The present paper discusses the problem of psychiatric comorbidity within the context of Kuhn's theory of scientific revolutions. A conceptual analysis is carried out to clarify the role of those philosophical implicit positions that influence the empirical problems related to comorbidity. Psychiatric comorbidity is an artifactual byproduct of the Diagnostic and Statistical Manual of Mental Disorders (DSM) classification because of its internal characteristics. The authors of the DSM-III tried to handle comorbidity by adding hierarchical exclusion rules, but many exclusion rules were deleted in subsequent DSM revisions for philosophical reasons (prevalence of empiricism on conventionalism). The consequent explosion of comorbidity rates led the DSM toward a scientific crisis. Three possible intra-paradigmatic solution strategies are considered (disorders lumping, reintroduction of exclusion rules, and liberalization of comorbidity). The first two strategies might be efficacious but theoretically unacceptable. The third would be in line with the empiricist requirements, but its practical effect (dramatic increase of comorbidity rates) would exacerbate the crisis of the DSM pressing for a "revolutionary" solution. Finally, the waited effect on comorbidity of three alternative models (etiopathogenetic diagnosis, spectrum, and dimensional diagnosis) is briefly considered.