A host of errors can occur at sentencing, but whether a particular sentencing error can be remedied may depend on whether judges characterize errors as involving a “miscarriage of justice”—that is, a claim of innocence. The Supreme Court’s miscarriage of justice standard, created as an exception to excuse procedural barriers in the context of federal habeas corpus review, has colonized a wide range of areas of law, from “plain error” review on appeal, to excusing appeal waivers, the scope of cognizable claims under 28 U.S.C. § 2255, the postconviction statute for federal prisoners, and the “Savings Clause” that permits resort to habeas corpus rather than § 2255. That standard requires a judge to ask whether a reasonable decisionmaker would more likely than not reach the same result. However, the use of the miscarriage of justice standard with respect to claims of sentencing error remains quite unsettled. In this Article, I provide a taxonomy of types of innocence of sentence claims, and describe how each has developed, focusing on federal courts. I question whether finality should play the same role regarding correction of errors in sentences, and I propose that a single miscarriage of justice standard apply to all types of sentencing error claims, when not considering on appeal under reasonableness review. Finally, I briefly describe how changes to the sentencing process or sentencing guidelines could also reflect certain concerns with accuracy.

It is customary at the USC Gould School of Law to commemorate the publication of books authored by members of the faculty. A while before the publication of In Doubt: The Psychology of Criminal Justice, Dean Robert Rasmussen summoned me to discuss a way to commemorate its release. The conversation quickly converged on the idea that rather than hold an event to celebrate the publication of the book, we should seize the opportunity to hold an earnest discussion about the core issues raised in it: What has brought the criminal justice process to its current state, and more importantly, where should the process go from here? In that vein, we invited leading figures working at the forefront of these questions to participate in a conference: Criminal Law at the Crossroads. We also invited the speakers to submit their papers for publication in a special Symposium by the same name, and they responded graciously. It is an honor to pen the opening article of this Symposium.

What causes crime and why crime rates vary over time and place—these are vast questions that dominate the careers of criminologists. The related question of what we can expect government agencies to do about crime—that is, what we can hold government responsible for—occupies much of our civic discourse. My subject here is deceptively more modest: how we identify and address one major problem that government agencies, most obviously criminal justice agencies are supposed to resolve: the elusive phenomenon called recidivism. In this Essay I will undertake some admittedly impressionist reflections on recidivism. I will suggest that because of the power and salience of the term in our discourse, we need to be more self-critical in deploying it. Turning to more pragmatic concerns of criminal justice, I will review how variable and contingent are the formal definitions of measures of recidivism, and I will address the need for sensibly self-critical stipulations of the meaning of the term in order to make the most of any pragmatic use of the term feasible. But first, to suggest what a multimeaning term recidivism is and what a complex phenomenon it may signify, I beg indulgence for a quick narrative of developments in California.

The American death penalty is often described as anomalous, distinctive, or exceptional in the sense that at present, in the early years of the twenty-first century, the United States is the sole Western democracy that retains the practice of capital punishment. However, a second aspect of American exceptionalism in this context has largely escaped notice. The United States has chosen not merely to retain the death penalty while its peer nations have abolished it; rather, the United States has embarked on nearly 40 years (since 1976) of intensive, top-down, constitutional regulation of the practice by the federal courts, led by the U.S. Supreme Court. The choice of regulation in the place of mere retention has produced a complex web of interactions among the federal judiciary and state and local legislatures, executive officials, courts, and of course activists on both sides of the issue and the general court of public opinion. Close study of these interactions generates a compelling and dynamic story that sheds a great deal of light on the death penalty itself—on its functions and meanings in American society and politics, on its history, and on its future.

Thirty-two-year-old Eric Rinehart was a former police officer and member of the Indiana National Guard. He was going through his second divorce, he had custody of his seven-year-old son, and he had no criminal record. During this time, perhaps against his better judgment, he began two sexual relationships with young women, aged sixteen and seventeen. Although the young women were much younger in age, both of Rinehart’s sexual relationships were consensual and entirely legal. Under Indiana state law, the legal age of consent for sexual intercourse is sixteen.

During the course of his relationship with one of the young women, Rinehart lent her his digital camera after she suggested, based on her past experiences with other partners, that she use it to take provocative photographs of herself. When she returned the camera, Rinehart found pictures of the young woman engaged in “sexually explicit conduct.” Following this event, Rinehart photographed the same young woman engaged in similar sexual activities. In addition, Rinehart created “short videos of himself and [the second young woman] engaged in sexual intercourse.” All the photos and videos were taken with the knowledge and consent of his sexual partners. All of the images were uploaded onto Rinehart’s home computer, but none were distributed to a third party, nor was there evidence that Rinehart intended to do so.

