In this study, we assessed the effect of a transfer of stimulus control procedure on the acquisition and cross‐setting generalization of manual signs with four profoundly retarded individuals. Two individuals were trained to produce appropriate signs to verbal instructions, whereas the other two were trained until manual signs were controlled by visual stimuli (objects). Data obtained from three probe situations (two for one individual) constituted the dependent variables of the investigation. The results, gathered within a multiple‐baseline design across signs showed that (a) individuals acquired the trained signs and (b) maintenance and generalization across settings (classroom, ward) and persons (teachers, ward staff) occurred, but was highly variable between and within individu