Contesting Marshall Sahlins on Kinship



In a recent two-part article on the nature of kinship, Marshall Sahlins maintains that performative criteria for kin-reckoning are at least as salient as procreative ones, and that, at conception, an individual is endowed with a wide circle of kin, including the ancestral dead. For both reasons, he argues, there is no warrant for granting privileged status to what anthropologists have called ‘primary’ kinship. The contentions here are that performative criteria are modeled upon procreative ones; that ties to ancestral figures are seen as antithetical to procreative ties; and that, therefore, all kinship constructs are derived from nuclear family relationships. Evidence in support of these contentions is provided from the Mae Enga, Fiji, the Trobriands, and Aboriginal Australia.

A recent brief contestation between Marshall Sahlins (2011, 2012) and myself (Shapiro 2012) raises questions of fundamental importance to kinship theory. This being so, it seems to me to be worth at least a full-length article, not just an abbreviated exchange of remarks.

The gravamen of the debate comes to two considerations. First, Sahlins argues that what have been called performative criteria in kin reckoning – e.g. those based upon co-residence, or commensality, or name-sharing – are at least as salient as procreative ones in human communities, and that for this reason there is no warrant for granting privileged status to procreatively-derived relationships. My own contention, by contrast, is that there is every reason to do so, because the former are usually modeled upon the latter. Second, Sahlins argues that at conception, at least outside the West, an individual is endowed with a wide circle of kin, including not only the living but, as well, the ancestral dead; and so procreatively-derived relationships, far from being the building blocks of other kinship notions, are imbedded in them and are somehow – Sahlins does not say exactly how – secondary to them. I have absolutely no quarrel with the first statement, but the second, I submit, involves a serious distortion of the ethnographic record. In particular I shall argue that notions of ancestral kinship are viewed as antithetical to those involving primary kinship and are thus also derived from them. In the course of my arguments, I shall emphasize some of the ethnographic cases he does, not only in his exchange with me but, as well, in his well-known commentary on sociobiology (Sahlins 1976).1

More specifically, the following section focuses on Mervyn Meggitt's outstanding analyses of sociality among the Mae Enga of Highland New Guinea (see following references), because these provide Sahlins with some of his favourite ethnographic examples (e.g. Sahlins 1965:107, 1968:105–08, 1976:32). There I deal with both of his contentions. I also bring in Fijian materials pertinent to the first contention, because Sahlins has carried out fieldwork in Fiji and has provided us with an important ethnographic statement based upon this research (Sahlins 1962). In the two succeeding sections I focus upon the second contention.

Mae Enga Kinship Notions and Fijian Comparisons

In an important argument for recognition of a ‘third party’ in addition to the mother and the father in native ideologies of embryological development, Sahlins (2012:675–76) emphasizes Meggitt's report (1965a:163) that the Mae Enga assign considerable significance to ancestral spirits which are held to enter a foetus. Thus, as soon as a Mae Enga child comes into the world he or she is incorporated into a network of ascendant kin, in addition to contemporaneous kin other than his/her parents. Here are Sahlins' own words: ‘[W]hat is significant about the make-up of Enga children is the relative devaluation of the father's substantive contribution to the foetus in favour of the spiritual bestowal of the patrilineal clan ancestor – which is also to say the so-called primary kinship of fatherhood is secondary to the extended brotherhood of the clan’ (2012:675). Note that this statement contains its own contradiction, viz., if the ‘brotherhood of the clan’ is ‘extended’ – as, I shall show, it most certainly is -, how can it also be primary, as Sahlins implies it is? He clearly has a major problem with the conventional anthropological distinction between ‘primary kin’ and others, and he indicates this by regularly enclosing ‘primary’ in quotes. The implied argument is that the primacy of procreatively close kin is a Western concoction not shared by other people. But if this is so, how can we account for the fact that in Mae Enga kin classification such kin are said to be ‘true’ members of their respective classes – people whose membership in these classes is rendered by native idioms which Meggitt (1964b:193) translates as ‘without a doubt’ and ‘completely.’ This, it should be noted, makes them ipso facto primary, according to Mae Enga notions and those of the ‘structural semantics’ of Scheffler and Lounsbury (1971), something which Sahlins notes but plainly does not understand.

This analysis can be pursued further. In most cases, those who are members of their classes ‘without doubt’ in fact include entire local patrilineal groups (Meggitt 1964b:192–93). But note that membership in such a group is determined by one's father's membership and that, as we shall see shortly, membership in more inclusive patrilineal groupings is idiomised by paternal images of generation. But this is not all. Meggitt (1964b:193) notes that the ‘true’ subclasses vary contextually: sometimes only the genealogically closest members of their respective kin classes are allotted ‘true’ status, with others normally so allotted left out. Such contextual variation of focal membership in kin classes has been reported several times (e.g. Feinberg 2004:81; Marshall 1976:214; Shapiro 1981:38–40) and is in fact probably the usual pattern. More generally, the focal status of genealogically close kin is one of the most secured ethnographic findings we have, Sahlins (among many others) to the contrary notwithstanding; indeed, it is all-too-apparent in his own materials from the Fijian island of Moala. His words:

A … notable distinction within the kindred is between near and distant kin. … [H]ere Moalans have … categorical phrases which distinguish degrees of kin distance. They commonly made a distinction between “true relatives” … and distant, “relatives by descent” … Generally speaking, the offspring of one's own grandparents are “true relatives” … I have pressed people, moreover, into a further distinction between [primary kin] and [collateral kin]: the former [are] “very true relatives” … (Sahlins 1962:157).

James Turner, working elsewhere in Fiji, has more recently come to much the same conclusions:

The Fijian word whose meaning corresponds most closely to the English term kinship, veiwekani, can be used to refer to any sustained, friendly, and irrevocable relationship between two persons or groups. … [But t]he term veiwekani is sometimes used in a more restricted sense to refer to relationships based on the sharing of common substance (veiwekani dina, true relatives). Such relationships are understood to be a consequence of the process of conception (Turner 1991b:184, emphases added).

Several pages later in his exposition, Turner confirms the distinction found by Sahlins between ‘true relatives’ and ‘very true relatives’ (Turner 1991b:191). And he notes elsewhere (Turner 1991a:15) that, in the Fijian community he studied, an individual may be referred to teknonymously as ‘X's father’ or X's mother' – which is to say that parental kin terms in their unmarked form are taken to refer to the closest kin of their respective classes. As Scheffler and Lounsbury (1971:11–12) have pointed out, this is a recurrent phenomenon in the ethnographic record. Further, Capell and Lester (1945:177), surveying all the Fijian material available to them at the time they wrote, report expressions they translate as ‘distant kinsman’ and ‘distant kinswoman,’ and these seem to pertain to people who are not procreatively or affinally close.

