Chapter 2

Updated 2026-07-02

"The baby, assailed by eyes, ears, nose, skin, and entrails at once, feels it all as one great blooming, buzzing confusion." — William James, The Principles of Psychology (1890)

Watch two kids in the back seat of a long car ride play twenty questions. Is it alive? Is it bigger than a breadbox? Can you eat it? Each question splits the world in two and throws half of it away. Twenty good splits can find one thing among a million. It is, formally, the search strategy a computer scientist would choose, invented from scratch, for fun, by people who cannot yet do long division.

---

In a developmental psychology lab in 1993, a three- or four-month-old sits on her mother's lap in front of a screen. She can't speak, can't point, can't be asked anything. What she can do is look, so looking is what gets measured. The screen shows her a photograph of a cat. Then another cat. Then another — different colors, different poses, different cats. Her gaze, long and serious at first, gets shorter with each one. There is no politer word for it: she is getting bored.

Then the screen shows a dog.

The gaze comes back. She stares the way she stared at the very first cat — as if to say, that one's new. And for that stare to make any sense, something else has to be true: all those different cats — fat ones, thin ones, sitting, walking — must have been filed together somewhere behind those eyes, in a pile the dog doesn't belong to. She has no word for cat. She has no words at all. Somewhere behind her eyes, without ever deciding to, she'd already sorted them.

Babies a few months old do this with cats and dogs, with male and female faces, with categories nobody would think to teach them.¹ The experiments are slow and patient and depend entirely on the fact that boredom is real and measurable at twelve weeks old. The conclusion is hard to overstate: sorting is not something children learn to do. It is closer to something they cannot help.

---

In the cloud forests of New Guinea lives a drab brown bird called the Vogelkop bowerbird, and the male builds, for courtship, one of the strangest structures in nature: a woven hut of sticks, a meter across, with a lawn of cleared moss in front of it. On the lawn he arranges his collection. Beetle wings, berries, flowers, fungus, bits of charcoal — and he does not heap them. He sorts them. Blue things in one pile. Red things in another. Orange in a third.²

When researchers mix the piles as a test, the bird puts everything back where it belongs, and he is not casual about it; he will pick up an offending berry and carry it from the wrong pile to the right one like a man returning a misfiled book.³

Young males are bad at this. Their piles are muddled, their color judgments approximate, and the females — who tour the neighborhood inspecting lawns before choosing — are not kind about it. Sorting well takes years to master, and sorting well is, for this bird, more or less the whole of seduction.

A bird sorting objects into piles by kind, getting better at it with practice, because his future depends on the piles making sense to somebody else.⁴ Whatever the urge to sort is, it is older than us.

---

Ask a four-year-old about a baby kangaroo taken from its mother at birth and raised by goats. Will it grow up to be a good hopper, or a good climber? Will it have a pouch? Hopper, the children say. Pouch. It makes no difference who raised it, what it eats, that it has never once seen another kangaroo — a kangaroo is a kangaroo somewhere underneath, and the underneath wins.⁵

Or walk slightly older children through a stranger story. Scientists take a raccoon, dye its fur black, bleach a white stripe down its back, sew in a sac of foul-smelling stuff — and the photograph the children are shown at the end is, to any eye in the room, a skunk. Is it a skunk now? No, they say, with the patience you'd extend to somebody a little slow. It looks like a skunk. It's a raccoon. The youngest children waver on questions like this; by seven or eight the verdict is firm, and it stays firm for the rest of everyone's life.

Insides beat outsides. Origin beats upbringing. What a thing is lives somewhere deeper than anything you can see or change about it. Psychologists call the belief essentialism, and it isn't taught any more than the cat-sorting was — it arrives early and spontaneously, ahead of anything school could have installed. And notice what never gets this treatment: no child on earth believes a Snickers is a trading candy underneath, down where it really counts. The grip is reserved for the categories the child takes herself to have discovered rather than built — kinds of animal, kinds of plant and, soon enough, kinds of people.

---

What do all these sorters have in common — the girl with the candy, the parents with the fire list, the baby with the cats, the bowerbird on his lawn, the seven-year-old holding the line on the raccoon?

Not correctness — or not one kind of it. The candy isn't sorted by what candy is; it's sorted by what the week needs — eating, trading, hiding — and it answers to nothing but the week. The fire list isn't an inventory of the house; it's an escape plan. The baby's cat-pile is something else: a bet about the world, useful exactly insofar as the next cat behaves like the last one, and the world gets a vote in whether it pays off. And the raccoon isn't a bet at all, as far as the child is concerned — it's a fact, sitting deeper than the evidence of her own eyes. One instinct, sorting at least three ways for three different reasons: piles built for a purpose, piles answerable to the world, piles believed in outright. Every one of them is still built — for something, by somebody, from somewhere — and a four-year-old runs all three before dinner without feeling a seam.

And only some of the piles get let go of lightly. The trading pile empties and is never rebuilt; nobody mourns it. The bowerbird's careful lawn gets scattered by a rival and rebuilt without complaint. Piles built for a purpose die with their purpose, easily, because somewhere below the level of ever thinking about it, the child knows she made them. But tell her she's wrong about the raccoon and she will not re-sort. She'll defend it — patiently, indefinitely — because that pile never felt made in the first place. It feels found. The instinct comes with two grips, and it picks between them by one unspoken question: is this a pile I built, or a pile I discovered?

That first, looser grip — sorted with conviction, held lightly — is the instinct's better half, and a four-year-old with a tub of toys has it for free. Keeping it is another matter. The piles on the living-room floor have one limitation: they live in one head, and they vanish. To share a category with strangers, to make it outlast the evening, somebody eventually has to write it down. A label. A shelf. A form with boxes on it. And a written-down category is a different creature from a living-room pile — it can do things no pile can do, organize libraries and cities and sciences, and it cannot do the one thing every four-year-old finds effortless.

It cannot let go.

What happens when the sorting instinct freezes solid — when the piles get carved into shelves and schedules and forms that millions of people have to live inside — is that the same impulse that once sorted candy by whim starts sorting people by rule. And the forms are not built by the candy half of the instinct. A box for what you are is the raccoon conviction with infrastructure behind it: a category that presents itself as found rather than made, and invites everyone standing in front of it to believe the same.

The fire list dissolves when the question ends. The candy piles don't survive November. The instinct that built them survives everything. A pile in one head is easy to let go of. What happens when that same instinct gets poured into concrete — a building, a border, a database — built to hold not one child's candy but millions of people at once, and never designed to let go?

---