Although the devastating consequences of schizophrenia have long been known, interest in preventive intervention has only recently emerged. The shift in focus toward early treatment has been encouraged by findings suggesting that the longer psychosis remains untreated, the poorer the prognosis, and by the recent introduction of novel antipsychotic medications with a more benign side effect profile than conventional neuroleptics. In this paper, we argue that interest in prevention has outpaced the necessary scientific and ethical underpinnings for clinical trials involving the schizophrenia prodrome. Specifically, we maintain that the prodromal phase of schizophrenia is, at present, essentially a retrospective construct and that, as a result, the defining signs and symptoms currently in use must be validated in naturalistic, longitudinal studies. In particular, it is essential to establish solid base rates for schizophrenia in prodromal individuals before early treatment can be effectively evaluated. Additional ethical/scientific issues discussed include: (1) the need for an exit strategy (i.e. the determination of when to discontinue treatment in an individual who does not develop schizophrenia), (2) the advisability of pharmacological interventions that specifically target neurocognitive deficits, and (3) the possibility that antidepressant medications may be as effective or more effective, with fewer side effects, than antipsychotic medication for prodromal individuals.