This thesis demonstrates that all of the practices which define and produce the Tswana family involve dimensions of risk, conflict, and crisis – glossed as dikgang (sing. kgang) – that also threaten to undo it. Dikgang need constantly to be addressed in the right ways by the right people, in a continuously adaptive process of negotiation. Efforts to negotiate dikgang are also fraught, and often produce further problems in turn. I show that Tswana kinship is experienced, generated, and sustained in a continuous cycle of risk, conflict, and irresolution; and that it creates and thrives on crisis. In a kinship system renowned for its structural fluidity, I demonstrate that these processes chart the limits of family, and define relationships within it. I further suggest that understanding kinship in these terms provides unique insight into the effects of public health and social welfare crises – like the AIDS epidemic – which may work to strengthen Tswana families, rather than simply destroying them. However, governmental and non-governmental interventions responding to such crises operate according to different assumptions about the stability and fragility of the family, and its incapacity to cope with crisis. The thesis argues that the frustrations such interventions typically face may be traced back to divergent understandings about what constitutes and sustains family, and the role of conflict and crisis in that process. The effects of such interventions are linked to the ways in which they enable, invert, disrupt, or bypass everyday practices of kinship among the Tswana, and instantiate practices and ideals of kinship from elsewhere. I argue that holding these intervening agencies and families in the same frame illustrates suggestive links between the spheres of kinship and politics on both national and transnational levels.