Legal scholarship has long treated substantive criminal law and evidence as two separate and distinct fields. The former largely concerns itself with evaluating substantive criminal law rules by reference to various animating theories—most prominently, those of deterrence and retributivism.1 Scholars, students, and policymakers laud or condemn doctrines based on notions of “just deserts” or ideas about the incentives they create for those disposed to commit a crime. When it comes to the numerous evidentiary and other rules that determine the course of prosecutions and proof, however, the conversation is different. Here, questions of reliability, evidential worth, and accuracy in fact-finding dominate the debate. References to the deeper concerns of deterrence and retributivism, and the significance of various evidentiary and procedural rules toward the program of one or the other, are by and large absent.
Our article “Mediating Rules in Criminal Law”2 challenges this conventional divide between evidence and substantive criminal law theory. Our claim is that the traditional understanding of evidentiary rules in criminal law as geared overwhelmingly to truth in fact-finding is incomplete. Evidentiary rules, we argue, also perform a deeper, systemic function by mediating latent conflicts between criminal law’s deterrence and retributivist objectives. They do this by skewing errors in the application of the substantive law to favor whichever theory has been disfavored by the substantive rule itself. So, for example, if retributivism dominates the substantive law of insanity, special evidentiary rules governing the presentation and proof of that defense might cabin it in a way that responds to deterrence concerns by making it more difficult to invoke successfully. These “mediating rules” of evidence do this, moreover, without undercutting retributivist objectives as significantly as would redrawing the substantive defense itself. How is this so? In the next few pages, we will sketch the outlines of our theory and offer a brief illustration.1. How Evidentiary Rules Mediate
To understand the functioning of mediating rules, one needs to understand the dissimilar relationships of deterrence and retributivism to the different rules that go into any liability and punishment determination. Two fundamental traits of those theories play an especially important role in our analysis: retributivism’s general agnosticism toward the allocation of errors in the substantive law’s application, and deterrence’s overarching concern with expected sanctions.
Take retributivism first. All varieties of retributivism care primarily about one thing: doing justice in the particular case. Retributivism holds that an offender should be punished “because, and only because, [he] deserves it.”3 Retributivism’s case-focused orientation shapes its stance toward the various rules that determine criminal liability and punishment. For the retributivist, substantive and sanctioning rules are—or at least should be—generally more important than evidentiary rules. Substantive and sanctioning rules are the primary determinants of whether any given instance of criminal liability satisfies the “just deserts” requirement. Poorly-crafted, overbroad substantive rules mean that some individuals who in fact do not deserve punishment may be punished anyway; poorly-crafted, underinclusive ones mean that some who in fact do deserve it may not be. Overly harsh or unduly lenient sanctioning rules mean that some wrongdoers will be punished out of proportion to the gravity of their offenses. The retributivist, therefore, will want to craft substantive and sanctioning rules to track the “just deserts” criterion as closely as possible.
When it comes to evidentiary rules, things are more complex. Like substantive rules, evidentiary rules are important to retributivists insofar as they ensure that the truly deserving are held liable and punished while the truly undeserving are not. They are important, in other words, to the extent that they can achieve ultimate factual accuracy of outcomes in individual cases.
Under conditions of uncertainty, however, minimizing adjudicative errors—by which we mean applications of the law to erroneously determined facts—always involves tradeoffs. That is to say, where one can never know with absolute certainty whether an individual did or did not commit a crime, virtually all evidentiary rules designed to increase the factual accuracy of a legal determination of guilt will at the same time decrease the factual accuracy of a legal determination of innocence. Requiring the government to prove guilt “beyond a reasonable doubt,” for example, means that criminal convictions must be as factually accurate as practically feasible. This standard greatly increases the chances that a defendant found guilty will in fact have committed the crime in question and greatly decreases the chances that factually innocent defendants will be convicted. By the same token, though, it also greatly increases the chances that many factually guilty defendants will not be convicted. In short, by decreasing the incidence of false positives (erroneous convictions of the factually innocent), a “reasonable doubt” standard increases the incidence of false negatives (erroneous acquittals and non-prosecutions of the factually guilty). Sticking with standards of proof, one could decrease the incidence of false negatives by, say, lowering the standard to one of preponderance of the evidence. This would result in more convictions of factually guilty defendants. But it would result in more convictions of factually innocent defendants as well. False negatives would go down, but false positives would go up.
This aspect of fact-finding in the face of uncertainty about factual guilt leaves retributivism with little to say about the precise evidentiary rules that should govern the imposition of liability and punishment. For retributivists, both false positives and false negatives violate just deserts. Both therefore are instances of injustice in the individual case. Among the various combinations of rules that promote accurate fact-finding and reduce the risk of error in adjudication, retributive theory expresses no strong preferences for whether such rules skew more toward the side of false positives or more toward false negatives.4 Tinkering with evidentiary rules in a way that decreases one at the expense of the other thus does not much matter to retributivism. Any rational system aiming at the pursuit of truth should be acceptable.
