THE C HA RACTER OF KINSHIP EDITED BY JACK GOODY 49n CAMBRIDGE UNIVERSITY PRESS CAMBRFDGE LONDON ■ NEW YORK ■ MELBOURNE Genetrix: Genitor : : Nature: Culture? J. A. Barnes We salute Meyer Fortes for his achievements not only in Hie intensive investigation of the Tallensi and Ashanti of Ghana but also in the comparative analysis of diverse social and cultural systems, notably in his Lewis Henry Morgan lectures published as Kinship and the Social Order (1%9), Although he has long remained a steadfast defender of the strategy of concentrating anthropological field resources on the study of peoples whose material artefacts are simple (Fortes 1958 a), he has always been sensitive to the light these studies shed on patterns of living found in industrial societies. It is therefore appropriate in the present context to discuss some of the issues that arise in kinship studies when we endeavour to compare social and cultural patterns in many different societies including our own. My thesis is that this comparison suggests a reformulation of the relaiion between kinship and nature. I focus on putative physical relations rather than on relations of social parenthood-It can be argued that in anthropology and sociology comparative analysis is impossible without including, either explicitly or by implication, the society to which the analyst himself belongs and the culture whose concepts and categories he uses to ihink with. This view has been expressed by Schneider (1972: 41-%} who, in recent years, has appealed forcefully for the study of kinship as part of a cultural or symbolic system and who has provided the most uncompromising account of Such a system in the Western tradition. He says: The next problem.. .is the old one of how comparison can be conducted on a cultural level if it is assumed that each and every culture may be uniquely constituted., .our own culture. - always serves as a base-line for cross-cultural comparison. Without some comprehension, however botched, distorted, biased, and infused with value judgments and wishful thinking, both good and bad, our own culture always remains the baseline for all other questions and comparisons. In part, this is because the experience of our own culture is the only experience which is deep and subtle enough to comprehend in cultural terms, for the cultural aspects of action are particularly subtle, sometimes particularly difficult to comprehend partly because they are symbols nut treated usually as symbols but as true facts. Schneider implies, so it would seem, that even when we are comparing, say. unilineal systems found in different parts of Africa, as in Fortes' classic 61 GcncirLx. Genitor.. Nature : Culture? paper on 'The structure of unilineal descent groups' (1953 «}- thee is an implicit, comparison between the various African notions of unilineal descent and filiation and similar notions current in the Western tradition. In the passage cited however. Schneider is talking about cultural symbols, and it is not clear whether he would argue that in comparing, say, forms of social organization we are similarly forced to begin our analysis with forms prevalent in our own society. Indeed, at other places in his paper he draws a distinction between, on the one hand, 'the scientific facts of biology' and 'biology as a natural process' and, on the other, the cultural symbols that may perhaps (or perhaps not) be derived from these scientific facts. This suggests lhaL he sees science as the study of nature, i.e., as natural rather than social science, and that he makes a distinction between 'science', dealing with facts, and 'culture', dealing with symbols. But if there are facts of nature and biology that can be demonstrated scientifically, as he maintains (he seems to have the processes of human reproduction in mind), then presumably there are other scientifically ascertainable facts about where people live, who they work with, who commands whom, and so on which can provide a framework for a comparative analysis of social organization that is not linked distinctively with any particular society, not even our own. The distinction Schneider draws is widely used in social science. It is somewhat akin to the contrast between objective and indexial meanings used by the phenomcnotogists. and to that between 'objective' and 'subjective' social class by students of stratification. The same distinction is presented in another form in what Naroll (1964: 306) calls Goodenough's rule: what we do as ethnographers is, and must be kept, independent of what we do as comparative ethnologists (Goodenough 1956: 37), The closest analogue to Schneider's contrast is found in the distinction between etic and etnic categories, labels which some social anthropologists have taken ovw from linguistics (see Goodenough 1970 b: 98-130), In Pike's (1967: 37-72) formulation (he elic-emic contrast is unashamedly positivisl. The scientific linguist observer, with his objective categories, is contrasted with the speaking actor who uses subjective categories to produce and decipher meaningful utterances. Inter-language comparison is implicit in Pike's scheme. As used in anthropology the cross-cultural and cross-societal emphasis has been retained but the positivisl implication of the contrast has been played down. Instead we have the emic categories of thought of the actors contrasted with the etic categories of analysis of the observer, neither set necessarily more real or true than the other. I have argued that this contrast can be applied without modification only in the 'colonial* or laboratory situation such as Pike had in mind (Barnes l°67 />. 