Volume 86 1977 > Volume 86, No. 2 > The culture of gender in Pukapuka: male, female and the Mayakitanga 'Sacred Maid', by Julia Hecht, p 183-206
                                                                                             Previous | Next   

- 183

The Pukapukan origin myth may be summarised as follows: The birth and socialisation of the first Pukapukan, Mataliki, was supervised by two male gods, Tamayei and Te Atua Vaelua. At the latter's direction, Mataliki sought and obtained a wife, Te Vaopupu, from another land and returned with her to Pukapuka. They had four children, two male and two female. The elder son and daughter mated producing the original chiefly line; the younger son and daughter mated producing numerous offspring who, when disaster struck, divided themselves into the two matrilineal moieties. 1

In their recent article on “Male and Female in Tokelau Culture,” Huntsman and Hooper describe the brother-sister relationship as the dominant male/female dyad in Tokelau culture, and contrast this with the situation in Pukapuka where, from their understanding, a greater emphasis is placed on the husband-wife dyad. 2 This myth would appear to support Huntsman and Hooper's assertion that the husband-wife relationship is emphasised in Pukapukan culture as opposed to the brother-sister relationship in Tokelau, since brother and sister are virtually turned into husband and wife. Furthermore, there appear to be no other oral traditions that focus on the brother-sister dyad. Yet the brother-sister relationship is by no means insignificant in Pukapuka, and does figure, if in veiled form, in the origin myth. I will return to this at the end.

In this article I shall present data on the culture of gender in Pukapuka, drawn from my field research on the island between 1972 and 1974. I shall use this information to reinterpret aspects of the account of kinship given by the Beagleholes in their Ethnology of Pukapuka, 3 and to show that the brother-sister relationship probably occupied a crucial position in - 184 Pukapukan symbolic and social organisation.

I begin with an outline of Pukapukan social organisation and a brief review of the Beagleholes' data, followed by a general treatment of male and female in Pukapuka. I then go on to present additional information on collateral kinship in Pukapuka with a view towards examining its significance in a wider Polynesian context.

I shall be discussing primarily traditional concepts and practices, using data from the Ethnology of Pukapuka and from my own fieldwork. There are occasionally problems with time reference, for the Beagleholes make reference to an undated past as well as to their field period of 1934-35, and are not always clear about what time period they intend. This problem is exacerbated because I did my fieldwork at yet a more recent time and elicited additional information about the past as well as collecting data on the contemporary culture and society. In this paper I have attempted to use the present tense exclusively for contemporary usage, but it is not always simple to establish contemporary as opposed to past when dealing with ideas. Therefore, although my choice of tense may at times appear inappropriate, I have made a considerable effort to be consistent.


Pukapuka is an atoll in the northern Cook Islands. Its three islets have a total area of between 1250 and 1800 acres. 4 In January 1974, 761 people resided on the atoll. Pukapuka's nearest neighbour, Nassau Island, 5 42 miles south-east, was occupied by an additional 118 Pukapukans in 1974. The next nearest of the Cook Islands is Manihiki, 286 miles north-east. Pukapuka's nearest neighbours to the west are Tokelau and Samoa, both somewhat under 400 miles away. Pukapuka is 715 miles north-west of Rarotonga, the administrative centre and capital of the Cook Islands.

A few details of traditional cultural ideology and social organisation provide necessary background for the treatment of Pukapukan gender that follows. Since pre-contact times there have been three villages. These villages are of primary importance in the contemporary social organisation so that with few exceptions group activities are now based on village membership. Yet, there are positive indications that kin-based categories were at least equally as important as village afflliation in the past.

Patrilineal, matrilineal and cognatic descent principles were recognised. Membership in the and their subdivisions the wakavae ‘patrilineal or burial categories’ was validated by common place of burial with the sociological father, and was therefore a tentative affiliation during a person's lifetime. 6 Membership in the wua ‘matrilineal moieties and lineages’ and - 185 their subdivisions the keinanga ‘matrilineal sublineages’ was given by the premise that mother and child shared blood substance and therefore was established at birth. There were seven subsuming twenty-one wakavae, and four matrilineal wua subsuming thirteen keinanga. 7 While burial and matrilineal categories held nominal title to land, wuanga (or maelenga, in traditional usage) ‘cognatic descent categories’ were directly involved in the exploitation of sections of property. These cognatic descent categories were usually known by the name of an apical male ancestor and appear to have been about five generations in effective depth.

Modern social organisation is almost exclusively based on cognatic principles and village affiliation. However, there are titles which still inhere in particular burial lines, and documents suggest to some extent when and how the village unit gained ascendancy over its constituent burial groups.

The Beagleholes' stated purpose was “to study the life of the Pukapukan people with reference to the ways and customs of pre-missionary times” and they did not relate their findings to any general theory of culture and social structure. 8 Their presentation of the data is directed towards describing constituted social groups and actual social interaction, primarily in the present but with frequent reference to earlier practices.

In the Ethnology of Pukapuka, there is no systematic treatment of cultural symbolism nor were cultural principles about the constitution of the Pukapukan world directly considered. For example, although the Beagleholes postulated the historical primacy of matrilineal groupings, they did not consider the implications of the matrilineal focus, 9 and while they noted some denotations of the terms for matrilineal and patrilineal groupings, they did not treat them as part of a meaningful system. 10

Consideration of cultural principles, based upon eliciting and analysing symbolism, leads to a fuller analysis of several issues that the Beagleholes either did not investigate fully or found contradictory or obscure. These issues focus on cross-sex and collateral kin, for which the terms mayakitanga, ilamutu, yinākava, kainga wakamā and wale atua were noted. 11 The analysis of these terms is a focus of this paper.

The Beagleholes' understanding of these terms was doubtless obstructed by the way they divided their material in analysis. While political organisation was indeed included in the general social organisation section of the Ethnology of Pukapuka, within that section it was analytically separated from kin and lineage organisation, a division which masked the breadth of the congruence between village and kin organisation. For example, every Pukapukan belonged to one of the three villages by affiliating with a tuanga kai ‘food share unit’, 12 yet these units in the past were based on - 186 wakavae ‘burial subcategories’, with wives affiliating to their husband's ‘food share units’ and villages. Similarly, almost all statuses within the village polity were based on prior affiliation with smaller kin-based categories, primarily and wakavae ‘burial categories’, though male elders of matrilineal categories were also representatives in village meetings. 13

In relation to my study here, the a priori division of “politics” and “kinship” as separate institutional domains, which at least in this case have little ethnographic validity, has obscured the significance of collateral relationships in Pukapuka. Where kinship terminology for collaterals appears to be inexplicable in the Ethnology, 14 it is now seen as part of a system which had previously escaped analysis because the terms are used to refer to kinsmen in contexts which are not obviously those of kinship. 15 As these terms and their associations are intimately connected with concepts and roles of men and women, a re-evaluation of these kinship relationships is required for the comprehension of male and female in Pukapukan culture.


The Pukapukan view of the atoll divides the islets into te vai, lower ‘wet (swamp)’ surrounded by te wenua, higher ‘(dry) land’. The dryness of the higher coconut ‘land’ is also associated with males and the periphery in varying contexts; while the wetness of the lower ‘swamp’ is associated with females and the centre or interior in context. These linked associations of male, dry, up, outside and female, wet, down, inside are played out in Pukapukan social patterns and inform cultural concepts.

