In their provocative Article “Mediating Rules in Criminal Law,” Professors Richard Bierschbach and Alex Stein tell us that rules of criminal liability and rules of evidence, often pushing in opposing directions, collectively determine the quantum and mix of deterrence and retribution that a given punishment practice delivers.1 An example could be the excuse defense of duress. This doctrine is designed to prevent punishment of some actors who are not morally at fault for their behavior (or at least comparatively less at fault than others) because forces not of their own making (and about which we have some sympathy) drive them to commit socially harmful acts. The doctrine might undermine deterrence by granting actors permission, past a point, to abandon efforts to resist offending. However, the evidentiary rule in most jurisdictions that requires the defendant to carry the burden of proving an excuse defense tilts matters somewhat back towards deterrence. Making the defendant prove duress filters out many of the weaker (and even fabricated) duress claims, which might arise in cases in which a defendant was perfectly capable of resisting pressure to violate. Bierschbach and Stein urge us to analyze many areas of criminal law doctrine with this methodology.
Their effort is an intriguing forward move in criminal law scholarship that seeks to remove barriers—perhaps of use to lawyers and law teachers—that have impeded full understanding of the system of social regulation and practices that produces punishment. In a variety of styles and methodologies, scholars have been opening our eyes to how classification-driven distinctions between substantive criminal law and criminal procedure, liability rules and evidentiary rules, law and social norms, and law and politics block us from clearly viewing the social practice of punishment.2
As I worked with Bierschbach and Stein’s methodology, however, I became convinced that it is flawed. The approach requires us to force components of doctrine onto the authors’ deterrence-retribution ledger, yet, we can see as we are doing this that many of those components do not belong there. By the time we are done, we cannot escape the conclusion that the ledger itself is too wooden and simple to account accurately for the theoretical structure of a given locus of criminal law doctrines.
The only option this conclusion leaves is to see Bierschbach and Stein’s analytical framework as an “as-if” tool for the criminal law: measure and examine a doctrine’s effects with a binary deterrence-retribution scale, but disregard a doctrine’s purposes or other effects. Such a deliberately artificial stance could only be useful for the normative implications that the authors say are subsidiary to the positive aims of their project. But the normative value of this posture appears doubtful. Their description of a struggle over public policy between advocates of deterrence and retribution is not likely to be accurate if their binary framework for punishment’s justifications proves to be overly simplified. If the framework of Bierschbach and Stein’s thesis fits neither doctrine nor politics, it is not likely to meet their aspirations of changing scholarly analysis or redirecting public debate.
I will illustrate my disquiet with Bierschbach and Stein’s positive account by discussing two areas of doctrine that they offer to support their thesis. I will then explain why I think their analysis, even if considered solely in terms of doctrine’s surface and effects, does not seem promising as a means of advancing normative debate.
Consider their account of perjury law.3 As a normative matter, most would agree that all lies, even all lies under oath, are not equally condemnable. Bierschbach and Stein sensibly assert that retributive objectives would counsel punishing only serious lies that pervert legal processes. Perjury statutes punish much more: essentially any intentionally uttered falsehood that is even hypothetically material. Many jurisdictions, however, narrow the reach of this rule by requiring the prosecutor to supply corroborating evidence in perjury cases. Bierschbach and Stein see this as an example of a retributive countermeasure in evidence law to a deterrence measure in the substantive law of liability. They say the corroboration rule cuts back the class of cases in which liability will be imposed to something more like those with serious lies.4
This account ignores that legislatures and courts have created and interpreted corroboration requirements in perjury law for other reasons entirely. These requirements serve two purposes, both utilitarian: first, they guard against erroneous convictions in cases that often turn on elusive matters such as the defendant’s knowledge, intent, and memory while testifying, and involve easily fabricated claims by accusers of the A-said-B-said variety; and second, they reduce the risk of overdeterring voluntary testimony, which is a socially valuable activity.5
It might turn out that perjury cases that fail for lack of sufficient corroboration just tend to be perjury cases that involve less blameworthy lies. But we do not know this empirically and probably could not easily find it out, since most of the cases that fail for insufficient corroboration are cases that prosecutors decline to charge. The more significant problem is that Bierschbach and Stein’s analytical framework is artificial. In considering law that counteracts a liability rule, they limit themselves to the law of evidence. I do not see why their claim about proper perspective on conduct rules should not extend to examining all applicable decision rules, as even some of Bierschbach and Stein’s doctrinal examples suggest.6 Once we take this step, however, we see that the argument that we should confine our perspective to the surface and effects of doctrine, while overlooking its development and purposes, proves too much. If a liability rule looks like it administers a strong dose of deterrence, then any rule that narrows liability can be deemed to have “retributive” countereffects. It could as easily be said that the right to unfettered cross-examination makes perjury law “less retributively overbroad” by increasing the likelihood that a defendant will persuade a jury to side with the defendant in a battle of A-said-B-said.
Now consider how the thesis works when applied to accomplice liability rules, another of their doctrinal areas.7 To deter joint efforts to break the law, accomplice liability rules cast a very wide net (most any form of assistance triggers full liability for the offense assisted) and are very punitive (in most jurisdictions, the punishment for helping is the same as for doing). Bierschbach and Stein identify a number of procedural and evidentiary doctrines as retributivist countermoves to these expansive liability rules.