1970). If the actors speak only their own language, think only in their own terms, and draw only upon a locally-generated slock of 'knowledge' of their environment, then the flow of information is only one-way. J. A. Homes The observer may well modify his etic analytical categories in the light of what the actors do and say, but they do not alter their ways of thinking and acting because of assertions made about their behaviour by the observer. This is the paradigm situation of inquiry in natural science, the principle of indeterminacy notwithstanding. Until a couple of decades ago anthropological fieldwork in distant colonies approximated to it, though even then there were substantial and critical differences from the typical scientific laboratory. These laboratory-like conditions have not persisted and, following the end of colonialism in its classic form, they are probably now gone for ever. Instead there is two-way communication between actors and observers, so that the actors begin to take over not only the material artefacts brought by the observers and their compatriots but also their languages, concepts and social institutions, changing them in the process. In particular they take over and adapt some of the jargon and some of the content of science. Once this happens the observer in the remotest jungle begins to face the same difficulties as his colleagues working in the metropolis have always faced: the facts that the language of science, and of social science in particular, is also to some extent the language of the people and that the findings of science, and even its techniques of inquiry and verification, are continually seeping into popular consciousness. In general, then, though his reasons are different from mine. I accept Schneider's point that the categories and concepts of the observer's own culture are the starting point for comparative analysis in social science. If this is so. how does it affect the study of kinship? There are many issues wc might take up but I want here to consider just two related matters: how valid is the distinction Schneider draws between culture and natural science; and how does kinship, in contrast to other aspects of social and cultural life, relate to nature? I use the standard triple distinctions between genetic or carnal father, genitor and pater, and between genetic or carnal mother, genetrix and mater, stressing that the statuses of genitor and genetrix are defined, if at all, in terms of local doctrines about the process of human reproduction (Barnes 1961: 297-8; 1964: 294; Goodenough 1970 b: 27). Fatherhood and motherhood are used as cover terms. At first glance Schneider's position seems to be paradoxical. He seeks to establish science as distinct from culture and yet to insist that a comparative science of cultures has to be rooted in a particular culture, the culture of the investigator. He appears to make natural science free of culture but to query the possibility of meta-categories for analysing cultures. But this apparent paradox can be quickly disposed of by referring to his book on American kinship, where he makes a fourfold contrast between (I) what he calls biological facts, (2) formal science, (3) informal elhnoscience and (4) certain cultural notions which are put, phrased, expressed, symbolized by cultural notions depicting biological facts, or what purport to be biological Gtnetrix : Genltor a Nature: Culture? J. A. Barnes facts' [Schneider 1968: 114- 15). I find it confusing lo use 'biology", the name of a science, for phenomena that exist independently of efforts tn study them, and therefore re-label (I) nature. Category (4), of which 'broken heart' and heartache* are examples from American culture, need not detain us. This category contains what in more traditional language might be called extensions of kin usages that are perceived by the actors as being metaphorical, figurative, symbolic; all the parishioners know that the village priest is not 'really* their father. We can concentrate on categories (I) (2) and (3). Formal science, category (2), is part of American culture as much as categories (3) and (4). Indeed Habermas (1972) argues that the salient diagnostic Feature of contemporary culture