publisher colophon



Michael D. Lieber

It is evident from the preceding chapters that the circumstances surrounding movement of a community or part of it from one place to another vary. There is a difference between the kind of movement that has resulted in enclaves of Rotumans in Suva and the kind of movement that has resulted in Bikinians living on Kili. We ought to have a terminology that reflects these differences.

I suggest the following terms for referring to population move­ments and their effects as they are reported in this volume. “Reset­tlement” refers to a process by which a number of culturally homogeneous people from one locale come to live together in a dif­ferent locale. To the extent that people form an identifiable com­munity—identifiable to themselves as well as the observer—we can describe it as a resettled community. Except for the Nukuoro, all the groups described here are resettled communities.

We can distinguish two types of resettled communities on the basis of the processes by which communities come to be resettled: planned movement of a group of people, whose destination is de­termined by some outside agency, and movement undertaken by individuals without the intervention of an outside agency. The first process is here termed “relocation”; the second is “migra­tion”.1 According to these definitions, Bikinians, Banabans, South­west Islanders, Southern Gilbertese, Ambrymese on Epi, Kapinga­marangi homesteaders in Metalanimwh on Ponape, and Tikopia live in relocated communities. The Rotumans, the Kapingama­rangi 343 in Porakiet, and the Ambrymese on Efate are migrant com­munities.2 The Orokaiva, who migrated because of the eruption, resettled in their homeland after about a year, and we shall see that they resemble a relocated community more than a migrant one. The Nukuoro, who have migrated to Ponape, form no com­munity at all.

Once these distinctions have been made, the question naturally arises: Must the two processes of resettlement somehow yield cate­gorically different results? From the data reported in the previous chapters, the answer is: It depends. Murray Chapman made this fact clear to the participants in the symposium that resulted in this volume. The nature of this contingency constitutes Chapman’s crucial contribution to the symposium and to this volume. He pointed out that the patterns of mobility of a community only makes sense (to us or to the community) in terms of some larger system of which the community is a part. His discussion of com­munity mobility forced us to focus our attention on the levels of the larger system in which movements of people occur: the neigh­borhood, the village, the district, the island, the colonial territory. Each type of resettlement—relocation or migration—can make a difference in the history of a community insofar as it implies that a certain relationship exists between the community and the larger system.


Movements of people always occur as part of a relationship be­tween persons, groups of persons, or categories of persons. These relationships occur at various levels of the social system. A woman might go to a neighboring village to visit her mother, for example, or a chief from an outer-island village might go to a port town to negotiate with colonial officials. These two movements not only involve different relationships, but they also imply relationships that make sense at different levels of the social system. Social systems that form the contexts of movement can themselves have very different structures, and these differences are associated with differences in the nature and regularity of movements of persons within the systems. The foregoing chapters amply illustrate this point.

Robert McKnight has demonstrated that in precontact Palau, 344resettlement was common: Palauans had institutionalized meth­ods for dealing with immigrants in the villages where they ap­peared. In addition to the occasional canoeload of castaways, warfare sometimes dislocated entire villages or districts: “Villages and districts defeated in war fled to friendly communities and were sometimes absorbed by the villages that received them and the abandoned lands divided by the victorious villages” (Kane­shiro 1958:301). This is highly reminiscent of the New Guinea Highlands system described by Watson (1970), where resettlement is a recurring feature as a result of warfare and the competition between political factions within villages. When a victorious vil­lage drives a defeated village from its territory, the losers flee to villages where they have kin, affines, and friends. The process of resettlement in the new village is much as McKnight has described it for precontact Palau. The immigrants are integrated into the host village under the authority of a village ‘big man’, beginning a transformation of the immigrants’ identity that usually culminates in a merging of the emigrants’ identity with that of their hosts in two or three generations.

Despite differences of structure at the village and lower levels of the intervillage system, New Guinea Highlands societies and the Palauan social system are similar in some fundamental respects. In the New Guinea Highlands, relations between villages are com­petitive; intervillage relations are expressed through warfare and alliance, as well as through economic reciprocity; resettlement is a constant feature of the system. The position of a village in relation to other villages is not fixed but varies with time as competitive relations are played out. Thus, when we look at the larger set of relations in a multivillage district, the ordering of relations at time A is different than it is at a later time B. It is possible that at time B, one or more villages present at time A will no longer exist. The state of the larger (district) system at any given time depends total­ly on the states of the relationships among the component villages of the system at the time. Erik Schwimmer has shown in chapter 11 that these relationships fluctuate between hostility and al­liance.

Palauan villages, like their New Guinea counterparts, func­tioned as discrete units in political relations. Even so, each village was part of a confederation, and the villages of a confederation were ranked from highest to lowest. The highest-ranking village of 345a confederation controlled political activity for all the lower-ranking villages, especially decisions concerning the making of war and peace. According to McKnight (1960), the ranking vil­lage could and did engineer armed conflict between villages with­in its own confederation. Thus there is evidence of intraconfedera­tion competition among villages, although this was not ideally the norm. The four large confederations that made up the larger Pa­lauan social system continually competed with one another for rank. Villages within a confederation fought as a unit and were represented as units within the confederation. Palauan political relations appear to have been more highly ordered than those of the New Guinea Highlands; confederations were permanent and internally ranked on Palau whereas there were no stable confeder­ations in the Highlands, where villages were unranked and inter-village relations were constantly in a state of flux. Like the New Guinea Highlands, however, relationships between Palauan con­federations did vary with time. The state of the entire Palauan system at any given time is an outcome of the competitive relation­ships between confederations. This outcome is not predetermined by any higher level of organization. Furthermore, each village, whatever its position in a confederation, is a social and cultural unit in terms of political relations; its identity is maintained within its confederation and in its relationships with villages of other con­federations, especially in its role as a military unit. Immigrants to a village represent both a resource and a danger. They can add to the fighting strength of the village and to the esoteric knowledge of its political leaders; but until their commitment to the village is assured, they represent a danger. Significantly, in neither New Guinea nor Palau are there migrant communities: migrants either return to their old village or are absorbed by the host village.

Colonial systems show a much different structure than that of the New Guinea Highlands or of precontact Palau. The colonial system is characterized by a hierarchy in which local communi­ties, villages, and districts are under the control of a politically and militarily powerful group. This group is itself organized, and its organization bears little resemblance to that of the local com­munities it controls. The colonial system contains components of different orders with a suprasystem (a high-level organization that controls and coordinates the subsystems) whose organization is different from the subsystems it controls. It is a differentiated, 346hierarchically organized system in contrast to the New Guinea and Palau systems, whose largest local components are not con­strained by any sort of suprasystem. It is in the hierarchical social system that we find relocation, relocated communities, and migrant communities. Relocation is not restricted to colonial sys­tems, however; it becomes possible in any society that has the req­uisite organization to conceive of and implement it. Moreover, the character of migration in the Palau-New Guinea Highlands type of system is sharply different from that of migration in the colonial system.

Migration, like relocation, is closely geared to the needs of the higher orders of the colonial system such as administrative agen­cies, commercial organizations, and missions. This point is clearly illustrated by the Rotumans, Kapinga, and Ambrymese on Efate, all of whom have migrated to those islands where wages or cash for services are readily available. Members of all three groups were recruited at one time or another by commercial organiza­tions, missions, or administrative agencies. Relocation provides even clearer examples of how mobility within colonial systems contributes importantly to their maintenance. The Southwest Is­landers, for example, constituted a problem for the German ad­ministration on Palau: expensive shipping was necessary to service the atolls. The typhoon of 1905 offered a pretext for solving the problem by relocating the islanders. The Tikopia situation paral­lels that of the Southwest Islanders in that a natural disaster of­fered the administrative agencies and a commercial organization an opportunity to solve two problems at once: servicing the isolat­ed island and recruiting needed labor by reducing the population. The Ambrymese relocation took place under similar circum­stances, although the magnitude of the natural disaster that prompted it was perceived differently by the administration than by the Ambrymese. The worst disaster, a typhoon, occurred once the Ambrymese had been moved to Epi; the commercial interests were quick to take advantage of it by recruiting the Ambrymese as laborers. The experiences of the Southern Gilbertese and the Ka­pingamarangi homesteaders, like that of the Ambrymese, illus­trate what happens when relocation is a function of the colonial administration’s perception of an impending emergency—in these two cases, overpopulation. Also involved in the relocation plans for these two groups were government programs that called for ethnic integration of larger administrative districts. 347

The relocations of the Bikinians and the Banabans clearly were prompted wholly by the interests of the colonial hierarchy. Bikini was used for bomb testing and Ocean Island was exploited for its phosphate resources. At first appearance, the Orokaiva situation seems to differ from the others in that relocation was only tem­porary and people returned to their home territories little more than a year after the eruption. But the location and internal ar­rangements of the new villages and, subsequently, their economic organization reflect a new relationship with the colonial ad­ministration. As happened in the Tikopia, Ambrymese, and Ka­pinga cases, the administration was quick to take advantage of a natural disaster to implement its own plans for the affected peo­ple. The Orokaiva did not simply return to living in their old villages just as they did before the eruption. After their return to their home territory, the Orokaiva constituted a resettled com­munity comparable to the Bikinians and Ambrymese.

The differing structure of the two types of social macrosystems discussed here have another implication for movement and its consequences: the maintenance of ethnic boundaries (see Barth 1969). The Palau-New Guinea Highlands type of social system is an ethnic boundary-dissolving system vis-à-vis the emigrants. Because of the kinds of factors that provoke resettlement within the Palau-New Guinea system, return is at best problematical for the emigrants. If they are to remain in the new community, the maintenance of their ethnic boundary poses serious difficulties for themselves and their hosts. In communities where personal identi­ty and rank depend on land rights, attendant kinship connections, and demonstrable commitment to coresident villagers, the maintenance of ethnic identity by resettled people denies them permanent access to these. Although the immigrants may be under the aegis of a ‘big man’ at the outset, ‘big men’ do not live forever, nor are they politically powerful forever. The favor of a ‘big man’ or a chief does not really offer dependable access over the long run to village resources. Meanwhile, their hosts might want them to remain in the village for any number of reasons: to provide fight­ing strength, labor, special knowledge, or marriage partners. If the immigrants’ continued presence is deemed important to their hosts, then the maintenance of their ethnic boundary is likely to be regarded as diluting any permanent commitment and posing a threat to their hosts or at least some considerable faction of the host community (Watson 1970:116–117). Intermarriage with the 348hosts, and the consequent incorporation of children into the in­heritance system, often initiates the process of boundary dissolu­tion. The dissolution of ethnic boundaries of migrant groups seems to be complete within about three generations both in precontact Palau and the New Guinea Highlands.

