publisher colophon



Alan Howard
Irwin Howard


In his introduction to Ethnic Groups and Boundaries, Fredrik Barth points out that even a drastic reduction of cultural differ­ence between ethnic groups in culture contact situations does not correlate in any simple way with reduced relevance of ethnic iden­tities (Barth 1969:32–33). The point is well taken, and there is considerable evidence to support his contention. One can go far­ther and assert that without regular and persistent contact ethnici­ty is socially irrelevant, for, as Barth cogently argues, the existence of ethnic groups depends less on the sharing of a common culture than on the maintenance of social boundaries. For social bound­aries to be actively maintained, they need to be continually vali­dated, and this requires regular interaction with members of out-groups.

Given these premises, the effects of European colonization on ethnicity in the insular Pacific are of particular interest. Prior to European contact, many Pacific islands experienced very little in­teraction with peoples of a substantially different cultural back­ground. An occasional canoeload of other islanders might make a landfall from time to time, but as a rule they were either driven off, killed, or absorbed into the local population through inter­breeding (see chapter 2). Particularly within the large culture areas of Polynesia and Micronesia, such immigrants were likely to 162exhibit only slight differences in language and customs from the host population.

The arrival of Europeans in the area, and the subsequent estab­lishment of colonial regimes, radically altered this situation. Not only did the Europeans inaugurate regular firsthand contact, but they also initiated and institutionalized boundary-maintaining mechanisms designed to distinguish ethnic groups. Furthermore, social privilege in many instances was allocated on the basis of ethnicity. In the early contact period this resulted in what was primarily a European-native dichotomy, but as time passed differ­ent indigenous groups were brought into regular contact with one another and immigrant laborers were brought from outside the region (from India, China, and elsewhere) into the crucible of plantation, mining, and urban communities. Additionally, inter­breeding between Europeans and indigenous populations gave rise to a half-caste or part-European group. The result has been the de­velopment of polyethnic societies and an opportunity for social scientists to study ethnic groups in the making.

One such group is the Rotumans, who currently form an ethnic enclave within Fiji (see map 8). The processes by which Rotuma developed into a hinterland community to Fiji’s urban centers have been documented elsewhere (Howard 1961). Our focus in this chapter is on the adaptation of Rotumans to the social milieus of four such urban areas. Here we are concerned with the degree to which they have formed viable ethnic communities, the organi­zational forms that have developed, and the extent to which ethnic consciousness has been created under varying conditions. Special emphasis is given to an analysis of the Rotuman community in Vatukoula, since it is there that the processes germane to our thesis have been most intense.


Although a substantial literature has developed over the years dealing with such topics as race relations, minority group studies, and ethnic studies, most research and theory have focused on rela­tions between groups whose boundaries were clearly defined or treated as unproblematic. Few studies have centered on the pro­cesses by which a people who share a common history are trans­formed into an ethnic group within a larger social system. Accul­turation studies in anthropology, while dealing with processes of change, have generally dealt with alterations in culture content, social transformations within a group, or the significance of change for acculturating individuals. Barth’s recent effort pro­vides some promising leads, but it falls short of projecting a theory of ethnic group development. In this section we attempt to build on Barth’s formulation; specifically, we postulate a set of processes that lead to the development and crystallization of ethnic bound­aries and, by implication, to the formation of ethnic groups. After presenting data from the Rotuman case, we conclude the chapter with a consideration of specific variables that hasten or retard the relevant processes. 163


The theoretical paradigm we are advocating begins with two distinct populations who are unaware of each other’s existence. Initial awareness may occur either through direct contact or in­directly through intermediaries, but in either case the first bits of information provide the basis for the development of ethnic cate­gories. If information flow is slow and irregular, these categories may remain vague for a time, but with regular contact informa­tion input is accelerated, generating preliminary stereotypes. Barth points out that the features taken into account in generating ethnic stereotypes are not necessarily based on “objective” differ­ence, but that “some cultural features are used by the actors as signals and emblems of differences, others are ignored, and in some relationships radical differences are played down and de­nied” (1969:14). He suggests two types of information of rele­vance to the establishment of ethnic dichotomies: one consists of the diacritical features that people look for and exhibit to show identity, such as dress, language, house form, and general life­style; the other involves evaluative criteria for judging behavior and the products of behavior. Barth’s basic message, however, is that “ethnic categories provide an organizational vessel that may be given varying amounts and forms of content in different socio-cultural systems” (1969:14). As such, their social existence is in­dependent of culture content but depends instead on the mainte­nance of social boundaries.

Ethnic boundaries may not emerge with clarity as soon as cate­gories develop, however. In the early stages of contact such boundaries may include extensive “shadow areas” in the form of ambiguous situations, role discrepancies, and obtuse or overlap­ping 165 diacritical features. During these stages social relations may involve the two populations more as ethnic aggregates than as ethnic groups. Such is particularly likely to be the case when there are no clearly demarcated geographical boundaries between the populations. The distinction between an ethnic aggregate and an ethnic group is that with the former, ethnic designation is subor­dinate to other identity principles in the organization of a popula­tion’s social life while with the latter it is superordinate. Barth’s comments concerning polyethnic social systems (1969:17) are what we have in mind in considering ethnicity as superordinate:

Common to all these systems is the principle that ethnic identity im­plies a series of constraints on the kinds of roles an individual is al­lowed to play, and the partners he may choose for different kinds of transactions. In other words, regarded as a status, ethnic identity is superordinate to most other statuses, and defines the permissible con­stellations of statuses, or social personalities, which an individual with that identity may assume.

The crucial question from the standpoint of the development of ethnic groups can thus be phrased: Under what conditions does ethnicity become the superordinate symbol of identification with­in a social system? Our position is that the fundamental conditions underlying the transformation of an ethnic aggregate into an eth­nic group are (1) the development of an ethnic community, that is, a localized interactive network consisting of individuals of the same ethnic designation who are emotionally committed to the symbols of their common heritage and formally organized for the purpose of pursuing common goals; and (2) the formation of eth­nic consciousness. Ethnic consciousness may be defined as a spe­cial case of ethnic awareness, that is, a recognition by an individ­ual that his ethnicity is a significant factor in ordering his social relations. When ethnicity assumes a position of primacy for the in­dividual in structuring his interactions, whether with others of his own ethnic category or outside it, his awareness may be said for our purposes to have become consciousness.

Ethnic consciousness may develop on an individual level in response to a number of circumstances: these include overt dis­crimination by others, a sense of superiority or inferiority, or status ambiguities that can be resolved by giving primacy to eth­nicity. Collectively, ethnic consciousness emerges as a result of 166repeated messages circulated throughout networks of kinsmen, friends, and neighbors to the effect that other identity criteria are less significant for structuring interpersonal relations than ethnic differences. The redundancy of these messages serves to structure both social interaction among ethnic cohorts and an ideology of “we-ness,” the sharing of a common social fate. The structural manifestations of these messages are the extension of close per­sonal bonds characteristic of kinship and friendship to all who are members of the same ethnic category and the restricting of one’s personal relationships to people within that category. That one member of the category is shamed, offended, or honored implies shame, anger, and honor for all vis-à-vis nonmembers. To the ex­tent that nonmembers of an ethnic category view members as in­terchangeable, the redundancy of the relevance of ethnicity is like­ly to be reinforced. For example, when the message that an indi­vidual lost his job or was abused because of his ethnicity circulates through a network of people of the same category, indignation and emotional solidarity are more likely to be engendered than if other identity variables are acknowledged to have played a part. The notion of sharing a common fate, if accepted by members of an ethnic category, takes on the character of an ideology by which people interpret their relationships within and without the net­work of ethnic cohorts. At this point, we can say that an ethnic group has emerged.1

