Volume 66 1957 > Volume 66, No. 1 > Cultural evolution or cultural change - The case of Polynesia, by H. B. Hawthorn, p 18-35
                                                                                             Previous | Next   

- 18

THE STUDY OF cultural evolution has always been a major occupation of archaeologists and prehistorians, but among cultural anthropologists it has had peaks and troughs of fashionable attraction. There has been an increasing pre-occupation with this topic among cultural anthropologists of recent years; to such an extent that many of the earlier warnings about the difficulties of a sociological treatment of evolution are going unheeded. The purpose of this paper is to re-state these warnings in the context of a critical evaluation of some examples of work in Polynesia, and to show that their relevance is ever with us and cannot be dismissed lightly.

We take the study of cultural or social evolution to mean the examination of the ways in which contemporary and prehistoric cultures or societies emerge from previously existing and radically different ones. Characteristics must be compared temporally, and must differ in major and definable ways so that one may speak of one type of society giving place to another type. The differences may have reference to such characteristics as structure, technology, and specific inventions such as writing. The study of evolution, in the large, is concerned with piecing together the historical examples in such a way that a grand scheme for the development of human society is revealed, based upon universal cultural laws.

As against the evolutionary approach we would contrast that of studies in cultural change (social change, acculturation) which have many of the same ultimate objectives, but which proceed along quite different lines. Such studies are concerned primarily with an exposition of change built upon small-scale examination of observed or historically recorded developments. These usually cover too small a time-span to represent a radical change of the whole society, but where radical change has occurred, the aim is to demonstrate the historical sequence in such a way that each important event in the sequence is adequately known. Analysis of cultural change based on the historical descriptions is achieved by reference to the functions and disfunctions of the society. The theoretical aim of culture change studies is to build up a corpus of sociological laws, which will be of the same order as and closely related to laws arrived at in sociology or in anthropology by the analysis of contemporary functioning societies.

In this paper we suggest (1) that the methods and aims of the study of cultural evolution set traps for the scholar which are almost impossible to avoid (and which have not been avoided in recent studies), - 19 (2) that the study of culture change is more likely to yield acceptable generalisations than the study of evolution, and (3) that the latter cannot proceed without a basis of knowledge laid by the former. We suggest that there may not be a pattern of evolution in the historical sense, but that there may be laws of human behaviour of a more abstract kind. And finally we point out that, since the hoped for laws will be derived from the study of societies and cultures as abstractions, even the study of culture change must be prepared to arrive at models which will be independent of actual or historical points of reference.

Because of the limitations of length in a paper such as this, these issues will be raised in an exploratory way only. The need for such restatements and explorations arises from the frequent misuse of method by anthropologists today, particularly in relation to evolutionary studies. We begin by demonstrating this, for the major, over-all point to our paper lies in our conviction that anthropology, of any kind, which is written in an innocence of scientific method will almost certainly be bad anthropology.


We take as our text for criticism and evaluation a recent study of Polynesia which attempts to discover a widespread and repeated sequence of change. In this study Irving Goldman (1955) simplifies and extends the explanation offered by Edwin Burrows (1939:21) that one of several causes of change in political and kin alignment in Polynesia was “warfare arising from rivalry over land or ambition for enhanced status.” Goldman postulates that “fundamental changes in the Polynesian systems of status appear to have produced major modifications in many basic Polynesian institutions and cultural values.”

As Goldenweiser (1937) pointed out, there is an inherent difficulty in the demonstration of evolutionary hypotheses from observations of contemporary cultures. The circularity of argument which Goldenweiser sees as the greatest and final obstacle need not, of course, always be vicious. Granted that the evolutionist must assume that differences in contemporary societies express past evolutionary changes, and that he cites his ability to arrange the societies according to these differences as part of a proof of the evolutionary process, the analysis need not be confined to this, nor is it always compelled to operate with this circularity. Archaeological sequences, accounts of cultural changes from oral history, or the analysis of processes in contemporary life, may be invoked to supply data for checking the sequences postulated in an evolutionary reconstruction. Finally, the constructed model may meet the requirements of fitting the known facts from the pertinent ethnographic field and the pertinent sciences. The circularity of its beginnings may in these various ways be overcome.

Yet this is not easy to do; and other difficulties which face evolutionary reconstruction become apparent in actual instances.

First, the demonstration of cultural change cannot stand abbreviation. The most general statements of cultural change are either truisms or so challenged by exceptions that their point is lost. Perhaps the - 20 terminology of evolutionary studies as contrasted with cultural change studies reflects also a division on this matter of abbreviation and exactitude.

Secondly, evolutionary sequences are generally taken to mean a progression between states that are arranged in a rank order or scale. Institutions and values which have any complexity are inevitably hard to classify and to arrange in rank order. One may arbitrarily list them in orders of intensity, complexity or whatever, but it is a non sequitur for the essentially static scheme to be mistaken for a proof of sequence. Thus we are told by Goldman that “in these illustrative injunctions to rulers we see the emergence into explicit form of an ethical code . . .” but since we are unrewarded with any indication that anything is really known about an emergence, Goldenweiser's stricture is fully pertinent.

