My understanding is that physicists generally accept the notion of locality, that a phenomenon cannot be affected by `spooky action at a distance.' But how has that be reconciled with EPR experiments? Here is an example that emphasizes the problem.
If I understand it correctly, vertical spin and horizontal spin are a pair of properties, like position and momentum, and are dual in the sense of the Heisenberg uncertainty principle. In the famous experiment, I take a particle with zero spin and cause it to fission, sending two sub-particles A and B off in different directions in such a way as to preserve the (zero) spin of the system. If I send particle A through a device that measures vertical spin, I will get an answer, `up' or `down'. If I subsequently make the same vertical measurement on particle A, I will get the same answer. Particle B will give the opposite answer (so the spin of the system is preserved.) But a vertical measurement destroys any information about horizontal spin; after a vertical measurement, a subsequent horizontal measurement has a 50-50 chance of yielding `left' or `right' as an answer.
Question 1: If I do a vertical spin measurement on particle A followed by a horizontal spin measurement on particle A, and then a horizontal spin measurement on particle B, am I certain to get opposite horizontal spin indications for particles A and B? (If not, why does this not violate the conservation principle governing spin?)
If so, it seems by induction that I could extend the above procedure to perform an arbitrary sequence of intermixed horizontal and vertical spin tests on particle A, and then perform a test on particle B with the same orientation as the last test on A. After doing so, I would get opposite spin indications from the two particles with probability 1.
If the above is true, it seems to imply that the quantum states of particle A and B remain tangled at a distance, even if I switch the horizontal spin of one of the particles back and forth by interposing vertical spin measurements. The particles cannot store an arbitrarily large amount of information, i.e., how to respond to arbitrarily long sequences of horizontal and vertical tests. But they respond in consistent ways with probability one to queries made when they are far apart.
Question 2: How is this reconciled with the notion of locality?
Slashdot Mirror
If I understand it correctly, vertical spin and horizontal spin are a pair of properties, like position and momentum, and are dual in the sense of the Heisenberg uncertainty principle. In the famous experiment, I take a particle with zero spin and cause it to fission, sending two sub-particles A and B off in different directions in such a way as to preserve the (zero) spin of the system. If I send particle A through a device that measures vertical spin, I will get an answer, `up' or `down'. If I subsequently make the same vertical measurement on particle A, I will get the same answer. Particle B will give the opposite answer (so the spin of the system is preserved.) But a vertical measurement destroys any information about horizontal spin; after a vertical measurement, a subsequent horizontal measurement has a 50-50 chance of yielding `left' or `right' as an answer.
Question 1: If I do a vertical spin measurement on particle A followed by a horizontal spin measurement on particle A, and then a horizontal spin measurement on particle B, am I certain to get opposite horizontal spin indications for particles A and B? (If not, why does this not violate the conservation principle governing spin?)
If so, it seems by induction that I could extend the above procedure to perform an arbitrary sequence of intermixed horizontal and vertical spin tests on particle A, and then perform a test on particle B with the same orientation as the last test on A. After doing so, I would get opposite spin indications from the two particles with probability 1.
If the above is true, it seems to imply that the quantum states of particle A and B remain tangled at a distance, even if I switch the horizontal spin of one of the particles back and forth by interposing vertical spin measurements. The particles cannot store an arbitrarily large amount of information, i.e., how to respond to arbitrarily long sequences of horizontal and vertical tests. But they respond in consistent ways with probability one to queries made when they are far apart.
Question 2: How is this reconciled with the notion of locality?