The content of the unifying ideology may vary from group to group, but it always involves a common symbol or set of symbols. The key symbols may be racial features, religious practices, a monarchy, or common acceptance of some kind of charter myth, for example. Inasmuch as symbols and ideology are involved, we regard the formation of ethnic groups as very much a cultural pro­cess as well as a structural one, although we agree with Barth that once a group is formed its culture content may change drastically without the boundaries of the group being affected. For Rotu­mans, the dominant symbol of their shared ethnicity is the island of Rotuma itself: any person may claim to be a Rotuman if one or more of his known ancestors was born on the island and shared in the core social and cultural life that characterizes the society.2

As reported in an earlier article on conservatism among the Rotumans, the emergence of a consolidating ideology is rooted in the bicultural experience of nontraditional leaders, that is, in­dividuals 167 whose prestige accrues from success in Western occupa­tions and professions (Howard 1963a:73–74). These leaders are people of influence because they are educated Rotumans among uneducated Rotumans; their success in the outside world is ac­knowledged by other Rotumans as significant. Their influence and high status are located within the Rotuman community and de­pend on its existence. Moreover, they are leaders because of a demonstrated commitment to Rotuma, a commitment that has become highly conscious as a result of European education and experience in a Europeanized society. Having learned the mechanics of European culture, they have also learned to evaluate their own society in abstract terms as, for example, these terms are used in school to describe models of law and social organization (such as the government of Great Britain). Possessing an intellec­tual idiom for perceiving a society, educated Rotumans have often been struck by the inconsistencies between ideology and behavior in Western societies, as compared to a far greater consistency in Rotuman values and behavior, and between Western (particularly Christian) ideology and Rotuman behavior. Their education has therefore tended to foster an idealism about their own society while their experience has provided means to implement their ideas in community action.

Before describing the circumstances that have generated a Ro­tuman ethnic group in Fiji, we present in the following section some aspects of social life on Rotuma that are relevant to our basic discussion.3


From Cession in 1881 until Fiji was granted independence, Rotu­ma was administered by Great Britain as part of the Colony of Fiji. The decision leading to this arrangement was based on ad­ministrative convenience rather than on any existing ties between Rotuma and Fiji. In language, culture, and physical type Rotu­mans are clearly distinct from Fijians, resembling more closely than the latter the Polynesians to the east. Administration of Ro­tuma (which lies some 300 miles north of the Fiji group) was in the hands of a district officer who was responsible to the commis­sioner and, ultimately, to the governor of Fiji. In addition to his administrative duties, the district officer had the power of second-class 168 magistrate and presided over the Council of Rotuma, which was composed of the paramount chiefs of the island’s seven tradi­tional districts, an elected representative from each district, and the senior medical officer on the island.4

The traditional social organization is based on a system of bi­lateral kinship. A key concept is kainaga, which in its broadest sense refers to kinsmen and in a more restricted sense to the bilat­eral descendants of an ancestor holding rights over a particular parcel of land. Kainaga, in the restricted sense, are the major landholding units. In each traditional district, a limited number of kainaga hold the right to a chiefly name, some being eligible for paramount chieftainship within the district, others not. Districts are divided into hoaga, which comprise from three to seventeen households (with an average of ten). Hoaga are essentially work units, whose members have an obligation to assist one another in times of crisis and on ceremonial occasions. The most basic socio-economic unit in Rotuma is the kau nohoag ‘household’ (essential­ly persons sharing a common hearth and comprising a common consumption unit, since food is easily the most important con­sumable commodity). The modal ‘household’ consists of a nuclear family with one or more relatives of either spouse (39.3 percent) or a nuclear family by itself (29.6 percent). Persons who are not members of a nuclear family (widowed and divorced persons, or­phans, offspring of unwed mothers, unmarried adults) tend to have a high rate of residential mobility, moving from household to household. Almost every Rotuman man is an agriculturist, at least while living on Rotuma. Even those engaged in wage labor main­tain gardens to provide their families with food. A man is judged primarily in his role as provider, and to be a good provider means to bring home more than enough food for his family’s needs. With the exception of wage earners, this means being a competent and industrious farmer and harvesting available copra. The women on Rotuma have as their major tasks the care of children, keeping the household clean and presentable, and supplementing the family food supply by fishing on the reef. This sexual division of labor is not rigid, however, and cooperation between husbands and wives on domestic tasks is the rule rather than the exception.

The traditional kinship-based socioeconomic organization is crosscut by geographical divisions. Within Rotuma the sharpest in-group/out-group distinctions are essentially territorial. We 169found it rather striking that stereotypes held by persons of each district paralleled those between ethnic groups elsewhere. These stereotypes typically focus on alleged behavioral differences; for example, the people of one district are ridiculed as being like chickens—that is, marrying with kinsmen who are genealogically closer than deemed appropriate. For each district (and in some cases for each village) it is possible to elicit a stereotype that has currency and is essentially shared. Furthermore, it soon became clear to us that territorial proximity plays an extraordinary role in structuring social relations on the island. It is a general rule that people who interact frequently as neighbors, especially as ‘work unit’ mates, manifest a strong solidarity; correspondingly, clashes between neighbors often precipitate a residential move by one or the other. Even close kinship ties are rarely strong enough to over­come long-term geographical separation.

Two other organizational principles crosscut those of kinship: one is religion and the other is the formation of voluntary associa­tions. Voluntary organizations are formed mainly for the purpose of playing such European sports as soccer, rugby, and cricket. They are generally ephemeral organizations, lasting only as long as interest in a particular sport is salient. As a matter of conve­nience they tend to be strongly influenced by territorial pattern­ing. For all practical purposes the only religious groups represent­ed on the island since Cession have been Methodists and Catholics. The division between these groups largely coincides with a pre-European political division and, therefore, also has a strong ter­ritorial patterning. Rivalry between the two religious groups was intense enough to provoke a war just prior to Cession, and religion has remained a significant factor in ordering social relations on the island to the present day. Cross-religious marriages are frowned upon, and when they do take place one of the partners usually is required to convert. Even here, the power of territorially based solidarity is manifest: it is the person who takes up residence in the spouse’s village who changes religion.

Ethnicity is another factor considered by Rotumans in account­ing for behavioral differences on the island. The obvious cases are when Europeans, Fijians, or Indians are involved. Attitudinally, there is a hierarchical structure of stereotypes for these three groups. While Europeans are regarded as superior and are afford­ed deference (although they are also seen as an enigma), Rotumans 170regard Fijians and Indians as of lesser status than themselves and sometimes treat them with mild disdain. Because the number of such cultural aliens on Rotuma has always been very small in the past, Rotumans have not been under pressure to differentiate themselves as an ethnic group while confined to the island. Al­though they developed relatively clear conceptions of other ethnic categories, their conception of “Rotuman” remained vague. In large part it remained vague because the great majority of people on the island rarely if ever interacted with non-Rotumans, and so the interfaces between ethnic groups remained shadowy. It was only after people gained a sense of what it is like to be treated as a Rotuman (rather than as a farmer, a man from the district Oinafa, a chief) that a sense of ethnicity crystallized. Our argument is that this did not occur until substantial Rotuman enclaves developed in Fiji. In recent years, as the circulation of people between Rotuma and Fiji has increased to the point that most adults on the island have spent some time in Fiji, awareness of Rotuman ethnicity has spread throughout the population. Even so, such ethnic identity is salient only in Fiji as a basis for self-identity and for ordering social relationships.


