I'm not sure. Depends partly on what one means by 'modern science.' Caloric theory and Bode's Law have been debunked, but were legitimate science with evidence to support them.

I think the problem with the characterization isn't whether science can overturn old ideas. Whether we view it as an overturning or a refinement seems a semantic argument to me. The thing is that even when science has overturned old ideas, that hasn't been its goal. There's nothing in science that says, "Everybody's too comfortable, we have to switch things around a little." Rather, things have had to change (overturn or refine, I dont' care which phrase is used) because while the existing theory seems to explain X and Y very well, it's a little weak on Z.

k