"Any approach to scientific inference which seeks to legitimize it and answer in reponse to complex uncertainty is, for me, a totalitarian parody of a would-be rational learning process."