Volume 66 1957 > Volume 66, No. 4 > Variations in Polynesian social organization, by Irving Goldman, p 374-390
                                                                                             Previous | Next   

- 374

ONE OF THE leading questions in evolutionary theory concerns the transformation of kinship-organized societies into social systems governed primarily by political and territorial principles. While this question has received due attention, 2 the processes by which this transformation occurs have yet to be adequately described and set within a completely satisfactory theory. There are, to be sure, inherent difficulties in reconstructing processes that occurred mainly beyond the direct observation of historians. Still, we have reason to believe that many of these difficulties can be mitigated by appropriate theory and method. To suggest procedure and theory towards an understanding of these processes is the intent of this paper. I shall draw my illustrative material from Polynesia where political and territorial systems were becoming established at the time of European contact in the late 18th century.

Dr. Goodenough has explained the rationale of studying evolution within a culture area. To what he has already said, I should like to add the following: The comparative study of historically-related societies composing a culture area brings us directly to the central problem of evolution, namely, variation. It is through an understanding of the characteristic variations of particular social systems that we can hope to reconstruct evolutionary processes, directions and stages. As long as we are assured of the genetic relationship of variant types, the reconstruction of their patterns of development need not be much more speculative than the usual functionalist analysis of a single society. Elements of social organization being relatively orderly follow a logical pattern of development and so are particularly amenable to historical reconstruction, 3 providing they can be shown to be variant forms.

The choice of an island area, such as Oceania, is a fortunate one. As a result of their relative cultural isolation, islands are more sheltered from the overwhelming effects of diffusion, and variant forms are more clearly observable. In this respect, the Polynesian islands, the most removed from continental masses, are most favourably situated for our - 375 purposes. 4 We find in Polynesia one other favourable condition for evolutionary study, the existence of genealogical traditions that, used circumspectly, have historical value.

Finally, the culture area approach adds historicity to evolutionary theory that commonly suffers from undue abstractness. Historical detail gives substance to theory and perhaps new insights.

The theory I am proposing is briefly this: The main course of cultural evolution in Polynesia has been governed by status rivalry. A constant in every Polynesian society, status rivalry seems to have provoked stresses to which the status system, along with other components of the social structure, adapted. These adaptations rearranged the total social structure producing new structural types. Each type, in turn, had its own pattern of adaptation, so that, once established, a type tended to develop along new lines. Having described what I consider to be the broad lines of cultural evolution in Polynesia in a previous paper, 5 I shall confine my discussion to a more detailed examination of those changes in social organization that seem to have been involved in the evolution of political systems.

We may consider first a hypothetical model of a proto-type Polynesian social organization. The model has been constructed by the usual method of comparing and evaluating widespread and generalized features against narrowly distributed and specialized features, and relating these to what we know about the functional characteristics of Polynesian social organization. Underlying similarities of Polynesian social organization are, in fact, so marked that the model is, in the main, actually characteristic of many Polynesian societies; and its principal elements are discernible in all. The model can be best comprehended from a developmental point of view. Thus, we visualize a founding family, settled in an area, expanding by branching into a group of related families forming an enlarged descent group. The descent group, in turn, enlarges by branching until a group of related descent groups form a tribe. The tribe fills out a demarcated ecological or politically bounded zone. The head of the tribe is the senior male of the senior descent line, and the head of the descent line is the senior male of the senior family. Seniority of descent defines social status, along with political, economic and religious leadership. Genealogical links as well as the concentration of the principal sources of leadership along the lines of male primogeniture imparts virtually an organic unity to the tribe.

In the literature on Polynesia the descent group has been commonly called a “lineage.” The term is apt even though the Polynesian lineage is atypical in that it is associated with a bilateral kinship system and - 376 normally lacks exogamy and strict sex linearity. It is neither patrilineage nor matrilineage but a “Status Lineage” in that it traces descent along status lines, usually through male seniority. 6

From this idealized model we can begin to chart a series of variations. From one point of view, the variations range from a strong emphasis upon patrilineal descent to complete bilaterality at all levels of social, economic and political integration. From another point of view, we see variations in the degree of genealogical unity among members of descent groups and of tribes. These two modes of variation—in linearity of descent and in the genealogical unity of descent groups—are, in fact, related. It is the thesis of this paper that both interact with the most dynamic aspect of Polynesian social structure, the system of social status. Polynesian social structure was unusually sensitive to changes in the system of status because virtually all significant directive functions were theoretically concentrated in the senior descent lines. In presenting this thesis, I must emphasize its tentative character, and note the fact that data on Polynesian social organization are seriously lacking in detail.

