This is part of what Hacking is getting at, I think, when he talks about semantic contagion. The idea of having one's legs amputated might never even enter the minds of some people until it is suggested to them. Yet once it is suggested, and not just suggested but paired with imagery that a person's past may have primed him or her to appreciate, that act becomes possible. Give the wish for it a name and a treatment, link it to a set of related disorders, give it a medical explanation rooted in childhood memory, and you are on the way to setting up just the kind of conceptual category that makes it a treatable psychiatric disorder. An act has been redescribed to make it thinkable in a way it was not thinkable before. Elective amputation was once self-mutilation; now it is a treatment for a mental disorder. Toss this mixture into the vast fan of the Internet and it will be dispersed at speeds unimagined even a decade ago.
Michael First, the editor of the Diagnostic and Statistical Manual, is quite aware of this worry. When I asked him how the DSM task force decides what to include in the manual, he told me there were three criteria. One, a diagnosis must have "clinical relevance"—enough people must be suffering from the condition to warrant its inclusion. Thus more data must be gathered on apotemnophilia before a decision is made to include it in the next edition. Two, a new diagnostic category must not be covered by existing categories. This may turn out to be the catch for apotemnophilia, because if the data suggest that it is a paraphilia, it will be subsumed into that category. "People have paraphilias for all kinds of things," First says, "but we do not have separate categories for all of them."
Three, a new diagnostic category must be a legitimate "mental disorder." What counts as a disorder is hard to define and, in fact, varies from one age and society to the next. (Consider, for example, that homosexuality was defined as a mental disorder in the DSM until the 1970s.) One way DSM-IV marks off disorders from ordinary human variation is by saying that a condition is not a disorder unless it causes a person some sort of distress or disability.
However, the fuzziness around the borders of most mental disorders, along with the absence of certainty about their pathophysiological mechanisms, makes them notoriously likely to expand. A look at the history of psychiatry over the past forty years reveals startlingly rapid growth rates for a wide array of disorders—clinical depression, social phobia, obsessive-compulsive disorder, panic disorder, attention-deficit hyperactivity disorder, and body dysmorphic disorder, to mention only a few. In trying to pinpoint the causes for this expansion one could, depending on ideological bent, point to the marketing efforts of the pharmaceutical industry (more mental disorder equals more profits), the greater diagnostic skills of today's psychiatrists, a growing population of mentally disordered Americans, or a cultural tendency to look to psychiatry for explanations of what used to be called weakness, sin, unhappiness, perversity, crime, or deviance. But the fact is that none of these disorders could have expanded as they have unless they looked a lot like ordinary human variation at their edges. Mild social phobia looks a lot like extreme shyness, attention-deficit disorder can look a lot like garden-variety distractibility, and a lot of obsessive-compulsive behavior, as Peter Kramer told me, "verges on the normal." The lines between mental dysfunction and ordinary life are not as sharp as some psychiatrists like to pretend.