Again, if comparable cultural elements can, even though arbitrarily, be placed in a rank order, what can be inferred from this as to cultures as wholes? Perfectly homogeneous cultures could, ignoring their other elements, be placed in the list so as to occupy corresponding positions. What happens with a culture which is not uniform, one which has multiple patterns relating, for example, to different classes? Obviously distortion would result if it were to be ranked in an evolutionary scheme on the basis of one of its patterns, ignoring alternatives and variants.

Other queries arise in relation to the use of historical data. Boas (1896) suggests that analysis be confined to the history of linked societies as a first step in evolutionary reconstruction to safeguard against the uncritical interpretation of superficial similarities, a safeguard which is hardly as necessary today as it was sixty years ago. But several other difficulties can beset the follower of Boas' suggestion. There is still the barrier that isolated data from different periods or different cultures do not necessarily provide material for an analysis of process. (“In the Marquesas . . . some chiefs had begun to usurp land which they had held in custody for the tribe. In Mangareva this process had gone to the point of complete expropriation.” (Goldman, 1955:689). Is this one process?) And the obverse of this is the objection that hypotheses are insufficiently general if they can be applied only in like societies: they must be applicable in unlike societies to qualify for acceptance as general theory about the manner of culture change.

It was the claim of Boas that evolutionary studies, which began within one culturally linked region, could be broadened into general theory, when enough regions were covered. “Thus by comparing histories of growth general laws may be found.” (Boas, 1896:279). The success of this procedure depends on formulating hypotheses which are of general application but which permit of adequate testing; otherwise it cannot lead to temporal prediction, or go beyond the explanation of special events, such as those of Polynesian prehistory.

Burrows' article, cited above, confines itself to historical generalisation and interpretation. It deals with the differentiation of the relationships between government and kinship. “In some parts of Polynesia,” he wrote, “the tracts governed as political units were populated by - 21 groups also regarded as kinsfolk, whether by blood, marriage, adoption or rationalization. In other parts the higher chiefs did not regard all their subjects as kinsfolk, nor were the common people in a territorial-political units necessarily related among themselves.” He makes the case, supported by ethnographic statement, oral history and some linguistic evidence, that some regions had coincidence of kin and political units and others had an intermingling; and that “progressive encroachment of border over breed seems to have been the rule in Polynesia. As territorial units grew larger and stronger, kinship grouping became simpler and vaguer; for in both areas of intermingled breed and border, complex ramified kinship grouping was either absent, or the larger groups were vague in conception and limited in function. Several processes that favoured change in this direction have been suggested: inter-marriage, adoption, migration, and—perhaps most powerful of all—warfare arising from rivalry over land or ambition for enhanced status.”

The causality suggested here is fairly complex. Burrows offers also the comments of Buck and Emory, each stressing different geographical factors in different regions. For Mangaia and Mangareva, Buck adds exogamy. Some loose ends and contradictions remain; on the one hand the meagre resources of atolls, on the other the rich resources of New Zealand, are each held to offer little incentive to change.

Yet the analysis is a broadly acceptable account of differentiations in what might earlier have been a homogeneous Polynesian culture. In order to form a basis for a major contribution to general cultural change theory, the family and governmental processes would need to be studied in more detail. The contradiction above could probably be harmonized quite readily; it would not be as easy to fit in Mead's summary of the conflict between kin and governmental institutions in Manua:

“Relationship allegiances undermine the growth of village authority, mock at village pride, contravene the whole village system . . . But the household, like the village, is a local unit . . . Every member of the household is knit up with the social and industrial life of the village. For the behaviour of all, the matai is responsible to the fono. In consequence, the fono is only too happy to see natural ability over-ride blood claims, and the residence unit assert itself against the blood units.” (Burrows, 1939:20).

There we see the stress on quite different values: appreciation of natural ability and of efficiency, and local pride. Mead's statement forces the recognition that the link between kin and political bonds is not a unitary one; we are reminded that in most societies authority is expressed through different combinations of the two; and we conclude that while Burrows' reconstruction of Polynesian prehistory is a guide for more detailed work based on a closer analysis of kinship and government, the unanswered questions as to how a consideration of the points raised by Mead would affect the reconstruction, make it unsatisfactory as a foundation for a broader study of cultural change in Polynesia.

In order to proceed toward the goal of general theory, the student of cultural change finds it necessary to formulate hypotheses which - 22 relate to action in general. Insofar as Goldman does this, he selects one of the several motives inferred by Burrows; this one, “ambition for enhanced status,” emerges in Goldman's work as a fundamental motive in Polynesia. He thus makes causality much simpler, and bases a number of hypotheses on the assumption that “Rivalry raises issues and provokes conflicts that can never be fully resolved. It promotes a sequence of culture changes . . .”

