Trying to clear up some confusion (clock synchronization)

[Dale]

Look. It is a good assumption, and it is silly to make any other assumption. But simply because an assumption is a good one doesn't make it not an assumption.

Here are the invariant facts:
A light pulse is emitted from side A when the pointer reaches side A
A light pulse is emitted from side B when the pointer reaches side B
Both light pulse are received at the same time in the middle

Those three facts are compatible with many explanations, not all of which include light going at C.

 

[rede96]

You're arguing in a circle. If you want to prove that the pulses are sent simultaneously, you can't assume that they are sent simultaneously.

In your case, an example of how tricky this is, is you say something like: I have pointer at rest, that is measured straight, and tips are equal distance from the pivot. I apply torque to the center. Without assuming isotropy, you cannot assume that the torque, leading to motion, propagates at the same speed towards booth pointer ends. It could be that motion reaches one tip earlier. You may say that you can detect this, but by anisotropic effects on how you propose to determine this, you can't. You have to come up with an experiment which cannot be reconciled with any conceivable anisotropic model. As I noted, that has actually been proven impossible, in that there is a known family of anisotropic models that are experimentally indistinguishable from standard SR.

Here are the invariant facts:
A light pulse is emitted from side A when the pointer reaches side A
A light pulse is emitted from side B when the pointer reaches side B
Both light pulse are received at the same time in the middle

Those three facts are compatible with many explanations, not all of which include light going at C.

Right, I think I understand this now. Basically what it boils down to is that any test of simultaneity between two events, involves sending signals to verify the events were simultaneous is some frame of reference. (Or signals to synchronise clocks etc.) And any signals sent from the events are sent in just one direction, in which case there is no way to test for isotropy, as that would also involve sending more signals.

I'm not sure I have worded that very well, but it seems to make sense.

The only thing I would say is that taking my thought experiment as an example, there is a limit as to how far out any delay in the ends triggering the light sources simultaneously can be. I can test this statistically and find the process variation using a number of different set ups. (e.g. different materials, different lengths, alternating the pointer orientations etc. ) It wouldn't tell me exactly of course, but I should be able to agree a certain limit within a certain confidence interval.

Anyway, thanks again to everyone for their time, as always it is very much appreciated.