In quantum renormalization, one distinguishes between the (unobservable) bare electron and the observable properties. The justifications that I can find for the procedure seem to rest purely on mathematical and empirical grounds: the methods are in line with a consistent mathematical theory, it can survive different measurement scaling, and it gives the right answers. However, long, long ago (so long ago that I cannot find a reference to this), I read an attempt to justify the methods which said that the bare electron increased the probability of virtual positive particles forming (and disappearing) around it, so that the on-the-average positive shell reduced the charge of the bare electron when we measure it outside of this shell. The fact that I have not read this elegant explanation recently leads me to suspect that it has been discredited, despite its charm. (Of course, I don't know how this would be adapted for gravitation, but probably no one else does as well, so that's OK, innit?) Has it? If it hasn't, why doesn't it appear? If it has, is there any other known physical reason for the individual steps of methods of renormalization to work?