Psychiatric nosology is widely criticized, but solutions are proving elusive. Planned revisions of diagnostic criteria will not resolve heterogeneity, comorbidity, fuzzy boundaries between normal and pathological, and lack of specific biomarkers. Concern about these difficulties reflects a narrow model that assumes most mental disorders should be defined by their etiologies. A more genuinely medical model uses understanding of normal function to categorize pathologies. For instance, understanding the function of a cough guides the search for problems causing it, and decisions about when it is expressed abnormally. Understanding the functions of emotions is a foundation missing from decisions about emotional disorders. The broader medical model used by the rest of medicine also recognizes syndromes defined by failures of functional systems or failures of feedback control. Such medical syndromes are similar to many mental diagnoses in their multiple causes, blurry boundaries, and nonspecific biomarkers. Dissatisfaction with psychiatric nosology may best be alleviated, not by new diagnostic criteria and categories, but by more realistic acknowledgment of the untidy landscape of mental and other medical disorders.