Engels was Right: Early Human Kinship was Matrilineal

Mother and baby, Hadza tribe, Tanzania
Mother and baby

The earliest human institution was not the nuclear family. The latest research now indicates that it was the communistic, female-centred clan.

Engels argued this in 1884. After a century of establishment denial, including complicity by 'Marxists', it now turns out that he was right after all.

Submitted by hedgehog on September 23, 2012

by Chris Knight

In 2008, the Royal Anthropological Institute published a scholarly volume entitled Early Human Kinship.[1] It stemmed from a 2005 workshop held in Gregynog, Wales, as part of the prestigious British Academy Centenary Project ‘From Lucy to Language: The Archaeology of the Human Brain’.

It’s common knowledge among anthropologists – part of our origins myth – that Frederick Engels published an influential essay on world prehistory, The Origin of the Family Private Property and the State, shortly after the death of his friend Karl Marx[2]. Engels argued that kinship was originally matrilineal. In association with matrilineal descent, Engels continued, every child was raised not just by its parents alone but by the whole clan.

In the early years of the twentieth century, especially following the Russian revolution, these ideas became associated with ‘Bolshevism’ and deemed so inflammatory that, outside the Soviet Union, they were effectively suppressed. As politicians and clerics railed against the threat to ‘marriage and the family’, their counterparts in the universities re-established anthropology on a completely new footing, defining it in opposition to Darwinism on the one hand, Marxism on the other.[3]

To my knowledge, that workshop in Gregynog was the first time for something like a hundred years that professional anthropologists had summoned up the courage to re-open these debates. In my talk and eventually published chapter, I argued quite bluntly that Engels was essentially right: we now have compelling evidence that early human kinship was indeed matrilineal.[4] To my delight and astonishment, most of my professional colleagues participating in the Gregynog workshop tended to agree.

But why did we have to wait so long? Why, for the best part of a century, were academics barred from even debating questions of this kind? To explain that, we must take a step back into history.

In the immediate aftermath of World War I, anthropology in Britain and America became established for the first time as a professional discipline. This new discipline was set up with a political mandate: to overturn the materialist paradigm established in the pre-war period by Lewis Morgan, Frederick Engels and Karl Marx. In particular, the mission of the new anthropology was to debunk the theory – until then accepted by almost everyone – that the primordial human social institution had been not the individual family but the matrilineal clan. The last prominent exponent of the ‘matrilineal group motherhood’ theory of human origins was Robert Briffault, who published his encyclopaedic three-volume work, The Mothers, in 1927. As it turned out, Briffault’s monumental effort – nowadays comprehensively forgotten – was the final gasp of the Morgan-Engels paradigm. It served as a red rag to a bull. The principal founder of the new professional discipline of anthropology – the LSE’s Bronislaw Malinowski – responded as if personally affronted. Group motherhood, he declared, is not only a false but a dangerous idea. Society was, is and will always remain founded on individual marriage and the family.[5]

Most Marxists today prefer to ignore this whole controversy, regarding it as somewhat embarrassing. After all, didn’t Engels celebrate not only communism in general but sexual communism in particular? Didn’t he talk about group motherhood and primitive matriarchy? That kind of nonsense can’t possibly be right. Why should today’s Marxists feel committed to the details of Engels’ lurid speculations – when these in turn were based on conjectures by Lewis Morgan which everyone now knows to be wrong?


Since Engels’ actual words are today scarcely remembered, let me begin with a direct quotation. In the following passage, Engels celebrates Morgan’s ‘rediscovery of the original mother-right clan’:

‘The rediscovery of the original mother-right clan has the same significance for the history of primitive society as evolution has for biology and Marx’s theory of surplus value has for political economy. It enabled Morgan to outline for the first time the history of the family. Clearly this opens a new era in the treatment of primitive societies.’[6]

Engels went on to describe ‘the mother-right clan’ as ‘the pivot around which the entire science turns.’

Engels based his ideas not simply on Lewis Morgan. An equally important source was Marx’s early notes. In 1844, he and the young Marx had declared that ‘the immediate, natural and necessary relationship of human being to human being is the relationship of man to woman’, adding: ‘From this relationship the whole cultural level of man can be judged’.[7]

When Marx discovered Morgan’s work, he took it as confirmation of his own youthful ideas. In The German Ideology, Marx and Engels had contrasted primitive communism with the subsequent emergence of property, ‘the nucleus, the first form of which lies in the family, where wife and children are the slaves of the husband’.[8] Engels’ ideas in The Origin of the Family didn’t come out of the blue. He based them on a lifetime of collaborative intellectual work with Marx, much of which had never been published or completed.

According to Morgan, the domestication of cattle disempowered women, breaking up their earlier sisterhood. Previously, women had cared for their children within a co-operative framework, sharing the burdens with their mother and other female relatives. One down-to-earth traditional custom made this possible. Throughout her life, a woman remained living in the same household with her mother and siblings. Children would be raised by a confederacy of mothers, daughters and sisters, backed up by male kin and possessing sufficient solidarity and authority to prevent any husband from taking them away or breaking them apart. A certain loose kind of ‘marriage’ was consistent with these arrangements, but in-marrying spouses had to agree to visit their brides within these women’s communal home. No woman could be privatised by her spouse or taken away.

