This study focuses on a well-known but yet elusive concept: (technological) lock-in. We summarize what is known about the nature of lock-in and offer a critical view on history- dependent processes based on recent contributions to the literature. We discuss if lock-ins are really inescapable, especially when innovation is concerned. Also, we address the question if lock-in is a well-defined concept at all. To offer a fresh view on lock-in and to tackle the issues just raised, we employ the replicator dynamics model. By making a parallel between monopolization in the replicator dynamics and the occurrence of lock-ins, we show that the convergence of a system to a given outcome can be reversed, under certain conditions. We highlight the need for a more precise demarcation of the conceptual boundaries of lock-in and path dependence, both from the formal and the empirical side, and suggest that further structural features - for example users heterogeneity - may play a relevant role in affecting the outcome of dynamic allocation and competition processes.