Home Page
Howard-Rensel Papers

3. Land, Activity Systems, and Decision-making Models in Rotuma

Alan Howard

[Published in Ethnology 2:407-440, 1963]

The island of Rotuma is located at approximately 12° S. lat. and 177° E. long., on the western fringe of Polynesia. Politically, Rotuma has been governed as part of the Colony of Fiji since its cession to Great Britain in 1881. [1] Although its linguistic affiliations remain somewhat of an enigma (Grace 1959; Capell 1962), the culture of the island reveals a closer affinity with Samoa, Tonga, Futuna, and Uvea than with Fiji or the Melanesian islands to the west. Descriptions of selected aspects of Rotuman culture and social organization have appeared in print (cf. Howard 1963a), but as yet no comprehensive account exists. The ethnographic information contained in this paper is presented in lieu of a more thorough description which I hope eventually to produce. The main purpose of the present article is not meant to be ethnographic, however, but ethnological. Specifically, it is my intention to use data from Rotuma as a basis for discussing some of the theoretical and methodological issues posed by societies emphasizing bilateral kinship organization.

In recent years bilateral, or cognatic, kinship organization has become a focal point in the comparative study of social structure. At stake are some basic issues, including the whole conceptual apparatus of structural analysis, and it seems likely that when these issues are resolved, structural theory will have undergone some profound changes. The fundamental issue is not in my opinion a conceptual one, e.g., descent vs. filiation, or cognatic vs. bilateral vs. nonunilinear, but rather concerns the nature of social systems as such. In short, is it best (or simply necessary) to regard social systems as "statistical" models as Leach (1960) has at times suggested and Murdock (1949, 1960) has continually maintained, or to consider them as "mechanical" models as Goodenough (1961) and Sahlins (1963) have suggested? [2] The issue goes deeper still; it implicates the problem of explanation that lies at the core of the Homans and Schneider vs. Needham debate. Indeed, there would be justification for regarding it as the fundamental issue in social anthropology today. Let us review the arguments advanced in support of either side.

Murdock, in defending the statistical approach, asserts that to regard structural rules as exerting pressures upon individuals is to reify the concept of social structure, and endorses a statement by Leach (1960:124) that "social structures are sometimes best regarded as the statistical outcome of multiple individual choices rather than a direct reflection of jural rules." Murdock (1960: 9) supports his position with the assertion that such an approach has "the enormous advantage of making possible the utilization of psychological principles and of scientific knowledge concerning the dynamics of cultural change in the interpretation of social systems."

One of the most explicit proponents of the opposite approach has been Sahlins. In a review of Social Structure in Southeast Asia, Sahlins accuses Murdock of looking at social structure from "the inside out," and asserts that the more fruitful approach would be to examine it from "the outside in" (i.e., the relationship between units in a political system). The emphasis would then be upon "ideology and mechanism." From this point of view, according to Sahlins (1963: 45), "political groupings of descent order seem to form a continuum: dogma ranges between an extreme emphasis upon patriliny to a mere emphasis upon common descent (nonunilinear) groups."

It would seem, then, that we are forced to choose between a view of society as a statistical model resulting from a multitude of individual choices and encompassing the totality of social behavior, and a mechanical model based upon intergroup relations and including only ideal, or alternatively, politically oriented, behavior. If this were the case, it would constitute a most unfortunate situation, since both positions have shortcomings. On the one hand, mechanical models are superior to statistical models as conceptual instruments for scientific analysis; on the other hand, any model which dismisses a large portion of social behavior as irrelevant leaves much to be desired. Fortunately, however, these are not our only alternatives. Goodenough (1961: 1343), in discussing Murdock's 1960 article, makes a plea for an alternative approach:

It is high time...that we develop a typology that is completely free of statistical and functional considerations, using only structural or formal ones, based on the criteria and principles by which people make membership decisions (as distinct from the kinds of alignments which tend to result from the making of these decisions under a particular set of stable conditions).

Let us examine Goodenough's statement from the standpoint of the requirements for constructing the kind of typology he suggests. To begin with, of course, we would need a series of ethnographic descriptions containing adequate data. The question then becomes, what constitutes adequate data? For an answer to this question we can turn to Goodenough's other writings, beginning with his monograph on Truk. It is here that we find his first published statement (Goodenough 1951: 10) describing the goal of ethnographic description in terms of "trying to give the reader a basis for learning to operate in terms of the culture described in somewhat the same manner that a grammar would provide him with a basis for learning to speak a language." Expanding upon this theme in a later paper (Goodenough 1957), he suggests that an ideal ethnographic description of cultural behavior would constitute a formulation of what one would have to know in order to act in a culturally appropriate manner in given social circumstances.

The essence of this approach is that ethnographic description should provide sufficient data to permit the construction of decision-making models in which the culturally perceived alternatives are designated, the principles (or factors) which are determinate for choosing between alternatives described, and the relationship between the factors specified (as in factor analysis). In other words, a mechanical model based upon individual choice and predictive of actual behavior would ideally result.

It is worth noting that these premises underlie the newly emergent field of cultural ecology (cf. Frake 1962: 54), and are implicit in recent studies focused on the cognitive principles underlying folk taxonomy (e.g., Frake 1961). Social anthropologists have been slow to take the cue, however, and it is one of the purposes of this paper to present the results of an experiment with using this same logic as a basis for analyzing a wider aspect of social phenomena, in this case land tenure. But before presenting the data it is necessary to consider some basic concepts.


It is not my intention to review here the various definitions of social structure that appear in the literature. The point I want to make at the outset is that any definition that treats societies as though they were uni-structural models conflicts with the premises we have put forth. The objections to the single structure approach have been succinctly stated by Nadel (1957:153):

...it seems impossible to speak of social structure in the singular Analysis in terms of structure is incapable of presenting whole societies; nor, which means the same, can any society be raid to exhibit an embracing structure as we understand the term. There are always cleavages, dissociations, enclaves, so that any description. alleged to present a single structure will in fact present only a fragmentary or one-sided picture

The essence of what I am suggesting is this: If we accept the ideal that ethnographic description should permit the reader to act appropriately in the society being described, the question "What are the principles of social structure?" is inappropriate, for what we would have done is to postulate an entity (social system) and then ask how it is built. It would seem the more appropriate question should be, "What are the principles that structure behavior under given circumstances?" This is an entirely different question and releases us from the enigma of reification. The fact is that societies are not systems in the proper sense of the term, if by "system" we imply a conceptual order in which the specified units are in determinate relationship with one another. For one thing, the units we are forced to use when employing the uni-structural approach--e.g., positions, roles, groups, and institutions; or, alternatively, social relations, norms, and institutions--are themselves abstracted from average behavior (or expectations of behavior) in a wide variety of situations. They are therefore unlikely to be predictive for any single situation, and hence fall short of the desired goal. In sum, the concept of society as a single system is based on an analogy (Radcliffe-Brown 1952: 178187), and to base a science on an analogy hardly seems satisfactory. Nor can we be satisfied with assuming partial determination. As Nadel (1957:144) puts it:

...if we are concerned (as in fact we are) with discovering the mutual determination of a plurality of elements, as in a "system," we shall naturally assume, to begin with, that such a determination exists, in which case we must also assume the possibility of a complete state of this kind, a state wholly determined by the interrelation of all elements, and hence unchangeable while the elements are what they are.

Instead of conceiving of a society as having a social structure, I would suggest we conceive of social behavior as being structured by participation in given activities within which behavioral choices (decisions) are regular and predictable. Our "systems" would then best be regarded as activity systems, the relevant units being the principles (or, methodologically, factors) that are predictive of choice among behavioral alternatives.

How, then, shall we define the concept of social structure or, as I would have it, the structure of an activity system? First, it is evident that, if consistency is to be maintained, one criterion that must be met is that any definition be congruent with the notion of human behavior as decision-making. Taking this into account, I would suggest that the structure of an activity system be defined as a set of interrelated principles by which the participants performing an activity (or series of activities) determine who shall make decisions (or be held responsible for them) on issues that implicate all, or a part, of the group (two or more participants).

To develop comprehensive models of decision-making behavior we need to go further. Namely, we must specify the principles by which the persons selected by structural rules actually make decisions. To the extent that these are cognitively shared I would choose to call them cultural principles. An example may help to clarify these points.

Consider the case of a man, his wife, and child engaged in preparing a feast at Rotuma. Two orders of problem are involved. The first is: Who shall decide what is to be done? The principles used to resolve this problem are structural principles. The second order of problem is: What shall be done? For example, what food should be used and how shall it be cooked? Who shall perform the various tasks, and who should be served first? The principles used to solve these problems are cultural principles.

To extend the example further, let us consider an actual situation that recurs in Rotuma and examine it. Suppose the child had just returned home from an extended holiday, or had just received some form of public recognition (such as first communion). These would be equivalent circumstances as far as the preparation of the feast is concerned, since in either case it would be regarded as an expression of parental (or familial) recognition of the youngster. If we were to ask a great many Rotumans, "Who should make decisions?" and "What procedures should be followed?" under such conditions, a high degree of consensus would no doubt emerge, from which we could derive an "ideal" or "jural" decision-making model such as the following:

Two structural principles apply: a husband is entitled to make decisions in this activity system when a problem involves himself and his wife, and parents are entitled to make decisions in preference to children. In other word., the man is entitled to make decisions affecting all three of them, or himself and either of them. The woman is obliged to accept her husband's decisions, but in matters affecting only herself and her child she is entitled to make the decisions.