We may turn now to a consideration of the variations in the Polynesian systems of social status. I have arranged 18 Polynesian societies in the following order which I believe to be roughly an evolutionary sequence, starting with Ontong Java at the bottom and ending with the Hawaiian Islands at the top: Traditional societies—Ontong Java, Tokelau, Tikopia, Pukapuka, Manihiki-Rakahanga, Maori, Tongareva, Uvea, Futuna, Manu'a. Open societies—Niue, Easter, Mangaia, Marquesas. Stratified societies—Tonga, Mangareva, Tahiti, Hawaii. I have stated my reasons for regarding this as an evolutionary sequence in a previous paper, already referred to, and I shall therefore, not repeat them here. The hypothetical sequence is based upon criteria of “politicalization of status.” By this term I mean the degree to which status has acquired the attributes of coercive power in political and in economic terms. I regard the sequence as a gradient or continuum that divides into three distinctive types. The first type I have called “Traditional” because it represents, by and large, the prototype system of social status in Polynesia. In the Traditional type social status is derived primarily from seniority of descent preferably through males, and resembles a pyramid of graded rank based upon degree of genealogical relationship to the senior lines. The ten societies in this group are not all alike. Each is a variant of the central tendency. Those societies clustered at the middle are to be regarded as most characteristic of the type, and those at the ends as the most aberrant. Thus, - 377 Ontong Java, Tokelau and Pukapuka are in transition from status based upon age to status based upon seniority of descent. Tongareva, Uvea, Futuna and Manu'a, on the other hand, are an ascending series moving towards the next type, the “Open,” characterized by an open contest for power and status through political and military means. In the Open societies—except for Niue, a problematical case—the tradition of rank through seniority of descent persisted, but the practices had changed markedly. Among the Open societies the Marquesas illustrate a transition to the last type, the “Stratified.” In the Stratified societies control over lands and with it political power had become the decisive source of status, even though the traditional forms of hereditary rank were still highly respected. However, hereditary rank alone gave little save prestige unless combined with land ownership. The fundamental status division was between landed and landless. Among the three status systems the Stratified appears clearly as a new type, while the Open, though typologically significant, is to be regarded as transitional. Thus, there is a question as to whether Samoa should be grouped with the Traditional or with the Open, but there is little doubt as to the placing of the line between Stratified and Open.

In sum, Polynesian status systems appear to have developed along these lines: (1) The criteria for social status shifted from age, to seniority of descent, to unstabilized power over people and, finally, to control over lands and thereby to relatively stabilized power over people; (2) the prerogatives of status evolved from an emphasis upon prestige and leadership to an emphasis upon prestige and power; (3) the hierarchical ordering of society evolved from relative equality to graded rank, to temporary cleavages between weak and powerful and, finally, to a more rigid division between landed and landless. In all these changes the tradition of hereditary rank through seniority persisted.

These basic changes in systems of social status were, I believe, instrumental in the transformation of the total social structure, except for the kinship system which remained consistently bilateral. This is not to say that status was a sole determinant. Economic, political, military and other factors were also involved in these social changes. I am suggesting, however, that it was the status system that played a central role because it was the pivot, so to speak, of the entire social structure. Under the influence of status rivalry and the politicalization of status, the status lineages tended to become more bilateral as well as less cohesive. The village organization, an expression of the common interests of kin groups, was replaced by dispersed hamlets as families attached themselves to the households of powerful chiefs (Open societies) and to the estates of landowners (Stratified societies). Tribal organization tended to achieve its greatest significance in the warlike and contentious Open societies only to be obliterated by conquest and the fundamental rearrangement of landholdings that followed in the Stratified.

- 378

In the Traditional societies descent groups formed themselves around a line of firstborn, preferentially males, and conducted all essential regulatory economic, social and religious functions except that of exogamy. Community organization was of the village type, village councils representing the interests of households and family lines and of lineages if more than one occupied a village.

ONTONG JAVA: Hogbin refers to “joint family” (manava kanaka—“male belly”) that included all persons who could trace their descent through males back to a common ancestor five or six generations. 7 The joint family was the core of the fishing group and owned coconut groves. Its headman was its eldest male who was often the priest. A number of joint families made up a tribe and occupied a main village. There were also matrilineal features. Residence was matrilocal; and a descent line of women owned taro gardens and houses in the main village. It is noteworthy that status considerations altered the alignment of kin. Thus, if the wife's family were poor, residence was patrilocal; if both sides were poor, it was matrilocal; and if both were wealthy, it was bilocal. Well-to-do families had totems and were cohesive, being able to attract membership through the maternal side. Poor joint families, having few economic interests in common, had no totems and were loosely held together. Tribal unity was mainly religious and chiefly authority was weak; but the chieftainship grew stronger after one joint family had succeeded in asserting its domination.

TOKELAU: 8 The descent group has been identified as a “kindred.” Although descent was mainly patrilineal, residence was largely matrilocal and property rights were to all intents and purposes determined bilaterally. Maternal kin, for example, determined many social functions. Leadership, on the other hand, went usually in the male line, according to principles of age seniority but with some tendency to recognize seniority of descent. Tokelauan society seems to present a compromise between a lineage principle, expressed in the importance given to lineal descent from a common ancestor with preferential patriliny, and a kindred principle expressed in the equating on the community level of maternal and paternal kin. As long as effective leadership remained with the land-controlling local group, the kindred principle was most important. The lineage principle, on the other hand, regulated the arikiship, essentially a priestly office, but with capabilities, as we have seen, of gaining in secular prominence.

TIKOPIA: To quote Firth: “The Tikopia lineages are patrilineal, membership being traced through the father along the male line to an original male ancestor. The lineage alignment is an important element in the Tikopia social structure. Though it does not regulate marriage . . . it is the basis for land-holding and use, it provides significant units for many kinds of exchange of goods and services at marriage, funerals, initiation ceremonies; and within the clan structure it has a basic part to play in the religious system. The senior man in each lineage acts as its major representative in public affairs.” 9 The Tikopia term for lineage is paito (house). The paito were formed by branching and were hierarchically ranked. They were subdivisions of a “clan” (kainana) which is preferably translated as - 379 “tribe” in accordance with usage elsewhere in Polynesia. The village was mixed but had a preponderance of paito of one kainana. The entire island was divided between two districts that were rivalrous and showed tendencies of overpowering the kinship bonds. In this respect, Tikopia foreshadowed events and developments that emerged fully in the more politically developed Polynesian societies.