Colonizers create ethnic boundary-maintaining systems by first imposing peace on the subject peoples. Although a colonial regime may act for humanitarian reasons, it is also true that local warfare interferes with profitable commercial operations. Labor recruit­ment and the maintenance of a labor force require relatively sta­ble island populations with a modicum of mobility. These popula­tions both supply labor and consume manufactured goods. Thus suppression of warfare within its territories is always one of the first programs of a colonial regime. In Oceania, accomplishment of this objective made intervillage and interisland travel a much less risky affair than it had been in precolonial days. The elimina­tion of warfare might not end interethnic hostility, but it does allow a resettling group to maintain itself as a community by pre­venting subjugation, or the threat of it, by surrounding groups. Given the need of a colonial government to maintain a flow of per­sonnel within its territories for such purposes as labor recruitment, mission work, education, and so forth, the protection of emi­grants, including resettling groups, is ensured. Ethnic boundaries are effectively frozen as a result. This is not to say that individuals or families of one group might not be absorbed by another through marriage or other means. Barth (1969) has shown that there can be a flow of people back and forth between one ethnic group and another that does not in any way obliterate the boundary between the groups.

The role of missions in the maintenance of ethnic boundaries must be emphasized. Although colonial regimes have the power to enforce peaceful relations among subject ethnic groups, missions typically form contexts in which peaceful and cooperative interac­tion is both possible and desirable. The context is legitimized by a consistently stated ideology that makes it possible to play down ethnicity, replacing its unifying role with that of membership in a congregation (and opposition to other religious sects). Even the missions can be forced to recognize ethnicity in their programs, however. Periodic services in the Rotuman language on Fiji and the growth of ethnic congregations among resettled communities 349on Ponape are examples of such recognition (Lieber 1968b: 135–137). Even in these instances, the contexts in which ethnicity is recognized emphasize its distinctiveness only as part of the larger whole that is the church. In other words, the church is an environ­ment in which interethnic interaction occurs in an atmosphere of safety; ethnic boundary maintenance poses no threat in that con­text. By contrast, commercial interests can use ethnicity to stimu­late competition, as in the gold mines in Vatukoula, where man­agement encourages competition between Rotumans and Fijians in order to increase production.

Whether the different processes of resettlement yield categori­cally different results, then, depends in the first instance on the structure of the larger intercommunity system in which the move­ment takes place. Although migration can lead to the resettlement of all or part of a community in a nonhierarchical system (such as that on Palau), there are no permanent migrant communities in such a system. Permanent migrant communities are to be found in the more differentiated, hierarchical systems such as states and colonial regimes, which typically act to ensure the maintenance of ethnic boundaries. Relocation is peculiar to this latter type of so­cial system.

In addition to the structure of the system in which movement takes place, the structure of the relationship between the moving group and the larger system also determines the consequences of the movement, whether it is a migration or a relocation. This point is amply illustrated in the foregoing chapters. The decision to relocate is always part of an asymmetrical relationship between some superordinate government agency and a local community. The decision can range from forced relocation, such as the reset­tlement of Ponapeans on Saipan after the Sokehs rebellion against the Germans (Bascom 1950), to a series of delicate negotiations that result in a joint decision, such as occurred in the Tikopia resettlement. It is obvious that the colonial administration has the upper hand in such negotiations, because at the very least it usual­ly must fund and provision the resettlement program. Moreover, the administrative agency, at the very least, controls the alter­native sites to which the emigrants can move. It thereby controls the ecological-demographic environment to which the community must adapt.

Despite the power differential in the relationship, forced migra­tion 350 is rare; some sort of negotiation almost always occurs in the relocation situation. This is extremely important, because the con­tent of the negotiations includes not only the decision to move and the details of implementing the decision, but also the nature of the relationship between the colonial administration and the com­munity (though it is rarely discussed explicitly). It is this latter aspect of the negotiation process that can prove significant for the future of the relocated community. The Bikinian relocation is an obvious example. The history of that community can be seen, from one point of view, as a series of negotiations between the Bikinians and the U.S. Trust Territory administration whereby Bikinians have sought, with some success, to persuade the administration to adopt their view of the relationship. Once this was accomplished, the focus of negotiations shifted to attempts to get the administra­tion to act in accordance with this point of view. The Banabans and the Orokaiva have similarly attempted to persuade their re­spective administrations to accept their view of the relationship, but the outcomes have been different. Unlike the Bikinians, Tiko­pia negotiated to prevent the colonial administration from becom­ing involved in their internal affairs and to keep Nukufero as isolated as possible from other communities. The Southern Gil­bertese situation seems to have been somewhere between these two extremes—that is, between asking the administration to take re­sponsibility for the support of the community and asking it to ig­nore the community.

The difference between the relocated Southern Gilbertese and the other relocated communities is the result of two factors. First, the relationship between the community and the administration was worked out in detail before the actual resettlement. The new community was to have exactly the same relationship to the ad­ministration as the communities from which the emigrants came. Second, the major problem of the emigrants was an internal one: the formation of a community by people who had not lived to­gether previously. The administration could do nothing to solve this problem.

The difference in the outcomes of the relationships between the Bikinians, Banabans, Orokaiva, and their respective colonial ad­ministrations raises a problem complementary to that of the struc­tures of larger social systems. Although the structure of the colo­nial system can make certain processes (such as negotiation) in­evitable, 351 the outcomes of the negotiations still depend heavily on who the colonialists happen to be. The notable success of the Biki­nians in gaining concessions from the Trust Territory administra­tion, as opposed to the frustrations of the Banabans who attemp­ted to do the same, might have depended less on their relative negotiating skills than on the fact that Bikinians were dealing with Americans while the Banabans were dealing with British (Barnett 1953:93). Until anthropologists begin to study seriously the dif­ferent sorts of colonialists, we shall not be in a position to test this idea.

Migrant communities differ from relocated communities in two important and related respects. First, the migration process does not require nor does it usually involve negotiation between the migrants and the colonial administration. In every case presented in this volume, the formation of a migrant community begins with a very small enclave of migrants. They provide a nucleus around which growth occurs as later migrants come to join their families, obtain wage work, and so forth. Community formation is neither a conscious goal nor a particular problem for the migrants or the colonial administration, at least at the outset. Second, migrant communities have a different relationship with the colonial ad­ministration and the colonial system as a whole than do relocated communities.

Relocated communities almost always constitute administra­tive units within the colonial system. The extent to which the com­munity is willing to allow the administration to involve itself in community affairs and the extent to which the administration is willing to be involved in the community’s affairs vary according to the actual negotiations between the community and the admin­istration. We have seen that the range is from dependency (Bikini) to noninvolvement (Tikopia). In either situation, however, the community has a special position within the administrative hier­archy; there is a directness of communication between the com­munity and administration that bypasses the usual hierarchy from local community to district legislative body to the various levels of colonial bureaucracy. The administration, for its part, has a stake in the outcome of the relocation because the expense and effort have to be justified by whatever agency planned and undertook the program. Moreover, the outcome (or some image of an out­come) of the program can be crucial to administrators whose 352careers are at stake. What the relocated community makes of this relationship may largely determine its subsequent history. For one thing, the relationship outlines the parameters of adaptation for the community. The Bikinians, for example, sought to establish a relationship of dependency on the Trust Territory administration from the outset of their settlement on Kili. The size, quality, and availability of taro land were far less crucial to their adaptation than were the form and location of Kili’s shoreline and harbor. What mattered was whether boats could land supplies from ad­ministration ships, not how much taro could be grown (see Barnett 1953:88–89).

Migrant communities have no special relationship of this kind with the colonial administration. By and large, they do not con­stitute administrative units within the colonial administration or the colonial system: their status as ethnic communities is relevant to the administration, if at all, only for census purposes. The ad­ministration can ignore their existence, as is true of the Ambry­mese on Efate, or it can actively resist attempts by the community to establish itself as an administrative entity, as has been the ex­perience of the Kapinga in Porakiet. Migrant communities’ invisi­bility within the administrative hierarchy does not mean they have no place in the colonial system, however. If they are not seen by the administration as communities, they are at least seen as eth­nic groups both by the administration and by their neighbors. The Rotumans and Kapinga provide examples of this: their status as “Polynesians” in non-Polynesian areas gives them, as individuals, certain privileges such as administrative jobs and career advance­ment. Porakiet, for example, has been the administration’s showplace on Ponape for tourists.

In the absence of any direct relationship with the colonial ad­ministration, the parameters of adaptation for the migrant com­munity are set by lower levels of the colonial system: specific eco­logical conditions and relations with neighboring communities, commercial organizations, and missions. Commercial interests and missions are, of course, important in the adaptation of relo­cated communities, as the Tikopia, Southern Gilbertese, Banaban, and Orokaiva examples attest. Nevertheless, the relationships be­tween these organizations and relocated communities appear to be different from the relationships between the same organizations and migrant communities. The differences can be attributed to the 353structures of the colonial systems in question. The British, Austra­lian, and Japanese colonial administrations have been far more in­volved with commercial organizations and missions than have the Americans in Micronesia. Relationships of the Orokaiva to the Australian administration, and of the Banabans, Tikopia, and Southern Gilbertese (since 1963) to the British administration, have necessarily involved them as communities, with commercial interests operating with the sanction of the administration. In each case, involvement of the community with local commercial interests has been an important part of what has been negotiated in the community-administration relationship.

Involvement of the relocated communities with missions has also depended on the relationship between administration and mission. Education as well as medical care in the Mount Lam­ington area has been delegated to the missions by the administra­tion. Schwimmer demonstrates that Orokaiva contact with the missions did not become intensive until after the eruption; this is the point at which negotiation with the administration becomes meaningful. In the British colonies only education has been con­trolled largely by missions. By contrast, the American administra­tion has kept education and medical care within the administra­tive sphere, with missions providing these services to only a small proportion of the population. In short, the extent to which relocat­ed communities have been involved with commercial interests and missions has been largely a function of the relationship between the community and the colonial administration. Thus the con­straints on the relationship between these communities and com­mercial and mission interests have been political as well as economic.

