Nature" makes a choice, namely which result--spin-up or spin-down in the spin-1/2 case. The combination of the two is what seems to influence Bob. If Alice chooses orientation [itex]\alpha[/itex] and gets spin-up, then Bob will definitely not get spin-up at orientation [itex]\alpha[/itex]. But Alice's choice alone has no influence on Bob; there is nothing she can do that makes Bob's result more likely to be spin-up or spin-down. So Alice cannot use the correlations for FTL communication.

 

[Derek Potter]

So Alice cannot use the correlations for FTL communication.

Well, she can but the information will remain encrypted until she sends her results to decode it. She could even send them FTL but they will remain useless until her second set of results is sent. And so on ad infinitum. There's always a master key left over that has to be sent by carrier pidgeon.

 

[mfb]

I don't see why that should be the case. Like I said, I don't see the relevance to pointing out that something COULD be explained by X if we know that X is not, in fact, the explanation.

Let me use an analogy: does a falling apple show that spacetime is curved? Let's ignore the philosophical part about reality and models.
We know curved spacetime gives a very good description for a falling apple. But so does Newton.

The experiment is in agreement with nonlocality and our best theory for the universe explains the observations with nonlocality (or other counterintuitive things depending on the interpretation), but based on this experiment alone we cannot conclude that anything is nonlocal.

 

[Derek Potter]

Let me use an analogy: does a falling apple show that spacetime is curved? Let's ignore the philosophical part about reality and models.
We know curved spacetime gives a very good description for a falling apple. But so does Newton.

The experiment is in agreement with nonlocality and our best theory for the universe explains the observations with nonlocality (or other counterintuitive things depending on the interpretation), but based on this experiment alone we cannot conclude that anything is nonlocal.

It's not intended to prove nonlocality. It's intended to prove that nonlocality entails FTL.