PUKAPUKA: 10 There were maternal as well as paternal lineages, the latter being most important. The paternal lineages (po) did not of course regulate marriage, but controlled garden lands, coconut groves and houses, all inherited through the father. The paternal lineages had their own gods and priests and their secular leaders. Residence was patrilocal and succession to priestly title went in known paternal lineages. Each of the four lineages named a chief, one of them the high chief. The four lineages used to be equal, but subsequently one chief established his as the dominant lineage. Village membership cut across lineage lines. Village groups were important economic units; they divided fishing grounds, and owned lands on the three islets of the atoll as well as taro reserves. An informal village council handled property distributions and organized group labour. Villages were patrilocal.

MANIHIKI-RAKAHANGA: 11 According to native traditions, the two linked atolls were settled by a single biological family. Tribes (matakeinanga) were subdivided into sub-tribes (tukuwhare). At one time a single ariki who was a priest-chief, ruled. Subsequently, the arikiship was divided, one ruler representing the welfare of the land and food growing on it and the other holding power over the phenomena of nature. The arikiship descended in the line of male primogeniture. Tribal chiefs (whakamaru) were the custodians of tribal lands, responsible for their periodic redistribution. We do not know how the whakamaru was chosen. According to Buck, “It is to be presumed . . . that the whakamaru was appointed by succession in the male line from the leading family in each group.” 12 This would follow from the genealogical evidence that the first office of land distributor (tuha whenua) descended in the senior lineage, where “a female break in the ariki line was usually fatal to the succession.” 13 Households were, of course, bilateral in that land was inherited from both sides. Even so, since residence was predominantly patrilocal and since family heads were the senior males, the lineage principle operated on the household as well as on the descent group level. The four matakeinanga occupied one village, maintaining their divisions.

MAORI: 14 The descent group (hapu) was not purely unilateral, but it was built around an ancestor of note. Lineage membership was ideally traced through males and residence was preferentially patrilocal. When it was convenient to do so, however, the relationship to the distinguished ancestor could be traced through women, and residence could be matrilocal. The highest-ranking ariki, however, were expected to be descended from a line of senior males. A large hapu occupied a village (kainga); and if there were several hapu each occupied a fixed quarter of the village. Related hapu formed a tribe (iwi). The most closely co-operating social unit was the - 380 extended family (whanau) commonly patrilocal. Insofar as land tenure went, there were a series of jurisdictions that recognized tribal rights, lineage rights and household rights. Among these, however, the hapu, usually the occupant of a village, a co-operative community, was for that reason the social unit of outstanding importance. Maori represents the status lineage in a most characteristic form.

TONGAREVA: 15 From Buck's account, Tongareva was very similar to Maori. Male primogeniture established the main status and leadership positions; descent was patrilineal and residence patrilocal in the main. We are given no term for descent group. It is to be assumed, however, that the local groups occupying a common territory and claiming lineal descent from the same ancestor are the equivalent of a lineage. “Every district occupied by a number of families had its senior family which automatically supplied the district chief through primogeniture. When all the districts in an island or division of a large island combined for mutual protection or aggression, one chief by reason of the common descent from one ancestral family was senior to all the others. His pedigree showed that his descent was from the senior family in each generation, and he was thus senior to all the other chiefs whose families had branched off from the main stem (toro ahuru).” 16 As in New Zealand, circumstances of wealth, status, or lack of male offspring led to matrilocal residence, and even to the absorption of the son-in-law into his wife's kin. These were exceptional cases, however, and did not affect the structure of the lineage. The patrilocal household was the landowning group, but actually arbor-culture with coconut trees favoured strong individual ownership. Tongareva was also exceptional (for a Traditional society) in lacking villages. Families lived among their coconut groves to guard the trees against theft, but were able to get together to discuss common affairs.

UVEA: 17 The descent group (kainga) included all who could trace descent from a common ancestor usually through males. Burrows defines the kainga as a “lineage” and has remarked that the degrees of relationships included in it were not rigidly fixed but were recognized only insofar as they were serviceable. The services they rendered included shares in produce of the common land and co-operation in feasts, attending births, marriages, deaths. The lineages were the land-owning group. Chiefly lineages carrying high prestige were larger because they attracted more families. At the same time, unwieldy lineages tended to split up. The village was an important co-operative unit sharing and allocating taro beds.

FUTUNA: 18 Burrows has defined the descent group (kutunga) as a “kindred” in recognition of the fact that property and status sometimes passed through a line of females. Patriliny and patrilocal residence were, nevertheless, the preferred modes of transmitting rank, titles and property. Futunan social structure resembled that of Uvea in all fundamental respects so that there is no real cause to speak of one as “kindred” and the other as “lineage.” Both have status lineages.

