ANTROPOLOGICA 59-62, 1983-1984: 299-307 The premise of equality in Carib societies Kathleen J. Adams The ethnographic situation presented in the Guianas is one of many small groups, some of which appear and disappear, change location, combine or separate. This pactern has persisted for hundreds of years of Western contact (Riviere 1969a: 27-28). From sources of different time periods, Gillin (1948: 804-817) records over two hundred named groups representing more than ten language divisions for the Guianas as a whole. Most of these ace reported as living in the interior forests as groups that had fled Western encroachment along the coast and the large rivers. Burt's survey (1965) depicts a similar contemporary situation for the Guianas, a number of small groups organized for survival whacever the conditions of regional relations. Caribs among them seem to emphasize this pattern. The central common element in chese small-scale societies is cross cousin marriage in a two-section system (Gillin 1948: 849, Murdock 1951: 433; Oberg 1955: 478; Riviere 1974a: 640). The emphasis is on direce marriage exchange which must be understood in terms of the tendency for highly consanguineous marriages and concentrated kinship relatedness. These societies are characterized by a “simple, undifferentiated structure, tightly knit yet flexible within its own parts” (Butt Colson 1971b: 88). Designed for group survival at a minimal level (Oberg 1955: 479), the commonly occurring kinship system can organize a very small umber of people. ‘Analysis of Carib relations tends to focus on male kinship, especially the asymmetry between son-in-law and father-in-law (Riviére 1974a; Armellada and Buct Colson 1976; Adams, n.d.). To launch his adult life, a Carib youth leaves home and apprentices himself toa potential spouse's facher, usually his mother's brother, In the perimeter of the household, the young man’s status remains ambiguous until, perhaps ten or more years later, he becomes a father himself and eventually separates his family from that of his father-in-law. While a man has lived in (at least) two households, his wife remains in her natal household until the couple establish their own. AUTHOR'S NOTE: | would like to thank Peter Riviére and Peter Kloos whose comments inspired this re-analysis of Miller's leadership and the role of women in Barama River Carib society.As one with no wife, a man does not negotiate with his father who-has no wife - to.give him. As a wife taker, a man relates to the more powerful mother's brother, whose daughter he desires for a wife, Brothers tend to cooperate in order to release the supply of wives from the control of the older generation, Alsoas a wife keeper, a man relies on the political support of his brothers. In turn, men as wife givers relate from positions of power to the following generation of sisters’ sons. Women are given and women are kept as men conduct politics among themselves. From another point of view, women remain in a mainstream of generational continuity by long-term relations to their mothers and their daughters (Adams 1978). Men's relations which are formed and dissolved in each generation conform to and promote this continuity among women. Political asymmetry between Carib men is never resolved, Rather generational change-over and population dispersion realign relations. A man matures from one with no wife toa wife taker, toa wife keeper, to a wife giver, interacting at each life course stage with different men. A wife giver is always more powerful than a wife taker, but these terms refer to a dialogue in process. Once the wife is given and taken, the relationship between father-in-law and son-in-law is structurally concluded. Beyond the wife giver stage, a man relies on the care of his female relatives, his wife and her sisters, and his daughters. Inequality beeween men emerges in conflicting kinship claims to the same woman. A son-in-law tries to assert a wife status of a woman whose father claims her as his daughter. The woman's location, either in her father’s household or in her husband’s household, establishes her identity by fiat. Equality between men is based on equivalent relationships to the same woman —the same mother, the same sister, the same wife (as a category of potentials). Brothers who marry sisters keep their equivalence beyond those brothers who do not. Unrelated men who marry sisters initiate a basis of equivalence. Men's relations to other men are defined through women, assumed to be passive at least with regard to the formation of households and settlements (Gillin 1936: 136). ‘Adult men move their households about freely in order to associate with other men. In general, cooperation develops among brothers, actual siblings and parallel cousins, and these kinsmen tend to live near each other. And in each generation, these groupings of brothers are equivalent or replaceable among themselves —allowing little basis for the distinctions of leadership. The tenuous quality of leadership among Carib populations is marked (Butt 1965-1966; Dole 1966; Riviére 1981: 23-26, Schwerin 1980: 47; Thomas 1982). When told what to do, Caribs tend quietly to leave for less congested environs, or only temporarily delay this plan. Sons-in-law wait for the time they can remove themselves and their families from the households of their fathers-in-law. A leader, such as John Miller among the Barama River Caribs of Guyana, can emerge through marshalling kin relations, but such political brinkmanship is an anomaly. It is temporary at best and may prove co be in the service of continuity in Carib society. John Miller is prominent in Gillin’s 1931-1932 ethnography. Miller is described as Sawari's headman, "a tall virile fellow with little bloodshot eyes and a crooked left arm” Gillin 1936: 101). He is a “forceful personality, known fat andwide by reputation in the Barama country” (Gillin 1936: 120), The basis of this claim is Miller's success in surrounding himself with people. His settlement, Sawari, had 47 people, almost a fourth of the entire Barama River Carib population at that time (Gillin 1936: 110). At Sawari, Miller's village clearing included 27 people (Gillin 1936: 109). In addition, 13 people lived at John Cook's clearing (Gillin 1936: 111). Some distance between Cook's and Miller's clearings suggests the degree to which friendship had been replaced by competition between them. A third small clearing was inhabited by a single family: a husband and wife and their four children (Gillin 1936: 110). In Sawari's several clearings Miller had around him eleven other adult men (see Gillin 1936: Figure 13). Miller's relationship to his current father-in-law seemed to lack the obsequiousness common to a son-in-law, Indeed with a view to pursuing further domains of power, Miller was learning shamanism from his father-in-law. A few years after the death of his first wife, relations based on a link between Miller and her father were not in evidence. However, the deceased wife’s sister Mary, who had moved to Sawari when her husband died, remained there after her sister died. She cared for her sister's four small children as well as a deceased brother's child (Gillin 1936; 117-118). Miller was most distantly related to ewo married men who did not seem to have brothers of their own. These men no longer lived with fathers-in-law. Nor did they have the human resources to claim independence at this stage in their lives. One of these men had no daughter (as yet), the other's daughters were still young children. It would be some time before either of them could plan to include a son-in-law in their households. As mentioned, one of these men had his own small clearing, a satellite to Miller's. The other was included in that of John Cook. Miller chose to call these two men "younger brother,” although so distantly related they could also be categorical cross cousins, Gillin records that their mothers’ mothers and Miller's mother's mother were sisters, but certainly other relationships could have been traced as well. The presence of these two men in Miller's settlement evokes the greatest generational depth in figuring kinship. The ties between each of these men and Miller are traced through cwo generations of women. Despite such matricentered kinship memory, the wives of these men are unconnected to any other adults in Gillin’s chart (1936: 113, Figure 13). Ina similar manner, Miller's father-in-law's current wife is linked with her sister, but their kinship relations with other members of the settlement are not pursued. In Gillin’s chart and analysis women are wives, mothers, sisters and daughters That is, someone's wife, someone's mother, someone's sister, someone's daughter. The kinship information to support a consideration of women who have mothers, fathers, brothers, sisters, sons, daughters and husbands is absent. The image of women’s inconsequence is enhanced by ignoring direct kinship relations among women in this highly interrelated population. Gillin’s analysis expresses a young man's point of view (Tiffany and Adams 1985). By exploring a common feminine genealogy, Miller could readily co-opt as“young brother” the ewo men in a waiting stage in their own adult lives and so add to the population size of Sawari. Had Miller considered these two men “cross cousins,” they would have called for more complex stratagems in the marriages between living individuals to incorporate them into the settlement. Either way, kinship choice is devised by simplifying and selecting kinship knowledge, particularly to the exclusion of information about women. Such kinship informa- tion as reported is a legend for the political relations between Miller and adult males at Sawati. Eight men were more clearly Miller's age mates and these he addressed as “cross cousin,” in this case wife's brother or sister's husband. Two men were the brothers of Miller's younger wives (who were sisters); their third brother was not yet an adule. Another man was the (tentative) husband of one of Miller's two sisters. This man and his two sisters were Miller's father's sister's children, first marriage choices for Miller's sibling set according to Carib organization. However, Miller had not married in such a manner. These two sisters were cach married to one of wo brothers (Cook and his half-brother) with no recorded relation to Miller. These two brothers’ two sisters were also present and each married to one of two brothers whose third brother was married to Miller's other sister. Miller's influence in his settlement diminished beyond the husbands of his sisters and the brothers of his wives. Essentially bankrupt in human resources due to the death of his first wife, John Cook had come to live near Miller (Gillin 1936: 127-129). Cook married Miller's father’s sister's daughter. This woman was an appropriate marriage choice for Miller himself, and so by extension a sost of brother relationship was initiated between Miller and Cook. After rebuilding his “capital” by having some children, Cook now presented a challenge to his fragile relationship wich Miller and to Miller’s leadership in general. Based on a mature marriage and growing children, Cook claimed an independenc estate in people. Cook's attempts to relocate himself threatened the ties among the sibling groups of his sisters and his wife and her sister, distributed between his and Miller's clearing, At the time of Gillin's field work, Cook was not able to separate chese sibling groups. Ten years or so into the process of establishing a large settlement, Miller had built a coalition, not, however, with brothers or sons-in-law. In an earlier episode in their lives, Miller's brother Alec and