The Night They Voted Homosexuality Out of the DSM

On December 15, 1973, the Board of Trustees of the American Psychiatric Association voted to remove homosexuality from the Diagnostic and Statistical Manual of Mental Disorders. The vote was 13 to 0, with two abstentions. A disease that had been treated with electroshock therapy, chemical castration, aversion conditioning, and involuntary institutionalization was, in the space of one meeting, no longer a disease.

The decision was immediately challenged. A group of psychiatrists demanded a referendum of the full APA membership. In April 1974, the membership voted. 5,854 to 3,810. Homosexuality stayed out.

Sit with that for a moment. A category that had been used to justify decades of suffering; that had destroyed careers, marriages, and lives; that had been enforced through some of the most brutal interventions in psychiatric history; was removed not by a scientific discovery, not by a new finding about neurology or genetics or brain chemistry, but by a vote. A show of hands. A ballot. The membership of a professional organization decided, democratically, that a condition was no longer a condition.

This is the single most revealing event in the history of the DSM. Not because the decision was wrong. The decision was correct. Homosexuality is not a mental illness. The revealing part is the mechanism. The DSM changed because people voted. And if a diagnostic category can be voted in or voted out, then the DSM is not a scientific document. It is a consensus document. Those are different things, and the difference matters for every diagnosis currently in the manual.

How It Got In

Homosexuality entered the DSM in 1952 as a “sociopathic personality disturbance.” The classification reflected the prevailing psychiatric consensus of the era, which treated homosexuality as a deviation from normal psychosexual development. The theoretical basis was psychoanalytic; Freud had actually been ambivalent about whether homosexuality constituted pathology, but his American followers were not. The American psychoanalytic establishment in the 1940s and 1950s was aggressively pathologizing, and figures like Charles Socarides and Irving Bieber built entire careers on the premise that homosexuality was a treatable illness caused by faulty parenting. Domineering mothers. Absent fathers. The usual suspects.

The treatment protocols that followed from the diagnosis were not gentle. Aversion therapy paired homosexual stimuli with electric shocks or nausea-inducing drugs. Patients were shown images of same-sex couples while being shocked, then shown images of opposite-sex couples while the shocks stopped. The logic was Pavlovian and the execution was barbaric. Lobotomies were performed. Institutionalization was routine. Conversion therapy in various forms persisted for decades, and its echoes have not stopped; it remains legal for minors in the majority of American states as of this writing.

All of this was authorized by the DSM. The diagnosis was the license. Once homosexuality was in the manual, every intervention aimed at curing it had institutional legitimacy. Insurance covered the treatment. Courts cited the diagnosis. Parents committed their children on the basis of it. The diagnosis was not a description. It was a weapon.

The Science That Wasn’t There

The diagnostic category of homosexuality was never supported by the kind of evidence that the DSM claims to represent. There were no biomarkers. There was no neurological finding. The psychoanalytic theories about causation were speculative and unfalsifiable; you cannot design an experiment to test whether a domineering mother causes homosexuality, because the category of “domineering” is defined after the fact to match whatever the patient’s mother was actually like.

Evelyn Hooker’s 1957 study was the first major empirical challenge. She administered standard psychological tests to thirty homosexual men and thirty heterosexual men, then gave the results to expert clinicians and asked them to distinguish the two groups. They could not. The homosexual men showed no greater incidence of psychopathology than the heterosexual men. The study was methodologically sound and its conclusion was straightforward: homosexuality, by itself, was not associated with psychological impairment.

Hooker’s study should have been sufficient. It was not. The study was published, cited, discussed, and then largely ignored by the psychiatric establishment for fifteen years. The DSM-II in 1968 kept homosexuality in the manual, reclassified as a “sexual deviation” alongside fetishism and pedophilia. The reclassification was a demotion, not a removal. The institutional inertia was immense. Careers had been built on treating homosexuality. Treatment programs existed. Textbooks had been written. An entire clinical infrastructure depended on the diagnosis being real.

What Actually Changed the Vote

The removal of homosexuality from the DSM was not driven by Hooker’s research, though the research provided ammunition. It was driven by activism. Gay and lesbian activists disrupted APA conferences starting in 1970. They showed up at panels, confronted presenters, demanded to be heard. The most famous disruption was at the 1970 APA conference in San Francisco, where activists interrupted a session on aversion therapy and shouted down the presenter. In 1971, Frank Kameny grabbed the microphone at the APA conference and told the assembled psychiatrists: “Psychiatry is the enemy incarnate. Psychiatry has waged a relentless war of extermination against us. You may take this as a declaration of war against you.”

The pressure campaign worked not because the psychiatrists were intimidated, but because it created political conditions in which the existing evidence could finally be heard. Robert Spitzer, who would later become the architect of the DSM-III, began meeting with gay activists and reviewing the evidence. His conclusion was that homosexuality did not meet the DSM’s own definition of mental disorder, which required either subjective distress or impaired social functioning. Many homosexual people experienced neither. The distress they did experience was caused by stigma, not by the orientation itself.

Spitzer proposed a compromise: remove homosexuality as a general category but retain a new diagnosis called “sexual orientation disturbance” for people who were distressed by their own homosexuality. The compromise was adopted by the Board of Trustees, and when the membership referendum upheld the decision, the deal was done.

