On the strange arithmetic of ancestry: why your foremothers outnumber your forefathers, what that says about how humans mate, and where the falling birth rate is taking all of us.

Around five thousand years ago, something reached into the human family tree and pruned it from one side only. The maternal branches went on fanning out, generation after generation, wide and green. The paternal branches kept dead-ending. Line after line of fathers simply stopped, until across much of the world the diversity of the male lineage had collapsed to as little as one seventeenth of the female. For a stretch of the Bronze Age it is as if, genetically speaking, there were seventeen women reproducing for every one man.

That sentence is startling, and it is also easy to misread, which is the reason for this whole article. The number is real, drawn from the actual variation in living DNA. What it measures is not what it first sounds like. So I want to take the long way around, starting with the simplest and most vertiginous fact I know about being human, and using it to earn the harder ones.

The thesis is short to say and takes the rest of the piece to defend. You are descended from roughly twice as many women as men. That asymmetry runs all the way down, from the size of a single reproductive cell to the shape of human courtship to the population curve now bending toward its first peak in modern history. Along the way I will keep doing one thing on purpose: separating the numbers that are solidly established from the round figures that are useful pictures laid on top of them. The pictures are worth keeping. They just should not be mistaken for the measurements.

One in fifteen

Start with a question that sounds like a bar-room puzzle. Of all the human beings who have ever lived, what fraction are alive right now?

The most careful public estimate comes from the Population Reference Bureau, whose demographers Toshiko Kaneda and Carl Haub have updated the calculation for years. Their 2022 figure is that about 117 billion members of our species have ever been born. Against today's roughly 8 billion living, that puts the living at about 6.8 percent of everyone who has ever drawn breath, close to one in fifteen.

What to notice: you are one of the lit ones, one of roughly one in fifteen. Then slide the origin of modern humanity backward and watch the total swing by billions, which is the honest heart of the estimate.

The first thing to say about that 117 billion is how much art goes into it. Kaneda and Haub are refreshingly plain that the number is, in their words, "part science and part art." We have solid census data for a sliver of history and essentially nothing for the rest. For more than ninety-nine percent of the human story there was no one counting births, so the estimate leans on assumed birth rates stretched across tens of thousands of years, and small changes in the assumptions move the answer by billions. Push the starting date for "behaviourally modern" humans deeper into the past, from 50,000 years ago toward 190,000, and you fold in vast numbers of short, high-mortality lives, and the total climbs by something like 8 billion. The headline is robust in its shape and soft in its digits.

It helps to see how the sausage is made. The calculation starts by choosing a beginning, usually somewhere around 190,000 to 200,000 years ago for anatomically modern Homo sapiens, and then divides the long human past into eras with assumed birth rates. For almost all of that span the birth rate had to be punishingly high just to keep the species in existence, because the death rate was punishingly high to match. Infant and child mortality did most of the killing. Through most of history something close to half of all children died before reaching adulthood, so a birth rate that looks enormous by modern standards, forty or fifty births per thousand people per year, translated into only slight growth or none at all. That grim fact is why the total runs to so many billions: a species that barely grew for tens of thousands of years still had to keep replacing its dead, and every one of those short lives counts in the tally of the ever-born.

The estimate also exists partly to correct a specific myth. A claim floated around the middle of the twentieth century that the living made up a huge share, sometimes quoted as seventy-five or even ninety-five percent, of all the humans who had ever lived. The demographer Nathan Keyfitz and, later, Carl Haub set out to check it, and the real figure came back an order of magnitude smaller. What the number kills stone dead is the cocktail-party version of that claim, that the living now outnumber all the dead combined. They do not, and it is not close. The dead outnumber the living roughly fourteen to one. You belong to a small, bright, recent minority of your own species. Hold onto that framing, because the rest of this article is about how unevenly the paths to becoming one of those 117 billion were distributed between the two sexes.