According to custom (i.e. recognized cultural principles), a pig should be killed and cooked in an earthen oven or koua, which is prepared by the man. The woman should cook all the other foodstuffs which are not to be baked in the oven, while the child is freed from performing any tasks. When the food has been cooked, the child is supposed to be served first and given the head of the pig, which in Rotuma is the prestigeful part of the animal.

This model is "ideal" in the sense that virtually all Rotumans would agree that it is how things should be done. It is "jural" in the sense that unless something interferes, thus requiring other ideal principles to be taken into account, such behavior is invariably met with approval.

But how predictive is this model for the way in which such a feast would actually be carried out in Rotuma? Is it a satisfactory "grammar" in Goodenough's sense? Rather than answering these questions, I shall only make the methodological point that, in order to construct a predictive model, an investigator would have to observe a number of feasts held under these circumstances and from his observations would have to derive empirically the principles of action that best fit (i.e., are retrospectively predictive of) the way things are actually done.

There is one important point that should be made clear since it is at variance with the generally accepted point of view. Whereas most anthropologists have dealt with such aspects of the socio- cultural order as rules of succession and residence as structural principles, within the present frame of reference they are classified rather as cultural principles. Only when priority of decision-making is involved are we dealing with structural principles as I have defined them. In other words, the problem of determining who shall choose a successor to chieftainship is resolved by employing structural principles, but the problem of selecting the grounds on which he shall be chosen involves the use of cultural principles. Likewise, within the present context, kinship principles may be either structural or cultural. They are structural when used to determine decision-making priority; they are cultural when used as a means of recruiting assistance. While this specialized usage is somewhat unorthodox, it is necessary if logical consistency is to be maintained.

I now propose to describe certain features of Rotuman social organization that have a bearing on some of the issues under discussion. After presenting the historical background, I intend to describe the "ideal," or "jural," model which Rotumans portray as pertaining to land tenure, and then shall consider actual behavior patterns when questions involving land arise. The latter will be formulated as decision-making models, i.e., guides to behavior that appear, on the basis of my ethnographic data, to be predictive under defined circumstances. I have chosen to focus upon land tenure for two reasons. In the first place, the transmission of property rights is a crucial aspect of the descent group controversy; and, second, it is an example of an area that has usually been treated as part of a total social system but which, in my opinion, is better regarded as a related set of activity systems, each with a distinct decision-making model.

It should be stressed that the formulation of decision-making models does not, by itself, constitute an adequate substitute for complete ethnographic description. To be adequate as an ethnographic account data relating to attitudes and values affecting land, the use to which it is put and its significance in the lives of the Rotumans would have to be presented. More important still, a statistical record of the incidence of each kind of issue and the decisions actually made would need to be included. Such material has been omitted for the sake of brevity and because I have not yet completed a statistical analysis of the relevant data. I hope to include these in a future publication. The information analyzed to date, however, supports the conclusions that are drawn regarding Rotuman practices.


The best early account of land tenure in Rotuma is that by Gardiner (1898). Even at the time of Gardiner's visit land tenure had already changed to a considerable degree and was in a state of transition, but he was able to obtain sufficient information to make a fairly comprehensive reconstruction of the former model. His account is consistent with all other available data and therefore appears to be reliable.

According to Gardiner, the aboriginal landholding unit was the ho'aga, which was a kinship community under the direction of a titled chief, the fa esho'aga. The chief acted as steward (pure) of the land, and it was his duty to divide it among the component households of the ho'aga for planting purposes. He also settled disputes if any arose within the ho'aga. Although these units were organized into districts (seven at the time of Cession) under the political leadership of a paramount chief (gagaj es itu), the latter had no claim in the land, although in some cases he was given first fruits and was called upon to settle disputes between ho'aga. Indications are that ho'aga were exogamous and that the residence pattern was bilocal, depending upon such factors as available land, relative status, and the nature of the marital arrangement (elopement, arranged marriages, etc.). Succession to the chiefly title took place within the male line, although it is not clear whether primogeniture or senior status was the dominant principle.

As a result of acculturation with Europeans, which became increasingly intense during the nineteenth century, land tenure in Rotuma was considerably altered. The cumulative effect of three factors in particular--a decline in population, the development of a commercial economy, and the establishment of missions--led to the breakdown of the ho'aga as a kinship unit and the individuation of land holdings. As adjacent ho'aga amalgamated in order to maintain adequate manpower in the face of depopulation, the kinship unity of the group was lost, and ho'aga evolved into political units composed of a number of localized (usually adjacent) households. At the same time, the growth of the copra trade gave men a more permanent interest in the land than previously, and with the encouragement of the missionaries and traders a concept of individual ownership (i.e., the right to allocate and dispose of land) was fostered. In time the right of the fa esho'aga to distribute the land gave way to the rights of the user, and ho'aga land was divided into individual holdings with each household head assuming the rights of a pure over the land he worked. This undoubtedly did not take place as a consciously executed plan, but rather as a gradual process involving a growth of vested interests in specific blocks of land and a loss of authority on the part of the fa es ho'aga resulting from the intrusion of non-kinsmen into the group. The process of fragmentation which took place was furthered by frequent sales of land by the Rotumans among themselves for money, pigs, and various other items.

The establishment of the Colonial Administration following Cession also had a great impact on matters pertaining to land. Of greatest importance was the establishment of the administrative structure itself, which provided for the systematic handling of transactions and disputes. Within this structure the District Officer (previously the Resident Commissioner) is the supreme decision-maker and interpreter of custom. The Government further complicated the situation by introducing a land tax, the payment of which has come to be regarded as a means of legitimizing claims.

During the first two decades following Cession the Resident Commissioners followed a policy of resolving stewardship disputes by dividing the land among the disputants. This policy was pursued until early in the twentieth century and perpetuated the process of land fragmentation that had begun long before Cession. As the point of diminishing returns was reached, and land division was no longer feasible, a declaration of joint rights was favored in cases in which clear decisions could be made for neither plaintiff nor defendant. With the virtual elimination of land division, in conjunction with an expanding population, communal interest in land was increasingly perpetuated.

The handling of many land matters, however, has remained unaffected by these changes. It is only when disputes arise, transactions between nonrelatives take place, or taxes are not paid that the Administration gets involved, and for every land matter settled by the Administration many others are settled by the parties concerned among themselves.

In 1960, when field work was being carried out, Rotuma was a prosperous society. Its population had increased from a low of slightly more than 2,000 in 1915 to approximately 5,000 in 1960. The island itself, which is quite small (approximately 17 square miles), supports about 3,000 of these; the majority of the remainder now live in urban centers in Fiji (cf. Howard 1961). Although those who remain on the island are almost completely reliant upon the copra trade for money, their standard of living is high in comparison with most people in the non-Western world, and the land, which is incredibly fertile, is the source of their wealth.


The basis of Rotuman kinship can be designated by the word kainaga, which in its broadest sense means "kind, sort, variety, species, class" (Churchward 1940: 235), in other words, belonging to the same category. It can be used to describe people of the same nationality, or in a more limited sense to designate persons "of the same blood," i.e., consanguinity. Since kinship is traced bilaterally in Rotuma, a person's kainaga, in this latter sense, constitutes a personal kindred, a grouping which becomes functionally operative during life-crisis ceremonies and when an individual becomes critically ill. The term kainaga may also be used in a still more restricted sense, indicating common descent from an ancestor who has resided at, and held rights in, a given house site or fuag ri. [3] Each person is considered to have rights in the fuag ri of his eight great-grandparents. [4] Every site is named, and a person usually describes his affiliation by such a statement as "I am a member of the Halafa kainaga." Associated with each fuag ri are sections of bush land, presumably those over which the ancestor held rights, and to claim membership in a given kainaga is to claim rights in these lands. The person who lives on the fuag ri, and controls the land, is the pure. He (or she) is obligated to grant usufruct privileges to any member of the kainaga. If a pure is unreasonable or overly stingy, the kainaga have a right to hold a meeting and depose him in favor of another person. If he dies or otherwise leaves the ancestral fuag ri, the kainaga should hold a meeting to select a new man. At a kainaga meeting social relations are structured according to the principles of kinship, i.e., senior males are required to be least restrained in expressing their opinions, while the expectancy that others will acquiesce to their decisions increases to the degree that they owe them respect. The prescription for selecting a new pure is that the senior male of the kainaga should be chosen, seniority being based upon age in one's own generation and one's father's seniority in the parental generation. Theoretically, then, succession goes from elder brother to younger brother, to eldest son of elder brother, to younger son of elder brother, to elder son of younger brother, to younger son of younger brother. A woman may become pure only if there are no eligible males, and her eldest son is expected to succeed her provided she has no brothers with sons.

Some fuag ri carry with them chiefly titles, to which the men who become pure have claim. When titled they are known as as togi (successor to the name) and are entitled to the privileges, and burdened by the responsibilities, that go with the particular name. [5] The assumption of a title is not automatic, however, but requires a ceremony (hül 'umefe) in which the symbol of chiefly status, a short-legged eating table ('umefe), is turned upright, after which kava is partaken, the candidate anointed with oil, and a feast eaten.

That a person's eight kainaga correspond, in the formal model, with his kindred can be illustrated by reference to life-crisis ceremonies. When such a ceremony is about to take place, the eight fuag ri are supposed to be gathering points of Ego's relatives, each individual going to the one he has in common with Ego. If anyone has more than one kainaga in common he may choose among them at his discretion. The pure of each kainaga is in charge of the group that congregates at his fuag ri, and he decides what shall be done. He may call a meeting sometime before the ceremony is to take place, and allocate tasks. On the day of the ceremony each person brings his contribution of mats, food, kava, etc., to the appropriate fuag ri, where these goods are amassed. When all arrangements have been completed, the kainaga departs as a unit, under the leadership of the pure, to the fuag ri at which the ceremony is to take place. [6] Here they are ceremonially received and make their presentations. Any person who attends the affair, if he is not a member of any of Ego's kainaga, should go as the guest of one of them. Each group therefore regularly contains not only consanguineal relatives but affines and unrelated persons as well. These latter, it must be stressed, go as guests rather than as members of the kainaga.