father were party to the homicide in which Miller incurred a deformed arm, and all three served jail terms in Georgetown Gillin 1936: 120). But neither father nor brother is mentioned at Sawari in 1932-1933. Alec lived at Hairi, a secclement some ten miles away (Gillin 1936: 120). Miller's daughters by his firse deceased wife were very young at this time, and Miller had no children, as yet, by his younger wives, Elizabech and Charlotte, Rather, Miller held together a kin group of men who were cross cousins to him, either actual cross cousins or those by marriage. All of these men were related through 2 female with a more immediate kin relationship to Miller. Retaining his deceased wife's sister as a wife, Miller had in his settlement two more groups of sisters —his own sisters and his younger wives, sisters to each other, and their younger sister. Miller's seetlement grew with the arrival of three other sets ofsisters, including Miller's father's sister's two daughters and Miller's sister's two sisters-in-law. Also, Miller's father-in-law David had remarried a widow whose widowed sister, in the course of events, came to live near her at Sawari and subsequently took David's son as a husband there (Gillin 1936: 124). Of the many deaths which occurred among the Barama River Caribs at this time, chat of a spouse of adult age was the most disruptive to sociery. A woman, it seems, upon the death of her husband, returned to her sister's clearing, A man would go where he could obtain another wife. Miller took advantage of both of these tendencies in the management of his settlement’s population size. Miller's first wife's sister arrived at Sawari upon the death of her husband and Miller himself became her husband. Miller's current father-in-law’s wife's sister similarly moved to Sawari when her husband died. Although much younger, Miller's wives’ brother became her husband, Miller's sisters joined his clearing. The husband of one had died. The other's husband had left her. Miller helped to secure another husband for each of them at Sawari. Miller's accomplishment is contrary to Gillin’s analysis of the building blocks of Carib society: "brothers tend to group together to form a settlement and assist one another in social and economic affairs in the settlement” Gillin 1936: 97). Groups of brothers were found in Miller's settlement, but Miller had no actual brothers there. By holding together sets of sisters, Miller was able to expand beyond that social group of brothers and their wives possible with the Carib principle of direct marriage exchange. Miller, without brothers, was able to attract men who added themselves as husbands to women in Miller's settlement. These men brought siblings and their spouses, establishing a complex society under Miller's purview. Miller had a double wife taker relationship, at least a nominal one, with his father-in-law, but shared a wife keeper status with none of his contemporaries. Nor was Miller a wife giver. With the rest of the men, Miller had neither a common status with regard to the same woman nor a competing claim with regard to the status of the same woman. The premise for these men's cooperation with Miller's leadership was negative, if not null. Had he so wished, Miller could have initiated a “brother” relationship with his sister's husband's brother who desired the second of Miller's younger wives for his own wife, Her father said that he had “given her” to Miller Gillin 1936: 119). Miller would not release her. Miller's sister's husband and his brother were Cook's brothers (half brothers), and such a marriage would have added momentum to Cook's assertations of independence as well as the separation of sister sets. Of course, women may serve as passive links between men who are political actors. The relationship between son-in-law and father-in-law may be conceptual- ized in this manner, although the potential for the daughter's influence is great Women's presence alone does not convey participation in political arenas. What is the manner of women’s political participation in this Carib example? What did sisters have to gain by supporting Miller's leadership among men? The Barama River Carib example suggests that women and their reproduction add a dimension to the Carib political charter. Existing marriages among living individuals and their children are more critical than references to lineages forestablishing kin groups. An adaptation to dispersion, small Carib societies tend to be based on no more than three generations, and adjacent generations tend to be in charge of their interrelations without reference to the interests of a third generation. Usually fathers-in-law and sons-in-law relate to each other in the absence of a political role for grandparents or even sons-in-law's parents. However, the daughter/wife does inrervene between father-in-law and son-in-law. At stake is her ability biologically and socially to reproduce children, her interest in which is maximized in Carib society. Her courtships are overseen by her natal family who continue to protect her as a young mother. Only as an experienced mother with growing children does she leave her parents. Population producers and their surrogates, women and their sisters, gain a measuce of social determination as their interests converge with those of men, whose political careers are based on the management of population size. Thus, women and their children are a basic unit of Carib society, identified in the immediacy of dispersion and relocation. Collective leadership of extended fraternal groups is a feature of Carib society understood in theory, by anthropolo- gists and by the Caribs themselves. With three generations or more of population stability, such groups of brothers, matried to their cross cousins, can emerge. Among Carib populations, however, disruption and dispersion seem to have been more common than stability. Sisters and mothers and mother’s sisters represent sources of assistance and refuge. Under these circumstances, women neither give nor take relatives. Sisters, as kin keepers, care for parents and husbands, and as mothers they produce relatives, ie. children. Carib society regenerates itself again and again, And in this process, women, especially sisters, and che sibling sets they reproduce are a basis on which political relations among men can be elaborated. A Carib population can theoretically grow from one woman, Equality between women, like that between men, is based on an equivalent relationship to the same woman —the same mother. The inequality between men does not seem to occur in women’s relations. The principle of regeneration is chartered in the mych, a Carib version of the Noah story, presented here. Among the Barama River Caribs, Noah is a woman A dog would come to people’shouses. He was very mangy and skinny. Every time the people would scorn the dog. They would yell at him to go away. The dog would return to the forest and appear at another house. At some houses he would appear as an old man who had many sores and could hardly walk. It was all che same. From house to house, the people would chase him avay. ‘One day, he came to a place where there wasa lady alone. There were no children, nobody else. He appeared like a man, an old man with many sores. He seid, “Hello, my daughter.” ‘The lady gave him a seat and he sac down, The lady studied him and thought, “My father could look like this.” ‘With wacer, she washed his sores and wiped his skin. He sat down, In good time, he parted his hair with his hand. He had no comb, bur his haic shone like glass, The lady said, “Daddy, please take food.” She gave him food and sweet cassiri (a fermented beverage) to drink. (This lady's husband had gone hunting. He did not know anything), When the man was finished eating, he said, “Alright, daughter, I am going awa;She said, “Alright, walk good.” “The man went a good way. The lady went to where he had sat and found a piece of arrow. She said, “Eh, my father lefe chis thing.” She ran behind him hollering all che time. ‘The old man said, “Aleight, girl, bring it.” He told her what was going to happen with the fittle piece of arrow. He said, "This is a great, great thing I left for you. [ did nor forget it. Beceuse you followed me, 1am going to tell you everything that is going to happen. My daughter, this piece of stick is going ro be a big boat for you. This boat will get large by degrees. After so many years, it should be big, Put all kinds of plants in the boar. Do not mind that there is no water, put everything in. Take every animal by pairs. Pack everything in the boat. My daughter, later on rainis going to fall night and day without stop. The rain will not stop until all che trees are gone and the world has been killed with water, That is all 1 will tell you.” ‘And the old man was gone on his way. ‘The husband came back from hunting, and the lady cold him what happened and showed him the stick that would be « big boat, Sheexplained that her father told her how it would ger large and she was to put all che plants inside, She was o rip out all of the cassava and pur it inside, Then it was going to rain night and day and never stop. Water was going to pour out on the yard, and not one tree would be left to dimb. ‘Years went by and more and more the boat grew. The lady planted all the plants inside. She pur in all che types of animals. The rain fell night and day as the old man had said it would. Plenty of people were sporting, drinking cassiri."The message was sent all about, but nothing happened, There was more and more waer, and more and mare people were sporting, The family, che sister, all were told good-bye. The wife and busband went inside the boat, and che door locked. All the people said, “Let us come in.” ‘The husband said, "No, I do not know how to open the doot ‘There was more and more water. The people climbed upon the house roofs. More and more of the people drowned, Then all of them drowned. This big ocean covered the whole bush. There were no trees, just a boat and one vine rope from the boat to the sky. Al the time the ocean was lifting up the boat, and afterwards che boat was left hanging up in the sky. God sent down a crow to see if the water had dried. The crow saw a lot of human beings who had drowned. The crow started to eat them and did not bother co go back. God forgot him. God sent a waracabra to see if the water had dried. Fine red ants ate both of the waracabra’s feet down to the bones. The waracabra wanted to come back, but he could nor. God forgot him. God senca pigeon to see if the water had dried, The pigeon flew over everything and did not lighton the ground. Nothing troubled him and he went back to tell God that the water had dried. “The boat was sent back to the land. God opened the door, and every type of animal came out and fended for themselves. God made a building for the two people and told them to plant. Alll the plants came back quickly. What was planted today could be used tomorrow. In this narrative, many themes are expressed: ethnic distinctions, a woman's relation to her father and her husband. Among them, a metaphorical reference to the cassava plant is unmistakable. The asexual or vegetative reproduction of the cassava plant, as cultivated, represents human reproduction as well. Beginning wich one woman, children bud off, and from the daughters among them grandchildsen bud off and so on. ‘A benefactor appears as a mangy dog or sick old man, a powerless Carib status. ‘The old man exchanges a seed for the care given him by one woman, the exception among many. The benefactor is supernatural, “his hair shone like glass.” In other versions he has "boots and a hat,” suggesting