The Defendant’s Burden: the Onus Probandi and the
Anomaly of Affirmative Defenses
National Autonomous University of México
In any sophisticated system of inquiry, we expect to find several features intimately tied to questions of error and error avoidance. Above all, we want a system a). that will produce relatively few erroneous beliefs (without resort to the skeptical gimmick of avoiding error by refusing to believe anything) and b). that, when it does make mistakes, will commit errors that tend to be of the less egregious sort than of the more egregious sort (supposing that we can identify some error types as more serious than others). Finally: (c). we want to have mechanisms in place with the capacity to eventually identify the errors that we have made and to tell us how to correct for them. In short, we want to be able to reduce errors, to distribute
those that do occur according to our preferences, and to have a self-correction device for identifying and revising our erroneous beliefs. This is, of course, an unabashedly Peircean view of
the nature of inquiry, though one needn’t be (as I confess to being) a card-carrying pragmatist
to find it congenial. Most of the papers collected here will deal with the problem of error as it arises in the context of scientific research. That is fair enough since most of us are philosophers of science. In my remarks, I want to look instead at some of the ways in which the epistemology of error intrudes into legal fact-finding. Like science, the law is an institutionalized, collective effort to find out the truth about matters of high interest to human beings. Unlike science, however, the law has been at the outer periphery of the field of vision thof epistemologists in the 20 century. I’d like to bringer it closer to the center of focus since I
believe it raises genuine epistemological issues not posed by scientific research. If, like me, your general notion of criminal justice was shaped principally by Hollywood movies and what you learned in high school civics, then you probably suppose that, at least in theory, 1). the defendant in a criminal trial is presumed innocent until proven guilty, 2), the prosecutor alone carries the full burden of proof, and 3), a conviction requires proof beyond a reasonable doubt that the defendant is guilty.
Even if you’ve done jury service, you might still believe these things since juries are routinely instructed that such are the ways of the legal world. Consider three examples of Model Federal Jury Instructions about the burden of proof:
As a result of the defendant's plea of not guilty the burden is on
the prosecution to prove guilt beyond a reasonable doubt. This
burden never shifts to a defendant for the simple reason that
the law never imposes upon a defendant in a criminal case the
burden or duty of calling any witness or producing any 1evidence.
It is the prosecution's burden to prove the defendant guilty
beyond a reasonable doubt. That burden remains with the
prosecution throughout the entire trial and never shifts to the
defendant. The defendant is never required to prove that he is 2innocent.
… the defendant in a criminal case never has any duty or
obligation to testify or come forward with any evidence. This is
because the burden of proof beyond a reasonable doubt
remains on the government at all times, and the defendant is 3presumed innocent.
These statements would appear to be unambiguous: ―the burden never shifts to a defendant,‖ ―the defendant is never required to prove that he is innocent,‖ and the defendant ―never has any duty … [to] come forward with any evidence.‖ As I have learned painfully over
the last few years, ―never‖ almost never means never when it comes to the law.
In this essay, I want to look at a class of situations where our untutored hunches about how proof functions in the law are stretched to, even beyond, the breaking point. What we will come to see is that defendants are frequently asked to come forward with evidence, that they are often obliged to prove their innocence, and that the burden of proof is as apt to fall on the defendant as it is on the prosecutor. More than that, we will see that there are various epistemic mechanisms in the law --especially those associated with the notion of the burden of proof—that find no obvious counterparts in other, more familiar, activities where the truth is sought, such as the natural sciences. This is one reason why I think that legal epistemology, as I call it, ought to be of much broader general interest to epistemologists than it is usually thought to be.
