>PS I overlooked the error you made here - a confusion that is often >seen. As Eric points out, SRT is *not* (never was!) limited to >inertial states. It's the same as for Newtonian mechanics. SRT uses >Poincare's PoR, which refers to Newtonian reference systems; and we >can switch reference system whenever we like, using the LT. Thus the >twin problem is a trivial exercise in SRT. The twin paradox however >was aiming Einstein's *General* PoR.
I still don't know why you say that. Are you interpreting the general principle of relativity to be "absolute acceleration is not detectable", while the limited principle of relativity is "absolute velocity is not detectable"?
Depending on what you mean by "acceleration", the general principle may be true or false. If you mean *proper* acceleration, then of course that is detectable: Put a heavy mass on a spring. If the spring is uncompressed and unstretched, then there is no proper acceleration in the direction of the spring. Alternatively, you can set up a rectangular coordinate system with metersticks, and plot the trajectory of a light beam or a tossed ball: if the plotted path is curved, then the coordinate system is accelerating.
If by "acceleration", you mean *coordinate* acceleration, then whether you are accelerating or not is relative to whatever coordinate system you are using.
In any case, in what way do you think the twin paradox causes problems for the principle of relativity?
I think you missed the point of Einstein's dialog that you referenced. I don't see it as an admission that the twin paradox is a consistency challenge for his theory. I take it as a tutorial about his principle of relativity, and the twin paradox is a thought experiment that is used to highlight distinctive features of it.
