Well, it probably will, over a million year time scale -- but it's not going to go far, or fast, in solid glass unless something pretty strong is driving it (like the impact of a fast neutron or fission fragment); with lenses only 50 years old, where the index of the thorium glass hasn't changed enough to bollix up the prescription, the anions won't have moved enough to be a big problem. Wherever the anion lands after the initial event that forms it, is pretty much where it will stay. And since there are electrons throughout the glass, it's pretty easy to redistribute those a bit to match the anion up with *some* electron and then find another one elsewhere -- and so forth, net result being you reduce the overall concentration of color centers and the glass gets clearer again.