Well, part of what motivated the question was having an experience of doing just that and later having the filesystem go bad. But that's no proof that the fsck was the cause of the degradation; in fact I doubt it just because I can't see how a purely passive operation would introduce errors. The comment about the clean bit shows that an fsck can have an effect, but even that argument doesn't show how it's a harmful effect. I was running the fsck just to confirm that a filesystem I believed to be OK (I had just created it with a "cp -a") was really OK. Admittedly I was being lazy by not unmounting the filesystem, but I couldn't see why I would need to.
The reason to do it really is any activity on the filesystem will be seen as corruption. As far as it ruining the system, I feel like it would but the more I think about it, you're correct. It would be a passive operation which would incorrectly report corruption on your desk, but wouldn't corrupt the disk itself. Which means this has really become academic since running fsck on a live filesystem wouldn't give you an accurate report anyway. (Althouhg I'm curious. When I get home I'll try it on my RH laptop and see what happens). Ben