This seems to be a poor starting-point. Thus if we try a substitute statement that is almost the reverse: “The aristocratic doctrine of hereditary rank provides a resolution for the issues and conflicts of rivalry in Polynesia. Thus rivalry had little to do with promoting cultural changes . . .” we have a statement which has plausibility, perhaps almost as much as the original. Yet, like the original, its vagueness of concept and its ambitious scope place it beyond clear proof or disproof. Perhaps a better hypothesis on the relation of status and conflict in Polynesia could be derived from the statements of the balance between co-operative and competitive pressures framed by Mead and Mishkin for Samoa and Maori. (Mead, 1937).

An example of interpretation of the processes operative in one Polynesian society is supplied by Buck for Mangaia (1934). Buck presents a fairly detailed reconstruction of some aspects of Mangaian culture. Mangaia is a small volcanic island, with several immigrant groups. The culture is characterized by some land coveting and considerable rivalry, though it appears in terms of stark power and survival rather than status. There is a relationship of religion to this rivalry; in part, religion is subordinated, though the relationship is complicated. This study is one of the outstanding examples of historical reconstruction in Polynesian anthropology. It is a classic attempt to use oral history and inference as the basis for a study of cultural change. If we grant the validity of inferences from oral history, we may believe it capable of yielding conclusions of a general nature. Yet its use is by no means foolproof. As Beaglehole writes below, it may not have been in itself intended as a balanced work. Warfare is probably exaggerated; certainly its role in Mangaian cultural processes is sharply highlighted. The theses Buck has propounded in relation to Mangaian history cannot readily be extended to all Polynesia.


Some of the requirements of method we consider necessary for the general study of cultural change apply more strongly than others to the problems of this topic in Polynesia. Those of greatest pertinence are: (1) proper and balanced selection of societies and close definition of social units and processes if the inferences are to extend beyond the instances given; (2) definition and discovery of themes and patterns in variable and complex cultures; (3) propositions on the sequences of culture change framed so as to permit clear proof or disproof.

Both Burrows and Goldman have been limited by practical considerations in their selection of societies. Burrows takes those which are “best described” and some others, and touches on “most parts of - 23 Polynesia.” Goldman has not given criteria of selection other than having the societies illustrate three “historical phases” which he calls Traditional, Open and Stratified.

In each case the sorting is uneven in relation to habitat, and although Goldman mentions initially that ecology is linked to some of the cultural sequences, this variable then disappears from his equations. From the facts that the majority of Goldman's atoll societies are classed by him as Traditional in character; that no atoll societies are classed as Stratified; that three out of four of his Stratified group are of high topography, there seems no justification for excluding these variations from the argument. Some of the simple adjustment theses once propounded by Buck in relation to Maori clothing (1926) and to Polynesian cultures generally (1944) and others suggested by Burrows and Emory (1939) must be incorporated in any complete discussion of long-term culture change in Polynesia.

A fully rational basis of selection would not be easy to construct for this region; students of the cultures within the area have not set out the basic definitions which would enable it to be done. Buck writes often of “groups of societies” where adjacent cultures have a high degree of similarity and a common recent history, and often in broad terms characterizes the group as though it were uniform. But even the broadest approach would lead to regarding the Maori as a group of societies, rather than as one, and for purposes of establishing sequences of culture change, there could be no proper evasion of cultural divisions such as those established by Skinner (1921) for that group. Some of the Maori instances noted below fail to fit Goldman's generalisations, and we suggest that it is likely that not only one but eight or more separable Maori cultures failed to fit some of the generalisations.

Goldman notes comparable variation in the case of the Marquesas. It would have been profitable had he made different use of this in his analysis, instead of offering this variation as a substantiation of evolutionary development in that group of cultures; even then he twice cites a single pattern as characterizing the Marquesas. Thus he writes that the “Marquesas had uneven systems (of food redistribution) favouring the chiefs or warriors” and in diversification of gods, where tribal variation is to be expected, “the Marquesas approached the Stratified” plenitude of gods and cults.

In parts of Polynesia the political tribe is co-extensive with a distinctive culture, and for most comparative purposes this should be considered the unit. To date, Polynesian anthropology has pursued few of these basic matters of cultural boundary far enough to obtain the answers needed by the comparative student who wants to use whole cultures as his points of reference.

There is a similar need for analysis of variations within each separable Polynesian culture. Variant patterns have been described for many of them although the nature of the variations has not been set out precisely and systematically in any instance within our knowledge.

In a personal communication, Ernest Beaglehole writes: “There is no doubt in my mind that Polynesian cultures each represent a - 24 synthesis of variant or contradictory themes. Pukapuka, for instance, combines benign casualness with episodic brutal violence. I suspect that investigators have chosen to emphasise one or the other according to their personal predilections . . . I am perfectly sure that when Buck wrote his ethnology of Mangaia he deliberately emphasised the violent and warlike aspects of that culture very largely because in his estimation at that time there had been no previous systematic study of the role of warfare in a Polynesian culture. I do not think, however, that he meant this study to be a final and well-rounded version of Mangaia.”

The task of testing a hypothesis by means of the published ethnographies to see if it fits a particular culture is thus rendered hazardous; even when method is the most scrupulous, the procedure must contain a high element of guess and arbitrary choice.