Walter (1975:185) reports much the same, adding that the ‘distant kinsman’ class, in the part of Fiji in which he carried out fieldwork, may be in turn subdivided – into kin who are ‘a little far only’ (Walter 1975:188) and an apparently unlabeled residual subclass. The ‘little far only’ subclass refers to the children of cross-cousins (ibid.), which is to say that it is based upon procreative ties.

So Fijians seem to have a notion of what anthropologists conventionally call ‘the personal kindred,’ and this notion is entirely comparable with those found in the West. The same is true of the Mae Enga: Meggitt (1964b:192) tells us that their language has a pair of contrastive terms, one of which he translates ‘other people.’2 Elsewhere, and quite some time ago, Sahlins (1963:40) suggested that such ideas are universal, and I think he is very nearly right. But this is not at all his present position, which insists upon a sharp distinction between ‘a naturally given set of “blood” relationships’ – note the quotes surrounding this expression, suggesting that it is no more than a Western construction – and ‘a culturally variable system of meaningful categories’ (Sahlins 1976:22–23). By my lights this distinction is egregiously overstated, and Sahlins is entirely wrong in insisting upon it.

Sahlins does not tell us whether Fijian kin classification has an ‘own'/'classificatory’ distinction, as its Mae Enga analogue and so many others do. For this we have to thank Nayacakalou (1955:40). The latter also reports (1955:48) that, although this system generally follows the pattern conventionally labeled ‘bifurcate merging,’ the father's brothers are subclassed according to their ages relative to the father: his older brothers are called, more specifically, ‘big’ or ‘senior’ ‘father,’ whereas his younger brothers are ‘small’ or ‘junior’ ‘fathers.’ Similarly, the mother's sisters are subclassed according to their ages relative to the mother by the use of the same modifiers. Parents are thus the focal members of their respective kin classes. It is surely worth mentioning here, moreover, that Nayacakalou is an anthropologist who is also a native Fijian. Hence the conclusion that these semantic distinctions are not anthropological impositions on non-Western thought, and the suggestion that people around the world do indeed distinguish between primary kin and others.

Nayacakalou's findings are consistent with those of Thompson (1940:56–57), Geddes (2000 [1945]:66–67), and Ravuvu (1971:481), the last named also an anthropologist-cum-native Fijian. Geddes (ibid.) adds that one may distinguish his/her own children from others of the ‘child’ class by a modifier which he translates as ‘true’ (ibid.:67) – something which we have already encountered in regard to the Mae Enga. The same modifier may be used to distinguish actual siblings from parallel cousins, and first cross-cousins from those less closely related (Geddes 2000 [1945]:68).

In that part of Fiji in which Hocart (1929:33) worked, by contrast, there is no ‘big father’ or ‘big mother' term. Rather, all parents’ siblings, real and classificatory, are classed as ‘little fathers’ or ‘little mothers.’ That is, they are still defined relative to the parents but they are labeled as a semantically reduced version of them. This is such a common pattern worldwide that Murdock (1957), in his cross-cultural research, coined a special label for it, viz. ‘derivative bifurcate merging.’

Further evidence of such distinctions comes from those parts of Fiji with a matri-moiety system. Thus Capell and Lester (1945:192) report a special term ‘applied to all male members of the father's moiety and generation.’ In other words, one's father is the focus from which this kin class is derived, much as in English one's father provides a model for calling a Roman Catholic or Orthodox Christian priest ‘father.’ Capell and Lester (ibid.) also note that the father's elder brother rates a prefix which they translate as ‘great,’ presumably signifying his age relative to one's father, who perforce provides the focus of his kin class. They do not tell us, however, whether the father's younger brother is prefixed by a label translatable as ‘small,’ though this is often the case elsewhere in Fiji, as we have seen (see also Capell and Lester 1945:198). In this same area the ‘mother’ term is subject to the same modification, based on age relative to the mother, who perforce provides the focus of her kin class (ibid.) This too applies elsewhere in Fiji, as already observed.

Quain's findings, which are unusually rich, mostly support what we have already encountered. In the part of Fiji in which his fieldwork was carried out an adjective meaning ‘small’ may be appended to kin terms to signal certain classificatory as opposed to procreatively close members of a particular kin class, who perforce provide the foci of that class (Quain 1948:247, 280). Here he echoes Hocart. Quain (ibid.) provides an example: ‘ “Small” brothers are those of the same [matri-]moiety who have grown up in the same village.’ By contrast moiety mates ‘whose acquaintance is of short standing and who do not know what relationship term their parents applied to each other qualify their brotherhood by calling themselves “brothers in moiety” ‘ (Quain 1948:248). But Quain (1948:247) also notes that an adjective meaning ‘real,’ the opposite of ‘small’ in kin classification, is applied not only to the procreatively closest member of a kin class but, as well, to any other member with whom one has ‘[a] traceable biologic tie or tradition of established kinship.’ Later (1948:290) he includes, presumably in the latter category, those who are said to be ‘land-together'- i.e. people outside one's region who intermarry regularly with those within it. I suspect that what is involved here is rather like the contextual variation in focality we found in the Mae Enga case and with Sahlins' Fijian materials, in which ‘real’ subclasses can be subdivided into those who are ‘very real,’ pace Sahlins and Turner, and those who are, as it were, just ‘real.’ In any case, Quain (ibid.) also tells us that ‘[e]xtremely distant [kin-class] relationships may be qualified with … “big,” which refers to the scope of extension’ (emphasis added). This seems understandable, if ethnographically unusual, though elsewhere Quain (1948:244) suggests that genealogical knowledge is far too shallow to deal with procreative relationships beyond the grandparental level. Be this as it may, his material makes it clear that certain performative criteria – particularly those having to do with co-residence and historical relationship – supplement procreative notions, at least in this part of Fiji.

But there are limits to such supplementation. When a child is adopted its relationship with its natal parents is maintained, and it is said to be ‘the child of’ these parents – not, apparently (Quain's data here are not as detailed as one might wish), of its adopters (Quain 1948:293–94).

Also pertinent are Quain's findings in regard to the cross-cousin relationship. First cross-cousins are not supposed to marry, and, accordingly, may distinguish each other from others of the ‘cross-cousin’ kin class by an expression which he translates as ‘brother-sister cross-cousin’ (1948:268). In short, they liken their relationship to that between siblings, between whom marriage is absolutely forbidden.

Quain (1948:280) further notes that only one's real parents are addressed by kin term alone. In the case of ‘small parents,’ he tells us, ‘the kin term is used as an honorary title preceding their personal names.’ It is not, I think, too much to say that such nonfocal parents are honorary parents, somewhat like godparents, step-parents, and foster parents closer to home.

Ravuvu (1971:481), like Turner, notes that certain kin should be addressed and referred to teknonymously. In his example one's older brother is rendered as the ‘father’ of his – the brother's – first-born son, further evidence that the primary meaning of the ‘father’ term is indeed supplied by one's genitor. In the same vein he adds that the ‘father’ and ‘mother’ terms ‘used without a qualifier are always used specifically for [one's] father (genitor) and mother (genetrix) respectively’ (ibid.) He further notes that the FZ may be referred to as ‘father,’ to which a special prefix is added, and that the same relative may address Ego as ‘little father’ (Ravuvu 1971:482).