Migrant communities, like relocated communities, face the problem of recreating an infrastructure. In every case in this vol­ume, the infrastructure has evolved as the community has evolved. Beginning with a few individuals or families who are later joined by relatives and friends, the community infrastructure develops by gradual accretion as more facilities are needed for more people. Moreover, friends and relatives who join the enclave obtain such things as wage work through those who already hold jobs. Two of the migrant communities have developed more or less permanent ties with commercial operations; these ties parallel the relation­ship between Tikopia and Lever to some extent. The Ambrymese 354acquired land from a local planter in exchange for labor, and the Rotumans at Vatukoula are housed and fed by the mining firm for which they work. In both instances, the company has developed a relationship with the community, but the relationship is some­what different in each case. Ambrymese, when they are acquiring land for their community, negotiate as representatives of a com­munity. The planter is interested in the Ambrymese as a communi­ty only in that their settlement and stability ensure a dependable labor supply. Other than land and its products, the planter fur­nishes no other facilities of an infrastructure. Moreover, not all wage earners in the community work for him. By contrast, the Rotuman community at Vatukoula is to a great extent an artifact of company policies that provide Rotumans with en bloc housing and stimulate competition between Rotumans and Fijians for the purpose of increasing production. In neither case has the colonial administration played a role in developing and sanctioning the re­lationship. The constraints on the relationships are purely eco­nomic.

Except for the situation of the Rotuman community at Vatu­koula, relations between commercial interests (or missions) and the migrant communities seem to be institution-individual or institution-interest group, rather than institution-community. Al­though the missions can be forced to recognize ethnicity, they tend to deal with individuals and families rather than with the com­munity in their day-to-day educational, medical, and ritual activi­ties. Thus, for missions, ethnicity is largely a matter of ethnic categories.

Interethnic relations seem to be far more important in the adap­tation of migrant communities than in the adaptation of relocated communities. In every case in this volume (with the exception of Rotumans at Vatukoula), migrant communities have been forced to depend on reciprocity with neighboring groups in order to secure access to at least some of the resources crucial to the crea­tion of a community infrastructure.3 On the other hand, not one of the relocated communities described here has had to depend on its neighbors for its infrastructure, at least for as long as it main­tained its relationship, as a community, with the colonial ad­ministration.

There is, then, a significant difference between relocated and migrant communities. The different processes of resettlement do 355yield different processes of community formation, owing to dif­ferences in the relationship between the community and the colo­nial macrosystem of which it is part. These differences include the relations between the community and the colonial administration, relations between the community and other agencies such as busi­nesses and missions, and relations between the community and other ethnic groups. The differences between the positions of relo­cated and migrant communities within the colonial system deter­mine different adaptive strategies, and these are, in turn, broadly determined by the different levels of the system within which the community has to operate. Yet, given these differences, the crucial question posed by Homer Barnett remains: Can a knowledge of these differences, plus specific knowledge of the relationship be­tween the social macrosystem and microsystem, allow us to predict what type of stability or change to expect in the organiza­tion of the microsystem and the way people in the community con­ceptualize themselves and their surroundings?


A comparison of the Nukuoro and Kapingamarangi resettlements answers Barnett’s question firmly in the negative. Given the same relations with the same colonial administration and roughly the same amount of land in the same location, the two attempts at re­settlement have yielded entirely different results. Prediction of change and stability in social organization and culture is impossi­ble if we know only the relationship between the community and the social macrosystem of which it is part. It is also clear from this comparison, as Vern Carroll so succinctly points out, that stability and change in social organization cannot be separated from stabil­ity and change in culture. Both the Nukuoro and the Kapingama­rangi have replicated their cultures on Ponape; because of this, the Nukuoro community falls apart whereas the Kapinga community does not. The Rotuman data afford a similar comparison. Only one of the four enclaves of Rotumans in Fiji replicates anything similar to the community organization of the home island. The question raised by the Howards and by the Nukuoro and Kapinga data is a fundamental one for social science: What does it take to make a community?

Throughout this volume the concept of community has been 356used mainly as a sociological construct—namely, as a coresident group of persons with some sort of formal organization. The con­trast between the Nukuoro and Kapinga on Ponape suggests that using this definition might be foolhardy. A human community is, clearly, more than just a sociopolitical unit. It is also a complex unit within the culture of its members. Exactly how complex it can be, especially to its own members, is clear in Martin Silver­man’s account of the Banaban meeting.

Living together in some organized fashion is not merely a socio­logical fact; it is a cultural fact as well. Living together means something to people. The meaning of living together depends on people’s living together in a particular way, not just any way at all. The meaning of living together includes the definition of the relationships within which individual careers are played out, rela­tionships such as those of kinsmen, affines, friends, rivals, neigh­bors, and colleagues. It must also include definitions of the social settings in which the relationships take place and the premises that make the relationships and their settings either vital or negligible, comfortable or dangerous. Sociological arrangements of people—or, more properly, people’s models of their sociological arrange­ments—express these meanings. These models inevitably act as constraints on the formation of a resettling community.

The Howards confront the question of community formation for the Rotumans, outlining cultural, sociological, ecological, and demographic variables as the important parameters. The How­ards go on to ask whether the importance of each variable differs with the particular community or whether one variable structures the relevance of the others in every case. The data presented in the preceding chapters all point to the Howards’ second alternative as having comparative validity. Although the meaning of living together can be expected to vary from one community to another, the relationship between the meaning of living together (the cul­tural variable) and the sociological, ecological, and demographic variables will be roughly the same in every case. The magnitude of the constraints on the replication (or formation) of a community that are exercised by sociological, ecological, and demographic variables will vary as a function of the cultural variable. The premises that structure people’s perceptions of living together will make specific demographic or sociological or ecological problems more critical or less critical to the formation of a community. The 357contrast between the Nukuoro and Kapinga resettlements clearly illustrates this point.

Carroll has shown that for the Nukuoro, the anxiety and mis­trust with which they regard their social relationships can be ex­pressed in two alternative ways: by constant “scanning” of one another and testing of their relationships or by total dissociation (see Carroll 1970). The nucleated village settlement pattern can be seen as an expression of the need for constant scanning, and one’s prolonged absence from the village (without having left the island) quickly raises queries from one’s fellows. The Nukuoro, conse­quently, find it exceedingly difficult to form cooperative groups, even kin groups, for any purpose. Land, for example, is held on an individual basis. By contrast, the Kapinga regard social relation­ships as normally solidary; hostility and mistrust are temporary problems that will eventually be dealt with successfully by the parties concerned. They are group oriented, and most land on the atoll is held by kin groups. Cooperative labor is typical of Kapinga daily life, including fishing, house and canoe construction and repair, and bridge and pier construction. Although borrowing and lending have a decidely pejorative evaluation on Nukuoro, the Kapinga define these actions as essential to being human. The nucleated village settlement pattern on Kapingamarangi and in Porakiet expresses the Kapinga perception of social interaction as solidary and satisfying. This becomes quite clear when contrasted with the Kapinga homestead program in which more than seventy people were relocated to Metalanimwh on Ponape between 1955 and 1956. Homestead land was allocated in two areas—thirty-four of the sixty homesteads near the lagoon and the other sixteen 5 miles inland. The inland homesteads had many more productive food plants growing on them at the outset then did the homesteads near the lagoon. The inland homesteads were dispersed in such a way that people working on them had to live in isolated hamlets. All but two of the sixteen homesteads are continually vacant, and the inhabitants of the other two spend about half their time in Porakiet. Given the Kapinga cultural model of living together, residential dispersal was an insuperable barrier to communica­tion. On the other hand, although the residence pattern of the Nu­kuoro makes concentration impossible, as Carroll has shown, the pattern here means something different than in the Kapinga case: it is the Nukuoro replication of their model of living together. 358Given the opportunity to choose, Nukuoro elect to live apart from one another. From the point of view of the Nukuoro, there is nei­ther cultural nor social change; there is only change in the fre­quency of adoption of an arrangement that is already present in the cultural system. The residential dispersion and the consequent lack of strong ties between households and the lack of enough pro­ductive land to support a large population make community for­mation impossible. The antecedent condition that accounts for these demographic and sociological factors is clearly the Nukuoro premises defining the meaning of living together. By contrast, the Kapinga, with an allotment of usable land only half the size of that originally granted to the Nukuoro, have formed a community in Porakiet whose population almost equals that of the atoll. Given Kapinga premises about living together, the lack of produc­tive land was a problem requiring innovative adaptive strategies, but it was never a problem with respect to forming a community.

The Howards state that the Rotumans appear to be in an inter­mediate position between the Nukuoro and the Kapinga, because they neither coalesce nor fall apart as a community. As a sociolog­ical statement, this is clearly true. Yet a comparison of residence patterns and group stability of the Rotuman enclaves on Fiji and on Rotuma itself reveals that the enclaves on Fiji have replicated traditional Rotuman premises that define living together and the social arrangements that express these premises. In so doing, Ro­tumans on Fiji have formed (what are from their point of view) four communities. On Rotuma, there is much movement of fami­lies from one house and from one district to another. As personal relationships in one neighborhood become strained or explosive, members of one or more households simply move to another house where they have residence rights. With this sort of mobility, the population of a neighborhood is fluid (Howard 1970). Thus it is hardly inevitable that one will see any particular person at any given time, except as required by ceremonial obligations. Partici­pation in social relationships on Rotuma occurs in various con­texts, including those of the household, the neighborhood, the de­scent group, the district, the church, and the island. The multiplic­ity of options each person has in descent group membership, land rights, residence, and various voluntary associations means there is no fixed set of relationships that demands permanent commit­ment and participation.

What is significant about the Rotuman enclaves in Lautoka, 359Levuka, and Suva is that they do provide a variety of contexts for social interaction and participation much like those on Rotuma (household, ceremonial, recreational, occupational, religious) while maintaining the fluidity that characterizes relationships on Rotuma. Religious services in the Rotuman language, Rotuman clubs, weddings, funerals, visiting, and interhousehold reciprocity are all contexts of interaction that symbolize community. These have all been replicated, along with the expectation that the par­ticipation of any given person in any of them will not be uniform over time.