All of these associations of male and female appropriately converge in the patterns of economic activity. In geographical terms, men work on the periphery and women work in the centre. On land, women tend crops “down in the (wet) taro swamp” in the interior of the village islet, while men harvest coconuts and guard over the women on the periphery “up on the (dry) land.” As is to be expected, these distinctions are not absolute in practice. Nevertheless, approximately 80 percent of the stewards of dry land are men and, although men initially excavate and assist with the mulching of the swamps (which can perhaps be interpreted as making dry land into swamp), the routine planting, weeding and harvesting of taro are almost invariably women's work.

Both men and women may fish in the lagoon and on the reef where fishing often involves going into the water and diving. Traditionally only men might go beyond the reef into the open sea as in the saying: “The bird of Mataliki (the man) goes on the sea, the bird of Taua (the woman) stays on the land.” 16 Thus the ocean, though wet, appears to be a male domain. But it is probably the ocean surface in particular that is intended, given injunctions to keep canoes dry on top of the ocean.

- 187

The economic contributions of female and male on land balance each other and are symbolic respectively of female regeneration and of male continuity (associations that appear elsewhere). Women provide taro from the interior swamps for their children, kin and affines, and for larger distributions of food. Taro has a short life cycle, as do human beings, but it is easily renewable. Men provide coconuts from the periphery. Although it may no longer be productive, the coconut tree lives across generations of men, and leaves a mark of their passage. Even when trees die, the stumps are often left to serve as useful markers in support of land claims. An informant talks about the coconut tree, source of the paradigmatically male contribution:

The coconut tree remains: it is we who will die. We planted the coconut tree. We ate when the tree fruited. Our children came, and they ate the fruit of our coconut tree. The parents and the children eat during their lifetime. It is said of our coconut tree: it is an eating tree for the parents, and for all of us (children).

In recreational activities higher periphery and lower centre are contrasted. The call “up . . . down” announces the union of male and female in the dance. Today when Christian hymns are sung in formal competition, the men of each village sit up on benches facing towards the centre and their women sit on mats down on the ground, facing outward towards their men.

Traditionally in the political context, while men gathered outside in the public arena, women were excluded from participation. Chiefs, priests, members of te pule ‘the village corvée units’, food dividers, elders and all other representatives in the village meeting were men. Along with exclusively male participation most village statuses were associated with burial categories. Chiefs and priests represented specific burial categories, with four of the burial categories having aliki ‘chiefs’ at their head, and others having langatila ‘lesser chiefs’ as their leaders. Of these positions, only village corvée service and attendance and participation at village meetings are now open to women. In addition, while within the internal organisation of the matrilineal sublineages the leaders were the oldest members of either sex, male elders represented the sublineage outside in village meetings. 17

Periphery and centre also have connotations of strength and control which are particularly evident in cultural attitudes towards sexuality. When asked what is the difference between men and women, informants usually say, “Men are strong; women are weak,” yet this is not a statement of what should be but of what generally is. Strong women who produce a lot of taro and children are praised in chants and one of the large taro swamps is named in commemoration of such a woman. A strong woman, a good worker and an able sportswoman is prized as a wife. In appropriate contexts, then, female strength is positively valued; it is simply that men are generally stronger.

Women, however, are said to be stronger than men in coitus specifically because the female maintains her strength after the male loses his at climax. Sexuality is generally regarded as positive and pleasurable, yet female sexual capacity arouses some male anxiety. Female sexual strength, or - 188 more specifically aggressiveness, is usually negatively valued and even dangerous. In a story about the woman Wekeli, she is so pleased with her lover that she suggests to him they meet again. He replies that she is “bold”, and he tells her not come near him again, “you have had all you will get. Your words sound bad, so bad that all of the men of the island, and the women too, would get angry if they knew of them.” 18 The sexual insatiability of a woman is said to have brought on the great wave that swept away most of the population about 400 years ago. 19

Yet in contexts of culturally approved licence, the sexually aggressive female was the appropriate instigator and her role positively sanctioned. Women, for example, were expected to initiate the period of sexual licence following the abstinence required before and during fishing competitions and women, as well as men, organised poloaki ‘communal dates’ 20 which began as social gatherings out in the bush for eating, singing and dancing, until later when couples paired off and left the group to engage in private sexual intercourse.

Assertions of female sexual aggressiveness and insatiability suggest that female physiological strength was coupled with a lack of control that rendered it a social weakness. In order to channel uncontrolled female strength for the benefit of society it was necessary that it be controlled by males, who while guarding the periphery also controlled the centre.

Wakatane and wakawawine denote the attainment of sociological manhood and womanhood. As a person of specific sex, it is presumed that one will act accordingly, not only sexually but in other gender-specific behaviour as well. Men demonstrate their manhood most prominently by publicly displaying their strength in sports competitions and a man reduces his stature by not participating. Women demonstrate their womanhood most saliently by producing and caring for children. They are not derided if they do not join in sports competitions and they may not have participated traditionally. Most sports definitely asserted to have been traditional, wrestling and dart-pitch, are still exclusively masculine.

Wakatane and wakawawine may be appropriately glossed as ‘maleness’ and ‘femaleness’. The cultural associations of maleness/femaleness are clearly dry/wet, up/down, periphery/centre. The association of strength/weakness is more complex because a contrast between general physical and sexual strength is confounded so that statements about male and female strength initially appear contradictory. When self-control is included within the concept of strength, then the situation becomes more understandable.

- 189

With this background, I now go on to look at specific relationships, first those of male and female as sexual partners, spouses, and parents.


The attributes defining ‘(dry) land’ and ‘swamp’ and basic gender distinctions also relate to the realm of sexual relations and conception. Women are ‘wet’, specifically during sexual relations, menstruation and childbirth; in contrast men are ‘dry’. In coitus, although not necessarily in foreplay the man should be on top and the woman underneath. While the female remains strong, the male ‘dies’, i.e., detumesces, at climax.

Informants say that the wua ‘seed’ contained in the yua ‘fluids’ of both parents must combine to create the child. One act of intercourse, in which the genetrix contributes valewu ‘female sexual fluid’ and the genitor contributes nene ‘semen’, is necessary and sufficient for conception. The genitor's role is recognised as necessary to the creation of the child, not just forming it as has been reported for many societies with matrilineal descent ideology.

The Beagleholes' informant said that the child was composed from the semen of the father and the blood of the mother. It is not completely clear how the semen and blood relate to concepts of ‘seed’ and ‘fluid’ above. While the genetrix is said to contribute both ‘fluid’ and ‘blood’, no contribution is recognised from the genitor beyond the ‘seed’ contained in ‘semen’.

Both the Beagleholes' and my informants gave several opinions about the determination of the sex of the child. Some said that a male was composed mainly from semen, and a female mainly from blood or the mother's ‘seed’. Others said that if a child of either sex bore the likeness of one parent more than that of the other, then the child was composed primarily of that parent's contribution. 21

The period of gestation is called te ikonga, from iko ‘to germinate’ as in ‘it is the semen which makes the child germinate’. 22 Pregnancy is recognised by the cessation of the menses or diagnosed by a Pukapukan medical practitioner who feels the woman's abdomen to see if potopoto toto ‘lumps of blood’ have formed in the womb. Apparently, and this is my inference, the idea is that intromission of semen causes the coagulation of blood to form these lumps. Pregnancy is also recognised by dreams of the mother, father, or a kinsman, which, following the paradigm, are of a taro swamp for a female child and of a coconut palm for a male. When pregnancy is diagnosed, the mother, if unmarried, is said to know which of her suitors is the genitor, although I am not clear about what data inform her knowledge.

