Family forms

Emmanuel Todd, The Explanation of Ideology: Family Structures and Social Systems (Blackwell, 1985, French original from 1983) was written a long time before the author became a prominent contrarian political pundit. But the book was also academically controversial when it appeared, making very broad claims about how historic family types affected modern political ideologies and electoral outcomes in countries across the world.

Todd’s structural anthropology breaks down family types on three axes: equality/inequality, authority/liberty and exogamy/endogamy. It’s a very male-centred approach, so this is almost all focused on the experiences of fathers and sons, making some of Todd’s definitions idiosyncratic. Equality and inequality concerns whether all sons inherit equally from their father or whether the father has testamentary freedom to favour one son over others. Authority/liberty is about whether one or more of the sons continues to live in the family home after marriage or whether they all establish separate households. Todd also breaks exogamy/endogamy down in three ways: preferred marriage to parallel cousins (father’s brother’s daughter), to cross-cousins (father’s sister’s daughter) or a prohibition on cousin marriage.

Todd starts with four kinds of families found in premodern Europe, all exogamous:

• Community family where all married sons remain in the father’s household and the brothers share inheritance equally (authority and equality)
• Authoritarian family (in other literature called stem family) where only one married son remains in father’s household while the others disperse and there is unequal inheritance by brothers (authority and inequality)
• Egalitarian nuclear family where sons form new household on marriage and brothers share the inheritance equally (liberty and equality)
• Absolute nuclear family where sons form new household on marriage and brothers do not share the inheritance equally (liberty and inequality)

Todd then adds in an additional dimension based on exogamy/endogamy and identifies seven ideal anthropological types of family structure. He divides the world into regions based on the predominance of these and then proceeds to make claims about which political ideologies fit best with these. It’s often unclear what evidence he’s using or what period he’s talking about, so it can be hard to be sure whether he’s basing his classification on anthropological studies, premodern historical data or nineteenth and twentieth century censuses. Some of his claims about how family structures affect political views are more plausible than others, especially since he sees several possible ideologies as associated with each family type:

• Exogamous community family: communism
• Authoritarian family: political Catholicism, social democracy, Nazism
• Nuclear family (egalitarian or absolute): liberalism, non-totalitarian dictatorship
• Endogamous community family: Arab socialism, fundamentalism
• Asymmetrical community family: caste system, communism
• Anomic family: individualism, communitarianism
• African systems: military regimes.

I do think there is something interesting lurking about how family dynamics affects people’s wider expectations of society, but it’s hard to pin it down as specifically as Todd tries to do. In this post, however, I’m less interested in the political aspects of Todd’s argument and more in one of his axes: (male) authority/liberty as reflected in neolocality or instead residence with one’s father after marriage.

The anthropological literature tends to conflate neolocality with the nuclear family in somewhat unhelpful ways. Firstly, you can have the nuclear family in non-settled societies, such as hunter-gatherer societies, without having neolocality in the sense of shared or separate dwellings. Secondly, in societies with low life expectancy, you can end up with predominantly nuclear families in practice even if the ideal is Todd’s community family (all married sons remain in the father’s household) or authoritarian family (only one married son remains in the father’s household).

For example, here are figures from households in Nippur under the Kassite dynasty (taken from a Middle Babylonian collection of administrative texts, legal documents, and letters from more than five hundred texts and fragments dating between 1359 and 1224 BC), compared to Medieval Tuscany and Roman Egypt.

Table 7.2 Percentages of attested households by type for the servile-worker population in Kassite Nippur compared with general populations of medieval Tuscany and Roman Egypt

(Widell 2013), Table 7.2

It’s plausible that the multiple-family household, containing two or more conjugal (nuclear) families is the preferred form in all these societies, because otherwise you’d expect it to have near zero occurrence, but it’s still relatively uncommon for purely practical reasons. Relatively few men, in particular, live long enough to get to be grandfathers. That also makes it hard in terms of evidence to distinguish virilocality (wife moves to live with husband) from patrilocality (wife moves to live with husband, who is himself living with his father).

You can’t therefore always easily tell a nuclear family from a community/authoritarian family in certain stages of their cycle and that’s a particular issue with more limited evidence. For example, the “oldest nuclear family” was identified based on a shared grave (grave 99) containing a man, a woman and two children biologically related to both of them (Haak et al 2008). These graves probably reflect the aftermath of a violent raid on the settlement, with the burials carried out by  survivors of the raid. The four family members could have been a permanent nuclear family, the start of a planned community family, or just the part of an existing community family that died. But given the age at death of the man (40-60), woman (35-50) and the children (one 4-5, one 8-9), it’s also at least possible that you instead have here a different part of a community family being buried together: a paternal grandfather, his daughter in law and her two sons.