Associated with each fuag ri are also ancestral ghosts, or atua, who at once protect the living members of the kainaga from outside harm and act as instruments of justice when internal squabbles occur. Thus an individual engaged in an illicit love affair may feel relatively safe from discovery if he takes his lover to an ancestral burial ground somewhere in the bush, or he may, if he feels he has been unjustly denied rights in kainaga land, implore the atua to see that justice is done. In some cases, special personal powers possessed by an ancestor are considered to be transmitted to his descendants by virtue of the power of his atua. These may include healing powers, the ability to catch turtles (a prestige food), or special prowess as a fisherman. It is generally held that the more recipients of such power the more diluted it becomes, and an alternative is for power to be personally transmitted from live individual to live individual through a special ritual. When this is done the donor loses the power and the recipient gains it in an undiluted form.

With regard to land tenure, the concept of kainaga rights has been complicated by the advent of individual ownership following European contact. This has resulted in a range of concepts to describe rights in land, consisting of the following seven concepts:

l. Hanua ne togi, which refers to land acquired by an individual through purchase. Rights thus obtained are undivided and the purchaser is undisputed pure.

2. Hanua ne na, which refers to land acquired by an individual through gift. The only distinction from hanua ne togi lies in the method of acquisition; rights are undivided and the recipient is undisputed pure.

3. Hanua ne haisasigi, which refers to land in which a sibling group and/or their known descendants own rights. Thus land which is individually owned by Ego becomes hanua ne haisasigi to his offspring (provided there are more than one) [7] and remains so in descending generations as long as genealogical connections to these siblings remain clear. When genealogical connections are obscured, the land becomes known as hanua ne kainaga.

4. Hanua ne kainaga, which refers to land over which all the descendants of ancestors who previously held rights own rights. Relationship may be traced back to any ancestor who held such rights, not necessarily to a specific person (such as a founding ancestor) or sibling group. Hanua ne kainaga is usually distinguished from hanua ne haisasigi in that the latter is generally reserved for cases in which genealogical connections are known, whereas the former is used where they are obscured. [8]

5. As ne hanua, which refers to land that "belongs" to a chiefly title. Since all members of the kainaga to which the title is relevant share rights in the land, functionally (within the formal model) as ne hanua is identical to either hanua ne kainaga or hanua ne haisasigi, as the case may be. The only distinction is the conception that the title holder is automatic pure.

6. Hanua ne 'on tore, which refers to land in which the collective descendants of a pure who had owned undivided rights own rights. The concept is more specific than hanua ne kainaga in that the latter does not include reference to an original ancestor who had enjoyed undivided rights; on the other hand it is less specific than hanua ne haisasigi in that it does not require knowledge of genealogical connections to the original ancestor.

7. Hanua pau which in its broadest sense includes any individually held land. In a more specific sense hanua pau refers to land in which the sole survivor of a kainaga owns rights. By implication, hanua pau (in its more restricted sense) was once hanua ne haisasigi, hanua ne kainaga, or hanua ne 'on tore. [9]

An interesting feature of this conceptual system is a built-in cyclical aspect. Individually held land (hanua ne togi, hanua ne na, or hanua pau) becomes hanua ne haisasigi in the second generation, providing the original pure has multiple offspring. The land remains hanua ne haisasigi until genealogical connections are obscured (probably in the fourth or fifth generation), when it becomes known as hanua ne kainaga or hanua ne 'on tore. This in turn may become hanua pau should all but one of the kainaga die out.

As a general rule, persons who are related are expected to be generous with one another. Thus if a person requires land for subsistence purposes, he needs only to ask the pure of one of his kainaga for the privilege, and the latter is obliged to grant permission if land is available. The fact that the land is to be used for subsistence purposes is sufficient justification. Once permission is granted, it is required by custom that the pure be given first fruits. If the recipient plants any permanent crops, e.g., breadfruit or coconut trees, he may expect use of these throughout his lifetime provided he does not otherwise violate the pure's hospitality, but upon his death the trees become part of the land. A request to cut copra requires specific justification, however, since it is tantamount to a request for money. The pure is expected to grant such a request to a member of his kainaga (i.e., the one over which he is pure) provided the suppliant's need is "legitimate" and provided his own needs, or prior commitments, are not interfered with.

Personal property may also be acquired from relatives, but here the emphasis is upon reciprocity. Reciprocal borrowing is an institutionalized pattern in Rotuma (between nonkin as well as between kinsmen) and is signified by the word fara. The request should be made humbly, in pleading fashion. Need is implied in such requests, and if the person being asked is approached in an appropriate manner he should not refuse unless he has a good reason. By its very character, fara applies only to useful goods rather than luxuries, and the nature of the commodity determines whether it is to be returned or not. Durable goods, such as tools, are expected to be returned when they have fulfilled their utility. Perishable goods, such as food or copra, cannot be returned, but in either case the obligation to reciprocate is implied in the borrowing. One of the most frequent uses of fara is to obtain ceremonial exchange goods, particularly mats and pigs, for presentation at life-crisis ceremonies.

In addition to this institutionalized pattern of borrowing, a person may request assistance (faksoro) [10] from kinsmen and/or nonkin. A person may faksoro labor to build a house, cut large quantities of copra, or prepare a feast. The person who asks for labor creates an obligation to reciprocate when others come faksoro to him, and he is required to feed the laborers during the time they are working for him.


Having described the ideal model, let us now go on to consider four types of activity associated with land. These are usufruct, succession, transactions, and disputes.


When land is unlimited and person A makes use of a given plot, this need not concern person B (provided no other special significance has been attached to the land) since he can readily find land for his own purposes without undue hardship. Where land is limited, however, if A uses a piece of land, this restricts the land available to B. On this basis we can postulate that conflicts of interest, and hence the frequency of decision-making circumstances, are likely to be greater where the ratio of population to usable land is such that each person cannot obtain all of the land he can use. Under such circumstances, a set of rules, consisting of both structural and cultural principles designed to regulate usufruct is necessary if conflicts are to be avoided. As a derivative postulate, it is to be expected that the rules governing usufruct of scarce commodities, or products of limitless value, will be more exacting than those involving plentiful commodities of limited value. To state this another way, we would expect a person with decision-making rights over a piece of land to be willing to grant usufruct license [11] over land in inverse proportion to the value of the anticipated yield. When the commodities involved are of limited value, the person desiring to use them may need little or no justification for making his request, but when the commodities involved are valuable, justification on the basis of ideal cultural principles is likely to be required.

In order to understand the significance of the above postulation for Rotuma, a distinction must be made between coconut trees as property and plantation land. The Rotumans, unlike some other Pacific Islanders, do not distinguish rights over trees from rights over land, in that the pure of the land is automatically pure over the trees on the land; but at the same time there is a considerable difference in the willingness of individuals to grant license over the two. Many men, for example, are pure over lands with far more planting space than they personally need or can possibly use. Men with such extensive holdings freely grant usufruct privileges to friends and neighbors, as well as to relatives--often for no better reason than that the land is conveniently located [12] or for the pleasure of working near comrades. Granting usufruct privileges under these circumstances is an inexpensive way of raising prestige and bolstering status through a display of generosity, which is a primary Rotuman virtue. This same pattern of generosity and permissiveness extends to any tree fruits not specifically planted, such as oranges, mangos, fava (Pometia pinnata) and vi (Spondias dulcis). These may be taken in moderation by anyone passing through the bush without asking the permission of the owner. It is also regarded as permissible for an individual, in want of a drink, to climb a tree to get a few young coconuts for this purpose. The only exception to this free use of tree products for subsistence reasons occurs when a tree has been made taboo (fapui), usually by binding a coconut leaf or a coconut to the trunk of the tree. When a tree is so marked it is a warning to others not to take fruit from it, the implication being that the pure himself needs the fruit for a special purpose.

Granting license over coconuts for copra is an entirely different matter. Since copra brings money, and money is a means to wealth, a person could not possibly take another's copra without a conflict of interests. Wealth, unlike subsistence crops, is cumulative, and each person can use all the copra he has. A pure may therefore be expected to grant usufruct privileges over coconut trees only with good reason. Thus, whereas the use of land for subsistence purposes is usually freely granted and no explanation is demanded, requests for copra must be justified on the basis of cultural principles. Several considerations enter into a request for copra, including relationship, need, previous obligations, and relative wealth.

Relationship may be used to justify requests for copra cutting rights in two ways. In the first case, a person may make his request on the basis of his belonging to the kainaga with which the land is associated. The logical basis for such a request may be stated in this way:

Formerly a lineal ascendant of mine was pure over this land. Since parents are expected freely to grant license over property to children, and to their children's children, and to their children's children's children, etc., the original pure would freely have granted usufruct rights to me. So here I am.

A request based upon this principle involves a direct relationship to the land and only an indirect relationship to the pure of the land. The evidence, however, leads me to believe that requests are rarely made on this basis.

