This article illustrates the shortcomings of an objectivist epistemology in publicly accounting for technological accidents. Public inquiries convened in the aftermath of accidents tend to operate with such an objectivist approach and, as a result, usually assign blame to either or both of two causal culprits: technical malfunction and socio-organizational failure. Following Downer, I argue that a constructivist understanding of technological failure opens the possibility of a third type of cause—one that is epistemological in nature. Public inquiries frequently fail to (...) recognize these “epistemic accidents” and are thereby limited in their capacity to effect positive change. Using the case of a 2003 railway accident near Sydney, Australia, and the public inquiry that was convened in its wake, this article argues that recognizing the epistemological origins of failure holds promise for a more constructive way of accounting for and learning from technological accidents. (shrink)