Linguistics, Genetics and Archaeology: Internal and External Evidence in the Amerind Controversy

Greenberg (1987a) argues, on the basis of his technique of multilateral comparison, for a classification of New World languages into three groups, including the ‘superfamily’ Amerind. Greenberg and his collaborators defend this classification using linguistic, genetic and archaeological evidence. In this paper, we assess the validity of all three types of evidence, and pursue the consequences of our assessment for multilateral comparison, for Greenberg's tripartite classification, and for the unity of the ‘Amerind’ phylum. We aim to show that the external evidence Greenberg cites is either not truly independent, or does not strongly support his conclusions, and that his linguistic methodology is both unreliable and statistically intractable. The last temptation is the greatest treason: To do the right deed for the wrong reason. (T. S. Eliot: Murder in the Cathedral)