If the mother is unmarried, her father or another kinsman will become the sociological father of the child, often through formal adoption. If the mother is married, her husband will usually be considered the sociological father, even if he is not the genitor. In order that he not unwittingly commit incest, the child will eventually be informed of the identity of his genitor.

- 190

On Pukapuka rules prohibiting marriage, which implies sexual relations, apply as well to sexual relations outside of marriage. In the past, these rules prohibited marriage between persons within the same cognatic and matrilineal descent categories. Marriage is no longer forbidden between people of the same matrilineal category, specifically the matrilineal sub-lineage, as it was formerly, but is still prohibited between persons related through male or female links closer than three generations removed from a common ancestor, i.e., second cousins may not marry but third cousins may. If not otherwise barred on the basis of cognatic links, affiliates of the same ‘burial lineage’ are allowed to mate.

Beyond these negative rules, choice of sexual partners is based on personal qualities such as beauty, skill and character. Although Pukapukans are willing to comment on the attributes of particular individuals, when asked to pick out the most attractive or desirable man or woman, they refuse to rank or compare their fellows asserting that “everyone is the same”. In the egalitarian society of Pukapuka today, it is considered wītoki ‘unbecoming or bold’ to be ostensibly better than anyone else. Perhaps there is a lesson to be learned from the story about a man who was so sexually desirable that women flocked to him, neglecting the other men. In the end, he and his matrilineal sublineage were banished from the island. The implication of this banishment appears to be that, given the blood identity of the sublineage, similar behaviour could be expected of any of them. 23

While personal choice of sexual and marital partners is and was the general practice, persons of chiefly status in the past had their mates chosen “from a family wealthy in taro and coconut trees” in order that the gift giving be on “an elaborate scale commensurate with the wealth and status of the two families concerned.” 24 However, in general there seems to have been little distinction in wealth between chiefly and other lines.

Courtship is usually a private sexual affair which also involves gift exchanges of food and ornaments between the man and woman. Only when the couple has decided to marry do they ask for their parents' permission, because it is not considered correct to involve one's parents in sexual affairs not intended to culminate in a publicly sanctioned union. Occasionally, a man intent on marriage might try to influence the parents of an unwilling woman, 25 but this tactic appears to have been less effective than supplicating an appropriate cousin as described below. Parents tried to ensure a good marriage for a child through beautification rituals or by refusing the child permission to live in a young men's or young women's house thereby limiting his or her pre-marital sexual activity. 26

- 191

While parents should not be involved until the courtship has progressed towards desire for a formally sanctioned union, the ‘cousin-in-avoidance’ 27 could play an important role:

A sure way for a boy of former times to get an unwilling girl to marry him was for the boy to work for the girl's wale atua. No word was said about courtship, but the male wale atua knew that the boy wished to marry his female kinswoman. If he was pleased with the boy, he sent a message to the girl, telling her to marry the boy; a girl would rarely disobey this command from her wale atua. A girl might similarly court a boy by working for his female wale atua. 28

There are few data on traditional weddings. The formal union itself was accomplished by the man's sleeping with the woman in her house and remaining there until morning instead of leaving before dawn. In the feast and exchange of gifts that followed, the Beagleholes note that there was no differentiation in gift-giving like that associated with the bride's side and the groom's side in Samoa. 29

Informants say that the families of bride and groom exchange the same kind of gifts “because there are both men and women in each group”. In each men contribute fish, coconuts, sennit and like goods, and women contribute goods of female labour or manufacture such as taro and pandanus mats. Among the descendants of the MF, MM, FF and FM of the bride and those of the groom, these male and female contributions are pooled. Thus a collection of goods is brought forward from each of the eight groups of wuanga ‘cognatic descendants’.

Some quantitative statements about past marriage choices may be made on the basis of statistics derived from a sample of 100 couples married during the early to mid-nineteenth century in the fifth to sixth generation following the great wave of 400 years ago. Since virtually everyone can trace relationships through the six sons of the high chief following the wave, I have ignored such relationships in these calculations. About 52 percent of all marriages were endogamous in respect to the wuanga, the five-generation span cognatic descent category, yet in only one case in the sample can a genealogical relationship between spouses be traced through solely female links. As noted above, there was a specific prohibition on marriage between members of one matrilineal sublineage, and such a relationship could be assumed quite simply on the basis of common affiliation which would not necessitate the tracing of a genealogical relationship. In 20 percent of the marriages the spouses appear to have been kainga wakamā or wale atua ‘prohibited cousins’ within a three generation range of a common ancestor, forbidden cognates today. Perhaps the frequency of such marriages attests to the typical shallowness of remembered or recognised genealogies involving female as opposed to male links removed a few generations.

- 192

Serial monogamy on divorce or death of one partner was common; levirate and sororate secondary unions were occasional. Sisters might be exchanged and two brothers might marry two sisters. However, no general pattern according to lineality is perceivable. While computerisation of the data might possibly produce some statistical patterns, I suspect these would be no more revealing than what is easily discernible from the data: lines tend to diverge and then coalesce after the lapse of four generations or so. This may be expectable simply given population size and isolation, which is not to say either that these marital patterns are insignificant or that Pukapukans are not aware of this tendency.

In 1974, 15.6 percent of the adult census population had been permanently separated or ever divorced. 30 Although discrimination between death and divorce as mode of termination of marriage is only rarely possible in the genealogies, the number of serial marriages suggests that the divorce rate was equally high in the past. The cultural emphasis on the husband-wife pair does not mean that it is a particularly solidary dyad. Indeed, the emphasis on the positive value of sexuality, rather than on the maintenance of the union itself, may contribute to the high frequency of separation and divorce, because if a marriage is sexually unsatisfactory, it is considered understandable, although at least today certainly regrettable, when one or both partners seeks satisfaction elsewhere.

Post-marital residence is about 60 percent virilocal. This percentage of virilocality corresponds to the association of with a definite locality. 31 While brothers tend to remain on the land of their , their sisters usually leave to join their husbands.

Adult brothers-in-law or sisters-in-law occasionally live in the same household, but usually with separate sleeping houses. It is rare that a man, his wife, his sister, and her husband live in the same household. In 1974 one of the only two such instances in 117 households was a case of sister exchange. There is often a marked tension between sisters-in-law, usually expressed outwardly in squabbles over household belongings and also friction over the exploitation of taro swamps. The brother-in-law relationship appears to be more formally regulated, e.g. by the contemporary rule that brothers-in-law may not wrestle. 32

Pukapukan kin terms for affines exhibit a so-called Eastern Polynesian elaboration. 33 Pukapukan kin terms are almost exclusively terms of reference, for people typically address each other by personal names, especially when they are of the same generation or senior to junior in adjacent generations. 34

Own spouse is (taku) vale ‘(my) spouse’, or the gender specific tane ‘man, husband’ or wawine ‘woman, wife’. Other terms used to refer to - 193 one's spouse are tokalua which suggests an aspect of complement and partnership (toka- ‘counting prefix for human beings’ lua ‘two’), and tomatoukau, literally ‘our (plural, exclusive) person’. 35

Affines of ego's generation, whether by own marriage or kinsman's marriage, are collectively called tao tangata, but may be distinguished as taokete tane and wawine, respectively ‘wife's or husband's brother, sister's husband’ and ‘wife's or husband's sister, brother's wife’. Spouse's parents are matua angavai tane (male) or wawine (female), but spouse's grandparents are not distinguished terminologically from own grandparents, tupuna tane (male) or wawine (female). Ego refers to his children's spouses as unaonga tane (male) or wawine (female) and other affines of ego's children's generation are collectively known as konga tangata, specifically tua konga tane (male) or wawine (female) ‘children's spouse's siblings’. As with affines two generations above, there is no specific term for a grandchild's spouse. 36

A lack of specific terms of reference for affines beyond a limited generational span begs the issue of possible close genealogical connection between spouses, a fact perhaps to be correlated with the high percentage of consanguineous marriage.


