In a recent article, Andreasen and Flaum (Schizophrenia Bulletin, Vol. 17, No. 1, 1991) argued that greater emphasis should be placed on negative symptoms in the diagnosis of schizophrenia, leading to a less important role for positive symptoms. This article presents a counter-argument to this view. Positive symptoms are common and reliable and therefore highly useful diagnostically. First-rank symptoms, although not specific to schizophrenia, show good discriminability. No other type of symptom or investigative method can make such claims to usefulness. Although positive symptoms do not predict outcome, this is not a necessary function of diagnostic criteria. The predictive power of negative symptoms is, in any case, based largely on studies of patients with chronic disorder. Premorbidly impaired social development may interact with schizophrenia, worsening the prognosis. We believe positive symptoms have always been the essence of psychiatric disorder and should remain so. Increasing the diagnostic weight given to negative symptoms risks restricting the definition of schizophrenia excessively.