In the second instance, a person may make this request on the basis of his relationship to the pure himself. When this is the case, the important variable is social distance, which for the most part appears to be a function of genealogical and residential distance. Thus, while relations between first cousins are likely to be intimate regardless of residential distance, social relations between relatives further removed are apt to remain active only when they live in the same community, i.e., are in frequent face-to-face interaction. With near relatives it does not matter much whether the land involved is hanua ne kainaga or individually held, since license is based on their relationship to one another rather than an abstract conception of rights to the land. It is significant that under such circumstances gifts of money are often given as a substitute for copra rights--a further indication that it is the social distance between individuals that is of paramount importance, rather than the relationship of persons to the land.

A second principle generally required to justify a request for copra is need. The purpose for which the person requesting rights wants money is taken into account by the pure, and if he does not regard the reason as legitimate he may feel justified in refusing. A person requesting rights would also be expected to have exploited all of the more readily accessible resources, including his own lands and those of relatives closer to him. Not only the absolute need of the person making the request is taken into account, but also the relative needs of the two parties involved. If the pure has a legitimate need for money himself, or if he has a previous commitment to a closer relative, he can be expected to refuse.

Previous obligations are also taken into account in determining the legitimacy of a person's request for copra. On innumerable occasions relatives exchange gifts, give aid, and do favors for one another. A careful, though informal, account is kept by most persons of favors owed, and there is little reluctance to make requests for property rights on the basis of obligations so incurred. A request for a cutting of copra that is backed up by a reminder of a previous gift or favor carries much more weight than one made without such backing.

Relative wealth is a fourth factor. One of the prominent Rotuman values is that the rich give to the poor. Underlying this value, of course, is an awareness that a wealthy man is more likely to have a surplus which he can invest in establishing a reputation as a "good man." A pure with especially large land holdings is therefore more likely to be asked for copra-cutting privileges than one with meager holdings, and the refusal of a reasonable request by a wealthy man is much less easy to defend than a refusal by one with lesser resources.

Still another consideration may enter into the picture--the question of etiquette. The attitude displayed by the applicant is of primary concern in such matters. Rotuman custom requires a person making a request to do so humbly. Even though the pure may be indebted to the applicant, etiquette requires the latter to disguise what may, in fact, be a demand as a humble plea. This can be understood from the standpoint of social economics as a bartering of one's personal dignity (and hence status) in exchange for consideration of the request. It is very awkward for a Rotuman to refuse a man who comes fara in the customary manner, and to do so requires great delicacy. On the other hand, if a man were to make his request in the form of a demand, or without a show of humility, he would almost certainly be refused, for by implication this would be a challenge to the pure's decision-making rights.

With regard to usufruct, then, the ideal model is of little consequence for actual behavior. When land is wanted for subsistence purposes it is freely lent, and kinship makes almost no difference whatsoever. Nor are other cultural principles likely to be brought into play to justify a request. When copra-cutting rights are at stake, however, a situation requiring strategy is operative. Thus, although a pure may at times wish to give money or copra to close relatives, it is apparent that pure are not motivated to give away copra to more distant kinsmen if they can help it, and an applicant must bolster his request by bringing into play the cultural principles we have discussed. He may prepare for his approach by attempting to reduce social distance or by establishing obligations, but to the extent that he ignores the relevant cultural principles he facilitates the ease with which the pure can refuse his request. The pure, on his part, can attempt to increase social distance and avoid obligations with a person he suspects wants a cutting of copra, or he can keep his resources tied up so that there will be no surplus available. The situation can thus best be understood as a kind of game in which the principles of strategy involve social distance, need, balance of obligations, relative wealth, and etiquette. It should be added, however, that even if a pure "loses," he still gains prestige for his generosity.


For the purpose of considering succession it is convenient to divide Rotuman conceptions of land tenure into two broad categories. The first includes those types of tenure in which rights are shared by members of only one household and consists of hanua ne togi, hanua ne na, and hanua pau. We shall simply call these individually held lands. The second includes those types of tenure in which rights are shared by members of more than one household, and consists of hanua ne haisasigi, as ne hanua, hanua nekainaga, and hanua ne 'on tore. We shall call these communally held lands.

The problem of determining who shall succeed to stewardship over a land holding arises on two occasions. One is when the acting pure dies; the other is when an acting pure defects, either by moving to another locality or by leaving the island. The circumstances under which succession takes place may differ markedly depending on whether a pure has died or defected, and they require separate consideration. We can therefore consider succession from the standpoint of four categorically distinct circumstances. These are:

(1) cases in which land is individually held and the pure dies;
(2) cases in which the land is individually held and the pure defects;
(3) cases in which the land is communally held and the pure dies;
(4) cases in which the land is communally held and the pure defects.

Under any of these circumstances it is necessary for a new pure (or an acting pure) to be chosen, and therefore a decision-making situation arises. It must be determined who possesses the right to choose the successor (on the basis of structural principles) and what criteria should be employed (cultural principles).

When the pure of individually held land dies, provided he has not made his wishes known to the contrary, the right to decide who shall become pure belongs to the deceased's ascendant, sibling, or descendant of senior standing, i.e., to, as it might be termed, the senior member of the deceased's nuclear kindred. Seniority is determined by the structural principles of kinship previously described. The operative cultural principles are those described for the ideal model; priority begins with the eldest male child of the deceased and passes successively to the youngest male child, the eldest female offspring, the youngest female offspring, and, in the subsequent generation, to the eldest male child of the eldest male child, and so on. The relevant principles can, if need be, be extended indefinitely to establish priority. It is important to note that only lineal descendants of the deceased pure can legitimately claim rights in the land. Collateral relatives, including the pure's own siblings, are technically ineligible.

In actual fact, however, the transference of stewardship rights is usually automatic and does not require a meeting or a formal decision-making procedure. Provided the household does not dissolve following the pure's death, the senior member (who is eligible by the rules of succession described above) of the household automatically assumes pure's rights unless he is challenged. In most cases, if the pure had been elderly, he had already turned over active stewardship to a suitable person in his household, usually his eldest resident son. The son thus may already have taken over responsibilities such as paying taxes and keeping the land clean, so that succession upon death merely adds de jure status to a de facto situation. A common technique for assuring that succession will go to a specific person, e.g., an adopted child, is for the pure to register the land at the Government station in this person's name rather than his own upon purchase or reception of gift. He remains acting pure as long as he so desires but has given a legal buttress to his desires as to who shall succeed him.

If the household dissolves on the pure's death and he has made no definite provision for a successor, as may happen when no obvious successor is available, then someone must decide who shall take control of the land. Even under these circumstances, however, a formal meeting is not likely to take place. Usually a close relative (in the social sense), even though he may not be formally eligible, simply takes over unless challenged. If he is challenged, a formal meeting of concerned persons may be held, but such an event when individually held land is at stake is extremely rare. In other cases the land is allowed to fall into disuse while nominal stewardship is assumed by an absent relative.

In cases in which the land is individually held and the pure defects, there is likely to be little room for argument. The pure simply requests someone of his own choosing to take care of the land while he is gone, usually turning over both pure's rights and responsibilities. The only complications that might arise occur when the acting pure either defects or dies. If he defects, and can contact the original pure, he is obliged to do so and to allow him to name a successor. If this is not convenient he can turn over pure's rights to any close relative of the original pure. If the acting pure dies, the original pure may request someone else to take over the land, or he may simply allow it to fallow in anticipation of eventually returning to use it himself. In most cases the persons involved are very close relatives, usually part of the pure's nuclear kindred, and the possibility of conflict is minimized.

In cases in which the land is communally held and the pure dies, all persons with a stake in the land should ideally hold a meeting at which a successor is chosen. The person in charge of the meeting should be the senior member of the kainaga, based on the principles of seniority previously outlined. The same person is also formally entitled to assume stewardship if he so desires. But these ideal circumstances are rare. In most cases the majority of related persons are already established on their own lands in their own villages and are not likely to want to make a change. Particularly if the land is of low yield, they may be rather indifferent to the whole business. A meeting of the kainaga may not even take place, therefore, unless someone specifically raises an issue. The more usual circumstance is for the senior member of the existing household to assume pure's rights over the land, provided he is a member of the kainaga. If this person is a married woman and her husband is living with her, he acts as pure, while she remains pure de jure.

If the household disintegrates after the death of the pure, or if someone claims priority over the persons left in the household, a meeting of the kainaga is ordinarily arranged. Actually, this rarely involves all persons with potentially valid claims; only those who have an active interest in the land are likely to attend. Provided that the senior member of the kainaga can be determined, he is responsible for making the decision, and if there are disagreements, he arbitrates. But in cases where the relevant group is large or diverse--in Rotuman terms, if the kainaga has many sides--there may be no easy way to determine who is the senior member. Matters then depend upon the respective claims of various persons. If one person claims seniority and is unopposed, he automatically assumes the position of decision-maker. If more than one claim is made, the rivals present their cases before the kainaga, and if the overwhelming majority support one party's claim the opposing parties are likely to yield.

Often, however, factionalism develops, generally along sociometric lines, focused around two or more strong claimants. From this point on, the situation becomes one demanding strategy on the part of the interested parties. Each group puts forth its arguments--at a formal meeting if one is held and/or informally whenever an opportunity presents itself. The purpose is dual: to gain the support of those kainaga members who may still be unaligned, and to elicit popular support from the community. Support from men of chiefly rank is particularly solicited, and the unequivocal backing of the paramount chief from the district in which the disputed land lies is generally decisive during this stage of the game. Support from men reputed to be learned in genealogical matters is also valuable, since formally the matter is one of genealogical priority. Without the aid of written records, however, genealogical connections, as well as events pertaining to ancestral transactions involving the land in question, are matters of opinion and are rarely capable of substantiation. The crucial problem is actually one of gaining public support, and a plausible genealogical claim is only an initial qualification for entering the arena. There are two important aspects of strategy of great consequence for swaying public opinion. The first involves attempting to convince concerned persons of one's greater relative need for the land in question; the second involves portrayal of oneself as a better man in the Rotuman sense (i.e., more humble, kind, generous, etc.). The claimants thus must "walk a tightrope," inasmuch as they must make forceful claims while at the same time attempting to portray a humble, generous personality. To be successful at this game one must know, of course, his own relative strengths and weaknesses as compared with those of his opponents, and act accordingly.

