Replication is one of the most important tools for the verification of facts within the empirical sciences. A detailed examination of the notion of replication reveals that there are many different meanings to this concept and the relevant procedures, but hardly any systematic literature. This paper analyzes the concept of replication from a theoretical point of view. It demonstrates that the theoretical demands are scarcely met in everyday work within the social sciences. Some demands are just not feasible, whereas others are constricted by restrictions relating to publication. A new classification scheme based on a functional approach that distinguishes between different types of replication is proposed. Next, it will be argued that replication addresses the important connection between existing and new knowledge. To do so it has to be applied explicitly and systematically. The paper ends with a description of procedures how this could be done and a set of recommendations how to handle the concept of replication in the future to exploit its potential to the full.