It is time to return to the Mae Enga in earnest. Meggitt (1964b:194) tells us that the referential use of the ‘father’ term simpliciter, i.e. without modification, refers only to one's father: this too is a recurrent feature in the ethnographic literature: we saw this, for example, in Quain's Fijian materials. Other members of his kin class and local patrilineal group bear the suffix kore. The meaning of this suffix apparently escaped Meggitt, but subsequent scholarship in the area revealed it means ‘little’ and ‘thin,’ i.e. that it signals secondary status.3 Hence it is tactfully omitted in direct address – a strategy with wide ethnographic employment (e.g. Banks 1974:59; Mayer 1965:15; Nuttall 2000:46).

There is more of relevance that can be discerned in Meggitt's materials. Distant agnatic kin are said to be ‘people of the erect penis’ and ‘those begotten by the one penis’ of an agnatic ancestor (Meggitt 1964b:192) – i.e. the male procreative contribution provides a model for more remote agnatic linkage. It seems clear from the foregoing analysis of the Mae Enga kin classification that an individual distinguishes the penis that generated him/her from other penises, which is of course to say that rendering connexion to more remote agnatic kin in phallic terms is metaphorical, rather like my saying that George Washington is the Father of My Country. In Sahlins' terms, it is an idiom involving a ‘mutuality of being’ – something which, for him, is the core of human kinship. But we have already seen that this alleged ‘core’ is decidedly compromised for the Mae Enga by subclassification in kinship terminology.

There is further evidence for the importance of primary kin among the Mae Enga. Meggitt (1965a:169–71) notes that an individual may be attacked by the ghosts of deceased kin, and that ‘[t]he most formidable are those of members of his domestic group, especially of the father and mother and of siblings and children’ (1965a:169) – as if nuclear family members are bound by a supernatural link not shared, or not as strongly shared, with more distant kin (see also Meggitt 1965b:111). Moreover, in mourning ritual, ‘[t]he members of[the deceased's] immediate family withdraw from everyday life for several weeks’ (Meggitt 1965a:181; see also Meggitt 1965a:187–89, 192, 1965b:113). Further, in an informative if ghastly analysis of the incidence of the severing of parts of fingers in mourning, Meggitt (1965a:184) found that the vast majority of such self-inflicted amputations were carried out by the primary kin of the deceased. More distant kin merely ‘tear their hair and beards,’ or, more extremely, ‘slice their ear-lobes so that blood flows over their shoulders’ (Meggitt 1965a:182). Finally, when the body is prepared for burial, ‘[m]embers of the immediate family … are too grief-stricken to participate, so that the body is prepared for burial by [a]ny close relative’ outside the family (ibid.). One might reasonably call all this evidence of a genuine ‘mutuality of being,’ but, quite unlike the metaphors linking patrilineal groups to ‘one penis,’ it unites people within the nuclear family.

It bears stressing that the nuclear family is indeed a unit of Mae Enga sociality. Because of a remarkable fear of the pollution that allegedly emanates from women, men and adolescent boys do not normally reside with their wives and mothers (see esp. Meggitt 1964a; also Meggitt 1958:270, 274); instead, residence for them is situated in a local men's house, which is also the locus of sacred ritual secreted from most women. The existence of such ‘men's cults’ has led some Marxist and feminist scholars to conclude that the nuclear family is absent in such situations, as Engels (1972 [1884]) posited it would be without capitalism more than a century ago. I have attempted to counter these conclusions elsewhere (Shapiro 2009, 2011), chiefly by arguing that the criterion of round-the-clock, year-long co-residence is met nowhere. In the Mae Enga case, in addition to what has already been noted, there is a conceptualization of residential unity: the married couple and dependent children are said to be ‘those who habitually live together’ (Meggitt 1965a:19) – this presumably because an adult or adolescent male usually resides near his wife or mother though not under the same roof. This symbolic co-residence is reinforced by the notions that a man should build a house for his wife which he visits for conjugal purposes (Meggitt 1965a:23), and that ‘[t]he elementary family constitutes the basic gardening and pig-raising unit in Mae society’ (Meggitt 1965b:107; see also Meggitt 1958:266, 274, 276, 1965b:106). Here is Meggitt (1965a:126–27) in some detail on the matter:

The newly married couple wear their wedding dress and decorations for a month after the final [marital] exchanges. They should not consummate the marriage until this period is ended. Meanwhile the bride ought to think constantly of her husband so that her good thoughts counteract the polluting effects of her sex and protect him from the impending dangers of intercourse. … [Her husband] has almost no contact with her during these first weeks. Much of his time is spent in learning the spells that every young man must purchase – magic to protect him from the effects of coition … When the groom knows the details of the magic, his teachers advise him it is time to copulate with his wife. He takes her first to his new gardens and points out the plots that she will have to cultivate. The two then enact a brief planting ritual … to symbolize and ensure a long-lasting and economically successful marriage. He performs his protective magic and they copulate. After four or five days of meeting for intercourse in the bush, they resume everyday dress. The husband gives his wife a new digging-stick and a net-bag, then sends her to the gardens … (see also Meggitt 1965b:127).

It should be added, finally, that an individual has a variety of important relations with a network of matrilateral and other close kin (Meggitt 1965a:124, 212). Some of these have to do with warfare, which is endemic among the Mae Enga (see esp. Meggitt 1977). Sahlins (2011:236) places considerable emphasis on Meggitt's report of an adage ‘We fight the people we marry’ (Meggitt 1965a:101), which he apparently takes at face value. But this by no means should be treated as a comprehensive statement of Mae Enga rules of military engagement. It is true that warfare occurs between local patrilineal groups, largely over land, but when it does people take account of individual matrilateral relations outside such groups. Thus Meggitt (1965a:215) tells us that when a man and his sister's son find themselves in opposed camps they avoid aggressive action against each other and focus instead on men of the opposed group who are less closely related (see also Meggitt 1977:80–81). Indeed, it sometimes happens that a man ‘refuses to join an attack on the clan of his [maternal] uncle or [sororal] nephew and tries to warn his relatives of the danger’ (Meggitt 1965a:215; see also Meggitt 1977:80–81). He may even join his maternal kin in battle, even against his own clan – though in such a case he avoids conflict with his own brothers (Meggitt 1965a:37). In short, whatever ‘mutuality of being’ may be idiomised within the patrilineal group, this is trumped by close kinship; and this in turn involves or is generated from relationships within the nuclear family.