Many times Goldman has concluded that such and such a sequence has resulted in greater frequence or intensity of the pattern. “Property attitudes became more predatory,” “sexual orgies became more prominent,” “women gained more political rights,” and so on. With the existence of multimodal cultures, and of variant and even contradictory patterns, adequate basis for a statement that a sequence results in “more” of the pattern is hard to give. At any time it is hard to establish the credibility of such a statement; at times it is certain that even in a statement of the simple inception of a pattern, Goldman has not succeeded. Thus, in support of his postulated sequence: “Tapus came to be enforced by physical as well as by religious sanctions” he writes that “Mangaia and Easter Island had penal sanctions,” distinguishing this stage from the Traditional societies, in which “Tapu violations were self-punishing.” But Buck, who is the author Goldman cited on Mangaia, writes: “Punishment for infringement of tapus was supposed to be brought about automatically by the gods,” (Buck, 1934: 152) which is a clear statement of a dominant ideal pattern different from Goldman's citation. Buck then adds that “The human upholders of the divine laws were not always content to wait for punishment to come about by supernatural means,” and gives one instance of secular punishment.

The raising of a pattern of minor frequency to the dominant position could play hob with any resulting generalisation. Numbers of other instances could be cited in similar fashion to challenge Goldman's statements on tapu: among various groups of the Maori, for example the Ngapuhi, as cited by Maning (1876), violations of tapu were punished quite frequently by property seizure. The truth is that tapu there as well as elsewhere was a whole set of beliefs and practices which at any one time had most of the meanings which Goldman ascribes to different historical phases of tapu. These meanings, some of them contradictory, coexisted at the same time in the same place. Are some of these meanings to be treated like the “survivals” prominent in an earlier period of anthropology? We can wish that past ethnographers had given more exposition of the situations and occasions in which one or other pattern predominated; accounts which are - 25 probably incomplete and rather impressionistic have been written many times, however, and perhaps the descriptive literature is full enough to allow others to follow Steiner (1956) in making a full analysis of tapu and its functions, based on the relative frequencies of the different sorts of sanction among other things. Until something like this is done, however, there is even less justification for taking these co-existent variations and linking their emergence with different historical phases. Nor, when selection from the variant patterns is indiscriminate, can any scheme of sequence be devised which will allow the valid demonstration of hypotheses of evolutionary change.

The third requirement for the productive study of cultural change (after the separation of comparable cultural units, and the appreciation of intra-cultural variation) is that propositions be stated in terms which conceptually allow of clear testing. A study which falls short of the satisfactory fulfilment of this requirement will certainly be limited to historical generalisation. Formulation of such propositions is not easy: in any broad and general statement, cultural themes may be contradictory and categories overlapping.

We have given our opinion that Goldman's main assumption on the effects of status rivalry can almost be reversed and remain equally plausible; yet it is incapable of adequate historical testing. We reach a similar conclusion when we attempt to check his propositions of sequence in government, war and so on. Cases can be adduced to support them; cases can be adduced to counter many of them. If some of the propositions are varied, the variant proposition can also be supported. Thus, “religion became an official arm of government.” In all Polynesian societies secular authority was generally supported by the force of dogma and ritual. Would it be more true, then, to state that religion remained an official arm of government in the supposed transition from a Traditional to a Stratified stage? Actually, neither proposition is unambiguous and demonstrable. By what index do we identify “religion,” “government” and the point of becoming an “official” arm? In the changes occurring in Polynesian societies it does seem likely that something happened comparable to the probable meaning of Goldman's statement on the change of the relationship between religious and governmental institutions. However, all Polynesian religious institutions had complex relationships which include support as well as other sorts of interaction with secular institutions. Something of this complexity is indeed suggested by Goldman when he describes a priesthood which in Traditional societies served “primarily the interests of the ruling dynasty—although it could also oppose them . . .” But when such a contradictory possibility must be included in the main proposition, proof or disproof is hardly possible.

It would not be impossible to sort out the different processes lumped together in this way by Goldman. Such broad and vague hypotheses of culture sequence could have a use in facilitating the description of Polynesian history, but they cannot be the basis of a scientific study of cultural process. Until the contradictions are sorted out, the study cannot move beyond impressionistic formulations, replete with arbitrary conclusion and ambiguous analysis.

- 26

In the foregoing discussion of methods and obstacles, we offered a few instances where we held that Goldman had not made the best interpretation or inference.

The cultures which supplied these instances also yielded others, some of which we cite below in support of the contention that unless there is exactitude of definition and procedure, the resulting rough and ready study of process within one culture area is likely to amass as many errors as did the early evolutionists in their global reconstruction.


For some broad purposes it is legitimate to write of a generalised Maori culture. In making comparisons of cultural change the generalised picture is likely to omit the detail of vitally important distinctions.

Buck frequently drew attention to tribal differentiation. “In culture, physical character, and language, New Zealand resembles a Polynesian archipelago in that the various districts occupied by descendants from different ancestral canoes may differ as much as the individual island of a group.” (1949:2). The standard subdivision by Skinner into eight culture areas followed a slightly different principle but achieved impressive results. For this region which Goldman treats as a cultural unit, Skinner points out the existence of differences as profound as semi-nomadic life contrasted to settled residence, and other distinctions which imply corresponding major differences in wealth, and many features of social organization.