a Western identity and the old man's source of power as well. The seed is a “little piece of arrow” in this case, a Western coin in other versions. The seed grows into a boat into which the woman is to" putall kinds of plants,” especially cassava, The earth and the people are drowned by water, just as the cassava plant can be. When the water subsides, only the woman and her husband remain. While the animals fend for themselves, the gardeners’ endeavors are rewarded by rapid growth; the vegetative reproductive process of the Caribs embarking on society once again. The woman specifically says good-bye to her sister, and while her husband accompanies her, he “did not know anything,” It is she who treats a stranger as her father, obtains from him the gift of future life, and secures plants in the growing boat. When called upon to save kin, the husband confesses, “I do not know how to open the door.” Neither does the woman, Rather through naive acts of nurturing, she is able to reproduce kin, Closer to the mouth of the Barama River, the people of Sawari in the 1930s were first exposed to introduced diseases and Western economic influences which changed the settlement patterns and demography of the entire Barama River Carib population by the 1960s. Ic is suggested that John Miller was the headman of segments of Carib society, sisters and their husbands and children, exposed to this cycle of historical disruptions. These sisters returned to each other when left widowed, took care of deceased sibling's children, increased the number of daughters among their offspring (Adams 1981), and in general, emphasized kin keeping, In a time of greater stability, these sisters may have been more willing to leave each other and devote themselves to their own husbands and children. Miller rose to leadership through the manipulation of marriage relations which permitted sister sets to stay together. Quite a different image, including allusions to violence, was attached to Miller. He was thought to be the best hunter. Despite his deformed arm, Miller was “respected by all for his cunning, courage, and strength” (Gillin 1936: 121). Miller, apparently, was able to hold his setclement together, at least most of its population, until his daughters by his first wife found husbands and became mothers. Eventually, Miller, as a wife giver, confronted younger men as wife takers. Charlotte's group reported that in the 1950s John Miller had fallen victim to a 4anaima, the mythical avenger of the Caribs. After cutting a field that day, Miller had gone at dusk to bathe in the creck. Later he was found badly beaten and unable to speak. Two days later he died. The population at Sawari dispersed, and then some of them, including Charlotte and Elizabeth and their children, moved back. Miller, whose atypical leadership was made possible by the women’s part in the Carib political charter, was eventually defeated by women's roles. While Miller was able to manipulate networks of adults, he could not forestall the re-emergence of Carib society among his daughters and their children, Eventually, the generational sequencing of Carib society terminated Miller's leadership. While the accumulation of actual people in one location is a basis for male leadership, opposing population principles also operate in Carib society. These principles are introduced with women's roles in reproduction and distribution of individuals These principles operate in generational time and tend to undermine the political career of a leader such as Miller. Without an economic component in his politicalcontrol, the end of Miller's leadership was timed by the reproductive maturity of a new generation of Carib women, In 1970-1971, Sawari was abandoned. The Carib population, while increasing rapidly in size, had collected around Baramita Air Strip near the headwaters of the Barama River, in response to economic opportunities there. Others had left the river system for Matthew's Ridge some fifty miles to the North. ‘The Sawari population which had moved to the air strip included Charlotte, Miller's youngest wife in 1932-1933, Charlotte's two daughters (and che husband they shared), Charlotte's deceased sister Elizabeth's (also Miller's wife in 1932- 1933) son and daughter and their spouses and Charlotte's young son by Miller's older son whom she took for a husband when Miller died. Ochers from Sawari, including Charlotte's eldest daughter marrieu. to Charloree's brother, had gone to Matthew's Ridge. ‘At the air strip, Charlotee’s groups lived in close proximiry to each other and helped each other. Charlorce’s married son began to pay more attention to his wife's family, Charlotte's daughters’ husband and Charlotte's sister's daughter's husband (also married to two other women who were sisters) were distant relatives to each other, but they began to act like brothers. New marriages and new children recreated a Carib society. Abstract Carib social organization is explored using an example of leadership among the Barama River Caribs of Guyana. In this care, women’s contributions to social organization are examined in generational time as Carib strategies for survival. Principles of equality among men and among women are reviewed as mechanisms for population and generational continuity. Resumen El presente artéculo ofrece una investigacién sistemdtica sobre la organizacién Caribe en la que se muestra, ademas, un ejemplo de liderazgo entre los indigenas guyaneses del rio Barama. Elirabajo presenta un estudio de caso (a través de varias generaciones) de la contribucién femenina a la organizacién social Caribe como ertrategia de supervivencia. También se examinan los principios de igualdad entre hombres y mujeres como mecanismos de adaptacién poblacional y de continuidad generacional.ANTROPOLOGICA 59-62, 1983-1984: 309-329 Some aspects of structure in Kuikuru society Gertrude E. Dole Introduction Some 3,000 years ago a branch of Southern Caribs separated from the large block of Caribs in the Guiana region (Durbin 1972). Contemporary survivors of that southern branch include the Kuikuru, Kalapalo and Matipt-Nafuqué in the Upper Xingé region of Central Brazil. Because of their separate social and ecological history it is not surprising that these groups differ in many respects from those in the North. Significant differences occur also among the Carib groups in the Upper Xingt region. By the 1880s the Kalapalo and Kuikuru, for example, had been separate long enough to develop regular dialect differences and, as we shall see, Kuikura culture differs in numerous other respects from that reported for the Kalapalo. Sufficient data are now available co make comparisons among the various Carib societies and to permit an analysis of the variations. I therefore take this opportunity to present some details of Kuikuru organization in the hope of stimulating comparison. ‘The various aspects of a society may be structured according to different principles. Residence, for instance, may be matrilocal in a society that reckons descent patrilineally. Or people may reckon kinship through bilaceral filiation even though succession to leadership is patrilineal. In the following description, therefore, I will deal separately with leadership, economic organization, residence patterns, and kinship organization! 1 Most of the data in this paper represent conditions as they were during the period of my field work in 1953-1954, before the Kuikuru left their settlement on the twin lakes Kuikuru (or Kufikugu) and Lamakuke to the south of the Parque Indigena do Xingd. These data are supplemented by additional informacion collected by Robert Carneiro in 1975 at the ceservacion settlement of Afanitafagi, towhich the Kuikuru where moved in 1961.I want co stress at the outset that native models of cultural? patterns are only one of several kinds of information available. As a matter of fact Kuikuru informants seldom speak in terms of what must or should be done. Instead they state their interpretations of what is done. Their statements often represent not societal norms but individual informants’ versions of social practices and often differ from one informant to another. This generalization deserves some explanation. Kuikuru households, and to some extenc individual families, maintain a respectful, if not suspicious, social distance from one another. The extent of social isolation in this small supposedly communal society is perhaps a surprising contrast to the constant, sometimes relaxed and intimate, interfamily communica- tion among friendly neighbors that characterizes relations in some complex societies. Because of family isolation Kuikuru concepts of both ideal and actual behavior are influenced largely by particular experiences, and to the extent that each person experiences the details of social relations differently, his or her perception of tribal custom is unique. Hence informant statements do not necessarily represent either norms or practices that are common to the whole society. To insist on cultural cules in these circumstances runs the risk of imposing ethnographers’ concepts on the data rather than deriving patterns and concepts from the data, For these reasons] will give full consideration to observed practices as well as to native models or concepts of ideal behavior. By noting discrepancies between actual behavior and stated norms on the one hand and the range of individual variation on the other, I hope to clarify some issues regarding the nature of Kuikucu social organization, Norms and practices: leadership The community (ete, or ito)’ is a politically autonomous group, ie. tribe (see Dole 1968). Other indigenous Upper Xingu groups are referred to as ubuge, "my people,” in contrast to outlying Indian peoples, referred to as nikogo, “enemies,” and white people, who curiously enough are called Karaiba (Kagaifa in Kuikuru phonology). There is no political integration of these indigenous groups except for the influence of the Brazilian government through the presence of the Indian Service, FUNAI In the Kuikuru tradition of a minimal hierarchy of leadership, one person is recognized as headman, with a category of subordinate male and female leaders, or “helpers.” Both classes of leaders are called aneti: Although there in no special erm 7 Luse “culture” here to include not only symbolic systems, beliefs and concepts, but also social ‘organization and artifacts, » The term o16mo recorded by Basso as being the “Xingi Carib’ term for village or people (1970) is not the Kuikuru term for community but was heard among the Kuikuru only from’ in-marrying Kalapalo, The Kuikuru name for the settlement visited by Basso is Afanitafagi and not Lafatué Although the Kalapalo apparently refer to the Kuikuru as Lafatus, that is not the name by which they themselves or other Xinguanos refer to them but rather the name of a site they occupied in the 1950s.for headman, priority and deference are accorded the headman by acknowledging him to be the owner of the settlement plaza, a symbol of his authority co make political decisions and represent the group in interrribal contacts. In contrast with aneti all other members of the community are referred co as isandagi ot kamaga (reportedly a native adaptation of the Brazilian term "camarada”). ‘Although Kuikuru society is generally cognatic, succession ro leadership status is ideally patrilineal in an uninverrupted line of local ancestry. In practice, however, it seldom happens that a headman is succeeded immediately by his son. In 4 sequence of twelve headmen in recent generations, leadership passed only once directly ¢0 son. In most instances it passed to another family and was assumed