The Man Who Tried to Reverse-Engineer Human Nature
John Tooby, who died in November 2023, spent his career on a single, audacious bet: that the mind is not a blank slate shaped purely by culture, but a piece of evolved machinery — built by natural selection to solve very specific problems our ancestors faced on the African savanna, in small bands, over hundreds of thousands of years. Together with his wife and lifelong collaborator Leda Cosmides, he founded evolutionary psychology as a discipline, arguing that if you wanted to understand jealousy, or tribalism, or why people reason so badly about statistics, you had to ask not "what did this culture teach us" but "what problem was this circuit designed to solve, and under what ancestral conditions."
That reframing sounds modest. It wasn't. It meant treating the brain the way an engineer treats a device recovered from a wreck — not as a mystery to be described, but as a machine to be reverse-engineered, component by component, by asking what job each part was built to do. Tooby did this with unusual rigor and, as his Edge essays show, with an equally unusual willingness to turn that same cold instrument on the follies of the present, including the follies of scientists.
Why smart people believe stupid things together. One of Tooby's sharpest observations was about coalitions — the tribal instinct to bond through shared belief. He noticed something almost perverse: a political or religious group united by supernatural conviction can happily update its views on economics or climate, because those beliefs were never the glue holding the group together. But a coalition built on the claim of being "rational" and "scientific" has no such flexibility. To revise a belief there isn't just to update your model of the world — it's to betray the tribe. Question the consensus, even for good evidence, and you're not seen as a careful thinker; you're seen as a bad and unreliable ally, someone who might cost people their jobs, their friendships, their belonging. The result, Tooby argued, is that groups organized around "rationality" can be more resistant to correcting their errors than groups that never pretended to be rational in the first place — because admitting error there costs something no mere fact can outweigh: your place in the tribe.
Two jobs, one belief. This fed into a bigger idea he called the dual function of belief. We assume beliefs exist to track reality — to be tested, corrected, and improved by contact with the world, the way physics or chip design gets ruthlessly corrected by whether the bridge stands or the transistor switches. But beliefs also do social work: they signal loyalty, buy belonging, and coordinate us with the people we depend on. Tooby's point was that the less a domain gets tested against hard reality — and the more socially costly disagreement becomes — the more "network fixation" takes over, and belief drifts from being about truth to being about tribe. It's why he'd expect physics to stay comparatively honest while fields with looser feedback loops and higher social stakes are far more vulnerable to fashion, groupthink, and self-organizing collective delusion.
Compression is the price of thought. In "The Iron Law of Intelligence," Tooby framed the brain's whole design problem in a single, striking image: the universe is unimaginably vast and richly structured, while the brain — even the most brilliant one — is comparatively tiny. Evolution's answer was to build compression algorithms: mental shortcuts small enough to fit inside a skull but powerful enough to pay off in survival. The catch is that all compression is lossy. Every heuristic, every instinct, every "common sense" intuition is a bet that throws away almost everything in order to keep just enough to act on. Our minds aren't unreliable because they're broken; they're unreliable in exactly the ways you'd predict from machines built to be small and fast rather than complete and accurate.
"Culture" and "learning" are not one thing. Tooby was allergic to concepts that sound like explanations but are actually just labels. He pointed out that "learning" just means some interaction with the environment changed the brain's information states, by some unspecified mechanism — and "culture" just means information in one head somehow produces similar information in another head, again by some unspecified mechanism. Because we use one word for each, we assume there's one underlying phenomenon to study. There isn't. Trying to build a unified science of "culture," he said, is like trying to build a unified science of white things — eggshells, clouds, dandelion sap, human eyes, old MacBooks — objects that share a surface label and nothing else. The real work, in his view, was specifying the dozens of distinct cognitive mechanisms hiding behind those two convenient, misleading words.
We are the surviving improbability. Perhaps his most quietly staggering essay opened with the line that the most remarkable news in science is that he existed at all — and not just him. Of the roughly 5.5 billion people who've lived past puberty, he estimated maybe one billion would be here without modern sanitation, medicine, and market abundance. For nearly all of human history, most people died before finishing childhood or before raising a full set of children of their own. The comfortable modern default — parents assuming their children will outlive them — is one of the rarest conditions in the history of our species, a genuine triumph most of us never think to notice because we've never known anything else.
What ties all of this together is a kind of relentless, unsentimental curiosity — a refusal to let familiar words (learning, culture, belief, rationality) stand in for actual understanding, paired with a willingness to point the same unflinching analysis at scientists, tribes, and himself. He spent fifty years trying to read the source code evolution left in us, on the theory that