Some of the facts which fail to fit the sequences postulated by Goldman are given below.

Sequence postulated by Goldman:

1. “The attitude toward human life became more callous” (in the transition from Traditional to Stratified); and “combat became more cruel.”

Exception or contradictory statement in Maori ethnography:

For the (Traditional) Ngatiporou tribe, Best wrote that “cases are on record in which enemy prisoners were cooked alive, made to lie down in a prepared, heated steaming pit and covered over.” (1924:70). The many other exceptional instances of cruelty and callousness among this group that Goldman labels Traditional lead one to conclude that if any significant difference can be established, it would be in point of frequency, an attribute of Goldman's generalisations that is never satisfactory.

Sequence postulated by Goldman:

2. “Violence in mourning began to turn outward. Almost all Polynesian mourners showed grief by beating or gashing themselves. However, only in the Stratified societies did mourners for a chief or a king turn their violent expressions of grief against other people as well.”

- 27

Exception or contradictory statement in Maori ethnography:

Best records for the Tuhoe tribe: “It is said that the act of slaying a person would serve the purpose of allaying the grief of the sick person's relatives, who expected soon to lose him,” and makes a general statement that: “Slaves were sometimes sacrificed at the death of such a person” (of superior rank). (1924:Vol. 2, 54, 158.)

Sequence postulated by Goldman:

3. “Status entered more into the conception of the afterlife. The Traditional attributed no rank distinctions to the afterlife …”

Exception or contradictory statement in Maori ethnography:

Of the multiple Maori beliefs on this matter, Best writes that one set of beliefs is that the dead pass to one of two spirit worlds. “There is some evidence to show that the upper spirit-world was the aristocratic realm of the two, and that the majority of the people knew little or nothing of this conception.” (1952:82).

Sequence postulated by Goldman:

4. “Tapu became a political instrument” (in the transition from Traditional to Stratified). “Traditional: Tapu emphasized the sacredness of the high-born and was used to conserve crops.”

Exception or contradictory statement in Maori ethnography:

Certainly in some Maori tribes, tapu had all of these functions including the political. As one example, Taylor states for the Whanganui that “a chief anxious to obtain a fine large canoe belonging to an inferior who had offended him, merely called it by his own name, and then his people went and took it.” (1870:172).

Sequence postulated by Goldman:

5. Traditional: Tapu violations were self punishing.

Exception or contradictory statement in Maori ethnography:

Again for the Whanganui tribe among whom he operated a mission, Taylor gives an instance where violation of tapu caused the mission boat to be plundered. (1870:172). And, among the references in Goldman's bibliography, Buck wrote: “In serving a chief with food, it must not be passed over his tapu head, and the absent-minded person guilty of such a lapse was punished either directly or by the unseen agencies which guarded the chief's tapu.” (1949: 349).

There are numerous other points where we would make different interpretations from the one Goldman makes but where the clash of fact is not as clear. Thus, Goldman states that the Maori “offered only limited social mobility,” a statement that is in a sense true of every known society. The difficulty of giving this statement an unequivocal meaning epitomizes the futility of attempting an analysis of culture change with loose hypotheses. As a beginning it would have been - 28 helpful to distinguish between the formal classification of rank, with their social fictions, and positions of actual power. In respect of actual power, there was considerable upward mobility in some of those Maori tribes on which there is adequate information, and there were possibilities of greater downward mobility. The social fiction related to these changes could be altered afterwards to fit them. An example of upward mobility is offered by Te Rauparaha (Firth 1929:94); the examples of downward, following mésalliance, defeat or population decline, are countless.


Buck's study of Mangaia furnishes data for a major part of Goldman's argument. In spite of this general support, there are some inconsistencies and apparent errors in Goldman's use of Mangaia as an example of the postulated open or middle stage in the cultural evolution of Polynesia.

In the first place, some of the claimed sequences, such as the importance of war and the unpleasantness of the after-world, both of which Goldman holds to increase in intensity with evolution from Traditional to Open to Stratified, fail to show Mangaia in a middle stage. Buck, whether or not he deliberately over-emphasizes the importance of war in Mangaia, places that society at one extreme of Polynesian variation in this respect. Other sequences cannot be maintained because the situation is apparently the same at the beginning as at the end of Mangaian history, and there is probably no way of arriving at relative frequencies. Thus, the challenge to hereditary claim to temporal power, which Goldman's hypothesis should find as meeting with final success, is attested as happening in the first years as well as at the end. (Buck 1934:43).

To deal with this sort of Procrustean fit, Goldman concedes that the process of development was more complex than his presentation of it, and that retrogressions may have taken place. This admission has frequently been offered in past decades as one of the correctives of simplistic or unilinear evolution. Surely this is not an achievement of greater sophistication about the evolutionary process. Failure to account for alterations or reversals of direction follows from the failure to construct models of cultural change that are adequate. Particularly when the scope of the inquiry is too wide, there is the danger that some processes remain hidden. Of course, when they reveal themselves in unexpected ways, the analysis should be radically revised; no understanding is gained by mere “retrogression.”