in industrialized societies is the belief that science is the only authenticated form of knowledge. Though both are part of Western culture it is possible, at least for classical times and since the Renaissance, to draw a fairly clear distinction between professional scientific assertions and lay beliefs that, rightly or wrongly, are perceived as based on formal scientific inquiry (see S. B. Barnes 1969). Informal ethnosciencc embraces more than the latter category but it certainly includes it. Weber (1946: 139) notes this distinction, and it is well put by Evans-Pritchard. in a discussion of the views of Levy-Bruhl. The fact that we attribute rain to meteorological causes alone while savages believe that Gods or ghosts or magic can influence the rainfall is no evidence that our brains function differently from their brains.. .It is no sign of superior intelligence on my part that I attribute rain to physical causes, I did not come to this conclusion myself by observation and inference and have, in fact, little knowledge of the meteorological processes that lead to rain. I merely accept what everybody else in my society accepts, namely that rain is due to natural causes (Evans-Pritchard 1934: 21). Our yardstick, then, is our own culture, which contains a vast number of propositions perceived as science. Against it we compare other cultures, noting in what respects they resemble one another and how they differ, and endeavour to discover why this is so. How does a comparison of this kind work in the field of kinship? It is reasonable to expect that data from category (I) will impinge in special fashion on kinship data from categories (2) and (3). Despite the recent efforts of some cthologists to postulate a pan-primate basis for political order, and for much else as well, kinship remains the aspect of human culture with the closest links to the natural world. Indeed, in American culture, we are told, "kinship 1$ biology' (Schneider 1968: 116). Apes and monkeys may have dominance hierarchies and territories but, unlike men, they du not have representative government nor, as far ac we know, do they believe in God. Like us. however, they copulate, conceive and give birth, activities with which kinship has a close connexion, however problematic the qualities of the connexion may be. These activities, when performed by humans, are perceived as natural rather than cultural. Part of the basis for a comparison of ideas of kinship has then to be our own cultural notions about the reproductive process, some of which are derived directly from formal science but which include others that belong solely to ethnosciencc. The inevitability of beginning cross-cultural comparison by matching alien cultures against our own is well shown by the discussion in Maa a few years ago on virgin birth (Leach 1967; Spiro 1968; Douglas 1969 and references therein), and by earlier controversies about the ignorance of physiological paternity. The diverse beliefs about non-miraculous human reproduction found in pre-scientific cultures have been described many times and need not be repeated here (Ashley-Montagu 1937, 1949: Ford 1945; Leach 1961 b, 1961 c, 1967; Malinowski 1963; Meyer 1939: 1-16; Spencer 1949-50). The point I emphasize is that when these beliefs arc compared, the yardstick used is falsely presented, for we tend to assume that for ourselves no distinction between formal science and informal ethnosciencc is needed. We present our own view of conception as a single event, in which only one man and one woman are involved, and which triggers the whole sequence of gestation, as scientifically validated. We contrast this view with theories that the foetus forms and grows in the womb by receiving contributions via many acts of coitus not necessarily all performed by the same man. a view held, for example, by the Azande (Evans-Pritchard 1932: 410); or with other theories, found for example in Aboriginal Australia, whereby the process of gestation is neither initiated nor sustained by coitus; or with intermediate theories. These indigenous ideas are recorded in the ethnographic literature, but where do ours come from7 From formal science, or informal elhnoscience, or from a cultural heritage in which natural substances like blood and semen serve as symbols in statements that have nothing at all to do with natural science? Even if we prune away metaphorical ideas in Schneider's category (4), a moment's reflection shows that ideas in categories (2) and (3) are not as easy to pin down as may seem at first sight. At this point we can come to grips with a distinction between fatherhood and motherhood. Consider first fatherhood, Nowadays most educated people in the West have heard of genes and chromosomes and know thai the embryo draws its stock of chromosomes equally from its genetic father and mother. I guess that, in the sex-conscious culture of contemporary Britain, almost all adults believe