The class of cases where these surprises routinely occur forms what are usually known as affirmative defenses (hereafter ADs). Put succinctly, an AD arises whenever the defendant, instead of contesting the claim that he harmed someone else, usually concedes that point but insists that his actions were nonetheless legally blameless. ADs, I hasten to add, are not some recherché or esoteric area of the law. Fully one third of all criminals trials alleging violent acts 4turn on an affirmative defense. By focusing on this set of cases, I hope to persuade you that the law poses some epistemological puzzles that are both intriguing in themselves and are quite unlike the epistemic conundrums that those of us who grew up as philosophers of science are accustomed to dealing with.
It will doubtless be helpful if, before moving to the technicalities, I mention some typical examples of affirmative defenses. Since the Congress and every state legislature can invent its own list of defenses that the courts must recognize, there can be no comprehensive list of
1 1-1 Modern Federal Jury Instructions-Criminal P 1.02. 2 1-4 Modern Federal Jury Instructions-Criminal P 4.01. 3 1-5 Modern Federal Jury Instructions-Criminal P 5.07. 4 Kalven & Zeisel, The American Jury, p. 221.
them. Indeed, they vary dramatically from one jurisdiction to another. Here, however, are the more common and familiar defenses:
Defense of others, property or self Provocation
Insanity Mistake of fact
Consent Involuntary intoxication
Impossibility Third-Party Guilt
Diplomatic, executive, or legislative immunity Statute of limitations
Double jeopardy Military orders
Entrapment De minimis infraction
Plea-bargained immunity Incompetency
(I will leave unexplained for now the rationale for my division of defenses into these two classes. Its significance will emerge towards the end of my paper.)
The Two Paradigmatic Forms of Affirmative Defenses
In the list I just gave, the ADs were divided between those defenses that provide a justification for the actions of the accused and those defenses that, while they also get the defendant off the prosecutorial hook, are grounded purely in reasons of public policy and expedience rather than being in any way morally exculpatory. For our purposes, however, there is a more epistemically salient way of dividing up ADs. This bifurcation reflects the fact that different jurisdictions impose quite different sorts of probatory burdens on those who invoke an affirmative defense. Essentially these fall into two classes:
There is, for starters, what I will call the full-blown affirmative defense. It has the following
-the state must prove all elements of the crime beyond a reasonable doubt (hereafter: BARD).
-if defendant asserts an AD, he must prove to a demanding standard (usually the preponderance of the evidence, sometimes more) that his alleged defense is true.
-a failure by the defendant to prove his defense more likely than not (as judged by the jury) obliges the jury to convict him.
5Full-blown affirmative defenses are the rule in at least twelve states, all of which have
statutes requiring the defendant to prove his defense to a preponderance of the evidence and
5 Alabama, Alaska, Delaware, Illinois, Louisiana, Maryland, Ohio, Pennsylvania, Rhode Island, South Carolina, Texas, and West Virginia. In many more states, an insanity defense requires proof by the defendant to at least a preponderance of the evidence.
by several federal circuit courts. (Seven of the twelve federal circuits likewise require full-blown 6ADs for at least some of the defenses enumerated above.) Minor variants on the full-blown
defense can be found in Delaware, Georgia, and North Carolina, where the defendant must prove certain defenses ―to the satisfaction of the jury.‖ In Kentucky, the evidence for an affirmative defense must be ―convincing,‖ whatever that means. Delaware, Georgia, and
Oregon have required defendants to prove an insanity defense BARD, a requirement 7astonishingly upheld repeatedly by the U.S. Supreme Court. However these unconventional
standards are glossed, it is clear that all require of the defendant substantially more than that he raise a reasonable doubt about his guilt.
Secondly, many jurisdictions utilize what I call modestly demanding affirmative defenses. These
occur in about half the state criminal courts and in a slim majority of the federal circuit courts. Here are their principal features:
-the state is obliged to prove all the elements of the crime BARD.
-if the defendant asserts an AD, he is assigned
-a burden of producing sufficient evidence
-to raise a reasonable doubt that it is false that the AD applies to him.
-if the judge is persuaded that the defendant has met these two conditions, he will
-instruct the jury about the existence of the defense in question
-allow the jury to hear evidence relevant to the exculpatory defense and
-inform the jury that the prosecutor must prove BARD that defendant’s excuse is false.
