The concept of lock-in can certainly be listed among those weighing most heavily in the conceptual toolbox used by scholars of innovation and evolutionary economics. Processes of competitive diffusion, or choice between alternatives of ‘unknown merit’, are known to generate lock-in, that is, inflexible outcomes, and this finding has critical implication for the study of economic dynamics and history-dependent processes. In this chapter, we first summarize what is known in the economic literature about the nature of lock-in, and we discuss if lock-ins are really inescapable, especially when innovation is concerned. Second, we employ the replicator dynamics model, suggesting a parallel between monopolization and lock-in, and show that the convergence of a system to the dominance of a single alternative does not have to be inescapable; rather, it is strongly dependent on the regime and parameters characterizing the competition between alternatives. In particular, the interaction of positive reinforcements driving market selection and negative reinforcements occurring at the level of each individual alternative generates outcomes far from the lock-in into one uncontestable alternative.