When scientists use a taxon name like Mammalia, it is important that they talk about the same thing. But, what does it mean to be the same thing in different phylogenetic hypotheses? And, how is taxonomic reference maintained across hypotheses? Here, we discuss the differences between real and hypothetical clades, and how such a distinction relates to the sameness problem. Since hypotheses influence how we perceive things and pursue science, we find it important to have a functioning nomenclatural system for clades as perceived in phylogenetic hypotheses. As a solution to the sameness problem for such clades, we argue that a taxon name does not primarily refer to a single clade that somehow mirror the reality of branches in the tree of life. Instead we suggest that a taxon name refers to a set, or natural kind, of counterfactual and reconstructed clades.