If none of the factions is able to gain a clear-cut victory in the arena of public opinion, the case may be brought to the District Officer's Court for resolution, but in most instances popular support for one faction crystallizes sufficiently so that other claimant groups withdraw their efforts. If the losing factions remain convinced that their claim is indeed superior, they may take comfort in the knowledge that justice is immanent from supernatural sources (God and/or the ancestral ghosts), and they may announce as much to their successful opponents. At this stage, two alternative avenues of action are open to the unsuccessful claimants. They may strive to insure that their claims are recognized by the victor as legitimate, though inferior; or they may initiate a civil suit in the District Officer's Court. The former stratagem is the safer play and, if successful, at least opens the door to future manipulation. The method is to admit the victor's priority, accompanied by a humble apology for having raised opposition and an excuse such as genealogical ignorance for having done so. A plea is then made that the victor recognize the rights of all concerned and that, as pure, he be generous to all members of the kainaga. Unless the victor has been seriously antagonized during the course of the negotiations, or unless he genuinely regards his opponents' claims as false, he is likely to comply, at least verbally, with such a request. The alternative avenue, initiating a court case, is a more desperate measure and if unsuccessful is likely seriously to hamper the faction's chances in future claims. These events may or may not take place within the context of a formal meeting of the kainaga or of a series of such meetings. They are just as likely to occur during a sequence of informal communications over a period of time.

Leaving aside the matter of disputes for the time being, once the senior member of the kainaga has been determined, he is supposed, if the rules of selection are carried through to their logical conclusion, to select as pure the person with the most senior status who is willing and able to assume the responsibility, provided he does not accept it himself. More often than not, however, his choice is determined by social distance to himself, often under the guise that this is a substitute for his assuming the position.

Essentially the same procedures are involved when a chiefly title is at stake, except that in some cases the conflict of interests is enhanced, particularly if the associated land is of high yield. Also, since a title usually involves some leadership responsibilities, every effort is made to find an eligible male to fill the role. There are, in fact, no titled women in Rotuma today, although it is not regarded as beyond the realm of possibility. The decision-making model for selecting a chief may also be complicated by outside influences. In many instances a district chief (gagaj es itu) has exerted pressures to get a sub-chief (fa es ho'aga) of his own choosing selected, and in the selection of a district chief customary principles may be overridden by the District Officer (cf. Howard 1963b).

Holding a title which includes as ne hanua opens the door for manipulation, since stewardship of the land is inherent in the title. Thus the titleholder can go to live on a fuag ri other than the one pertinent to the title and thereby gain control of two sets of kainaga land. Were untitled persons to evacuate the kainagafuag ri, a relative could cause embarrassment by asking to assume pure's rights. These circumstances derive from the fact that living on a kainaga fuag ri lends legitimacy to a pure's position, even though failure to live on the land does not necessarily require a person to relinquish his stewardship.

In cases in which the land is communally held, the pure runs the risk of losing his stewardship whenever he leaves the premises. Theoretically, the kainaga should meet for the purpose of selecting a new pure, and if the land has simply been vacated this is likely to happen. One way for the pure to prevent this is to delegate authority to someone of junior status to himself, e.g., a younger brother or sister. In this way he can be assured of regaining control of the land if and when he so desires. The person to whom control has been relegated becomes acting pure and exercises de facto decision-making rights. The original pure may or may not continue to pay taxes on the land. Many choose to do so, even though they may be away from the island, as a means of reinforcing the legitimacy of their right to assume pure's rights should they return. Since the Rotumans are a peripatetic people, and have a high rate of mobility not only between districts in Rotuma but between Rotuma and Fiji as well (Howard 1961), the de facto stewardship of particular land holdings may be transferred several times in such an informal fashion.

In actuality, since many persons are pure over both communally and individually held lands, more than one problem may be involved when a stewardship is vacated. In most cases, if the pure defects, he chooses a successor as acting pure who takes over control of all the lands, regardless of the type of tenure. If a pure with multiple holdings dies, and no one takes over unopposed, the individually held lands and those communally held are treated separately, each according to the models already described.

There is one other instance in which pure's rights are transferred. It sometimes happens that a man with a large holding will grant rights over a section of the land that is under his control, usually to one of his near relatives. When this occurs, the grant of license is actually a permit for extended usufruct, often including the right to build a home on the land. At most, the grantee has privileges comparable to those of an acting pure; at least, he is a subordinate with usufruct rights.

Transactions Involving Land

Theoretically, any piece of land can be sold or given as a gift so long as persons with rights in the land agree to its disposal. From the time of Cession the Colonial Administration has taken a vital interest in land traffic on Rotuma and has passed regulations to control it (Howard 1962: 118-163). The administrative procedure that was evolved to handle land exchanges was described by Resident Commissioner Macdonald in a meeting of the Rotuma Council of Chiefs during September, 1907 (Minutes of the Rotuma Council of Chiefs, September 5, 1907):

If any Rotuma Native wishes to sell, make a gift of, or exchange land with any other Rotuma Native, the Rule is for the former to intimate the sale etc. to the Chief of his district, who will inform me before all the Chiefs to make the matter of the sale etc. public, in the District Meetings on the following day, and to request anyone who considers that he has a claim in the ground and who has not been consulted regarding the sale or given his approval to it, to lodge his objection with me within the space of one month after the date of the district meeting. If no objection is lodged within that time, the sale, gift or exchange can then take place, and in the absence of fraud or mistake holds good according to law. The sale must take place before the Commr.

Where land transactions are concerned, "actively asserted rights" in land are of special significance. When individually held land is put on the market there is little chance for problems to arise, but when communally held land is offered, the pure must be careful to take into consideration any person who might lodge a complaint. In a great many cases, however, only a small proportion of the persons who could legitimately claim rights actually do so. For lands of moderate or low desirability, the group of actively interested persons tends to be limited to near relatives of the pure. Distant relatives generally place their stakes in more desirable lands and those under the control of pure who are closer to themselves. In order to put up a piece of land for sale, then, even a pure of kainaga land often needs to get permission from only a small group of near relatives. If he is the senior of this group he holds decision-making rights over them in any event and is thus not likely to face any formidable opposition. The same holds true when a pure wishes to make a gift of some land or to exchange it for another piece. Gifts may be given for a variety of reasons--as a means of passing land to a favored relative, as an act of generosity when a kinsman is in need, or as an act of gratitude for assistance "beyond the call of duty" by a friend or distant relative. Exchanges are generally a matter of expediency. They are usually based upon convenience of location, each pure trading a piece farther away from his home (or main bush lands) for a piece which is nearer. Regardless of the nature of the transaction, however, a pure must be quite sure of his position relative to other possible claimants if he is to avoid a legal dispute.

A tabulation of land transactions recorded at the Government station since Cession is presented in Table 1.


Land Transactions by Decade, 1881-1960


The statistics in Table 1 are of some interest. The virtual nonexistence of land transactions during the first two decades following Cession reflects the strong resistance of the first Resident Commissioners to any such dealings (cf. Howard 1962:134-143). During the period from 1900 to 1950 Government officers were more permissive, and the number of transactions recorded probably accounts for a high proportion of the dealings that actually took place. In addition to legal transactions, however, there can be little doubt that informal dealings have continually taken place, and the sharp decline in legal transactions between 1951 and 1960 probably reflects a reluctance by Administrators to handle them, rather than an actual decrease.

Disputes Involving Land

Disputes involving land can be divided into three categories: demands for usufruct, contested stewardship, and boundary disputes. Table 2 reveals the number of each kind of dispute per decade heard by the Land Court.


Land Disputes by Decade, 1881-1960

Usufruct Demands
Stewardship Disputes
Boundary Disputes

Let us take each type of dispute separately. If a pure refuses usufruct privileges to an individual who claims membership in the kainaga in question, and the latter regards the refusal as unjust, he can initiate a court case. The District Officer, if he decides for the claimant, can force the pure to yield. As can be seen from Table 2, however, except for one dramatic outbreak of cases, Rotumans have been reluctant to bring such cases to court. The reasons for the rash of cases during the 1911-1920 period are not clear (actually all the cases occurred between 1918 and 1920), but indications are that it was initiated by an interpretation of the law that encouraged demands for copra-cutting rights on the basis of membership in a kainaga. Previously--and subsequently--only claims in which land was wanted for subsistence purposes were given consideration.

Disputes involving stewardship generally arise in one of two ways: as the result of conflicting claims following the death or abdication of the pure of kainaga land, and as the result of an emergent conflict between relatives sharing the same land. The dynamics of the former instance have already been discussed. In the latter case the priority of rights is usually quite clear, and in most instances the subordinate party is likely to leave the holding, but in some cases a petition may be entered for a division of the land. In any case, court action is likely only when lucrative holdings are at stake and only after nonlegal alternatives have been exhausted by the disgruntled party. The pre-court strategy in the latter case parallels that of the former, depending upon manipulation of public opinion.