It’s also often difficult to distinguish patrilocality from neolocality if sons are setting up home near their father. (In fact, I’ve seen some definitions of patrilocality which include sons in separate dwellings but who are still working the land/farm together with their father). In archaeological contexts, patrilocality tends to be identified by strontium isotope analysis showing incoming women who’ve grown up elsewhere, for example, which will only show a general region. Even using ancient DNA from graveyards to produce family trees from skeletons isn’t going to distinguish between sons living in the same settlement as their father and sons living in the same household as their father.

Neolocality and women

Before wide-scale censuses, it’s therefore hard to be sure exactly who is living in the same household. But I think that neolocality (in the sense of separate dwellings, even if in the same settlement) may have been significant for wives’ experience in two ways. Firstly, it seems likely to lessen one form of domestic abuse: one where a new wife is essentially regarded as a servant or even a slave of her mother-in-law, having to do her bidding. When such a hierarchy of adult women within a household is a life-cycle norm, it perpetuates and reinforces poor domestic treatment. After all, if you have been subservient as a young bride, becoming a mother-in-law is when you get your chance of domination in return. Why would you want to give that up? In contrast, if you’re not sharing a household with your in-laws, you’re probably not preparing food jointly and eating together all the time, which is the most obvious setting for such hierarchical abuse.

Neolocality also potentially causes greater divergence between male family members’ economic interests. In a community family, there should be in theory no difference between fathers’ and sons’ interests in the household economy, since they’re all part of it. In an authoritarian family, there is a difference between the economic interests of the sons who won’t inherit the main property and the one favoured son; Todd points out how this weakens fraternal solidarity. But again, the financial interests of the father and the son who remains are similar.

In contrast, with neolocality (at least once there is no longer a joint farm/homestead), adult sons do not necessarily share the economic interests of their father. They may be concerned to use resources to set up their own households sooner even at the expense of the long-term benefit of the father’s property. And once they’ve left the father’s household, they have less interest in how that is doing economically.

In such a situation, especially if divorce is difficult, a wife becomes the only person who theoretically does share the same interest in her husband’s household economy throughout their married life. Their household wealth is tied together, which makes a wife her husband’s most reliable agent/deputy, if the legal system will allow that. She potentially has less incentive to cheat her husband’s business than either an unrelated agent/manager or another male relative of his has. A number of western European legal systems from the Romans onwards have therefore accepted this idea of a woman acting on her husband’s behalf in certain circumstances.

The origins of neolocality

Todd was interested in modern political patterns in The Explanation of Ideology, so in that book he didn’t discuss the historical development of these different family forms. Some of his later work may do so, but I haven’t looked at that. Here’s his 1983 map of European family types:

Todd himself claims (pp. 196-197) that looking worldwide the modern locations of family patterns and the ideologies they support are more or less arbitrary and they don’t match specific ecological zones or economic models. He does admit, however, that the community family model areas look mostly contiguous, with a probable origin in western Asia. Recent anthropological attempts to look at the evolution of marriage locality suggest that there’s considerable swapping between different patterns and that these changes aren’t consistent in direction, but vary between language groups (see e.g. Moravec et al 2018). Even that article probably underestimates the variety and change, not taking in account, for example, emerging evidence for Celtic matrilocality (Cassidy et al, 2025).

All that suggests that modern patterns (nineteenth century and later) reflect initial ecological constraints that are then substantially modified by cultural transmission from elsewhere (either voluntary or via conquest), probably with a lot of path-dependence thrown in for good measure. Judith Bennett (Bennett 2019), for example, discussing some of the earliest detailed evidence of medieval neolocality, from thirteenth-century England comments (p. 336):

I doubt that we will ever have more than a remote, elite-based view – from ecclesiastical, royal, and seigneurial sources – of the emergence of neolocal marriage in the English countryside. It is clear that manorial policies could favour nuclear households (for example, by insisting on primogeniture and taxing landless adults), and it is possible that the extensive cooperation required by open fields and common rights generated a similar pressure from neighbours. Other causes and precise chronology await further research, but we can be confident that, if this sort of marriage was already common in rural England at any time between 1086 and 1250, it was then much easier for people to achieve, simply because these were decades of rural expansion and land reclamation.

In other words, once patterns have become established they may persist despite ecological and economic changes. (Moravec at al 2018) reckons that in Indo-European groups, transitions between marital residence states happened on average around every 420 years, while Todd’s endogamous community family survives largely in (some) regions conquered by Muslims from the seventh century onwards.

The problem of identifying origins then becomes that for most of western Europe we have very little demographic evidence from households before the late Middle Ages. What I’m going to do here, therefore, is pull together some fragments to try and produce a hypothesis about where north-western European neolocality came from.