The cultural concepts concerning parenthood are most directly linked to the ideology and organisation of descent groupings. The child grows inside the female who nourishes the child internally with her toto ‘blood’. So the saying that “the blood of the child is the mother's blood” is a Pukapukan statement of fact to be literally interpreted. As blood determines to a large extent the character of the child, particularly his modes of activity in sport and sexual behaviour, it is this blood inheritance, the “strength and skill of the living body”, that issues from and belongs to the mother and matrilineage.

Should a woman's desire for either a child or its genitor be weak, she would not conceive in the first place, or the foetus would miscarry, or the child would be stillborn or die young. If a husband did not behave according to his wife's expectations, she could loimata ‘curse’ him with one of these eventualities. This is the Pukapukan correlate to the sister's power over his brother's issue reported for Tokelau and Samoa. 37 Kinsmen can also curse in Pukapuka, usually dead parents or grandparents who do so to punish a child's disobedience, but there is no mention of sister's or father's sisters cursing. A curse may bring sickness or death to the child or his children, or may cause a woman to bear only sons, or a man only daughters, so that they have no one to tend their swamp or coconut land, respectively. However, only a wife can specifically affect fertility.

The mother has primary responsibility for the child's life, not only in utero, but throughout its life: “She nourishes the child from her own body. That is why she is responsible for its life.”

- 194

While the mother was specially responsible for the child's life, the matrilineal sublineage as a whole was accountable to other groups for the child's actions. In the case of anti-social behaviour by a sub-lineage member, specifically homicide, the onus fell on the matrilineal sublineage to make restitution. 38 I would infer the crucial point to be that a group member had taken a life from another group, and his group should forfeit in turn.

The female remains sexually strong and the “strength and skill of the living body” belong to the mother and matrilineage; the male ‘dies’ at climax and the child belongs fully to the father and ‘burial lineage’ only after its death. The matrilineage controlled the lifetime of children, on the island; the ‘burial lineage’ claimed the mauli ‘soul’ for the afterlife, transcending the island and linking the generations through men. 39

The sociological father, or the genitor especially if he eventually hopes to claim the child, is expected to provide for the child from outside, by feeding the mother during her pregnancy. When the child is born, the sociological father claims the wenua ‘afterbirth’ and buries it in the wenua ‘(dry) land’ of his ‘burial lineage’, although not actually in the cemetery. A tree is usually planted over the afterbirth. Informants say that “the tree will grow well when planted over the afterbirth.” They explain that the connection is not one between the life of the tree and the life of the child, 40 but rather that the afterbirth is richly nurturant which appears to indicate a broader connection between ‘wet’ fertility and the ‘dry land’, and to imply that the ‘wet’ fertility of females must be introduced into the dry land if there is to be continuity between the generations.

Each ‘cemetery’ is associated with a ‘burial lineage’. One should be buried in the cemetery on the dry land of his sociological father. If a cemetery is situated in a shallow place such that graves begin to be dug into the ‘wet’, it should be relocated. One such cemetery is said to have been relocated three or four times. Yet the “chief's grave, unlike the grave of an ordinary person, was dug very deep, down to water level” and a special chant was recited as the bearers slung the body into the grave. 41 This singular burial in the ‘wet’ appears to associate the chief with the female domain, an association which I will treat more fully below.

There is little detail about the traditional view of death and the afterlife. The mauli ‘soul’ belonged to the father and the burial lineage in some sense. The ‘soul’ is present in the living body, but tied to it, and is liberated only in dreams or death. In dreams as well as death the ‘soul’ may journey to the ‘underworld’. 42

Each is associated with a definite locality, a strip of land and a cemetery. The wua and kainanga were associated with no particular place, - 195 having strips of taro swamp here and there in the centre of the island. Burial lineage activities marked boundaries: creating ties of continuity across generational lines, maintaining contact with the gods and the underworld through the burial lineage chiefs and priests and defending the island by challenging intruders from outside. Matrilineal concerns were with the lifetime, and with affairs internal to the island; for example, the intra-island fighting and homicide restitution were asserted to have been their province. 43

Matrilineal affiliation is normatively unchangeable; burial lineage affiliation is expressly subject to change. The mother-child relationship is seen as congenital; the father-child relationship is seen as conditional and essentially developmental. Maternal nourishment is a sharing of substance. Conduct follows without further actuation. This behaviour is itself a shared identity, the wua ‘skill or style’ of the particular matriline. Paternal care involves external, volitional action and requires action in return, the exchange of burial for nourishment and care to complete the transaction. The mother-child relationship presupposes shared identity and sharing behaviour. The father-child relationship depends on a balanced exchange.

While there is no elaborate genealogical tradition, efforts to delineate genealogical connections are primarily concerned with linkages to and through males, understandable when the necessity of creating links through burial is understood. In contrast, matrilineal continuity may be assumed and genealogy is not necessary as a mnemonic device or a charter.

The cultural premises about maleness and femaleness relate to a broader worldview in which the sexual associations of fertile wetness and lifeblood with the female, of death, burial and inter-generational continuity with the male are acted out in the disposition of the child to his parents as female fertile lifegiving is complemented by male death and burial. Formally the burial lineages maintained contact across the perimeters of Pukapukan physical and spiritual existence; the matrilines were concerned with activities interior to these boundaries, in mundane and secular contexts.


Siblings of opposite sex reciprocally referred to each other as kainga, or a brother referred to his sister as tua wawine and she to him as tua tane. Brother and sister should share freely according to the norms of kinship, that is, they should share everything except sexual favours. Kainga should not have sexual intercourse with or marry each other, nor should they discuss their love affairs, but otherwise they may interact without restriction.

Generally the brother-sister relationship is one of ease, lacking the pronounced avoidance/reserve/respect/honour component found in some other Polynesian societies. As married adults the relationship is primarily one of informal reciprocity, e.g., the brother occasionally sends fish or coconuts to his sister, and she provides taro from swamps to which her - 196 brother has no other access. When adult brother and sister are living in the same household, usually without the spouse of one or the other, which is an acceptable and common arrangement, they each provide from their respective domains for their household.

The oldest child of a sibling set is distinguished as the uluaki ‘first-born’ and should serve as the steward of all the lands of the sibling set and be looked to for guidance and be treated with respect by his younger siblings. This position is clear and straightforward when the eldest is male, and there is a presumption that the eldest child should properly be male, which is reflected in the practice of naming first all males in a sibling set in order of age and then all females. Even when, in collecting genealogies, I asked that siblings be named in exact order of birth, very few females were given as first-born children. The significance of this bias in ordering lies in the reluctance to associate females with primogeniture, the implications of which are more explicit in the situation of the chiefly sister and in the nature of prohibited cousinship.