-if the judge is not persuaded that there is a sufficient modicum of evidence for the AD,
-the judge will not allow the defense to raise it with the jury,
-the judge will not instruct the jury that the law considers the AD in question a form of exoneration, and
-the case will proceed before the jury as if the only germane issue is whether the prosecutor has proved the statutory elements of the crime BARD.
The sequence of events here is important to grasp correctly. The defendant claiming a modestly demanding affirmative defense must first persuade the trial judge --out of hearing of the jury-- that he has a plausible case to make. That requires the defendant both to produce
exculpatory evidence supporting his affirmative defense and to convince the judge that this
evidence reaches the stipulated bar of creating a reasonable doubt about the hypothesis that his AD is false. Like the full-blown AD, this lite-version of an affirmative defense imposes a burden on the defendant of satisfying a standard of proof, although the latter standard –raising 8a reasonable doubt—is clearly less onerous than the preponderance of evidence standard.
6 The First, Fourth, Seventh and Ninth Federal Circuits do not require full-blown ADs. 7 See Rivera v. Delaware, 429 U.S. 877 (1976), and Leland v. Oregon, 343 U.S. 790 (1952). Arizona
requires clear and convincing evidence for an insanity defense. 8 In some jurisdictions, it is enough to satisfy the requirement of raising a reasonable doubt if the defendant merely asserts that his action was in conformity with one of the available defenses. In most, however, the bare assertion of a defense is insufficient to put it on the table. There, there must be exculpatory evidence apart from defendant’s plea of an AD. Hereafter,
It may seem strange to speak of a burden of proof at all when the obligation on the defendant is as weak as merely raising a reasonable possibility that the he is innocent. But I am persuaded that there is no better way of describing it. Note that the defendant in these cases is not merely obliged to present relevant, exculpatory evidence. (This is the so-called burden of producing evidence.) That requirement alone would not be tantamount to imposing a burden of proof (although it would certainly be sufficient to give the lie to the claim that the defendant is never obliged to present any evidence). What makes this into a genuine burden of proof is that there is a clear specification of the quantum of proof necessary for the defendant to get his case before the jury: he must present enough evidence to raise a reasonable doubt about his guilt.
Both these ADs are puzzling in rather different ways. What they share in common is an implicit renunciation of the presumption of innocence, and of the related thesis that the prosecutor alone bears the burden of proof in a criminal trial. Those points are too obvious to require elaboration here. Instead, I want to focus on some other epistemic puzzles posed by these two forms of AD. I will begin with the full-blown version, where a defendant asserting a certain defense must persuade the jury that his AD is more likely than not, turning later to explore some of the puzzles associated with the more modest version of the AD. The Problem with Full-Blown ADs: Misunderstanding the Function of a
Standard of Proof
This practice of obliging the defendant to prove his exculpatory defense to be more likely than not poses conceptual problems aplenty. For instance, one might profitably explore how, if at all, such defenses can be squared with the presumption of innocence (since they appear to presume guilt rather than innocence) or how they can be reconciled with the thesis that the burden of proof in a criminal trial never shifts to the defendant. For today’s purposes, I prefer 9to leave those intriguing questions to one side. Instead, I want to focus specifically on an issue
that I think is more salient than either of those. It has to do with the kind of reasoning that lies behind the widespread belief that it is appropriate under certain circumstances to require that the defendant prove it to be more likely than not that his proposed defense or excuse is true. For purposes of illustration, I will focus on one familiar affirmative defense that of self-defense, although my remarks will apply alike to all those affirmative defenses that I earlier labeled as ―justificatory.‖
In a case of self-defense, the law recognizes that if A –having done nothing to provoke B--
is threatened or menaced by B, and if A has reasonable grounds to believe that B is about to do him grievous harm, then A is entitled to take whatever steps a rational person might deem necessary to protect himself, including, if necessary, immobilizing B by killing him. (I don’t
intend to explore the morality of the doctrine of self-defense, although it strikes me as plausible enough.)