Sahlins is utterly oblivious to these considerations. Indeed, his taking seriously the adage ‘We fight the people we marry’ is much like his wholesale subscription to Mae Enga homilies about distant agnatic linkage, noted above, in connection with his emphasis on ‘the spiritual bestowal of the patrilineal clan ancestor?’ In this case too he mistakes native platitudes with effective rules of sociality.

Here we need to take a closer look at Mae Enga religion, about which Meggitt provides relatively little information. He does tell us, however, that spirit-entry into the foetus occurs not at conception but four months later, i.e. at foetal quickening (Meggitt 1965a:163), and that such ancestors are the beings placated in male secret/sacred ritual (Meggitt 1965b:115). Women are barred from such ritual (ibid.) Further, it is considered offensive to the ancestors if a married couple copulates in the vicinity of the men's house and/or just after ancestors have been propitiated, for to do so would ‘antagonize’ them (Meggitt 1965b:115, 118). There is, then, a conceptualized separation between sexual reproduction and spiritual life: indeed, the two seem to be seen as antithetical in Mae Enga thought. But the positing of an antithesis requires a thesis, and this is supplied by physical generation. In short, all of Mae Enga sociality is modeled on primary kin ties – secondary ties as their extension, spiritual notions as their antithesis.

Sahlins and the Trobriands

I take it as obvious that an antithesis between spirituality and carnality is very much a part of the Great Religions of the West. This being so, the insistence on a radical West/Rest distinction, so crucial to Sahlins' thought (see esp. Sahlins 2008) finds no support in Mae Enga religion. Nor, as we have seen, do the Mae Enga or the Fijian materials lend themselves to such a distinction when it comes to the focality of primary kin in constructing other kinship-like notions. But what of Sahlins' claim that, in Enga thought, there is a ‘relative devaluation of the father's substantive contribution to the foetus in favour of the spiritual bestowal of the … ancestor?’ Surely this brings to mind the famous claims that Trobriand Islanders are ‘ignorant of physiological paternity,’ and that they have an ideology of conception which assigns agency to subclan ancestors (Malinowski 1929:170–85). The Trobriand case is especially pertinent here because Sahlins (1976:37–39) buys without question into Malinowski's report. His – Sahlins' – point is that this report is evidence – like the supposed lack of concern among the Mae Enga with individual paternity – of the unimportance of procreative links in non-Western thought. Here again, I shall argue, he utterly misconstrues the ethnographic record.4

Not long after Malinowski's claim, Rentoul noted a decided concern among Trobriand women with native measures ‘to expel the male seed’ after intercourse (1931:153; emphasis in original) in an attempt to abort pregnancy (see also Powell 1980:701; Senft 2009:221–22, 2011:33–34).5 In fact a very great deal of evidence has accumulated over the years indicating that the alleged ‘ignorance of physiological paternity’ in the Trobriands masks not only a religious dogma, as Edmund Leach (1967) concluded, but various other nuances that eluded Malinowski. Several of these are suggested by Rentoul's further remarks:

[T]ogether with this practical knowledge of physiological paternity, there has always existed the magico-religious explanation … {T]his is the Story of Birth, as it is believed by intelligent Trobrianders as thoroughly as a modern [Western religious] congregation would believe the curate's shy announcement that during the night the angels had brought him a little son … Presently the [spirit] will visit the woman and place upon her forehead a miniature babe … The babe descending the body of the mother will visit each breast for nourishment, then descending further will enter the womb, where it will remain until the day of its birth. In this process the father's part is simply ‘to keep open the way’ by sexual intercourse (1931:153).

Besides the contention that Trobrianders are indeed aware of the fertilizing power of semen, there is a great deal here that should be pursued. Note the apt comparison with Western religious notions connected with conception. Note too that the Trobriand father ‘keeps open the way’ by repeated acts of sexual intercourse: indeed, Malinowski (1929:207) reported as much. He noted as well that a child is supposed to resemble his/her father (Malinowski 1929:204–08; see also Lepani 2012:67–68; Senft 1995:221). The two doctrines are connected: the father is held to shape the child's features through repeated contact between his penis and the foetus, much as a sculptor increasingly gives shape to his creation with his chisel. Trobrianders refer to this process by an expression which Austen (1934:103) translates as ‘hammering,’ which corresponds fairly closely to our mildly obscene use of ‘banging’ and ‘pounding’ to refer to male thrusting in intercourse. A comparable sort of ‘hammering’ occurs in Trobriand garden magic, wherein a man repeatedly strikes with a pole or stick the soil of his wife's garden, expressly likened to her body (Brindley 1984:17–40; Eyde 1983:70–71). One of Mosko's informants ‘likened the soil to a womb and the stick to an erect penis’ (Mosko 2009:686). One could, I suppose, argue that it is the presumably ensuing rain and not the stick which fertilizes the garden, but this contention can be dispelled by appeal to Trobriand mythology. Here women have their ‘ways’ opened by rain – or, especially noteworthy, this – by stalactites, whose dripping ‘water’ effects the ‘opening’ (Malinowski 1948 [1916]:228). In case there is any doubt about the symbolism, we have the direct statement from Rentoul (1932:275) that the mythical stalactite ‘is looked upon as a phallic symbol’ (see also Barton 1917:109; Malinowski 1929:182–83; Senft 2011:18).

Several further points in Rentoul's exposition warrant emphasis. Note the conmingling of a ‘magico-religious explanation’ with a ‘physiological’ one, and that in the former case, the location of bodily contact is non-vaginal: the forehead.6 In this sense, at least, the two ‘explanations’ are antithetical, just as the Immaculate Conception, in Roman Catholic theory, is antithetical to the conception of the rest of us and, as such, was effected by Divine entry into Mary's ear. Moreover, the Trobriand spirit is said to come to the woman ‘presently.’ ‘Presently’ after what? In a contribution authored only two years after Malinowski's initial report of ‘ignorance,’ Reid (1918) argued that spirit entry occurs not at conception but at foetal quickening, as is the case with the Mae Enga. Austen (1934:108) reports this as well. This being so, some other action must be construed to cause conception, and there is every reason to believe it is coitus. A final consideration is that the ‘physiological’ theory was most plainly held by Trobriand women: this is in fact a recurrent theme in subsequent ethnographic and theoretical literature on the area (Austen 1934:104, 113; Hocart 1954:99; Mosko 1985:211; Powell, cited in Montague 1971:359; Senft 2011:33–34; Sider 1967:95–96, 105). Why, then, should Trobriand men be so concerned with the spiritual contribution to the foetus?

