I find this mode of reasoning confusing and unilluminating. Alliance theory does not, for me, shed light on the problem which I consider to be fundamental. It does not concern itself with the mechanisms and processes of social structure or reckon with the politico-jural determinants of kinship, descent, and marriage relations. This surely is the gravamen of the criticism made by Homans and Schneider of LCvi-Strauss's theory of cross-cousin marriage. Whatever objections may be raised against their theory, it follows in the tradition of Radcliffe-Brown's structuralism in seeking for the mechanism and process behind the customary usage. Heuristic, interpretative, and explanatory schemes, whether we call them models, theories, or what you will, are indispensable to our study. But caution is needed. "The price of the employment of models", Braithwaite warns, "is eternal vigilance" (1953 : 93). For if our models drift too far away from the empirical facts, we run the risk of relapsing into the habits of mind that almost ruined ethnological research and scholarship in Morgan's day.