In this paper I analyze a contentious debate between proponents of coalescence methods and proponents of concatenation methods for phylogenetic inference. First, I clarify claims that phylogenetics can be reduced to genetics, and that concatenation can be reduced to coalescence. Concatenation can be mathematically reduced to a special case of the coalescent model by specifying parameters in the coalescent model to match empirical assumptions made by concatenation methods. Concatenation proponents claim that empirically realistic application of coalescent methods will amount to concatenation, but these claims have not been adequately supported. The shape of this debate reiterates earlier arguments that resulted in the mathematical reduction of parsimony to maximum likelihood. Proponents of concatenation appeal to many of the same philosophical claims that were used to justify parsimony, while coalescence models are a natural extension of the maximum likelihood framework. In light of the historical development of methods of phylogenetic inference, these debates can be interpreted as conflict points between competing paradigms. The asymmetry in arguments about reduction suggests that concatenation (an outgrowth of total evidence) is a paradigm in decline.
Concatenation, Coalescence, and Concatalescence:
Conceptual Disputes in Phylogenetic Inference