According to Morgan, this arrangement – known technically as ‘matrilocal residence’ – originally prevailed in all human societies across the world. So what happened to change such time-honoured arrangements? According to Morgan, the key factor was the domestication of animals, notably cattle. A man possessing a herd could bargain with his future bride’s kinsfolk, allowing them to keep his cows in exchange for their daughter, whom he could now take away. Under the previous matrilocal system, his bride would have been free to remain with her natal family, benefitting from shared childcare, the children’s father being considered just a visitor. The introduction of cattle triggered a general shift from matrilocal to patrilocal residence. This change ‘reversed the position of the wife and the mother in the household’. Morgan explains:

‘She was of a different clan to her children: she had been pulled away to live with her husband. As well as being a different clan from her husband, she was isolated from her kin in the separate and exclusive house of her husband. Her new condition tended to subvert and destroy the power and influence which descent from the female line that the joint tenement houses had created.’[9]

Engels added political impact to all this: ‘The man took command in the home also; the woman was degraded and reduced to servitude; she became the slave of his lust and a mere instrument for the production of children.’[10] He continued: ‘The first class opposition that appears in history coincides with the development of the antagonism between man and woman in monogamous marriage, and the first class oppression coincides with that of the female sex by the male.’[11]


That is the theory, which few Marxists these days defend. From the moment of Malinowski’s onslaught during the 1920s, western Marxism began caving in. Soviet official communism paid lip service to the theory, perhaps mentioning it here and there, but did nothing to develop or advance it. Meanwhile, left-leaning intellectuals in the western world abandoned the paradigm wholesale. The consequences remain with us to this day. Go up to any ordinary member of today’s Socialist Party, Socialist Workers’ Party, Workers’ Liberty Alliance or whatever. Ask whether they defend Engels on the matrilineal clan. They won’t know what to say.

Of course, it could be that Engels was wrong. Maybe the whole theory of matrilineal priority was just a mistake. Maybe Morgan himself got it wrong. Perhaps he was over- influenced by the fact that he underwent an initiation ceremony conducted by Iroquois Indians, whose traditions had been strongly matrilineal. Possibly he just extrapolated from Iroquois tradition or even myth, incorrectly projecting what he was told onto human history as a whole. If Morgan got it wrong, then as Marxists we must put science first, abandoning outdated nineteenth century dogma.

But you might have expected the left intellectuals to probe a little. Shouldn’t they have asked questions about this sudden reversal in anthropology? Virtually all prominent thinkers in the 1880s and 1890s had accepted one or another aspect of Morgan’s theory. Well into the twentieth century, intellectual giants such as Emile Durkheim and Sigmund Freud continued to accept Morgan on the question of the matrilineal clan. Then, just after World War I, the new discipline of anthropology in the universities suddenly did an about-turn. Might we not have expected that Marxists – of all people – would be slightly suspicious of that? Mightn’t they have suspected an element of politics behind such a change? I find it quite staggering that out of all the New Left Review types, all the Althusserians, all the postmodernists, all those fashionable intellectuals who called themselves Marxists, not one sought to fight back. To this day, no one seems to have bothered to ask even the most basic question: What actual evidence is there that Morgan and Engels got everything so wrong?

As we review the historical evidence, it becomes clear that political passions were never far beneath the surface and ultimately played the decisive role. With regard to the topic of ‘primitive promiscuity’, Engels commented: ‘It has become the fashion of late to deny the existence of this initial stage in the sexual life of mankind. The aim is to spare humanity this “shame”’.[12] The reference here was to Edward Westermarck, scholarly defender of individual marriage and the family who was later to inspire the young Malinowski. Westermarck had chosen to turn public opinion against the theory of ‘primitive promiscuity’ by associating it with modern prostitution. To this, Engels retorted: ‘To me it rather seems that all understanding of primitive conditions remains impossible so long as we regard them through brothel spectacles’.[13]

Once Engels had incorporated Morgan’s findings into the socialist canon, however, no one could write neutrally on such topics any more. Morgan’s Ancient Society, as the American ethnographer Robert Lowie was later to comment,

‘attracted the notice of Marx and Engels, who accepted and popularised its evolutionary doctrines as being in harmony with their own philosophy. As a result it was promptly translated into various European tongues, and German workingmen would sometimes reveal an uncanny familiarity with the Hawaiian and Iroquois mode of designating kin, matters not obviously connected with a proletarian revolution’.[14]

Once Engels had endorsed it, Morgan’s theory was destined to become a casualty of the central conflict of the age. Social anthropologists may like to imagine that their discipline became shaped in its modern form quite independently of Marxism. It would be more accurate to describe it as moulded specifically in reaction against the ideas of Engels and Marx. ‘With Morgan’s scheme incorporated into Communist doctrine’, observes Marvin Harris, ‘the struggling science of anthropology crossed the threshold of the twentieth century with a clear mandate for its own survival and well-being: expose Morgan’s scheme and destroy the method on which it was based’.[15]

A widespread consensus developed on both sides of the Atlantic that regardless of the intellectual merit of Morgan’s ideas, ‘group motherhood’ was in any event too dangerous a concept to be allowed. A radio broadcast by Malinowski revealed his state of mind:

‘A whole school of anthropologists, from Bachofen on, have maintained that the maternal clan was the primitive domestic institution…. In my opinion, as you know, this is entirely incorrect. But an idea like that, once it is taken seriously and applied to modern conditions, becomes positively dangerous. I believe that the most disruptive element in the modern revolutionary tendencies is the idea that parenthood can be made collective. If once we came to the point of doing away with the individual family as the pivotal element of our society, we should be faced with a social catastrophe compared with which the political upheaval of the French revolution and the economic changes of Bolshevism are insignificant. The question, therefore, as to whether group motherhood is an institution which ever existed, whether it is an arrangement which is compatible with human nature and social order, is of considerable practical interest’.[16]

While denouncing ‘ideology’, Malinowski nonetheless saw it as his scholarly duty to ‘prove to the best of my ability that marriage and the family have been, are, and will remain the foundations of human society’.[17] He insisted that ‘marriage in single pairs – monogamy in the sense in which Westermarck and I are using it – is primeval’.[18] It’s worth remembering here that the Finnish historian of marriage attributed ‘marriage in single pairs’ equally to chimpanzees and gorillas, believing human marriage to have been inherited from a primate precursor.

