This social learning analysis of Community Safety Partnerships in Scotland will develop two sets of arguments – one empirical and one epistemological. The empirical argument is that the well-documented difficulties in partnership working (largely a result of the very different occupational cultures, structures, roles and functions of the agencies generally brought on board) are not only very much in evidence but that current ways of organising and structuring partnership working in Scotland are also very often not conducive to overcoming them. It will be argued that viewing partnership working through the lens of a relational social learning perspective (Etienne Wenger’s theory of communities of practice) provides a clear set of recommendations for resolving these problems. These empirical arguments shall form the main focus of the thesis but, given the theoretical perspective employed, a related epistemological argument also emerged and shall be developed. It is generally accepted in theoretical criminology (and elsewhere in the social sciences) that the ideas and mentalities of the discipline have been shaped by the institutional contexts in which actors were doing criminology or criminal justice work (whether as practitioners or as scholars). Therefore, it will be argued that Community Safety Partnerships are important not only as sites of criminal justice practice but also as new institutional spaces in which ways of thinking about crime and community safety have the potential to be transformed. The empirical and epistemological arguments are interrelated because it will only be where the problems of conflict and communication within partnerships can be positively resolved that their potential to become sites of thinking that transcend traditional criminal justice mentalities will be fulfilled.