In line with the general lack of importance attached to the relationship of brother and sister, the Pukapukan treatment of the chiefly sister was such that she has never been accorded ethnographic notice as a kinswoman, but has instead been treated exclusively as a political status symbol.

The Pukapukan mayakitanga ‘sacred maid’ was ideally the eldest daughter of the aliki ‘chief’ or, if necessary, another girl of the chiefly burial lineage. 44 Each of the chiefs of the four chiefly burial lineages had the right to designate a ‘sacred maid’, who was initiated before puberty, at about 10 years of age. She never married, remaining a guarded virgin for life, and she retained her title even if the father died and the succeeding chief named another sacred maid. Thus, a ‘sacred maid's’ relationship to the incumbent chief may have been either eldest daughter, eldest sister, or possibly even father's sister. In the context of high rank, mayakitanga was a kin term, or in any case a term for an important kinswoman.

The ‘sacred maid’ was said to be a “symbol of the power and dignity of the lineage”; 45 yet her great sanctity was coupled with the utmost passivity. She was thought in some way to ensure prosperity, but by no obviously active means, and indeed she seems to have done so by serving as the passive channel for the good will of the gods. The only specific function of the ‘sacred maid’ recalled by the Beagleholes' informants was that she accompanied her lineage's voyaging canoe in order to protect it, apparently not by actively supplicating the gods, but rather by simply encouraging their beneficence through her presence. She was the only female who was allowed to pass into the male domain of ocean-going canoes.

Goldman suggests without elaboration that the descent line of the ‘sacred maid’ was “cut off for the sake of economic continuity.” 46 I - 197 would further interpret the institution of the sacred maid not just as necessary to assure prosperity in strictly economic terms, but also as a statement about the vulnerability of the chief who straddled categories of sacred and secular power and was thus placed in a dangerous as well as a powerful position. The forfeiture of the secular and sexual productivity of the ‘sacred maid’ in order that she become a singular channel of sacred power was particularly necessary in order that the chief might command a secure position in the secular domain.

Shore's analysis in terms of formal and instrumental power in Samoa may provide some insight. He suggests that “Formal power derives its force from position, or dignified form and . . . instrumental power (is) associated with movement, action and utility.” 47 A diagrammatic presentation of the distribution of powers in Pukapuka might look like this:

Distribution of powers.

The chief was more active in secular affairs than the ‘sacred maid’, yet less so than the tūpele, the ‘elders’. All three statuses were based on ‘burial lineage’ affiliation. Beyond this the ‘sacred maid’ and elders were selected on the basis of single principles. The ‘sacred maid’ was supposed to be an eldest daughter and the most beautiful girl of her lineage, though it is likely that she was presumed to be the latter by virtue of her selection. The elders were the oldest men, and the oldest among them was the wola, the chiefly executive who also served as a priest. In contrast, the selection of the chief was based on several principles, on ascribed and achieved secular qualities and on sacred recognition. He was supposed to be of the most senior descent within his lineage but had to demonstrate ability as well. His selection had to be confirmed by the gods, by a rainbow sign in the case of one lineage, by breaking out in goose flesh, or becoming subject to hysteria or trance in the case of other lineages. “If the reign of the new chief was good and prosperous, the people knew that they had correctly interpreted the sign of the god.” 48

- 198

But prosperity in the sense of favourable notice by the gods was also dependent on the ‘sacred maid’. On the side of instrumental power and secular control, the elders held the unambiguous status. “The governing of the island . . . really lay with the group of old men who because of their age and presumed experience were considered better able to order affairs than a chief or his executive.” 49 The instrumental power of the chief varied according to personality, the support of his followers, and presumably his rapport with the elders. Successful chieftainship would have depended on a balance of formal and instrumental powers, with the chief most firmly aligned with the formal and sacred, epitomised by his ‘sacred maid’. While ordinary individuals who displayed super-human powers were normally banished, “it is usual to find the chief displaying supernormal powers, bestowed by the god of his lineage.” 50 However, the only chief reported to have exercised his secular power in a cavalier fashion was deposed. 51 Thus the chief is most clearly associated with the gods, but through the symbolism of the sanctified female, the mayakitanga. I would suggest that her most unambiguous sacred status as his representative freed him for more secular manoeuvring.

While I initially thought that the ‘sacred maid’ was named before puberty simply because she could not go through the rite of initiation which probably included a sexual act, it seems that while she did not undergo a full or ordinary initiation, she did undergo a partial one, remaining in the liminal phase in perpetuity. 52

Goldman notes that the initiation of the Pukapukan ‘sacred maid’ resembles that of chiefs elsewhere in Polynesia. 53 From the meagre information on Pukapukan ritual, her initiation was similar to that of Pukapukan chiefs, in that she was also lifted up and carried into the sacred enclosure. The ritual lifting of the ‘sacred maid’, literally from the female domain to a sacred status in the male domain, symbolically divested her of her sexual role. 54 But when she was carried to the sacred enclosure of her burial line, she was set down with the chant used at the burial of a chief. Once initiated she lived in seclusion for the rest of her life near her ‘burial lineage’ cemetery, although she might join more in mundane activities were another ‘sacred maid’ named before her death. She was fed rich food and was wakayawi ‘put on public display’ every few months. 55

The seclusion, fattening and display of the ‘sacred maid’ are clearly similar to the most elaborate general initiation, the kaitau wakayawi. New adults of each ‘burial lineage’ were secluded in the wale kaitau, a darkened house, for a month, fed fattening food, and then displayed for the admiration of the village. Those initiated in this manner were expected to make good marriages: “Only a man who was a good fisherman and - 199 rich in nuts and taro could hope to get a fat, fair wife, and only a fat woman would be able to get a kaitau husband.” 56 Yet, while the fatness and fairness of general initiands make them especially desirable spouses, the ‘sacred maid’, who would be the most fat and fair from her perpetual seclusion, is barred from marriage.

In contrast to the general initiation, the ‘sacred maid’ “died” at her initiation and did not go through any phase of re-aggregation or “rebirth.” She lived by her cemetery for the rest of her life. She did not go away as other sisters tend to do; in fact, she would have been the most localised of all individuals, male or female. The nature of her liminal status is clearly that she was the only person to belong explicitly to her ‘burial lineage’ during her lifetime, so that there was no ambiguity about her burial lineage affiliation. She is associated with the male domains of access to sacred power, burial, and defence on the ocean perimeter. Unfortunately, there are no data on her death and burial, but perhaps it is relevant that while she was in a sense “buried” above ground during her lifetime, the chief was buried below ground, that is, below water level, at his death. Her association with the male and sacred domains was complemented by his association with secular power and the female domain.


While there is no term other than matua tane for mother's brother, there were two ancient terms that focused on the relationship between a man and his sister's children: ilamutu and yinākava.