The empty half

Here is where ancestry stops being symmetrical. Everyone has one mother and one father, so a single generation back the sexes are perfectly balanced. Keep walking backward, though, and the two sides of your tree grow at different rates, because the men and women of the deep past did not have children in the same proportions.

The evidence is written in two special pieces of DNA. Mitochondrial DNA passes only from mother to child, so it traces a pure maternal line. The Y chromosome passes only from father to son, so it traces a pure paternal line. Because each is inherited whole and unshuffled, geneticists can estimate how far back the common ancestor of everyone's mtDNA lived, and how far back the common ancestor of everyone's Y chromosome lived, and then compare.

The reason these two molecules give such clean readings is that neither one is shuffled. Most of your DNA is a blend, recombined every generation from two parents, four grandparents, and a rapidly widening fan of ancestors, so no single stretch of it traces a simple line. Mitochondrial DNA and the Y chromosome are the exceptions. They pass down intact, mother to child and father to son, accumulating only the occasional random mutation. Those mutations act as a molecular clock, and by counting the differences between people alive today, geneticists can estimate how long ago all the maternal lines converge on one woman, and all the paternal lines on one man. These are not the first man and first woman, and they did not live at the same time or even necessarily meet; they are simply the points where the surviving lines happen to coalesce.

They are not the same age. In a foundational 2004 study, Rachel Wilder, Zahra Mobasher, and Michael Hammer found that the most recent common ancestor of human mtDNA is roughly twice as old as the most recent common ancestor of the Y chromosome. Read through the population genetics, that points to an effective number of reproducing females across human history around twice the effective number of reproducing males. Two mothers, loosely, for every father. That is the plain sense in which you descend from about twice as many women as men.

What to notice: the gap is not fixed. Scrub back to the trough a few thousand years ago and the male line thins to a fraction of the female one, then recovers toward the present.

The two-to-one figure is the deep-time average. The extreme comes later and sharper. In 2015 Monika Karmin and a large international team read the fine structure of the Y chromosome across the world and found a startling dip. Beginning after the spread of agriculture, somewhere between about 8,000 and 4,000 years ago, the effective size of the male line crashed while the female line held and grew. At the bottom of that bottleneck the ratio reaches its notorious extreme, an effective female-to-male size as lopsided as seventeen to one in the most affected regions, before recovering toward the present.

Why then, and why only the men? The timing points at farming and everything that came with it: heritable wealth, settled land, formalized status, and social power that could be concentrated and passed down the male line. A 2018 modeling study by Tian Chen Zeng, Alan Aw, and Marcus Feldman made the mechanism precise. The bottleneck, they argued, is not mainly about a few emperors with enormous harems. It is about competition between patrilineal kin groups. When descent and property run through the male line, whole clans rise and fall together, and when a clan loses, its male lineages can be wiped out wholesale while its women are absorbed into the winners. The Y chromosomes of the losing side vanish; the mtDNA carries on. Do that across a continent for a few thousand years and you prune the paternal tree to a fraction of its former width.

Notice what this does not require. It does not need every winning man to have dozens of children, and it does not need any single generation of dramatic violence. It needs only that the variance in male success be organized along family lines and sustained over many generations, so that the losing lineages accumulate and eventually disappear while the winning ones spread. The bottleneck is a statistical result of a social arrangement, not the signature of a particular conqueror. It also did not last. As the curve in the visualization shows, the male line recovered as those tightly clan-structured societies gave way to larger, more mixed populations, and the extreme trough is a feature of a specific window in the Neolithic and Bronze Age rather than a permanent condition of our species. The depth of the trough also varied by region, cutting deepest where patrilineal social structure was strongest, which is exactly what the clan-competition model predicts.

What to notice: no single number appears on the trees, only their shapes. The maternal side keeps reaching the present; most of the paternal lines stop somewhere along the way.

