In a Different Key: The Story of Autism

By John Donvan & Caren Zucker

16 min read

Why does it matter? Because every battle over autism has really been a battle over who belongs in the human family.

Most people assume the history of autism is a medical story — a condition waiting to be discovered, then understood, then treated. That's not quite what happened. What actually happened is a story about power: who got to name the condition, who got blamed for causing it, who got locked away because of it, and who eventually forced their way into the room where those decisions were made. Scientists mattered, but so did desperate parents, calculating bureaucrats, self-serving researchers, and eventually autistic people themselves, whose opinions about their own lives the system had spent decades ignoring. The label "autism" has been stretched, shrunk, weaponized, and reinvented so many times that the word itself is less a medical fact than a political history. Understanding that history means understanding something larger: how societies decide, slowly and badly and sometimes cruelly, who belongs.

The First Diagnosis Was Made in a Letter to a Mother, Not a Journal

In late September 1942, Leo Kanner — America's leading child psychiatrist — sat down to write a letter to a mother in Mississippi. He had been corresponding with Mary Triplett for four years, and she had finally lost patience. Why, she asked, had he never given her a real diagnosis for her son Donald? Kanner admitted she was right: he had never matched Donald to any existing label because no existing label fit. But in that same letter, he told her something he had told no one else. He had been quietly observing eight other children whose behaviors echoed Donald's in striking ways, and he had found a name for what connected them all: 'autistic disturbances of affective contact.' The defining feature, he explained, was an inability — present from the very beginning of life — to relate to other people, in children who were otherwise not significantly impaired.

That letter, dated September 28, 1942, was a carbon copy. It sat undiscovered in the Johns Hopkins archives for sixty-five years. Autism was first named not in a peer-reviewed journal but in private correspondence with a desperate parent. The official version of discovery runs from lab to journal to clinic. What happened instead was that a family's relentless search for answers pulled the diagnosis into existence — Mary and her husband Beamon not by design, but by refusing to stop looking.

Donald Triplett had given Kanner plenty to work with. Growing up in Forest, Mississippi, Donald had fixated on the famous Dionne quintuplets — five Canadian sisters who drew six thousand visitors a day in 1935 — not because he was moved by their survival story, but because they were a matching set of five, just like his paint bottles. Blue was Annette; red was Cecile. Always, regardless of context. His mind processed the world through pattern and sameness where everyone else processed social meaning. His mother Mary eventually described him in a letter as 'hopelessly insane,' because that was the only language available to her. The diagnosis of autism had not yet been invented.

Autism Is Older Than Its Name — and the People Who Had It Were Often Closer to Acceptance Than We Think

Autism has always existed. The diagnosis is new; the people it describes are not. And the surprise, when you trace their stories backward through the centuries, is not how badly things went for them — though often they did — but how often a small, tight-knit community found its own rough accommodation, centuries before any clinical framework existed to demand it.

Consider what happened to Hugh Blair of Borgue in 1740s Scotland. Hugh was a landed-class loner in his late thirties who compulsively washed his wig and then put it on backward. He collected twigs, feathers, and scraps of fabric from the ground and piled them in his attic. He wandered into neighbors' houses whether they were home or not, helped himself to whatever caught his eye, and attended every local funeral regardless of whether he knew the deceased. At dinner, when his cats plunged their paws into his spoon, he didn't push them away — he licked their paws clean. His neighbors, who knew perfectly well that his visits were never social calls, put up with him anyway. They had simply learned his ways.

His brother John eventually dragged him into court, arguing Hugh was too mentally incompetent to have entered the marriage their mother had arranged. Twenty-nine witnesses testified. The court agreed — ruling Hugh a 'natural fool' — and annulled the marriage. John got his inheritance. But the woman Hugh had briefly been married to refused to leave. She stayed. They raised two sons together. Hugh lived into his sixties surrounded by people who knew him and made room for him.

Uta Frith, one of the psychologists who helped establish autism's biological basis, studied the court records in the 1990s and called the evidence for autism 'unambiguous.' What the eighteenth century lacked was the word. What it didn't entirely lack — at least in one small Scottish community — was the basic human capacity to absorb someone profoundly different and keep going. The institution was not inevitable. The question was always whether the people around you had enough stability, enough familiarity, enough of whatever it takes to decide that you were worth the bother. Sometimes, remarkably, they did.

That informal logic — community as therapy, familiarity as tolerance — had no name and no theory behind it. What came next was the clinical era, when doctors would acquire the formal power to name what was happening. Some would use that power wisely. Others would use it to devastating effect.

