Systems, 451. Ch. Hose and Wm. McDougall, The Pagan Tribes of Borneo, I., 8o sq. Kroeber, Kinship in the Philippines, An throp. Papers, American Museum of Natural History, XIX., 1919 75 sq. E. M. Loeb, ms. on the Mentawei Islands.
For most of Oceania Rivers has given a convenient summary. Broadly, the Melanesians refrain from using personal names whenever there is a suitable term of relationship, while the Polynesians use personal names even for the closest relatives.
Of the Papuans, the Banaro, settled on a tributary of the Au gusta River, have been accurately described. The differentiation of the maternal uncle suggests a bifurcate merging system, but in other respects they have developed on generation lines, though anomalous matrimonial customs colour the terminology. An inter esting distinction is drawn between a paternal aunt who has and one who has not been exchanged for the speaker's mother : the former is called "mother," the latter "maternal uncle's wife." In the contemporary generation, siblings are distinguished according to seniority, cousins are treated as siblings, but are designated in accordance with the relative seniority of the connecting par ents. This feature is shared by such Melanesians of New Guinea, as the Jabim. These tribes fall into the bifurcate merging cate gory. Reciprocal terms occur for grandfather and grandson, ma ternal uncle and sister's son, parent-in-law and child-in-law.4 Morgan inaugurated both historical and sociological interpreta tions of the facts he scheduled. Starting from the axioms that kinship classifications were "independent of the mutations of lan guage" (p. 506) and that they could not be diffused except if the phonetic symbols themselves were borrowed, he accepted the similarity of systems as proof of racial identity. He thus inferred the unity of the Dravidians and the American Indians, and that of the Zulu and the. Hawaiians (p. 466). Apart from the question ableness of the premises this argument would lead to the absurd conclusion that the Eastern North American Indians stood racially much closer to the Melanesians and various Africans than to their congeners of the Basin area. However, by his collation
of specific resemblances, Morgan paved the way for sound his torical conclusions. For example, he recognized the likeness of certain Central Algonkian and southern Siouan systems (p. r79)— a fact now interpreted as the result of diffusion. Diffusion evi dently is the only possible explanation of the concentration west of the Rockies of such features as the discrimination between paternal and maternal grandparents.
Sociologically, Morgan viewed his nomenclatures as indices of family life. The "Malayan" as the simplest was considered the earliest ; and since it failed to distinguish the mother's brother from the father, Morgan concluded that the two were identical, i.e., that a man had access to his own sister when the system origi nated. The introduction of a term for maternal uncle was said to result from the prohibition of such unions when the Turanian Ganowanian system arose. Finally, the "descriptive" terminology, failing to merge any collateral relatives with the direct line of descent, was correlated with the rise of property rights and the parents' insistence on transmitting them only to direct descendants (1). 492f.). The scheme thus forms part of an evolutionary theory explaining the gradual rise of our civilized family out of more or less promiscuous beginnings.
These interpretations are now recognized as basically wrong. Morgan was not warranted in assuming that the Polynesian term translated "father" implied procreation : the facts simply reveal a common status term for all kindred of one generation. Simpli city here, as in many other linguistic phenomena, is a late develop ment, not a badge of antiquity. Three independent investigators, Sternberg in Siberia, Rivers in Oceania, Lowie in North America, have been able to show recent transformations of bifurcate merg ing nomenclatures in the direction of generation systems.