The most famous face people attach to this is Genghis Khan. A 2003 study led by Tatiana Zerjal found a Y-chromosome lineage carried today by a remarkable share of men across the former Mongol empire, on the order of eight percent of the region, which works out to something like sixteen million living men. The lineage's spread across the empire's old footprint, and its estimated age, led the authors to suggest it traces back to Genghis Khan and his close male relatives. It is a genuinely striking result, and it should be stated with its own honest hedge: the men are thought to descend from that lineage, no verified DNA from Genghis Khan exists, and the attribution rests on geography and timing rather than a sample from the man himself. The Khan is a memorable illustration of concentrated male reproduction. He is not the proof of it.

The eighty-forty line, and where it really comes from

There is a much-repeated version of all this that you have probably met. The psychologist Roy Baumeister put it in a 2007 talk and paper roughly this way: throughout human history, maybe eighty percent of women reproduced, but only forty percent of men did. It is a vivid line, and it has been passed around, sharpened, and sometimes presented as hard data, which it is not.

What it is, is a memorable gloss on the real two-to-one signal from the genetics, with the hedges (the "maybe," the deliberately round numbers) usually sanded off in retelling. Baumeister was encoding the same asymmetry the DNA shows, in a form a lecture audience remembers. Take it in that spirit, as a picture of a ratio rather than a measured percentage, and it points the right way.

The picture holds together because of a hard piece of arithmetic that no amount of caveating can soften. Every child ever born has exactly one biological mother and exactly one biological father. So the total number of children credited to all women equals the total credited to all men, which means the average number of surviving children is identical for the two sexes:

$\bar{c}_{\text{women}} = \bar{c}_{\text{men}} = \frac{\text{total children}}{\text{(one parent of each sex per child)}}$

The averages must match. What need not match, and does not, is the spread. Female reproduction is biologically capped and tends to be distributed fairly evenly: pregnancy takes the better part of a year, the fertile window is finite, and very few women have more than a dozen children. Male reproduction has almost no such ceiling. One man can father hundreds of children in the same span in which one woman bears a handful. When a few men monopolize mating, the arithmetic has to balance somewhere, and it balances by forcing many other men to zero. A childless woman is comparatively rare across history. A man whose line ends is common. And a line at zero is pruned permanently, because it takes descendants to carry a lineage forward.

What to try: switch from the more monogamous case to high polygyny and watch a spike of childless men rise at zero while a long tail stretches to the right. The average never moves. Only the variance does.

That difference in variance, not any difference in averages, is the engine behind the empty half of the tree. It is also the bridge to a much older question, and a more contested one. If male reproductive success is so much more variable, if the winners win so much bigger and the losers lose everything, then selection should have shaped the two sexes to approach mating differently. That is where the argument leaves arithmetic and enters evolutionary psychology, and where the caveats have to get louder.

Why the game is not symmetric

The root of the asymmetry is older than farming, older than our species, and almost absurdly small. It is the size of a sex cell. By definition, females are the sex that produces the large, costly, few gametes, the eggs, and males produce the tiny, cheap, plentiful ones, the sperm. This is anisogamy, and from it a good deal seems to follow.

Charles Darwin had already sketched the shape of this in his account of sexual selection: in many species the males compete and display while the females choose, which is why peacocks carry the tail and peahens do not. What Darwin lacked was a clean explanation of why the roles usually fall that way. Anisogamy supplies it. Because eggs are expensive and few while sperm are cheap and plentiful, a female's reproductive output is limited by the resources she can turn into offspring, while a male's is limited mainly by the number of mates he can obtain. The two sexes are therefore playing for different prizes, and the game rewards different tactics.

In 1972 Robert Trivers built the influential bridge from that cellular fact to behaviour: parental investment theory. The sex that invests more per offspring, he argued, becomes the scarcer reproductive resource and therefore the choosier one, while the sex that invests less competes more fiercely for access. In most mammals the female invests vastly more, through gestation and nursing, so the theory predicts choosier females and competitive males. The theory's elegance is that it also predicts the exceptions. In seahorses and some shorebirds, where males take on the greater parental burden, the roles reverse and it is the females who compete and the males who choose. The prediction tracks the investment, not the chromosome, which is exactly what you want from a theory.