Malinowski’s assertion that ‘monogamy’ must be ‘primeval’ fits uneasily with declarations such as the following: ‘I would rather discountenance any speculation about the ‘origins’ of marriage or anything else than contribute to them even indirectly….’[19] Notable here is Malinowski’s tactic of dissociating himself from evolutionary research while specifying his opinion as to the ‘initial situation’ for human kinship. Throughout much of the twentieth century, as Adam Kuper records, the strategy of smuggling in assumptions about ‘origins’ and ‘initial situations’ without having to justify them proved popular among anthropologists of virtually every school.[20]

In the United States, Franz Boas initially accepted the Bachofen-Morgan scheme according to which descent systems invariably underwent historical change from matriliny to patriliny and not the reverse. In support of this was a ‘complete lack of historically attested, or even inferentially probable, cases of a direct transition from patrilineal to matrilineal descent’.[21] Boas later came to believe, however, that he might discredit Morgan if he could find a single exception.

On Vancouver Island, the Kwakiutl Indians were organised in groups known as numaym, which Boas translated initially as ‘clan’ or ‘gens’. He explained that there was no consistent rule of descent: ‘The child does not belong by birth to the gens of his father or mother, but may be made a member of any gens to which his father, mother, grandparents, or great-grandparents belonged’.[22] Six years later, however, Boas changed his mind, attributing to the numaym now ‘a purely female law of descent’, albeit one secured ‘only through the medium of the husband’.[23] However, anything short of a purely patrilineal system switching to a purely matrilineal one might still have allowed Morgan’s evolutionist scheme to survive. Boas duly supplied the requisite categorical formulations. Although the Kwakiutl had today ‘a purely female law of descent’, he now proclaimed, the ‘organization must have been at one time a purely paternal one’.[24] For the very first time, a unilineal descent system had been found changing in the reverse direction from that stipulated by Morgan.

The loyalty of Robert Lowie to his great friend and teacher, Franz Boas, could hardly be in doubt. In 1914, however, even this ardent disciple admitted that the Vancouver Island data had been stretched to fit the case. Although ‘the Kwakiutl facts are very interesting’, as he put it, ‘it is highly doubtful whether they have the theoretical significance ascribed to them’.[25] Most awkward was the fact that the Kwakiutl numaym groupings central to Boas’ entire argument were not unilineal descent groups at all. Neither ‘matriliny’ nor ‘patriliny’ were applicable concepts. ‘For these reasons,’ as Lowie put it, ‘the Kwakiutl conditions do not seem to furnish a favorable test case.’ However, such scholarly reservations did nothing to stop Boas or his students from continuing to disseminate the myth. The ‘extreme interest in Boas’ handling of the numaym’, as the historian Marvin Harris comments in his analysis of the whole shameful episode,

‘stems from the fashion in which he and his students seized upon this case to destroy the supposed universal tendency for patrilineality to follow matrilineality and at the same time to discredit the entire historical determinist position. On the basis of this one drastically deficient case, there gradually diffused out of Schermerhorn Hall at Columbia, through lecture, word of mouth, article and text, the unquestioned dogma that Boas had proved that it was just as likely that patrilineality succeeded matrilineality as the reverse’.[26]

Meanwhile across the Atlantic, A. R. Radcliffe-Brown intervened with his celebrated article entitled ‘The mother’s brother in South Africa’.[27] It was this intervention which by general consent – at least among British structural-functionalist anthropologists – buried once and for all the evolutionist theory that the mother’s brother relationship in patrilineal societies was a survival left over from an earlier matrilineal stage. Radcliffe-Brown’s specific target was a comprehensive monograph on the Thonga of Mozambique.[28] The following features seemed to require explanation:

• Although inheriting clan membership from their father, Thonga children on being weaned went to live in their mother’s brother’s village.
• A man without patrilineal heirs could require a sister to remain in his settlement, her male offspring continuing his lineage.
• Even when a man did have patrilineal heirs, his sisters’ sons could claim items from his own estate.
• The maternal uncle had a share in the bride price received for a sister’s daughter.
• The maternal uncle and not the father officiated at the sacrifices in a young man’s life-crisis ceremonies.

Henri Junod himself had interpreted these features as clear evidence that the Thonga were not straightforwardly patrilineal but were embroiled in a difficult and sometimes contradictory process of transition from matrilineal to patrilineal descent. For Radcliffe-Brown, it was axiomatic that any such ‘pseudo-history’ had to be repudiated. The various components of a social system should instead be explained in structural-functionalist terms – that is, by invoking fixed laws on the model of physics and chemistry.

He now proposed his celebrated structural functionalist explanation. The involvement of the mother’s brother in the upbringing of a Thonga (or Tsonga) boy has nothing whatsoever to do with past or present matrilineal descent. It’s just an expression of a fixed and invariant sociological principle – the principle of ‘the equivalence of siblings’. Since a woman and her brother are equivalents, any human child’s feelings toward its mother will naturally tend to include to her brother as well. Radcliffe-Brown termed this universal psychic mechanism ‘the extension of sentiments’, concluding after a few further observations that everything was now satisfactorily explained.

