An Evaluation of Cladistic and Character Analyses as Hypothetico- Deductive Procedures, and the Consequences for Character Weighting

Abstract

Although phylogenetic analysis based on synapomorphies is said to comply better with the hypothetico-deductive methodology advocated by Popper and other philosophers of science than other approaches to phylogenetic reconstruction, the role of parsimony in cladistic analysis is not consistent with that of the hypothetico-deductive method. Parsimony evaluates the form of a hypothesis but cannot decide on the relative support for hypotheses. Cladistic analysis is not a hypothetico-deductive procedure for testing hypotheses; it is an inductive procedure that summarizes the information from character analysis (the synapomorphies) and not a separate level in the hierarchy of deductively testable hypotheses of phylogenetic analysis. Synapomorphies are not hypothetico-deductive tests of cladistic hypotheses. In contrast, the testing of hypotheses of synapomorphy during character analysis appears to conform to hypo- thetico-deductive principles. Thus, hypotheses of relationship can be tested deductively only during character analysis. The use of parsimony to inductively generate cladograms is justified but it is unclear which of various procedures based on parsimony is the most appropriate. Because incongruence among synapomorphies can be resolved only by character analysis, emphasis should be placed on the most testable characters as in Neff's (1986) a priori method of character weighting that is based on testing and corroboration. (Character analysis; character weighting;