Courts are notoriously skeptical of recantation evidence, in part because of finality concerns, but also in part because judges tend to treat such statements as less reliable than the original, in-court testimony that was previously given. This essay argues that such skepticism is unwarranted, or is at least exaggerated, because it is based on a misunderstanding of the dynamics of what the essay refers to as the "perjury sword" - credible threats by police and prosecutors to bring perjury charges against witnesses who wish to recant prior statements. In particular, this skepticism reflects a blindness to the ease by which use of the perjury sword can induce false testimony and contribute to wrongful convictions.

This essay sets forth the argument for reforming our approach to recantation evidence. While acknowledging some of the challenges police and prosecutors face, particularly when dealing with crimes committed in high-crime, urban contexts, and in domestic violence cases, the essay suggests a variety of strategies to mitigate the worst abuses of the perjury sword. These strategies include reducing witness exposure to perjury charges, expanding the recantation defense, and reconsidering how courts evaluate recantation testimony. While it is imperative that witnesses testify truthfully at trial, serious problems arise when perjury sanctions are deployed to discourage honest recantations. Trial truth is optimal, but truth delayed is better than no truth at all. The legal system must permit the responsible consideration of recantation evidence, even recognizing that doing so has a cost in terms of finality.

The essay concludes by questioning the legal system's strong attachment to monistic accounts of truth. Perhaps we would be better served if we embraced a more pluralistic conception of what constitutes truth in criminal justice, if not in the world more generally.