With the benefit of hindsight, what are we to make of this short essay? Let me begin by recalling some fundamentals of matrilineal versus patrilineal descent. Underlying everything is the fact that the alternatives, for any woman, involve a choice. To put her brother first in her children’s lives is to put her husband second; conversely, if she puts her husband first, her brother must come second. It really is as simple as that: you can either have brother-sister unity as the fundamental principle or else you can have husband-wife unity, but you cannot have both at the same time.

From this it follows that brother-sister unity cannot possibly be ‘a universal sociological principle’. Matriliny presupposes brother-sister unity; patriliny entails just the opposite. In its pure form, it requires men to let go of their sisters, instead monitoring the fidelity of their wives. Where the husband’s rights prevail, the wife to that extent yields control over her fertility to him and his kin, weakening her bond with her brother, enhancing paternity certainty and thereby favouring patrilineal descent.

Let us suppose, however, that after marriage, a woman chooses to remain with her brother. This can only be at the expense of her bond with her husband – reducing paternity certainty and hence favouring matrilineal descent. To sum up: the very factor invoked by Radcliffe-Brown as an alternative to the matrilineal complex – namely, brother-sister solidarity – turns out to be a covariant feature of that complex itself. To invoke ‘brother-sister unity’ as an explanation for the mother’s brother relationship is no more than to invoke an aspect of the matrilineal complex while concealing it under another name.

The renowned comparative anthropologist G. P. Murdock long ago poured scorn on Radcliffe-Brown’s whole approach:

‘In the first place, the alleged principles are mere verbalizations reified into causal forces. In the second, such concepts as ‘equivalence of brothers’ and ‘necessity for social integration’ contain no statements of the relationships between phenomena under varying conditions, and thus lie at the opposite extreme from genuine scientific laws.’[29]

Exactly so.

Ironically, Murdock’s subsequent historical research on the Thonga confirmed that they were indeed in the throes of transition from matriliny to patriliny just as Junod had originally claimed.[30] In conformity with Morgan’s scheme, the rise of alienable property – typically cattle – is the crucial factor cementing marital bonds at the expense of brother-sister solidarity throughout much of sub-Saharan Africa. As everyone now agrees, ‘the cow is the enemy of matriliny’.[31] Following Murdock’s cross-cultural comparative work, a recent phylogenetically controlled analysis has confirmed a negative correlation between African matriliny and cattle-owning.[32] In their most recent analysis of matriliny as daughter-biased investment, the researchers comment: ‘the two factors Morgan identified, heritable wealth and paternity uncertainty, remain central to our understanding of variation in matriliny and patriliny in human social organisation’.[33] I should emphasise that this isn’t controversial. Among anthropologists who specialise in such topics, it’s the consensual position.


Despite this, when scholars turn to ultimate origins, thy tend to return to the old stereotypes. Most continue to invoke paternity certainty as key to the process leading from Plio-Pleistocene hominin to modern Homo sapiens.

In typical versions of the story, paternal investment is linked directly to the sexual division of labour, food sharing, lengthy juvenile dependency, ovulation concealment and continuous female sexual receptivity. The idea is that since the human female produces such unusually helpless and dependent offspring, she needs a man to provide long-term pair-bonding commitment and support. The catch is that no man should enter such a contract unless confident that his partner will be faithful to him in return. ‘In evolutionary terms’, as Terrence Deacon puts it, ‘a male who tends to invest significant time and energy in caring for and providing food for an infant must have a high probability of being its father; otherwise his expenditure of time and energy will benefit the genes of another male’.[34]

Dating from the 1960s and 1970s, this scenario has become in effect the Standard Model of Human Evolution. Textbooks on human evolution go so far as to impute monogamy to Homo erectus:

‘Females may have had difficulty providing food for themselves and their dependent young. If H. erectus hunted regularly, males might have been able to provide high-quality food for their mates and offspring. Monogamy would have increased the males’ confidence of paternity and favored paternal investment.’[35]

Evolutionary psychologist Steven Pinker uses the same idea to explain why the sexual double standard is natural and inevitable:

‘Sexual jealousy is found in all cultures…. In most societies, some women readily share a husband, but in no society do men readily share a wife. A woman having sex with another man is always a threat to the man’s genetic interests, because it might fool him into working for a competitor’s genes’.[36]

In response to such dogmatic statements, Stephen Beckerman and Paul Valentine have assembled counter-evidence from a substantial number of lowland South American societies. In their book, Cultures of Multiple Fathers, they demonstrate how – in direct refutation of Pinker – the paternity of a woman’s baby becomes partitioned among multiple males. They quote this passage from a classical account of the Xocleng (previously Kaingang):

‘Klendó’s daughter, Pathó, is my child’, said Vomblé. ‘How do you know,’ said I, ‘since Klendó also lay with her mother?’ ‘Well, when two men lie with a woman they just call her child their child.’ But not only do men feel that their mistress’s children are their children, but people whose mothers have had intercourse with the same man, whether as lover or husband, regard one another as siblings’. [37]

If such ‘partible paternity’ (as the authors term it) were found in only a tribe or two, it could perhaps be dismissed as an aberration. However, the institution is widely distributed across Lowland South America and found among peoples whose traditions diverged millennia ago – as evidenced by the fact that they live thousands of kilometres apart, speak unrelated languages, and show no indication of having been in contact for centuries. The authors continue:

