THE CLASSICAL psychiatric nosology for schizophrenia1 for various reasons has never been fully satisfying to those who use it. Such descriptive labels as paranoid, hebephrenic, etc, do not reliably describe actual behavior, especially in the chronic stages.12 According to Arieti,2 in the later stages, it is difficult to distinguish a catatonic from a paranoid and different types converge toward one another. This can be seen in the fact that hospital records of individual patients, through the years, often carry a series of different diagnoses, the impression of each new psychiatrist being different from the last.
If classical diagnostic subcategories provided a basis for selection of treatments or if they were useful in predicting the outcome of a specific treatment such as tranquilizing drugs, their utility value would make them worthwhile. However, there is little evidence that such diagnoses have other than a theoretical