Boundary disputes differ socially from the previous categories in that they generally take place between nonrelatives. While the original boundaries between ho'aga were clearly marked, in the process of fragmentation that took place following European contact, new boundary markers had to be established. Large trees, rocks, and other features of the natural environment were used. Most of these boundaries were never recorded in writing but simply represented agreements reached between individuals at the time of division. This left the door open for expedient persons to try their luck at expanding their territories, and there can be little doubt that many took advantage of the opportunity. Not all boundary disputes, however, resulted from calculated opportunism. Genuine misunderstandings arose as a consequence of poorly defined markings and lack of knowledge. The marked tendency toward increasing boundary disputes (until 1950), when contrasted with the relative consistency in the frequency of stewardship disputes, can probably be accounted for by the fact that knowledge of boundaries has been handed down by word of mouth and that with each new generation the probability of misunderstanding and faulty knowledge has increased.

The District Officer's Court constitutes a distinct activity system. Within this system the District Officer, as Magistrate, is the supreme decision-maker. He may or may not use advisers, but the responsibility for arbitrating disputes is entirely his. Rotuma's isolation from Fiji's main administrative centers and the inadequacy of the ideal model for resolving actual disputes have inevitably left Land Court decisions up to the personal discretion of each District Officer, and each official has evolved his own model for deciding disputes, often combining Western and Rotuman cultural principles.

Among the main principles used in resolving disputes are the following: (1) the right of the pure to allocate and dispose of land; (2) the rights of kainaga members in kainaga land; (3) the rights of an occupant; (4) persuasiveness of factual arguments put forth by litigants; and (5) evaluation of the integrity of litigants.

All five of these principles generally enter into the resolution of stewardship disputes (plus any principles from European culture the District Officer feels to be applicable). They obviously do not permit impersonal decisions. The first two principles, in fact, are in conflict with one another. The Rotumans will argue their case on either grounds, and each District Officer must decide which principle to emphasize. There has been considerable alternation historically, some officials weighting the rights of the pure quite heavily relative to the rights of kainaga members and others doing the reverse. Often this was conceived in terms of individual versus communal ownership, and the particular official gave precedence to the type he personally favored. According to the Rotumans, the third principle is irrelevant; if a person with priority enters a claim, it makes no difference that the occupant has spent many years cultivating, and perhaps has improved, the land. The person with priority should be pure. In settling actual disputes, however, there is little doubt that European officials have given considerable weight to occupancy, particularly when the value of the respective claims was in doubt.

The fourth principle is more complicated than it might seem on first consideration. Since most of the officials responsible for arbitrating disputes have been learned neither in Rotuman custom nor in the Rotuman language (appointments have generally been for only a couple of years), they have had to rely quite heavily on advisers. In most cases, of course, the factual information was not subject to verification, and the personal influence of the advisers was great. From the standpoint of the litigants, then, strategy called for influencing the advisers so that the supporting information would be interpreted in a light favorable to themselves. Kinship and personal obligations were frequently brought into play, and thinly disguised attempts at bribery have been common. There are several cases on record of advisers being dismissed when District Officers have discovered what had been going on, and some have preferred to make their own way without any.

The fifth principle is, of course, a completely personal one, but it is evident that it is sometimes the most important. In such a confusing game of strategy and manipulation, often the only clear impression the official gets is that one of the litigants is a nice, friendly, unabusive person (or so he appears), while his opponent seems aggressive and abusive (or at least is simply less of a nice person). I do not believe that officials consciously used this principle in their arbitrations, but their summations make it clear that in not a few cases this factor was decisive.

In usufruct and boundary disputes only the third, fourth, and fifth principles are likely to be brought into play, but otherwise the strategy is the same. As a result of the indefiniteness of these principles there has been little historical consistency in the official decision-making model--a fact which helps to account for the substantial number of reversed decisions that are found in the Land Court records (cf. Howard 1962: 145-157).

It still remains to explain the marked decrease in land disputes (and transactions) during the 19501960 period. Interestingly enough, during this period the District Officer, with the exception of some intermediary periods, has been a Rotuman. In 1945 the first Rotuman was appointed, and during his four years in office he heard some 27 disputes, or an average of 6.75 cases per year. Between 1936 and 1944, under European District Officers, the average was 22.9 cases per year. In 1953 a second Rotuman took over and during his first year heard three disputes. In 1954 he heard one more case, and has not convened the Land Court since. Since 1954 some fifteen disputes (thirteen concerning boundaries and two concerning stewardship) have arisen, but none had been dealt with by 1960. The existence of most of these disputes, in fact, was discovered during an inquiry by a visiting Commissioner. The District Officer professed knowledge of only five of them.

The reduction in land cases heard before these Rotuman District Officers is probably dependent on several factors. It might be of some significance, for example, that their personal and intimate knowledge of Rotuman affairs inhibits many persons who might otherwise be willing to try their luck with spurious claims. Perhaps, too, they are less prone than European officials to think of land tenure as consisting of a set of abstract principles and are more prudently aware of the arbitrariness with which decisions must be made. Lacking the protection of impersonal law, they are reluctant to hear cases. But the most significant factor leading to an avoidance of land matters has undoubtedly been the fact that these men themselves are involved in a network of kinship relations and cannot dissociate themselves from obligations with ease. They are therefore placed in a situation of role conflict. If, in a dispute involving one of their relatives with a nonrelative, the decision were to be made in favor of the former, the people would accuse him of favoring his own kinsmen. If, on the other hand, the decision were to be made in favor of the nonrelative, his kinsmen would regard this as a breach of kinship obligations. A situation involving conflict between two relatives also involves role conflict.

The dynamics of kinship maximizes the possibility that these conflicts will occur. In Rotuma it is a maxim that one who is wealthy or influential has many relatives, but one who is poor and insignificant has few; and the bilateral emphasis of filiation permits almost unlimited possibilities for claiming relationship where it is expedient. It can be expected, therefore, that when a Rotuman is appointed District Officer, his "active" kainaga increases many times over. The process of establishing active relationship, called by the Rotumans re kainaga (lit., to make into a relative), simply involves the giving of a gift or showing some special consideration which is nearly impossible by custom to refuse without giving offense, accompanied by a claim of kinship affiliation. Naturally reception of such a gift entails corresponding obligations.

The official's situation is complicated still further by the fact that most Rotumans consider decision-making to be a personal business, rather than attributing the process to the impersonal implementation of law. When a European official made an adverse decision, a losing litigant would ordinarily shrug his shoulders and attribute it to the arbitrator's ignorance of Rotuman ways, but when a Rotuman makes the decision it is tantamount to a proclamation of personal animosity. There can be no doubt that the prestige of a Rotuman District Officer is likely to decline in direct proportion to the number of decisions he has to make under such circumstances.

The first Rotuman to occupy the position was, to all appearances, a conscientious man who, finding the role of arbitrator insufferable, chose to relinquish his position in favor of a lesser job in Fiji. The current District Officer has evidently attempted to reduce conflict by avoiding cases involving civil litigation between parties. The decrease in the number of land transactions during the same period, manifest in Table 1, can be interpreted as part of the general reluctance of these men to handle land matters of any kind.

In passing, it is worth noting that from the "holistic" point of view administrators and administrative procedures create a problem as to whether they should be included in or excluded from the system. Within the present frame of reference no such problem exists.


What can we say now about the Rotuman kainaga? Is it a cognatic descent group or a corporate group in any sense? If we were to use the ideal model as our guide, it is clear that the concept of the kainaga as a landholding unit fits the definition of a "sept" (i.e., the nonunilinear equivalent of a sib) by Davenport (1959: 562) or the more precise definition of a "sept" by Ember (1962: 966) as "a dispersed descent group which has only one clan segment" (i.e., land in only one locality). If we were to adhere to this view, and consider Rotuman social organization in terms of a single social system, it would be possible to conceive of Rotuman society as composed of a number of "septs" (i.e., kainaga). But such a model would help us very little in understanding or explaining Rotuman behavior. As far as reality is concerned, it would appear to be little more than an elegant illusion.

If we confine ourselves to analyzing activity systems and the composition of actual groups, however, the questions become irrelevant. We can then take an operational point of view and simply ask ourselves what are the principles that determine group composition when given activities are being carried out. Instead of beginning with a definition and trying to fit indigenous (ideal) group concepts into it, we can then proceed empirically and derive our analytical concepts on the basis of the principles that determine group membership, as Goodenough has suggested. After all, groups do not exist per se. They come into existence during the performance of activities and dissolve when the activities are completed. They are far more fluid than our traditional concepts seem to imply.

As a concept, the Rotuman term kainaga is better understood as a cultural principle, used in variety of situations by individuals as a means of legitimizing their participation in certain activities, e.g., choosing a pure, rather than as a group of one kind or another. As should be clear from my discussion of the data, other principles must be taken into account to explain the composition of actual groups. When viewed in this perspective there is no contradiction between the kainaga as a landholding unit and the kainaga as a kindred. One need not even regard them as homonyms. They are simply two applications of the principles of filiation, as manifested in the performance of different activities. In each case the resultant composition of actual groups is modified by the other principles brought into play within the relevant activity system.

Let us consider further the relationship between the ideal and actual (i.e., predictive) decision-making models. It should be clear from the evidence cited that in Rotuma the ideal model for land tenure has a differential predictive value in different circumstances. Where usufruct is concerned it is of almost no value, whereas in cases of succession following the death of a pure of kainaga land it is moderately predictive. What, then, is the significance of the ideal model to the Rotumans? To answer this question we must consider it from two points of view.