Humans complicate the picture in an interesting direction, because our children are unusually helpless for unusually long. A human infant cannot cling, walk, or feed itself for years, which raises the value of a second investing parent and helps explain why human fathers, unlike the males of most mammals, tend to stick around and provision their young. That shared investment pulls the two sexes closer together than the raw logic of egg and sperm would suggest, and it is part of why human reproductive skew is milder than that of a species where males contribute nothing but genes.

So far so tidy, and here the honesty has to kick in, because the classic supporting pillar has cracked.

The measured human reality is milder than the fruit-fly caricature, and that is the point to hold. Humans are mostly, if imperfectly, pair-bonding. Fathers in our species invest a great deal by mammalian standards. The variance gap between the sexes is real and shows up most sharply in societies where a minority of men hold many wives, and it is far from the seahorse-scale reversal or the fruit-fly-scale male free-for-all. The asymmetry is a tilt in the playing field, not a cliff.

Desire, and the argument about it

If the field is tilted, the tilt should leave traces in what people say they want. The psychologist David Buss made the boldest version of that claim in his 1989 study across thirty-seven cultures and in his book The Evolution of Desire. His headline findings are much cited: averaged across those cultures, women placed more weight than men on a partner's financial prospects in thirty-six of the thirty-seven, while men placed more weight on youth and on physical cues associated with fertility. Buss and David Schmitt later folded this into Sexual Strategies Theory, the idea that people carry a menu of short-term and long-term mating strategies and deploy them by context rather than following a single fixed script.

It is worth being concrete about what the studies actually report, because the caricatures run in both directions. Across cultures, when people rank what they want in a long-term partner, the top items are strikingly similar for men and women: kindness, intelligence, dependability, and mutual attraction sit near the top for nearly everyone. The sex differences show up further down the list and are differences of degree, a few rungs of relative emphasis, not opposite priorities. Buss also reported an asymmetry in jealousy: in several studies men were somewhat more distressed by imagined sexual infidelity and women somewhat more by imagined emotional infidelity, which he read as tracking the different reproductive threats the two sexes historically faced. That jealousy finding, it should be said, is among the more contested in the whole program, with critics arguing the effect is fragile and sensitive to how the question is posed.

These are among the most replicated findings in the field, and they are also among the most fiercely argued over, and an honest tour has to give the critics real space rather than a footnote.

The reasonable position, I think, is neither to wave the preference research away nor to treat it as a blueprint for how anyone must behave. There is a robust, cross-culturally repeated tilt in stated priorities that lines up with the reproductive asymmetry we have been tracing. The tilt is modest, it softens as societies grow more equal, and it says nothing about what any individual owes anyone or wants for themselves. Evolution shaped a set of nudges, working through the same evolved mismatch between ancient wiring and modern life that I traced in Cognitive Biases and Decision-Making. It did not write anyone's life.

The oldest game, on a new board

Everything so far has been about the deep past, but the asymmetry did not retire when the Bronze Age ended. It simply found new arenas, and the newest of them is a screen in your pocket. Over the last fifteen years courtship has moved online at a speed that has few precedents. A 2019 study by Michael Rosenfeld and colleagues found that meeting online had become the single most common way American couples get together, overtaking friends, family, school, and the neighbourhood. Whatever else the apps are, they are now the marketplace where a large share of pairing begins.

And the marketplace has a shape that should look familiar. When the OkCupid co-founder Christian Rudder dug into his own site's data for his 2014 book Dataclysm, he found that female users rated roughly eighty percent of male profiles as below medium in attractiveness, and that messages and attention concentrated heavily on a thin slice of highly rated men. Later looks at swipe data told the same story in sharper form. The attention a mating market hands out is not spread evenly; it piles up on a few, which is exactly the high-variance pattern that pruned the paternal tree, now playing out in real time.