It’s worth starting with the basic ecological/economic constraints. Community family structures are the most obvious basic model for settled patrilineal agriculturalists. In general, it’s more resource efficient for independent peasants to have a community family structure (in terms of dwellings, food preparation and giving an additional labour force for the holding). It’s also easier for such a household to defend itself from external enemies than if living in scattered nuclear families. In areas with restricted additional agricultural land available (such as late medieval England), neolocal marriage is not the ideal strategy, because it makes it so hard to marry.

From that angle, the stem/authoritarian family looks like an obvious if unfair adaptation of the community family when there’s not enough land to divide up, or at least not without a severe shock to standard of living or social status. Fauve-Chemoux 2006 argues that stem families tend to be associated with  agro-pastoral mountainous surroundings.

Neolocality, in contrast, only makes economic sense when there’s a lot of additional land available. As Sabine Huebner (Huebner 2017) comments (p. 12):

From studies by family historians, sociologists, and social anthropologists on societies of early modern and modern periods, we know that composition and size of household depended on the economic basis on which the family made its living, because different forms of property and economic activities required different kinds of labor. In regions with low population density and abundance of land, neolocality was common and young couples established their own households upon marriage. In those areas where land was scarce there was a tendency to form complex households with several generations living under one roof. The fact that exigencies rarely changed with time or political rulers, afforded family forms some regional continuity.

For the ancient and medieval period, however, we also need to take cultural diffusion and/or conquest seriously as causes for change, because they take peoples and practices out of their original ecological zones. The most likely reason for an initial spread of community families (including into regions with a lot of spare land, like Russia) is the influence of Proto-Indo-European speakers. Haubrichs 2007 discusses kinship terms in PIE as well as proto-Germanic and PIE includes terms that make most sense within a community family, such as the term for grandfather also meaning head of the household, and for paternal uncle as a “little father”.

There also seems to be agreement that the western Roman elite were neolocal, although Roman Egypt, the only part of the Roman world for which we have reliable evidence at household level, had a community family model among the non-elite. We have very little evidence for non-elite family structure in ancient and medieval Western Europe, but the two Carolingian period studies of peasant households in these terms that I’m aware of (Devroey 2003, pp. 64-64, Ring 1979) show relatively high levels of joint families in a couple of estates in the Paris basin and central Italy, suggesting that the community family may have been the peasant ideal then.

On the other hand, as Bennett shows, peasant family structures  could easily be constrained by ecology or by lords’ preferences. Elite family structure can probably tell us about the desired family model, even if it’s not one that’s necessarily obtainable by many families. In other words, if elites want community families, you should expect to see them exist.

If you look at Todd’s map of Europe in very broad terms of cultural flows, then one possible schema would be an initial spread of PIE-speaking community family cultures from the east, followed many centuries later by western Roman neolocal influence that leaves a mark in e.g. Romania, but doesn’t dislodge authoritarian/stem families in more mountainous regions. That might possibly explain elite neolocality in northern France, which Le Jan 1995 pp. 334-339 shows can be seen in Merovingian and Carolingian sources. But it doesn’t strike me as an adequate explanation for elite neolocality in England, which was much less Romanized and then had a post-Roman influx of culturally dominant Angles, Saxons etc, from areas of authoritarian families.

Instead, it’s worth looking at sources of land in early medieval England and Francia. Charles-Edwards 1997 p. 195 points out the existence of a norm by the early eighth century that noblemen received royal grants of land for life at the age of about twenty-five to enable them to settle down and marry. That pattern of court service by young noblemen followed by royal landed reward is also known from seventh-century Merovingian kingdoms, and provides an obvious context for a cultural shift to elite neolocal marriage. If royal favour/service gets you land and that favour is possibly available to several brothers, then you no longer need to stay in your father’s household until his death.

There’s also a small amount of archaeological evidence that might suggest a gradual move away from elite joint families, as shown by Loveluck 2013 pp. 253-254:

By the tenth and eleventh centuries, the principal venue for public feasting was a large hall at both palaces and important estate centres, in both West Francia and Anglo-Saxon England…The majority of elite secular centres possessed multiple rectangular buildings for residence and dining, between the seventh and ninth centuries, in both West Francia and England. Hence, the emergence more generally of a focal core of principal large hall range and ancillary residential buildings was a significant change in the internal social organisation of settlements.

This hypothesis also suggests why the change to neolocality happens in specific kinds of early medieval polities only: “weak” states based on the redistribution of land as a royal favour, rather than strong states offering mainly paid office as a reward, where it’s likely that only one son from a family will get rich and the communal family then forms around him.

