I can't prove it, but I'm pretty sure that people gain a selective advantage from believing in things they can't prove.