On the Naturalness of Phylogenetic Classification