The question of whether child and adolescent mental disorders are best classified using dimensional or categorical approaches is a contentious one that has equally profound implications for clinical practice and scientific enquiry. Here, we explore this issue in the context of the forth coming publication of the DSM-5 and ICD-11 approaches to classification and diagnosis and in the light of recent empirical studies. First, we provide an overview of current category-based systems and dimensional alternatives. Second, we distinguish the various strands of meaning and levels of analysis implied when we talk about categories and dimensions of mental disorder--distinguishing practical clinical necessity, formal diagnostic systems, meta-theoretical beliefs and empirical reality. Third, we introduce the different statistical techniques developed to identify disorder dimensions and categories in childhood populations and to test between categorical and dimensional models. Fourth, we summarise the empirical evidence from recent taxometric studies in favour of the "taxonomic hypothesis" that mental disorder categories reflect discrete entities with putative specific causes. Finally, we explore the implications of these findings for clinical practice and science.