MANU'A: 19 According to Mead, the descent group (aiga) “. . . is a curious bilateral grouping in which all the descendants in the male line are - 381 balanced against the descendants of the women in the family.” 20 The male line, the most important land-holding group, was normally localized while the female line was dispersed. Nevertheless, status leadership and property rights moved freely in either line; residence was ambilocal, one-third of marriages resulting in matrilocal households. Manu'a seems to be the most bilateral of the Traditional societies. This bilaterality together with the absence—apparently through decay—of primogeniture and the encroachment of the village organization upon the traditional prerogatives of rank gives to the Manu'an descent group a different character from any we have yet described. In Manu'a it seems evident that the lineage principle had yielded to the pressure of status mobility. In place of lineage a kindred structure developed, in which local interests rather than traditional genealogical relationships ruled the formation of kin groups. According to information compiled by Williamson, the situation was essentially the same in Western Samoa. 21


In the Open societies, where the traditional forms of status and leadership had given way to a more mobile system, bilaterality, as might be expected, tended to become more strongly established. Attendant changes included the loss of village organization, growth in political importance of tribe along with the formation through military conquest of a multi-tribal organization. Land-holding tended to shift out of the lineage to be directly controlled by smaller kin groups. Where information is adequate it reveals with some clarity how status rivalry encouraged bilaterality, and at the same time dispersed the lineages. Military conquest and the growth of political power over lands and people were the specific instruments of these changes.

NIUE: 22 While data on social structure are unfortunately unclear, they suggest bilaterally extended families rather than lineages. According to Loeb, Niuean society was organized into “families” (fagai—“people who eat together”). In earlier times one “family” occupied a village; but more recently, villages have included several fagai. The fagai owned property in common and was supervised by a family head; residence was patrilocal and a man preferred to have his eldest son inherit the largest land portion. Presumably, in this instance, Loeb was referring to a household. In another context he identified the Niuean social system with the “Melanesian type (that) goes back only three or four generations, but in all collateral lines.” 23 Niuean genealogies went back five generations, and at the most seven. The island was divided into two warring endogamous “moieties,” one of which emerged victorious and established chiefship (patuiki) over the entire island. Warfare was chronic involving villages and families in feuds. In time of war people abandoned their villages to scatter among the coconut groves. Niue had neither priesthood nor aristocracy.

EASTER ISLAND: 24 Because of early depopulation, information on social organization is meagre and therefore of uncertain reliability. Metraux, - 382 evaluating scant and discrepant data, has pictured a division into ten paternal descent groups which he calls “tribes” (mata). Each tribe at one time had its own territory which was known by the name of that group. “Later the tribes became more scattered and isolated members or entire households of one tribe lived within the limits of another district.” 25 Frequent wars scattered the defeated populations which were then absorbed into the districts of their conquerors as subjects. Thus, some tribes had no districts of their own but were scattered and intermixed. Tribes were divided into apparently localized lineages that were the principal land-holding groups. An ariki-mau, a descendant of the senior lines, was mainly a religious figure. He had high social status, but the active authority was with the warriors who had achieved their position of tribal chief by valour. It seems probable that Easter Island had a lineage system that had become dispersed in warfare. There is no information on bilatcrality other than the observation that children inherited from both parents. Residence, on the other hand, was evidently patrilocal. “Village” sites are referred to in reconaissance, but we have no description of a village organization.

MANGAIA: 26 As described by Buck, the formal social structure of Mangaia resembled that of the Traditional societies. Seniority in the male line established the line of ariki households; descent lines and tribes were strongly patrilineal and patrilocal, and rank was graded according to genealogical relationship to the senior line. The lineage principle defined leadership and land rights and established the bases of blood unity within the tribe. Moreover, the Mangaian lineages were uniquely exogamous.

This was the theory of Mangaian social structure. In the departure of practices from this idealized version we begin to see more fully some of the main processes of transformation of Polynesian social structure. The details of Mangaian history leave little doubt that the traditional forms were the shadow and that the substance was a sharp struggle for status and power. In the course of wars the arikiship was reduced to an honorific position, and the active authority was seized by successful warriors, so that it was no longer hereditary. Vanquished peoples deprived of their hereditary lands were attached to the victors' households as dependents. The seized lands were then distributed among the leading warriors. Weak families sought refuge among protective kin of either the maternal or the paternal side. Thus, wealth and power attracted dependents and realigned local groups which, under the circumstances, were hardly patrilineal let alone kin-unified. Land holdings, dispersed and subject to laws of conquest, were administered by appointed warrior chiefs. Mangaia lacked villages, their place apparently having been taken by the settlements of warrior families. How these historical events reshaped the social structure has not been brought out in the literature on Mangaia. That the lineages were disrupted seems evident from the facts. We are interested, however, in the nature of the disruption. On this point we may say that the lineage fragmented. It continued to regulate prestige-giving hereditary rank but not the more significant sources of power and authority. These, we must assume, came more strongly under the jurisdiction of the heads of bilateral extended families. In short, the lineage persisted in theory for the society as a whole and survived as a functioning institution for the ariki lines only. Since the ariki had lost their authority, the lineages accordingly lost their significance.

- 383

MARQUESAS: 27 Although the social structure is not fully described, the evidence for bilaterality is positive enough. There were only two social units, tribe and household, the enlarged descent group apparently not being present at all. Households were run bilaterally and the tribe was evidently a kindred. However, rank went by primogeniture, in theory, irrespective of sex, so that the lineage principle regulated one aspect of status. In fact the genealogies of some important families went back twenty-five generations. But, as in all Open societies, political influence measured in wealth and control over people was a more significant attribute of chieftainship than hereditary rank. Although the tribe was a kindred and administered by a kin chief, there were desertions for cause. People joined other tribes so that tribal organization came to be increasingly political, even though the fiction of kinship was maintained. As in Mangaia, popular and effective leaders established centres of population growth around themselves and in this way advanced their own status and influence. In some parts of the Marquesas, chiefs overstepped the narrow line between stewardship over kin lands to landlordism. In so doing, they paved the way for the “feudal” land holding systems of the Stratified societies. There were no villages, their place, as in Mangaia, having been taken by the enlarged households of leading chiefs.


