How can we register the participation of a range of elements, extending beyond the human subject, in the production of HIV events? In the context of proposals around biomedical prevention, there is a growing awareness of the need to find ways of responding to complexity, as everywhere new combinations of treatment, behavior, drugs, norms, meanings and devices are coming into encounter with one another, or are set to come into encounter with one another, with a range of unpredictable effects. In this paper I consider the operation of various framing devices that attribute responsibility and causation with regard to HIV events. I propose that we need to sharpen our analytic focus on what these devices do, their performativity-that is, their full range of worldly implications and effects. My primary examples are the criminal law and the randomized control trial. I argue that these institutions operate as framing devices: They attribute responsibility for HIV events and externalize other elements and effects in the process. Drawing on recent work in science and technology studies as well as queer theory, I set out an analytic frame that marks out a new role for HIV social research. Attentiveness to the performative effects of these devices is crucial, I suggest, if we want better to address the global HIV epidemic.