To count as improvements, however, such deviations from the human norm would have to be pointed in very particular directions rather than at random; and they would continue to presuppose the existence of a largely undisturbed anthropocentric frame of reference to provide humanly meaningful evaluative generalizations