The Most Damaging Theory in Autism's History Was Built by a Man With No Medical Credentials

Bruno Bettelheim had no medical degree, no psychology training, and no clinical credentials of any kind. He was an Austrian lumber merchant who had earned a doctorate in art history. What he had instead was genuine charisma and eleven months as a prisoner at Dachau and Buchenwald. By the 1950s and '60s, this man — operating under the title 'Dr. B' despite never having attended medical school — had become the country's most trusted voice on the human psyche, the dominant public explainer of autism, and the architect of a theory that told mothers their children's condition was their fault.

His 1967 book, The Empty Fortress, made the argument systematic. Drawing on his camp experience, Bettelheim mapped Nazi psychology onto autism symptom by symptom. Prisoners averted their gaze to avoid detection; autistic children avoided eye contact for the same reason. Prisoners retreated into repetitive behavior to preserve their sanity; autistic children stimulated themselves compulsively. The analogy was simple and complete: if the guards had broken those men, then mothers had broken their children. He insisted he never explicitly called mothers Nazis. That was technically true in the narrowest sense, and entirely misleading about what his readers understood.

The cruelest part of this story is not the theory itself but how it was deployed. Rita Tepper, a New York mother who had studied sociology, worked in child welfare, and read widely about psychiatry, sat in a circle of folding chairs at Bellevue Hospital while a social worker guided her and other mothers through confessions of their emotional failures. The program's premise was that the mother had to be treated alongside the child, because she was part of the problem. Rita was too conscientious — too credentialed — to dismiss the expert consensus. When the social worker asked about her earliest feelings toward her son Steven, Rita told the truth: she had once thought her jaundiced, spike-haired newborn looked like a chicken. The social worker's expression told her they were finally getting somewhere. Her honesty became evidence against her.

The theory's most damning origin belongs not to Bettelheim at all. The 'refrigerator' metaphor was coined by Leo Kanner — the same man who had discovered autism and correctly identified it, in print, as 'inborn.' Kanner introduced the image in a 1949 article and was quoted promoting it in Time magazine. It was only after he framed autism as the product of cold, emotionally withholding parents that the rest of psychiatry began taking his discovery seriously. Scientific recognition was the reward for abandoning a correct idea. The consensus that destroyed families like Rita's was built on a capitulation.

Parents, Not Experts, Broke the Consensus

Who actually breaks a scientific consensus? The official story is that better data displaces bad theory — a cleaner experiment, a larger study, the weight of accumulated evidence. What happened with the refrigerator-mother theory was entirely different. The theory wasn't overturned by the psychiatric profession correcting itself. It was demolished from outside, by parents who had every reason to be furious and eventually found someone with the credentials to give their fury scientific form.

Bernard Rimland was a Navy psychologist in San Diego who had never heard the word autism until his son Mark was born in 1956. When he finally tracked it down in a college textbook, the explanation waiting for him was that his wife Gloria had caused it — through emotional coldness, psychological withdrawal, the particular cruelty of the 'refrigerator mother.' Rimland didn't believe it for a moment. Not because the evidence contradicted his experience of Gloria, though it did. Because there was no evidence. He was a data man, and he could see that nobody writing about refrigerator mothers was offering any statistical backing at all.

So he spent two years assembling what no one had assembled before: every documented autism case he could find — 230 of them — gathered through library visits, handwritten letters to researchers in New York, London, and Amsterdam, and a near-photographic memory deployed because photocopying was too expensive. When he pulled the data together, the theory collapsed almost immediately. Nearly every mother of an autistic child was also raising children without autism. A pathologically cold mother who only damaged one child out of several made no logical sense. Psychotherapy aimed at the mother consistently failed to help the children; in one study, the 29 children who received what was described as high-quality therapy made no progress at all, while 13 who received little or no treatment managed to reach school readiness. The data wasn't even close.

His 1964 book, Infantile Autism, won a major academic prize and came with a foreword by Leo Kanner himself — the very man who had coined the refrigerator metaphor, now quietly making amends by endorsing his own corrector. Bruno Bettelheim refused all cooperation, writing that he would give Rimland 'no help' and dismissing him with a parting psychoanalysis: feelings, Bettelheim explained, were simply unimportant to a man like Rimland. He wasn't entirely wrong about that. What Rimland felt was useful. What he trusted was numbers. And the numbers said the theory was fiction.