by a mature male whose father also had been headman at some time in the past. Thus leadership alternates among several patrilines and lineal succession is usvally delayed, From this record, it can be seen that the norm of patrilineal succession represents only the right to succeed, subject ¢oa number of conditions. In actuality, leadership is largely achieved. If a son does succeed, it is not necessarily the oldest but rather one who shows leadership qualities, including a desire to lead. His authority must be validated by earning the respect of che community. In principle, sons of former leaders in other tribes who have married into the Kuikuru settlement are not eligible for the position of headman, even though they may be effective leaders in many internal economic and ceremonial activities. Informants explicity denied that one such man, the son of a former Mehinaku headman, would succeed in the evenc of the current headman’s death. Nevertheless, afer the latter died in 2 measles epidemic the “Mehinaku” did assume the leadership role and was explicitly acknowledged as having been headman by at least some Kuikuru informants a generation later, although other informants named a different man as having been the successor. This incident illustrates an important aspect of Kuikuru social organization. Not only may there be a discrepancy between tradition and practice, and wide differences of opinion about who should or will succeed, but also as a supreme example of the variable nature of Kuikuru society there may be lack of agreement about who is headman at a given time, as was the case in 1954 and again in 1975. In the latter year a young grandson of a former outstanding headman was acknowledged by many to be owner of the plaza and the proper headman. He represented the group in intertribal ceremonies but apparently felt himself to be too young or inexperienced to exert his authority by haranguing the community. At that same time an older and overtly more ambitious man, who himself was grandson of a former headman and nephew of another, undertook to harangue the group and in other ways assert his own right to the position, Opinion in the society was sharply divided as to whether the latter was in fact headman, some saying that he was too old and already “dead.” Resentment at what was seen by some as an attempt to usurp the authority of the younger man led toa temporary division of the tribe into two local groups. Both competing head men remained in the same division, however, and the seceders later rejoined that group. Although this temporary split illustrates the fact that family groupssometimes disagree about public issues, harbor resentment and talk against one another in private, Kuikuru social structure has no provision for formal factions. Even during the temporary separation factions did not crystallize, nor are Kuikuru aneti leaders of factions as reported for the Kalapalo (Basso 1975). Kuikuru anet¥ are not che ones who usually express interfamily tensions; on the contrary they area class of industrious adults who exemplify Kuikuru ideals of peaceableness and cooperation. When a legitimate and acknowledged headman fails to exercise leadership in times of crisis, alternate mechanisms are used to maintain the integrity of the society. On one occasion in 1954 some informants expressed frustration with an ineffectual headman by claiming that the rea/ leader was another person who, like the resident headman, was both the son and grandson of former headmen. Ironically that person no longer resided in the community, having left che Kuikuru settlement many years previously, after missionaries and journalists had made him notorious as the supposed grandson of the British explorer Colonel Percy Fawcett. Still, no effort was made to recall him, and political decisions were arrived at through the supernatural mechanism of divination by a shaman (see Dole 1964, 1966), and some aspects of local leadership were assumed informally by other aneti, Succession to the status of subordinate anes? mainly pacritineal, but a woman also may acquire the status from her father and may transmit it to her son, Here again there is uncertainty and disagreement, even among aneti themselves, as to who is and who is not a leader. With characteristic unwillingness to assign high status, one informant asserted that only an aneti/himself knows whether or not he is one. Like the position of headman, subordinate leadership status is in part achieved and depends largely on personality and performance. Economy Although no parcels of real property are owned either by subgroups or individuals and boundaries between Upper Xingii tribes are indefinite, the area regularly exploited by the Kuikucu is claimed as their tribal terricory. Interestingly, their concept of territoriality extends far back into the past and many miles from their settlement, for they still claim ownership of the very large Lake Tafonuno, on which their ancestors settled perhaps two centuries ago on their way into the Upper Xingii basin, They continue to use this lake as a major source of fish and maintain that other groups should ask their permission before exploiting it. 1 will describe the organization of economic cooperation in terms of successively more inclusive groups of people who are linked by successively weaker kin ties, beginning with che intimate family groups. In contrast to the primary economic importance of extended family households as reported among the Kalapalo (Basso 1973: 43; 1975: 210), much of Kuikucu economy is conducted by and for individuals and small families. Staple food crops are usually cultivated and processed by the members of the nuclear family. Not infrequently, however, members of an extended family join together to perform tasks such as harvesting, and preparing manioc and piqui fruit. The most common and most efficient ofthese work groups are patrilocal and fraternal extended families. Fishing is sometimes done by individuals or separate nuclear families, but more frequently by small groups of men who are friends or close ki Other cooperating groups comprise members of multifamily households, formal friends (ato), and isogender (same-sex; see Dole 1957: 144-145) age mates. Within the multifamily houses various nuclear or extended family groups, whether closely related or not, share the use of major immovable equipment. All the families ina house store manioc flour in a single basket silo and draw from it for daily meals, use a single storage rack, and often cook on a common hearth. Groups of volunteer workers, sometimes including virtually the entire adult male population, cooperate to perform large tasks such as gathering and preparing food for a festival, hunting monkeys or grasshoppers, cutting and transporting woodskin canoes or house- posts, and constructing houses or fish weirs. On these occasions workers are recruited with the promise of a festive meal as their reward. A basic principle of Kuikura economics is that each person owns the product of his or her labor. Men own garden plots as long as they are maintained in cultivation, and crops are owned by the women (or men) who harvest them. Persons who organize projects with group labor own the products of thar labor. Most material goods are not inherited. At che death of the owner, significant items, including houses, should be destroyed or buried with the deceased. However, sources of perennial food harvests, such as fish weir emplacements and groves of fruit trees, may be inherited, usually by a son of the deceased owner, and in some circumstances even houses are used after the death of the builder-owner. Native songs and dances may be purchased by individuals, while those of foreign origin are acquired by learning them and paying performers with festive food. These also may be inherited by a son, or by a daughter in the case of the few women’s ceremonies. Trading, both domestic and intertribal, is conducted by individuals and not by kin groups or the tribe as a whole. Goods and services may be exchanged directly or through the partially standardized medium of beads, obviating the necessity of maintaining close or long-term interpersonal ties for the purpose of completing delayed exchanges. However, a man tries to cultivate a special friend (ato), usually not a close relative, in each tribal settlement to facilitate intergroup trading and visiting. (For further details on ownership and exchange, see Dole 1959). On the whole the economic organization, like leadership, is characterized by individual initiative and nuclear family cooperation, supplemented by voluntary cooperation of other individuals who are compensated by direct payment. Residence Each of nine or more multifamily houses shelters from one to seven nuclear families. In contrast to the Kalapalo, among whom the household is equated wich an extended family (Basso 1973: 50-51), the organization of Kuikuru households varies greatly. Some comprise all or parts of several extended families, and conversely, nuclear families that are closely related are often dispersed among several households, even among separate tribal settlements.Ideally a couple lives with the husband's family, patrivicinal residence according to Carrasco’s typology (1963). From the ideal of patrilocality it follows that there is an expectation of forming patrilocat extended family residence groups. However, because of many exceptions to the ideal pattern few such groups occur. One of these exceptions is the usual trial marriage period of initial uxorilocal residence. The length of this period varies from a few days to years and depends on a number of socio-economic factors, including the demands of life crises, the strength of the bride's family, and the politics of family relations. In some instances the couple continues to reside uxorilocally for many years or indefinitely while the groom fulfills his obligations of bride service. In any case, a couple may live with the wife’s family for some weeks or months both before and after the birth of a child, since giving birth ideally should take place in the prospective mother’s parents’ home. As a result of these variable residence arrangements, a variety of family residence groups is formed, including virilocal and uxorilocal extended families, ambilocal fraternal extended families (groups of martied brothers, married sisters, or both), virilocal polygynous families, and composites of two or more of these types. Aside from the predictable changes under conditions already mentioned, individuals and families change residence in response to other contingencies such as disasters, overcrowding and interpersonal tensions or outright hostility resulting from failure to cooperate, or suspicion and accusation of delicts or witchcraft Kuikuru frequently respond to such situations by leaving their house (or settlement) and taking refuge with friends or relatives elsewhere. Thus in spice of an ideal of ultimate patrilocality, residence is in fact irregular and unstable with respect co both settlement and focal relatives. It is difficult co discern a representative pattern of residence at any one time. Nevertheless, if one considers residence in the context of the widest circle of kin, the data not surprisingly show a tendency toward uxori-patrilocality in contrast to the pattern of matrilocality attributed to Caribs in the Guiana region, Descent, filiation and kin groups If residence conformed to the ideal of pattilocality, the local group would be a patriclan in Murdock’s terms (1949: 66), comprising the families of married males but not their married sisters or daughters. As we have seen, however, uxorilocal residence and frequent