Relationships Throughout History

The Alliance Era

For most of recorded history, a wedding joined two families rather than two people. The couple stood at the center of a contract they had not written and could not easily break, closer to the terms of the deal than to the parties making it. Their assignment was to be joined, exchanged, and made to produce: an heir, an alliance, a claim on land. Whether they wanted each other was a secondary detail, and often not a detail anyone bothered to record.

A caveat shapes everything that follows: this is largely a history of paperwork. We know about Mesopotamia, Rome, England, and colonial America because those societies wrote things down and the writing survived, which tilts the record toward the literate, the propertied, and, after antiquity, the Western. The couples who left documents were disproportionately the ones with property worth contracting over. The further you get from the archive, the blurrier the picture, so the societies that kept the best receipts are best read as a sample rather than a census.

The contract was rarely even one man to one woman. We tend to picture the past as a tidy row of one-husband-one-wife households with the modern couple at the end of the line. Widen the lens past the literate West and the row breaks up. At scale, the anthropological record shows a human default far messier than the row suggests.

The 85% figure sounds like a world of harems, and it was nothing of the kind. Even where the law let a man take several wives, most marriages stayed one-to-one, because most men could neither afford nor command more than one. Polygyny in practice was a privilege of the top: a few high-status men absorbed extra wives while a matching tail of low-status men got none. The pattern is older and more durable than any wedding rite. The marriage market has always been stratified, sorted by rank and wealth and power, and formal polygyny is only what that sorting looks like when the law lets it surface. Remove the law and the sorting continues in silence. It disappears from the written record for most of what follows, then returns with force at the end.

When marriage surfaces in the written record, it surfaces as commerce. The oldest receipts are marriage contracts pressed into clay in Mesopotamia, and they read like what they were: transfers of property with terms attached. The groom’s side paid a bride-price, the bride’s side brought a dowry, and the consent that made it binding belonged to the fathers rather than the couple. It was written down for the reason any high-value transaction is written down, so that ownership, obligation, and the legitimacy of any children could be proven later before an authority. The Code of Hammurabi, around 1750 BC, put the logic bluntly: a marriage without a contract did not count.

No tablet, no wife.

The penalties protect the deal, not anyone’s feelings. A betrothal was a contract with money down, and Hammurabi wrote the remedies for breaking it in the flat register of commercial law.

The dowry rule pointed the same way. It kept a divorced wife from destitution, and it also stopped her family’s capital from being quietly absorbed into her husband’s line. Marriage here was infrastructure for moving property and producing legitimate heirs, and the law was the enforcement layer that kept the infrastructure honest.

One provision reveals what the whole apparatus is for. A husband could divorce a wife who bore no children, but returning her dowry made divorce costly, so a concubine was often the cheaper option (infertility, needless to say, was always blamed on the woman). The rule manages a consequence of sex rather than sex itself: the legitimate child, and the property and paternity that ride on it. Almost no early society policed desire directly. What they policed was what desire produced.

Rome ran the same logic through a more elaborate legal machine. The stated purpose of marriage was children; the legal minimum ages were twelve for girls and fourteen for boys; a valid union needed the consent of the couple and their fathers. Most women in fact married between fifteen and twenty and men in their late twenties, with elite families reaching for earlier matches to lock in dynastic alliances. Roman law even split marriage into two forms: cum manu, which moved the wife and her property into her husband’s control, and sine manu, which left both with her birth family. Sine manu gradually won out, and the reason it won is telling. Families preferred the version that kept their capital from leaking into the husband’s line. When the deciding question in a marriage is which household keeps the assets, the arrangement is a merger.

Rome then pushed the logic somewhere that feels startlingly modern, making marriage a matter of state. In 18 BC the emperor Augustus’s Lex Iulia required men to marry by twenty-five and women by twenty, penalized the unmarried, let childless widows forfeit half of what they stood to inherit, and gave the parents of several children preference for public office. This is a state treating marriage as infrastructure, something the polity needs built and maintained at a certain rate, like roads or granaries, and treating the childless citizen as a policy problem to be fixed. The lever it reached for was inheritance, the same one every earlier arrangement had pulled. Augustus did not try to make Romans want to marry. He made not marrying, and not producing heirs, expensive, because what the state cared about was not who desired whom but who would inherit.