The Stratified societies achieved greater stability of status as compared with the Open societies. They did this by retaining genealogical claims to rank and by endowing rank with more permanent attributes of power among which the more or less outright ownership of land was the most important. Economic stratification has obvious connotations for social structure in the sense that it creates sharp social cleavages. In the Traditional societies kin unity was relatively strong; in the Open it had weakened, even though the fiction of unity survived. In the Stratified societies kin unity broke at the class line. Commoners were a distinct and, from a kinship point of view, alien class. The data unfortunately are not always clear enough to illustrate the full implications for social structure of these major political and economic changes. But available evidence suggests that the fragmentation of lineages, noted in Mangaia, was the common process. That is to say, that the ranked segment of the population held to the lineages, whereas for the commoners the units of organization were household and kindred. In all of the Stratified societies villages did not exist; and in Tahiti and in Hawaii even the tribal organization had given way to a district organization that was primarily political.

TONGA: 28 Patrilineal lineages (haa) consisted of a nucleus of related chiefs about whom were grouped status-inferior relatives, the lowest and most remote of whom were commoners. The lineages tended to be localized through patrilocal residence and were hierarchically ranked according to the status of their leading chiefs. Membership was sometimes sought in the maternal lineage if that ranked higher. The theory of Tongan social structure corresponded to that of the Traditional societies insofar as primogeniture governed graded rank and established the principal leadership - 384 positions within a kin-unified corporate body, the haa. At the same time, Gifford has noted that: “Everything points to the necessity of a line of powerful chiefs for a nucleus about which the lineage groups itself. Without such chiefs it appears to wilt and die, and its membership gradually aligns itself with other rising lineages. This process of realignment naturally contravenes the rule of patrilineal descent, which theoretically and largely in practice determines lineage membership.” 29 Chiefs “owned all the land and could take what they pleased from their tenants and even kill them.” 30 They headed greatly enlarged household establishments that included inferior relatives as menials, as well as wife's kinsmen.

The question of land ownership is of first importance because it is land tenure that determines the true character of a descent group. If the meagre data on this subject from Tonga are at all correct, they support our view of the haa as a quasi-political structure in which power motives had begun to replace kinship as the basis of organization. Alexander distinguishes the Tongan land system from that of New Zealand and Samoa calling it a “fully developed feudal system” in which the “head of the clan has become a landlord.” 31 Nobles received land from the “King” in return for services and tribute. Land grants as a reward for political services endowed their recipients with many prerogatives of nobility. The degree to which the kin ties between commoners and chiefs had become more theoretical than real is brought out by items of information about commoners such as these: their uncertain occupancy of land; the unusual and capricious cruelty with which they were treated; their lack of voice in the fono; the belief that they had no surviving souls; 32 and their lack of redress against chiefs. There were no villages, the population being dispersed among the land holdings of the nobility.

MANGAREVA: 33 Here, the transformation of the descent group seems to have gone much further than in Tonga. Buck has noted that there were no family lineages, but extended families that grew in numbers until they filled out an ecological zone. Though bilateral, these extended families traced their ancestry through the lines of male primogeniture back to their most distinguished founders. Patrilocal residence maintained them as a local group and strengthened what to all intents and purposes was a lineage organization. The theory of social structure was, therefore, traditional. The practices were otherwise. Mangarevan history, as revealed by its genealogical traditions, has been one of chronic warfare in which senior lines were disarranged and in which war leaders assumed great power. At one time commoners ruled the island. It was presumably in the course of these struggles that land was alienated from the commoners and held exclusively by a landed gentry. Victorious chiefs rewarded allies with land grants; the defeated were expropriated and the weak yielded up their lands to protectors. While seniority of descent gave prestige, valour and land ownership gave authority. It was authority and power around which social groups formed. People grouped themselves around strong chiefs irrespective of kinship. The chief's domain included kinsmen, conquered people, refugees, and groups upon whom alliances had been forced. While some land descended in family lines, most holdings seem to have been negotiated leaseholds. There were no villages, - 385 the population being distributed among the plantations. Landless commoners were lowly-regarded fishermen and lived in scattered households along the coast. In short, the lineage persevered in the kingship and royalty, but was evidently a fiction for all others.

TAHITI: 34 For all the abundance of published material on the Society Islands, data on social structure are surprisingly scarce and far from clear. The best single source seems to be the Arii Taimai chronicle as recorded by Henry Adams. Yet neither Taimai nor any other source for that matter refers much to family organization. In the main the paucity of information on family life is probably the customary bias of early ethnography. It is also likely that Tahitian society had become so politicalized that the traditional groupings of lineage and tribe had become more or less unrecognizably absorbed into the district organization. According to Taimai only one of the districts of Tahiti was occupied (in her time) by a “clan,” presumably a kin-unified people. Within this district of the Teva, one lineage held, through seniority, the highest ranking ariki title and was genealogically linked with other leading families in the district. Elsewhere in Tahiti, wars had disturbed the occupancy of districts, their lines of seniority and the lines of authority. The districts were called mataeinaa, a term that connotes tribe elsewhere in Polynesia. This suggests their origin. Handy has pointed out that tribes may have been divided into lineages because the prefix ati that commonly identifies a lineage appears in Tahitian place names although no direct evidence links them with any present family. 35 As everywhere in the Stratified societies, commoners (manahune) were a landless and alien group. About their social structure we know nothing at all. The raatira were an independent landed gentry who offered the usual tribute and services to the ruling arii and were secure in their tenure. Theoretically, the raatira were the younger branches of arii lines but this has not been genealogically demonstrated. As for the arii, they held on to their prestige, but their power and authority depended upon their wealth and their ability to control their own districts and those they had conquered.

