On the sociology of lost ideas in schizophrenia research and the watering of gardens.

What follows is the third in a series of guest editorials that we plan to publish in the Schizophrenia Bulletin. Other readers are encouraged to submit short commentaries on schizophrenia-related topics that they feel merit attention and discussion. Responses to the issues 'aised in the editorial below will also be welcome. Although written by an NIMH staff member, the editorial does not necessarily reflect the opinions, official oolicy, or position of the National Institute of Mental Health.-Jhe Editors. Just as there are fashions in etiquette, sexual mores, and dress styles in our culture, so there are also fashions in schizophrenia research. Some research ideas are in fashion, while others are considered unfashionable or old fashioned. The fashion cycle runs not in 2-to 3-year trends, like women's dresses, but rather in 10-to 15-year trends, like men's ties. A psychiatric Rip Van Winkle, asleep for 20 years, could quickly update himself on the current fashions by skimming the psychiatric journals or by sitting in a grant-review committee meeting. Unfashionable ideas draw scathing barbs from the anonymous reviewers of potential journal articles, and the wise men of the review committees say, "Yes, the methodology is sound but (furrowing their collective brow) the idea isn't very good." Low marks. Rejected. Approved but unfunded. One consequence of these fashions is that potentially valuable research ideas get discarded for years at a time, stored away in the attic with old-fashioned narrow ties. Often the discarding process is sanctioned by a prestigious textbook of psychiatry which declares the idea disproven and therefore dead, when in fact it may not be so at all. But once the obituary of the research idea has been put into print, it gets passed from textbook to textbook and everyone knows it is Truth. Let me illustrate. Virtually all textbooks of psychiatry say that it is a proven fact that schizophrenia is universal and occurs in approximately equal incidence