The book's last pages included a 76-question diagnostic checklist Rimland intended for fellow researchers. Parents read it as a personal invitation, tore the pages out of library copies, and mailed them to San Diego by the hundreds. Each one got a handwritten reply and usually a long-distance phone call. The scientific demolition of mother-blaming turned, almost by accident, into the infrastructure of a movement. The people who changed what the world believed about autism weren't the experts. They were the parents the experts had been blaming.

Institutions Didn't Just Fail Autistic People — They Erased Them

While researchers and parents were fighting over causes, people like Archie Casto were simply disappearing.

When Archie Casto walked through the iron gates of Huntington State Hospital in 1919, he was five years old and knew five words. His thirteen-year-old sister Harriet watched their mother collapse into uncontrollable sobbing and say that some things are worse than death. Then Archie was gone, and Harriet was given a rule: never mention him to anyone. He had no brother. She learned to keep the secret.

Once inside, Archie ceased to have a history. No letters home. No photographs. No school file ever opened in his name. The only official trace of his existence appeared every decade, when a census taker arrived at the gate and recorded his name and age by law. A few digits in a ledger, proving only that he was still alive somewhere behind the fence. The institution didn't just confine Archie. It made him uncountable. You cannot advocate for people you cannot see.

Fifty years passed. He lost all five words. His smile eventually disappeared, replaced by what Harriet called his stony look. His teeth were pulled — a standard institutional method for managing children who bit. The state moved him to another hospital more than a hundred miles away without telling his family. Harriet, who had never learned to drive, lost contact entirely. He vanished twice: once from his family, once from his sister.

This was the mechanism the eugenics-inflected shame culture had quietly engineered. Doctors told parents that institutionalization was mercy; parents told neighbors the child had gone away for his health; neighbors learned not to ask. Each silence reinforced the next. By mid-century, researchers trying to understand where autistic people had gone found the answer in state hospital rolls. Kanner tracked down the first eleven children he had ever written about and found five had spent their entire lives in those facilities. A 1967 British study put three-fourths of autistic adults inside institutions. They weren't missing from history. They were in it, just locked out of the parts anyone could read.

Archie got out in 1988, at seventy-four, only because his elderly sister had finally learned to drive and a researcher named Ruth Sullivan confirmed what Harriet suspected: her brother had autism, a diagnosis that had existed for four decades without ever reaching him. At eighty-one, he saw the ocean for the first time. When he died in 1997, more than a hundred people attended his funeral. The hospital that held him for nearly seventy years is now a Walmart.

The Number That Launched an Epidemic Was Admitted, by Its Own Author, to Be Arbitrary

Think of prevalence statistics the way you'd think of a property boundary. The line on the map looks authoritative — coordinates, legal language, official stamps — but somewhere in the chain of title, someone stood in a field and made a judgment call about where to drive the stake. Victor Lotter drove that stake in 1964, and he was honest enough to say he wasn't sure it was in the right place.

Lotter was a South African psychologist assigned to conduct the world's first systematic count of autistic children, working from a list of 78,000 kids in Middlesex County, England. After months of school visits, hospital tours, and home visits that exposed him and his wife to more family suffering than any prior researcher had witnessed, he narrowed the pool to 61 children who all showed some autistic traits. Then he did something remarkable in its candor: he ranked them from most to least impaired, drew a line under the 35th name, and declared everyone above it autistic. His explanation for why the line fell there and not elsewhere: those above it were simply the ones 'it was thought should be included.' His word for the cutoff was 'arbitrary.' Thirty-five children out of 78,000 became 4.5 in 10,000 — the founding prevalence figure that every subsequent epidemic claim would be measured against, often cited as though it were a fixed property of nature.

Lotter himself warned that small changes in his assumptions could have pushed the number meaningfully higher or lower. He published those warnings in plain language in 1966. They were almost universally ignored by the people who needed his number to anchor their arguments. That's the uncomfortable pattern this history keeps producing: a researcher makes an honest, hedged estimate; the policy apparatus strips the hedges off; the number hardens into fact. The number that anchored an epidemic was, by its author's own account, a graduate student's acknowledged guess.

Fear Proved More Contagious Than Any Vaccine

The vaccine panic wasn't primarily a scientific failure. It was the inevitable result of a medical establishment that had spent decades dismissing parents — blaming them, ignoring them, treating their careful observations as hysteria — suddenly producing a rogue researcher who called their instincts a moral issue worth fighting for. The science was thin from the start. What held the movement together was something older and more combustible than data.

