For most of the twentieth century, the demographic spine of the West was a simple shape: nearly everyone paired off, most paired off young, and most stayed paired. By the 1950s in the United States, the median age at first marriage for women was around 20 and for men around 22. More than 95 percent of adults eventually married. Divorce existed but was rare; cohabitation outside marriage was a fringe arrangement; nonmarital childbearing was roughly 5 percent of births. The system was not utopia—much of it ran on coerced compromise, restricted female options, and quiet suffering—but as a coordinating mechanism it worked: it routed almost every adult through a long, household-forming pair bond.
That spine has bent. The median age at first marriage in the United States has risen above 28 for women and 30 for men. The marriage rate has fallen roughly 60 percent since 1970. Cohabitation has gone from rare to standard: most first unions are now cohabiting, not marital. Nonmarital births sit near 40 percent of all births. Sexual frequency among adults under 35 is declining. A growing share of young men—roughly 30 percent in some surveys—report no sexual partner in the prior year. Fertility has dropped well below replacement across nearly every wealthy society. And loneliness, especially among men under 30, is rising on every measure.
The temptation is to flatten this into a single story: feminism, or pornography, or smartphones, or capitalism, or hookup culture, or the death of religion. None of these mono-causes survives the data. Andrew Cherlin's work shows that the institutional meaning of marriage has shifted from a cornerstone people built their lives on, to a capstone people achieve after assembling stability elsewhere. Sara McLanahan's "diverging destinies" framework shows the change is sharply class-stratified: college-educated Americans still marry, still have children mostly inside marriage, and still report relatively stable unions, while non-college Americans have seen marriage collapse and unstable cohabitations replace it. Kathryn Edin's ethnographic work shows that low-income women want marriage intensely but treat it as a luxury good they refuse to attempt until financial conditions are right—conditions that, for many, never arrive.
What links the data points is a coordination failure at the collective scale. Pair bonding is a coordination problem: it requires two people to commit at roughly the same time, in roughly the same place, with roughly compatible expectations, backed by a social environment that makes the commitment legible and defensible. The twentieth-century system solved this by force—through religious authority, economic dependency, kin pressure, and a narrow range of socially acceptable life scripts. When those scaffolds were removed, no replacement coordination layer was built. Each generation since has been asked to negotiate, from scratch and individually, what previous generations received as a default. The result is not a society of liberated romantic agents. It is a society in which a growing share of adults cannot find, hold, or trust the kind of bond they say they want.
Law 3—Connect—is the law most violated here, and most loudly. Humans evolved as obligately pair-bonding cooperative breeders. The infant cannot survive on one parent alone; the female cannot easily provision a slow-developing child without an allocator partner; the male, statistically, lives longer and behaves better with a partner than without one. Sarah Hrdy's evidence makes the cooperative-breeding picture unambiguous: humans are not built for solo parenting, and they are not built for serial low-trust mating either. The current arrangement—high search costs, low commitment, asymmetric expectations, and a near-absence of kin reinforcement—runs against the species design.
Law 2—Think—is violated by the diagnostic discourse around this data. The dominant framings (women's liberation broke marriage; men became weak; apps ruined romance) are each true in narrow slices and useless as wholes. What the data actually shows is a class-stratified institutional collapse with multiple proximate drivers and one shared distal driver: the absence of a load-bearing social architecture around the dyad.
Law 4—Plan—is violated at the policy level. Almost every Western government still designs tax, housing, childcare, and labor policy as if the standard household were a married pair with children. The standard household no longer is. The mismatch is not neutral. It taxes the disappearing form and subsidizes nothing in particular.
Law 0—Humility—is the most useful frame here. No one knows yet whether the new equilibrium stabilizes, collapses, or rearranges into a form we don't yet recognize. The data tells us the old machine is broken. It does not tell us what the next one looks like. Anyone confident about the answer has stopped reading the data.