Rotuman emigration to Fiji in substantial numbers has been rela­tively recent. The census of 1921 shows only 123 Rotumans, or 5.5 percent of the total Rotuman population, residing in Fiji (Fiji Legislative Council 1922). Fifteen years later the figure had risen only to 273 persons, representing 9.7 percent of all Rotumans. Since 1936, however, the rise has been rapid—to 569 persons in 1946 (17.2 percent) and 1,429 persons in 1956 (32.3 percent). The biggest Rotuman concentration in 1956 was in Ba Province, the site of a large gold mining industry. Most Rotumans living in Ba reside in Vatukoula, where the mine is located, or in the nearby town of Tavua. In 1956, when the Fiji census was taken, the Rotuman population of Ba totaled 669. The second largest con­centration was in Suva city, with 372 Rotumans. Third came Lautoka township with 71 Rotumans, then Levuka township with 56 Rotumans. These four locations accounted for 81.7 percent of all Rotumans living in Fiji at the time. Vatukoula not only con­tained the most Rotumans in absolute terms during 1956 but also 171showed the highest ratio of Rotumans to others (103 per 1,000); next came Levuka (37 per 1,000), then Lautoka (10 per 1,000), and finally Suva (1 per l,000) (McArthur 1958).

By 1961, when we conducted our census of Rotumans in Fiji, the overall number of Rotumans in Fiji had swelled considerably. Increases were taking place selectively, however, with Suva and Lautoka absorbing almost all additional migrants and Vatukoula and Levuka remaining nearly constant. Thus the 1966 Fiji census shows 986 Rotumans in Suva, an increase over 1956 of 165 per­cent, and Lautoka shows an increase to 187 Rotumans for an in­crease of 163 percent (Zwart 1968). These were, in effect, open towns from the standpoint of Rotuman migrants. The Rotuman population of Levuka, on the other hand, decreased by 14 (–25 percent) and that of Vatukoula decreased by 3 (–0.6 percent). These were closed communities from a migratory viewpoint. Dur­ing the same period the population of Rotuma increased by 7 per­cent. We shall refer to Rotumans in each of these urban areas as constituting an “ethnic enclave”—that is, members of an ethnic category who are residentially embedded in a sociopolitical unit dominated by others.

Before going on to a comparative analysis of Rotuman com­munities in Fiji, it is important for our argument to describe brief­ly the social structure of ethnicity in Fiji at the time of our study. One may gain a good initial picture of ethnic divisions from the dominant European perspective by referring to the census cate­gories used. The 1956 census lists seven categories: Chinese and part-Chinese, European, part-European, Fijian, Indian, Rotuman, and Other Pacific Islander. Broadly speaking, and again from a European point of view, these groups may be arranged in three major status categories with Europeans at the top, part-Europeans intermediate, and native populations (including Indians as well as Fijians and Rotumans) at the bottom. The Chinese are generally less visible socially and their rank is less clearly defined. There are, however, refinements within these groups, one being that the Poly­nesian Rotumans are generally regarded as more advanced than the Melanesian Fijians.5 The key population from the standpoint of ethnic mobility within this system is the part-European group. Because they are racially mixed, social entry into this group is less rigidly bounded than those based on “pure” race. Thus an educat­ed Fijian remains just that, unless he happens to have a European 172ancestor and shows at least some European racial features; he can then pass as a part-European and probably increase his social privilege. It is significant for our purposes that Rotumans enjoy a distinct advantage over Fijians and Indians with regard to this mobility channel. As Polynesians, they were favored as mates and mistresses by European men, so a high proportion of Rotumans have a European ancestor. But apart from that, their physical type is closer to that of the stereotypic part-European, making it easier to gain acceptance without resorting to genealogical credentials. This circumstance probably has retarded the consolidation of a Rotuman ethnic identity within Fiji in some respects. It was easy enough, while numbers were small, for Rotumans to pass for part-European, particularly since a high proportion of early emigrants were in professional roles such as teachers, and medical officers. The situation in Levuka during 1960 was probably indicative of this early phase.


Levuka, on the island of Ovalau, was the original capital of Fiji when the colony was formed. After the capital was shifted to Suva, Levuka remained an administrative center (the location of the eastern commissioner, who holds jurisdiction over Rotuma), but its importance slipped as a commercial and trading town. The population of Levuka in 1956 was 1,535, including 56 persons registered as Rotumans (McArthur 1958).

At the time of our study only six fully Rotuman households ex­isted in Levuka. Three of these were headed by men of professional status. A fourth was headed by a physician, Dr. Kautane, who ranks as the senior assistant medical officer on the island of Ovalau.6 The other two Rotuman households were headed by a clerical worker and a postman. In addition to these, there were two Rotumans (living with non-Rotuman spouses) and two Rotu­man men, each of whom had a Rotuman mother and a European father. The community was rounded out by seven student board­ers and five Catholic nuns.

A significant feature of the Levuka enclave is that most of the residents were assigned to their positions; they did not opt to go there in search of employment or to be with relatives. In fact, most of the residents are functionally nonkinsmen. This distinguishes 173Levuka from the other communities to be discussed, in which kin­ship has played an important role in expanding and organizing the enclave. As a corollary to this, Rotumans in Levuka are geograph­ically scattered instead of being clustered in a neighborhood.

The Rotuman enclave in Levuka has no formal organization, and no exclusively Rotuman clubs have been formed. Dr. Kautane is the unquestioned leader of the enclave, but strictly in an in­formal fashion. He is the one to whom people go for advice regard­ing things Rotuman, and he serves as a critical link with the home island by transmitting and receiving information. It is to him that Rotumans outside the Levuka enclave look when mobilization of resources is required. His primary credentials are extraordinary prestige within the broader community and relatively lengthy res­idence in Levuka as well as compelling personal characteristics. He is one of three native members of the Masonic lodge in Fiji and a member of two primarily European clubs; his closest friends are European and part-European. He owns his own well-furnished and spacious home, which serves as a hostel for Rotuman school­children studying in Levuka. Dr. Kautane is fluent in English and Fijian, although Rotuman remains the predominant language within his household.

Socially, then, the Rotuman enclave in Levuka forms a loosely knit network with Dr. Kautane as the major node. Interaction is most frequent among the professional men and their families, although there are occasions, such as births, when most members of the network are present. But these occasions are rare, and what is more important, most persons include in their intimate network several non-Rotumans. Also of relevance is that Levuka is a small town, and, particularly among the professionals, people are placed socially more by their positions than their ethnicity. As a result, the ethnic boundaries circumscribing Rotuman ethnicity in Levuka are permeable. Whatever centripetal forces are generated by a common language and sense of kinship are more than ba­lanced by such centrifugal forces as professional association, interethnic organization, and neighborhood scatter.


Unlike Levuka, Lautoka was a rapidly expanding town during our period of research. A new wharf had just been completed, and in 174addition to being the commercial and administrative center for one side of Viti Levu, Lautoka was beginning to serve as a major international seaport as well. Previously, the town centered main­ly on the Colonial Sugar Refinery and had served as a market town for the sugar plantations which occupy much of the land around it. The 1956 census showed a population of 7,420 for Lautoka, in­cluding 71 Rotumans (McArthur 1958); but by 1960 the popula­tion had climbed above 10,000, and the number of Rotumans had more than doubled. Our questionnaire on residential mobility revealed that the Rotuman population in Lautoka includes few short-term visitors, particularly very few of those from Rotuma who intend to return to their home island. In this respect it con­trasts most with Suva, where a high proportion of households in­clude short-term “guests.” As in Levuka, the Rotumans in Lautoka are residentially scattered, but the Lautoka community does con­tain a core network of closely related families.

Although there are no formal Rotuman organizations in Lautoka, the level of Rotuman-oriented activity is higher and the formalization of leadership is somewhat greater than in Levuka. A monthly service is held in the Rotuman language at the local Methodist church with the two Rotuman preachers in Lautoka presiding. Unscheduled meetings of the entire Rotuman communi­ty in Lautoka are called every month or so by Mekatoa, the ac­knowledged leader of most of the families in Lautoka. Very little business is discussed at these gatherings according to Mekatoa, but he believes they are necessary to keep the Rotuman communi­ty together. Because of the larger population, and owing to the greater degree of kin relatedness than in Levuka, there are more births, marriages, and funerals to bring people together and rein­force their sense of Rotuman identity, but these still occur at ir­regular intervals and with much less frequency than in Rotuma proper. In an attempt to perpetuate Rotuman identity among the children growing up in Lautoka, a night school was organized some years ago to teach them the essentials of Rotuman custom, but the venture did not take and dissolved from unknown causes.

