This article argues that, in public and policy contexts, the ways in which many scientists talk about uncertainty in simulations of future climate change not only facilitates communications and cooperation between scientific and policy communities but also affects the perceived authority of science. Uncertainty tends to challenge the authority of climate science, especially if it is used for policy making, but the relationship between authority and uncertainty is not simply an inverse one. In policy contexts, many scientists are compelled to talk about uncertainty but do not wish to imply that uncertainty is a serious challenge to the authority of scientific knowledge or to its substantial use in policy making. "Boundary-ordering devices," the contextual discursive attempts to reconcile uncertainty and authority in science, depend critically for their success on their "dual" interpretation: at a general level across a boundary and differently on either side of it. The authors empirically identify a range of such boundary-ordering devices in the climate field.