Contextual learning starts from the premise that learning cannot take place in a vacuum, but should somehow be connected with real world attributes to make sense to learners. Today, digital media tend to bring about new dimensions of context: internet connections and mobile devices enable learners to overcome restrictions of time and location, and neglect the physical boundaries and limitations of the learning environment. This calls for reconsidering contextual learning. This paper takes a theoretical stand by conceptualising the notion of learning context in the light of its virtualised extensions. It explains the historical and pedagogical backgrounds of contextual learning and reviews existing models that deal with context parameters. The paper identifies and discusses the constituting components of context for learning and it demonstrates how attributes of virtual representations affect the nature of context. The overall purpose of the paper is re-establishing the notion of contextual learning in the light of emerging digital media and making explicit the various dimensions involved. (Contains 1 table and 3 figures.)