What to try: drag the inequality up toward the dating-app estimate and read how little the least-liked half of men receive, then overlay national income to see that the swipe market can out-concentrate an entire economy.

There is a temptation to quote a hard number here, and it needs resisting with the same discipline as the eighty-forty line. The often-repeated figure that dating apps run a "Gini coefficient" of about 0.58, higher than the income inequality of most countries, comes from a 2015 blog analysis by a writer working under the name Worst-Online-Dater, not from a peer-reviewed study. Take it as an illustration of a real shape rather than a measurement of it. What is better established is the direction: on the female-choice side of the app economy, attention is markedly concentrated, which is the same long tail from the offspring chart, now made of likes instead of children. The Bronze Age skew has a digital echo.

Balance matters here too, and it cuts more than one way. The apps that concentrate attention also widen the pool, letting people meet far outside the small circles that used to bound courtship, and they have produced an enormous number of real and lasting relationships. Swiping is also not marrying; a rating on a screen is a low-stakes snap judgment, and it would be a mistake to read the harsh arithmetic of first impressions as the arithmetic of who ultimately partners and raises children. The market's entrance is far more unequal than its outcome.

Alongside the apps sits a quieter change that ties the mating story back to the population story: young adults, in several rich countries, are simply having less partnered sex. Drawing on decades of survey data, the psychologist Jean Twenge and others have documented a decline in sexual frequency among young people, with a rising share reporting no sexual partner at all in the previous year. The journalist Kate Julian gathered the strands into a 2018 essay that named the pattern a "sex recession." Japan has reported some of the most striking figures of low sexual activity among its young adults. Less partnered sex, later relationships, and more hours spent alone with a screen all point, gently, in the same direction as the falling birth rate.

The through-line from the Bronze Age to the smartphone is not that nothing has changed. It is that the underlying asymmetry, the higher variance of the male side of the reproductive ledger, keeps finding fresh expression, while the thing built on top of it, whether and when people form the partnerships that lead to children, has begun to shift in a way no earlier generation ever saw. That shift is the subject of the rest of this article, because it is bending not just individual lives but the size of the species.

The great descent

For almost the entire span we have been discussing, the human population grew, because on average each generation more than replaced itself. That is over. In the span of a single lifetime the birth rate has fallen off a height that would have been unimaginable to anyone born in 1950, and the fall is the hinge on which the whole demographic future turns.

The measure to watch is the total fertility rate, the average number of children a woman would have across her life at current rates. The threshold that matters is about 2.1, a touch above two because not every child survives to have children of their own. Above 2.1 a population eventually grows; below it, eventually shrinks. In the early 1950s the global figure sat around five. In 2024 it is roughly 2.25, hovering just over the line. The United Nations projects it drifting toward something near 1.6 by 2100.

What to try: toggle the regions. The single world line hides two almost opposite stories, a late and lofty descent across sub-Saharan Africa and an outright collapse across East Asia.

The descent has not been perfectly smooth, and one bump is worth naming because it shaped the modern intuition that birth rates are a thing governments can steer. After the Second World War much of the Western world saw a sharp, temporary rise, the baby boom, that ran through the late 1940s, the 1950s, and into the early 1960s before reversing hard. For a couple of decades it looked as though the long decline had paused or turned. It had not. The boom was a blip on a falling line, and by the 1970s fertility across the West was dropping again and has kept dropping since. Remembering that the last great reversal was temporary is a useful check on anyone who promises the next one will not be.

Demographers have a name for the shape of this, the demographic transition: societies move from high birth rates and high death rates, through a bulge where deaths fall first and births stay high, to low births and low deaths at the far end. The interesting question is what pushed each society along it, and here the causes are genuinely multiple and reinforcing.

