Children aged 5-6 years performed a problem-solving task designed to i
nvestigate the source of cognitive gain observed after peer social int
eraction. In a joint problem-solving context, pairs of subjects had to
work by using a tool whose function was to facilitate and regulate th
eir solving activity. The role of each partner in the dyads was establ
ished by the experimenter in the ''socio-regulated interaction'' group
but not in the ''free interaction'' group. A control group of single
children was also included in the present study. The assumption was th
at role allocation would facilitate semiotic mediation and lead to dif
ferentiation in the solving activities. The ''free interaction group''
obtained the highest performance in the joint problem solving context
, whereas a delayed post-test revealed that ''socio-regulated interact
ion'' leads to the highest improvements in individual performance. The
se findings are discussed with reference to the activity-regulation sy
stems and the appropriation mechanisms of regulation.