Schizophrenia is a clinical syndrome of both extraordinary importance and extraordinary complexity. Its conceptual history contains many perspectives on the "essential" nature of the illness. For example, Kraepelin in 1919 emphasized primarily onset and course, although he also stressed the importance of some symptoms such as changes in affect and volition. Bleuler in 1911 took a more cross-sectional approach and attempted to identify fundamental characteristic symptoms, especially stressing fragmenting of thought processes. Schneider's (1959) approach was cross-sectional, stressing a group of "first-rank symptoms." DSM-III and its successors attempted to achieve a synthesis of these concepts. Nevertheless, heterogeneity in the clinical presentation of schizophrenia is certain, and heterogeneity in pathophysiology and etiology is likely. Although we can now define a particular construct of schizophrenia with reasonable agreement, the construct must be recognized as provisional and based on a need to achieve consensus about definitions rather than on an understanding of pathophysiology and etiology. The major challenge confronting the student of schizophrenia is to identify its mechanisms and causes in order to develop improved strategies for treatment and prevention. Several different approaches have been proposed to achieve this goal. Early attempts to explore and validate the construct of schizophrenia stressed descriptive and epidemiological techniques; the "validity" of a given construct of schizophrenia would be determined by evaluation of familial aggregation, course and outcome, response to treatment, and laboratory tests. This earlier approach to validation is now complemented by one that draws on techniques from neuroscience and attempts to understand schizophrenia in terms of underlying neural mechanisms. While the earlier approach conceptualized schizophrenia primarily in terms of a single disease entity, the second approach is particularly useful for the exploration of subtypes or dimensions. Research strategies for the study of schizophrenia have been developed to explore its heterogeneity. Three different competing models are discussed: (1) A single etiopathological process leading to diverse manifestations, similar to multiple sclerosis; (2) multiple disease entities leading to schizophrenia by different etiopathological processes, similar to the syndrome of mental retardation; and (3) specific symptom clusters within schizophrenia reflecting different disease processes that come together in different ways in different patients. Each of these models has strengths and weaknesses for the identification of etiology and pathophysiology.