Here, I think, we need to recall that the Trobriand subclan is construed to be a part of one of the four clans that, in native theory, have always existed, that emerged from the Underworld at the Beginning (Eyde 1983:67–68; Hutchins 1980:21; Malinowski 1935:165, 342). Despite its ‘matrilineal’ character, it is thus much like Aboriginal Australian patri-clans, which Stanner (1960:253) aptly called ‘sacramental corporations of a perennial order.’ Malinowski (1929:179) himself noted that the ‘ignorance’ theory ‘gives a good theoretical foundation for matriliny: for the whole process of introducing new life into a community lies between the spirit world and the female … ‘This ‘new life,’ he further tells us, is held to be a re-incarnation of an old one, which merely housed its spirit, and that this spirit has always existed and will continue to do so, going to yet another individual after the demise of its present host. He notes further that the spirit is specific to a particular subclan, from which it cannot be alienated (Malinowski 1929:182, 1948 [1916]:219–20). From this perspective, then, each subclan is self-generating through a process in which sexual intercourse has no place, given the incest barrier within each such group (see Moore 1964). Hence even this classic formulation leaves room for doubt about the alleged ‘ignorance.’ It suggests, rather, that the ‘magico-religious explanation’ is employed when discourse is about enduring corporations, and that, in other areas of conversation, a ‘physiological’ one is maintained.

That the two ‘explanations’ are antithetical is also suggested by Malinowski. Thus he tells us that ‘as a means of testing the firmness of their belief [in spirit-entry], I sometimes made myself … aggressively an advocate of the … physiological doctrine of procreation’ (1929:185). When he did this, ‘I was sometimes astonished at the fierce opposition evoked by my advocacy of physiological paternity’ (1929: 186; emphasis added). Further evidence of a similar ‘fierce opposition’ is provided by Fortune's account of the neighbouring Dobuans (Fortune 1934). There is considerable mixing between the two peoples; hence

The Dobuans know the Trobriand belief that procreation is from the reincarnation of spirits of the dead, not from the biological father. They say bluntly that the Trobrianders lie. The subject is not brought up between Trobrianders and Dobuans as it has been the subject of anger and quarrel too often in the past. My Dobuan friends warned me not to mention the matter in the Trobriands before I went there. Once I was there I deliberately made the experiment. The Trobrianders asserted the spiritual belief, just as Dr. Malinowski had published it. But the head of every Dobuan in the room immediately was turned away from me towards the wall. They affected not to hear the conversation; but afterwards when they had me alone they were furious with me (Fortune 1934:239).

The suggestion, surely, is that, at least on the Trobriand side, something more is at stake than a knowledge of the real facts of life. It is pertinent to note here that the debate Fortune intended was almost certainly between groups of men, for it is apparently Trobriand men, not women, who deal with outsiders. Indeed, although the denigration of women in the Trobriands hardly compares with its Mae Enga analogue, the two sexes in the former area are not at all on a par. Political leadership is a male monopoly, and magical power and kula fame are mostly male achievements (Campbell 2002:160 et seq.; Glass 1986:50; Malinowski 1922:81; Mosko 1995:774 et seq.; Weiner 1977:67, 1988:139–57). Indeed, Montague (1983:38–39) tells us that in Trobriand theory women are construed as animal-like, not quite the real human beings that men are held to be. Now this is especially remarkable in view of the fact that even those Trobriand men who insisted on the ‘magico-religious explanation’ with Malinowski were entirely explicit on the ‘physiological explanation’ in accounting for animal reproduction (Malinowski 1948 [1916]:227–29).7 So it makes sense that men, being non-animals par excellence, and as such leaders of enduring corporations, should wish to sustain the fiction that their reproduction is non-carnal, and that they should be ‘fiercely opposed,’ at least in public encounters like those with anthropologists and other foreigners, to any suggestion to the contrary (see Austen 1934:103–04; Montague 1971:359; Powell 1969:652; Rentoul 1932:275; Senft 1995:216).

This conclusion is supported by more recent research in the Trobriands by Jerry Leach (quoted in Glass 1986:47). The following statement from him is especially remarkable in the present context:

Trobrianders believe in spirits of the dead who reincarnate themselves [within] their … matrilineal group[s]. The formal belief seems to deny males any role in reproduction, and the Trobrianders convinced Malinowski that their religious belief was a true statement of their actual knowledge about [conception]. … [H]owever, males are [in fact] recognized as part of the reproductive process … The public denial of this seems intended for the ears of the spirits[,] who jealously guard their pre-eminent role in the formation of new human beings, but it has led the world to believe that the Trobrianders do not associate intercourse with [conception] (emphases added).

Highly pertinent here is the research of Glass (1986, 1988), based upon Malinowski's publications and unpublished field notes, as well as Glass’ own examination of Trobriand art in various museums. In his 1988 article he shows that topographic features in the Massim area are endowed with blatant sexual associations. The one of greatest importance for our purposes involves a particular shoreline, which is designated as momola, an expression which, following Malinowski (1929:167, 1948 [1916]:223), he translates as both ‘semen’ and ‘female sexual discharge.’ [I]t was generally by bathing in the momola that women announced that they had become pregnant' (Glass 1988:63–64). Malinowski (1948 [1916]:223) insists that ‘[t]he spermatic fluid … serves merely the purposes of pleasure and lubrication (see also Malinowski 1929:167).’ This conclusion, however, is gainsaid by Glass’ 1986 analysis of the artwork on Trobriand war shields. These shields contain more or less explicit images of phalluses, the female reproductive tract, coitus, semen, and human embryos. Glass (1986:54) also stresses the fertilizing significance assigned to water in Trobriand ritual, especially the fact that a woman usually communicates that she is pregnant by saying that she was entered by a spirit-child while bathing in the sea. More explicitly, the idiom reported by Malinowski (1948 [1916]:218) is that ‘a fish has bitten me.’ There is considerable evidence (see esp. Glass 1986:54, 58; Senft 2011:31; Malinowski 1929:172–76) that fish represent for Trobrianders spirit-children and, less certainly, the ‘water’(= semen)-surrounded phalluses from which they are implicitly held to emanate. As for the ‘biting’ metaphor, Malinowski (1929:340–42) tells us that erotic biting is an expectable part of love-making in the Trobriands. Thus Glass (1986:58) is led to the following conclusion: ‘What is overtly negated on land (male fertility) takes place … on the seashore, momola (semen …) through water and fish … which [are] linked to the phallus …’ Glass (1986:60) even has an explanation for the covert nature of this symbolism: ‘The Trobrianders,’ he tells us, echoing Jerry Leach, ‘were very guarded about articulating their knowledge of paternity for fear of offending “the ears of the spirits” ’ (emphasis added; see also Senft 2011:29–30).8 So, as with the Mae Enga, it is the spirits, the ancestors, who are ‘fiercely opposed’ to any talk and/or action which participates in carnal generation. I would add only that the ‘negation’ of Trobriand ‘male fertility’ on land makes sense in that parcels of land are associated with particular subclans and hence corporate immortality, whereas its representation in near the sea links it with kula and individual renown (see esp. McDougall 1975).

It is surely pertinent here to note that, as Senft (2009:221) has recently observed, ‘the female world’ was ‘[o]ne of the few ethnographic niches Malinowski left.’ Nearly all his major informants were men – hardly unusual for a male ethnographer, especially, at the time, a single one, but, nonetheless, almost certainly connected to this ‘finding’ of ‘ignorance of physiological paternity.’