Hard demography sat underneath the moralizing. Childbirth killed women of prime marrying age often enough that, from about the second century BC, Rome ran a chronic shortage of marriageable, fertile women. In market terms a proven childbearer was the scarcest asset in the room, which pushed divorce and remarriage upward as men competed for the women likeliest to give them surviving heirs. The alliance market was not sentimental about its own churn, pricing women even for the risk that they might die producing the very heirs that made them valuable.

For all its machinery, the alliance system had one conspicuous blind spot. Nowhere in it did the wishes of the two people being married count for much. Closing that gap took the next thousand years, and it happened slowly, unevenly, and largely by accident.

The Church Rewrites the Rules

When the Roman order gave way to the medieval Church, the institution inherited the alliance system intact and then, across centuries, quietly edited it. The Church’s strangest innovation was a single requirement: consent. It spent generations insisting that what actually made a marriage was the freely given “I do” of the two people, not the father’s signature and not the transfer of property. In daily practice families went on arranging and pressuring as before, and a peasant girl’s “free” consent was rarely very free. Written into doctrine, though, the principle was radical and load-bearing: the individual now stood inside the institution rather than beside it as one of its terms. The Church also tightened the rules against marrying close kin, which historians argue helped dissolve the very clan networks alliance-marriage had been built to serve, though the mechanism is debated. Either way, a wedge had been driven between the family’s interest and the couple’s, and it never fully closed again.

It is tempting to read consent as the moment love entered marriage. It did nothing of the kind, and the gap between the two is the thing worth getting right. Consent did not invent love. The assumption that it did, that people in the past simply did not feel what we feel, collapses the moment anyone looks for the evidence.

Love was never the anomaly. Marrying because of it was.

Romantic love is ancient and nearly universal. What is historically strange is the notion that love should be the thing that triggers a marriage. For almost all of the past people fell in love constantly and married for other reasons: land, alliance, labor, security, a parent’s preference. Love and marriage were two systems that occasionally overlapped, not one system under two names. The proof sits inside the very checklist the anthropologists used to detect love: elopement. Nearly every culture that felt romantic love also kept a name, and usually a punishment, for running off with the person you loved instead of the person you were meant to marry. A culture needs that category only when love and marriage regularly point in different directions.

The medieval elite turned that gap into an art form. In the eleventh century the troubadours of southern France built a whole code around it, courtly love: an idealized, consuming passion aimed deliberately outside marriage, adulterous by design, since marriage was business. Its rulebook, De Amore, written around 1184 in the circle of Marie de Champagne and attributed to Andreas Capellanus, opens its rules of love with a line that gives the game away: “Marriage is no real excuse for not loving.” The most famous woman of the age put the same conviction without the courtly decoration. Héloïse, brilliant and in love with her teacher Peter Abelard, made her case against marrying him in a letter that has never stopped being quoted.

She was the articulate edge of a wide tradition, not a lone eccentric. Love flourished across the medieval world; it simply flourished mostly outside the marriage contract, in songs and affairs and the elaborate fictions of the courts.

Because love and marriage ran on separate tracks, the marriages could be coldly practical about things like age, which is the source of the period’s most lurid reputation and where most of that reputation falls apart. The stock image is the medieval child bride, married at twelve. It barely survives the records. The genuinely young brides were royal women wed in their early teens for reasons of state, precisely because their marriages were the highest-stakes alliances of all. Ordinary Italian and English town girls married in their mid-twenties, peasant girls in their early twenties. The dramatic older-man, younger-girl match was largely an Italian pattern rather than an English one, which leaves a quiet joke buried in the Western canon: the most famous teenage marriage in our literature is Romeo and Juliet’s, and Shakespeare set it in Verona. In an English parish it would have raised eyebrows. When English marriage records start surviving in bulk, they settle the matter. Of a thousand entered in the archdiocese of Canterbury between 1619 and 1660, exactly one bride was thirteen, and 966 of the thousand were nineteen or older at their first marriage.

