ABSTRACT
Perhaps the most fundamental problem in operationalizing Hsu's scheme lies in the very concept of "dominant kinship relationship." Is there an unequivocal way of defining dominance so that i t satisfies the demands placed upon i t by the scope of Hsu's hypothesis? Even a precursory examination of the papers contained in this symposium should convince the reader that the answer is no. Dominance cannot be determined in the abstract, but only in relation to well-defined contexts and specific hypo thetical variables. The difficulty in applying Hsu's concept is that i t fails to take into account some important analytical distinctions. For example, a fundamental distinction can be made between defining a dominant rela tionship on the basis of importance for the social system and defining i t on the basis of importance for the psycho-social development of individuals growing up within the society. I f one takes the systemic view, stress is likely to be laid upon such matters as the control and transmission of property, integration of the domestic group and larger social units, mar riage and residence choices, and the like. From this vantage point the
Dominant Kinship Relationships in Fiji and Rotuma 97 possibility must be allowed that a triadic or even larger set of relationships can more appropriately be selected as the unit of analysis than any of the dyadic sets that compose it . I n this regard Levi-Strauss's argument for the inclusion of a pair of spouses, their children, and the wife's brother in the fundamental kinship set for most societies certainly cannot be summarily dismissed. I f one takes the developmental approach, stress is more likely to be laid on such matters as the importance of particular classes of people as role models, as dispensers of reward and punishment, and as objects of hostility or affection. This point of view involves complications also. For example, developing children engage in a sequence of relationships, and the one that is dominant at one stage in development may not be at a later stage. Thus, during infancy the mother is apt to be most important in any society, but as the child matures, his father, siblings, and perhaps others may become more central to his development. Furthermore, males may go through a different sequence than females. At a particular point a boy's mother's brother may provide his most important role model, while his sister may look to her mother. We must also acknowledge the possibility that non-kin may assume socialization roles of primary significance, as peers or nursemaids do in some societies. Another problem, one that is relevant to either view, has to do wi th the dimensions of interpersonal relations. For example, i f we take the systemic view, do we rely solely on formal qualities of interaction or do we take into account informal, or subinstitutional, qualities? Or, i f we take the developmental approach, do we concentrate on cognitive or affective learning? I t is quite possible that the most important person for shaping cognition is not the same person that is most important for shaping emotion.
