"Why not? There would be a cause for the initial impetus to change and a cause for the particular way of recombining.

That is what I am questioning."

Suppose we have a sequence that goes CGTAGGT on two identical DNA strands. Say we irradiate the strands and they both, er, mutate (yeah, I know...) One strand ends up as CGATGGT and the other ends up as CGTGAGT. We can strongly suspect that the radiation triggered the mutations (you could do controls of course to confirm that) but it's a bit harder to explain why they ended up different and not the same. Maybe we bent over backwards to ensure that they were irradiated identically and that all other conditions were identical for both. But therein lies the rub. It's impossible, to all intents and purposes, for either us or the environment to get identical conditions for both, when you think right down to subatomic particle level. All that jiggling and shooting about and jostling of atoms and molecules in the surrounding medium... The most reasonable position to take would be that they recombined differently because they were subjected to ever so slightly different environmental conditions. I don't think it's reasonable, just because we can't actually pin down the cause, to assume that was was no cause. I suppose that if you repeated the experiment thousands of times there's a chance that you'd find the numerical spread of the new combinations fitted some kind of statistical "randomness", but, even if that were so (and it might not be), it would elucidate precisely nothing about what was going on (as I probably haven't). It wouldn't show that there were no causes, that's for sure.