That raises the obvious question. If children were not being married off young, why were Europeans marrying so late? The most influential answer, and one of the most contested, came from the demographer John Hajnal, who in 1965 drew a line across the map from Trieste to Leningrad. West of it, he argued, sat a marriage pattern unlike most of the world’s: women marrying in their mid-twenties, men a few years older, a tenth to a fifth of people never marrying at all, and newlyweds expected to start a household of their own rather than move in under a parent’s roof. East of the line, marriage came early and almost universally, young couples folded into large multi-generational families.

The engine behind late Western marriage was money rather than romance. Starting an independent household, which Western marriage now demanded, took years of saving, usually earned during a long stint in service on someone else’s farm or in someone else’s house before a couple could afford a place of their own. Marriage was an economic threshold crossed once you could pay for it, which is why so many crossed it late and some never crossed it at all. Love ran alongside the whole time; the songs and the elopements prove it. It simply did not set the clock. And the apparatus of public vows, witnesses, and recorded legitimacy did in the Church era exactly what the clay contracts had done in Mesopotamia: it made sure that when a child arrived, everyone knew which marriage it belonged to. Consent was the new ingredient.

For the couple this was progress and limitation at once: a genuine say in the marriage, but not yet a marriage built around their own happiness, and no culture yet honest about what courting couples were already doing in the meantime.

The Early Modern Turn

If the Middle Ages carry an undeserved reputation for child brides, the early modern centuries carry one for prudishness, the buttoned-up world of stern sermons and unbroken chastity before the wedding. That reputation is just as false. In the rural English parishes of the late sixteenth century a startling share of brides walked to the altar visibly pregnant. The demographer P.E.H. Hair went through the parish registers and put the figure at roughly a quarter of all brides. For couples already headed toward marriage this was close to routine rather than a scandal grimly tolerated. Premarital sex carried little real shame on its own. What carried shame, and what the community and its courts moved against, was its consequence arriving in the wrong order: a child with no marriage behind it, and no one clearly on the hook to support it.

The court records make the logic almost comic. Elizabethan church courts routinely hauled in couples who were already married and fined them for “antenuptial incontinence,” the offense of having conceived a child before the very wedding they had gone on to hold. Couples merely courting, and clearly bound for marriage, were often left alone. By modern lights the priorities are backwards: the system reached past the unmarried lovers to punish the married ones, retroactively, for the timing of a pregnancy. It makes sense only once you see the target, which was never the act of sex but the question of whether a binding contract stood behind the child. A couple who supplied the contract late still got docked for collecting early.

Every age polices the consequences of sex — not the sex.

The sharpest evidence against the “puritanical past” is the group the fantasy was named for: the actual Puritans of colonial New England. They practiced bundling. A courting couple, unmarried, would be allowed to spend the whole night in bed together under the girl’s parents’ roof, the parents treating it as a sensible way for two young people to get through a long, cold night and learn something about each other. New England was, in effect, hosting supervised sleepovers.

Underneath the sex, the shape of marriage was slowly shifting. The old late-marriage pattern held: women marrying in their mid-twenties, men a couple of years older, four or five children over a lifetime, a sizeable minority never marrying. But the unions forming now increasingly looked companionate, grounded in the couple’s own consent, close in age, light on parental control, and set up in a household of their own. Even the official liturgy registered the change slowly. When the Church of England fixed its marriage service in the 1662 Book of Common Prayer, it still ranked the reasons for marriage in the old order, with companionship placed last of the three.

The feeling was already running ahead of the doctrine, and further ahead of the law. English common law still ran on coverture, the doctrine that folded a wife’s legal identity into her husband’s: she could not readily own property, sign contracts, or keep her own wages. A marriage in this period could be a genuine love match and a legal erasure at once, the woman choosing her partner freely and, in the eyes of the law, half-disappearing the moment she did. The state was slow even to register the couple’s new power. Clandestine, consent-only marriages were not outlawed in England until Hardwicke’s Marriage Act of 1754, and in Scotland not until 1856, which is why eloping English couples bolted north to Gretna Green.