Mekatoa has resided in Lautoka since 1939 and is employed as a fitter for the Public Works Department. He is acknowledged by all but three families to be the informal leader of the Rotumans in Lautoka. The three families who do not recognize Mekatoa’s lead­ership broke with him after an incident involving kinsmen in 175Vatukoula and now look to one of their own for leadership. Mekatoa also serves as coordinator of the Rotumans within the Methodist church. As a leader, he enjoys neither the legitimacy of Rotuman chieftainship nor the charisma of Dr. Kautane in Levu­ka. His main credentials, in fact, come from his long-term resi­dence in Lautoka and familiarity with the local scene. Whenever a new Rotuman family comes to Lautoka, they are expected to in­form Mekatoa of their arrival and intentions; he then keeps them informed about Rotuman affairs.

Discussions with Mekatoa indicated that keeping the Rotuman community together in Lautoka takes a strong conscious effort on his part; without it, he says, the community would dissolve and Rotuman custom would be neglected. The factional dispute men­tioned above is only one indication of the tenuousness of group solidarity. Although the Rotumans in Lautoka are more organized than those in Levuka, they do not form a cohesive group. Ethnicity there has not yet clearly emerged as the primary basis for structur­ing social relationships, although it is clearly of significance.


Suva is the city of Fiji. It is the center of government, commerce, and entertainment and by far the most cosmopolitan of Fiji’s ur­ban areas. The population of Suva in 1956 was 37,371, of whom 372 were Rotumans (McArthur 1958). Residentially, Rotumans concentrate in a few clusters in different parts of the city; general­ly the clusters are formed around acknowledged kinship ties. The range of occupations represented among Rotumans in Suva is greater than in any of the other communities, and the degree of residential fluidity is greatest there. Persons coming to Fiji from Rotuma are most likely to spend their initial time in Suva, either because it is the center for services they are seeking (medical, gov­ernmental, educational) or because it offers the most by way of ur­ban contrast with Rotuma. The entire picture, reflecting that of the general urban milieu, is one of considerable social, economic, and residential fluidity. Suva is the place where Rotumans come to seek their fortune, so to speak, and for many this changes on a dai­ly basis.

Whereas the Rotuman enclaves in Levuka and Lautoka could be considered as singular loosely knit networks, in Suva it would 176be more accurate to characterize the social arrangement as con­sisting of several closely knit networks within a rather open-ended system of relationships. For one thing, class differences based on Europeanization and educational and occupational differences are more pronounced in Suva than elsewhere in Fiji and they are reflected in contrastive life-styles. There are also several Rotuman clubs to be found in Suva, some of which are exclusive to district of origin in Rotuma and help newcomers adjust to the city, al­though others are open to all Rotumans and serve as sports clubs as well as fraternal organizations. Both the Methodist and the Catholic churches in Suva regularly perform services in the Rotuman language, and each sponsors Rotuman-oriented ac­tivities such as bazaars and bingo.

Leadership within the Suva community is essentially informal, as in Levuka, but it is multiple. Several Rotuman men with high positions in the professions or in government reside in Suva, and each is looked up to by a portion of the enclave. They are asked for advice on issues pertaining to their competencies, but none is ac­knowledged by all to be their spokesman. Several attempts have been made to organize the entire community, but all have been short-lived. It seems that internal differences of interest are too great, and the pressures from outside too little, to sustain solidari­ty. Nevertheless, it is far easier for an immigrant to remain wholly within a Rotuman social world in Suva than it is in either Levuka or Lautoka since the variety of Rotuman-held jobs encompasses the entire range of services available without going beyond the boundaries of the ethnic enclave. This is made possible by the larger size of the Suva enclave and by residential clustering in parts of the city.

Suva thus seems to provide conditions conducive both to open­ing and to closing ethnic boundaries. Among the Europeanized professionals and white-collar workers, it is often expedient to minimize one’s Rotuman background and pass as a part-European or to leave the whole question of ethnicity unspoken. Some mini­mize their affiliation with other Rotumans, including kinsmen, in order to reduce the drain on their accumulating resources. For these individuals Rotuman ethnicity plays a minimal role in struc­turing their social life. For others, however, the fact of “Rotuman­ness” becomes paramount. They are aware that the vast majority of people in the city are ethnically different from themselves and 177speak languages they do not understand. They confine all signifi­cant social relations to the Rotuman enclave and come to see the contrast between Rotumans and non-Rotumans as the most signifi­cant ones in their social worlds.


Vatukoula grew up as a result of a gold mining operation begun in 1935 by three mining companies owned by overseas European in­terests. Initially it was assumed that the mining operation would be short-term and so it was based on open-cut work, but later on the lodes were found to have depth and underground shafts have sustained a commercially profitable operation. Two of the com­panies ceased operations in 1959, leaving the Emperor Gold Min­ing Company in complete control. At the time of our study the EGMC’s management formed the effective government for the en­tire community in the classic style of colonial enterprise.

The mine management explicitly divides its employees into eth­nic categories as follows: Europeans, part-Europeans (actually limited to Euronesians, or mixtures between Europeans and Pa­cific Islanders), Fijians, Rotumans, and Indians. Each ethnic group has been allocated living quarters supplied by the manage­ment. The quarters allocated to Rotuman workers are insufficient for their needs, and many are forced to reside 10 miles away in Tavua until additional housing is made available by the mine management. Unfurnished houses in Vatukoula are assigned to in­dividual workers and their families; the worker is responsible for the upkeep of the house and pays a modest rent. A worker is not permitted to sublet his house, and when he leaves the mine’s em­ployment he is obliged to vacate. The house is then reallocated by the mine’s management. Thus, although residence itself is quite stable in Vatukoula, there is an aura of impermanence within the community.

Although wages are the main basis of support, land for cultiva­tion is made available by request to the company. Despite the per­petuation of subsistence activities by almost all the Rotuman households, a fundamental alteration has occurred in the relation­ship between people and capital in this new environment. In Ro­tuma, a person’s descent group has use rights over his land and can make legitimate claims on it for copra cutting and residence 178sites. In Vatukoula, on the other hand, the sole criterion legitimiz­ing control of capital goods (house and cultivated land) is merit with the company. A result of this altered situation is that kins­men, including parents, may be considered parasitic in Vatukoula if they stay in a household to which they do not materially con­tribute. The critical distinction is that wages do not involve prior capital, and they can be accumulated. Traditional rights are therefore not involved in the same way, and the provision of sup­port is likely to be interpreted by a wage earner as an act of bene­volence rather than one of obligation. Nevertheless, Vatukoula had the lowest percentage of nuclear households and the highest percentage of expanded households of any of the Rotuman en­claves studied.7 This follows from the traditional Rotuman rule that those who are well off ought to nurture those who are not, and since employment in the mines is tantamount to being well off for Rotumans in Fiji, relatives are drawn to them. The net result is a high degree of intrahousehold conflict and strains on relationships that are more severe in Vatukoula than elsewhere. At the same time, some informants believe that wage earning tends to reduce disputes between households that stem from the system of land tenure on Rotuma. They point out that on Rotuma, when a man needs money he must take coconuts for copra off family land, thereby creating competition for limited resources, whereas in Vatukoula, as one man put it, “We earn our money by our own sweat and it is clean money.” When asked what he meant by “clean money,” he explained that it was free of the dirt of land problems and the potent curses that accompany family disputes.