that conception occurs when a spermata-zoon penetrates an ovum. But what sort of knowledge is this? Surely most of us know as little about the physiology of human reproduction as Evans- Gcnetri.x: Geititor:; Nature : Culture? Pritchard knows about meteorology. We believe these processes lo occur because we believe also that al some poinl in I he past long-forgotten scientists discovered (hat this is what ready happens. We assume that though the discovery of genes and chromosomes is post-Darwin, the Tact that conception is a unique event and not a prolonged process has been scientifically established for a iong time. The view that conception and gestation can Follow a single act of coitus is indeed consistent with Aristotle's account of reproduction in Generation of Animate, Book 2, and is implied in his statement in llie History of Ant mate that 'if the second conception take place at a short interval, then the mother bears that which was later conceived and brings forth the two children like actual twins. .. The fotlowing is a striking example: a certain woman, having committed adultery, brought forth the one child resembling her husband and the other resembling the adulterous lover' (585;|). Thus the doctrine of 'one child, one genitor" has been part of the Western tradition for more than two thousand years. Yet although the presence of physical resemblances between some, though not all. children and their mother's husbands calls for an explanation, it does m>l necessarily demand a theory of universal monopatemity. The dominance of a monopatemal theory cannot have been determined by the we I glit of evidence, for apart from resemblances there was little material evidence available until |he seventeenth century. Spermatozoa were discovered accidentally in 1677 by Ham. though their connexion with fertilization remained unknown. Mammalian ova were discovered, alsy accidentally, in a pregnant bitch by van Bacr in 1828 and in 1353 Newport claimed to observe spermatozoa entering an ovum. Not until IS75 were the male and female pronuclei in spermatozoa and ova identified by Oscar Mcrtwig. who described how they combine (Meyer 1939: 123. 137-S, 189-192). Thus for most of the historic period in the West, the uniqueness of physical paternity was a cultural construct for which there was very little conclusive evidence-Even so, this doctrine was modified by a belief in 'maternal influences', the idea that events experienced by a pregnant woman are reflected in the constitution of her child. The belief forms part of several indigenous theories of procreation (e.g. Levi-Strauss 1966 b: 76) and is certainly still present in contemporary Britain. It is exemplified for animals in the story told in Genesis, chapter 30, verses 25-43, about Jacob changing the colour of the lambs borne by La ban's ewes. 'Maternal influences' may always have been restricted to elhnoscience, old wives* tales, but orthodox formal science long entertained the related idea that Weismann (1893: 383) calls 'telegony'. the notion that the physical characters inherited by an individual are influenced not only by his (or her) own father but also by other men by whom his mother may previously have had children, Dobzhansky (1970: 420A) attributes this belief to Aristotle and it was supporled, for plants and /. A Parties animals at least if not for humans, by Darwin (1875: 435-7; sec Morton 1821: Zirkle 1935: 117 and 1946: 119; Parkes 1960: 242) in conformity with his thesis of pangenesis. Thus whereas most pre-scientific beliefs about multiple physical fatherhood identify as genitors men with whoin a woman has had intercourse during a given pregnancy, telegony ascribes physical paternity to her earlier mates as well as to the man who initiates the pregnancy. The doctrine of telegony lives on among animat breeders but has been abandoned by orthodox science, as has a later suggestion of a naturally-occurring poly paternal process called 'somatic fertilization'. According to this hypothesis, substances may be absorbed in the female genital tract after copulation; these evoke the production of factors which may exert an influence on the embryos of subsequent ma tings (Austin and Walton 19(50: 393; Parkes I960: 242). ]n the laboratory, however, the fusion of two embryos at the eight-cell stage has been achieved, producing tetra-parentai mice. Chimeric mice with even more complex constitutions have been bred and studied (Tarkowski 1961; see Wegmann 1970; Mullen and Whitten 1971 and references therein). Indigenous assertions of human poly paternalism m nature have thus been vindicated for some mammals in the laboratory. Indeed there is evidence that double fertilization sometimes occurs naturally in humans (Benirschke 1970: 40-5). Human polypaternalism seems therefore to be compatible with the available scientific evidence. Tetraparental mice and other chimeras produced in the laboratory receive their diverse