Although Vatukoula is the only coresident Rotuman enclave, it is the mining company rather than the Rotumans who created the possibility of coresidence. The location and allocation of housing are not decided by Rotumans but by company bureaucrats. Al­though coresident households, neighborhood relations, Rotuman-language church services, and some political structures character­istic of the home island are replicated at Vatukoula, the residential fluidity and multiplicity of options that characterize Rotuma and other Rotuman enclaves on Fiji are clearly lacking. Older relatives who are an economic liability, for instance, might be forced to leave. Rotumans at Vatukoula have a community that partially replicates the Rotuman model; Rotumans in Lautoka, Levuka, and Suva also have communities, and these also partially replicate the Rotuman model, but in a different way. In other words, the issue is not why Rotumans replicate a community at Vatukoula and not elsewhere, but why they replicate their community in a different way at Vatukoula than in Suva, Levuka, and Lautoka. Demo­graphic and sociological contraints on community formation take on a different character in this context. Residential dispersion, for example, poses a serious barrier to the formation of a coresident community, especially in Suva, which is already crowded. How­ever, the spread of the Rotuman population does not form any ser­ious barrier to communication among Rotumans in Suva. More­over, the residential dispersion of Rotumans found in Lautoka, Levuka, and Suva maintains the spatial and communicational fluidity that is traditional in Rotuman social relationships. The fact that a Rotuman in Suva can obtain almost anything that he or she needs without going outside of the ethnic group suggests that despite a lack of a stable, coresident Rotuman settlement, there is in fact a Rotuman community in Suva, at least from a Rotuman point of view. 360

The Howards’ argument for the relevance to the formation of ethnic consciousness of an opposing ethnic group takes on a new significance in the argument presented here. In Vatukoula, it might well be the presence of Fijians and the intense rivalry be­tween them and the Rotumans that allows the Rotuman rigidity of residence and association to be maintained without tearing the Rotuman community apart. When personal relations become strained, Rotumans cannot get away from one another without giving up their jobs. Competition between the two ethnic groups, in other words, compensates for the lack of fluidity in the Rotu­man community at Vatukoula. The Howards note that this intense competition seems not to be apparent elsewhere on Fiji; they point out, for example, the common phenomenon of reciprocity between Rotuman and Fijian households in the other enclaves.

The Southern Gilbertese provide the clearest example of what it takes to form a community. Unlike the other resettled communi­ties described in this volume, this community consists of people who had not lived together previously. The formal organization of the community had been worked out before relocation. Districts, political organization, spatial arrangements of houses and horti­cultural areas, and the like were all mapped out before the move. The administration provided building materials, seedlings, cut­tings, and basic subsistence needs during the period of initial set­tlement. In addition to mobilizing the large amount of labor needed to lay out the village, construct houses, and clear and plant land, the essential problem for the Southern Gilbertese was that of creating the social relationships that make a community. Various organizational modes were tested during the period of early settle­ment in order to establish these personal relationships. When the first wedding was performed in the community, the parents of the couple invited the entire community to the wedding feast (Knud­son 1965). In the Southern Gilberts, wedding guests are normally restricted to kinsmen of the couple. The father of the groom, who was also the chief magistrate of the community, used a kinship model of the community in deciding to issue the invitation. It is this kind of testing that finally leads to the solution of the problem of organization.

In attempting to organize the work of clearing and planting on Sydney Island, two alternative modes of organization were sug­gested: a single work group of community men working on each 361family’s land in turn, or each household working only its own land. This problem became a major issue, as each alternative im­plied a different conception of community relations and each had its own adherents. Both positions were based on traditional, though different, norms of Gilbertese culture. Communal organi­zation of work projects is traditional in the Gilberts, although this was rarely applied to the clearing and planting of privately held land. Although the model of organization was an established one, its proposed application was novel Moreover, the proposal ran counter to the axiom that a person’s land is his own and those without rights of ownership or usufruct have no right to make decisions concerning its disposition. Both sides in the dispute held fast to their positions, and the tension finally erupted into a perma­nent schism between the two groups of settlers. It was this schism that finally organized the community.

The ideological nature of the schism is demonstrated by the fact that although the settlers were all Protestants at the outset, one of the groups converted wholesale to Catholicism after the split. In other words, the schism was given religious legitimization (see Berger 1969). That the schism is a permanent and necessary fea­ture of the community as a whole is demonstrated by the fact that when the second relocation (to Ghizo Island) occurred, both groups moved (although not at the same time) and maintained their spatial separation in the new locale. The ideological dispute articulates two different but related conceptions of community based on different levels of norm of the same cultural system. In this sense, the Southern Gilbertese community affords another demonstration of Schwimmer’s point that a culture is not a mono­lithic entity. The same event or issue can be the subject of different ideologies. These ideologies, in turn, can become political weap­ons, as the two Gilbertese ideologies were.

The schism plays the same role in the Southern Gilbertese com­munity that opposing ethnic groups play in the Rotuman com­munity at Vatukoula and in the Southwest Islanders’ community on Palau. The presence of a competing or hostile ethnic group pro­motes solidarity within the Rotuman and Southwest Islanders’ communities in the face of divisive factors. In the Southern Gil­bertese resettlement, where the community solidarity that results from long histories of personal relationships is absent, opposing groups provide one another with a basis for permanent solidary 362relations—members of each group are united in their opposition to the other group. The principle of solidarity through opposition to some out-group is the same in all three cases.

The limiting case for our hypothesis regarding the relationship among ecological, demographic, sociological, and cultural vari­ables is that of the Bikinians on Rongerik. The atoll was simply too small and too unproductive to support the Bikinian population. The communal system of allocating labor and distributing food, traditionally used for organizing feasts, was employed at first only as a temporary expedient and then as the only available means of preventing famine. Even this tactic was inadequate for coping with the low productivity of the atoll’s resources.

The food shortages on Kili contrast with the situation on Rongerik. The coconut and taro resources might have been ade­quate for the community’s needs had they been carefully cultivat­ed and exploited. The Kili food shortage was to a great extent the result of political conflict within the community: any land alloca­tion proposal provoked conflict between those who held land on Bikini as headmen and those who wanted land (and its attendant headman status) of their own. The communual allocation of labor on Kili failed to provide the incentive for careful cultivation that was furnished by land division. At the same time, the community’s strategy of maintaining a dependency on the Trust Territory ad­ministration would have been ill-served by a comfortable adapta­tion to life on Kili. In other words, the food shortage was in great measure a result of a compromise made to avoid conflict within the community. It also became a political weapon in the struggle between the community and the colonial administration. Kili’s ecosystem had little to do with the food shortage.

If cultural premises that define living together give shape to the demographic, sociological, and (within limits) ecological prob­lems for the resettled community, then we must explain an ap­parent paradox. In each of the resettlements just described, the process of replicating the social relationships that are structured by those premises leads to change in the relationships. The paradox disappears, however, if the change is examined at the rel­evant level of the cultural system. We can distinguish three levels of the cultural system at which change occurs (hereafter referred to as levels of change) in the communities described in this volume. First, there is change in life-style—alterations in the way 363people allocate their time and resources. These alterations corre­spond to the learning of new skills and subsistence activities (and the novel application of old skills) and the adopting of new roles and relationships in order to adjust to a new ecological and social context. The first level of change can be seen in all the com­munities described here. Wage labor and the use of money for sub­sistence purposes, with their implications for the allocation of time and resources, can be seen in all but the first Southern Gil­bertese resettlement. The adoption of new skills, new types of en­tertainment, and new plants and planting techniques is common to all the atoll populations that moved to high islands. The learn­ing of new roles for interethnic relations is common to all the reset­tled communities. So, for example, a Kapinga man on Ponape who has a full-time job may be absent from his community for eight or more hours a day, five or six days a week, whereas his atoll con­gener would not. The time that the wage earner can spend with his family and friends has changed in a patterned way. Moreover, if he is required to wear a shirt and slacks to work, he has to invest more money in clothes than would a man who does not hold a job.

The second level of change involves the way people organize their social relationships with one another within the community. Organizational change can be seen as an outcome of the im­plementation of new strategies for deciding and acting. Although this sort of change is well exemplified below, what needs to be clarified at the outset of the discussion is the distinction of life-style and social organizational changes as different levels of change. The distinction implies that life-style, particularly the way people organize their time, energy, and other personal resources, is some­how embedded in the larger context of how people in a communi­ty organize their personal relationships with one another. We have seen, for example, that the Kapinga in Porakiet have undergone many changes in life-style without much change in the way people organize their social relations within the community. We have also seen that many of the changes in life-style constitute strategies for securing resources from outside the community; these re­sources flow back into the community as goods and services that people bring to their relationships with one another. The changes in life-style, therefore, preserve the social organization of the com­munity, as has been well illustrated in chapter 3. Moreover, it is the way that community social relations are organized that gives 364shape to its life-style changes; who learns which new skills and in what contexts is determined as much by the community’s rules for assigning roles as by the opportunities to learn the new skills. Thus, for example, it is no accident that Kapinga men hold full-time wage work in Porakiet nor that women in the community control the expenditure of money. The traditional domain of Ka­pinga men is at the peripheries of the village (outer islets, the lagoon, the reef), bringing back to the village the means of sub­sistence. Women are traditionally responsible for scheduling meals, planning the amount of food needed, and timing its collec­tion. It is the women in Porakiet who do the shopping, even though it may require a trip to town to do so. This constitutes a life-style change for women, but not a social organizational change. If, on the other hand, Kapinga men went to town to do the shopping (or routinely did it on the way home from work), this would constitute a social organizational change; specifically, the change would be one in the organization of male and female roles.

People’s models of their community’s social organization fol­low logically from premises that define persons, categories of per­sons, relationships between persons and categories, and the set­tings in which these relationships occur—premises defining the meaning of living together (as a kind of shorthand). These pre­mises form the cultural context in which people’s models of their community’s social organization are nested. Thus the premises that define the meaning of living together are of a higher logical order than models of social organization that follow from them; the relation between them corresponds to that of a class and its members (Bateson 1972:279–308). Change in the premises defin­ing the meaning of living together constitutes the third level of change discussed here. In the hypothetical case of the Kapinga man doing the grocery shopping, the social organizational change in male and female roles would lead us to infer that there has also been a change in the premises defining males and females. We would want to test this inference by looking for other changes in the distribution of responsibilities by sex. If there were such changes, and they all had a similar pattern—men taking over responsibilities formerly assigned to women—we would be justi­fied in describing the changes as having occurred both at the sec­ond level (of social organization) and at the third level (of premises defining the meaning of living together). As we shall see, change at 365the level of social organization often conserves higher-level pre­mises defining living together, but social organizational changes can also be a condition for change at the higher level of premises.