In the first place, the model does provide a general guide to action, but with the implicit qualification of "all other things being equal." This is true of any ideal model, of course, but the degree to which a given model accounts for potential influences can vary considerably. In some societies the formal rules covering a given activity system are comprehensive, and hence the models are highly predictive of actual decisions. In other instances this is not the case. Reasons for the discrepancies can be varied. In Rotuma, for example, it appears that complications were introduced by acculturation with European society and that this had the effect of dramatically altering actual decision-making procedures while only slightly affecting the ideal model. Reasons may also be sought in social psychology. The Rotumans give every indication of being made uncomfortable by tightly prescribed rules in any behavioral area. They seem distinctly to enjoy the measure of personal freedom for manipulation which their social organization affords them, and they have ardently resisted attempts by the Administration to tighten the rules. Undoubtedly, other explanations could be found for other cases.

Second, the ideal model provides a set of legitimizing principles that people employ in defending their claims. As such it is not a model for making decisions at all, since it is called into play only after decisions have been made. To be sure, other principles than those incorporated in the ideal model are used to substantiate claims, particularly in informal disputes, but a claim in accordance with the ideal model carries greater weight, and in a court case it is likely to be decisive, all other considerations being equal.

What then can we say about jural rules in Rotuma? Are they determinants of behavior or merely the outcome of multiple individual choices as Murdock suggests? Upon careful examination the question turns out to be a false one, for there are clearly two separate issues involved. The first concerns the principles actually employed in making decisions. As we have seen, in Rotuma these principles may or may not have jural implications, i.e., they may or may not be contained in the ideal model. Second, there is the question of sanctions. It seems to me that the only plausible test to determine which principles are functionally jural is to observe what happens when they are ignored or contradicted. From this standpoint, the only aspect of "juralness" that is significant is the expectation of socially induced punishment or reward, should one or other mode of behavior be adopted. So conceived, it is possible to consider the anticipation of consequences as an additional cultural principle which an actor takes into account when choosing between alternative forms of behavior. When seen in this perspective it becomes impossible to consider jural rules as only "causes" or only "effects." They are determinants to the extent that they are incorporated into actual decision-making models; they are effects to the extent that individuals within the society sanction them.

Methodologically, we must begin to analyze social behavior, as Murdock suggests, with statistical analyses of individual choices. But if we are to presume, as I think most social anthropologists would, that social behavior is systematically structured, we cannot be satisfied with a mere statistical count. Rather the statistics are a means by which we can test the validity of certain principles, but eventually we must be able to derive mechanical models that approach a predictive value of 100 per cent, in a way comparable to the manner in which a competent structural analysis of a language allows us to predict the forms utterances may take. But for reasons already stated, we shall never be able to approach this goal as long as we regard societies as uni-structural models, i.e., as systems.

In this paper I have continually spoken of "choices" and "decisions" as though they were synonymous with "behavior," a usage which requires further explanation. As a beginning qualification, it is obvious that social anthropologists must place some limits on what they are attempting to predict, and comprehend, lest they become involved in individual psychology or even physiology. I think most social anthropologists would agree that the "behavior" they are interested in is limited to social behavior, i.e., behavior that has significance for others. The question then becomes, how do we decide what is significant for a given group? To answer this there is simply no alternative to an intensive analysis of the cognitive world of our subjects. We must learn how they categorize behavior, how they distinguish types of behavior from another. It may be that a whole range of behavioral acts, which to the observer might appear quite distinct, are to our subjects only insignificant variations within a distinct category, as phonetic variations are to a phoneme. In describing residence patterns, for example, we must discover whether the choice of residence at marriage in the groom's father's home is considered the same or different from the choice of the groom's mother's home; or whether building a house next door is considered the same as or different from moving into the same house. It is methodologically unwarranted to assume the validity of a distinction between patri-virilocal and matri-virilocal residence, or to equate moving into the same household with building next door. If we want to predict residential choices, we must learn what our subjects conceive the alternatives to be and then attempt to derive the principles upon which they base their choices.

It may turn out that kinship considerations are only partially predictive, and that other principles are required to construct a satisfactory decision-making model. But only if our models accurately reflect the cognitive worlds of our subjects can we hope to reach the ideal of ethnographic description so elegantly put forth by Goodenough. For this reason I regard semantic studies of kinship terminology and research on folk taxonomy of major importance to social anthropologists. In this soil genuinely systematic analysis is germinating.

From this discussion it should now be reasonably clear what I mean by a "choice" or "decision." I am referring to the selection (consciously or unconsciously) of a mode of action which can be considered, on the basis of one's evidence, as cognitively distinct from another mode of action. This may involve distinguishing categories of persons and/or types of behavior. In Rotuma, for example, I have inferred that "near relatives" are distinguished from others, and that this distinction is of value in predicting choices when it comes to asking for copra-cutting rights. It was also intimated that requests accompanied by humility were distinguished from demands, and that this, too, had a predictive value.

Lest the reader gain the wrong impression, let me make it clear that I am not advocating a fragmental approach to society, nor suggesting that an activity system can be understood apart from other features in the society. In practice it is always necessary to explore behavior in many activity systems in order to understand any one of them, as every field worker knows. I am only suggesting that the degree of interrelationship is an empirical problem that must be studied separately for each case. To assume functional unity is philosophy, not science.

Let us turn now to the problem of explanation. The dilemma here has been posed as one of socio-cultural explanation versus psychological explanation. But this issue, too, seems to dissolve once we abandon the notion of societies as whole systems and turn instead to explaining social behavior. As Lounsbury (1962: 1302-1310) has cogently pointed out in his discussion of the cross-cousin controversy, certain distinctions between types of cross-cousin marriage have been ignored. Where cross-cousin marriage is used as a mechanism for maintaining political association between groups it may occur statistically infrequently, though incorporated into the formal ideology. In other cases it may constitute an ideal model for individuals in choosing their mates. I would argue that here we have two different kinds of activity system, one political and the other domestic, each responding to different influences and each requiring its own explanation.

To be credible, scientific explanation of recurrent events must be testable, which is another way of saying that it must apply to operationally definable phenomena. In our science this limits us to explaining social behavior. But before discussing the relative merits of sociocultural and psychological explanation, it is necessary that we define these terms as they apply within the current conceptual framework. For our purposes, the most plausible distinction that can be made between the concepts "socio-cultural" [13] and "psychological" is to regard the former as pertaining to those decision-making principles which are cognitively shared and overtly agreed upon by the members of the group under observation, and to regard the latter as pertaining to those decision-making principles which are covert but statistically shared as the result of comparable socialization experiences. Operationally, of course, these definitions are matters of degree rather than absolute. Given this distinction, socio-cultural explanation would then lie in suggesting mechanisms that produce particular socio-cultural principles; psychological explanation would lie in suggesting mechanisms that produce particular psychological principles. Both types of explanation would involve an exploration of ecological or external (to the group; to the individual) influences and an examination of the internal relationship (set) of the principles with one another. Thus an adequate explanation of behavior within an activity system would include several facets, including an account of the origins of the stimuli to the activity, an accounting for the cognitive distinctions employed, and devices to explain why certain decision-making principles are favored over alternatives. Conceived in this way, the necessity of choosing between socio-cultural and psychological explanation is eliminated, for both of them, plus historical explanation, would almost certainly be required in the majority of cases.

Leaving aside diachronic explanation, the degree to which socio-cultural or psychological mechanisms can satisfactorily explain a particular case would seem to be an empirical rather than a theoretical issue. It boils down to what one must know to predict behavior accurately. In forming groups to perform particular activities, for instance, it may be possible in some societies to predict with nearly 100 per cent accuracy who will be included on the basis of kinship alone. In other societies there may be greater room for individual choice, and psychological factors may influence decisions. I am suggesting that socio-cultural principles may be thought of as setting the limits of acceptability, but that the degree of freedom permitted an individual (i.e., the number of alternative choices possible within those limits) differs from society to society, and from activity system to activity system within each society. [14] To the degree that choices are limited by socio-cultural principles one would expect socio-cultural explanation to suffice; conversely, to the degree that they permit choice, one would expect psychological principles to be required in addition.

Having said this, let us turn once more to the question of typology. If my arguments are granted, it follows that any attempt to construct a typology of complete societies, or whole "social systems," is futile. For a typology to be useful it must be based on a scheme of classification that reflects determinate mechanisms. This requires units that are genuine systems, which whole societies are not. I would therefore suggest that we aim at a typology of activity systems, in which the criterion of classification would be based upon the structural and cultural principles that are predictive of actual behavior.


In the introduction of this paper I made the assertion that in their concern for "cognatic" kinship organization social anthropologists are facing the basic issues of their science. In concluding I shall attempt to evaluate the significance of this concern in its historical perspective.

If one examines the history of science, it is not difficult to discern a certain cyclical aspect to the relative emphasis on abstract theory and data-gathering empiricism. Modern social anthropology, for example, has its roots in the crude data collection following the Age of Discovery, to which little scientific theory was applied. Evolutionism was social anthropology's first genuine theoretical framework, and within its conceptual structure the data, or at least a large portion thereof, were neatly ordered. The prominent theorists of the period, e.g., Morgan and Tylor, spent their greatest intellectual energies elaborating the conceptual tools of evolutionary theory, but in so doing they became further and further removed from the data they were dealing with. Empiricism gave way to conceptual elaboration, and incongruities were either ignored or explained away as "survivals."

Then, toward the end of the nineteenth century, came Boas, and a new period of empiricism was initiated. Boas pointed out both logical flaws in the evolutionary argument and the inconsistency of the data. By presenting cases which did not fit the evolutionary scheme he demolished with empirical evidence the theoretical foundations so elaborately constructed by his predecessors. His insistence was that we needed better data if we were to construct valid theories.

