There has been an increasing number of policy-driven vulnerability assessments in southern Africa over the past five years (see for example www.wahenga.net and www.sarpn.org.za). As pointed out by Patt et al (this volume, Chapter 1), vulnerability assessments encompass a very diverse set of purposes and conceptual and methodological approaches. Stakeholders, whose needs to which assessments are designed to respond, are often concerned with very specific development issues, sectors or institutional purposes. It has been suggested, however, that vulnerability, that is, the inability of groups or individuals to secure well-being in the face of climatic stress, is best understood in terms of the totality of interacting environmental, social, economic and political stressors rather than through a focus limited to single stressors (Leichenko and O’Brien, 2001). Food insecurity in southern Africa is, for example, driven not only by failing food production, but also by threats to livelihoods derived from multiple non-agricultural sources, such as social networks, employment, forest products, trade and migration (Scoones et al, 1996; Ellis, 2003; O’Brien and Vogel, 2003; Wiggins, 2005). How does one ensure that assessments driven by diverse and sometimes relatively isolated development concerns generate sound understanding of vulnerability, including the interaction between multiple dimensions of vulnerability? Such understanding is critical to formulating holistic, rather than piecemeal, adaptation measures that may effectively address the causes of vulnerability.