The second level of change is evident in the Nukuoro, Rotuman, Southern Gilbertese, Orokaiva, and Bikinian communities. We have already seen that living in dispersed, noninteracting house­holds is one way of expressing Nukuoro premises of living togeth­er. For the Rotumans on Fiji, the incorporation of Fijians into household reciprocity relations and the formation of Rotuman clubs implement Rotuman assumptions about the maintenance of their social relations with one another and their personal mobility within these relations. In the Southern Gilbertese situation, schism and its religious expression stem from two different levels of mean­ing inherent in traditional patterns of landholding and labor allocation.

The Orokaiva resettlement is another clear illustration of change at a lower level that follows from and conserves a higher level of premise. In this case, the Orokaiva view of the colonial ad­ministration as a trading partner in the traditional sense, that is, as an equal in an exchange relation, structures (or restructures) not only the Orokaiva political stance vis-à-vis the administration but also their external and internal economic relations. The change from subsistence farming to growing coffee for cash is, from the Orokaiva point of view, an economic strategy to strengthen their competitive position with the administration. Coffee and cash re­place the traditional articles of exchange without changing the way the Orokaiva perceive the structure of the trading-partner re­lationship. This strategy, however, induces other kinds of change. Joint ownership of land and the complex fabric of ownership and use rights traditional in Orokaiva subsistence farming make stable cash cropping difficult. These land tenure practices gradually gave way to fee-simple ownership. This complex set of changes in social organization is symbolized by the ‘new age’ ideology; yet Schwimmer shows that underlying this ideology is a very tradi­tional set of premises about interethnic relations. The ideology and its associated economic practices are expressions of these premises at the levels of social organization and life-style.

The Bikinian resettlement offers an even more dramatic exam­ple of organizational change. Bikinian social organization appears to have undergone a spectacular transformation in a 25-year peri­od. 366 Matrilineages have ceased to function as landholding social units and what appear to be cognatically organized groups are taking their place. But has there really been a fundamental change in the process of replication of the community’s structure? Clearly the answer is no. Bikinians continue to assume that prestige, in­fluence, and authority are outcomes of the responsibility a person can assume for the destinies of others. They continue to assume that ownership of land, through which one provides the resources that maintain others in a dependent relationship, is the where­withal to demonstrate one’s ability to handle responsibility. If the political organization of the Bikinians is viewed as a game—a competition for high rank according to prestige and authority—then responsibility for a group of kinsmen continues to be the criterion of eligibility to play the game. What has changed are the criteria for deciding who is a legitimate dependent and how authority is to be transferred from one generation to the next. These changes are not entirely new, however, because use rights over land were distributed cognatically on Bikini and patrifilial inheritance of land was already a traditional strategy of land transfer (when one could get away with it). As in the Nukuoro situation, the change in Bikinian social organization is one in which an alternative that is initially one of low incidence becomes more or less standard practice in a new social context.

The role that the social macrosystem plays in changes at this level, whether it is the colonial administration, other colonial agents, or other ethnic groups, is far from clear. Certainly, we can­not talk about “causes.” In no case is there any evidence that the changes described here are caused by policies, decisions, or ac­tions of groups with whom the resettled communities are in con­tact. Nor can we categorically state that change at this level is the result of adaptation to the new environment. For the Rotumans on Fiji, it does seem clear that their relations with Fijians and their formation of Rotuman organizations are a response to demo­graphic problems—finding housing and maintaining households and communication with other Rotumans in a crowded urban area. To this extent, these changes are adaptations; but it is not at all clear that there is no available alternative, such as arranging en bloc housing in a section of Suva.4 In the Orokaiva case, the changes described by Schwimmer are adaptations to a relation­ship with the colonial administration. Given the Orokaiva deci­sion 367 to establish a “trade partnership” with the administration, the adaptations are necessary ones. Yet Schwimmer demonstrates that the decision to establish the relationship was not a necessary one and that the Orokaiva have not adopted all the programs and policies required by the administration. The stability of the rela­tionship depends on each party being able to maintain its own perceptions of the relationship; this, in turn, depends on each side ignoring a good deal of evidence that would contradict its view of the relationship.

The failure of the Nukuoro on Ponape to articulate any organi­zation above the household level, the formation of the Southern Gilbertese community through schism, and the transformation of Bikinian corporate kin groups are only indirectly related to the en­vironments in which these communities resettled. For the South­ern Gilbertese and Bikinian communities, neither the specific locale and its ecosystem nor the relationship between the com­munity and the colonial system have any effect on the organiza­tional changes that followed resettlement. What is crucial in both situations is that the settlement is a new one and that certain organizational decisions were not made before resettlement. In effect, not having considered how labor was to be allocated on Sydney and how land was to be distributed on Kili before moving subsequently provided an opportunity for differences of opinion on these questions to be expressed. It was a new ball game in both communities. People had an opportunity to shape the community to their own advantage in a manner that would have been impossi­ble on their home islands. The colonial administration provided the opportunity by carrying out the resettlement; otherwise its role in the changes was negligible. The physical and social environ­ment played a slightly more important role in the change in Nukuoro organization. The possibility of wage work, acquisition of homestead land, and marriage with spouses of other ethnic groups on Ponape afforded the Nukuoro the opportunity to live in­dependently of one another.

Maintaining the premises that define the meaning of living to­gether can necessitate change at some lower level of the social sys­tem; however, the premises themselves can change. This is the case in the Tikopia, Ambrymese, Kapinga, Banaban, and South­ern Gilbertese communities. In each instance, premises about what living together means become a basis of more or less con­scious 368 controversy regarding issues that demand action by the en­tire community.

The controversies in the Tikopia, Ambrymese, and Kapinga communities involve the relationship between the resettled and the home communities. In the Tikopia resettlement, various ad­justments had to be made to a new ecosystem, to the relationship with Lever Company, and to the need for concerted action on community projects in the absence of the chiefs. These ad­justments required innovative responses by the Tikopia. Changes in work schedules and tasks, uses of cash, and the organization of decision making were necessary in Nukufero. These were changes in life-style and community organization. The fact that these adaptive changes were both necessary and successful made them even more problematic to the Tikopia, who had created a new community with the conscious intent of exactly replicating their social system on the Tikopia model. Given the manner in which they conceptualize their system—as an assemblage of customs, units of behavior, and organization—complete replication was impossible under the circumstances. Thus it is the very necessity of these adjustments that makes them problematic from the per­spective of Tikopia custom. The conflict between necessity and custom has permeated discussion and decision making in the com­munity, finally calling the concept of custom itself into question. As Eric Larson points out, some villagers began to ask whether it was possible to preserve the spirit of custom without exactly du­plicating its details. This question, which has not been resolved, is crucial for the community; it represents a change in the idea of custom for all the villagers. The change is one of differentiation of levels—between intent and practice, between ideals and strategies for implementing them.

The ramifications of this change in the concept of custom are wide for two reasons. First, the change opens the way to numerous strategies for living together. Second, differentiating the concept of custom into the levels of ideals and strategies for implementing them makes possible the conscious evaluation of new strategies in terms of how they affect other ideals inherent in other customs. Moreover, by making the level of ideals explicit, conscious evalua­tion and discussion at that level (and possibilities for further change at that level) become possible, as we have already seen in Martin Silverman’s description of the Banaban meeting. Under 369these conditions, the phrase “Tikopia and Nukufero are the same” acquires a new order of complexity through the differentiation of the levels at which they are the same (or different). Yet even a change in the idea of custom does not call into question the still higher-level premise that there is a uniquely Tikopia order of thinking and acting. The change does, however, make possible conscious reflection on what that order really is.

In the Ambrymese community on Efate, changes in both life­style and community organization underlie a change in high-order premises. An initial shortage of land and a consequent need for cash for subsistence, combined with the availability of wage work, had two important results. First, the Ambrymese established rela­tionships with a local planter, securing jobs and, later on, land by lease and sale. Second, the resettled Ambrymese were unable to replicate traditional land tenure patterns. Wage labor restruc­tured the villagers’ life-styles, especially the time spent in the vil­lage. More important, exchanging labor for grants of land re­quired a well-organized community effort. This effort is connected with the impossibility of replicating traditional land tenure prac­tices on Efate.

Traditional land tenure practices were inapplicable on Efate, initially because there was too little land and later because the land that was acquired belonged to the village as a whole rather than to individuals. Although the failure to replicate traditional land tenure practices is common to all the migrant communities discussed in this volume, this fact is crucial here because land disputes so often result in sorcery allegations on Ambrym. From the beginning of the resettlement, a major source of conflict was absent. In this respect, Maat Efate is comparable to the Southern Gilbertese and Bikinian situations; it too was a new ball game. Many fewer deaths were attributed to sorcery in Maat Efate. For the villagers this was salutary enough but it also changed mobility patterns in a decisive way. Sorcery scares on Ambrym resulted in a periodic exodus of people, especially males, out of the village for periods of up to several years. The bargaining power of the villag­ers on Efate that enabled them to secure land from the commercial planter in exchange for their labor depended on a stable popula­tion of males in the village. Not only are the need for a stable work force and the absence of sorcery mutually reinforcing, but the sta­ble village population and lack of potentially homicidal conflict 370allow the villagers to explore, in a way that would have been im­possible on Ambrym, potentials for community action that are already inherent in their social system.

The permanence of the Maat Efate community became a con­tinuing issue as people on Ambrym exerted constant pressure on their relatives on Efate to return or at least to clarify their status with regard to land rights on Ambrym. The studied ambiguity of the Efate villagers only sharpened this pressure. This issue ar­ticulated for the Efate villagers the differences between the two communities as they examined their commitments to each. Al­though the details of this series of changes might not have been ap­parent to Efate villagers, the net result certainly was. They have evolved two models of their own social system; although the pres­ence or absence of sorcery has been the stated difference between the models, it was not the only one. Sorcery has implications for how each community works. In other words, sorcery became a metaphor for the two different models of the Maat community as the implication became clear to the villagers that sorcery is an out­come of certain social contexts. Maat Ambrym villagers are not unaware of the two competing models of the community, as was demonstrated by their efforts to defend their own version.

The relationship between Maat Efate and its immediate en­vironment is important at the lower levels of change. The ecological-demographic problems of land shortage for the villag­ers and labor shortage for the commercial planter make possible the exchange of land for labor. It is the willingness of the planter to negotiate land transfers for a reliable labor force that makes it both possible and necessary for Maat Efate villagers to act as a corporate community in the negotiations. The disappearance of sorcery from the village cannot be regarded as inevitable just because one major source of conflict was eliminated. The necessi­ty for corporate community action constrains divisive activity such as sorcery. Once such a positive feedback relationship be­gins, it is self-reinforcing, at least insofar as the elimination of in­tracommunity sorcery is concerned (Maruyama 1963).

