The coordination in open multi-agent systems is difficult to achieve, mainly because agents are autonomous. Social commitments have been proposed as a new coordination paradigm, suitable to describe agent communications or interactions. As some authors have pointed out, commitment enforcement is necessary, i.e., autonomous agents that violate commitments must be punished. In related work this is done by means of agent organizations or institutions, of which several different models have been proposed so far. In this paper we go one step further and we extend commitment theories to define roles and organizational structures using social commitments. We thus obtain a unified model of coordination in multi-agent systems: the expected behaviour of an agent can be defined using social commitments in both organizational and nonorganizational contexts.