Now take deterrence. In contrast to retributivism, deterrence is system-focused. It sees the purpose of criminal liability and punishment as averting social harm by imposing costs on undesirable conduct. As such, deterrence cares only about the incentives that the criminal law creates in the form of expected penalties. This means that the scope of a given criminal prohibition or the sanctions actually imposed for a given crime need not be exactly right. The punishment may be high or low, and the crime definition may be overbroad or too narrow—in either scenario, deterrence can still achieve the “right” expected penalty through evidentiary and procedural rules that raise or lower the barriers to conviction. An evidentiary rule that makes it more difficult to obtain a conviction—such as a stringent “beyond all reasonable doubt” standard—erodes the expected penalty by causing an offender to discount the likelihood of actually being punished. A rule that makes it easier to convict—such as one requiring a defendant to prove any affirmative defense by a preponderance of the evidence—does the opposite. False positives and false negatives thus matter to deterrence because they affect the expected penalty by affecting the probability that the penalty will be imposed.
The asymmetrical orientations of deterrence and retributivism with respect to evidentiary rules provide a degree of flexibility in the crafting of criminal law doctrine that commentators have overlooked. Given that these theories react to those rules in different ways, the combinations of rules that go into any liability and punishment determination can be adjusted to perform a mediating function between deterrence and retributivist objectives. Retributivism, for instance, might mandate a crime definition (a substantive rule) that is somewhat too narrow and a punishment (a sanctioning rule) that is somewhat too lenient from a deterrence perspective. Deterrence objectives could nevertheless be accommodated in such a case by relaxing the normal evidentiary requirements governing conviction in a way that pushes the expected penalty back up toward the acceptable range for deterrence. So long as those requirements do so without undermining rational fact-finding, retributivists should not object to them.2. An Example from the Law of Defenses
We see this dynamic in the varying standards of proof commonly applied to the two most important categories of criminal law defenses: justifications and excuses. The basic substantive distinction between the two types of defenses is straightforward. Justifications are conduct-focused defenses. They mark out circumstances under which conduct that would otherwise be criminal is socially acceptable and deserving of neither criminal liability nor censure, usually because the conduct prevents even greater harm or furthers a greater interest than that sought to be protected by the criminal prohibition. Excuses, by contrast, are actor-focused defenses. Excuse defenses mark out scenarios in which, although the conduct at issue is still socially unacceptable and deserving of censure, some peculiar characteristic of the actor diminishes or even eliminates his responsibility and blameworthiness. While the conduct remains wrong and something to be discouraged, the criminal law deems the actor an inappropriate candidate for punishment.
Just as important for our purposes is a basic procedural distinction governing the use of these two categories of defenses. In most jurisdictions, core justification defenses—such as self-defense and necessity—once raised by a defendant, must be disproved by the prosecution beyond all reasonable doubt. By contrast, also in most jurisdictions, core excuse defenses—such as insanity, duress, and provocation—must be both raised and subsequently proved by the defendant by a preponderance of the evidence (and in some instances by an even higher standard).
This divergence is puzzling. It cannot be explained by the dictates of constitutional criminal procedure, which generally leaves jurisdictions free to treat all affirmative defenses—justifications and excuses alike—more or less the same. Nor is it satisfying to understand these different rules as a function of the degree to which a given defense turns on subjective elements that would be difficult for prosecutors to prove by reference to external, objective facts, as some have suggested. The more subjective the defense, these arguments go, the more sense it makes to place the burden on the defendant as the party with the best knowledge of and access to the relevant evidence.5 But these arguments do a poor job of accounting for positive law. Most core excuse defenses like duress and provocation stand or fall on a reasonableness inquiry that makes them primarily objective in nature.6 To the extent that the reasonableness inquiry takes into account the subjective situation of the defendant, as it does in many jurisdictions, it does so for justification defenses as well.
A more robust understanding of the doctrine starts with the importance of the conceptual differences between justification and excuse to deterrence and retributivism. It then considers the ways in which shifting the burden of proof helps to mediate between the demands of those two theories. Conflict between deterrence and retributivism in the creation and definition of justification defenses is minimal. Most retributivists, for instance, would believe that there is nothing blameworthy in responding to an unprovoked attack with proportional force in order to protect oneself, or in breaking down the door of an empty cabin in the woods in order to keep from starving. Likewise, deterrence supports such defenses because, properly cabined, they increase social welfare (or at least do not reduce it). Self-defense (and defense of others), for example, deters would-be aggressors much more cheaply than would an “always retreat and call the police” rule. Necessity, for its part, expressly conditions the availability of a defense on taking an action that minimizes social harm and, in many cases, disallows the defense if the actor could easily have avoided the dilemma to begin with.