It is in this context, moreover, that we can understand why Malinowski (1929:5) reported that the Trobriand father is said to be tomakava, ‘a stranger’ – a point on which Sahlins (1976:38) places some emphasis in his attempt to dissociate kinship from procreation. Weiner's rendition of the same term as referring to ‘ nonclanspeople’ (1976:53–54) is probably closer to the truth. But the term is also applied to people in mourning regardless of clan or subclan affiliation (Seligman 1910:720), so, whatever its focal significance, its widest application would seem to be something like ‘anyone who is outside the sphere of normal social relations in the situation at hand.’ In other words, it has a subjunctive, ‘as if,’ character: mourners, though really kin, are treated as if they were ‘strangers’ – just as, for example, in the Mae Enga case, all men of a clan are said to be ‘begotten by a single penis’ (even though people really know that several penises were involved), and a man my father calls ‘brother’ is regarded as if he were my father (even though I know he really isn't).

I return to this point in the next section, and in my concluding remarks. For now it should be noted that the Trobriand father is not really a ‘stranger.’ Here is Weiner on the matter:

Malinowski … placed great emphasis on the classification of [the] father as tomakava, … ‘stranger’ rather than own kinsman (veyola tatola). My informants said that no one would ever call their father tomakava. They said [instead] that he was the most important kinsman (veyola) they had. It was only in conversations or debates concerning … rights of a [subclan] where a man as father would be referred to as tomakava. … Even if other people call our father tomakava, we know that he is not tomakava to us (1976:124, emphasis added; see also Sider 1967:103–05).

In short, the Trobriand father is no stranger to his children; not only does he bring them up: he has a substantial role in their very generation. This is confirmed by recent ethnographic research in the area. Thus Mark Mosko, in an email communication to me dated 13/8/12, notes a Trobriand expression toli una'i, which can be used to refer to one's mother and one's father as a couple. He further reports that una'i means ‘to conceive.’ A day later, also via the Web, Kathy Lepani provided the same information. I think we can safely conclude that the Trobrianders are no more ‘ignorant of physiological paternity’ than the Mae Enga are. And, finally, the ‘third party’ in both Trobriand and Mae Enga theory is antithetical to the first two.

Links to Aboriginal Australia

It is hardly accidental that Malinowski ‘found’ Trobrianders to be ‘ignorant of physiological paternity’ not long after Spencer and Gillen (1899:265) reported that among the Aboriginal people of Central Australia ‘the idea [is] firmly held that the child is not the direct result of intercourse, that it may come without this, which merely … prepares the mother for the reception and birth … of an already-formed spirit child who inhabits one of the local totem centres.’ Indeed, it would be impossible to overstate the impact of Spencer and Gillen's ‘discovery’ on subsequent social theory.9 Suffice it to say here that it very largely framed subsequent discourse on Aboriginal sociality; the question became, ‘Is population X aware or unaware of the male role in conception?, when, as I shall show, it should have been, ‘Under what conditions are the people of population X inclined to emphasize physical or spiritual generation?’ (see Burridge 1973:108–09). Ashley Montagu's well-known compilation of the evidence (1937) is remarkably naïve in this regard, and entirely self-serving: every instance of ‘awareness’ is explained away as a result of European contact (Suffern 1938:138). This is not acceptable, for two reasons. First, any empirical discipline must deal with the pertinent data at hand; things may have been different in the past, but this possibility cannot be assumed in order to salvage an hypothesis. Second, native time constructs, which should most certainly be considered here, are not mere renditions of the past but have social employment: this has been documented in a considerable literature (e.g. Bloch 1977; Layton 1989; Morphy and Morphy 1984), and its significance will be clear once again shortly.

The first step in discerning what is really involved is the discovery, at least for some parts of the continent, that such theory is not about conception at all: rather, it pertains to birth or foetal quickening. The latter possibility was first suggested by Pink (1936:288) and Suffern (1938:138), the latter in an article which is all but unknown. More recently, Scheffler (1978:5–13) has surveyed the ethnographic literature and found both doctrines present but in different populations, and he concludes that the ‘ignorance’ claim is entirely misplaced. There is thus an immediate alignment of the Aboriginal materials with those from the Trobriands and Mae Enga, noted above. Moreover, to the extent that we can generalize from Scheffler's findings to the entire continent, we must assume that something other than the ‘spirit-finding’ experience is posited as the cause of impregnation. It is hardly difficult to imagine what this is.

A particularly informative case is provided by a relatively recent revival of the ‘ignorance’ claim for Aboriginal Australians by Tonkinson (1978), on the basis of his fieldwork in a part of the Western Desert.10 Tonkinson made it clear to his informants, he tells us, that he was interested in pre-contact ideas on embryological development. His words:

Since the men had told me that it was appropriate to question people about spirit-children and their part in procreation, I subsequently pursued this topic and talked with many informants, mostly males, about their knowledge of what the ‘old people’ had said about the subject in pre- contact times … I repeatedly stressed that I wanted to know about traditional beliefs. … [A] fairly consistent set of beliefs was obtained, and they were notable for the absence of physiological referents (Tonkinson 1978:83; first and second emphases added, third in original).

So Tonkinson's inquiries were authorized by certain Aboriginal men. I shall pursue the circumstances that led to this authorization shortly. But it should be noted now that, although he assumed his eliciting conditions were uncontextualized, he was in fact inquiring into a realm of discourse dominated by male authorities and male informants, much like Malinowski in the Trobriands. In the same vein, the distinction he urged on his informants between pre-contact and post-contact knowledge, I shall argue, was read by them as a distinction between spirituality and carnality, or, much the same, I shall also argue, a distinction between male discourse and its female counterpart.

Tonkinson does not hesitate to tell us of the circumstances that led to this inquiry. Initially in his fieldwork, he concluded that a connexion between coitus and pregnancy was indeed posited. His first informants on the matter were ‘a middle-aged married man … , his wife (whom I questioned separately), a second middle-aged married man and an older man’ (Tonkinson 1978:82). Here are his results:

[T]he data provided by the first three informants were substantially the same … These people maintained that sexual intercourse is a necessary prerequisite for conception … The female informant stated also that the mixture of semen and blood, together with what the mother eats, feeds the child growing in the womb. … My attempt to elicit information from the older man came to an abrupt halt as soon as I mentioned semen and menstrual blood. He admonished me forcefully that such things were ‘dirty’ and very distasteful and not fit topics for conversation. … [H]e was actively anti-Christian in his attitudes, so [missionary influence] was not the reason for his insistence on terminating this particular discussion. In fact, he was adamant that his views reflected what was the Law, traditionally … (ibid.; all emphases added).