In Porakiet, changes in life-style and the organization of rela­tions with the home atoll of Kapingamarangi generated a series of issues that called into question the definitions of ‘Kapinga person’ and ‘community’. The gradual growth of a core of permanent res­idents committed to careers on Ponape, and distinguishing them­selves 371 from transient residents, resulted in a differentiation of atoll and Porakiet life-styles. The differentiation is reflected in organi­zational changes in the relationship between the two communi­ties. Porakiet had been a colony of the atoll under the authority of the atoll chief, but by the early 1960s it was recognized by res­idents of both communities as politically independent from the atoll. These changes left the relationship between the two com­munities and its future ambiguous in two ways: To whom did the land in Porakiet belong? And to what extent were members of one community financially responsible for major projects conducted in the other?

When housing improvement loans became available to Ponape residents through a local cooperative, some of the permanent resi­dents in Porakiet sought to secure loans. To do so required that they put up land as collateral. This would have necessitated a land division of the village, which was proposed at a village meeting. Relations with persons and agencies outside the village (which normally generate resources for personal relationships within the community) are characteristic of the Ponape life-style of the Ka­pinga. Kapinga interest in housing loans in no way diverges from this pattern, yet its implication for the relationship between the atoll and Porakiet—that village land should belong to a few families—contradicted the notion that Porakiet was a place for all Kapinga people. The relationship between the village and its con­text generated this issue, and its resolution involved redefining that relationship.

The issue of the financial and moral responsibility of Porakiet residents for major projects on the atoll and, later, the issue of the responsibility of Kapinga homesteaders for major projects in Porakiet are intercommunity issues. To resolve them, the concepts of ‘Kapinga person’ and ‘community’ were invoked, discussed, and redefined. As we have seen in chapter 3, the definition of com­munity that eventually resolved these issues was, in effect, a redefi­nition of the context of the community. The redefinition of ‘Ka­pinga person’ and ‘community’ involved a recognition of the larg­er geographical and social system within which people who define themselves as Kapinga live together. The redefinition of these con­cepts interpreted the facts of the geographical and political separation of the Kapinga communities while preserving premises about personhood and responsibility at a still higher level. 372

The Kapinga, Tikopia, and Ambrymese cases have fundamen­tal similarities. In all three communities, systematic changes in life-style and organization distinguish the home community from the resettled community in a manner recognizable to members of both. This distinction, combined with the commitment of people in the resettled community to its maintenance, renders the rela­tionship between the paired communities ambiguous. The ambi­guity comes to the fore unavoidably when people’s commitment to the home island is somehow called into question. The question of commitment is brought about by specific issues in each case. The resolution of the ambiguity involves change at the level of pre­mises that define a specific concept—community in the Kapinga and Ambrymese cases and custom in the Tikopia case. In each case, redefinition involves a symbolic differentiation at the level of premise that interprets the differentiation of the life-style, organization, and demography of the home island and resettled community. In the Kapinga and Tikopia situations, the change of the premise is one in which the crucial symbol is differentiated in­to higher and lower levels of abstraction. In the Ambrymese case, the differentiation is between models of Ambrymese community. In all three cases, the symbolic differentiation redefines the com­munity as part of a universe larger than that of the home island.

The Banaban resettlement was preceded by several decades of change on Ocean Island involving both life-style and community organization. These changes resulted in the relationship between the Banabans and the commercial phosphate firm mining Ocean Island. This relationship altered the infrastructure of Banaban daily life. Electricity, running water, machines, and conveniences of various sorts provided by the mining company were part of Ba­naban daily life on Ocean Island by World War II. Moreover, conversion to Christianity eroded a ritual system that had orga­nized the relations between kinship, the ecosystem, economics, and locale on Ocean Island. Although the resettled Banabans were able to replicate their land tenure patterns and many patterns of personal relations, interest in replicating a traditional system was qualified by conflict over the specific content of that system; peo­ple no longer knew what it was. The Banaban dilemma was based on the close connection between what they decided to do about themselves and their relations with colonial authority and com­mercial interests. 373

Resettlement on Rambi necessitated making decisions about land distribution and land use, house and road construction, set­tlement pattern, public utilities, and so forth, all of which implied that there be models of organization on the basis of which these decisions could be made and implemented. Unlike the other com­munities described in this volume, the Banabans were in the posi­tion of having to invent, test, and refine their organizational mod­els rather than being able to replicate comprehensive, existing models. This process, which Silverman calls “testing out,” is essentially a matter of survival for the community from the begin­ning of the resettlement. The Banabans develop organizational forms around which to structure activity; they test the forms by deciding, acting, interpreting the consequences of their actions, and modifying the original construction on the basis of their ex­perience. This process is illustrated in chapter 6.

To decide how royalty money from Ocean Island land was to be distributed, the relationships between land and people, between various categories of people, and between government and people were discussed at great length. Categories of people, especially, were the subjects of examination and definition (or redefinition); relations between the categories, such as workers and farmers, workers and old people, were scrutinized in order to identify how the social order forced the victimization of some people by others. The Banabans are trying to arrive at some clear picture of a hu­man order within which sensible action maintains justice in social relationships. But in the conscious examination of these categories and the relations among them, the categories and their relations have been subjected to change. The results of any decision based on this examination and redefinition are necessarily fed back into the next decision. Given that this process is both continual and necessary for the survival of the community, change in the pre­mises defining living together is inevitable.

Change in Southern Gilbertese premises that define living to­gether stemmed from the attempt to reconstitute the traditional communal meetinghouse. Traditionally, seating areas in the meet­inghouse were ranked according to membership in a landowning kin group (Lundsgaarde and Silverman 1972). To resolve the heated conflicts over which families would get which seating areas, the emigrants agreed to designate all seating areas as of equal rank. Although the agreement avoided the short-term prob­lem 374 of squabbling over which family would get which seat, it also reduced the channels through which prestige aspirations could be expressed. The traditional basis of rank in Gilbertese communities is a historical relationship between kinship and land. This rela­tionship could not exist on Sydney and Gardner Islands because the land had no history that was meaningful to settlers and be­cause the reconstitution of the meetinghouse divorced rank from land and kinship. In terms of rank, which was a traditional mode of organizing the relations between persons and between families, the land on Sydney and Gardner Islands had neither a history nor a future. Thus, unlike the Banaban and the Tikopia situations, the home island and the new island could not be metaphors for each other. The meaning of real property as a set of relations among persons had been irrevocably changed and the symbolic connec­tions of land to relations among families severed. This implies either that the assumption that families are ranked had changed or that this assumption would have to be realized by a different strategy. In fact, this implication is a major reason for the decision by the Southern Gilbertese community to request a second reloca­tion.

Although the organization of the Southern Gilbertese communi­ty offered few channels for the expression of prestige aspirations, the world outside the community had no such apparent limita­tions. The experience of community members working and living off the island demonstrated that relations within the larger colo­nial system offered a variety of options for accumulating personal resources that could be brought to relations within the communi­ty. The decision to request a second relocation was a community decision made after much discussion. The reason given for the request—that life on Sydney Island, although comfortable, pro­vided no opportunities for “advancement”—has two implications. First, the reasoning implies that although the relationship among land, kinship, and rank had changed, the assumption that persons and families are differentiated by rank had not. Expressions of rank shift from relations based on resources held in the communi­ty to relations based on resources acquired from the larger social context of the community. Second, the symbolic connection be­tween community and place had been severed, in that one no longer necessarily implied the other. Moreover, this connection had been severed in several other contexts, both as part of the relo­cation plan and as an adaptive exigency. 375

A particular place in the meetinghouse no longer had any con­nection to the relations between persons or between families; also, it was no longer connected to particular places on the island, that is, to specific parcels of land owned by specific families. The or­dering of relations between persons and families, therefore, no longer had any necessary connection with place on the island. Moreover, as an adaptive necessity, the places from which emi­grants come are largely irrelevant to the formation of relation­ships with others in the new Southern Gilbertese community. To the extent that these ethnic differences are allowed to become rele­vant to everyday interaction, the formation of a single, solidary community becomes that much more difficult. This is one reason for requiring the emigrants to renounce all rights to land and rank on their home islands. To the extent that place of origin and place on the new island have become irrelevant to the organization of social relationships, the problems of forming a new community have become largely those of establishing predictable personal re­lationships. Once established, therefore, the community is por­table. Unlike the atoll model, in which the community is identified in terms of a particular place (the island), the new community is not identified in these terms. The decision to relocate a second time clearly implies the portability of the community and its irrel­evance to a particular place in any terms other than instrumental ones.

For the Tikopia, Kapinga, Ambrymese, Banabans, and South­ern Gilbertese, it is clear that change in the premises defining liv­ing together is preceded by and in some way depends on changes of a lower order. These lower-order changes are of two kinds—life­style and organization of social relationships. In all five cases, both types of change are designed to cope with specific problems. All these changes are the result of trial and error. People become committed to certain patterns of change because of their success in meeting adaptive needs, be it adaptation to a new ecological and social environment, as is seen with the Kapinga, Ambrymese, Tikopia, and Banabans, or adaptation to strangers who comprise the community, as in the case of the Southern Gilbertese. At these levels of change, the relationship between the community and its larger social and ecological context is important in every instance.

It is not unusual that changes in life-style and social organiza­tion in communities might ultimately contradict people’s expecta­tions of one another, expectations based on experience and un­conscious 376 assumptions about their relationships on their home islands. Ambrymese have reason to expect that a certain number of children will die each year as a result of sorcery. Kapinga have reason to expect help from other Kapinga when their community undertakes large and expensive projects. Tikopia have reason to expect that persons of high rank will be preeminent in making community decisions. Yet sorcery is virtually absent from the Maat Efate community, whose organization has changed in order to deal with outside agencies. The Kapinga distinguish themselves as members of different, politically autonomous communities. The Tikopia realize that those in the best position to make decisions for the Nukufero community are those with the most experience in dealing with commercial and colonial agencies.

