From Multivariate Statistics to Natural Selection: A Reanalysis of the Plio/Pleistocene Hominid Dental Material

Multivariate statistical techniques have had a natural application to the problem of taxa definition in hominid studies. The fundamental problem in taxonomic reconstruction of, first, sorting individual fossil specimens into homogeneous, distinctive groups and, second, of determining the taxonomic relations implied by those groupings, has a format immediately translatable into the framework of techniques such as multivariate cluster algorithms for sorting data into groups, discriminant analysis for measuring distinctiveness of groups, and factorial procedures for forming construct variables that are sensitive to morphological complexes signifying species uniqueness. Yet despite an increasingly sophisticated conceptual and analytical framework for the study of hominid fossil materials, establishing a generally agreed upon taxonomic classification for those materials has remained elusive, with alternative classifications apparently difficult to confirm or disconfirm in any generally accepted fashion as indicated by still current controversy over hominid classification.