It needs to be noted that Aboriginal people in the Western Desert and elsewhere refer to the mythically-generated transmission of local custom as ‘the Law’ in Aboriginal pidgin English. The obvious other thing to note here is that while the female informant was, apparently, quite comfortable with mentioning semen and menstrual blood, the men were not, and the oldest of the three, presumably the one with the most spiritual knowledge, was blatantly uncomfortable discussing carnal generation. It is surely not the case that he did not know about it; but, to borrow from the preceding Trobriand analysis, it ‘angered his ears’ to hear of it.

But Tonkinson was undeterred, and he pursued the matter in later fieldwork. His words once more:

I … returned to the procreation topic, with the same middle-aged informant and a younger man … Both thought that the role of semen was probably always known and again maintained that intercourse was necessary for conception to occur. … It was at about this juncture of the interview that the same old man who had silenced me on the subject of semen and menstrual blood seven years earlier joined us. Within a couple of minutes I decided to discontinue my line of enquiry because the new arrival remained uncharacteristically silent and looked ill at ease …

Within an hour, a man came and told me that there would be a big meeting later the same day to chastise my two informants for having talked ‘danger words.’ It transpired that my older informant had just told many people … that we had been discussing a dangerous and forbidden topic … Later the same day several people mentioned to me that they had heard about some ‘bad talk.’ … One man suggested that these topics (meaning menstruation and pregnancy) might be alright for women to talk about with one another, but are definitely not ‘men's business.’

About sixty men attended the meeting that afternoon, held out of earshot of the camp. … One asked, ‘Are we like stallions and dogs? …

The men seemed hesitant to bring the meeting to the central topic, so I addressed them and admitted that I had asked some questions about semen and menstrual blood – and in so doing, referred to both using native terms … A chorus of older men immediately affirmed that these were the ‘danger words’ and that no one wanted to hear any further mention of them. Then, amid nods and mutterings of strong agreement, my old informant who had told the camp of the conversation that morning took the floor and told me that, according to traditional Law, the only thing relevant to the topic of procreation is spirit-children. He said that he and the other men did not know anything about this other business (physiological) at all, and they didn't want to hear it because it has nothing to do with … their Law (ibid.; all emphases added).

Which is to say, again following Trobriand idiom, it ‘angered the ears of the spirits’ – indeed, almost literally: only men were present, and the meeting was held away from the main camp, which, I shall guess from other materials on Aboriginal symbolism (see esp. Munn 1962), is construed to be the domain of women. Note too the evocation of bestiality – ‘Are we dogs?’ – and its counterpart in the Trobriands. And note, finally, the association of the carnal aspects of reproduction with women, and the utter urgency of the claim that men's interest in the process is – or should be – confined to its spiritual aspects – again with decided counterparts in the Trobriands. It is surely not that these men do not know about ‘this other business’: it is that, by their lights, polite people (those who've had at least a glimpse of the sacred world, i.e. grown men) do not – should not – speak of it. Hence when some of these men authorized Tonkinson to broach the matter of spirit-children to people, they had already – and very definitely – structured his enquiries, which perforce have absolutely nothing to do with any ‘ignorance of physiological paternity.’ Instead, as do those of Ashley Montagu and many others, they mistake the salient oppositions carnal generation/spiritual generation and female discourse/male discourse for the naïve post-contact/pre-contact one.11

These conclusions are strongly supported by other ethnographers working in the Western Desert. Thus Poirier (2005:134) notes a ‘cosmological’ as well as a ‘biological’ account of what she takes to be conception among her informants, only to tell us subsequently that the ‘cosmological’ factor is held to occur at the time (or just before) ‘the mother experiences morning sickness’ (ibid.), and that it is linked to ‘an unusual event that occurred in early pregnancy (Poirier 2005:136; emphasis added). Both of these statements suggest that the ‘cosmological’ factor is held to take effect not at conception but subsequent to it, though probably somewhat before quickening. Still, the total theory espoused by Poirier's informants is in accord with Scheffler's generalization, noted above, of a native distinction between conception and spirit-entry, and it seems a fair guess that the posited cause of the former is coitus.

Myers' male informants refused to discuss menstruation, pregnancy, and childbirth, all of which they classed under an expression which he translates as ‘shameful matters’ (Myers 1979:364; see also Myers 1986:122–23). Such ‘matters,’ Myers (1979:364–65) tells us,

are considered “shameful” … because they imply … similarity to animals. One should behave in a manner unlike animals, that is, with shame. … [T]he promiscuity of dogs is a subject of derision … , and people who have intercourse with wrong categories of kin or who copulate indiscriminately (or without regard to the presence of others) are said to be ‘like dogs.’

The similarity to Tonkinson's findings, and Malinowski's, should be plain.

Yengoyan (1978:110) tells us that ‘[i]n relating copulation to conception, … males note that there is no connection between the two activities. ‘But,’ he soon adds,

this version of conceptual knowledge is almost specifically confined to … males. Those same males note that the dominance of spirit implantation, and copulation is primarily a nurturing force. Among women, the two factors are [mostly] reversed … In some cases a woman noted a strict one-to-one link between copulation and conception with no intervening factor. Others stressed copulation as primary, but noted that spirit matter must be present to provide form … (Yengoyan 1978:110–111).

The similarities between this and what Malinowski elicited from Trobriand men is striking, particularly in the attribution, in both cases, of ‘a nurturing force’ to copulation. Also in this vein is the assignment of bodily form to males – through coitus in the Trobriand case, via spirit-finding by fathers in the Western Desert .

Even more remarkable are the results of Sackett's research (Sackett 1977). He found the same gender associations as other Western Desert ethnographers. Thus he tells us that, although his informants posited a connexion between sex and foetal development

there was a noteworthy difference in men's and women's willingness to respond [to my queries]. The former showed a reluctance to discuss semen, menstrual blood, and the like, and in general endeavoured to change the subject or at least divert its direction. Women appeared to be less concerned in this regard (Sackett 1977:161–62).

Furthermore, Sackett pursued the matter of human reproduction with an elderly male informant. Here is part of the exchange:

  • Q. How does a spirit-child enter a woman?
  • A. Her husband might dream it.
  • Q. What about coitus? Can he put it in that way?
  • A. No! The spirit-child enters in another way.

This last remark should be taken quite literally: the spirit-child is said to enter though ‘the stomach, foot, or mouth – never the vagina’ (Sackett 1977:170, emphasis added; see also Fry 1933:252). I submit that the patent emotional tension here is much the same as that which led Malinowski to claim that the Trobrianders were ‘ignorant of physiological paternity,’ and why Sahlins is able to posit, for the Mae Enga, ‘the relative devaluation of the father's substantive contribution to the foetus.’ In all three cases, it now seems plain, what we have is a clear and emotionally charged distinction between physical and spiritual generation (Shapiro 1988) and an association of women with the former and men with the latter, though this is less obvious in the Mae Enga case.12 In none of these instances is there anything even barely resembling a lack of knowledge of paternity. The whole subject, really, is nothing more than a product of Victorian fantasy which has been resurrected by anthropologists like Sahlins and now figures in popular discourse, including the tourist trade (Senft 1998), and ‘radical’ feminism (see Eller 2000:93–95 for an important critique).