One feature of these marriages the modern eye misses entirely, because we measure them against a lifespan our ancestors never had. They ended sooner than the vow implied, routinely, and usually by death rather than divorce, which was nearly impossible for ordinary people. Death churned the marriage market in much the way divorce does now.

The same demographic pressure could produce the opposite institution. In China, by the historical scholarship, widow remarriage was progressively restricted from the Sui dynasty onward, and the Ming and Qing states went further, granting public honors, the commemorative chastity arches, to widows who refused to remarry at all. Same shortage of surviving spouses, opposite official answer: one culture recycling its widows briskly back into the market, the other building monuments to those who stayed out of it. Both are ways of managing the property, lineage, and obligation bound up in a marriage rather than anything to do with the widow’s own desire. Underneath every version sat the children, who in most pre-modern economies were labor and old-age insurance as much as kin. Child mortality was staggering by our standards, so families had to over-produce births just to see a few reach adulthood, which kept fertility rather than affection at the quiet center of the arrangement even as affection was finally being invited to the table.

Companionate marriage had arrived in feeling. What it still lacked was the thing that could make it fully voluntary: a way to survive, and even thrive, without marrying at all. That was about to be built in factories.

The Love-Match Revolution

Then the ground moved, and it moved because of work. Industrialization pulled production out of the household and into the factory and the office, dismantling the economic logic that had underwritten marriage for five thousand years. A household had always been a unit of production, a farm or workshop that a couple and their children ran together and neither could easily run alone, and marriage was, among other things, how you staffed it. Wage labor broke that. Once a person could earn a living alone, selling their hours to an employer, the household stopped being a business that required a spouse. Marriage did not vanish, but its footing changed: it became a choice rather than a necessity, which is a far more precarious thing for an institution to be. On top of the wage came the thing that had always been present and never in charge. People could marry for love now, because they no longer had to marry for anything else. A new expectation arrived with it. A marriage was suddenly supposed to make you happy, and happiness had been promoted from lucky accident to job requirement.

No one caught the new standard earlier, or put it more plainly, than Jane Austen. In November 1814 her niece Fanny Knight wrote to her aunt for advice about a suitor she was cooling on, and Austen’s reply became the most quoted thing she wrote outside her novels.

She was in her late thirties and unmarried when she wrote it, which lent the advice its weight. The standard she named, affection as the one non-negotiable, would have baffled almost everyone in the four thousand years that came before her.

A wage made marriage optional. And optional is fragile.

The gap between the old rules and the new reality shows up starkly in the period’s numbers.

The law still carried floors written for a world of dynastic alliances, where marrying off a fourteen-year-old could make cold sense. The behavior on the ground belonged to a different civilization, one where people waited into their mid-twenties because marriage had become a decision about a life rather than a family’s move on a board. Floor and practice had simply stopped speaking to each other.

The law lagged the couple a second way, and this one fell entirely on the wife. The companionate ideal had her choosing her husband freely, while coverture, carried intact from earlier centuries, folded her legal self into his the instant she chose. A love match that was also a legal erasure could not sit easily inside a marriage increasingly sold as a partnership, and in England the contradiction began to give.

The market these newly legible wives were sorting through was lopsided in its own right, and not only by class. Census accounts of mid-nineteenth-century England point to a real numerical mismatch between the sexes, on the order of 9.1 million women to 8.6 million men, with perhaps thirty per cent of women over twenty unmarried at any given moment. Contemporaries wrung their hands over “surplus” or “redundant” women as a moral crisis. Stripped of the moralizing it was something plainer: a supply imbalance in a marriage market, the kind of shortage that leaves real people unpaired no matter how a culture prefers to explain it.

