I have a charged insulating marble -- let's say I used an electrostatic generator to get it up to 1 coulomb, and it weighs 1g (don't make me say 9.8mN). I have it in an insulating "box" whose walls and ceiling are very far away. I have it sitting still for a long time, so its E field is ~1/r^2, and since it's perfectly still, the outward forces in all directions balance out (and apparently aren't as strong as the rigid forces keeping the marble together), so it stays put. But then, my cat jumps on the table and jostles the box -- so the marble moves ever-so-slightly to one side (say, the x direction). So at that instant, the charge is slightly off-center, but it takes a finite amount of time for the field to update: so there is now a net force in the x direction (since E=1/x^2, and x<<1, it is even rather large), causing it to accelerate more in that direction, and the process repeats: this creates a current in the x direction, which in turn creates a B field in the Z direction (eventually), which would cause an acceleration in the y direction, so there is no "braking" force to be found. The marble accelerates to 0.9999C and blasts through the side of the box, nearly crippling the cat. Now, what *should* have happened? Why don't we observe this with every perturbed charge? As far as I can tell, this is the accepted answer for classical calculations that include self-fields, even after renormalization (See http://en.wikipedia.org/wiki/Abraham–Lorentz_force, or Dirac's paper on classical self-fields).