In the different studies on the etiology of mental retardation (MR), a diagnosis has been reached in a remarkably variable proportion of cases. This is due to differences in methodology (population examined, age of the patients, degree of MR, time and protocol of investigation, expertise), method of classification (according to timing of defect, type of defect or both), and, in particular the different use of the term diagnosis. The latter varies with the employed criteria (causal vs. pathomorphological vs. pathophysiological), with the degree of exclusiveness (single cause vs. multifactorial/predisposing background factors) and particularly with the degree of implied certainty (definite vs. provisional). All these factors contribute to the fact that in general, the results of different clinical studies on the etiology of MR are not strictly comparable and cannot be pooled. A dynamic method of classification is proposed which combines both time of occurrence and type of defect, and includes the degree of certainty of a diagnosis. The method allows identification of patients which need follow-up and reconsideration of their cases.