It is neither necessary nor inevitable that such contradictions will become manifest to people in the community. The premises that people hold about themselves and their relationships are, more often than not, highly abstract and unconscious. The expec­tations of behavior and attitude that follow from these premises are not always met on their home islands even in the best of cir­cumstances. Even when expectations are constantly contradicted by experience, people can ignore the contradictions or fail to see them for what they are, as is illustrated by the contradiction be­tween personal and status relations on Kapingamarangi Atoll de­scribed in chapter 3. Change in the higher-order premises that define the meaning of living together, in other words, does not automatically follow a certain accumulation of change at lower levels of life-style and organization. Change at the level of premise is not dynamic in any sense of that term, nor is it the result of “im­pact” or “forces” or any other such metaphors that anthro­pologists are wont to borrow from physics. Contradictions are a matter of human perception, of the ordering of information that people have about themselves and the contexts they act in.

When decisions that must be made by a community hinge on resolving the contradictions between experience and expectations following from highly abstract premises, the contradictions be­come apparent and unavoidable. The context of making decisions about specific issues is the crucial condition of change in higher-order premises in all five communities. In the Banaban, Kapinga, and Tikopia communities, people explicitly discussed and rede­fined premises about living together. In the controversy over 377specific issues, the premises were consciously articulated and ex­amined before being redefined. People moved from the issue to the premise and back to the issue in order to decide on a course of ac­tion. As Silverman points out, the decision on a specific course of action concretizes the change at the higher order. In the Ambry­mese and Southern Gilbertese communities, the implication for the higher-level premise, rather than the premise itself, is under discussion. We can infer the change at the higher level from the manner in which people discuss the implications and the decisions they make about them.

The particular issues that bring about discussion, decision, and action in each community serve two important purposes. First, as has been pointed out already, such issues direct attention to the ex­istence of a contradiction that makes explicit either a higher-level premise or its implications; the resolution of the contradiction in­volves changing the premise. Second, resolutions of these issues are moments in the history of the community during which past decisions and adaptive strategies, community organization, peo­ple’s expectations of one another, and the premises on which their expectations are based are teased apart, examined, and pulled back together in order to reach sensible decisions. The issues thus serve as focal points around which people in the community or­ganize their experience of change and articulate their own history in relation to current community needs.


Given the levels of the cultural system at which change occurs and the processes by which change occurs at each level, the problem of predicting change must be differentiated accordingly. Prediction at the lower levels, especially of life-style, seems to be possible, but not necessarily easy. For example, one could predict that given the Orokaiva decision to grow coffee as a cash crop, changes in eco­nomic organization, land tenure practices, and kin relationships must follow. Given the Bikinian decision to allocate land by household, certain changes in political organization and the struc­ture of corporate kin groups were likely. Given an initial shortage of food-producing plants in Porakiet and the Kapinga need for cash, certain changes in Kapinga economic activity, and thus in life-style, must follow. But the nature of the givens in each of these 378examples constitutes a major problem. In the case of the Kapinga, knowing the structure of the Japanese colonial system on Ponape, the skills of Kapinga as fishermen, and the distribution of plant resources on Ponape allows a fair prediction of Kapinga adaptive strategies and their consequences for how the Kapinga allocate their time and resources. One does not need to know much about Kapinga culture in order to make such a prediction. In the Oro­kaiva case, we are dealing with an entirely different sort of given. First, the decision to grow coffee is an outcome of a decision to es­tablish a relationship with the colonial government; this decision is an outcome of a “contingency in its purest form,” as Schwim­mer has demonstrated. Moreover, unless one sees the structure of the Orokaiva-administration relationship from the Orokaiva point of view, predictions of change based on the idea that the Orokaiva have “accepted” government programs must prove false. In the Bikinian case, allocating land by household was one of three possi­ble arrangements; the organizational implications of each de­pended on Bikinian assumptions about power, responsibility, and land and on political conflict and its resolution within the com­munity. Nothing in the relationship between the community and the administration or in the relationship between the community and the Kili ecosystem would lead us to expect that land distribu­tion would even be a problem to the Bikinians.

Prediction of change at the higher levels of a culture might be theoretically possible if we had adequate ethnographic accounts of social and cultural changes attendant on resettlement and a thorough ethnographic analysis for each community of the pre­mises that structure people’s perceptions of themselves and their environment. In practical terms such prediction is impossible. More often than not, discovery of the premises of the culture under consideration comes precisely through analysis of the very sorts of change we are trying to predict. Moreover, in the analysis of higher-order changes, we are dealing with the types of contingen­cies that Schwimmer’s analysis copes with, although they are not all so dramatic as a volcanic eruption. The specific issues that form the crucial context of change described here are, by and large, contingencies. One could hardly assert, for example, that the proposal for a land division of Porakiet and the request for an assessment of villagers for roofing materials for the atoll church would be inevitable issues in the Kapinga community on Ponape. 379Given the history of the Kapinga settlement on Ponape, one can understand why these issues were important in changing Kapinga concepts of ethnicity and community, but prediction in such a case is out of the question.

The severe limitations on our ability to predict change in reset­tlement situations can give little comfort to governmental and other agencies contemplating relocation schemes. In the best of circumstances, adjustments to a new locale are not easy, nor are all the variables affecting it readily apparent. The provision by the responsible agency of adequate living and subsistence facilities for the emigrants, difficult enough to procure, is only the first step. What is adequate initially may not be sufficient subsequently, as we have seen in the Southern Gilbertese case. Natural population increase or higher aspirations may render the relocated communi­ty’s facilities inadequate in a relatively short time. It can only be considered a lucky coincidence for the Southern Gilbertese that their decision to request a second relocation came at a time when the British administration was looking for ways to integrate its Gilbert and Ellice Island Colony with its Solomon Island Colony as a prelude to emancipation of both. Kapinga homesteaders relo­cated in Metalanimwh on Ponape were not so lucky; nor were the Bikinians relocated on Rongerik. One of the two tracts of land allocated to the Kapinga had very few productive trees growing on it. Even with the input of labor required to provide subsistence for the homesteaders, it would have taken two or three years for most of the homesteaders to produce an adequate food supply. The administration had agreed to provide food for only the first year of the program. By the end of the second year, all but seven of the homesteads had been abandoned.

Even when adequate subsistence and living facilities are as­sured—and what appears adequate to the administration may not be adequate for the emigrants, as we have seen in the Bikinian case—the administration has very little control over the relations be­tween the resettled group and other ethnic groups who may be their neighbors. It can, of course, try to provide an area for reset­tlement that is isolated from other groups. It can also ensure that the relocated group is not annihilated by its neighbors. Other than these strategies, the relocating agency can do little to ensure even cordial interethnic relations. The idea that interethnic integration can be effected by political programming, strategic resettlements, 380and propaganda has become popular among colonial regimes in Oceania. The program of dissolution of ethnic boundaries (what McKnight calls “ethnocide”) within a system whose very presence serves to maintain, if not to freeze, such boundaries is both con­tradictory and naive.

On the whole, the migrant communities’ adjustments to their physical and social environments have been more stable and less conflict-ridden (both internally and externally) than have those of the relocated communities. This difference owes to the difference between the resettlement processes and to the fact that migrant communities have fewer variables affecting their adjustments. Migrants settling in a new location tend to be few in number, and their immediate problems are food and housing. Relations with neighbors, learning new skills, and assimilating new roles are adaptive responses acquired on the way to ensuring adequate sub­sistence. The early migrant enclaves tend to be organized by households and relationships between households; that is, kinship relations are the major organizational tools of migrant enclaves, at least until the enclave grows to the point where some higher level of organization is necessary to maintain it. Such problems are dealt with as they arise—for example, the Ambrymese organi­zation of the community in their dealings with commercial firms and the Kapinga’s political reorganization of Porakiet to maintain order and initiate work projects for a growing population. The growth of the community at every stage is geared to the problems that arise in the new environment. The structure of the communi­ty is the result of a series of decisions (and their ramifications), in­formed by their culture, as the need for collective action arises. The migrant community, in other words, is the result of an evo­lutionary process. Throughout the period of establishment and growth, we have seen that the migrant communities manage quite well without the aid of, or often the notice of, the colonial admin­istration. Moreover, the demands of maintaining a relationship with the administration are not present to complicate the various issues with which the communities must deal.

The relocated community, by contrast, has to cope with the ne­cessities of housing, subsistence, community organization, and the relationship with the administration simultaneously and immedi­ately. Miscalculations regarding the environment (physical or so­cial) by the administration or the community will be felt quickly. 381The Bikinians’ famine on Rongerik and the Southwest Islanders’ position as an instant pariah group are good examples of over­sight. Rectifying such errors (if that is possible) can be costly both to the community and to the administration, as U.S. Trust Terri­tory officials have learned, and as the British government, now facing a huge lawsuit by the Banabans, may well learn. The extent to which the problems of community organization are connected with the community’s relationship to the administration will make the problems of both more complicated. The Banaban relocation and its aftermath furnish a good example of this, especially in con­trast to the Tikopia, whose community in the Russell Islands has developed without much interference from the administration.

It should be apparent at this point that with all the possibilities of oversight and miscalculation inherent in planning a relocation, a critical variable that need not be left to chance is the relation­ship between the community and the administration. This vari­able is at once abstract and intensely practical. Besides all the many details of the move and the establishment of the community that have to be worked out and agreed on, the details—and, more important, the form—of the relationship between the parties must be clear to both. Each side needs to know what the other believes the relationship to be, as well as what the other expects by way of concession, compliance, and action in the future. Usually, all these matters are under negotiation, but such negotiations are im­plicit and fraught with ambiguity rather than explicit and clear. Perhaps the best model of clarity in this volume is the Southern Gilbertese relationship with the British administration. With all their other problems, including creating a community from a col­lection of strangers, if the Southern Gilbertese community’s rela­tionship with the administration had been ambiguous, the resettle­ment probably would have been impossible. In contrast, the Biki­nians’ relationship with the Trust Territory at the outset was a model of ambiguity.

It is obvious from the data presented here that ambiguity in the relationship between the relocated community and the adminis­tration can be exploited by both sides, despite the vast difference of political and economic power between them. The exploitation of the Banabans by the British and the skill and success by which Bikinians gain concessions from the U.S. Trust Territory exempli­fy this point. Administrations might have an initial advantage in 382negotiating with their subject populations, but with a more sym­pathetic and effective international press, with a highly critical coalition of Third World nations making its presence felt in the United Nations, and with various other organizations, such as public interest law firms, willing to take up the cause of an exploit­ed community, the initial advantage can be quickly neutralized.