Thus, in Tahiti, too, the lineage principle persevered among arii families and perhaps among the raatira; but by all indications its functions had become restricted to the transmission of rank and prestige. Even in this restricted setting, the most significant lineage function of allocation of lands was coming to be decided on political grounds. Among the raatira, for example, the first born (matahiapo) was entitled only to those lands on which the marae had been placed; other lands (taken by conquest or given as a reward) as well as other properties were bequeathed orally according to the will of the legator. There were no villages, since the population was dispersed among the plantations of the gentry.

HAWAII: 36 Features that appear but vaguely in the Tahitian accounts, such as the political character of districts and the absence of tribes and lineages show up strongly and unmistakably in Hawaii. At the time of European contact there no longer seem to have been tribes or lineages and the population was grouped into political districts whose leadership and land tenure patterns shifted with each change in administration. In Hawaii, more - 386 than elsewhere in Polynesia, land holding had become a reward for services rendered to an ascendant chief, and territorial divisions were arbitrarily formed as land grants. The chief alii granted large tracts to subordinate chiefs who, in turn, allocated subdivisions to their followers. Commoners were farm labourers who, if maltreated, moved to another landholder and added to his strength.

Royalty based upon seniority followed the lineage principle through descent lines traced equally through males or females. Genealogical relationship to royalty also regulated rank at court and the allocation of administrative offices. It was on the really vital issue of land tenure that the lineage principle broke.

Some interesting suggestions as to earlier social conditions in Hawaii are brought out in Handy and Pukui's study of kinship and social organization in the rugged and relatively isolated district of Ka'u on Hawaii. There the tribal organization with a strong feeling for kinship unity within the district persisted until late. In Ka'u the term makaainana meant tribe, as in other parts of Polynesia; whereas elsewhere in Hawaii this had become the term for commoner. “The folk of Ka'u regarded themselves as one tribe (makaainana) bred from a single parental stock ('ohana), 37 The 'ohana was the organization of the common people. A kindred, it established bases for economic co-operation and for common social activities. The 'ohana headman (haku—“director”) was an elder male of the senior branch of the kindred. He presided over food exchanges and worship, he planned communal activities, was responsible for welcoming the alii and for getting together the levy for the annual makahiki. Originally, Handy believes, the 'ohana was localized, but in the course of time and through intermarriages it spread throughout the various political subdivisions and became widely dispersed. In its recognition of seniority it bore some resemblance to the traditional lineage. However, its headman had no title of rank, and since there were no other titles, we must assume that seniority was definitely not a principle of ranking. Residence in the 'ohana was preferentially matrilocal and descent was bilateral. The 'ohana is of particular interest because it seems to suggest the way in which Polynesian descent groups would normally have developed in the absence of the organizing principle of graded rank. Thus, its appearance among the commoners of Hawaii serves as some corroboration of the central thesis of this paper.


We have encountered in Polynesia lineages, resembling patrilineages linked with a bilateral system and, therefore, rarely exogamous and rarely strictly unilineal. They were, nevertheless, true lineages in that they traced lineal descent through seniority (preferably male) and conducted the principal social, economic, religious and political regulatory functions of an organized kinship body. The lineages grew by branching, each new branch maintaining its link with the parent trunk through its senior line. Because of the principle of seniority the lineages were hierarchically ordered, as were their individual members. In addition, lineages tended to be unequal physically, i.e. with respect to natural resources and to numbers. The relationship between these two - 387 dimensions—of rank, and of size and resources—was a variable one; but in the end the physical dimensions predominated. Another form of lineage growth was by splitting. The stem broke from the trunk and began a new growth elsewhere. Status was a common motive for splitting. Younger branches avoided the stigma of falling into commoner status by starting a new lineage. Both forms of growth, branching and splitting, occurred concurrently, influenced by a combination of direct economic factors together with status-political factors. Strong leadership restrained splitting by overt political effort and by intelligent economic organization (irrigation and fish ponds, for example) and regulated branching.

These growth processes merit the closest attention for their intimate bearing upon political evolution in Polynesia. They incorporate virtually every dynamic element in Polynesian life, but above all, the rivalries inherent in a hierarchical social order that always granted leeway for social mobility through achievement.

Rivalry threatened each lineage from within as well as from without. Internal rivalry either split or weakened a lineage if the conflict was not readily resolved. A common solution was to divide the rule between the sacred (ariki) authority and the secular. In that event, the latter usually became dominant, thereby bringing direct political powers to the fore. If on the other hand, lineages were weakened by inner conflicts they were subject to attack and possible absorption, another event that hastened political evolution. Inter-lineage, or more often, inter-tribal hostilities were common enough, coinciding usually with a state of affairs in which internal disturbances were prominant. Warfare had an economic motive, but this motive was itself a product of chiefly ambitions and the status aspirations of the gentry. The consequences of warfare upon Polynesian social structure were overwhelming, not merely through the implication of conquest, but through the internal rearrangements attendant upon the prominence given to warriors and strong leaders. Warfare affected the social structures of conqueror as well as of conquered.

