I guess my tendency is to think essentially that the new wrinkles won't do the job if the old major idea didn't, and so you have to try something different. Then maybe they can all be combined in some coherent piece.
I think philosophers can do things akin to theoretical scientists, in that, having read about empirical data, they too can think of what hypotheses and theories might account for that data. So there's a continuity between philosophy and science in that way.