Again, in a number of themes and patterns, Mangaia manifests some variation. In most of these, such as those which bear on the relative dominance of the priesthood and the warriors, of the religious and the secular, no simple dominance of one pattern over the other can be gleaned from the material. Buck gives contrary instances and at times offers somewhat contradictory generalizations. In others there are clear alternatives, but Goldman makes no exposition of his method of selection of one or the other of them.

- 29

Some detail and some other points are presented below.

Sequence or characterization given by Goldman:

1. “Warfare became more prominent and more serious … It was virtually state policy in all the Open and Stratified societies, where the motives were usually predatory …”

Exception or contradictory interpretation from Buck:

Oral traditions give a long list of battles from the beginning of Mangaian history, and include statements of causes, with all the usual variants found in other parts of Polynesia: insult, revenge, anticipation of revenge, and so on. Buck's data do not show any apparent change over the period of Mangaian oral history, although the postulated operation of status rivalry would seem to demand change. And many of the wars do not make sense if the “goal was to displace an enemy from his land” as Goldman infers. (Buck 1934:47). The contrasting theme of peace is not presented by Goldman. Yet Buck wrote: “The maintenance of peace was of paramount importance to the ruling Temporal Lord, for with the shedding of blood his reign automatically ended.” (1934:124).

Sequence or characterization given by Goldman:

2. “Religion became an official arm of government … In Mangaia, for example, priests who sought political advantage for themselves lost their religious immunities.”

Exception or contradictory interpretation from Buck:

Buck's most relevant passage limits this liability to wartime: “When war broke out, however, the power of the warrior became supreme and the sacred nature of his office did not protect the Shore High Priest's life if he interfered with temporal matters.” (1934:117). And Buck also shows variants which are pertinent: “In important matters the tribal god was consulted and his decisions as made known by the priest were implicitly obeyed. The tribal priest thus exercised considerable authority on the policy of the tribe.” (1934:152). That Mangaian religion was not merely an instrument for political ends is shown by the voluntary self-sacrifices of men of high rank, indicating something like piety as well as family devotion. (1934:49).

Sequence or characterization given by Goldman:

3. “Sexual orgies became more prominent” (in the transition from Traditional to Open to Stratified).

Exception or contradictory interpretation from Buck:

Buck gave no clear reference to any developing prominence. On the contrary the condemnation of adultery was “not less severe” than among the Maori

- 30

Sequence or characterization given by Goldman:

4. In relation to religion, Goldman treats of Mangaia as homogeneous.

Exception or contradictory interpretation from Buck:

In reference to the place of departure for the underworld, Buck writes:

“Doubtless the different tribes maintained their own ideas on the subject,” and points out the worship of different gods, or at least differing orders of primacy of the gods, in the various Mangaian tribes.


There are, then, certain canons of method and the use of evidence which have not been followed rigorously by recent students of evolution. The lack of discipline is all the more remarkable when we recall that the main canons were set forth by Boas in his 1896 essay, which has been available to students as a classic guide in methodology ever since. Despite Boas' leadership, so sloppy were many of the following studies of historical reconstruction that Radcliffe-Brown and Malinowski led the whole British group of social anthropologists away from pursuit of these interests, and among that group sloppiness and historical reconstruction were, rather unfairly, considered to be synonymous. In the course of the succeeding controversy, both Radcliffe-Brown (1952) and (less rigorously) Malinowski (1945) produced statements which out-lined standards of procedure which are applicable to diachronic studies as well as to synchronic functional ones.

Optimistically, many anthropologists believe and hope that these issues are now dead and buried. One of them is: the historical dimension is everywhere admitted to be important, and many British functional studies are just as diachronic in outlook as any account of evolution (Leach, 1954; Barnes, 1954). But the preceding criticism will have demonstrated that the canons of the discipline are still not fully absorbed. This may be partly a result of the broad scope of anthropology, which teases us with Olympian fancies, and partly a result of the inadequate emphasis upon scientific method in our training programmes. Lack of discipline is certainly not confined to evolutionary studies; but here the facts are more difficult to control, and the temptations of grandiose architecture are well-nigh irresistible.

Specifically, the rules to which we have already drawn attention are:—

  • (1) Circularity in method should be avoided. Goldman set up a hypothesis to the effect that three types of Polynesian society were historically connected. He then assigned cultural elements to each of these three types, appointed historical examples as representative of each type, and found the elements he was looking for in each historical example. He phrased this as if he were showing, in each case, that there was a historical developmental sequence. All he was doing, where his facts were right, was to show that there were in fact historical
- 31
  • examples of each type. This in itself says nothing about historical connections. The goal of demonstrating connection was approached more closely by Burrows with the use of some evidence from linguistic reconstruction, oral tradition and a closer study of processes.
  • (2) Generalisations about particular societies, if they are to be used to demonstrate historical change, must be established as valid. This usually involves reference to such matters as the generality (in terms of frequency) of an idea or custom; the possibility of variations and alternatives (even the contrary of the generalisation) being true; and the logical probability that the statement is true in the light of known evidence. Often evolutionary studies, for lack of data, can only attempt the last of these three methods.
  • (3) Explanations about cultural development which purport to interpret actual historical change are suspect when they are monistic. This has been said so often that it is boring to write it again; yet Goldman's argument is a thinly veiled monism.
  • (4) Categories (societies and cultural elements) must be sharply defined in order that institutions and cultural characteristics may in fact be compared from one society to another.