Working out the relationship between a community and a colo­nial administration is a delicate process subject to misunderstand­ing in the best of circumstances and intentions—especially when the negotiating parties have different cultures and each party’s view of the other, as well as of their relationship, is vague at best. Added to this problem is that of different modes of communication in which proper deference, for example, might be construed as agreement. It is clear that some skilled third party, an interpreter familiar with both cultures, is necessary in such negotiations. But the interpreter’s relationship to each party needs to be at least as clear as the relationship between the negotiating parties. Social scientists have played these roles, but anthropologists in particular have become very sensitive about the ethical implications of their roles in negotiations between colonizers and colonized.

Monitoring the progress of the establishment of a new com­munity is also extremely important. There are at least two ways in which this can be done. One way is for periodic surveys to be con­ducted by the responsible agency, but this approach has several drawbacks. There is a tendency for administrators to be overly op­timistic about the progress of the community, and this is under­standable. Administrators planning and carrying out a relocation have personal stakes in its success. Failure of the program (or the image of failure) can damage their careers. There is a tendency to gloss over problems in making reports to superiors. One corrective measure is to have representatives from another agency take part in the surveys. This has its own drawback, however, as it opens up the possibility of complicating the situation with interagency politics. A second monitoring strategy is to assign an adminis­trative representative to the community for as long as it takes the community to establish itself or even on a permanent basis.5 This was the strategy used in the Southern Gilbertese relocation (Knud­son 1965). The British colonial agent participated in the estab­lishment of the community, acting as coordinator and trouble­shooter with the colonial office. The community’s needs were communicated to the office headquarters on the spot rather than 383going unreported until they developed into crises, as happened with the Bikinians on Rongerik. This strategy has its own draw­backs, however, for it opens up the possibility that the administra­tor might become a pawn in community factional struggles, as seems to have happened on Rambi Island. But if the administra­tion’s positions and policies are clear to begin with, this sort of thing is far less likely to occur.


The resettlement of communities is an occurrence ancient in human history, although the study of such communities is recent. Comparative study of relocated communities was initiated by Homer Barnett, director of the Pacific Displaced Communities Project. The project was designed to study and compare twelve resettled communities in Oceania, six of which are reported in this volume. The project grew out of Barnett’s theories of the process of culture change, specifically his theory of innovation as the basis of all culture change (Barnett 1953).

A culture, like any other viable system, must have the capacity for change. The process of innovation within a society, which Bar­nett has described, analyzed, and illustrated in detail, generates the variability within a community’s culture that represents its ca­pacity for change. Individual persons innovate, whereas the framework of community organization, belief, and interpersonal relations constrain innovative activity. A change in the physical, social, or geographical context of a community is necessarily a change in the constraints on innovative activity. Therefore, ac­cording to Barnett (1961), resettled communities are natural lab­oratories for the study of culture change.

Barnett’s design for collecting field data (see the appendix to this volume) was ordered in terms of variables that influence change and stability. The variables are subdivided into those ex­ternal to the community and those internal. These variables, which are seen as a set of constraints that either induce or inhibit innovative activity, correspond to variables of the macrosystem or the microsystem. On the basis of the comparative issues discussed in this volume, the outline of cultural variables in the appendix will afford anyone studying resettled communities in the future an extremely valuable guide for the collection of essential data.

The communities described here vary widely in terms of the cir­cumstances 384 of resettlement, their new environments, the relation­ships they have with the colonial system, and the specific changes they have undergone. Some are migrant communities; others have been relocated. Some communities resettle because of crisis; others do not. Some communities consist of the entire population of the home island; others comprise only part of the home island popula­tion. Two of the relocated communities consist of members of more than one ethnic group. It is regarding problems of communi­ty formation, the maintenance of and the changes in cultural sys­tems, and the relations between the community and its context that comparison is most fruitful.

Resettled communities are phenomena of complex, hierarchi­cally ordered social systems such as state and colonial systems. In these social systems, retention of ethnic identity is possible in a context of mobility of communities (or parts of communities). We have distinguished two processes of resettlement: migration and relocation. Also, we have determined that there are significant dif­ferences between migrant and relocated communities. First, the position of each type of community within the colonial (or state) system is different. Relocated communities are administrative units within the colonial system; they have ongoing administrative and political ties to the administration, its agencies, and nongov­ernmental institutions connected with the administration. Relo­cated communities negotiate with these agencies as a community. Adaptations that relocated communities make to their new con­texts depend heavily on these negotiations. Migrant communities are not administrative units within the colonial system and are rarely in a position to negotiate as communities with colonial agencies. The adaptation of a migrant community depends largely on the relationships that community members, as individuals, are able to establish with the lower echelons of the colonial system—missions, commercial firms, and members of other ethnic groups.

The differences in position of migrant and relocated com­munities within the colonial system result in different processes of adaptation for each type of community. Migrant communities de­velop gradually through accretion of newcomers to a small core of emigrants. Ties of kinship and friendship are the major organiza­tional modes at the outset. Newcomers are socialized to the new environment by older emigrants, who procure jobs and other aid for them and introduce them to other individuals and institutions 385outside the community. Kinship and friendship give way to other modes of organization when the population reaches a size that makes this organization inadequate for coping with issues affect­ing the entire population, as is seen in the formation of a council in Porakiet and the election of a headman in the Rotuman communi­ty at Vatukoula. These higher-level organizational modes, usually adaptations of home island models, serve primarily to deal with intracommunity problems, although they may also represent the community to the outside, as in the Maat Efate group’s negotiat­ing for land.

Relocated communities have different adaptive problems: a substantial population is transferred to a new locale in a relatively short time and must create its infrastructure immediately. More­over, problems of recreating a social organization that is adequate for coping with the new locale arise simultaneously with those of housing and feeding the population. In every case, the adjustment of the relocated community requires a heavy initial dependence on administrative or other agencies charged with responsibility for the community (for example, Lever Company in the Tikopia case and the missions in the Orokaiva case). At the outset, negotiation with the administration is crucial to the survival of the communi­ty. Moreover, the negotiations set a pattern for the relationship be­tween the community and the administration following the initial resettlement. Also, in every case, the problems of initial settlement release a tremendous amount of innovative activity in both subsis­tence and organizational spheres. This innovative activity in­cludes the adaptation of old practices to new contexts, such as the Bikinians’ use of communal organization of labor and food distri­bution on Rongerik and Kili, and the introduction of new prac­tices, such as coffee growing by the Orokaiva and the ‘council’ in Nukufero. Of course, this trial and error process of “testing out” is characteristic of both migrant and relocated communities, but it is more visible in relocated communities because of the immedia­cy and complexity of adaptive problems and the relatively short time people have to develop solutions. It should be noted here that Barnett not only predicts that the rate of innovative activity will rise sharply in such situations, but he also accurately predicts the form the activity will take; this has been called “testing out” and ethnographically described in detail by Silverman (Barnett 1953: 80–89; Silverman 1971). Finally, relocated communities are far 386less dependent for their maintenance on relations with other eth­nic groups than are the migrant communities.

For both migrant and relocated communities, the problem of community formation is ultimately the same. It is a problem of re­creating a set of relationships among people in a new context. Of all the variables listed in the appendix to this volume and dis­cussed above, the one that is most crucial to the formation of a community is the culture with which it starts its new settlement. In a community’s culture, the information people have about themselves and their environment, along with the structuring of perceptions of the ecosystem and of people and organizations out­side the community, come together in a more or less coherent sys­tem of premises. The adaptive problems of a community are struc­tured as much by the community’s perceptions of itself and its relations as by the “objective” facts these perceptions interpret (and which are transformed into what the community considers facts). Geographical spread of population, available land, the presence of other ethnic groups, and the nature of colonial agents become problems or nonproblems in community formation ac­cording to what it means to live together in that community. These meanings are the assumptions people have about their rela­tionships with one another; they are the (largely) unconscious premises that define people, relationships between people, and community. Because these assumptions define living together, they must also define what constitutes the problems of living to­gether, as we have seen in the data presented here. Resettlement presents a host of novel problems to a community; much of a com­munity’s innovative activity replicates in novel ways the assump­tions about living together. New skills, artifacts, and statuses are transformed into socially meaningful activities as they become re­sources or means of access to resources to bring to relations within the community.

Finally, the systemic nature of change in both the cultures of the resettled communities and the colonial systems of which they are part is the same. The structure of the colonial system might re­quire that its subject peoples be mobile, especially when large-scale administrative projects and commercial operations require large labor forces. Resettlement is one way of securing labor. In general, as Chapman pointed out to participants in the relocation symposium, the maintenance of the colonial system might require 387more or less constant changes in its constituent populations. In every instance reported in this volume, resettlement has contribut­ed in some important way to the maintenance of the colonial sys­tem. We see exactly the same principle operating in the resettled community (or the microsystem): maintaining the higher orders of a system often involves changes at the lower orders. This was true of the Nukuoro, Bikinians, Rotumans, Southern Gilbertese, and Orokaiva. We can add to this list the Banaban, Kapinga, and Ti­kopia instances of change in higher-order premises. For the Ka­pinga, change in the concept of community conserves the concepts of personhood and responsibility (Lieber 1974:91–92). For the Banabans, concepts of ‘farmer’, ‘worker’, and other roles and rela­tionships between those roles are subjected to change in such a manner that each is still consistent with the higher-order concepts of land and freedom. When the Tikopia concept of custom begins to change, the premise that underlies it—that there is a distinctive order that is uniquely Tikopia—remains unquestioned. This type of change is an evolutionary process that is typical of viable sys­tems, whether they are cultural or biological.


I would like to acknowledge the generosity of Homer Barnett, Vern Carroll, David Schneider, and Martin Silverman, who provided valuable criticisms and suggestions on earlier drafts of this chapter.

1. This use of the term “relocation” in no way departs from its normal usage. Ambiguity is introduced when the definition of the term is expanded to in­clude migration.

2. The maps in this volume differentiate each kind of community with the label “relocation” for the relocated communities and “movement” for the mi­grant communities.

3. By “infrastructure” I mean those resources necessary to maintain the com­munity, such as locally produced foods, access to land, materials for con­struction, techniques of production, and so forth.

4. Alan Howard, in a personal note, reports that the Rotumans in Suva are now seriously discussing the possibility of arranging en bloc housing.

5. Ideally, the representative would be defined as a community advocate (and would not be punished by the administration for acting as such). The com­munity advocate should also be replaceable if the community believes that the advocate is not adequately representing its interests. 388

Next Chapter


Additional Information

MARC Record
Launched on MUSE
Open Access
Creative Commons
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.