Re:No hints about c (Score 2, Informative)

The speed of light in a vacuum cannot change, by current definition. The meter is currently defined as 1/299,792,458 of the distance light travels in a vacuum in one second. Of course that hasn't always been the definition: the basis for the meter has been, in historical order:

(1) The quadrant of the Earth.

(2) The length of a particular metal bar.

(3) The wavelength of a particular atomic spectral line.

(4) The speed of light and the frequency of an atomic clock.

Each change improved the reproducibility of the best length measurements, given the technology at the time the change was made.

The observations that suggest alpha varies are based on comparing wavelengths of light from different atomic oscillations, potential distance standards similar to (3) above. They appear to vary relative to each other. If you want to attribute this to varying c, which one is the reference yardstick?

For our present technology the most reproducible clocks and yardsticks are atomic oscillations. If these lack relative constancy and you choose the frequency of one as your time standard and the wavelength of another as your length standard, you will apparently observe a changing value of c. However, the direction and magnitude of the change will depend on which pair you choose. If we had really independent distance and time standards (and it was clear which was which) it would make sense to consider c an experimental quantity. Since we don't we have just chosen one standard (a particular oscillation of cesium), and c is a defined constant.

Similarly, the electromagnetic quantities "epsilon nought" and "mu nought" were once experimental quantities, but are now by definition exactly {10^7/(4 Pi c^2)} and {4 10^-7 Pi} respectively. This means that the coulomb is no longer defined electrochemically: it is a derived unit, not a fundamental one in the SI system.