Although the historical school which Boas founded never developed a systematic theory of culture, the emphasis nevertheless shifted as time went on away from sheer empiricism toward conceptual elaboration. Perhaps the climax was reached in Kroeber's culture element list with its tremendously detailed scheme of classification.

The focus shifted again in the 1920s, stimulated by Malinowski. Although he did not base his call for empirical data on an assault against historicism as Boas had done with evolutionism, Malinowski (1922: xvixvii) nevertheless decried the lack of concern that had been shown for "dying" cultures and registered a plea that they be studied in their entirety while there was still time. He personally engineered, as a teacher, a wave of such studies, particularly in Africa. At about the same time, of course, Radcliffe-Brown was laying the theoretical foundations of functionalism by applying Durkheimian logic to anthropological data. It is significant that of the well-documented societies, which were required for demonstrating the functional model, most were in Africa where Malinowski's students had ventured, and also, significantly, that almost all of these emphasized unilinear kinship principles. The crux of my argument is that this early preoccupation with unilinear descent has tended to mask the difficulties inherent in the functional model and that the recent concern for basic issues has been stimulated by attempts to fit cognatic societies into the functional framework. To do this some theorists have attempted to elaborate the basic model conceptually, while others would simply treat the exceptions as special cases. It is symptomatic, perhaps, that the arguments contained in the literature have begun to sound more and more like theology, and less and less like science. What I am urging, in effect, is that we formulate our problems so that they can be resolved by empirical data, that we define our concepts operationally, and that we spend our efforts in acquiring the necessary information. It can do no good to wail over defunct societies, documented by an earlier generation of field workers with other aims, leaving us with a legacy of information too incomplete for our own purposes. To argue over what the facts really were seems futile indeed. Every science loses a portion of its data irretrievably through time, but there is no more reason for scientifically oriented anthropologists to deplore "lost" societies than for astronomers to deplore lost cosmic explosions. If the principles that govern human behavior are, as most of us presume, constant and immutable, we should be able to unravel them as long as there are human actors behaving in diverse circumstances. In any case, however, it would be far better to base our judgments on a handful of adequately documented societies than to play guessing games with five hundred.

In closing, I would like to offer this final qualification: I do not regard the argument contained in this paper as being new. Theoretically, I have merely tried to draw to their conclusions logical premises already contained in much of the work being done by contemporary ethnologists. Neither do I consider the description of Rotuman land tenure presented herewith as dramatically different from other ethnographic accounts. Many ethnographies do, in fact, offer descriptive accounts which permit one to construct decision making models for at least some activities. I am merely suggesting that we formulate our issues so that they are resolvable, and produce more of them. Only then will we be able to build our generalizations on a solid foundation and perhaps stop spinning our wheels in the sand.


[1] Field work was carried out among the Rotumans between October, 1959, and June, 1961, of which twelve months were spent on the island of Rotuma and nine months among the Rotuman colonies in Fiji. The field work was supported by the National Institute of Mental Health. I would like to acknowledge the assistance in the field of Irwin Howard, whose contributions toward the collection and analysis of field data were considerable. This paper was written and accepted for publication while I was a temporary lecturer in social anthropology at the University of Auckland. I have since accepted the position of cultural anthropologist at the Bernice P. Bishop Museum in Honolulu. [back to text]

[2] For a discussion of statistical and mechanical models cf. Lévi-Strauss 1953: 528-531. [back to text]

[3] It is the custom for sleeping houses to be built on raised platforms of stone and earth. [back to text]

[4] I was never able to get a clear statement as to which of one's great-grandparents, eight fuag ri one should choose, apparently because rights were never challenged on this basis. When pressed, informants, generally agreed it would be most appropriate to select the residential fuag ri on which the great-grandparent was raised. [back to text]

[5] For a discussion of the leadership aspects of chieftainship see Howard 1963. [back to text]

[6] Since the ceremony takes place at one of Ego's eight ancestral fuag ri, only seven groups must make the trip when ideal conditions are met. The kainaga from the fuag ri where the ceremony is being held gathers beforehand and acts as the host group. [back to text]

[7] In actual practice land held in common by a sibling group is often referred to as hanua ne togi or hanua ne na, signifying the method of acquisition by the parent and emphasizing the exclusiveness of rights in the parental generation. Most informants agreed, however, that such land is properly designated as hanua ne haisasigi. [back to text]

[8] This generally occurs in the fourth or fifth generations. [back to text]

[9] In 1959 the Government of Fiji drew up an ordinance to deal with Rotuman land tenure. Since this was done after consultation with Rotumans the definitions employed in the Bill are worth noting. Actually the seven types of tenure I have described are here reduced to three:

"hanua ne kainaga" means land held by that family community of Rotumans known as a kainaga, the members of each kainaga holding the land in undivided ownership and the acknowledged head of the family being the pure (or overlord) of the land;

"hanua pau" means land which is vested in a single individual Rotuman by sale or gift with the intention of creating hanua pau, or by an instrument deposited with the District Officer...

"hanua ne 'on tore" means land which is vested on intestacy in the first, second, and third generations of descendants of a deceased owner of hanua pau, as hanua ne 'on tore, when there is no single individual Rotuman in whom the land vests as hanua pau, such descendants taking life interests in undivided shares in such land, and the last survivor of them taking the land as hanua pau:

Provided that the limitation to three generations shall not apply to hanua ne 'on tore in existence at the commencement of this Ordinance... [back to text]

[10] The word faksoro also means to apologize, or to ask to be excused. Cf. Churchward 1940:195. [back to text]

[11] By "license" I mean the exercise of personal will as opposed to restraint, in this case concerning the disposition of property. Used in this sense "license" may be distinguished from "rights" in that it is meant to be purely descriptive of behavior, without jural implications. [back to text]

[12] Land holdings in Rotuma tend to be dispersed and consist of irregularly shaped blocks of land of varying sizes. Some pieces are no larger than a hundred square feet or so, while others may be several acres. In most cases a man works three or four plots at a time, each of these in a separate location. The more distant ones may require an hour's walk or more through the bush, whereas the nearest may be only a few yards from his dwelling. [back to text]

[13] To remain consistent, the term "socio-cultural" should here be construed as including "structural" and "cultural" principles. [back to text]

[14] For an excellent illustration of how group composition can vary from activity system to activity system within a single society see Groves 1963. [back to text]


Capell, A. 1962. Oceanic Linguistics Today. Current Anthropology 3: 371-428.

Churchward, C. M. 1940. Rotuman Grammar and Dictionary. Sydney.

Davenport, W. C. 1959. Nonunilinear Descent and Descent Groups. American Anthropologist 61: 557-572.

Ember, M. 1962. Political Authority and the Structure of Kinship in Aboriginal Samoa. American Anthropologist 64: 964-971.

Frake, C. 1961. The Diagnosis of Disease Among the Subanun of Mindinao. American Anthropologist 63:113-132.

----1962. Cultural Ecology and Ethnography. American Anthropologist 64:53-59.

Gardiner, J. S. 1898. Natives of Rotuma. Journal of the Royal Anthropological Institute 27:396-435, 457-524.

Goodenough, W. H. 1951. Property, Kin, and Community on Truk. Yale University Publications in Anthropology 46:1-192.

---- 1957. Cultural Anthropology and Linguistics. Report of the Seventh Annual Round Table Meeting on Linguistics and Language Study, ed. P. L. Garvin. Monograph Series on Languages and Linguistics 9:167-173. Washington.

----1961. Review of Social Structure in Southeast Asia, ed. G. P. Murdock. American Anthropologist 63:1341-1347.

Grace, G. 1959. The Position of the Polynesian Languages Within the Austronesian (Malayo-Polynesian) Language Family. International Journal of American Linguistics 25: iii, Supplement.

Groves, M. 1963. Western Motu Descent Groups. Ethnology 2:15-30.

Howard, A. 1961. Rotuma as a Hinterland Community.Journal of the Polynesian Society 70: 272-299.

---- 1962. Change and Stability in the Rotuman System of Land Tenure. Unpublished doctoral dissertation. Stanford University.

---- 1963a. An Annotated Bibliography of Rotuman Materials (mimeographed).

---- 1963b. Conservatism and Non-Traditional Leadership in Rotuma.Journal of the Polynesian Society 72 (in press).

Leach, E. R. 1960. The Sinhalese of the Dry Zone of Northern Ceylon. Viking Fund Publications in Anthropology 29:116-126.

Lévi-Strauss C. 1953. Social Structure. Anthropology Today, ed. A. L. Kroeber pp. 524-553 Chicago.

Lounsbury F. 1962. Review of R. Needham, Structure and Sentiment. American Anthropologist 64: 1302-1310.

Malinowski, B. 1922. Argonauts of the Western Pacific. London.

Minutes the Rotuman Council of Chiefs. Retained in the Central Archives of Fiji and the Western Pacific High Commission, Suva.

Murdock, G. P. 1949. Social Structure. New York.

---- 1960. Cognatic Forms of Social Organization. Viking Fund Publications in Anthropology 29: 114.

Nadel, S. F. 1957. The Theory of Social Structure. Glencoe.

Ordinance No. 13 of 1959. An Ordinance to Provide for a Land Commission in Rotuma, the Registration and Dealing with and Transmission of Land and Matters Incidental Thereto. Fiji.

Radcliffe-Brown, A. R. 1952. Structure and Function in Primitive Society. London.

Sahlins, M. 1963. Review of Social Structure in Southeast Asia, ed. G. P. Murdock. Journal of the Polynesian Society 72:39-50.

Home Page
Howard-Rensel Papers