Reliable contraception belongs near the front. The United States Food and Drug Administration approved the first oral contraceptive, Enovid, in June 1960, and the decades that followed saw fertility fall across the Western world as, for the first time in history, sex and childbearing could be reliably separated by choice. Set alongside it the single most powerful correlate demographers know, the education of women. As average schooling for girls rises from near zero toward six years and beyond, fertility tends to drop by something on the order of forty to eighty percent, an association that holds across very different cultures. Add women's entry into paid work, urbanization that turns children from farm labour into a cost, and falling child mortality that means parents no longer need many births to keep a few children, and the direction is overdetermined.

Some of the sharpest declines were pushed by the state directly. China's one-child policy, introduced in 1979 and enforced for more than three decades, drove the country's fertility down hard and left it with a distorted age structure and a surplus of men that it is still reckoning with. The policy was relaxed to two children in 2015 and to three in 2021, and the striking thing is what happened next, which was almost nothing: births kept falling anyway, because by then the deeper forces of urbanization, education, and cost had taken over from the mandate. When the government finally wanted more children, it found that permission was not the binding constraint. That is a preview of the trouble every pro-natal policy runs into.

Then there are the frictions of modern life that show up in the numbers. Children have become expensive in time and money. First births keep arriving later, which mechanically lowers completed family size. Housing costs press directly on the decision: one study attributed roughly fifty-one percent of the decline in United States birth rates in the 2000s and 2010s to the rising cost of housing. And underneath the economics sits a shift in what people want. A 2024 Pew survey found that forty-seven percent of American adults under fifty said they were unlikely ever to have children, and the most common reason given was not money or circumstance but simply that they did not want to. That last finding matters, because it means the descent is not only something happening to people. It is also, increasingly, something people are choosing.

The crest

Put those falling birth rates into the machinery of a population projection and something appears on the horizon that no living human has ever seen: a peak.

What to notice: the line that only ever climbed is now drawn cresting and turning down, and every recent revision has lowered the top.

The United Nations World Population Prospects 2024 puts the crest at about 10.3 billion in the mid-2080s, after which the global population begins, for the first time in modern history, to decline. It is worth watching the trend in the projections themselves, not just the projection: each recent UN revision has shaved the peak lower and pulled it slightly earlier, as the birth rates that feed the model keep coming in below expectation. A separate and lower path from the Institute for Health Metrics and Evaluation, published in the Lancet, puts the peak earlier and lower still, around 9.7 billion near 2064, followed by a steeper decline.

The future is not evenly distributed, and in several large countries it has already arrived. China's population began falling in 2022. Japan recorded roughly 900,000 more deaths than births in a recent year and continues to shrink. Italy has been logging its fewest births since national records began in 1861. Alongside shrinking comes aging: close to twenty-nine percent of Japan's population is over sixty-five, and South Korea crossed the threshold into a "super-aged" society, with more than a fifth of its people over sixty-five, in December 2024.

Why does any of this matter beyond the accounting? Because a society's age structure is the quiet foundation under its economy and its politics. A young, growing population has many workers supporting relatively few retirees, and the sums are easy. An old, shrinking one inverts the ratio: fewer workers carry more pensioners, health and care costs climb, and the pay-as-you-go systems that most rich countries built on the assumption of endless growth come under strain. None of this is destiny, and productivity, later retirement, and technology can absorb a good deal of it. But the arithmetic sets the terms of the problem, and it is the arithmetic that aging demography is rewriting first.

Where growth remains, it is concentrated. Sub-Saharan Africa carries most of the world's remaining increase, still descending its own fertility curve from a great height. And even where birth rates have already fallen below replacement, populations keep growing for a while through demographic momentum: the mothers of the next twenty years are already born and already counted, so the size of the young generation carries the total forward even after the per-woman rate drops. Momentum is why a country can fall below replacement and keep growing for decades, and why, once it turns, the turn is so hard to reverse.

What actually moves it

Faced with all this, wealthy governments have reached for their checkbooks, and the results are a sobering lesson in how little a determined state can buy in this particular market.

