Towards a logical model of social agreement for agent societies