CHICAGO — America has reached a point where almost half its population is described as being in some way mentally ill, and nearly a quarter of its citizens — 67.5 million — have taken antidepressants.

These statistics have sparked a widespread, sometimes rancorous debate about whether people are taking far more medication than is needed for problems that may not even be mental disorders. Studies indicate that 40% of all patients fall short of the diagnoses that doctors and psychiatrists give them, yet 200 million prescriptions are written annually in America to treat depression and anxiety. Those who defend such widespread use of prescription drugs insist that a significant part of the population is under-treated and, by inference, under-medicated. Those opposed to such rampant use of drugs note that diagnostic rates for bipolar disorder, in particular, have skyrocketed by 4,000% and that overmedication is impossible without over-diagnosis.

To help settle this long-standing dispute, I studied why the number of recognized psychiatric disorders has ballooned so dramatically in recent decades. In 1980, the Diagnostic and Statistical Manual of Mental Disorders added 112 new mental disorders to its third edition, DSM-III. Fifty-eight more disorders appeared in the revised third edition in 1987 and fourth edition in 1994.

With over a million copies in print, the manual is known as the bible of American psychiatry; certainly it is an invoked chapter and verse in schools, prisons, courts, and by mental-health professionals around the world. The addition of even one new diagnostic code has serious practical consequences. What, then, was the rationale for adding so many in 1980?

After several requests to the American Psychiatric Association, I was granted complete access to the hundreds of unpublished memos, letters, and even votes from the period between 1973 and 1979, when the DSM-III task force debated each new and existing disorder. Some of the work was meticulous and commendable. But the overall approval process was more capricious than scientific.

DSM-III grew out of meetings that many participants described as chaotic. One observer later remarked that the small amount of research drawn upon was “really a hodgepodge — scattered, inconsistent, and ambiguous.” The interest and expertise of the task force was limited to one branch of psychiatry: neuropsychiatry. That group met for four years before it occurred to members that such one-sidedness might result in bias.

Incredibly, the lists of symptoms for some disorders were knocked out in minutes. The field studies used to justify their inclusion sometimes involved a single patient evaluated by the person advocating the new disease. Experts pressed for the inclusion of illnesses as questionable as “chronic undifferentiated unhappiness disorder” and “chronic complaint disorder,” whose traits included moaning about taxes, the weather, and even sports results.

Social phobia, later dubbed “social anxiety disorder,” was one of seven new anxiety disorders created in 1980. At first it struck me as a serious condition. By the 1990s experts were calling it “the disorder of the decade,” insisting that as many as one in five Americans suffers from it. Yet the complete story turned out to be rather more complicated. For starters, the specialist who in the 1960s originally recognized social anxiety — London-based Isaac Marks, a renowned expert on fear and panic — strongly resisted its inclusion in DSM-III as a separate disease category. The list of common behaviors associated with the disorder gave him pause: fear of eating alone in restaurants, avoidance of public toilets, and concern about trembling hands. By the time a revised task force added dislike of public speaking in 1987, the disorder seemed sufficiently elastic to include virtually everyone on the planet.

To counter the impression that it was turning common fears into treatable conditions, DSM-IV added a clause stipulating that social anxiety behaviors had to be “impairing” before a diagnosis was possible. But who was holding the prescribers to such standards? Doubtless, their understanding of impairment was looser than that of the task force. After all, despite the impairment clause, the anxiety disorder mushroomed; by 2000, it was the third most common psychiatric disorder in America, behind only depression and alcoholism.

Over-medication would affect fewer Americans if we could rein in such clear examples of over-diagnosis. We would have to set the thresholds for psychiatric diagnosis a lot higher, resurrecting the distinction between chronic illness and mild suffering. But there is fierce resistance to this by those who say they are fighting grave mental disorders, for which medication is the only viable treatment. Failure to reform psychiatry will be disastrous for public health. Consider that apathy, excessive shopping, and overuse of the Internet are all serious contenders for inclusion in the next edition of the DSM, due to appear in 2012. If the history of psychiatry is any guide, a new class of medication will soon be touted to treat them. Sanity must prevail: if everyone is mentally ill, then no one is.

Mr. Lane, a professor of English at Northwestern University, is the author of “Shyness: How Normal Behavior Became a Sickness.”