shifts from one community to another, prevent the emergence of such a clan structure, The entire community is generally regarded as a single kin group. Kin ties are reckoned bilaterally and co this extent kinship is characterized by filiation alone rather than by descent, but there are many irregularities in the application of this principle with che result that kin ties are often ambiguous, as will be explained + "Extended family residence group” is used here because the groups of related families who live in the same house often differ in composition from the extended families that cooperate in food production and useA consideration of kin groups also reveals variation, ambiguity and flexibility of membership from the largest co the smallest units. With respect to the tribe, membership is somewhat indefinite, and informants may disagree on the tribal status of individuals. Although former members of the community currently residing elsewhere are usually still regarded as tribal members, persons who have married into the Kuikuru community from other tribes may be classed variously depending in part on their compatibility and the length of their residence with the Kuikuru, Offspring of intertribal marriages tend to be classed as Kuikuru if the father is Kuikuru and if the family resides with the Kuikuru, suggesting a patrilineal emphasis in tribal membership. The kin community is not necessarily the same as the maximal kindred for any person. Members of one’s personal kindred are referred to as ufisuingi, a general term that may be glossed as kindred, relative, sibling or brother. In its widest meaning one's ifisuings? includes those persons, living and dead, both in the Kuikuru community and elsewhere, to whom a genealogical relation is either known or stipulated. The Kuikuru claim no totemic ancestor, do not identify any lineages, and have no corporate descent groups other than extended families. All extended families, whether patrifocal, matrifocal of bilateral, are shallow, compri- sing no more than two generations of adults. ‘A major source of ambiguity arises from the fact that in this, as in other cognatic societies, people characteriscically recognize relationships to one another through two or more sets of genealogical links. A person may be one’s cross (and hence affinible) relative through one parent and at the same time a parallel relative through the other parent. Choices of which set of kin ties to recognize often vary according to one’s purpose. Recognition of genealogical ties sometimes shows the influence of sameness of sex and length of residence in the community. An emphasis on patrifiliation is shown by the fact that individuals sometimes recognize kinship to half-siblings and cousins only through male relatives. Another source of irregularity is that kinship is quantifiable, in the same sense that Kariera, for example, speak of “close-up” and “far-away” kin (Radcliffe-Brown 1930). In some instances reckoning of relationship appears quite inconsistent, as when one woman classified differently each of three own brothers in the same context, referring to one as “cross cousin” (zfgj), another as “brother” (ufis#), and the third as “unrelated” (tel0). In sum, although kinship among the Kuikuru is generally cognatic, the manner of reckoning kin ties is flexible and varies both from one informant to another and from one situation to another. Kin terms The pattern of reference terms® in general use among the Kuikuru is Bifurcate 5 Initial w- indicates the first person singular form; i- indicates third person singular. ® Data on kin terms were obtained by several methods, including 1) observation of native use, 2) participation in the kinship system by vireue of being spontaneously assigned fictive kinship status byGeneration. That is, the three medial generations alternate between the Bifurcate Merging and Generation types. With few exceptions the reference terms are used in address as well. As wich other classificatory nomenclatures, use of these terms is characterized by a considerable amount of polysemy. For example the generic term for brother, ifisuingi, may also refer to male parallel cousins, to all siblings and parallel cousins, to any relative in one's generation, or to all one’s relatives. Ie will be seen from the list in Figures 1 and 2 thar separate terms are used for affines, including parents-in-law, spouses, and siblings-in-law. (A child-in-law is usually referred to by name or through teknonymy by combining terms for child and spouse.) It is important to note that none of these affinal terms bears the meaning of cross relative. This point will be discussed further in connection with the Cross Cousin pattern of nomenclature. In general communication all cousins are referred to with sibling terms. However, at the level of greatest specificity these terms contrast with another term, ufgi, that may be tentatively glossed as cross cousin of either sex. This term occurs among the Kalapalo as #fa¥, which Basso represents as allogender cross cousins only and potential spouses (1970: 407). In either case the term corresponds to the distinction between cross and parallel relatives in the first ascending and descending generations. Its use is chus a feature of the Dakota-Iroquois component of the Bifurcate Merging pattern. Ir does not constitute the so-called Dravidian pattern identified and described by Fison and Morgan (Morgan 1871: 582) and named Cross Cousin by Hocart (1928: 180-182; see Dole 1957, 1972). The Cross Cousin pattern makes a rigorous distinction between all cross and parallel relatives in the three medial generations according to marriage category or type of genealogical ties, no matter how distant. By contrast the Dakota-[roquois pattern of terms among the Kuikuru and wherever it occurs, classes all cousins according to the sex of their inking parent relative to Ego’s linking parent alone, without regard to ocher genealogical ties. This pattern obscures the distinction between cross and parallel ties beyond first cousins and results in an inversion of the Cross Cousin pattern for cousins beyond the first degree. In the same way the Bifurcate Merging pattern classes all avuncular relatives, near and distant, according to their sex relative to Ego’s linking parent, and nepotic relatives are classed according to the sex of their linking parent relative to Ego, all without regard to genealogical ties. Finally, in the Cross Cousin pattern, terms for cross relatives also function as affinal terms; there are no special affinal terms such as are used among the Kuikuru and Kalapalo. The Kuikuru translate the term sfaj as “primo,” the Portuguese word for cousin, or as child of father’s sister or mother’s brother (etsi mukugu, auajti mug), contrasting it with MSiSo/D and FBSo/D. This term is of special interest for several reasons. For one thing itis rarely heard and its use is restricted primarily to Private conversations. Nevertheless, in spite of its illusive character, the cross cousin term is a key to a common heritage of kinship organization among Carib the natives, 3) the “genealogical method" of obtaining genealogies and asking for relations among the persons named, 4) eliciting from many informants theie kin terms for all other Kuikuru, and 5) asking for terms for kin types as represented in a traditional diagram of relationships.FIGURE 1 KUIKURU KIN TERMS OF REFERENCE: CONSANGUINEAL KIN Generation Term Meaning 43 apitsikuengi great-grandfather, other male relatives kokojokuengi ‘great-grandmother, other female relatives api gcandfacher, other male relatives 42 aupigi kokojé grandmother, other female relatives isi apaju father, father's male relatives unui, yut father; sometimes also father's male relatives ammatiu mother, mother’s female relatives +1 uisi mother; sometimes also mother's female relatives auaju mother’s brother, mother's male relatives Ojoget mother's brother, mother's socially close male kin esi father's sister, father's female relatives brother, male relative; younger brother (male speaking) vfitiano older brother, older male relative (male speaking) vfisuing? brother, male parallel cousin, sibling, cousin uufonifi orofono own sibling vuinansu sister, female relative 0 fast older siscer/female relative (female speaking) wikene younger sister/female relative (female speaking) kikeneko baby sister (female speaking) jaja older isogender sibling; sometimes also older sibling or cousin of either sex ufgj cross cousin pami® umugu son, child (male speaking); son of isogender sibling or cousin (u)mukugu + son, child (female speaking); son of isogender sibling or cousin vindisi daughter, daughter of isogender sibling or cousin vufati daughter of allogender sibling or cousin ufatyi son of allogender sibling or cousin a young child, allogender child ipi son, male relative, young child (male speaking) iit daughter, female celative, young child (female speaking) ufigi grandchild, relative vititi namesake grandchild 3 ufigikuengi ‘great-grandchild * In its bound form in combination with other relationship terms, -jogu has a much wider range of meaning and may be used to refer to virtually anyone who is in close kinshipor friendship relation wich the speaker. This usage expresses affectionate license and appears to correspond somewhat to our use ofexpressions such as “funny face,” “old thing” or "my old lady (man),” for socially close relations. ° In the Upper Xingti pami appears to be a word borrowed from the Bakairf, whose dialect forms a separate subdivision of Upper Xing Carib (Durbin 1977: 35). It is used by the Kuikuru in the same manner as ufgj. Basso’s assertation chat pam is not a kinship term among “Xingii Carib” and is used differently from ifag (1970: 412) does not apply ro the Kuikuru and should be restricted to the Kalapalo. FIGURE 2 KUIKURU KIN TERMS OF REFERENCE: AFFINAL TERMS Generation Term Meaning, +1 iffisofo parent-in-law ilo, -iso husband fits wife ufamerigi isogender sibling-in-law ° uakene allogender sibling-in-law ufafitst husband's brother, wife's sister ifitsimbigi former spouse, separated peoples. Cognate terms are used for cross cousins or siblings-in-law, actual or potential, among many other Carib peoples. | have listed 15 such groups (1969: 112), to which may now be added the Trio (Riviére 1969a: 284), Kalapalo (Basso 1970), Maroni River Carib (Kloos 1971: 133ff.), Galibé (Arnaud e Alves 1975), and Panare or E'fiapa (Villalén 1978). From the wide distribution of similar terms with the same meaning it is clear that a common term must have been used among Caribs before the separation of the Upper Xingu branch. Further, the use of a common cerm for cross cousins and/or siblings-in-law suggests that a Cross Cousin pattern may have been formerly used among Caribs, as it is today among the Yanomamé, Jivaroan groups and Campa. Its replacement by sibling terms in general usage is correlated among the Kuikuru with a loss of the distinction between cross and parallel cousins. Marriage Marriage among the Kuikuru is not structured by a prescriptive rule or by any proscriptions other than the genealogical definition of incest. The prohibited categories include, in addition to primary and lineal kin, parents’ siblings, one's first parallel cousins, and siblings’ children. Generally the same prohibitions apply also to extramarital sex relations. An occasional breach of these incest rules illustrates the permissive nature of Kuikuru society, for incest is aot punished, nor is the deviant couple reproached, although people are said to be “unhappy” about it. Exogamy and endogamy in the absence of other marriage rules it is necessary co discuss exogamy andendogamy’ specifically