In 1905 it became apparent that the speed of light was independent of the source (which was, and still is well established by experiment). The idea that gives relativity its special significance is that it claims that the speed of light is independent of the speed of the observer. That is, all observers regardless of their relative velocities will measure the speed of light to be the same.

In the Sagnac experiment, light is split and sent in opposite directions around a rotating plateform (using appropriately placed mirrors). The beams are then combined and by interference a very accurate measurement of phase can be made.

Now classically since the speed of light is independent of the source (but not independent of the observer), the light travelling against the rotation is received such that:

(1)      c₁=c+v

where v is the velocity of the rim. The light travelling with the rotation is received such that:

(2)      c₂=c-v

This assumes that the center of rotation is at rest, though it is actually general since, if the center of rotation was in motion relative to an absolute frame, differences in the flight time of light based on this would be cancelled out due to the closed paths.

Now, if light speed is independent of the observer, it is trivial that the results of the experiment should be:

(3)      c₁=c₂=c

a prediction that *every* relativist would have made before 1913. As it turned out, the Sagnac experiment confirmed that Eq.(1) and Eq.(2) were, in fact, correct, supporting the classical concepts of an absolute rest frame (as Sagnac understandably claimed).

After the experiment was performed it was proposed that since the observer was in a rotating frame (non-inertial) Eq.(3) was not required.

However, imagine that the experiment is done at radius 'r' which yields an acceleration (non-inertialness) of a=v²/r. At this radius the difference in observed light speeds is '2v'. Now double the radius so that the acceleration is a=v²/(2r). At this radius the difference in observed light speeds is still '2v'. The difference is completely independent of the acceleration. Thus by simple logic it can be seen that when a=0, the frame thus being inertial, that the difference is still '2v'. Therefore, Sagnac's original assessment was correct that the speed of light is not independent of the observer.

One might argue that in a calculation from the lab frame relativity would indeed predict the anisotropy, which is probably true, but strictly by relativity, the observer on the rim must not see this anisotropy. Furthermore, since the information is being directly measured and instantaneously shared with the lab frame, the viewpoint of the rim is actually being observed in the measurement. This also leads to a contradiction since both viewpoints cannot exist simultaneously in this experiment.

It can thereby be deduced that the theory of relativity is wrong.