To meet the methodological demands of the rules, a writer would have to use a great deal of evidence and rigorous formulation of categories. No evolutionary study, attempting to relate societies both historically and qualitatively, could be regarded as successful until the following minimum requirements had been accomplished:—

  • (1) The first point is to demonstrate an actual historical connection of definiteness and magnitude sufficient to place the matter beyond doubt. It must be possible to use evidence, independent of the qualities of the societies, to demonstrate the proposition that “Society A developed from Society B or from a Society C which was of the same type as Society B.”
  • (2) The second point is that each of the societies under examination must be fully known, to the extent that a functional analysis is possible. This is because our interest in processes leads us to examine institutions as they operate.
  • (3) If an objective of the study is to demonstrate qualitative development, the categories to be compared should be expressed in such a way as to constitute a graded scale, and it should be possible to identify, with reasonable certainty, just where on this scale a given institution should lie. Furthermore, any proposition of this kind should be capable, conceptually, of variation and reversal in order to enable it to be tested. It is tempting to overlook this canon.

Of these three canons, the first may have considerable utility if it is followed alone. In that case we have a cultural historical study which demonstrates historical change, and from which certain broad conclusions may be drawn which are appropriate only to the societies and historical circumstances under immediate examination. This is a limited but valuable task.

- 32

All three canons are important if anything is to be said about cultural evolution, for evolution implies some kind of graded and qualitative change, and cross-temporal cultural comparison, which cannot be examined by any lesser method.

But if the objective is to go further, and to use a knowledge of historical examples of evolutionary development as basis for the construction of sociological laws of universal validity, the data derived through the employment of the above three methods are insufficient. They may be useful as illustration, but cannot in any sense constitute proof. It may be helpful to set out in summary manner the various procedures which are available to us as anthropologists, and to assess, briefly, the distance they can enable us to travel along this particular road.

  • (a) The methods of (1) to (3) above can, of themselves, serve not only as illustration, but also to assist in the formulation of hypotheses, which then remain to be tested.
  • (b) Certain features of observed societies may be selected for attention, not as isolated elements, but because together they constitute either a systematic institution within a wider social system, or else characteristics which, in the abstract, form a model of such a social system, conceived from a particular theoretical viewpoint. Thus, Spoehr (1947) isolated the systematic institution of kinship among a number of Indian cultures of the south-eastern United States, and Redfield (1934) abstracted characteristics of types of Mexican social systems.

The important point to note about such studies is that we are no longer dealing with evolution, according to our original notions, but with cultural change. True, Redfield's later work suggests that the folk-urban continuum contains within it the germs of universal laws, and it might be possible to infer that Redfield hoped that all folk societies at all times and in all places had the characteristics of his model, and that the process of transformation into urban societies was always in accord with the processes he abstracts. But this, to our way of thinking, is the most dubious part of Redfield's presentation. The valid part of Spoehr's and Redfield's contributions consists of ideas which may be formulated in this way: “Given defined conditions of change, x, then a defined social system, y, will change in certain prescribed ways which result in z.” In other words, if either x and y are not present, z cannot be presumed to follow.

Such a formulation is universal, in that it applies to all situations in which x and y occur. But it is not universal in the sense of applying to all societies at a similar broad stage of development. Furthermore, history cannot always tell us whether x and y were in fact applicable, for there is always the expectation that information is incomplete, and there is moreover the expectation that the model formulated for the Tuhoe can only be applied to the Ngapuhi at the cost of major distortion.

In other words, Spoehr and Redfield have used their observation to - 33 create abstract models, but the validity of the models has nothing to do with the question, can the models be applied to further societies? Abstract theory is stimulated by history, but is seldom tested by it.

It should also be noted that, insofar as field work is to be used to validate such a general proposition, it has to be carried out according to this method, within the framework of a single culture. The narrower and the more closely defined the culture, the easier it is to isolate variables, and the nearer we get to empirical validation. The method is to compare two or more variants of the model, in which the same variables have changed, all others remaining constant. The variants are then combined into a single diachronic model which gives the theorems of process.

Now although the model is diachronic, stating that z emerges out of y, it is extremely difficult to find historical situations in which the variables are truly limited; it is more common to find change on a wider front, so that again we are in danger of being left with general interpretations rather than precise theorems. It is possibly more profitable to compare two variants within a single contemporaneous culture, in the manner of Redfield and Fei & Chang (1945).