What is involved has been theorized by Aijmer (1992) in his concept of ‘animation’ in foetal development. His words:

‘Animation’ may be as important as are notions of conception. The word animation may be used to signify … various cultural institutions that have to do not only with designating of a human existence onto a foetus but, furthermore, into more precise notions of individuality and personhood; we are dealing here … with ideas of ‘souls’ and the development of souls … (ibid.:10; see also Hiatt 1990).

Concluding Remarks

Sahlins' attempt to provide a definitive statement on human kinship is thus egregiously flawed, because he denies that what anthropologists have long called ‘primary kinship’ is indeed primary. That he is in good company – Marx and Engels, Durkheim, Lévi-Strauss, Leach, Needham, Schneider and their many admirers – is beside the point. The demonstrable fact is this: other kinship notions are modeled upon – they assume – primary kinship notions – either as their extensions, as both the Fijian and the Mae Enga cases demonstrate, or as their antitheses, as can be seen in the latter case and in the Trobriands and Western Desert. It is fine to question , as these and other scholars have, the assumption of primacy; but this assumption, it turns out, has an empirical foundation of the most solid kind. I can do no better in this connexion than to end with Lounsbury's conclusion in his seminal analysis of Trobriand kin classification:

The earlier ethnographers, who naively relied upon our notions of the family and genealogical connections, were favored by luck and good intuition. It turns out that they were on fairly firm ground after all. The theorists who have contested this view, on the other hand, may still have some more rethinking to do (Lounsbury 1965:183).


  1. 1

    Most of Sahlins' reply to me consists of a lengthy chunk of what was then his forthcoming book on kinship (Sahlins 2013). The first version of this essay was written before the book appeared. I have since read it in its entirety. I need to note that nothing in it obviates any of my remarks here.

  2. 2

    Meggitt (1964b:192) translates the ‘kin’ term as ‘people of the same clans,’ and one of the ‘nonkin’ terms as ‘other clans.’ This gives the mistaken impression that, for the Mae Enga, kinship depends exclusively upon patrilineal group membership, which, as we have seen, is not the case. The tendency to render all social relations in ‘group’ terms seems to be fairly widespread, including both anthropologists and some of the peoples they study (e.g. Keesing 1971; Shapiro 1981:118; Wagner 1974).

  3. 3

    I am indebted to an anonymous reviewer of an earlier version of this essay for this information.

  4. 4

    My analysis here is excerpted from a more comprehensive study of Trobriand sociality which I have yet to publish. I should note also the help via email I received from several contemporary scholars who have worked in the Trobriands – Kathy Lepani, Mark Mosko, and Gunter Senft. In more recent email communication Lepani and Mosko have reminded me that there is more to Trobriand notions of paternity than insemination, on which, they seem to think, I place too great an emphasis. But my analysis here is not intended as an exhaustive treatment of these notions; my concerns, rather, are simply to establish that one of them is insemination; and that spirit-entry occurs later in gestation.

  5. 5

    Malinowski (1929:197), by contrast, says that he ‘can say with complete confidence [that] no preventive means of any description are known, nor the slightest idea of them entertained.’ It seems fair to suggest that his ‘confidence’ on the matter was grossly misplaced.

  6. 6

    Malinowski (1948 [1916]:216) also notes that the spirit is sometimes held to enter vaginally, though it is unclear if and how this notion articulates with the idea of entry through the forehead. Elsewhere (Malinowski 1929:175) he tells us that the positing of vaginal entry is a ‘much less authoritative belief.’ In view of the foregoing discussion I suspect it is held mostly by women. It should be noted that Senft's male informants mentioned vaginal entry as well as entry through the head (Senft 2011:31–32). But the former mode, apparently, only occurs if the woman ‘swims somewhat carelessly’ (Ibid.: 31). Malinowski (1929:176) also mentions ‘the skin of the abdomen’ as a posited point of entry.

  7. 7

    Later Malinowski (1929:192) claimed that this conclusion was incorrect. Still, he seems not to have inquired deeply into Trobriand theories of animal reproduction, and he concludes that ‘animals are not subject in this, as in many other respects, to the same causal relations as man.’

  8. 8

    Thus the idea of sea water seems to disguise sensitive subjects in two related ways. Here it serves as a substitute for marital coitus. In another context the substitution is for brother/sister incest, ‘the supreme taboo’ in the Trobriands (Malinowski 1929:514). Thus Malinowski tells us that the brother (or, apparently less frequently, mother's brother) of a woman desirous of becoming pregnant is supposed to assist her by bringing a vessel of sea water to her and placing it beside her as she sleeps (Malinowski 1929:176). Presumably both actions – the impregnation of women by ‘sea water’ while bathing and the prestation of such ‘water’ by her brother – would, absent these substitutions, ‘offend the ears of the spirits.’

  9. 9

    Some idea of the intellectual climate can be gained from Belier (1997), Kuklick (2005, 2006), and Stocking (1995:87–98). The last-named quotes Malinowski (1913), who ‘estimated that half the theoretical work in the years since its publication had been based on Spencer and Gillen's book, and that nine-tenths had been “affected or modified by it” ‘ (Stocking 1995:96). Malinowski's remarks are typically hyperbolic, but the point is made.

  10. 10

    This analysis of Tonkinson's findings draws heavily, and without further acknowledgement, on an earlier effort of mine (Shapiro 1996). It is modified here to take account of a somewhat different context of scholarly debate as well as of pertinent literature which has appeared since my initial statement.

  11. 11

    This is an easy conflation, because in religious thought non-carnal generation is rendered as temporally antecedent to carnal generation, just as pre-contact is temporally antecedent to post-contact. In the Aboriginal Dreamtime the Creative Beings create non-carnally, largely by externalizing themselves into features of the landscape (Munn 1970). In the Trobriand case Original Creation is also accomplished non-carnally, i.e. through the emergence of an asexual brother/sister pair from the Underworld (Malinowski 1929:182). In the Bible God ‘originally’ creates through speaking and, in the case of Adam, from mud (Old Testament), then, in the New Testament, recreates the human world through the sacrifice of an asexual Being.

  12. 12

    It is also something of a simplification for the Western Desert. Poirier (2005:135) reports that in some cases the mother is the spirit-’finder.’ Moreover, Western Desert women have secret/sacred rituals of their own, from which men are barred, but these are deemed by both sexes to be of a lower order of sacredness than the corresponding male rituals and less profoundly concerned with the generation of humanity and nature (e.g. Berndt and Berndt 1944:242 et seq.; Hamilton 1980:15–16; Myers 1986:146–47, 252; White 1975:132–33; Yengoyan 1989:178–79).