Ilamutu was used for “a man's adopted child (tama kokoti) of either sex who is the blood child of a female blood relative of his own generation” and yinākava was apparently used specifically for the daughter of a man's sister. 57 The Beagleholes were dubious about the specificity of the usage of yinākava for a man's sister's daughter:

The word yinakava is archaic and is found only in old chants. It is applied to anything or anybody which is tapu. Hence the god house on a double canoe or a sacred structure is yinakava. Similarly one who stands in the relationship of cousin-in-avoidance to the speaker is yinakava to the speaker. A man might call his cousin-in-avoidance my yinakava, but only under stress of emotion as in mourning ceremonies; in a love chant one's cousin-in-avoidance may be addressed, by poetic metaphor, as yinakava. 58

A sister, the ‘sacred maid’ and sister's ‘sacred’ issue suggest the Oceanic perspective on the sister or the line issuing from her as “spiritually pre- - 200 dominant”. 59 But what is striking in the case of the mayakitanga and ilamutu is that there were no descendants of the sister. The ‘sacred maid’ remains a virgin and the ilamutu is adopted into his mother's brother's line. This would have its greatest meaning when viewed in the context of high rank. If the sister's line were viewed as “superior”, the institution of the sacred maid or the adoption of the sister's issue would have prevented any line from being superior to that of the chief. 60

This seems to be carrying Polynesian rank a little too far for an apparently egalitarian atoll society. However, a look at cousins in non-chiefly strata reveals that children born of senior sisters may have been significant and problematic. This provides further support for the view that the ‘sacred maid’ carried a strategic symbolic load for the rest of Pukapukan society.


Avoided cousins of opposite sex were called kainga wakamā and wale atua. Burrows included Pukapukan “cousin-avoidance” in his column

Diagrammatic representation of the Beagleholes' description of Pukapukan sibling and cousin terminology (after E. and P. Beaglehole 1938:259).
- 201

for brother-sister avoidance, noting that there is no brother-sister avoidance as such in Pukapuka. 61 And indeed the restrictions between kainga wakamā were similar to the types of avoidances that apply to brother and sister elsewhere in Polynesia. Kainga wakamā could not eat from the same platter, although they might have spoken to each other if it were absolutely necessary to do so. 62 However, the restrictions between wale atua were more severe; they avoided each other completely and spoke only through intermediaries. 63

These behavioural and terminological distinctions suggest that these were two types of kin. The first pattern the Beagleholes present to account for this distinction is diagrammed in Figure 2.

First cousins of opposite sex whether cross or parallel were reciprocal kainga wakamā. Second and more distant cousins of opposite sex were invariably reciprocal wale atua, If this formulation correctly represents the traditional system, then the terms and other behaviours correspond fairly simply with generational removal from common ancestry, and the pattern of progressively strict avoidance would seem to maintain the exogamous nature of the cognatic stock. The labelling of kinsmen also highlights what is indubitably the most solidary dyad: same sex siblings and cousins are reciprocal taina whatever the relationship of intervening generations.

But the Beagleholes suggest a second formulation to account for the cousin terminology. They note that some informants thought the “wale atua relationship arose only when there was a first-born female with brothers and sisters in the generation level from which descent was traced.” This second formulation is diagrammed in Figure 3.

They hold that this

. . . characterization would have more weight if there were other patterns in Pukapukan society pointing to the importance of the first-born female in the formulation of kinship patterns and practices. Since there are no such patterns, it is best to regard Veti's belief as incorrect and merely an example of the confusion that holds today regarding kinship structure. 64

It is here that the Beagleholes' division of “politics” from “kinship” has done their analysis an injustice, for even if they failed to consider her in their pages on kinship, they had already discussed the mayakitanga as an important first-born female. In this chiefly case, the first-born

- 202
A second formulation of cousin terminology (after E. and P. Beaglehole 1938:260).

daughter was a ‘sacred maid’, and had no issue. In the case of other lines, a first-born sister, while not in herself sacred, gave rise to ‘sacred’ issue in the wale atua ‘god house’ relationship. If there was no eldest sister, the descendants of the sibling set were only kainga wakamā to each other, subject to a mild avoidance that maintained the integrity of the cognatic descent category for as long as common ancestry was remembered.

Perhaps the practice of ignoring birth order in genealogy, listing all males first, was one way of obviating the stringent avoidance of the wale atua, which one informant protested would have been an unpleasant way to have to behave. But were the logic of these relationships taken to its extreme, making every first-born female a ‘sacred maid’, the practice might not have been merely unpleasant but also impossible to maintain in a restricted population.

While the first formulation seems to be based on the maintenance of exogamous cognatic stocks, the second formulation is concerned more directly with the interplay of sex-line duality and senior-junior relationships. The crux of the issue may be that wale atua had a range of reference, including ‘god house’, ‘issue of first-born female’, ‘distant cousin of opposite sex’. The metaphoric links between these referents may be more substantial than I have been able to demonstrate.


With these further data on male and female symbolism and statuses, the Pukapukan origin myth can be subjected to more illuminating analysis. The initial filiative link of male-male with the gods taking on paternal roles in relation to Mataliki and the initial male-female dyad of male autochthone-female outsider are represented in Figure 4 below.

The original chiefly line of Pukapuka derives from an incestuous union between the older son and older daughter of Mataliki and Te Vaopupu. A second incestuous union, between their younger son and younger daughter is seen as the origin of the matrilineal moieties. Thus, in the myth, two brothers and two sisters become instead two husband-wife dyads (see Figure 4):

- 203
The Pukapukan Origin Myth (adapted from E. and P. Beaglehole 1938:221-4, 240-1, 375-7).

To make some sense out of the myth, the implicit mayakitanga ‘sacred maid’ has been added although she is not taken account of in the text. The chiefly line replicates the initial male-male filiative link (gods-Mataliki) by sacralising the sister, and thus not only obviating incest, but also eliminating sister's children as progeny of another male. Because of the yanga kikino ‘evil works’ of their parents, said to be all this incestuous propagation, Mataliki struck the land with lightning and forced the children of Punga and Punga Momoto to flee. Those who sheltered from the storm on the land became the matrilineal moiety the Wua Kati or Land Creatures. The rest sheltered in the sea and became the other matrilineal moiety the Wua Lulu or Sea Creatures. Through his punishment Mataliki divided his grandchildren into the matrilineal moieties which presumably were initially exogamous categories.

Kati also means ‘fishing line’, is associated with uila ‘lightning’ and has connotations of maleness. Lulu like wua has the denotation of ‘female generative organs’. The Wua Lulu has strong connotations of femaleness and is said to be “more wua” than the Wua Kati. Because of this, the moieties are often referred to in discourse as the “Wua Lulu and Kati”, leaving the wua off the Wua Kati.

This division also replicates the initial male-female pair for the Wua Kati are sometimes regarded as autochthones and the Wua Lulu as outsiders, or later comers to the island. The usual male-female, outside-inside, sea-land contrast appears to be turned about in the associations of the matrilineal moieties but the contradiction is at least partly resolved when it is realised that the genesis thus takes further account of the male autochthone of Pukapuka and his foreign bride, the female from across the sea.

- 204

The origin myth is also formally congruent with the cousin terminology as given in the first formulation. Brother and sister are incestuous in the myth and behaviourally unrestrained. Separation occurs in the second generation in the myth as does the avoidance relationship arise between cousins.

On each side of the genesis, chiefly and matrilineal, there are both males and females. On one side the male-male regeneration of the aliki is guaranteed by the symbolic fertility of the mayakitanga, and by the implication that if she were to bear children their sanctity would threaten the status of the chiefly line. On the other side the stewardship of male elders of matrilines is dependent on their reproductive sisters. While the brother-sister relationship is finally submerged, it is nevertheless implicit and powerful on both sides of the duality of and wua in symbolic and social organisation.