Within the mining community itself, internal residence change is most often the result of house promotion. Thus whenever a house becomes vacant within the Rotuman allocation, workers with less desirable homes are given an opportunity to occupy it in order of merit with the company. This generally starts a chain response—a worker vacates his house in order to occupy another, someone in an inferior structure moves into his, and so on. Ulti­mately, this may result in someone who has been residing outside the company town in Tavua obtaining a company house. One consequence of this system is that job status within the company is directly translated into a highly visible form of social rank. This contrasts with Rotuma, where there is far less congruence between social status and quality of housing, and herein lies what may be a fundamental metaphoric distinction between the two communi­ties. 179 On Rotuma, social status often is symbolized in acts of social deference; in Vatukoula, it is the kind of house one resides in that conveys one’s social standing. Correspondingly, on Rotuma social merit is judged largely in terms of the degree to which a man uses his resources in the service of relationships and for community benefit; in Vatukoula social merit is very strongly (though not unequivocally) tied to the position a man holds in the mining com­pany.

These shifts in perspective are part and parcel of an adjustment to a wage-oriented market economy and away from an economic system based on subsistence and ceremonial redistribution. Al­though Rotuma itself is involved in the money economy of Fiji and the rest of the modern world, on the island money has been adapt­ed to the traditional system rather than having transformed it (see Howard 1970). In Vatukoula Rotuman custom has been adapted to the pressures of a capitalistic society; this is particularly evident in the way ceremonial events are handled. The most relevant so­cial aspect of such events on Rotuma, the ritualized redistribution of food, mats, and other items, is precisely the feature that came under heaviest attack in Vatukoula. On several occasions known to us, persons in Vatukoula refused to participate in ceremonial (redistributive) exchanges at weddings and other events involving close kinsmen and insisted on giving a cash gift instead. The motives behind such deviations from custom seem to be based on a growing economic conservatism oriented toward maintaining a life-style commensurate with one’s rank in the company and a cautious but nevertheless intense desire on the part of some leaders to raise the Rotumans’ standard of living and esteem vis-à-vis other groups. Characteristically, every leader or would-be leader has a scheme of some sort for improving the economic well-being of the Rotuman community. Rather than being aimed at accumu­lating more goods, these plans are calculated to save money. This preoccupation appears to characterize Rotuman attitudes when dealing with collective assets, not only in Vatukoula but on Rotuma as well. The Rotuma Development Fund and the Rotuma Cooperative Association, for example, both have accumulated substantial assets which, despite prodding by the colonial govern­ment, remain unspent. In neither case are the Rotumans willing to eliminate the copra taxes and high prices on goods, despite the fact that these are genuine burdens on the population.

One can only speculate about the reasons for this disposition. 180Perhaps it has to do with pride. We believe that to Rotumans the accumulation of money is symbolic of a capacity to master the so­cioeconomic system that has been imposed on them. The meta­phoric power of the symbol lies, we suspect, in the measure of in­dependence that is predicated on having capital reserves.

The concern of Rotumans for retaining independence and con­trol over their own affairs has been expressed in several ways in Vatukoula, often to the dismay of the mine management. This fea­ture of Rotuman coping tactics is evident in the view held by Mr. Carson, a European, the mine’s welfare officer. From our field notes come Carson’s observations.

Mr. Carson feels that one of the problems in his relations with the Rotumans is that they tend to allocate themselves more power than they actually have. An example of this problem is that the Rotumans believe they should have the power to allocate housing. The mine management assigns housing facilities on the basis of seniority of merits. The Rotuman community has various other criteria of seniori­ty that the mine management does not recognize, and this is the basis of the conflict. Mr. Carson states that the heads of the Rotuman com­munity approached him once and wanted their native minister to have a house better than he deserved by his other merits. After a good deal of consideration, Mr. Carson pulled all available strings and got him the house in question. This was all done with the recognition by Mr. Carson that the minister was a man of great value to the com­munity as a whole. He confides that he is still feeling the dissatisfac­tion of his superiors from that move. The Rotumans have come to him recently and not only told him who should go into a given empty Rotuman house, but have declared that a vacant European house across the field should be let to a Rotuman family.

This concern for housing, incidentally, suggests that although Rotumans have accepted the symbolic significance of housing for social status they are unwilling to yield completely to the mine’s unilateral right to assign that status.

Another illustration of this desire to control their own destiny is the Rotuman mess hall, which is run exclusively by Rotuman shareholders. Each worker has a card that is punched every time he has a meal. At the end of each month, the cards are totaled and a list is sent to the company. The company then subtracts that amount from the individual’s wages and turns it over to the Rotu­man mess; profits are then distributed to the shareholders. What is 181significant about this is that the Rotumans are the only ones in Vatukoula who take care of their own food. The part-European and European mess are run by contract to a Chinese caterer; the Fijian mess is taken care of by the company. The advantage en­joyed by Rotumans in their arrangement lies not only in profits but also in the capacity to allocate jobs within the mess to Rotu­mans.

This ability of the Rotumans to organize, and the attitudes underlying their quest for control, can be better understood in the perspective of the way leadership has evolved within the commun­ity. The first pure ‘headman’ (a person with the right to make deci­sions for a collective) was Tafaki, who was also the first Rotuman to be employed at the mines (in 1939).8 He had a reputation in 1960 for having been too weak in his dealings with the mine man­agement. Tafaki’s headmanship ended with his discharge from the company after he left his wife and family and ran off with another woman.

After a brief interval, Riamkau, an electrician with the com­pany and a man of strong character, was chosen as ‘headman’ by the Rotuman employees. In a short time he had gained a commit­ment from the company for better housing, but his aggressive manner also generated some antagonism within the community. Then Chief Tausia, one of the seven paramount chiefs from Rotu­ma, visited Vatukoula in 1950 and appointed another man, Vai, as ‘headman’. Our informants claimed that this move was unpop­ular but encountered no overt opposition. Vai remained ‘head­man’ until his death in 1960. He was described as a weak leader, somewhat like a Rotuman chief whose concern is more with ritual honor than with the instrumental exigencies of leadership. It seems evident that despite Vai’s formal role as ‘headman’, Riam­kau, who assumed a chiefly title in the mid-1950s, retained a great deal of influence in the community and was the dominant political force. Thus when Vai returned to Rotuma in 1959 to discuss the effects of an ill-fated land commission, Riamkau took over in his absence and immediately introduced some dramatic structural changes. He appointed a committee composed of one man of chiefly descent from each district on Rotuma and then held a meeting of the entire community and obtained a confirmational vote.9 Upon Vai’s return, Riamkau turned the role of leader back to him, but the committee remained operative. 182

Interestingly, the resultant structure very nearly duplicated the social structure on Rotuma. Thus the ‘headman’ in Vatukoula was put in a very similar position to the district officer on Rotuma, and the committee corresponded to the Council of Chiefs. Even the monthly meetings, which rotated among committee members’ households, paralleled the Rotuman custom of rotating host dis­tricts. After Vai’s death, a meeting of the entire community was held in the Rotuman hall (built by the mines for the exclusive use of the Rotuman community) for the purpose of selecting a new ‘headman’. Riamkau was elected. Acting on a proposal by one of the defeated candidates, the committee then passed a motion limiting the term of the ‘headman’ to two years. The inference was that Vai, who had been in the office for ten years, would have been replaced under such an arrangement.

