Purpose: The purpose of this paper is to portray social learning in cluster initiatives (CIs), namely, to explore, with the lens of the communities of practice (CoPs) theory, in what ways social learning occurs in CIs and discover how various CoPs emerge and evolve in CIs to facilitate a collective journey in their learning process. Subsequently, the authors address the following research questions: In what ways does social learning occur in CIs? How is social learning facilitated through the emergence and evolution of various CoPs in CIs? Design/methodology/approach: The study applies the abduction approach for the interpretation of the collected data and attempts to create the best explanations for the observations on the basis of the CoP theory. The qualitative study of four CIs helped to identify various ways that social learning occurs in CIs and the role of the identified CoPs in the process. Social learning is portrayed as a collective journey within and between CoPs, where the interactions of their members deepen their level of involvement and help them to enhance learning in their CoP. Findings: The paper shows ways that social learning occurs in CIs and describes the role of CoPs. It identifies three types of CoPs in CIs: participants, cooperators and locomotives. Additionally, it documents different ways of social learning in CIs, namely, one-way or two-way information transfer and raising awareness; demonstrating and inspiring; or motivating and educating. It also shows that while potentially every member of a CI has access to these practices, only a limited number of members are actually involved. Social learning in CIs is selective and some CI members accept their role as more peripheral in their CI. Research limitations/implications: The research shows the application of the CoP theory to the analysis of social learning in CIs, a peculiar type of clusters. It describes how CoPs in the studied CIs varied in terms of the occurrence of learning. Furthermore, it reveals how social learning related to the level of involvement of CI members, namely, with an increase of involvement, the members formed more selective CoPs and strengthened their social learning. Nevertheless, the qualitative approach in the study and the specific sample of the CIs chosen for the analysis do not allow a generalization of the findings. Practical implications: Although in different CoPs social learning occurs in different ways, at each stage of the development of CIs, the learning process is carried out on the basis of interactions created among members. Therefore, it is important to support the "soft" forms of cooperation within CIs – involving members in activities and developing interactions. In addition, to ensure the growth of their entire CI, coordinators should create conditions for the development of existing CoPs into higher forms, which better support learning. They should also adopt boundary-spanning roles between various CoPs to strengthen social learning in CIs. Originality/value: The literature on CIs, which are peculiar forms of clusters, is still underdeveloped. The research fills in the gap concerning the ways social learning occurs in CIs. It shows that selectiveness can be observed in this process, and emphasizes the role of interactions developed through CoPs and the benefits offered by them. The study applies the CoP approach. Consequently, it expands the theoretical base in view of the generally lacking studies on social learning in CIs in the literature on clustering. Because the CoP theory has rarely been applied in the management literature, it also augments this specific field. [ABSTRACT FROM AUTHOR]