‘It is difficult to come to any conclusion except that partible paternity is an ancient folk belief capable of supporting effective families, families that provide satisfactory paternal care of children and manage the successful rearing of children to adulthood. The distributional evidence argues that it is possible to build a biologically and socially competent society – a society whose members do a perfectly adequate job of reproducing themselves and their social relations – with a culture that incorporates a belief in partible paternity.’[38]

Not only is the belief compatible with successful reproduction. It may even help babies to survive. Among the hunting and gathering Aché, children with one extra father are significantly more likely to reach maturity,[39] a correlation confirmed by a longitudinal statistical study among the Barí.[40]

Can a woman really help her baby by taking extra lovers during pregnancy? The answer seems to be yes. The explanation is probably that additional fathers contribute additional provisions and more protection against infanticide. It is not in a woman’s interests to encourage the men in her life to engage in contests over biological paternity. From a woman’s standpoint, the truth is that her current husband may become injured, die or abandon her. In any event, she may have good reason to switch to a new man. If her new mate suspects that he is not this or that child’s father, her existing offspring might suffer infanticide or abuse. Loss of a wanted child is enormously costly to any human mother, making it best not to divulge but precisely to confuse accurate paternity information, taking extra lovers to distribute illusions among multiple males.

Whether these males contest or collude depends on the balance of costs and benefits involved. Where males strive to contest paternity, females may have an interest in driving up the costs. Beckerman and Valentine view the range of variation as reflecting ‘a competition between men and women over whose reproductive strategies will dominate social life’. In small-scale egalitarian societies, they continue,

‘women’s reproductive interests are best served if mate choice is a non-binding, female decision; if there is a network of multiple females to aid or substitute for a woman in mothering responsibilities; if male support for a woman and her children comes from multiple men; and if a woman is shielded from the effects of male sexual jealousy. Male reproductive interests, contrariwise, are best served by male control over female sexual behavior, promoting paternity certainty and elevated reproductive success for the more powerful males. This profile implies that men choose their own or their sons’ wives, and their daughters’ husbands; that marriage is a lifetime commitment and extra-marital affairs by women are severely sanctioned; and that this state of affairs is maintained by disallowing women reliable female support networks, or male support other than that of the husband and his primary male consanguines.’[41]

In humans as in other sexually reproducing species, neither sex is likely to succeed in imposing its strategies to the exclusion of those of the opposite sex.[42] Yet there are situations that may give the edge to one side or the other. In the human case, when male strategies dominate, patrilineal descent and patrilocal residence are the order of the day, with female autonomy correspondingly curtailed. However, as Beckerman and Valentine explain, the reverse outcome must be recognized if we are to grasp the parameters:

‘Where women clearly have the upper hand, uxorilocal [matrilocal] residence predominates; women’s husbands are often chosen for them by their mothers, or they choose their own husbands; when a woman’s husband dies, his children tend to be brought up by their mother, her brothers, and her new husband; women have broad sexual freedom both before and after marriage; the idea of partible paternity is prominent, with women having wide latitude in choosing the secondary fathers of their children; women usually make no secret of the identity of these secondary fathers; and the ideology of partible paternity defuses to some extent potential conflicts between male rivals – antagonisms that are seldom helpful to a woman’s reproductive interests in the long run’.[43]


Beckerman and Valentine are Darwinian anthropologists who can hardly be accused of having Marxist sympathies. But I cannot help thinking that if Engels were alive, he might have been encouraged by their results. Here, for the record, is Engels on the subject of male sexual jealousy in evolutionary perspective:

‘….animal societies have, to be sure, a certain value in drawing conclusions regarding human societies – but only in a negative sense. As far as we have ascertained, the higher vertebrates know only two forms of the family: polygamy or the single pair. In both cases only one adult male, only one husband is permissible. The jealousy of the male, representing both tie and limits of the family, brings the animal family into conflict with the horde. The horde, the higher social form, is rendered impossible here, loosened there, or dissolved altogether during the mating season; at best, its continued development is hindered by the jealousy of the male. This alone suffices to prove that the animal family and primitive human society are incompatible things; that primitive man, working his way up out of the animal stage, either knew no family whatsoever, or at the most knew a family that is nonexistent among animals’.[44]

Engels accepts that a male gorilla might strive to hold on to any females it has succeeded in acquiring. But applied to the human case, mothers would then be denied access to any but the most isolated and intolerant males. The point stressed by Engels is that only a decisive social breakthrough could have solved this problem:

‘For evolution out of the animal stage, for the accomplishment of the greatest advance known to nature, an additional element was needed: the replacement of the individual’s inadequate power of defence by the united strength and joint effort of the horde.… Mutual toleration among the adult males, freedom from jealousy, was…. the first condition for the building of those large and enduring groups in the midst of which alone the transition from animal to man could be achieved. And indeed, what do we find as the oldest, most primitive form of the family, of which undeniable evidence can be found in history, and which even today can be studied here and there? Group marriage, the form in which whole groups of men and whole groups of women belong to one another, and which leaves but little scope for jealousy’.

As I have argued elsewhere[45][46][47], not all of Engels’ revolutionary speculations look out of place today.


