In this paper we develop a formalization and algorithms that can manage the evolution of several ontologies from different contexts, using automated reasoning. It is in general difficult to maintain consistency between several ontologies, but we focus on developing computationally efficient ways of achieving this. Our formalization uses both the notions of several local contexts and of a sequence of states. We believe such a system can become a component in for example a distributed knowledge management system or some other knowledge infrastructure that requires semantic autonomy, i.e. lack of centralized semantics, but presence of a type of semantic coherence. In this paper version we summarize our approach.