That also makes me wonder if Roman neolocality developed for similar reasons. There’s no obvious cultural influence from elsewhere that would explain why they developed that model, but if the early republic was handing out land in the Italian peninsula to citizen-soldiers to allow them to marry, that might develop the pattern. The varied power of post-Roman rulers might also possibly explain the survival of the community family in northern Italy, since Lombard rulers had much less land to distribute than Frankish or English ones did. But, of course, we can’t be sure when community families developed in Tuscany, since we have so little early medieval data.

Given this lack of data on household composition, I don’t think we’re ever going to know exactly when and why neolocality developed and spread in some parts of western Europe. But I think, as I’ve discussed above, we can say something about its effect on wives, and that suggests again that customs which develop to advantage some men can have positive effects for some women.

Bibliography

Bennett, Judith M. 2019. “Wretched girls, wretched boys and the European Marriage Pattern in England (c. 1250–1350).” Continuity and Change 34 (3): 315-347. https://doi.org/10.1017/S0268416019000328.

Cassidy, Lara M., Miles Russell, Martin Smith, Gabrielle Delbarre, Paul Cheetham, Harry Manley, Valeria Mattiangeli, Emily M. Breslin, Iseult Jackson, Maeve McCann, Harry Little, Ciarán G. O’Connor, Beth Heaslip, Daniel Lawson, Phillip Endicott, and Daniel G. Bradley. 2025. “Continental influx and pervasive matrilocality in Iron Age Britain.” Nature 637 (8048): 1136-1142. https://doi.org/10.1038/s41586-024-08409-6.

Charles-Edwards, Thomas 1997. “Anglo-Saxon kinship revisited.” In The Anglo-Saxons from the migration period to the eighth century: an ethnographic perspective, edited by John Hines, In Studies in Historical Archaeoethnology, 2, 171-204. Woodbridge, Suffolk: Boydell.

Devroey, Jean-Pierre. 2003. Économie rurale et société dans l’Europe franque (VI-IX siècles). Paris: Belin.

Fauve-Chamoux, Antoinette. 2006. “Family reproduction and stem-family system: From Pyrenean valleys to Norwegian farms.” History of the Family 11: 171-184. https://doi.org/https://doi.org/10.1016/j.hisfam.2006.11.007.

Haak, Wolfgang, Guido Brandt, Hylke N. de Jong, Christian Meyer, Robert Ganslmeier, Volker Heyd, Chris Hawkesworth, Alistair W. G. Pike, Harald Meller, and Kurt W. Alt. 2008. “Ancient DNA, strontium isotopes, and osteological analyses shed light on social and kinship organization of the Later Stone Age.” Proceedings of the National Academy of Sciences 105: 18226-18231. https://doi.org/doi:10.1073/pnas.0807592105.

Haubrichs, Wolfgang. 2007. “Germanic and Gothic kinship terminology.” In The Ostrogoths from the migration period to the sixth century: an ethnographic perspective, edited by Sam J. Barnish and Federico Marazzi, 143-182. Woodbridge, Suffolk: Boydell Press.

Huebner, Sabine R. 2017. “A Mediterranean family? A comparative approach to the ancient world.” In Mediterranean families in antiquity: households, extended families, and domestic space, edited by Sabine R. Huebner and Geoffrey S. Nathan, 1-26. Chichester: Wiley-Blackwell.

Le Jan, Régine. 1995. Famille et pouvoir dans le monde franc (VIIe – Xe siècle): essai d’anthropologie sociale.Histoire ancienne et médiévale, 33. Paris: Publications de la Sorbonne.

Moravec, Jiří C., Quentin Atkinson, Claire Bowern, Simon J. Greenhill, Fiona M. Jordan, Robert M. Ross, Russell Gray, Stephen Marsland, and Murray P. Cox. 2018. “Post-marital residence patterns show lineage-specific evolution.” Evolution and Human Behavior 39: 594-601. https://doi.org/https://doi.org/10.1016/j.evolhumbehav.2018.06.002.

Ring, Richard R. 1979. “Early medieval peasant households in central Italy.” Journal of Family History 4: 2-25.

Todd, Emmanuel. 1985. The explanation of ideology: family structures and social systems. Translated by David Garrioch. Oxford: Basil Blackwell.

Widell, Magnus, McGuire Gibson, T. J. Wilkinson, Benjamin Studevent-Hickman, and Jonathan Tenney. 2013. “Household and village in early Mesopotamia.” In Models of Mesopotamian landscapes: how small-scale processes contributed to the growth of early civilizations, edited by T. J. Wilkinson, McGuire Gibson and Magnus Widell, In BAR International Series 2552, 112-129. Oxford: Archaeopress.