with reference to the various units of Kuikuru society, beginning with the language block. Stated preferences in choice of mates differ from one informant to another, Some express a general preference for marriage with Carib speakers, and there is in fact a strong tendency toward language group endogamy. Others, perhaps influenced by the current availability of eligible mates, prefer marriage with someone in a particular non-Carib group. Expressed norms are permissive also wich respect to marriage within the household unit, Some informants indicate that marriage to a member of one's own household is acceptable, and marriages do occasionally unite persons who already live in the same house. With respect to the tribe, some people prefer marriage with other Kuikuru, whereas others prefer mates from a particular non-Kuikuru tribe. At one time in the recent past at least ten different tribal communities were represented in the Kuikuru settlement, Some of those outsiders had joined the Kuikuru as refugees from other small groups who suffered depopulation. Depopulation ‘The Xinguanos have a 300-year history of demographic disturbance through warfare, depopulation and merging of tribal groups. The Kuikuru themselves tell of Karaiba (white people) entering their region a very long time ago wearing blue and yellow kerchiefs (“bandas”) around their necks. These “bandeirantes” killed Kuikuru and other Indians with "swords" and destroyed crops and food stores. Many more natives died from lack of food. The ancestors of the Kuikuru fled to the very large Lake Tafonuno, on the eastern edge of the Upper Xingii basin. Karaiba came again and killed more people. Some of the Carib-speaking resisted, and five groups of them joined to form a single settlement in the region of the Kulisefu (Kuliseu) River, from which the Kuikuru later moved to the lake they named Kufikugu. The grandfather of the grandfather of a contemporary elder witnessed this latter massacre, which might therefore have taken place around 1755. These same incidents are recalled by the Kalapalo in a “legend” told to the brothers Villas Boas, a famous team of pioneers and Indian agents. According to the Kalapalo account, many generations ago a very cruel white man called Pai Pero appeared in their settlement, which was then very far from their present location. A song that his companions sang in chorus as they walked, which was learned and is still sung by the Kalapalo, implies that those companions were other Indians. When Pai Pero left he took some of the natives with him. He later returned without these natives and asked more to go with him. When the natives refused, he and his men shot some and killed many others with machetes. Others ran and hid, and > OF course the cerms endogamy and exogamy have meaning only when used with reference to specified groups or categories. | have used chem with explicit reference to language group, tribe, local group (community), household, and kin categories. It is not “impossible to specify” the extent of endogamy or exogamy, nor does it obscure the range of native choices as Basso asserts (1970: 402, 411). ‘On the concrary. analysis of endogamy and exogamy of specific social units clarifies the extent and nature of those choices.when Pai Pero withdrew, carried by his men in a hammock as was his custom, natives shot an arrow into his hammock, He never returned (Ferreira [1951]: 78)- These incidents are also recorded in historical sources. “Bandeiras” of gold seekers and Indian fighters invaded the region of the Rio das Mortes some one hundred miles East of the Upper Xingii basin as early as 1663. At tha time many Indians were taken away as prisoners. A century later a notorious third-generation “bandeirante,” Antdnio Pires de Campos, Junior, made several expeditions to the same region, where he decimated Indians with sword and gun (“a ferro ea fogo”) because they were reported to be a menace to pioneer settlers in the region. In 1775 Pires de Campos with his band of 50 Bororo Indians, who called him Pai-Pird, again attacked a group of natives on the Rio das Mortes and “caused a terrible loss of life among the natives.” On this occasion he was wounded by an Indian arrow and was carried back to his headquarters in a hammock by his Bororo companions. He died of the wound soon afterward (Ferreira [1951]: 36-37). ‘The striking correspondence among these three versions indicates that che ancestors of at least some Upper Xingit Caribs were on the Rio das Mortes two centuries ago and that they fled westward into the Upper Xingé basin. Ic also indicates that a large number of them died as a result of hostilities with pioneers. When the first echnologists surveyed the Upper Xingé region in 1884 there were still some 3,000 natives there (Galvéo und Simées 1964: 136). Buc in the present century pioneers have continued to explore the area in search of gold and diamonds. More scientific expeditions were followed by missionaries, adventurers and journalists. Finally, in 1945, the Expedigio Roncador-Xingu entered the region to open Central Brazil to air traffic. No sooner had the work patty acrived than epidemics of influenza and other introduced diseases spread among the Indians (Ferreira [1951]: 77, 86) ‘Asa result of these contacts the first half of this century was a period of drastic depopulation in the Upper Xing region. In 1954 only about 700 people remained and the number of settlements had decreased. However, in spite of radical depopulation in the region as a whole, the size of the Kuikuru settlement remained relatively stable for many decades, partly by absorbing remnants of other Carib groups as well as some Arawak-speaking Yaulapiti families. It should be recognized, however, that the population decrease that caused remnant groups to coalesce in the past century is only the latest episode in a long process of social disruption as ancestors of the Kuikuru fled or were driven out of the Guiana region and came into conflict with new neighbors. Ic is common practice for remnant families from moribund tribes in the region to take refuge with larger friendly groups. Coalescing of these groups has important implications for the incidence of endogamy. Since refuges would not usually have been close relatives of the categories prohibited in marriage, their absorption has brought potential spouses into the Kuikuru community and thereby increased opportunities for focal endogamy. Thus marriage of Kuikuru co offspring of immigrant families, which would have been locally exogamous before the merger, is now locally endogamous. To establish whether or not amalgamation has in fact increased localendogamy, it must be determined which marriages took place before the mergers and which took place afterward. Although such data are not directly available, I have arrived at an approximation. Starting with the stated tribal affiliation of parents and estimated ages of their offspring based on informants’ recall of events for which dates have been published, it is possible to calculate che approximate dates of marriages. By comparing those dates with published dates of tribal mergers, | have made a determination in each instance as to whether a marriage took place before or after a merger. These calculations indicate chat the proportion of local endogamy has increased since the beginning of this century. This increase coincides with the period of marked reduction in both regional population and number of settlements in the Upper Xingd and with the period of recorded mergers of remnant tribes with the Kuikuru. At present about 75% of the Kuikuru marriages are locally endogamous. Since the Kuikuru generally regard the entire community as a cognatic kin group, marriage within the local group may be seen also as kin-group endogamy. It has been argued that serious depopulation leads to increased exogamy (Basso 1970: 414), but chat reasoning overlooks the effect of tribal mergers in maintaining the size of surviving communities. Only if and when there are no eligible women in the community must members of a group “marry out or die out.” So far this condition has been avoided among the Kuikuru and Kalapalo, in pare through tribal mergers. Cross cousin marviage Preferences vary regarding parcicular kin categories as potential mates. A patcern preferred by some informants is the marriage of persons regarded as cross cousins of some degree (éfgj). A young person's parents speak with his or her father’s “sister” (etsf) or mother’s "brother" (auajt), who are said to be "speakers" (Aitofo) in arranging such a marriage. If a young man's ets#, for example, has a daughter it is understood chat the aunt may “give” her daughter to the young man and that he may have sexual relations with (éupits? ake, “namorar”) or marry C'casar") his ifgf At present, cross cousin marriage constitutes only a small proportion of Kuikuru marriages and few marriages of persons classed as first or second cross cousin endure long enough to produce offspring. Similarly, marriages involving the exchange of siblings or parallel cousins are rare now, but data from personal genealogies indicate that these may have been more common in the recent past. (A numerical analysis of these data appears in Dole 1983). Informants insist thac arranged cross cousin betrochals are not binding, and, in fact father’s sister and mother's brother seldom become one’s parents-in-law. Rather, young people are said to make their own choice of mates before the parents negotiate a betrothal. Understandably, marriage with a lover (ajo) is preferred by some. It is said that such marriages are more enduring than other types and that arranged cross cousin marriages are sometimes broken by ajo. Special cross cousin terms are both statistically and functionally correlatedwith che cross cousin marriage among a large number of relatively unacculturated tribal peoples (see Dole 1972), and it is clear that traditionally among the Kuikuru, parents have a right to claim a cross niece or nephew as the first spouse for their offspring and that the term ifgfis retated to the norm of cross cousin marriage here (see Dole 1957, 1969). The Kuikuru themselves believe that cross cousin marriage was more common in the past. One man volunteered a statement that in former times people married their ifgj regularly ("bem direito”) but that today men want to marry other relatives, including mother's “sister.” That informant was well aware of the implications of the change, commenting that it ruined (“atropalha muito”) the system of kin relations. As we shall see, itdoes indeed contribute to ambiguity in the classification of relatives. ‘An interesting parallel to this situation is reported from the Vaupés region of Brazil, where the highest ranked sibs among the Tucanoan Uanano take care to adhere strictly to the rules of both language group and local exogamy because “a deviation [from.this practice] would upset the terminological system. ‘It would create chaos,’ and ‘you would not know what to call anyone’ " (Chernela 1983). The Kuikuru disagree among themselves about how the term ifgf may, o should, be used. Some informants stated that it should not be used in address because of embarrassment (ifisu; cf. Kalapalo iftitisz, Basso 1970, 1975), but others indicated that it is used even in address and that no ffisw is associated with its use. As informants, Kuikuru do not use the term in discussing marriage. Even though the norms and practices of marriage, eligible mates, extramarital relations, genealogies and kin terms were discussed with numerous individuals, who spoke freely on these subjects, the term sfgf did not surface in these discussions. On the other hand, it was used spontaneously outside the context of marriage discussions in the responses of a few informants when asked to identify relations with other Kuikuru, listed alphabetically. ‘A survey of current Kuikure martiages shows a tendency to marty into adjacent generations, although parents’ own siblings and one’s own siblings’ children are still prohibited in principle. A man may marry a woman who is classed as “niece” (ufati) if she is a “litele distant,” that is, che daughter of a female cousin and not of own sister. As Riviére notes with respect to the Trio, ctoss-generation marriages appear to be an adaptation of the cross cousin marriage pattern (Riviére, 1966a: 739). 