While it has long been observed that hypothetical changes in the speed of light would be difficult to measure if certain other quantities changed with it in lockstep (and by "long" I mean 100+ years, going all the way back to the original debates around relativity), it is still meaningful to ask if the speed of light has changed relative to the quantities that it seems to related to.
If that turns out to be the case, it will mean that our definition is wrong, not that the speed of light isn't changing. Our definition is a constant because we believe it to be a constant. If we're wrong, it will need to change.
I could define Pi as 3 and it would be CONSTANT. We'd then argue over what shape a circle looked like, but my maths would certainly be easier than yours due to lack of irrationality.
But it does, it falls out of the geometric relationship between the three space dimensions and the time dimension. In a way, the speed of light can be thought of as the speed of time.
Think of the relationship between space and time as orthogonal, as a right angle -- which makes sense, since the time dimension is at right angles to the three space dimensions. If you move quickly through space, you can't also move quickly through time, and the relationship between your time and space velocities is just what you would expect for an orthogonal relationship:
t' = t √(1-v^2/c^2)
t = time at rest
t' = time at velocity v
c = speed of light
Einstein wrote the above in 1905, then his math teacher (Minkowski) read Einstein's paper, saw the above equation, and realized it meant time was a fourth dimension. Minkowski then famously said, "Henceforth space by itself, and time by itself, are doomed to fade away into mere shadows, and only a kind of union of the two will preserve an independent reality."
In response, Einstein said, "Since the mathematicians have invaded the theory of relativity, I do not understand it myself anymore."
But from that point to 1915, Einstein acquired much more math knowledge and used it to write the General theory.
Actually that argument was had and ended a long time ago; you can set pi to 3 and you get a thing called a spherical geometry. See, for instance, http://mathforum.org/library/drmath/view/55021.html . Between the word "spherical geometry" and what you find in that link you'll have the keywords to continue digging if you want to.
Perhaps the relevant point is that the speed of light is not a number, but a dimensioned (if that's a word) quantity; and now the question becomes: is the dimensioned quantity (either its numerical value, or the meaning of the units) changing? I agree that it's hard to know how to make sense of the notion of units changing, though.
t' = t √(1-v^2/c^2)
That's pretty clear and easy to interpret. So measurements of c may be messy and subject to controversy, but the geometry expressed in the above equation, and what it tells us about spacetime, is much more clear.