EPR revisited

[anuttarasammyak]

Observers in different inertial frames who were making observations in the vicinity of measurement at A hold a meeting and present their respective viewpoints. It turns out that each of them is making a rational judgment. As a conclusion of the meeting, they adopt the earliest possible reading of B’s clock, T−L/c where L is distanbe between A and B, as the common consensus. This is because B’s behavior corresponding to clock readings between T−L/c and T+L/c is unknown, and measurement at B may occur during this interval. The only choice consistent with this possibility is the earliest time, T−L/c.

Here, I assume that all inertial frames agree that the collapse occurs when the reading of clock B is the smallest, namely at T−L/c.
This is consistent with the viewpoint of each inertial frame in the sense that, in every frame, B is in the collapsed state only afterward (although the observation itself is performed at A).

Next, let us consider the situation from B toward A.
If we take the collapsed state of B as the starting point, then, due to the relativity of simultaneity, in some inertial frames the state of A appears to have collapsed at a time earlier than the observation. The earliest such time is when the reading of clock A is T−2L/c. Let us again assume that this is accepted by all inertial frames as well.

If we continue this kind of reasoning back and forth between A and B, tracing further and further into the past, we arrive at the conclusion that, ever since the moment when A and B were at the same location for the purpose of generating quantum entanglement, the states of A and B had in fact already collapsed.
What was regarded as an entangled state up until the observation by A is, after the observation, reinterpreted as having been a collapsed state extending retroactively into the past, back to the time of entanglement generation.
In this way, it seems that there is no longer any need to worry about spooky long-distance correlations.

 

[PeterDonis]

If we continue this kind of reasoning

We can't because it's giving you contradictions. First you assume the collapse takes place at a certain event on B's worldline; then you assume that the collapse takes place at a certain event on A's worldline; and then you conclude that the collapse takes place at a different event on B's worldline. What that tells you is that "this kind of reasoning" is not valid. As I already pointed out to you in post #14.

 

[Nugatory]

What was regarded as an entangled state up until the observation by A is, after the observation, reinterpreted as having been a collapsed state extending retroactively into the past, back to the time of entanglement generation.

The collapsed state is a mixture of $|+-\rangle$ and $|-+\rangle$; the uncollapsed state is a pure superposition of the two. These are different states that behave differently in experiments designed to expose the difference. What does it mean to “interpret” one as the other?

 

[PeterDonis]

The collapsed state is a mixture of $|+-\rangle$ and $|-+\rangle$

I would say that the collapsed state is either $\ket{+-}$ or $\ket{- +}$, whichever corresponds to the measured results. We don't use the mixture state to predict further measurements; we use whichever of the two kets corresponds to the result that was measured. That's what the collapse postulate says to do.