Community radio typically relies on volunteers to produce and present stations’ programming. Volunteers are generally drawn from stations’ target communities and are seen as “representatives” of those communities. It is with such volunteers and their role as representatives of stations’ target communities that this study is concerned. It poses the question: “what are the central concepts that typically inform volunteers’ knowledge of their target community, and how do these concepts impact on their perception of how they have gained this knowledge, and how they justify their role as representatives of this community?” The dissertation teases out the implications of these conceptualisations for a volunteer team’s ability to contribute to the establishment of a media environment that operates as a Habermasian ‘critical public sphere’. It argues that this can only be achieved if volunteers have detailed and in-depth knowledge of their target community. In order to acquire this knowledge, volunteers should make use of systematic ways of learning about the community, rather than relying solely on knowledge obtained by living there. In a case study of Katutura Community Radio (KCR), one of the bestknown community radio stations in Namibia, the study identifies key differences in the way in which different groups of volunteers conceptualise “the community”. The study focuses, in particular, on such difference as it applies to those who are volunteers in their personal capacity and those who represent non-governmental and community-based organisations at the station. It is argued that two strategies would lead to significant improvement in such a station’s ability to serve as a public sphere. Firstly, the station would benefit from an approach in which different sections of the volunteer team share knowledge of the target community with each other. Secondly, volunteers should undertake further systematic research into their target community. It is also argued that in order to facilitate such processes, radio stations such as KCR should recognise the inevitability of differences between different versions of “community knowledge”.