Lineage rivalry was an active potentiality in every Polynesian society, occurring even in those societies (Traditional) where lineages had segmented into status-equal parts. Invariably, the traditions speak of one lineage having become dominant by some forceful means. Normally, rivalry within lineages and between co-resident lineages was adequately contained by a village organization that offered representation and bound its members into a variety of co-operative enterprises. Thus, whatever weakened the village organization unleashed again the potential rivalries within a locality. Another condition that might theoretically have constrained lineage and tribal rivalry was the wide extension of kinship made possible by bilaterality. Kin ties actually criss-crossed the lineages and stirred conflicts of loyalty. In the Open societies Mangaia is an example of such conflicts in which the political ties of residence were expected to govern. In this stress upon the residence factor we see another example of political growth of the lineage.

- 388

Because the linear principle was that of primogeniture rather than that of sex descent, I have used the label “Status lineage.” The resemblance of the status lineage to a patrilineage is in a sense fortuitous. It arises from the idea that leadership is a male prerogative. Thus, as long as leadership moved by heredity in orderly descent lines, the status lineage followed the patrilineage principles. As the social bases of leadership changed, the character of the status lineage changed. It lost its patrilineal features, its kinship unity and eventually its economic, religious and social functions. In the end it fragmented; one section continued to define the royal lines, while the section in which the commoners had been included disappeared altogether and was replaced by a kindred. The restricted functions of royal lineages do not in fact constitute a lineage structure either, so that, in a real sense, one may say that the Polynesian lineage eventually evolved itself out of existence. In the place of the lineage structure a political system based upon the organization of power within a territory irrespective of blood ties arose. This event, accompanied by economic stratification, marked a decidedly new phase in Polynesian social and cultural evolution.

Kinship does not disappear because state-like political systems are formed. Kinship relationships take on new forms under these conditions. The details of what these new forms were has not been recorded. But, even if vaguely, we can see that in the Stratified phase two systems of social relations co-existed in each society—one a vestigial lineage type for the nobility, and the other a kindred for the commoners.

Our survey of the social structures with particular emphasis upon the descent groups of the three postulated types of Polynesian society has borne out the thesis of status lineages and their pattern of evolution surprisingly well. Even the apparent atypicality for Traditional societies of Ontong Java, Tokelau and Manu‘a seems to be related to the fact that these societies were also atypical in their systems of status. The first two were still stressing age seniority while Manu‘a had abandoned primogeniture.

We need to say something more about the role of social status in Polynesian social evolution. A status system has a double character. One side of it is social and functional. It organizes the patterns of leadership, give them the sanctions of tradition and symbols of authoritativeness. It acts as a spur to effort and to responsible action. The other side of status is personal and potentially anti-social inspiring motives of personal enhancement, and lust for power. Ideally, both sides co-exist in a balance of counterpoised tensions. When this balance shifts to give more force to the personal side the status system becomes strongly dynamic. This is what happened in Polynesia. Since it was the status system that gave the social structure its focus of organization, it was inevitably that a change in the status system would markedly alter the social structure. Polynesian social structure was particularly vulnerable to change because of the additional fact of kinship bilaterality. The essence of bilaterality is flexibility in rules of exogamy, in residence, inheritance, succession, degree of kin group unity and dispersal, and in the size of kin groups. Thus, when status - 389 rivalry, the personal side of the status system, upset the lineage structure and, along with it, related aspects of the social structure, it had bilaterality as an ally. When the lineage system had given way, bilaterality then served as the basis for a reconstitution of kinship relationships.

From what has been said it would follow that the common view in social anthropology that changes in marital residence are the necessary precursors of change in social structure insofar as linearity is concerned is, to say the least, inexact. The factors in charge are far more complex. If we have interpreted the Polynesian data correctly, it would be the status and the related political conditions that were most significant. Since both are part of social structure, the conclusion to be drawn is that some social structures contain within themselves inherent potentialities for change. Thus, one of the implications of the Polynesian study is that social structures linked with a formally organized hierarchical system are fundamentally unstable. If study in other culture areas should prove this to be the case, we should have at hand some very useful conceptions for a general theory of social evolution.