The compromise is revealing in its own right. “Sexual orientation disturbance” was later renamed “ego-dystonic homosexuality” in the DSM-III. It remained in the manual until 1987. The category essentially said: being gay is not a disorder, but being upset about being gay is a disorder. This is circular reasoning dressed in clinical language. The distress was caused by a society that pathologized homosexuality; the DSM had been a primary tool of that pathologization; and now the DSM was diagnosing the distress its own previous diagnosis had caused.

What the Vote Actually Revealed

The 1973 vote revealed something that the psychiatric profession has never fully reckoned with. The DSM is not a catalog of diseases discovered through scientific investigation. It is a catalog of conditions agreed upon through professional consensus. The distinction is not semantic. Scientific findings do not change by vote. The boiling point of water does not depend on how many chemists raise their hands. But diagnostic categories do, because diagnostic categories are not natural kinds. They are constructs; useful, sometimes necessary, but constructed by committees of professionals making judgment calls about where to draw lines.

This does not mean the DSM is useless. Many of its categories correspond to real patterns of suffering. Depression, schizophrenia, PTSD; these describe genuine human experiences that benefit from clinical attention. But the mechanism by which they enter and exit the manual is not the mechanism of scientific discovery. It is the mechanism of professional consensus, shaped by evidence but also by politics, by cultural attitudes, by insurance reimbursement requirements, and by the career incentives of the people on the committees.

The homosexuality vote makes this visible in a way that is hard to deny. A condition was in the manual for twenty-one years. It was treated aggressively. Lives were damaged. And then it was removed; not because someone found the cure, not because the underlying neuroscience shifted, but because the political and evidentiary landscape changed enough that the consensus flipped.

The Damage That Remained

The vote did not undo the damage. The people who had been subjected to aversion therapy did not get their nervous systems back. The people who had been institutionalized did not get their years back. The people who had been fired, divorced, disowned, or driven to suicide by a medical establishment that told them their sexuality was a disease did not receive apologies. The APA issued a formal apology in 2021; forty-eight years after the vote; acknowledging that the organization had contributed to discrimination against LGBTQ people. Forty-eight years is a long time to wait for sorry.

The treatment infrastructure that the diagnosis had authorized did not disappear overnight. Conversion therapy persisted. It persists now. As of 2026, conversion therapy for minors is banned in only twenty-two states and the District of Columbia. In the remaining states, a licensed therapist can legally attempt to change a minor’s sexual orientation using techniques derived from the same theoretical framework that the APA officially abandoned in 1973. The diagnosis was removed. The industry it built was not.

This pattern; diagnosis creates treatment infrastructure, diagnosis is removed, infrastructure persists; recurs throughout the history of the DSM. A diagnostic category authorizes an ecosystem of treatment protocols, clinical training programs, pharmaceutical interventions, and insurance billing pathways. Removing the category does not dismantle the ecosystem. The ecosystem has its own momentum, its own revenue streams, its own professional constituencies. It survives the death of the diagnosis that created it, the way a building survives the demolition of the scaffolding.

The Questions That Follow

If homosexuality could be voted out, what else is in the DSM by consensus rather than by evidence? This is not a rhetorical question. It is the question that every subsequent edition of the DSM has failed to answer honestly.

The DSM-III, published in 1980, attempted to address the problem by shifting to a descriptive, criteria-based approach. Instead of psychoanalytic theories about causation, the DSM-III listed observable symptoms and required a certain number of them for a specified duration. This was presented as a move toward scientific rigor. In practice, it was a move toward reliability; clinicians could now agree on whether a patient met the criteria; without a corresponding move toward validity; whether the criteria actually described a real, distinct condition. Two clinicians could look at the same checklist and reach the same conclusion, but neither could tell you whether the thing they were diagnosing existed as a discrete entity or was an arbitrary slice of continuous human variation.

The homosexuality removal exposed the machinery. It showed that the DSM’s authority rests not on scientific certainty but on the profession’s ability to maintain consensus. When the consensus holds, the categories look solid. When the consensus breaks; as it broke in 1973, as it has broken and reformed with every edition since; the categories reveal themselves as what they have always been: agreements. Useful agreements. Consequential agreements. But agreements, not discoveries.

The people who were shocked, drugged, institutionalized, and broken by the psychiatric establishment’s treatment of homosexuality were not treated for a disease. They were treated for a consensus. The consensus was wrong. And the mechanism that made it wrong; professional opinion, institutional inertia, cultural prejudice laundered through clinical language; is the same mechanism that produces every diagnosis in the current manual.

Gender identity disorder followed a similar trajectory; listed in the DSM-III, revised through subsequent editions, and replaced in the DSM-5 with “gender dysphoria,” a reformulation that attempted to shift the locus of pathology from the identity itself to the distress it might cause. The structural echo of the homosexuality removal is unmistakable. A category that pathologized a form of human variation was renegotiated through the same combination of activism, evolving evidence, and professional consensus that produced the 1973 vote. The question of whether gender dysphoria belongs in the manual at all is actively contested, and the contest is political as much as it is scientific, because the contest always has been.

That does not mean every diagnosis is wrong. It means every diagnosis is provisional. It means the confidence with which a clinician delivers a diagnosis should be tempered by the knowledge that the thing being diagnosed was designed by a committee, approved by a vote, and could be redesigned by the next committee. The DSM is a living document. This is presented as a strength. It is also an admission that the document’s contents are subject to revision not because the science advances but because the people writing it change their minds.