What to notice: if money bought babies, the cloud would climb to the right. It does not. The most committed spenders sit among the lowest rates.

Run down the scorecard honestly. France built one of the deepest family-support systems in the world over generations, and its rate still slid to about 1.62 in 2024. Hungary poured something near five percent of GDP into marriage loans, tax exemptions, and housing support; the rate rose for a while and then gave much of it back, settling near 1.38, with a good deal of the initial bump looking like a "tempo" effect, births moved earlier in life rather than added over a lifetime. The Nordic countries, long held up as proof that good policy could hold fertility near replacement, delivered the paradox that even there the rate fell, with Finland dropping to around 1.26. And South Korea, which has spent on the order of two to two hundred seventy billion dollars over roughly two decades, holds the lowest national fertility rate ever recorded, 0.72, ticking to about 0.75 in 2024.

The research consensus is more measured than either the boosters or the doomsayers. A careful 2021 review by Janna Bergsvik, Agnes Fauske, and Rannveig Hart found that one-off cash transfers mostly move the timing of births rather than the total number, while the more durable levers are the structural ones, generous parental leave and accessible childcare, which soften the fall without reversing it. No rich country has durably returned to replacement through policy.

Immigration is the one lever that reliably changes the headcount, and it deserves an honest accounting of both what it does and what it does not do. Bringing in working-age adults directly slows the aging of a labour force and props up the ratio of workers to retirees, which is why the countries with the most open immigration have generally aged more gently than the closed ones. What it cannot do is solve the underlying arithmetic at a global scale, for the simple reason that migrants come from somewhere, and every major sending region is now on the same descending fertility curve, often only a decade or two behind the receiving one. Immigration redistributes young people; it does not manufacture them. As the whole world approaches low fertility together, the pool that today's aging societies draw from is itself shrinking, which is why demographers treat migration as a way to buy time rather than a way to buy out of the problem.

There is also a cultural layer beneath the economics that the spending scorecards miss entirely. What people believe about the ideal family, whether children are seen as a natural stage of adult life or as one optional project among many, whether the surrounding society is built for raising them or quietly hostile to it, seems to move fertility in ways that a marriage loan cannot reach. This is the least quantified part of the story and the easiest to wave around irresponsibly, so I will keep the claim modest: the countries that have held fertility highest among rich nations are not obviously the ones that spend the most, which suggests that something in the texture of a society, and not only its budget, is doing part of the work.

The shape of the whole thing

Step back and the argument closes on itself. Look far enough into the past and the family tree is lopsided: you descend from about twice as many women as men, and for a stretch of the Bronze Age the paternal side was pruned to a fraction of the maternal, because male reproduction has always been the higher-variance game, with bigger winners and far more men left at zero. That same asymmetry, the tilt that came from a difference in the size of a sex cell, left faint and contested traces in how the two sexes approach courtship, traces real enough to measure and modest enough that they soften as the world grows more equal.

And now, in the span of a single lifetime, the game has changed underneath us. The very forces that gave women education, work, contraception, and choice have sent the birth rate through the replacement line across most of the planet, and bent the population curve toward a peak and a long descent that no policy has yet learned to reverse. The empty half of the tree was carved by concentrated male reproduction over thousands of years. The thinning tree of the future is being shaped by something new, hundreds of millions of individual choices, made mostly by women who now have them to make.

I have tried, the whole way through, to keep two things apart, and it is worth naming the discipline one last time. The striking numbers are real: the 117 billion, the twofold skew, the seventeen-to-one trough, the fall to 2.25 and the crest near ten billion. The round figures laid over them, the eighty and the forty, the tidy multipliers, the single tipping points, are illustrations, useful for holding a shape in mind and dangerous when mistaken for measurements. The family tree really is emptier on one side. How much emptier, and exactly why, is a question we should keep answering with the humility the evidence demands.