[The classification of pervasive developmental disorders and other disorders: toward the DSM-IV]

Forty years after Kanner's groundbreaking work and many publications later, the outlook for children with autism has not changed markedly. Progress in understanding the etiology, let alone in developing new approaches for treatment, has been extremely slow for this and other serious, early onset disturbances. Historically, disagreements over diagnosis and syndrome definition have impeded research efforts. Indeed, in many cases it is impossible, on reading research reports of even a decade ago, to know exactly what population was being studied. DSM-III has broken new ground in its emphasis on developmental disorders viewed within a multiaxial diagnostic scheme. Increased diagnostic precision should facilitate the application of rigorous bio-behavioral research methods, which, in turn may help increasingly to define homogeneous populations. Refinements of nosology should not be mere academic exercises. Rather, they should suggest hypotheses which are amenable to empirical test. DSM-III, in addition to pointing the way out of the era of psychiatric Babel, has suggested empirical studies which have challenged its assumptions. In the field of nosology, this should be considered a major criterion of success. Changes in nomenclature, while sorely needed, should be undertaken with appropriate skepticism and conservatism and should build upon the foundation provided by DSM-III.