From a EE.
When the the R/R shorts the windings together some of that that energy goes to create a magnetic field that opposes the field of the fixed magnets in the rotor thus reducing the total energy generated. Some of that energy however turns to heat in the wire.
The stator wires are covered in a very thin insulation coating so that more wires can be wound on. They are wound around iron cores stamped out of sheet steel (with sharp edges) so the windings are square not circular. As the copper heat up and cools down it stretches and shrinks at a different rate than the iron it is wound on so there is some rubbing which will concentrate on the corners even though the whole assembly is coated in an epoxy or similar encapsulant.
As soon as a corner gets worn down to the point that the wires touch it is all over. Once you have a single "shorted turn" like that all the energy goes to that one loop causing the wire to heat up like a stove top, shorting other turns together and propagating until that section of the stator blows like a fuse. Hence the pictures of just one section burned up on the failed stators with adjacent sections showing some signs of heating.
Transformers do the same thing when they overheat. The big power ones on poles go off like fireworks if a lightning strike's high voltage causes the insulation to fail, big purple flash and a shower of sparks.
Better quality insulation, careful winding and good protection of the iron core all help.
I suspect the original stators were built to a certain quality level that balanced cost and performance. I also suspect the replacements out there are built to an even lower price, this would explain the serial stator failures after a replacement. In a competitive market you can generally assume you get what you pay for, a lower price means they cut a corner somewhere. Likely they used thinner, lower temperature or poor quality insulation on their wires which is nearly impossible to detect and wont show up immediately in use.
The R/R has a lot to do with this too, my original failed short circuit completely, it was letting AC get to the battery and DC get to the stator, neither of which are good as it causes more destruction. Again they are designed to a price and to last as long as the warranty. A modern MOSFET R/R has very different failure modes as well as reduced power dissipation so is better all round.
There is nothing fundamentally wrong with the design, its crude but effective. If Ma KAW held their suppliers to a slightly higher standard or made the stator easily changeable we would not be having this discussion. As it is you have a weaker then normal component that is not easy to replace, gives us something to talk about.
'99 Vulcan "Victoria"