The committee arrangement created some problems for the mine management in their dealings with the Rotuman communi­ty. Many of the problems that arose in relations between Rotu­mans and the mine management required, in the latter’s opinion, more rapid decision making than was possible under the new ar­rangement. Furthermore, whereas Vai had been employed in Car­son’s department (a position virtually ensuring subservience), Riamkau is an electrician and works in a different part of the mine’s operation. As a solution, Carson proposed that Sosefo Holt, a young, rather Europeanized Rotuman, be appointed clerical assistant in his office to act as a liaison between himself and Riamkau. This proposal was rejected by the Rotumans, in large measure, we were told, because the Rotumans regarded Sosefo as a man who was strictly out for his own interests and would not adequately represent the community. It is likely, of course, that the mine management was well aware of the potential such an ar­rangement would have had for diluting Riamkau’s leadership and Rotuman solidarity in general. Riamkau had made it clear in his election platform that he was not afraid of the management and would try to push for the welfare of the Rotumans even if his posi­tion with the mine would be jeopardized. As the following passage from our field notes makes clear, he was tapping a basic Rotuman attitude:

Tomasi says that Vai was fine for dealing within the Rotuman com­munity itself, but he was too masraga ‘shy’, ‘respectfully deferential’ 183to present Rotuman views forcefully to the European administrators. Riamkau, on the other hand, will go all the way to the general man­ager if he sees fit and is not afraid to deal with the management on even terms. Tomasi expressed in his conversation that the Europeans are always trying to buy out Rotuman leaders.

Rotuman suspiciousness of Europeans as being clandestine manipu­lators out to get around the Rotuman people seems to be one of the Rotumans’ big leadership problems. The Rotuman leader who is well aware of European mannerisms and customs, and displays them pub­licly, is often suspected of lacking allegiance to the Rotuman com­munity. Another problem, leading to misunderstandings between Rotuman leaders and European administrators, is the reluctance Rotumans show in passing vital information to the Europeans for fear it will be used to their own detriment.

Despite expressions of overall solidarity, including firm dealings with the management and the refusal of Rotumans to work on a day following the death of one of their number, lines of cleavage do exist within the Rotuman community. These are generally kept out of the management’s view. In addition to district of origin on Rotuma, recognition of which has been made explicit in the for­mation of the committee, kinship and religion remain powerful organizational principles among Rotumans in Vatukoula. Kinship figures prominently in recruiting for jobs and in structuring in­formal relations, but it can also be divisive in that leaders are under pressure to favor their kin in decisions requiring impartiali­ty. Also, as previously reported, the expectations of visiting rela­tives concerning extended, dependent visits is frequently a cause of intrafamilial conflict. The Catholic-Methodist dichotomy also re­mains potentially schismatic but thus far has not resulted in fac­tional conflict. In general, it was our impression that church-oriented activities are somewhat less central in people’s lives than on Rotuma. For example, the Catholic group had not held a kato­aga ‘large-scale feast in honor of a notable event’ for nine years, the last time being upon completion of a new church. On Rotuma, during our year of fieldwork, two such feasts were held.

Despite these lines of cleavage, the overwhelming impression we received in Vatukoula was one of community solidarity and ethnic pride. In the mines, being a Rotuman seemed to be more important to people’s sense of identity than being from Oinafa, be­ing a Catholic, being so-and-so’s kinsman, or being a winder-driver. 184 People spoke of “Rotumans” in reference-group terms far more often in Vatukoula than elsewhere, including Rotuma, and were concerned with their reputation as an ethnic group in more active ways. They had clearly extended their idea of personal rela­tionships to include any person who could be identified as Rotuman.

The development of firm ethnic boundaries that has taken place in Vatukoula has resulted in sharpened ethnic stereotyping and a crystallization of intergroup attitudes. Let us now consider Rotu­man-other relations in this context.

If there is any dominant quality governing attitudes of others toward Rotumans and vice versa, it could be characterized as am­bivalence. On the whole, the Europeans at the mine and elsewhere in Fiji have high regard for Rotumans in comparison with other native peoples. This is reflected both in the high proportion of Rotumans employed in the mines and in their overrepresentation in positions of responsibility. European managers of various mine departments were nearly universal in their praise of Rotuman em­ployees. Despite such praise, it was our feeling that the general at­titude of Europeans was somewhat condescending, that the praise had an implicit (if not explicit) condition—in comparison with other native peoples. It was as if their assumption is that native peoples are generally rather hopeless and that Rotumans some­times surprise them.

Rotuman pride is something of an anathema to many Europe­ans precisely because Rotumans refuse to conform to the docile, childlike native of the European stereotype. Thus Mr. Dawson, the stock manager for the mines, openly dislikes the Rotumans. “They haven’t an ounce of brains, and besides, they hate Europeans,” he commented. When asked how they show their hostility, he could not pinpoint any specific actions, but his analysis made it clear that he equates hostility with refusing to accept European domi­nance unconditionally. The reasons he gives for the failure of Fi­jians to perform better at the mines is instructive. He attributes their lack of success to the refusal of most Europeans to join them in their work. Too often, he claims, Europeans tell the Fijians what to do and then go away, as if to show that they would never do that kind of work themselves. In his own dealings with Fijians, Dawson says that he gets right in there with them, “even if it means getting mud on my boots and getting my hands dirty”; as 185long as he is with them, “I’d match my Fijians against any group in Vatukoula” (our emphasis). He adds that he would rather have a not-so-smart fellow who is willing to learn as best he can than a smart one, because the smart ones are those who will fight for themselves and are not “behind you.” Mr. Carson’s complaints about Rotumans allocating too much authority to themselves, re­ported above, also illustrate the irritation caused to Europeans by Rotuman pride and self-respect.

The Rotumans, for their part, acknowledge the social superiori­ty of Europeans only inasmuch as it is associated with standard of living, education, and occupation. They do not acknowledge ra­cial superiority, nor do they accept everything culturally Euro­pean as superior to those practices that are culturally Rotuman. In short, they perceive no insurmountable barriers in their Rotuman ethnicity to achieving an acceptable position in the modern world.

One manifestation of the fluidity with which Rotumans per­ceive racial boundaries is the ease with which they slip into the part-European category after gaining an education and when it suits their purpose. The advantage of passing for part-European rather than Rotuman stems from European rather than Rotuman ethnic conceptions. In general, the Rotuman stereotype of part-Europeans is unfavorable; they are seen as pretentious, particular­ly since the behavior of several of the more familiar models is less than exemplary. But being a part-European provides the possibili­ty for entrance into the European social world in a way that being a Rotuman does not. There are six such people in Vatukoula, and their attempt to pass as part-European signals not only an aspira­tion to move up the ethnic hierarchy but also an alienation from the Rotuman community.

Relations between Rotumans and Fijians are likewise marked by strong ambivalences. In general, Rotuman attitudes toward Fi­jians parallel the attitudes of Europeans—a mixture of mild dis­dain with patronizing condescension. They see Fijians fairly much in the mold of indigenes quite a bit more primitive than them­selves. Yet Rotumans hold Fijian chiefs in high regard and show them the ritual courtesies they would show their own chiefs; in this sense they see themselves more as part of an indigenous world in which mana and other aspects of Malayo-Polynesian superna­turalism are significant considerations. Since Rotumans do not practice sorcery whereas Fijians do, the latter are a source of awe 186if not fear. In general, though, Rotumans in Vatukoula have come to see themselves as competitive with Fijians. In charitable ven­tures in Vatukoula, for example, Rotumans attempt to outdo Fi­jians (and other groups) in a massive, public presentation of their contribution.10

The prevailing attitude of Fijians toward Rotumans appears to be one of resentment. Thus it was reported to us by several sources that incidents of hostility between Rotumans and Fijians were not unusual and were caused in large measure by Fijian resentment of privileges enjoyed by Rotumans in the mines. During the previous year, following a massive layoff of personnel, Fijian antipathy to Rotumans reached a boiling point. The general consensus was that this occurred because only one Rotuman was among those dis­missed. Apparently the matter cooled after a ceremonial presenta­tion of kava by the ‘headman’ of the Rotuman community to the head of the Fijian community. The headmaster of the local school also stressed Fijian resentment of Rotuman achievement. He stated that Rotuman children appear to be much brighter on the average than Fijian children and this results in jealousy. Fijian teachers are unnecessarily harsh with their Rotuman students, he maintains, and will assign them all the unpleasant jobs, such as cleaning lavatories, while assigning the pleasurable ones to the Fi­jians. They never put a Rotuman child in charge of Fijians, but always do the opposite. Rotuman teachers are discriminatory in a reverse fashion, he says, but with somewhat less vigor.