constituents before the implantation stage, Jong before birth. A belief in the post-natal physical transmission of information and attitudes is implied in the expression 'He took that in with his mother's milk.' An earlier belief in a more specific and selective form of located transmission is suggested by Dobzhansky's (1970: 420A) statement, made in the context of an article on heredity, that'An ancient English law holds a man who seduces the wet nurse of the heir to the throne guilty of polluting the "blood" of tbe royal family.' I have been unable to trace this law. The closest comparable laws seem to be those listed during the reign of King jEthelberht of Kent about A.D. 600, whereby a man who seduced a maiden of the king's household had to pay fifty shillings in compensation, compared with only twelve shillings for the seduction of a girl occupied on menial tasks (Attenborough 1922: 5; Liebcrmann 1903: 3 and 1916: 7). These laws give special recognition to the king's entourage but make no reference to suckling or pollution. It may well be that the ultimate source for the alleged ancient law is merely the Mirror of justice* where it is said that one of the ways in which an adulterer may commit the crime of lese-majesty, 'a horrible sin', is by seducing the nurse suckling the heir of the king (Whitlaker 1895: 15). The Mirror was at one lime regarded as a true account of the laws of England before the Norman conquest but in Mailland's view was largely fabricated by Andrew Horn, fishmonger and Chamberlain of the City of London, in Geitelrix: Geuitor:: Nature : Culture? about 1289; it contains many wilful falsehoods and misstatements of law (Maitland 1895). The Anglo-Saxons may never have held the doctrine thai sonic kind of malign influence can be transmitted from a man by adulterous copulation to a laciatiiig woman and thence through her milk to her royal Foster-child But if the law never existed, at least the doctrine formed prt oF the imagination of a thirteenth-century fishmonger. Despite these contrary notions, the main stream of Western popular belief has clearly been 'one child, one genitor*. If there was no compelling scientific evidente for this belief the reasons for its persistence must be flight elsewhere, in the organization of social life and in other parts of Western culture, rather than in nature. As far back as we have knowledge. Western society, like most other human societies, has been organized on the premise of one child, one pater. Likewise the Christian faith of the Wesl stresses Ihe uniqueness of God the Father. The Holy Ghost impregnated Mary through her ear and was manifest in, or symbolized by, a dove al Christ's baplism, but neilhcr act makes the third person of the Trinity co-pater with the lirsl (see Jones 1951; Swete 1909 : 28-9, 45. 365-6; Gudeman 1972: 54). If we encountered this constellation of facts in a tribal society, surely we would have no hesitation in saying that the organization iff society and the major premisses of religion are reflected in myths about unique physical parenthood. Molherhood is different. Conception is an internal and microscopic event that we laymen believe scientists have investigated, whereas gestation and birth, and with them the relation of physical molherhood, are macroscopic processes lhat, in principle, anyone can sec for himself. Hence the descriptions of physical motherhood in diverse cultures do not vary as greatly as with fatherhood. The so-called denial of physical maternity is not homologous with the denial of paternity, except when applied to special myths for uninitiates, as for example in our own tale for children about storks bringing babies (cf. Spire- 1968; 260, n, 11), The denial of physical maternity usually means merely that the mother is thought to contribute nothing oF importance to the foetus during pregnancy* as for example was believed in ancient EgypL (Necdham 1959: 43) and is staled by Apollo in Aeschylus1 Eiimenidcs I lines 657-61), when defending Orestes against the charge of matricide. This lack of symmetry between the notions of genitor and genetrix is emphasized by Goodenough (1970 cr; 392) who says lhat 'procreation ussociates children directly with women but only indirectly with men' and that "Motherhood and fatherhood cannot be defined in Ihe same way for comparative anthropological purposes*. Fathers are not self-evident as mothers are. 'Genilor' is a social status, and societies vary greatly in the rights and duties, privileges and obligations, if any, lhat they associate with this status. If the status exists, there musl be a rule for identifying genitors. J. A. Barnes But for the status to exist at all there must be a theory of procreation that calls for one. or for several, and, for all cultures prior to the physiological discoveries of the late nineteenth century, this theory cannot be supported by scientific evidence. Even though Aristotle wrote his Generation