(c) Already the method of (b) is merging into the next two. The first of these is the extension of comparison beyond the confines of like societies. So far we have described the attainment of two types of conclusion: (i) historical generalisation and (ii) theorems of cultural change. To go beyond these by the use of the comparative method is the objective of writers such as Radcliffe-Brown, Levi-Strauss, and Homans and Schneider. Homans and Schneider (1955) provide a clear-cut example of the advantages of the technique. Specific propositions are advanced to explain the occurrence of matrilateral cross-cousin marriage by demonstrating that wherever it occurs certain features are correlated with it; similarly patrilateral cross-cousin marriage is explained; and the two sets of explanation, dealing with unlike societies, are shown to be consistent with a more fundamental and universal theory relating marriage rules to the relationships involved in kinship authority.

This generalisation is achieved at the cost of a high degree of abstraction, which involves making statements about universals of human behaviour which are capable of being formulated in a limited institutional context. As it stands it is a synchronic theory, but just because it is based on a universal statement, it is capable of being used to explain change. When it is applied in this way, it will not result directly in a theory of evolution; it will result in theories of cultural change.

Our main suggestion is that the comparative method, to be successful in reaching universal laws, must be able to subsume unlike societies and unlike institutions into a common basis of abstraction, which deals in such categories as principles of choice, structure of action, and the force of pressure to change (Belshaw, 1954).

(d) The validity of general laws may also be tested by reference to their logical consistency with the various abstract models which are - 34 set up to represent social systems. This possibility has not been pursued very far in anthropology, since we usually adhere to the concept that ours is a natural science, dealing with natural phenomena, and that a social system which has not been discovered to exist cannot have meaning for our science. This is not the place for an extended argument, but it is at least possible to repeat the suggestion (a) that our description of any social system involves the use of an abstract model, and the historical contextual materials frequently blind us to this, and that hence (b) we are concerned, not with historical reality, but with logical propositions derived from permutations and combinations of assumptions. If this is so, we can expect the importance of this method to increase in the future; again, in terms of cultural change rather than of evolution.

  • BARNES, J. A., 1954. Politics in a Changing Society: a Political History of the Fort Jameson Ngoni. Capetown, Oxford University Press.
  • BEAGLEHOLE, E. “Cultural Peaks in Polynesia.” Man, September, 1937.
  • BELSHAW, C. S., 1954. Changing Melanesia. Social Economics of Culture Contact. Melbourne, Oxford University Press.
  • BEST, E., 1924. The Maori. Polynesian Society Memoir, Vol. 5.
  • — — 1952. The Maori as He Was. Wellington, Government Printer.
  • BOAS, F., 1896. “The Limitations of the Comparative Method of Anthropology.” In Boas, F., Race, Language and Culture. New York, The Macmillan Company, 1940.
  • BUCK, P. H. (Te Rangi Hiroa), 1926. The Evolution of Maori Clothing. Polynesian Society Memoir, Vol. 7.
  • — — 1934. “Mangaian Society.” B.P. Bishop Museum Bulletin, No. 122.
  • — — 1944. “Arts and Crafts of the Cook Islands.” B.P. Bishop Museum Bulletin, No. 179.
  • — — 1945. “An Introduction to Polynesian Anthropology.” B.P. Bishop Museum Bulletin, No. 187.
  • — — 1949. The Coming of the Maori. Wellington, Whitcombe & Tombs.
  • BURROWS, E. G., 1939. “Breed and Border in Polynesia.” American Anthropologist, 41:1-21.
  • FEI, H-T. and CHANG, C-I., 1945. Earthbound China. A Study of Rural Economy in Yunnan. Chicago, University of Chicago Press.
  • FIRTH, R., 1929. Primitive Economics of the New Zealand Maori. London, Routledge.
  • GOLDENWEISER, A., 1937. Anthropology. New York, Crofts.
  • GOLDMAN, I., 1955. “Status Rivalry and Cultural Evolution in Polynesia.” American Anthropologist, 680-97, Menasha.
  • HOMANS, G. C. and SCHNEIDER, D. M., 1955. Marriage, Authority and Final Causes. A Study of Unilateral Cross-cousin Marriage. Glencoe, The Free Press.
- 35
  • LEACH, E. R., 1954. Political Systems of Highland Burma. A Study of Kachin Social Structure. London, London School of Economics and Political Science.
  • MALINOWSKI, B., 1945. The Dynamics of Cultural Change. An Inquiry into Race Relations in Africa. New Haven, Yale University Press.
  • MANING, F. E., 1876. Old New Zealand. London.
  • MEAD, M., 1937. Co-operation and Competition Among Primitive Peoples. New York, McGraw-Hill.
  • RADCLIFFE-BROWN, R. B., 1952. Introduction. In Structure and Function in Primitive Society. London, Cohen & West.
  • REDFIELD, R., 1934. “Culture Changes in Yucatan.” American Anthropologist, 36:57-69, Menasha.
  • SKINNER, H. D., 1921. “Culture Areas in New Zealand.” Journal of the Polynesian Society, Vol. 30.
  • SPOEHR, A., 1947. “Changing Kinship Systems.” Anthropological Series, Field Museum of Natural History, 33:153-235.
  • STEINER, F., 1956. Taboo. London, Cohen & West.
  • TAYLOR, Richard, 1870. Te Ika a Maui. London.