Pukapukan cultural emphasis falls most obviously on husband and wife, and particularly on their roles as father and mother. Huntsman and Hooper have pointed out that the husband-wife emphasis appears to accord with Burrows' distinction between Western and Central Polynesia: “The western center has richer terminology for collateral relatives. . . . The central center has richer terminology for affinal relatives.” 65 Congruent with a diversion of emphasis from the sexually taboo brother-sister dyad, Pukapukan oral traditions include love stories and chants with explicit sexual content. In contrast, in Tokelau where the cultural emphasis is clearly on the brother-sister dyad, the oral tradition is notable for its lack of sexual content. 66

But in addition it has been shown that the Pukapukan mayakitanga, ilamutu, yinākava, kainga wakamā and wale atua relationships are part of a solution to the problem of the brother-sister relationship, specifically addressing the issue of the senior sister. Reintegrating these terms into the discussion of male and female in Pukapukan culture has demonstrated that the cultural focus is not restricted to the husband-wife dyad.

Polynesian male/female dualism has a number of possible permutations, including the Tokelau emphasis on non-sexual aspects of male-female complementarity and the Pukapukan focus on specifically sexual aspects. In addition, the senior-junior authority axis, most unambiguous in its application to male-male relationships, may receive further cultural elaboration where the problem of the senior sister is taken into consideration. 67

As we move further east in Polynesia, what appeared as essentially “kin” relationships tend to be transformed in new ritual and political contexts. The example of the mayakitanga suggests that a sensitivity to - 205 cognate forms coupled with a suspicion of arbitrarily isolated institutional contexts should enrich pan-Polynesian comparison.


While I was doing the research on which this paper is based, plans were under way for an Association for Social Anthropology in Oceania monograph on male and female in Oceania. Although that project was abandoned, I should like to thank Dr Jane Goodale for the initial impetus to address these issues.

The paper is a revised and expanded version of portions of my dissertation (Hecht 1976), based on research which was generously supported by the National Science Foundation. My dissertation committee, Dr David M. Schneider, Dr Marshall D. Sahlins and Dr Michael D. Lieber offered inspiration and helpful criticism while I was writing.

I spent 1976-77 as a Visiting Lecturer at the University of Auckland and gave this paper in substantially its present form at a staff-student seminar. Dr Judith Huntsman and Dr Antony Hooper were particularly forthcoming with comparison and comment. The paper is intended partly as a companion piece to their “Male and Female in Tokelau Culture” (Huntsman and Hooper 1975). Mr Michael Goldsmith and Mr Richard Parmentier also made valuable editorial and critical suggestions on a later draft.

The Bernice P. Bishop Museum, Honolulu, has kindly allowed me to quote from the Beagleholes' unpublished manuscript (Beaglehole n.d.). Some of the “Myths, stories and chants from Pukapuka” are available in translation in the Human Relations Area File. I should also like to thank Dr Jeremy Beckett who generously presented me with the genealogies he collected in Pukapuka.

I am grateful to the Pukapukan people who have helped me with their interest and patience, and to my colleagues who have urged me to address my data to issues of wider significance. Of course I am responsible for the paper in its final form and for any errors in fact or analysis.

