The upcoming publication of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), provides an opportunity to revisit the seldom-addressed methodological issues in contemporary psychiatry. We think that DSM widely determines the scientific and clinical orientation of the discipline, and therefore provides a good vantage point to critique the current psychiatric methodology. The main scientific problem is a perseverative attempt at validating descriptively defined disorders that are standardized and simplified to achieve diagnostic reliability. Lack of a single psychiatric phenomenon that is valid, i.e. natural, for initiating any reduction limits research to inductive-probabilistic methods, basically correlational analyses. Furthermore, reduction in psychiatry is typically directed at basic sciences, neglecting general medical diagnoses as possible intermediary correlates. The subcategory Due to a General Medical Condition is elusive, and the biopsychosocial approach does no more than strengthen the brain-disease illusion surrounding DSM definitions by justifying psychiatry as a branch of medicine while failing to stipulate detailed medical assessment and discouraging psychopathology-based clinical reasoning. It is therefore no surprise that, although our understanding of the neural basis and mechanisms of behavior has improved along with advances in the neurosciences, not a single DSM disorder has been validated by the discovery of a specific cause, pathophysiology, or structural abnormality since the adoption of the descriptive approach in 1980. New knowledge involves single traits or dimensions of mood, thought, or behavior, none of which are specific to any disorder. The optimum approach today would be to redefine the discipline as neuropsychiatry.