Basic agreement between deterrence and retributivism regarding the impropriety of punishing justified conduct means that little need exists to mediate between the two theories through resort to special rules governing evidence presentment and proof. This is not so with excuse defenses. Here the divergence between the two theories is sharp. Deterrence judges excuses by the same normative criterion it applies to justifications: their effect on social welfare. For deterrence, recognizing an excuse defense makes sense only when it removes a chilling effect from an activity that society has no reason to chill—only, that is, when it creates the proper incentives for future action. As always, retributivism rejects this test in favor of an individualized and ex post approach. The recognition of an excuse is appropriate for the retributivist whenever there is something peculiar to the actor and his situation—a mental disease or defect that distorts his senses or a coercive threat of great bodily harm—that vitiates his blameworthiness and renders punishment unjust in the particular case. To the retributivist, excuses are concessions to human frailty that a liberal society gives to individuals whose conduct it still deplores.
The problem for deterrence is that while frailty may not be blameworthy, it is not something to be encouraged, either. And frailty leading to social harm is something to be affirmatively discouraged. Thus, whereas retributivism sees the provocation defense as appropriately calibrating punishment to the reduced blameworthiness surrounding “heat of passion” killings, deterrence sees it as eroding the incentives to keep one’s passions in check in the heat of the moment and to avoid situations in which one may lose control to begin with. Retributivism sees the insanity defense as rightly acknowledging the absence of meaningful culpability on the part of one who is unable to appreciate the criminality of his actions. Deterrence, meanwhile, sees it as creating incentives to fakery, undermining the seriousness of the crimes at issue, and emboldening potential criminals at large. A retributively-driven doctrine of defenses, in short, would recognize a broad range of generously-defined excuses; a deterrence-driven doctrine would be loath to recognize any.
The fact that positive law recognizes many excuse defenses illustrates the degree to which retributivist considerations have dominated this area. It also provides a more illuminating explanation for the divergent burdens of proof that jurisdictions commonly apply to justification and excuse defenses. It is not so much that the special proof rules accompanying excuse defenses enhance accuracy and efficiency in adjudication by placing the burden of persuasion on the party with the best access to information. Rather, they fulfill a quite different, mediating function: they give ground back to deterrence objectives where those objectives have lost out to retributivist considerations at the level of the substantive law. By altering the default rules to decrease the probability that defendants will escape liability by virtue of excuse defenses, these special evidentiary rules increase the expected penalty in all cases in which excuse doctrines might come into play.
This special proof regime accommodates deterrence considerations much better than would the general “proof beyond a reasonable doubt” framework. It does so, moreover, in a way that is more palatable to retributivism than would be tinkering with excuses at the substantive definitional level. Any substantive redefinition that would satisfy deterrence necessarily would come at the expense of retributivism; removing the defense from those whose circumstances mitigate their culpability would result in many instances of unjust punishment. Shifting to a preponderance standard under which defendants must carry the burden largely avoids this problem because it enhances deterrence not through changing the substantive scope of excuses but instead by altering the balance of false positives and false negatives in a way that retributivism can accept. Under a preponderance standard, more excuses are erroneously denied, but fewer are erroneously granted. Most important for deterrence, the absolute number of excuses granted goes down.7* * *
We have referred to mediating rules as special rules of evidence. What makes them special is that, unlike most garden-variety evidentiary rules (e.g., general rules of hearsay, character, opinion, etc.), these rules are not trans-substantive. Each of them is tied to and directs the application of a specific criminal law doctrine in an area in which the substantive law generates especially acute conflicts between deterrence and retributivist goals. Traditional understandings of evidence often seek to explain and justify such special rules by reference to the usual concerns over accuracy in fact-finding. Our theory adds to this understanding by providing a novel interpretive account of how these rules work to strike a rough compromise between criminal law’s competing visions of deterrence, on the one hand, and retributivism, on the other.
Many special rules of evidence thus might be seen as previously unrecognized members of the set of devices that criminal law uses to negotiate difficult tradeoffs between conflicting commitments—in the same family with prosecutorial discretion at charging, judicial discretion at sentencing, and jurors’ discretion in rendering verdicts. Unlike such devices, however, special rules of evidence entrench this process by creating discrete legal spaces that accommodate pluralistic goals not embodied in the substantive law itself. Our positive account, we note, raises a host of historical and motivational questions about mediating rules’ origins and evolution. It also raises fascinating questions about the normative implications of such rules: Are mediating rules a virtue or a vice? Are they something to be rooted out and eliminated, or something to be quietly tolerated or even encouraged? Our tentative view on this latter issue is that the compromise effect of mediating rules is generally a virtue for a liberal criminal law in which we are unlikely ever to have decisive agreement in favor of a deterrent or retributive program. That said, we leave to future research more definitive answers to these questions, uninhibited by artificial separations between evidentiary and substantive rules. The interactions between evidence and substantive criminal law are richer than both evidence and criminal law scholars take them to be.
Preferred citation: Richard A. Bierschbach & Alex Stein, Deterrence, Retributivism, and the Law of Evidence, 93 Va. L. Rev. In Brief 189 (2007), http://www.virginialawreview.org/inbrief/2007/09/24/bierschbach.pdf.
[Back to this issue]