Embarrassingly for proponents of the patrilocal band model, genetic data on sub-Saharan African hunter-gatherers indicates a long-term historical preference for matrilocal residence. Studies of mitochondrial versus Y-chromosomal dispersal patterns show that over thousands of years, hunter-gatherer women across this vast region have tended to reside close to their mothers following marriage, migration rates for women being lower than those for men.[47]

A census among the Hadza – bow-and-arrow hunters of Tanzania – showed 68 per cent of married women whose mothers were alive residing with them in the same camp.[48] ‘Across all societies’, concludes the major specialist on this topic, ‘the greater the dependence on gathering, hunting, and fishing, the less likely that residence is virilocal [patrilocal]’.[49] Hunting has the strongest effect and, contrary to proponents of the standard model, results in less patrilocality, not more.[50]

A study conducted in 2004 reviewed the evidence behind the standard doctrine that patrilocality is the hunter-gatherer norm. Most of the widely used classifications turn out to have been based on totally inadequate data and ignore insightful discussions that took place in early anthropology. The few ethnographies in which camp data are available show that individuals use a variety of kin and other links to decide where to live, the only discernible statistical bias being in favour of mother-daughter interdependence and proximity.[51]

An important finding is that choices vary at different stages of life. A husband who has already helped provision a child might then be trusted sufficiently for his wife to agree to move with him to his natal camp. But this shouldn’t obscure the fact that residence among hunter-gatherers is matrilocal at first. Whether in Australia, Africa or the Americas, a young hunter-gatherer bridegroom is obliged to visit his bride while she remains in her own camp, working strenuously for her as he surrenders to her kinsfolk whatever game he catches. This, after all, is the essence of ‘bride-service’ – the fundamental economic institution in any hunter-gatherer society.

To maximise incoming provisions, the young hunter’s in-laws will strive to keep him under close supervision and control. It therefore comes as little surprise to find that cross culturally, males contribute less when they can take their bride back to their own camp, more when they are obliged to visit their bride who remains living among her own kin.[52] Females, then, obtain the best deal when they remain following marriage with close kin.

On what grounds can it be claimed that this residence pattern characterized early human kinship? According to the ‘grandmother’ hypothesis,[53][54][55] the selective advantage of distinctively human postmenopausal lifespans is that it enabled older women to assist their adult children in caring for and provisioning grandchildren. In genetic terms, a woman can never be as certain of her son’s offspring as she can of her daughter’s. For grandmothers to invest preferentially in their descendants through sons, therefore, would not be an evolutionarily stable strategy. It comes as little surprise, therefore, that a recent analysis of 213 Hadza camp compositions found that a woman over 45 with grown children is more likely to be in camp with her daughter than with her son, more likely to be with her daughter if that daughter has children under age 7 years and more likely to be with her daughter if that daughter is suckling a baby.[56]
Of course, advocates of the patrilocal band model might counter that grandmothers could somehow find ways to encourage or enforce fidelity in their sons’ wives. But unless they can explain how certainty of paternity could have equalled or exceeded certainty of maternity during the evolution of postmenopausal lifespans, we must conclude that the grandmothering hypothesis tips the scales decisively in favour of matrilocal residence and matrilineal descent.[57]

Turning to the emergence of modern Homo sapiens, it is now widely accepted that our species evolved recently in Africa. From about half a million years ago, brain size began increasing exponentially.[58][59] An infant with an outsized brain imposes heavy burdens on pregnant and nursing mothers.[60] If Homo sapiens mothers proved able to afford to raise such extraordinarily slow maturing, ultra-dependent offspring, this fact alone testifies to the success of their alliance-building and reproductive strategies.
The question arises: what new source of energy were they exploiting? The spare provisioning capacities of the evolving human male might in principle have been available for exploitation by females, but it is important to recognize the difficulties. It is unknown for a non-human primate male to systematically provision a pregnant or nursing female. In the case of chimpanzees, adult males are interested mainly in females who are displaying an oestrus swelling – i.e. those neither pregnant nor nursing. Where a female is nursing an infant, there is some danger that local males who feel suspicion regarding the paternity of that infant may attempt to kill and eat it.[61][62] The effects of primate male infanticide on female fitness and on population size and viability are for obvious reasons not positive.[63][64]

Where male reproductive differentials and corresponding levels of intra-male conflict are high, nursing mothers must divert scarce energy and resources away from direct offspring care into fighting off harassment and guarding against infanticide. A primate or hominin population whose females had to cope with such behaviour might head toward extinction, even as a minority of its males achieved short-term reproductive success.

But the converse equally applies. According to current models, the ancestors of extant humans comprised a small population dwelling somewhere in sub-Saharan Africa. Although disagreements abound, genetic studies indicate that their population size at one point may have resembled that of a modern endangered species, possibly no more numerous than today’s mountain gorillas. What happened next was extraordinary. The population exploded, Homo sapiens soon colonising the globe.[65]

Population expansion on such a scale is inconsistent with mothers having to put up with infanticide, harassment or the unaffordable costs of male philandering and double standards. If unusually large numbers of unusually large-brained offspring were being successfully raised to maturity, the quality of childcare must have been exceptional.

We know what the optimal solution would have been. There can be no doubt that mothers would have done best by cooperatively resisting male sexual control, relying for protection on supportive male and female kin, motivating multiple suitors to work hard for them – and taking advantage of every available childcare resource. For those conditions to be met, early human kinship needed to be matrilineal.