As with their account of perjury law, the countermoves they describe are “retributive” only in the sense of lowering the probability that liability will be imposed under an otherwise highly deterrent rule. They discuss rules that make it difficult for the state to force accomplices to testify, require redaction of self-incriminating statements to protect co-defendants, and dictate the severance of co-defendants in some trials involving such statements. These procedural doctrines are rooted in theories of self-incrimination protections, none of which have to do with countering broad liability rules.8 They further discuss rules in many jurisdictions that require corroboration of accomplice testimony. Like the corroboration requirement in perjury law, these rules are justified by goals of reliability and accuracy, not by a goal of limiting the scope of accomplice liability. Bierschbach and Stein talk about the Supreme Court recently making it harder to admit “testimonial” hearsay under the Sixth Amendment (which, in some circumstances, may include statements of accomplices).9 But in that context the Court was working out the fundamental and inviolable components of the trial right, not worrying about the scope of accessorial liability.
Bierschbach and Stein’s idea that some of these limitations, whatever their origins, supply accomplices with bargaining chips in negotiating plea agreements, thereby reducing the force of accessorial liability rules, is empirically doubtful. Those who represent accomplices—in dealings with the federal government at least—would attest that leverage comes entirely from information and (lawyers would lament) is virtually nonexistent in the law.
Some of Bierschbach and Stein’s examples do not comfortably fit their theory because the theory itself does not fit with the punishment system it seeks to explain. Deterrence is one objective that might be pursued in a consequentialist program for punishment. But there are others, of course. These include minimizing the social costs of punishment—including the costs, of all types, that can follow from mistaken punishment of persons who did not commit the act in question, as well as from punishment of persons who committed the act but who are not blameworthy. Consequentialist aims also include exploiting the communicative and educative force of punishment practices.
Retribution is “purely” an objective probably only for a small subset of theorists who believe that punishment exists so that society can impose desert on those who are morally at fault, full stop. People are much more apt to be “negative” retributivists. Such persons believe that the criminal justice system exists to improve society (or at least to prevent it from getting worse) and that considerations of fault act as a moral constraint on the state’s ability to accomplish social objectives through punishment. Put another way: punish if it is useful and worth the cost, but not if it would undeservedly harm a person.
One of Bierschbach and Stein’s intriguing assertions is that there is an “asymmetry” between the deterrent and retributive effects of evidence rules, because a deterrence advocate cares about the distribution of errors between false convictions and false acquittals while an advocate of retributivism simply wishes that the sum of errors in both directions be minimized.10 Bierschbach and Stein (candidly stating that they choose not to explore the complexities of retributivism) do not deal with the claim that wrongful state imposition of punishment is a greater evil than an offender’s wrongful avoidance of punishment—likely the claim of many persons concerned principally with individual moral fault. If a retributivist does care about the distribution of errors, then she will not find asymmetry in evidence rules to be welcome. Holding the total error rate constant, she would prefer a rule that generates more false negatives than false positives to one generating more false positives than false negatives.
By eschewing (quite purposely, they say) historical and etiological questions about the mechanisms of their “mediating rules,” Bierschbach and Stein fall prey to the simplicity of their own thesis. The truth is that a given locus of conduct rules and decision rules in the criminal law should be seen not as engaging in a tug-of-war, but as pushing in a variety of directions in service of numerous social objectives, many of which are not at all in conflict with each other.
Bierschbach and Stein might be saying something else that is not affected by my criticism. They might be urging us to disregard intentionally stories that lie below the surface of the rules. Look at a criminal law doctrine, they might instruct, and just score it between its discouragement of prohibited acts and its limitations on punishment of persons who are not blameworthy. In doing so, consider only inputs and outputs. What did the people who went into the system do? Among those, which ones were punished at the end?
I can see only one purpose in assuming this posture. It cannot be of positive value if—as I have shown and would expect to continue to find if I kept at it—doing so requires describing many aspects of the law as things they are not. It might be of normative value, however, if (as Bierschbach and Stein suggest in their brief concluding discussion) it would accurately tell two camps in a debate how much policy currency each was receiving from its government. Better, this binary accounting exercise might move such debate forward by convincing participants that they were getting more than they had thought.
This benefit could only be realized if Bierschbach and Stein’s binary deterrence-retribution scale, whatever its deficits in describing criminal law doctrine, nonetheless accurately described the policy debate. But if we end up believing that the framework does not capture the complexity of justifications for doctrine, why would we think it could capture a public debate about those same justifications? Maybe there is some very significant acoustic separation between criminal law theory and public beliefs about the criminal law (like the one I understand Bierschbach and Stein to see between liability rules and allegedly quiet, subterranean evidentiary rules). Interesting hypothesis. Exploring this would have required diving into the depths of politics in much the same way Bierschbach and Stein were unwilling to dive into the depths of doctrine.
Preferred citation: Samuel W. Buell, Purposes and Effects in Criminal Law, 93 Va. L. Rev. In Brief 215 (2007), http://www.virginialawreview.org/inbrief/2007/10/15/buell.pdf.
[Back to this issue]