That question seems to be asked by not just one but two former chairman of the committees who authored editions of the fabled DSM. Dr. Allen J. Francis has been very critical of the currently being written DSM 5 ( Yes, they changed from Roman to Arabic numbers). An interesting family feud is depicted in this article from Wired on line.
The first time I came across this general thought was when I was discussing a case with a psychiatrist to whom I had referred a patient I thought might be chronically depressed.He said she was not depressed but in his view she was just a very unhappy person. I can't recall if he then proposed to treat her with the antidepressant flavor of the day or just talk her out of being unhappy or what.
While non psychiatry medicine which I will refer to as "real medicine" (somewhat tongue in cheek) sometimes uses a symptom counting method of diagnosis ( e.g use of a number of symptoms in the a major or minor category to reach a diagnosis) it has the cognitive luxury of being able to rely on physical findings, imaging techniques, sometimes fairly definitive blood tests and often definitive biopsy results.
However, our psychiatry brethren- as best I can tell as an outsider- depends on a symptom counting based diagnostic paradigm outlined in excruciating- to- read- details in the powerful DSM. For example,if a patient has five out of the nine official symptoms of depression over a two week period then they are diagnosed with major depression.Apparently for a while there was a "bereavement exclusion" for depression diagnosis but with the newest classification that has been removed, exemplifying how fluid the definition of a given disease can be.This descriptive diagnosis system was heralded as a marked improvement over the situation in the old days in which psychiatrists of different psychoanalytic schools would differ as to the diagnosis of a given patient.
Symptom listing,counting and matching was considered more scientific and importantly became available to non-psychiatrist physicians ( or any clinical health care provider). With a check list type diagnostic paradigm generally available and multiple psychiatric drugs being promoted by big pharma, primary care docs might conclude that "this psychiatry stuff is easier than I thought" and I suspect many did.Numerous dinner CMEoid sessions with primary care docs featured as the "thought leader"showing their fellow PCPs ( and NP and PAs) that psychiatric diagnoses were not something to be afraid of probably paid off for the sponsoring drug companies.