1. Allen, N. J., H. Callan, R. Dunbar and W. James (eds.), 2008. Early Human Kinship. Oxford: Blackwell.
2. Engels, Frederick, 1972 [1884] The Origin of the Family, Private Property and the State. New York: Pathfinder Press.
3. Harris, M. 1969 The Rise of Anthropological Theory London: Routledge.
4. Knight, C. 2008. Early human kinship was matrilineal. In N. J. Allen, H. Callan, R. Dunbar and W. James (eds.), Early Human Kinship. Oxford: Blackwell, pp. 61-82.
5. Malinowski, B. 1956 [1931] Marriage: Past and Present. A debate between Robert Briffault and Bronislaw Malinowski. M. F. Ashley Montagu, ed.. Boston: Porter Sargent.
6. Engels, F. 1972 [1891] Preface to the fourth edition. The Origin of the Family, Private Property and the State. New York: Pathfinder Press, pp. 27-38; p. 36.
7. Marx, K. 2000 [1844] Economic and philosophical manuscripts. In D. McLellan ed, Karl Marx: Selected Writings. Second edition, Oxford: Oxford University Press, pp. 83-121. p. 96.
8. Marx, K. and F. Engels, 2000 [1846] The German Ideology. In D. McLellan ed, Karl Marx: Selected Writings. Second edition, Oxford: Oxford University Press, pp 175-208. The quottion is on page 185.
9. Morgan, L. H. 1881 Houses and House-Life of the American Aborigines. Chicago and London: University of Chicago Press. p. 128.
10. Engels, F. 1972 [1884] The Origin of the Family, Private Property and the State. New York: Pathfinder Press. p. 68.
11. The same, page 75.
12. The same, page 47.
13. The same, page 51.
14. Lowie, R. 1937 The History of Ethnological Theory. New York: Holt, Rinehart & Winston, pp. 54-57.
15. Harris, M. 1969 The Rise of Anthropological Theory London: Routledge, p. 249.
16. Malinowski, B. 1956 [1931] Marriage: Past and Present. A debate between Robert Briffault and Bronislaw Malinowski. M. F. Ashley Montagu, ed.. Boston: Porter Sargent, p. 76.
17. The same, p. 28.
18. The same, p. 42.
19. Malinowski, B. 1932 The Sexual Life of Savages in North-Western Melanesia. 3rd. edn. London: Routledge, pp. xxiii-iv.
20. Kuper, A. 1988 The Invention of Primitive Society: Transformations of an illusion. London: Routledge.
21. Murdock, G. P. 1949 Social Structure. New York: Macmillan, p. 190.
22. Boas, F. 1891 The Indians of British Columbia. Sixth Report of the Committee on the North-Western tribes of Canada. Report of the British Association for the Advancement of Science 1890, p. 609.
23. Boas, F. 1897 The Social Organization and the Secret Societies of the Kwakiutl Indians. Washington: Report of the U.S. National Museum, 1895, pp. 334-5.
24. The same.
25. Lowie, R. 1960 [1914]. Social Organization. In Lowie’s Selected Papers in Anthropology. Cora du Bois, ed., pp. 17-47. Berkeley & Los Angeles: University of California Press.
26. Harris, M. 1969 The Rise of Anthropological Theory London: Routledge, p. 305.
27. Radcliffe-Brown, A. R. 1924 The mother’s brother in South Africa. South African Journal of Science 21: 542-55.
28. Junod, H. 1912 The Life of a South African Tribe, Vol. 1 Neuchatel.
29. Murdock, G. P. 1949 Social Structure. New York: Macmillan, p. 121.
30. Murdock, G. P. 1959 Africa: Its Peoples and their Culture History. McGraw-Hill Co., New York, p. 378.
31. Aberle, D. 1961. Matrilineal descent in cross-cultural perspective. In Matrilineal Kinship, D. M. Schneider, & K. Gough, eds. pp. 655-727. Berkeley & Los Angeles: University of California Press, p. 680.
32. Mace, R. C. Holden, 1999 Evolutionary ecology and cross-cultural comparison: The case of matrilineal descent in sub-Saharan Africa. In Comparative Primate Socioecology, ed.. Phyllis C. Lee. Cambridge: Cambridge University Press, pp. 387-405.
33. Holden, C., R. Sear and R. Mace 2003. Matriliny as daughter-biased investment. Evolution and Human Behavior 24, 99-112, p. 110.
34. Deacon, Terrence, 1997. The Symbolic Species: The co-evolution of language and the human brain. London: Penguin, p. 388.
35. Boyd, Robert and Joan Silk, 1997 How Humans Evolved. New York: Norton, p. 435.
36. Pinker, Steven, 1997 How the Mind Works. London: Penguin, pp. 488-90.
37. Henry, Jules, 1941 Jungle People: A Kaingang tribe of the Highlands of Brazil. New York: J. J. Augustin., p. 45.
38. Beckerman, Stephen and Paul Valentine 2002 Introduction. The concept of partible paternity among Native South Americans. In Cultures of Multiple Fathers: the theory and practice of partible paternity in Lowland South America. Beckerman, Stephen and Paul Valentine, eds, pp. 1-13. Gainesville, FL:. University Press of Florida, p. 6.
39. Hill, Kim, and A. M. Hurtado, 1996 Aché Life History: The ecology and demography of a foraging people. New York: Aldine de Gruyter, p. 444.
40. Beckerman, Stephen with Roberto Lizarralde, Carol Ballew, Sissel Schroeder, Christian Fingelton, Angela Garrison and Helen Smith, 1998. The Barí Partible Paternity Project: Preliminary Results. Current Anthropology, 391: 164-167.
41. Beckerman, Stephen and Paul Valentine 2002 Introduction. The concept of partible paternity among Native South Americans. In Cultures of Multiple Fathers: the theory and practice of partible paternity in Lowland South America. Beckerman, Stephen and Paul Valentine, eds, pp. 1-13. Gainesville, FL:. University Press of Florida, p. 11.
