The nomenclature and classification criteria for arthritis in children should be dealt with initially as separate issues, although they are undoubtedly intertwined. The classification criteria should aim to delineate homogeneous patient populations, yet should be flexible enough to incorporate advances in disease knowledge. It should be recognized that arriving at an international consensus for classification criteria will merely provide a set of operational definitions to facilitate research, and not a set of diagnostic criteria. Indeed the only point to obtaining consensus is to begin a process of systematic ongoing review of the criteria. The labels attached to any of these diseases should facilitate accurate communication. In view of the heterogeneous nature of childhood arthritis, consideration should be given to using a broad umbrella term such as juvenile or childhood arthritis only for communicating with the lay public. Medical nomenclature should be formulated to reflect accurately homogeneous subgroups of arthritis, and should not artificially proscribe a relationship between paediatric and adult disease.