  • ADAMS, Henry, 1947. Tahiti, Memoirs of Arii Taimai. New York. Scholars Facsimiles and Reprints.
  • BEAGLEHOLE, Ernest and Pearl, 1938. Ethnology of Pukapuka. Bernice P. Bishop Museum Bulletin 150.
  • BECKWITH, Martha, 1932. Kepelino's Traditions of Hawaii. Bernice P. Bishop Museum Bulletin 95.
  • BEST, Elsdon, 1924. The Maori. Wellington. Polynesian Society Memoir 5, 2 vols.
  • BUCK, Sir Peter H., 1932a. Ethnology of Tongareva. Bernice P. Bishop Museum Bulletin 92.
  • — — 1932b. Ethnology of Manihiki and Rakahanga. Bernice P. Bishop Museum Bulletin 99.
  • — — 1934. Mangaian Society. Bernice P. Bishop Museum Bulletin 122.
  • — — 1938. Ethnology of Mangareva. Bernice P. Bishop Museum Bulletin 157.
  • — — 1949. The Coming of the Maori. Wellington, Whitcombe & Tombs.
  • BURROWS, E. G., 1936. The Ethnology of Futuna. Bernice P. Bishop Museum Bulletin 138.
  • — — 1937. The Ethnology of Uvea. Bernice P. Bishop Museum Bulletin 145.
  • — — 1939. “Breed and Border in Polynesia.” American Anthropologist, 41:1-21.
  • COULTER, J. Wesley, 1931. Population and Utilization of Land and Sea in Hawaii, 1853. Bernice P. Bishop Museum Bulletin 88.
  • ELLIS, William, 1853. Polynesian Researches. 4 vols., London, H. G. Bohn.
- 390
  • FORNANDER, Abraham, 1916. Fornander Collection of Hawaiian Antiquities and Folklore. Memoirs of Bernice P. Bishop Museum 4.
  • — — 1919. Fornander Collection of Hawaiian Antiquities and Folklore. Memoirs of Bernice P. Bishop Museum 6.
  • GIFFORD, E. W., 1929. Tongan Society. Bernice P. Bishop Museum Bulletin 61.
  • GOLDMAN, I., 1955. “Status Rivalry and Cultural Evolution in Polynesia.” American Anthropologist, 57:680-697.
  • HANDY, E. S. C., 1923. The Native Cultures in the Marquesas. Bernice P. Bishop Museum Bulletin 9.
  • — — 1930. History and Culture in the Society Islands. Bernice P. Bishop Museum Bulletin 79.
  • HANDY, E. S. C. and others, 1933. Ancient Hawaiian Civilization. Honolulu, Kamehameha Schools.
  • HANDY, E. S. C. and PUKUI, Mary, 1950. “The Hawaiian Family System.” Journal of the Polynesian Society, 59:170-190; 232-240. 1951, 60:187-222.
  • HENRY, Teuira, 1928. Ancient Tahiti. Bernice P. Bishop Museum Bulletin 48.
  • HOGBIN, H. Ian, 1931. “The Social Organization of Ontong Java.” Oceania, I:399-425.
  • — — 1934. Law and Order in Polynesia. New York, Harcourt Brace & Co.
  • LINTON, Ralph, 1939. “The Marquesas.” In the Individual and His Society. ed. A. Kardiner. New York, Columbia University Press.
  • LOEB, Edwin M., 1926. History and Traditions of Niue. Bernice P. Bishop Museum Bulletin 32.
  • MACGREGOR, Gordon, 1937. Ethnology of Tokelau. Bernice P. Bishop Museum Bulletin 146.
  • MALO, David, 1951. Hawaiian Antiquities. Honolulu. Bernice P. Bishop Museum Special Publication No. 2. 2nd edition.
  • MARTIN, John, 1817. An Account of the Natives of the Tonga Islands of Wm. Mariner. London, J. Murray.
  • MEAD, Margaret, 1930. Social Organization of Manu'a. Bernice P. Bishop Museum Bulletin 76.
  • METRAUX, Alfred, 1940. Ethnology of Easter Island. Bernice P. Bishop Museum Bulletin 160.
  • MURDOCK, George P., 1949. Social Structure. New York, Macmillan Company.
  • SHARP, Andrew, 1956. Ancient Voyagers in the Pacific. Wellington, Polynesian Society Memoir 32.
  • SMITH, Percy S., 1902-1903. “Niue Island and its People.” Journal of the Polynesian Society, 11:80-106, 163-178, 195-218; 12:1-31, 85-119.
  • WILLIAMSON, Robert W., 1924. The Social and Political Systems of Central Polynesia, 3 vols. Cambridge, Cambridge University Press.
  • WILSON, Capt. James, 1797. A Missionary Voyage to the Southern Pacific Ocean. London, London Missionary Society.
1   Revised and expanded version of a paper read at a symposium on “Studies in Cultural Evolution: Oceania,” before Section H, American Association for the Advancement of Science, New York, December 30, 1956.
2   For Polynesia see Burrows 1939.
3   Murdock 1949.
4   Since this was written I have read Sharp 1956, and see his book as corroboration of the thesis that Polynesian culture is to be explained primarily from internal developments rather than from diffusion.
5   Goldman 1955. I have since made minor changes in the postulated evolutionary sequences.
6   Firth (1936) and others (Gifford 1929, Sahlins, in the accompanying paper) have used the term “ramage” to describe the characteristic Polynesian descent group. The merit of the term is that it brings out the branching character of the descent group. My objection to “ramage” is that it does not convey the essential point of status. Moreover, it will be more useful for comparative purposes to differentiate kinds of lineages rather than introduce new terms. In a more recent work, in fact, Firth (1956) has dropped the term “ramage.”
7   Hogbin 1931.
8   MacGregor 1937.
9   Firth 1956:53-54.
10   Beaglehole 1938.
11   Buck 1932b.
12   ibid. p. 55.
13   ibid. p. 29.
14   Based upon Best 1924, Buck 1949, Firth 1929.
15   Buck 1932a.
16   ibid. p. 45.
17   Burrows 1937.
18   Burrows 1936.
19   Mead 1930.
20   ibid. p. 18.
21   Williamson 1924.
22   Loeb 1926, Smith 1902, 1903.
23   ibid. p. 66.
24   Metraux 1940.
25   ibid. p. 120.
26   Buck 1934.
27   Handy 1923; Linton 1939.
28   Gifford 1929.
29   ibid. p. 30.
30   ibid. p. 127.
31   Cited by Gifford p. 171.
32   Martin 1817:I 105.
33   Buck 1938.
34   The principal works consulted: Adams 1947, Ellis 1853, Handy 1930, Henry 1928, Williamson 1924, Wilson 1797.
35   Handy p. 11.
36   The principal works consulted: Beckwith 1932, Coulter 1931, Ellis 1853, Fornander 1878, 1880, 1885, Handy 1930, Handy and others 1933, Handy and Pukui 1950, Malo 1951.
37   Handy and Pukui p. 172.