Social learning in multi agents system: an additional step toward intelligence

According to Lev Vygotsky, individual\´s learning is influenced by social interactions with his environment. This confrontation of points of view enables an individual to evolve during the time. The Vygotsky\´s theory gives a quite important place not only to social tools as language, but also to interactions and relationships between individuals. From this report, we try to see how, in a multi agents system, an "intelligent agent" can learn and more easily integrate a new environment and a new agent\´s community. We propose a three steps process to ensure a new adaptive agent integration, which gives importance to the language and the relationships between agents.