Andrew Wakefield's paper on twelve children — hedged throughout with phrases like 'might be' and 'possibly,' explicitly not claiming proof of a causal link — would normally have generated measured scientific discussion. Instead, it ignited a decade-long panic, and the reason why played out in the room at the Royal Free Hospital on February 26, 1998. When the hospital's own dean, Arie Zuckerman, rose to his feet and pounded the lectern — insisting hundreds of millions of doses had been given safely worldwide — he was doing exactly what medical authorities had always done when parents raised uncomfortable questions: asserting expert consensus, expecting deference. Wakefield did the opposite. He declared that the three-in-one MMR shot might be too much for some children's immune systems, called it a moral issue, and said that one more case of harm was one too many. A fellow panelist then mused aloud whether injecting three live viruses simultaneously might be 'an unnatural and unusual event.' The whole enterprise unraveled in half an hour.

To the twelve families whose children were in Wakefield's study, none of the counter-arguments that followed registered as science. They registered as more of the same. These were parents who had watched their children deteriorate, brought their observations to doctor after doctor, and been told they were wrong, naive, or in denial. Wakefield was the first person inside the medical world who treated their timeline — the MMR shot, then the symptoms, sometimes within days — as meaningful rather than delusional. When the WHO issued alarmed statements and the British press corps of experts lined up to call Wakefield an outlier, the families didn't see a scientific correction. They saw a wall of institutions defending themselves. They circled around him. The weaker the evidence got, the more the critique felt like persecution.

The same refusal to defer that had been right about refrigerator mothers, right about institutionalization, right that the medical establishment had failed their children, became the engine of a campaign that suppressed vaccination rates and endangered children. The tragedy is that the parents who wielded it had learned that lesson from a system that had given them every reason to.

The People the Movement Forgot Are Now Speaking for Themselves

Ari Ne'eman was a college sophomore in a necktie carrying a briefcase when he learned that NYU was plastering New York City with billboard-sized ransom notes — mock kidnapper letters signed 'Autism' and 'Asperger's Syndrome,' warning parents that the disease had taken their children. The campaign was designed to generate 700 million impressions. Ne'eman, operating out of a University of Maryland dorm room in December 2007, killed it in eighteen days. He mobilized fourteen established disability organizations, generated coverage in four major newspapers, and flooded NYU with 3,000 contacts until the university pulled every billboard. When a Washington Post reporter called to confirm the cancellation, Ne'eman put him on hold, pumped his fist, and came back sounding like someone who had been doing this for years.

What makes this victory complicated is what Ne'eman was fighting for, and who it left behind. The neurodiversity movement — whose founding provocation came from Jim Sinclair, an autistic activist who declared in 1993 that parents who wish their child didn't have autism are effectively wishing their autistic child didn't exist — rests on the proposition that autism is not a disease to be defeated but a neurological variation to be accommodated. For the tens of thousands of articulate, self-advocating adults who built communities on that idea, it was liberating. For parents like Liz Bell, it was something close to an insult.

Bell's son Tyler had the kind of autism that doesn't appear in magazine profiles. Low IQ, nearly nonverbal, prone to seizures, in daily digestive pain, requiring constant supervision so he wouldn't wander into traffic or a swimming pool at night. When Bell finally confronted Ne'eman at a 2009 forum, she put it to him directly: if Tyler could ever argue against cures the way Ne'eman was arguing against them now, she'd admit he'd been cured. Ne'eman didn't flinch or soften. He held his ground. Bell left respecting his integrity and thinking that his imperviousness to her son's reality was itself evidence of the condition he was claiming to represent.

The movement these two were arguing within had spent forty years fighting for the dignity of people who couldn't speak for themselves. Now the people who could speak were doing so — and the terms they chose sometimes made the most severely affected harder to see. Tyler Bell, and the tens of thousands like him who would never attend a forum or sign a petition, kept receding from the public argument about what autism is.

What We Still Owe the People We Learned to Count

The stranger on the New Jersey bus didn't read a policy brief. He didn't complete a sensitivity training. He just looked up one day and recognized Nicholas — had seen him enough mornings, enough times, to understand that this was someone who belonged on this route, in this neighborhood, in the ordinary flow of things. Not in a cure. Not in a settled definition. Not in a resolved argument between two movements that may never fully reconcile. In a neighbor deciding that someone counted.

Which is why the hardest thing the book leaves you with isn't the debates that got loud — it's the people who stayed silent. The severely affected adults who aged out of school funding at twenty-one and fell off every map: no housing, no employment, no movement fighting for them the way it once fought for children. In 2012, a Harvard study found that 500,000 autistic children were about to turn twenty-two, with no federal entitlement waiting on the other side. Most of them simply disappeared from the data — not because their needs went away, but because no one was counting anymore. The bus hasn't stopped for them yet.