  • BEAGLEHOLE, Ernest and Pearl, 1938. Ethnology of Pukapuka. Honolulu, Bernice P. Bishop Museum Bulletin 150.
  • —— n.d., unpublished manuscript collection of myths, stories and chants from Pukapuka, on deposit in the Bernice P. Bishop Museum Library, Honolulu.
  • BURROWS, Edwin G., 1938. Western Polynesia: A Study in Cultural Differentiation. Goteburg, Elanders Boktryckeri Aktieborg, Ethnologiska Studier 7, Gothenborg Ethnografiska Museet.
  • GIFFORD, Edward Winslow, 1929. Tongan Society. Honolulu, Bernice P. Bishop Museum Bulletin 61.
  • GOLDMAN, Irving, 1970. Ancient Polynesian Society. Chicago, University of Chicago Press.
- 206
  • HECHT, Julia, 1974. From Conception to the Grave: Double Descent in Pukapuka. Paper presented at American Anthropological Association meetings, Mexico City.
  • —— 1976. Double Descent and Cultural Symbolism in Pukapuka, Northern Cook Islands. Ph.D. dissertation, University of Chicago.
  • HOOPER, Antony, 1970. “Socio-economic Organization of the Tokelau Islands,” in Proceedings of the 8th International Congress of Anthropological and Ethnological Sciences, Tokyo.
  • HUNTSMAN, Judith, 1969. Kin and Coconuts on a Polynesian Atoll: SocioEconomic Organization of Nukunonu, Tokelau Islands. Ph.D. dissertation, Bryn Mawr College.
  • —— 1971. “Concepts of Kinship and Categories of Kinsmen in the Tokelau Islands.” Journal of the Polynesian Society, 80:317-54.
  • HUNTSMAN, Judith and Antony HOOPER, 1975. “Male and Female in Tokelau Culture.” Journal of the Polynesian Society, 84:415-30.
  • HUTCHIN, J. J. K., 1904. “Traditions and some words of the language of Danger or Pukapuka Island.” Journal of the Polynesian Society, 13:173-6.
  • MABUCHI, Toichi, 1960. “The Two Types of Kinship Ritual among Malayo-Polynesian Peoples,” in Proceedings of the 9th International Congress for the History of Religions, Tokyo.
  • —— 1964. “Spiritual Predominance of the Sister,” in Allan H. Smith (ed.), Ryukyuan Culture and Society: A Survey. 10th Pacific Science Congress Series, Honolulu, University of Hawaii Press.
  • MACGREGOR, G., 1935. Notes on the Ethnology of Pukapuka. Honolulu, Bernice P. Bishop Museum Occasional Papers, Vol. 11, No. 6.
  • MEAD, Margaret, 1930. Social Organization of Manua. Honolulu, Bernice P. Bishop Museum Bulletin 76.
  • PANOFF, Michel, 1970. La terre et l'organisation sociale en Polynésie. Paris, Payot.
  • SAHLINS, Marshall D., 1976. Culture and Practical Reason. Chicago, University of Chicago Press.
  • SCHNEIDER, David M. 1970. “What Should Be Included in a Vocabulary of Kinship Terms?” in Proceedings of the 8th International Congress of Anthropological and Ethnological Sciences, Tokyo.
  • SHORE, Bradd, 1976. “Incest Prohibitions and the Logic of Power in Samoa.” Journal of the Polynesian Society, 85:275-96.
  • TURNER, Victor, 1964. “Betwixt and Between: The Liminal Period in Rites de Passage,” in J. Helm (ed.), The Proceedings of the American Ethnological Society. (Reprinted as Chapter IV of V. Turner, The Forest of Symbols: Aspects of Ndembu Ritual. Ithaca, Cornell University Press, 1967).
  • VALERI, Valerio, 1972. “Le fonctionnement du système des rangs à Hawaii.” L' Homme, 12:29-66.
  • VAYDA, Andrew Peter, 1958. “The Pukapukans on Nassau Island.” Journal of the Polynesian Society, 67:256-65.
1   Two versions of the myth are given in E. and P. Beaglehole 1938:221-4, 375-7, and one of the versions is also in Hutchin 1904.
2   Huntsman and Hooper 1975:423 (footnote 9).
3   E. and P. Beaglehole 1938. The Beagleholes' fieldwork during 1934-35 in Pukapuka is reported primarily in the Ethnology of Pukapuka (1938). I spent thirteen months in Pukapuka, concentrating my efforts on the cultural symbolism of kinship and land tenure (Hecht 1974, 1976). Unless otherwise noted, all data in this paper are from my fieldwork. The Beagleholes' orthography is generally followed; the only deviation from their usage being my indication of vowel length with a macron. Stress is normally on the penultimate syllable. Very little has been written in the Pukapukan language and there are as yet no set conventions.
4   The 1250-acre estimate is the Beagleholes' (1938:17). The 1800-acre estimate was made by James Gosselin, Resident Agent in Pukapuka from 1972-74, and was based on a map paced by Mr Gosselin and Mr Pareura Katoa, the Agricultural Officer.
5   See Vayda 1958.
6   It might be appropriate to speak of cumulative patrifiliation through burial rather than patrilineal descent. But Pukapuka has been labelled “double descent” for so long that I hesitate to jolt the idiom in the context of the present paper and will let the significance of burial speak for itself for now. I have used the gloss ‘burial lineage’ to indicate that I am referring to the same group that the Beagleholes glossed as ‘paternal lineage’.
7   E. and P. Beaglehole 1938:225, 229.
8   E. and P. Beaglehole 1938:4.
9   E. and P. Beaglehole 1938:232-3.
10   E. and P. Beaglehole 1938:221, 229.
11   E. and P. Beaglehole 1938:237-9, 256-63.
12   Although sponsorship for village membership was mentioned in the Ethnology of Pukapuka (E. and P. Beaglehole 1938:219), the tuanga kai was not specifically noted. The tuanga kai is described in my dissertation (Hecht 1976); it is roughly comparable to the Tokelau inati (Huntsman 1969; Hooper 1970).
13   E. and P. Beaglehole 1938:227.
14   E. and P. Beaglehole 1938:256-63.
15   See Schneider 1970 which presents some problems of demarcating a vocabulary of kinship terms.
16   E. and P. Beaglehole 1938:309.
17   E. and P. Beaglehole 1938:226-7.
18   Beaglehole n.d., p. 62a.
19   E. and P. Beaglehole 1938:386. Marshall Sahlins (personal communication) has pointed out that the supposed “seismic wave” may mask a political upheaval: of the 15 men and parts of their families said to have survived the wave, seven of the men were a chief and six of his sons, who effectively divided the island among themselves following the disaster. From another angle, I suspect that the sexually insatiable Wanguna of the Pukapuka wave may be cognate to Anuna, the cannibal-ogress in one of Judith Huntsman's as yet unpublished collection of Tokelau folktales.
20   Poloaki is probably from the Cook Islands Maori poroaki; the terms yau and ati (E. and P. Beaglehole 1938:386) are no longer recognised.
21   E. and P. Beaglehole 1938:263-4.
22   E. and P. Beaglehole 1938:264.
23   See below on blood and matrilineal identity.
24   E. and P. Beaglehole 1938:295-6.
25   E. and P. Beaglehole 1938:295-6.
26   E. and P. Beaglehole 1938:284.
27   Wale atua, see below.
28   E. and P. Beaglehole 1938:261.
29   E. and P. Beaglehole 1938:295-6.
30   In contrast, “In 1971, of the 718 Tokelauans 20 years and older, only 14 (less than 2 percent) had been permanently separated or ever divorced” (Huntsman and Hooper 1975:423n).
31   See below on the association of and locality.
32   See note 62 below.
33   Burrows 1938.
34   Today Cook Islands Maori kin terms, especially those for siblings and cousins, are used to the virtual exclusion of traditional Pukapukan terms.
35   Tokotoukau is heard for ‘your (plural, exclusive) spouse (literally person)’. The singular forms (tokukau ‘my person’ and toukau ‘your person’) are not grammatical.
36   E. and P. Beaglehole 1938:257.
37   Huntsman and Hooper 1975:424; Mead 1930:137.
38   E. and P. Beaglehole 1938:228.
39   Goldman (1970:384) appears to have intuitively reached the same analysis: “The wua speak for the more general and real, for the more diffuse quality of birth; the po, for the more specific and indeed focal continuity of a line as it joins its ancestry, and occupies an immovable site in a cemetery.”
40   Compare E. and P. Beaglehole 1938:270.
41   E. and P. Beaglehole 1938: 302-3.
42   E. and P. Beaglehole 1938:325-30.
43   E. and P. Beaglehole 1938:228, 374, 394.
44   Mayakitanga is cognate to the Tongan, Uvean, Futunan and Tikopian terms for FZ and the Niuean term for Z.
45   E. and P. Beaglehole 1938:239, see also Mead (1930:14) on the Samoan taupou “female ornament of the chief's rank”.
46   Goldman 1970:386.
47   Shore 1976:284 ff. While analysis in terms of a continuum between formal and instrumental power seems to serve the Pukapukan data better than a series of contextualised oppositions like Shore's for Samoa, this may be an artefact of insufficient data or speak to the need for a closer analysis.
48   E. and P. Beaglehole 1938:240, 244.
49   E. and P. Beaglehole 1938:246.
50   E. and P. Beaglehole 1938:245.
51   E. and P. Beaglehole 1938:391-3.
52   See Turner 1964.
53   Goldman 1970:385.
54   E. and P. Beaglehole 1938:237-9.
55   E. and P. Beaglehole 1938:238.
56   E. and P. Beaglehole 1938:282.
57   E. and P. Beaglehole 1938:258-9. An adopted child was supposed to be treated as if it were a natural child and ties with natural kin lapsed. However, in almost all cases the relationship between natural and adoptive parents was close and the relationship between the adopted child and its natural parents was only slightly altered. Matrilineal sub-lineage affiliation was very rarely changed so that the adopted ilamutu would probably be in the same sub-lineage as its natural mother and adopting mother's brother, and would presumably be buried in the same cemetery as its natural mother.
58   E. and P. Beaglehole 1938:259.
59   Mabuchi 1960, 1964.
60   See Valeri 1972, which considers similar, if more complex, issues in the context of Hawaiian rank, and also Gifford 1929 on the Tongan Tui Tonga Fefine and Tamahā.
61   Burrows 1938:60.
62   Some of my informants think kainga wakamā is a general term for ‘brother-in-law’ and, as noted above, today a man may not wrestle with his brother-in-law. According to the Beagleholes, a man and his kainga wakamā's husband were not allowed to wrestle, but they made no mention of any restriction regarding the sister's husband. Professor Bruce Biggs (personal communication) points out that this may involve some native etymological confusion between two Proto-Polynesian roots: PPN *maa ‘shame’ and PPN *ma‘aa ‘cross-affine’. Wrestling is still the focus of interest at the end of the yearly sports competitions during the Christmas and New Year season.
63   E. and P. Beaglehole 1938:259-63.
64   E. and P. Beaglehole 1938:260. Hutchin (1904:176) and Macgregor (1935:18) both give wale atua as a term for the daughter of an older or oldest brother. The Beagleholes do not mention this and apparently found no support for it.
65   Burrows 1938.
66   Huntsman and Hooper 1975.
67   See Panoff 1970; Goldman 1970; and Sahlins 1976, each of which treats in varying manner and degree the issues of male/female duality and the senior-junior authority axis in Polynesia.