42. Gowaty, Patricia, 1997 Sexual dialectics, sexual selection, and variation in mating behavior. In Patricia Gowaty ed, Feminism and Evolutionary Biology: Boundaries, Intersections, and Frontiers. New York: Chapman and Hall, pp. 351-84.
43. Beckerman, Stephen and Paul Valentine 2002 Introduction. The concept of partible paternity among Native South Americans. In Cultures of Multiple Fathers: the theory and practice of partible paternity in Lowland South America. Beckerman, Stephen and Paul Valentine, eds, pp. 1-13. Gainesville, FL:. University Press of Florida, p. 11.
44. Engels, Frederick, 1972 [1884] The Origin of the Family, Private Property and the State. New York: Pathfinder Press, pp. 49-50.
45 Knight, C. 1991. Blood Relations. Menstruation and the origins of culture. New Haven and London: Yale University Press.
46. Knight, C. 2007 Revisiting matrilineal priority. In Jean Lassègue ed, Émergence et évolution de la parenté. Paris: Éditions Rue d’Ulm/Presses de l’École normale supérieure, pp. 25-43.
47. Knight, C. and C. Power, 2005. Grandmothers, politics, and getting back to science. In E. Voland, A. Chasiotis and W. Schienfenhövel (eds.), Grandmotherhood: The evolutionary significance of the second half of female life. New Brunswick: Rutgers, pp. 81-98.
47. Destro-Bisol, G. with F. Donati, V. Coia, I. Boschi, F. Verginelli, A. Caglia, S. Tofanelli, G. Spednini and C. Capelli 2004. Variation of female and male lineages in sub-Saharan populations: the importance of sociocultural factors. Molecular Biology and Evolution 21(9): 1673-1682.
48. Woodburn, J. 1972 Stability and flexibility in Hadza residential groupings. In Richard B. Lee and Irven DeVore, eds, Man the Hunter. Chicago: Aldine, pp. 103-110.
49. Marlowe, F. 2004 Marital residence among foragers. Current Anthropology 452 (45): 276-84.
50. The same, p. 80.
51. Alvarez, H. P. 2000 Residence groups among hunter-gatherers: a view of the claims and evidence for patrilocal bands. In Kinship and Behavior in Primates. B. Chapais and C. Berman, eds, pp. 420-442. New York: Oxford.
52. Marlowe, F. 2004 Marital residence among foragers. Current Anthropology 452 (45): 276-84., p. 281.
53. O’Connell, J., K. Hawkes and N. Blurton Jones, 1999 Grandmothering and the evolution
of Homo erectus. Journal of Human Evolution 36: 461-85.
54. Voland, E. A. Chasiotis and W. Schienfenhövel, eds, Grandmotherhood: The evolutionary
significance of the second half of female life. New Brunswick: Rutgers, pp. 81-98.
55. Opie, K. and C. Power, 2009. Grandmothering and Female Coalitions. A basis for matrilineal priority? In N. J. Allen, H. Callan, R. Dunbar and W. James (eds.), Early Human Kinship. Oxford: Blackwell, pp. 168-186.
56. Blurton Jones, N. G., K. Hawkes and J. F. O’Connell 2005. Older Hadza Men and Women as Helpers: Residence Data. In Hunter-Gatherer Childhoods, edited by B. S. Hewlett and M. E. Lamb, pp. 214-236. New Brunswick, NJ and London: Aldine Transaction.
57. Knight, C. and C. Power, 2005. Grandmothers, politics, and getting back to science. In E. Voland, A. Chasiotis and W. Schienfenhövel (eds.), Grandmotherhood: The evolutionary significance of the second half of female life. New Brunswick: Rutgers, pp. 81-98.
58. Ruff C B, E. Trinkaus and T W Holliday, 1997 Body mass and encephalization in Pleistocene Homo. Nature 387: 173-76.
59. de Miguel, C. and M. Henneberg, 2001. Variation in hominid brain size: How much is due to method? Homo 52: 3-58.
60. Foley, R. P. Lee, 1991 Ecology and energetics of encephalization in hominid evolution. Philosophical Transactions of the Royal Society, London B 334:223-232.
61. Hamai, M. with T. Nishida, H. Takasaki and L. A. Turner, 1992 New records of
within-group infanticide and cannibalism in wild chimpanzees. Primates 33: 151-62.
62. Hirawai-Hsegawa, M. and T. Hasegawa, 1994 Infanticide in non-human primates: sexual selection and
local resource competition. In S. Parmigiani and F. S. vom Saal eds, Infanticide and Parental Care.
Langhorne, Pennsylvania: Harwood Academic.
63. Butynski, T. 1982 Harem-replacement and infanticide in the blue monkey Cercopithecus mitis
stuhlmanni in the Kibale Forest, Uganda. American Journal of Primatology 3: 1-22.
64. Janson, Charles H. and Carel P. van Schaik 2000 The behavioural ecology of infanticide by males. In
Carel P. van Schaik and Charles H. Janson eds, Infanticide by Males and its Implications. Cambridge:
Cambridge University Press, pp. 469-494.
65. Jorde, L. B., M. Bamshad and A. R. Rogers, 1998. Using mitochondrial and nuclear DNA markers to
reconstruct human evolution. BioEssays 20:126–136.
66. Reich, D. E. and D. B. Goldstein, 1998 Genetic evidence for a paleolithic human population
expansion in Africa. Proceedings of the National Academy of Sciences of the United States of
America 9514: 8119-8123.
67. Forster, P. 2004 Ice Ages and the mitochondrial DNA chronology of human dispersals: a review.
Philosophical Transactions of the Royal Society of London, Ser. B 359, 255-264.