Jurors customarily do their work with very little by way of instruction from the court, other than about the law. This suggests that they enter the jury room with the relevant cognitive and interactional tools at the ready, drawn from everyday life. This paper focuses on a specific conversational device jurors use to do their work: conditional-contrastive inculpations, whereby the defendant’s actions are compared unfavorably to what a normal, innocent person would have done, with the implication that the discrepancy indicates (...) guilt. We examine the logic, variants, sequential precursors, and immediate consequences of this phenomenon in two real-life American criminal juries deliberating the same charges. This study offers a rare glimpse into the operation of real juries, and specifically the way in which they appropriate a practice from ordinary conversation in order to perform the unordinary work demanded of them by the legal system. (shrink)