As implemented in the United States, the adversarial system is a significant cause of wrongful convictions, wrongful acquittals, and “wrongful” sentences. Empirical evidence suggests that a hybrid inquisitorial regime can reduce these erroneous results. This Article proposes that the American trial process incorporate three inquisitorial mechanisms—judicial control over the adjudication process, nonadversarial treatment of experts, and required unsworn testimony by the defendant—and defends the proposal against constitutional and practical challenges. While other scholars have suggested borrowing from overseas, these three proposals have yet to be presented as a package. Together they might measurably enhance the accuracy of the American criminal justice system.

The enormous value of Dan Simon’s In Doubt lies not just in its nuanced exploration of the challenges to accurate criminal factfinding, but also in its challenge to us to rethink trials themselves. Even as we endeavor to give criminal defendants the means and license to raise reasonable doubts, we need to think more about when and how those doubts can be allayed. Just because most jurisdictions have not come out of the first round of play—the one in which defendants get the tools to poke holes in the cases against them—does not mean it is premature to consider what should happen in the second period: What tools should we give jurors to assess the alleged holes—the “reasonableness” of an alleged doubt? And how can the prosecution try to mend them? These questions do not simply go to the fairness and, to use Simon’s term, the “diagnosticity of the trial.” They also, as I hope to show here, go to the role that criminal trials will play in a world with so few of them.

The shortcomings of forensic evidence in the criminal justice system are now well known. But most scholarly attention has concentrated on “first-generation” forensic techniques such as hair or pattern analysis, bite marks, firearms, and ballistics. Moreover, most of the attention has centered on the investigative process, specifically the collection and analysis of evidence. This Essay turns the critical lens on scientific evidence in a different direction. It focuses on “second-generation” technologies—such as location tracking, biometrics, digital forensics, and other database-driven techniques, and it scrutinizes the adjudicative system—the “bail to jail” stream—rather than the investigative process. Ultimately, this Essay argues that almost every aspect of the adversarial process, as currently conceived, is ill-suited to ensuring the integrity of high-tech evidence. Specifically, the adversarial model demands individualized rather than collective inquiries, embraces secrecy rather than transparency, and privileges viva voce evidence over other forms of fact-gathering. Furthermore, it heavily depends upon the skill of counsel and in-court confrontation rather than out-of-court oversight and structural reform to address problems related to evidentiary integrity, and adopts rigid rules of finality grounded in part on an assumption that proof is always inconclusive. This Essay concludes that the eighteenth-century model of justice may be ill-suited to twenty-first-century evidence, and offers recommendations for a more reliable factfinding system.

The story has become all too familiar. Facing felony charges, an indigent defendant is appointed a public defender to represent him. Working under a crushing caseload of 500 felonies per year, the public defender has around an hour to dispose of the case. He meets with the defendant once, quickly advises him to accept the plea deal offered by the prosecutor, and moves on to the next defendant. No investigation; no interviewing potential witnesses—just take the plea.

The case described above was appealed to the Washington State Supreme Court in 2010, and involved the prosecution of a boy who was a mere twelve-years old. The boy’s case, combined with pressure from state defense organizations and local media, convinced the state supreme court to take action in 2012 and enact mandatory performance standards, including strict caseload limits, that all defense attorneys in Washington must comply with beginning September 2013. The standards were adopted after a long struggle with underfunding and overwhelming caseloads among Washington public defender offices, a problem not unique to that state.

Federal and state law confers broad discretion on courts to administer the criminal laws, impose powerful penalties, and leave serious criminal behavior unpunished. Each time an appellate court reviews a criminal verdict, it performs an important systemic function of regulating the exercise of that power. Trial courts do the same when, for example, they admit or exclude evidence generated by government investigators. For decades, judicial decisions of this sort have been guided by case law made during the Supreme Court’s Criminal Procedure Revolution of the 1960s and 1970s. It is becoming increasingly clear, however, that the rule-bound, essentially bureaucratic regulatory regime that emerged in the 1960s does not assure accurate outcomes, especially for poor and minority criminal defendants and victims. As an additional protection, this Essay urges criminal courts to stop resisting and to embrace regulatory innovations in wide use in other domains that foster self-improvement through continuous observation and rigorous analysis of data on the system’s own results and error rates. The Supreme Court’s own penalty-default and head-counting innovations point the way forward. A certain conception of liberty or individualism—which intersected with the Criminal Procedure Revolution but has not traditionally served the interests of populations most at risk from miscarriages of criminal justice—is the main obstacle to such reforms and should be abandoned.