What reshaped courtship in this stretch was, of all things, the automobile. The historian Beth Bailey traced the shift and gave it a name, the move from front porch to back seat. Courtship left the family parlor, where a young man “called” on a woman under her mother’s eye, and climbed into the car, where a couple “went out” on their own. The vocabulary changed with the venue. The word “date” in its modern sense first surfaces in print in 1896, in a Chicago Record column by George Ade, when a clerk named Artie confronts a drifting sweetheart: “I s’pose the other boy’s fillin’ all my dates?”

The new custom was so illegible at first that the law could not read it. The historian Moira Weigel, in Labor of Love, notes that in the early years of unchaperoned dating the authorities struggled to tell a date from solicitation, and working women who let men buy them dinners or gifts risked arrest as suspected prostitutes. Kinsey’s data caught the underlying behavior mid-motion: American women who came of age in the 1920s were markedly less likely than the pre-war cohort to be virgins at marriage, though the shift was smaller and more orderly than the myth of the wild Jazz Age suggests. A clear majority of women under thirty were still virgins at their weddings, and of those who were not, about half had slept only with the man they went on to marry. The car privatized chastity rather than detonating it. What loosened, as before, was the surveillance of what sex might produce, and it loosened at about the pace those consequences came to feel survivable.

Marriage had become a choice, made for love and judged by happiness, the most fragile basis it had ever rested on. One hard link still tied sex to its consequences. Severing it would let the whole bundle come apart.

The Great Unbundling

The conclusion is worth stating bluntly, because so much modern argument is built on getting it backwards. The version of marriage nostalgia treats as timeless (the young couple, the single male breadwinner, the house full of children, the wife at home) was not history’s baseline but a brief, freakish dip. American first-marriage ages fell to the lowest point in the country’s recorded history around the middle of the twentieth century, women marrying at just over twenty and men at about twenty-two, before the long climb back to today’s late twenties and thirties. The rest points the other way: the mid-twenties brides of medieval England, the late-marrying Hajnal West, the companionate couples saving for years before they could afford a household. Late, deliberate, economically gated marriage is the human pattern. The “traditional” 1950s marriage that a certain politics wants to restore was, measured against the whole span, the least traditional moment in it.

The 1950s were the least traditional moment on the timeline.

Then it came apart quickly. Everything marriage had bundled into one package, sex and cohabitation and childbearing and economic partnership and social status, unbundled into separate stages that can now be taken in any order, with different people, or skipped. The machinery is familiar: the contraceptive pill, approved in 1960; no-fault divorce, which California pioneered in 1969 before it spread nationwide; cohabitation sliding within a single generation from scandal to unremarkable first step. The law arrived at the same moment, and from a revealing direction. In 1965 the Supreme Court struck down a state ban on contraception in Griswold v. Connecticut, and Justice William O. Douglas reached for the highest language he had.

The ruling that wrapped marriage in the language of the sacred is the same ruling that pried sex loose from procreation in constitutional law. The instrument that would do the most to loosen marriage was handed down in a sentence calling marriage almost holy.

The last piece to change was how people found each other, and machines entered that business earlier than the internet did. In 1965 a pair of Harvard undergraduates launched Operation Match, running students’ compatibility questionnaires through an IBM 1401 on punch cards. The web version arrived three decades later: Match.com, founded by Gary Kremen and Peng Ong, went live in April 1995 as the first major online dating site. It had almost no company at first.

The pill did something the whole preceding history makes legible. Every era so far policed the consequences of sex far more than the sex: Hammurabi’s dowry and paternity rules, Rome legislating heirs, the Elizabethan court fining a pregnancy rather than a passion, the shotgun logic under a New England bundling bed. The entire apparatus existed to manage one thing, the child that sex might produce and the property, paternity, and obligation trailing behind it. Reliable contraception cut sex free of pregnancy, and cutting that single thread pulled the load-bearing wall out of the structure. Once sex need not produce a child, most of the machinery built around it, the contracts and dowries and legitimacy courts and chastity arches, had nothing left to hold up. We are standing in the room after the wall came out, arguing about what the space is for now.

Scarcity and competition are only part of it. The modern relationship is also asked to carry expectations no earlier arrangement ever placed on a single bond.