Epistemological disputes in the philosophy of science often focus on the question of how restrained or expansive one should be in interpreting our best scientific theories and models. For example, some empiricist philosophers countenance only belief in their observable content, while realists of different sorts extend belief (in incompatible ways, reflecting their different versions of realism) to strictly unobservable entities, structures, events, and processes. I analyze these disputes in terms of differences regarding where to draw a line between domains in (...) which one has warrant for belief and those in which one should suspend belief and thus remain sceptical. I consider and defend the idea that the precise location of this line is subject to a form of epistemic voluntarism, and argue that a Pyrrhonian reading of the basis of such voluntaristic choice is both natural and transformative of our understanding of these debates. (shrink)