Before concluding this section on ethnic relations it may be ap­propriate to comment on language use. In general, most Rotuman men learn to get on well in both English and Fijian. English is necessary to comprehend information passed down from man­agers and is clearly the status language in the overall community. Fijian, on the other hand, is frequently necessary to communicate with Fijian workers whose command of English is poor and who cannot be expected to learn Rotuman. Most Rotuman women learn Fijian, but they are less likely than the men to be ac­complished in English. This is because they are able to deal with shopkeepers, service suppliers, and in some instances servants in Fijian even if exchanges are with Indians, but they have less in­teraction than men with Europeans or other exclusive English-speakers.11 187


We believe the evidence we have presented demonstrates that the development of ethnic communities and ethnic consciousness varies markedly within the different social milieus in which mi­grants live. In this concluding section we discuss some of the vari­ables that appear to have had significant effects on these processes for Rotumans in urban Fiji. What we would like to account for by reference to these variables are (1) the degree to which individuals from a given ethnic category (in this case Rotuman) confine their meaningful social relations to persons of a like background, (2) the degree to which ethnicity provides a basis for formal organization, and (3) the degree to which ethnic identity becomes salient in ordering social relations with persons who do not share the same background. Loosely speaking, we believe our study suggests a rank ordering of the four communities with regard to the impor­tance of Rotuman ethnicity. In Levuka it has the least effect, in Lautoka and Suva it is intermediate, and in Vatukoula it is a domi­nant principle.12

The variables affecting ethnicity can be classified into three types, demographic, social structural, and cultural. We have al­ready mentioned the prime demographic variable required for the formation of an ethnic group in the sense we are using the phrase—the existence of an out-group, a people sufficiently contrastive in diacritical features to create a sense of in-group identity. For ethnic boundaries to be formed and actively maintained requires, as we pointed out in the introduction, regular contact with at least one other group. When Rotumans were confined largely to their home island, opportunities for interacting with non-Rotumans were highly restricted, thereby limiting the kinds of experience upon which a solid sense of ethnic identity could be based. In Fiji, however, Rotumans are in regular interaction with several distinc­tive out-groups.

The absolute and relative size of an ethnic enclave appears to have a significant effect. If the number of individuals in a group is small, the possibilities for organizing along ethnic grounds may be too restrictive, given a minimal number of roles that must be played in a viable organization. If all are kinsmen, of course, they may in fact form a tightly organized group, but chances are that 188ethnicity will play a salient role, especially if nonkin are required to fill crucial organizational positions. From the standpoint of other persons in the town, ethnic stereotyping becomes a conve­nient means of ordering social relations only when a sufficient number of persons become socially visible to provide a consistent set of expectations. If this is correct, there is probably a “critical mass” or threshold required for ethnicity to become salient. Thus we believe it is no accident that the boundaries around Rotuman ethnicity roughly follow size of Rotuman population in the four towns, being least distinct in Levuka and most pronounced in Vatukoula.

Relative size of population also may exert an influence, in­asmuch as it affects overall visibility. Even though several hun­dred members of an ethnic group may dwell in a city, if they are scattered and form an insignificant portion of the population they may be absorbed without their ethnicity becoming salient. One way in which this sometimes happens is for such people to be in­corporated into a more inclusive stereotype—as Scandinavians rather than Swedes or Norwegians, as Orientals rather than Japa­nese or Chinese, as Polynesians rather than Rotumans or Samoans. Scattered residence patterns may not only diminish ethnic salien­cy by making a group less visible socially; it also reduces interac­tion among members of the group and makes organization more difficult. It likewise increases interaction with members of out-groups who are neighbors or who perform localized services, gen­erating friendships and cooperative relationships across ethnic lines. Contrariwise, condensed residential patterns are likely to facilitate organizational potential and diminish meaningful exter­nal contacts. Stability of residence is likely to be another factor, since the crystallization of ethnic identity is probably facilitated by feedback within fixed communication networks. Also, if per­sonnel are continually changing, organizational potential may be hampered and leadership rendered more problematic. It seems clear that both the nucleated residence pattern and the relative stability of residence in Vatukoula have greatly increased the ca­pacity for organization of Rotumans there in comparison with those in Suva.

One further demographic variable seems worthy of mention al­though its effects are far from obvious—the degree to which a community is growing or declining in size. Our hypothesis is that 189growth through immigration tends to increase ethnic conscious­ness because of the continual need to socialize newcomers, a pro­cess frequently requiring the explication of boundary mecha­nisms.

With regard to social structural variables, one must distinguish between those that are imposed by sources outside the ethnic com­munity, particularly those prevailing in dominant sociopolitical groups, and those endemic to the ethnic enclave. To the extent that the dominant society makes ethnicity a major criterion for defin­ing social roles and social privilege, one would expect ethnic con­sciousness to be fostered. As we have already pointed out, in Fiji ethnicity has been the major criterion for allocating privilege, with the gold mining community in Vatukoula epitomizing the situation. Thus the Rotumans coming to Fiji stepped into a social structure that sought to classify them by race from the very begin­ning. The process was fostered by their distinctiveness from Fi­jians in racial type (closer to Polynesian) and language; it was probably furthered by European favoritism for Polynesians (with whom Rotumans are generally classified by Europeans) over the darker Melanesians (including Fijians). It was to the Rotumans’ advantage to accept if not nurture the distinction. However, the presence of the part-European, or Euronesian, group probably has had a reverse effect. What is significant about this ethnic category, in addition to the fact that it is second in ethnic rank to European, is that its boundaries are fuzzy; it is therefore easily permeated by those who look like they might have some European blood, speak English reasonably well, and display appropriate decorum. On looks alone, it is easier for Rotumans to pass into this category than any other ethnic group in Fiji. Rotumans who have acquired an education, and particularly those who are in professional or quasi-professional roles, have often elected to pass as part-European.

Social structural variables internal to an ethnic enclave may be equally important for the crystallization of ethnic identity. We might begin by considering a major point of articulation between the ethnic community and the larger structure—the allocation of jobs. It seems clear that in wage-earning, market-oriented soci­eties, one of the primary bases of shared interest is comparability of position in the occupational structure. Men who work together in parallel roles tend to identify with one another and share com­mon 190 concerns. This was particularly evident in the gold mines, where the work is frequently dangerous and where safety and well-being are directly in the hands of one’s work mates. In Lau­toka, Suva, and Levuka, by contrast, Rotumans are unlikely to be working together in such teamlike efforts. Our hypothesis is that the sharing of work roles greatly increases male solidarity and in turn fosters the development of ethnic solidarity.

Another variable favoring the development of viable ethnic communities is effective, legitimized leadership. Migration may create some difficulties on this score. Thus, on Rotuma, chieftain­ship is essentially localized and related to the land; Rotumans in Fiji are deprived of these criteria. If a man assumes a Rotuman ti­tle while in Fiji, which is possible, it will be a title from Oinafa or Pepjei or some other Rotuman district. Other members of the en­clave are from other districts, and although they may pay appro­priate ritual deference, they are unlikely to accept his secular leadership as legitimated by the title. Legitimation of leadership in Fiji has therefore become associated with elections, and com­petency with political effectiveness in the larger community. Whereas effective leadership appears to have a strong centripetal effect on ethnic solidarity, ineffective leadership tends to produce factional disputes along lines of existing cleavages. These may be along kin lines, as in Lautoka, prior locality in the homeland, or religion and occupation. We would therefore advance the proposi­tion that the creation of ethnic solidarity is inversely related to the number of salient divisive criteria within the community as well as the effectiveness and legitimacy of leadership.