1 would suggest further chat among the Kuikuru this form of adaptation may be an adjustment to the uccasional scarcity of cross cousins or other eligible mates in one’s own generation resulting from depopulation, as has already been documented for the Barama River Caribs (Adams n.d.) Parallel cousin marriage Even more disruptive to a system of cross cousin marriage is the marriage of parallel cousins, which occurs among the Kuikuru. To understand how this can occur we need only recall their general failure to distinguish cross from parallel relatives beyond first cousins. The use of sibling terms for all cousins allows for variable interpretation of relatedness. Marriage partners in many instances aremerely assumed to be related by virtue of the fact that their parents use kin terms fot each other. This fact has several interesting consequences: 1. The exact genealogical links to offspring of father’s “sister,” for example, may not be known. 2. The daughter of etsé (F"Si") may be a parallel as well as a coss cousin, 3. Ufgi as a term for the daughter of e*sf may even denote parallel cousins. 4, Hence marriage to wfgf may unite genealogically parallel cousins who are assumed to be cross cousins. In other words, although Kuikuru conceive of persons of che ifgf category as cross cousins they are often not, and are in fact often parallel kin, although in many instances the degree and kind of genealogical relation is not known. Telo ‘The inability to distinguish cross from parallel cousins beyond the first degree leads some people to prefer as mates persons who are not known or assumed to be related (te/o). The most general meaning of this term is “other, different,” but in the context of marriage it may be glossed as “non-kin.” Like kin terms, telo is quantifiable, as when a man refers to the daughter of a classificatory “sister” as "a little bit celo,” and a wish to marry a particular cousin may be rationalized by referring to the intended mate as telo. Unlike the so-called Dravidian system, referring to a relative in this way serves to place that person in a class of potential affines by equating him or her with outsiders rather than with cross relatives. It seems evident that in such instances telo is used purposefully as a device to imply social distance or lack of kinship bonds in order to express affinibitity. It suggests further that people who use zelo in this way may regard cross cousins as non-kin and therefore as potential affines. In this respect, it serves the same function as the traditional cross cousin term ifgj which it seems partially to replace. The same usage is found also among other Carib groups. As in the case of iff, cognates of telo are used with the meaning of cross cousin among the Kalifia (Ablbrinck 1931), Galibi (Arnaud 1968; Arnaud e Alves 1975), and Maroni River Caribs (Kloos 1971), all far removed in both space and time from the Upper Xingi Caribs, further supporting the suggestion that a tradition of distinguishing cross from parallel cousins is of long standing among these Caribs, and that a failure to distinguish chem is a departure from that tradition. Alliance: wife-givers, or kin vs. affines The fact that most Kuikuru marry persons other than cross cousins brings into question the existence of marriage alliances. Marriage alliance systems are usually maintained by continued intermarriage of units or kin categories that are identifiable either by separate location or distinctive labels, egocentric or socio- centric, The Kuikuru lack these structural patterns as we have seen. Without spatial separation or distinctive labels such as kin terms or groupnames, marriage alliances can be maintained only if the offspring of two families intermarty generation after generation. Since exchange marriages among relatives are rare among the Kuikuru, there are no regularly intermarrying units and no categories of affines “transmitted from one generation to the next” (see Dumont 1968: 205). Indeed, there is no recognized category of affines other than persons to whom actual or fictive marriage ties are traced. Included in the category of fictive ties are the special friends of parents-in-law, When a person marries, he or she acquires as parents-in-law not only his or her mate's parents, but also the latter's ato, persons with no previous kinship relations necessarily. It is clear that affinibility among the Kuikuru is not determined on the basis of a distinction between cross and parallel bonds as defined by ethnologists, or by membership in a particular category of relatives as in either cwo-line alliance o lineal systems. Rather it is determined on the basis of a variable concept of social distance measured in part by supposed or stipulated genealogical links, but also to some extent by interpersonal relations, place of residence and tribal affiliation. Any attempt to maintain the opposition of kin and affines as finite mutually exclusive categories is vitiated by the marriage of parallel relatives, the practice of unrestricted marriage to non-relatives, the assignment of fictive parents-in-law, and the common use of Generation cousin terms. Since cross and parallel ties are confused, except among the closest kin, and parallel kin beyond first cousins may marry each other, it is not the case here that all the paralle! kin of one’s parallel kin must be one’s own parallel kin, which is a diagnostic trait of the Cross Cousin two-line alliance system. Without a category of affines as opposed to consanguines, there is no division of wife givers or wife receivers. Of course cross cousins, persons from other communities, and non- relatives might be regarded as wife givers and receivers, but each of these categories is ambiguous or variable. No category of affinibles is characteristic of the society as a whole because both ideal and actual marriage practices differ from one person to another. Because cousins are included in the personal kindred and are also eligible as. mates, they can be seen as a type of consanguineal kin that are at the same time potential affines, in the same way as cousins were recognized as both kin and legitimate mates in many Stratified Segmented societies and in 19th-century England and America (see Dole 1957, 1965, 1972). Since the Kuikuru do not make a clear distinction between distant cross and parallel relatives based on marriageability, their kinship system, and specifically their use of ifgj and ze/o, do not fit either Dumont's structure of kin vs. affine (1953a, 1953b) or Yalman’s “closed circle” of kinsmen where there are no in-laws (1962: 553). Although Dumont characterized the Cross Cousin kinship system as dividing a society into kin and affines, he also suggested that “there is likely to be an affinal content in terms which are generally considered to connote consanguinity or'genealogical’ relationships (such as ‘mother’s brother’ etc.)” (1968: 205). A similar condition has been observed among the Trumai, another Upper Xingé tribe.In the field of kinship, between the two poles of consanguinity and affinity, there is a series of situations where an individual A can modify his relationships to an individual B, according to residence, factional aligament and individual desire. One can be a little, more, or less in a consanguineal relation (or affinal) (Monod-Becquelin 1978: 2) Ir seems probable that the distinction of kin from affines may be most pronounced and meaningful in strictly exogamous unilocal or lineal structures, where persons are classed as being either members of one’s own kin group or not members. By contrast, cognatic societies that have no prescriptive marriage rules but only prohibiced genealogical categories may class cross relatives as both kin and potential affines. In these societies a consanguine becomes an affine only as a result of marriage. Affinity is then a de facto category dependent upon marriage. It might be useful for the purpose of analyzing this type of structure to recognize a class of “consanguineal affines,” that is, kin with whom marriage is permitted. Summary and discussion Summarizing the data presented here, there are some indications of a native ideal of a patrifocal, if not patrilineal, structure, with patrilineal succession, patrilocal residence and a patrilineal emphasis in reckoning kinship. At the same time there are also suggestions of a two-line cross cousin marriage alliance system as an ideal, Alchough these patterns are quite compatible, the preponderance of observed practices do not conform to either pattern. In fact a salient feature of Kuikuru organization is chat, in spite of these central tendencies, social relations are marked by variation in individual ideal patterns and by deviations from those patterns in practice rather than by structural oppositions. Specific features thac differ from the ideal of a two-line system of kinship and marriage include the following: Cross relatives beyond first cousins are not distinguished from parallel kin Few marriages unite cross cousins. ‘The category of affines is distinct from cross relatives. There are no marriage alliances. Generation terms are commonly used for cousins. The cousin term used to designate eligible mates is of the Dakota-lroquois, not the Cross Cousin, pattern. 7. Avuncular and nepotic kin terms are of the Bifurcate Merging rather than Cross Cousin pattern. A comparison of norms with actual practices suggests that some of the variations in Kuikuru social relations can be interpreted as resulting from the relaxation of structural norms. In support of this suggestion, comparative linguistic dara, native statements about former practices, ethnohistoric evidence of demo- graphic disturbance, and statistical daca on increasing kin-group endogamy poinc to a change away from alliance structure with cither Dakota-roquois or Cross Cousin terms, such as has been documented for numerous other peoples (see Dole 1969). Data on marriage in the ethnographic literature have been notoriously aAYay Noinadequate, partly for lack of uniformity and precision in the use of the terms exogamy and endogamy, and partly because ethnographers have tended to take informant statements as representing actual customs. Nevertheless, it is clear that many societies with kinship nomenclature of the Bifurcate Generation pattern have undergone social changes that led to relaxation of kin-group exogamy and in some instances to the marriage of men with their parallel cousins or with daughters of sisters or female parallel cousins. All of these marriage patterns result in a terminological equation of parallel with cross relatives, thus removing the principal social distinction on which special cross cousin terms are based. Such a change is not necessarily very recent among the Kuikuru; it may well have occurred gradually over centuries as part of general social adjustments. Nor do T maintain that kin-group endogamy is the only cause of a change to Generation cousin terms. However, cross-cultural data demonstrate that it is a major