Now that I've read it, it seems the author missed an option: the criterion of evolvability. Strange, given that he discusses the diachronic method right at the beginning.