There are undoubtedly many features of culture that bear on degree of solidarity among ethnic enclaves, and to complicate matters they may operate at different levels in communication systems. For example, the degree to which a people perceive their customs to be compatible with those of other cultural groups is ob­viously relevant: if the enclave regards outsiders’ customs as repugnant, or vice versa, this is likely to inhibit assimilation and lead to rigid ethnic boundaries. But at a broader conceptual level, the very way in which cultural formulations about differences in custom are arrived at may be significant. In short, variations in epistemology of cultural difference may be of greater importance than the differences themselves. Whereas one group may postulate crucial differences to be racial (this seems to have been character­istic 191 of colonizing Europeans), another may hold supernatural belief systems, language, or custom to be crucial. Clearly, these different views have different implications for the formation of ethnic boundaries. A people may be able to change their language but they cannot readily change their physical characteristics.

Another significant variable has to do with the importance of being a member of a culturally cohesive community. Some cultur­al systems produce individuals who feel personally immobilized unless they are part of an integral community, or at least they derive great pleasure from being part of one. Those reared in such a tradition tend to form compact ethnic communities even when they are few in number. Other cultural systems place a premium on independence and the maintenance of social distance from others; individuals from a background of this type may self-con­sciously avoid forming close ties with other members of their ethnic category. The Rotumans are intermediate between these ex­tremes. They seek neither to converge with nor to avoid other Rotumans with any pronounced motivation.

At an even broader level, cultures may vary in the degree to which they emphasize abstract formulations of cultural dif­ferences. We have already proposed that Western-educated Rotu­man leaders are likely to be more conservative than chiefs without Western education precisely because they have learned to make abstract contrastive judgments about social systems and cultural styles (Howard 1963b). The point is that the very concept of in­tegrity of a cultural system may be of major significance. For the Rotumans, then, Western education has provided the cultural equivalent of a concept of tribal integrity in strongly unilineal societies. It has helped to provide clear criteria for inclusion in a social unit of a higher order despite the fact that the traditional system was characterized by groupings with highly permeable social boundaries.

It would seem, then, that in addition to the demographic and social structural variables that foster the development of ethnic communities, the emergence of an ethnic group is facilitated by the presence of individuals for whom ethnic identity not only becomes problematic but is of ideological import. Although such individuals may develop within an ethnic community, we believe it is more often the case that they are the products of isolation from their native cultural systems, with the very isolation height­ening 192 their ethnic awareness. Western schools are breeding grounds of such individuals by virtue of the degree to which they render one’s identity problematic (particularly for non-Occiden­tals) and the degree to which ideological solutions to identity prob­lems are encouraged. But ideological solutions are apt to remain idiosyncratic unless they feed back into communication networks like those provided by ethnic communities. The Vatukoula com­munity and its leadership exemplify this process. When this oc­curs, and an ideology gains acceptance, conditions are optimal for transforming an ethnic aggregate into an ethnic group.


We have argued in this chapter that the development of an ethnic group from an aggregate of individuals who are members of the same ethnic category is primarily dependent upon the develop­ment of an ethnic community and ethnic consciousness. Ethnic communities are defined as localized interactive networks con­sisting of individuals of the same ethnic designation who are emo­tionally committed to the symbols of their common heritage and formally organized for the purpose of pursuing common goals. Ethnic consciousness is defined as a condition in which ethnic awareness assumes a position of primacy in structuring social relations. For a collectivity, ethnic consciousness is assumed to emerge as a result of repeated messages circulated throughout the network of an ethnic community to the effect that other social dif­ferences are less significant for structuring interpersonal relations than ethnic differences. The combination of these conditions gen­erates a critical mass, or threshold effect, leading to the extension of individuals’ integrity circles to include all members who iden­tify themselves in terms of the relevant ethnic category (Howard and Howard 1964). This process results in the development of a secondary community in which the “we-feeling” characteristic of primary face-to-face groups is extended to other members of the ethnic category on the basis of an ideology.

We hypothesize that the major variables responsible for the transformation of an ethnic aggregate into an ethnic group are demographic, social structural, and cultural For Rotumans in Fiji the major demographic variables favoring the development of ethnic communities have been numbers of individuals and resi­dential 193 contiguity. Only in Vatukoula have these variables pro­duced a cohesive community. Social structural variables also have favored Vatukoula as a location for the genesis of Rotuman ethni­city, in large part because the management of the gold mines has used racial criteria as the primary basis for organizing labor. Demographic and social structural variables have therefore com­bined in Vatukoula to make it the primary place in Fiji for a criti­cal mass to be reached, allowing for the crystallization of Rotu­man ethnic identity. The cultural variable of prime significance has been the development of an ideology of cultural contrast, in­troduced by a Western-educated elite. Although the birth of the Rotumans as an ethnic group has taken place in Fiji, we expect that the ideology which gives it substance will eventually be ac­cepted by all Rotumans and that a general consolidation will be the result.

The Rotuman case may contain some unique features, but we believe that the processes analyzed here may provide the basis for a universal theory of ethnic group development.


Support for the original research was provided by the National Institute of Men­tal Health. The National Science Foundation and East-West Population Institute provided additional support for the analysis of data and writing up of the mate­rial. We are grateful to all three agencies. We also wish to acknowledge the ex­cellent editorial suggestions of Michael Lieber and the secretarial assistance of Helen Takeuchi.

1. Like E. K. Francis, we conceive of an ethnic group as a kind of “secondary community” in which the we-feeling characteristic of a primary face-to-face group is extended to others on the basis of an ideology. An ethnic group is, in Francis’s conceptualization (1947:399), “the most inclusive, cumulative, and realistic type of secondary community.”

2. It is of some interest that several Rotuman clubs in Fiji are named for prom­inent geographical features of the home island

3. For more extensive treatments of Rotuman society, see Howard (1963b, 1964, 1970).

4. This and what follows reflects the situation during the period of fieldwork in 1960–1961.

5. The category “Other Pacific Islander” is further differentiated for census purposes into Polynesian, Melanesian, and Micronesian, suggesting that these distinctions have a social significance for Europeans in a formal as well as informal sense.

6. Pseudonyms are used throughout this chapter. 194

7. In Vatukoula 21.7 percent of the Rotuman households are nuclear, with 65.2 percent lineally or laterally expanded; comparable figures for Suva are 23.3 percent and 63.3 percent, for Levuka 25.0 percent and 33.3 percent, for Lautoka 33.3 percent and 40 percent, for Tavua 45.0 percent and 45.0 per­cent.

8. The term used on Rotuma for ‘chiefs’ or ‘subchiefs’, who would ordinarily hold ritually assumed titles, is gagaja. The term pure on Rotuma, in addition to being used to designate the informal leader of a work group, is used to de­signate the steward of a landholding kainaga.

9. Tafaki, who had subsequently been rehired by the mines, was one of those appointed.

10. This sort of competition is also practiced among church parishes and among districts on Rotuma. In Vatukoula it is other ethnic groups that are the out-groups, but they are structurally isomorphic with the church parishes and districts in the competitive context.

11. This, incidentally, is another manifestation of relatively high Rotuman social status; quite a few Rotuman women have Fijian servants, but we know of no instances of the reverse occurring.

12. Our concern here, it should be made clear, is only with the development of ethnic organization and consciousness. The variables that maintain ethnic boundaries in established social systems over the long run may be of quite a different nature.


Map 9. Relocation of Gilbertese to Solomon Islands.

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.