determinant. Generation terms are not only statistically correlated with kin-group endogamy in world ethnography but are functionally related as well, whereas the Dakota-Iroquois and Cross Cousin patterns are correlated with local kin-group exogamy (Dole 1957, 1969; see also Murdock 1947) Tn the context of change the continued use of cross cousin terms is an example of culture lag, When a marriage practice lapses there is a tendency to retain the traditional kin terms to “clothe” a newer practice. An instance of this tendency occurs among the Guiana Carib Wayana, When a Wayana man marries a parallel cousin, he changes the status of her brother from parallel co cross relative by using the term for cross cousin (Lapointe 1970; 124). In this way relationships may be altered to conform to a structural norm and the society retains the fiction of conforming to the norm. The suggestion of modification in the use of kin terms as a result of changing marriage customs has been misinterpreted as evolutionist® by several social anthropologists, including Basso, who rejected it as unnecessary to an understan- ding of "Xingé Carib” kinship. Instead Basso attempted to explain the “internal logic of kin-class assignment in this system’ through a" ‘conceptual analysis’ in the fashion of Schneider” (Selby 1972: 307). Assuming a “common cultural model in use among the Xingti Caribs” she described an ideal two-line marriage alliance system based on the Kalapalo native model and attributed thar system without qualification to ail other Caribs in the region (1970; 404 ff.). However, the Kuikuru data differ in numerous respects from the structure represented for the Kalapalo. The alliance structure reported for the Kalapalo does not occur among the Kuikuru and the kinship nomenctatures differ somewhat both in form and use. There is clearly no single "Xing Carib” terminology and marriage pattern, Explaining the system of kinship cerminology and marriage of a society on the * This hypothesis has been misinterpreted as an evolutionist explanation of adaptive modification ina "sequence of developmental stages” (Basso 1970: 415-416). Far from being an example of evolution as measured by its hallmark, that is, increase in complexity, the loss of structural opposition characteristic of a two-line alliance system posited in this instance is a regressive or devolutionary step, because it represents a decrease in complexity. (For an evolutionist interpretation of kinship see Dole 1972)basis of a native model or “specific rules of behavior” (Basso 1970: 403) leaves several questions unanswered. In the Kuikuru instance it is not the native model or ideal that needs to be explained. The use of a distinctive cross cousin term and its relation to cross cousin martiage are not in question (see Dole 1957, 1969). In this instance the native model represents an ideal to which actual practice does not conform, and the model cannot account for the extensive deviations from it. Thus a description of Kuikuru organization in terms of the two-line model alone is of dubious utility and appears to be generated by the model rather than by empirical data of Kuikuru practice. What does need to be explained is the general lack of conformity of existing practices with a native model in boch marriage and kin terminology. Why do cross cousins seldom marry each other, and why do people commonly use the same terms for cross cousins as for parallel, instead of the distinctive cross cousin term? A related problem is a lack of internal consistency in the kinship nomenclature; the Generation pattern of cousin terms is inconsistent with the Bifurcate Merging pattern of terms used in adjacent generations. Why is the cross cousin term not used in general conversation here, whereas in other societies with two-line alliance structure cross cousin terms are a regular and overt part of the common kinship nomenclature? In other words why is the special cross cousin term so loosely integrated in che kinship organization? Basso explained the common use of Generation cousin terms as being the polite way to refer to potential affines, who would be embarrassed to hear themselves referred to as in-laws (1970: 412). One must ask, however, why it is awkward or impolite to refer to a potential spouse with a cross cousin term in Kalapalo or Kuikuru society and not in all other societies with cross cousin terms of the Dakota-Iroquois or Cross Cousin patcern. What is it in this instance that makes it embarrassing to specify affinibility? In view of the historical and ethnographic evidence for change in Kuikuru society, it seems unproductive co deny the relevance of social change to these problems. The model of a two-line marriage alliance is a static, internally consistent and well integrated structure, recognition of which is of undisputed value co ethnological science. In essence, it describes how a particular system should work. Buc the Kuikuru and most contemporary social systems are not as stable, internally consistent or well integrated as this model. A functional-dynamic approach may complement a structural analysis in reaching a beccer understanding of social systems that have undergone change. In a discussion of cultural integration, Robert Anderson, following Linton (1936) and Kroeber (1948), assumed that various parts of a culture tend to become progressively well integrated and that "older complexes and elements will be better incegraced inco their surrounding systems than newer ones.” He showed that cultures in che process of change could be identified by the occurrence of variant practices and that the degree of cultural integration could be measured by the incidence of “lesser and greater numbers of variant behaviors” (Anderson 1960: 51-52). These principles can be applied to the Kuikuru data, and specifically to the use of kin terms for cousins.In small societies, such as the Kuikuru, latent tensions are a common feature of the social distance that characterizes affinal relations. Hostility between affines is often minimized mainly by ritual avoidance and exchanges, and by the formal etiquette of marriage proposals. In the ambivalent relations among affines, referring to a person as an affine implies a negative connotation of being an outsider as well as the positive implication of attraction to a potential spouse. It is because of the negative connotation in this relation that the use of affinal terms causes embarrassment. Of course, if a community is exogamous, its members can discuss potential affines freely without being heard by them and hence without embarrassing them It is only when prospective mates live in the sane community that such discussions would be overheard by the porential affines and the use of cross cousin terms with their connotation of potential affinity would be a source of embarrassment. Then people must be circumspect and use affinal terms only out of hearing of the referents. If the Kuikuru regularly practiced local exogamy, their eligible mates would not hear the affinal terms used in discussions of prospective marriage because they would be in separate local groups. However, for at least a century and probably much longer the Kuikuru, and the Kalapalo as well have from time to time incorporated remnants of other previously separate groups and have married within their own communities to some extent. Ic is a very common custom to extend consanguineal terms in address to all members of a community as a gesture of courtesy and as a gentle way of manipulating relations. Since eligible mates live in the same group among the Kuikuru and Kalapalo today, sibling terms are extended to them in this way. And since public use of affinal/cross cousin terms in this situation is awkward, it is not difficult co see that “sibling” terms might be extended to cross cousins in reference as well as address. Although Basso rejects a consideration of change as contributing co an understanding of Bifurcate Generation terminology there are inconsistencies in her material that can be reconciled only on the basis of variation and change. People who are referred to as ifay among the Kalapalo are said to be “the only kinsmen considered marriageable” (1970: 409), buc at che same time “marriage with an ifag is not an ideal” (1970: 411). In another place we read that only chose ufau sic] who are not “close” actually marry. Thus, children of biological siblings af opposite sex, though classed as martiageable never seem to actually marry. Only those distant persons..are actually married, that is, persons outside one’s factional and household group, but usually within linguistic and kinship categories (Becker 1969: 65-66) The genealogical relations of persons referred to as ifag (or ifau) are not clear, for “non-kin who are married may be called ifau, but no claims are made to a kinship relation” (Becker 1969: 81). And finally, “the Kalapalo cannot clearly state why ifandaw [the plural form of ifag] are appropriate spouses and lovers, and they are unable to explain why they identify certain kinsmen as potential affines and designate others as unaffinible” (Basso 1975: 209, emphasis mine).From these statements, it seems clear that the Kalapalo have a tradition of cross cousin marriage but that current practice does not conform.closely.to that tradition. These statements reflect a variable condition similar to that among the Kuikuru. The relevance of individual variation and change to a social system is expressed by Firch in distinguishing beewcen social structure and social organization. It is becoming increasingly clear that in order to understand both change in structure and change in detail, we must look to a closer study of the setting and results of individual choice and decision, as they affect activity and social relations (1954: 17). To overlook the factor of change is to close one avenue to an understanding of the individual choice and decision” that are characteristic of Kuikuru society. Abstract The Kuikuru are one of three Carib-speaking tribes inthe Upper Xingu basin in Mato Grosso, Brazil. Their cognatic social organization is very permissive, and behavior is characterized by an unusual amount of variation in both norms and actual behavior, with few strict regulations and fewer formal penalties for nonconformity. Nevertheless, some of its social relations are ordered by concepts of opposition and lineality. This paper outlines ideal and actual practices in spheres of leadership, economic organization, residence and kinship. Lineal and cognatic features are described, and the evidence for two-line marrige alliance structure is examined. Some characteristic features of the system of kinship and marriage are analyzed in the context of demographic and social change. Resumen Los Kuikuru constituyen una de tres tribus de habla Caribe dela cuenca del Alto Xingu, Mato Grosso, Brasil, La organizacién cogndtica de estos indigenas no es muy restrictiva y el comportamiento social se caracteriza por una variacién excepcional no s6lo en cuanto a las normas, sino también en cuanto a la conducta actual. Hay pocos reglamentos estrictos y no se penaliza la disconformidad. Sin embargo, algunas de sus relaciones sociales se organizan por conceptos de oposicién y aun la linealidad. Este trabajo presenta un esbozo de las practicas ideales y reales en cuanto al liderazgo, la organizacién econémica, la residencia y el parentesco. Se sertalan los rasgos lineales y cogndticos y se examina la evidencia que existe en cuanto aalianzas matrimoniales de dos lineas. Se analizan algunos rasgos caracterésticos del sistema de parentesco y de las alianzas matrimoniales en el contexto del cambio demogrifico y social.