What matters for our purposes is that relevant legislative bodies have specified that one is legally blameless if one genuinely acted in self defense. There is, in other words, no criminal
when I refer to modestly demanding ADs, I will be alluding to the latter sort of case. The former one raises none of the problems that I will be discussing. 9 I have explored such questions briefly in my ―The Presumption of Innocence: Material or Probatory?‖ Legal Theory, Fall 2006.
liability associated with such actions. A person who harms another in self defense is as innocent of a crime as if he had done no harm at all. That is the law.
The relevant epistemic question obviously is this: if action in self-defense is innocent 10behavior, then why must the defendant prove it more likely than not that he so acted? If we
believe, as a matter of general policy, that the state must prove one’s guilt in the case of an
alleged crime, why does the defendant have to prove his innocence in the case of self-defense? More specifically, in a usual criminal case (where an AD is not involved), the state must prove defendant’s guilt beyond a reasonable doubt, whereas in a case of self-defense, the defendant
must prove his innocence by a preponderance of the evidence. What we have in play here are two different standards of proof, BARD and PoE, and two different parties carrying the burden, the state and the defendant respectively.
It is widely accepted that proof BARD is an exacting standard. It requires a very impressive proof and it does so for a compelling reason: we regard the mistakes associated with a criminal trial verdict –a false acquittal and a false conviction—as exacting very different costs. As a
society, we have reached a social consensus that false convictions are much more egregious than false acquittals. We have adopted an exacting SoP because we believe that a demanding standard is, other things being equal, more likely to reduce the probability of a false conviction than a less demanding standard would. Put differently: with BARD we expect a low rate of false positives, and are willing to absorb a relatively high rate of false negatives, if necessary, in order to keep false convictions to an acceptable level. By contrast, the preponderance of the evidence standard implicitly but unequivocally denies that one sort of error is more egregious than the other. The Supreme Court has held on numerous occasions that the costs of a false conviction are so serious that no one can be convicted in a criminal trial unless the standard in play is BARD. No prosecutor could successfully argue that a jury was bound to convict a defendant provided that they thought his guilt was more likely than not. Indeed, no prosecutor could successfully argue that the defendant had to prove anything. But, when we turn to trials that involve full-blown affirmative defenses, all these familiar rules change. Where the affirmative defense is concerned, the prosecutor doesn’t have anything to prove, except that the defendant committed an act whose commission the defendant himself concedes. The defendant, by contrast, must prove his innocence to a preponderance of the evidence. Failure to do so means a verdict of guilty.
The thesis I want to argue is simply this: if the rationale for the selection of a standard of proof like BARD is that we believe that standard incorporates a considered social consensus about the respective costs of errors, then the utilization of any other standard of proof for
determining guilt and innocence violates the social contract that alone undergirds, and makes rational, the selection of the accepted standard. That is to say, with certain possible exceptions to be discussed below, the coexistence of rival standards of proof of guilt and innocence in a system of criminal justice system speaks to a profound confusion (on the part of both legislators and judges) about what a standard of proof is and about whence its rationale derives.
10 In general, a defendant asserting self-defense in a murder case must show (to a preponderance) each of the following: a). that he had reason to believe that he was under an imminent threat of death or serious bodily injury, b). that he had not negligently put himself in this situation, c). that he had no legal alternative to avoid the threatened harm, and d). that it was reasonable to believe that his action would avoid the threatened harm.
The argument is a straightforward one. It begins with the uncontroversial observation that the principal function of a standard of proof is to capture our shared social perceptions of the relative costs of the two sorts of mistakes to which criminal trials are subject: to wit, false convictions and false acquittals. Above all else, a standard of proof is a mechanism for distributing the
errors that are likely to occur. A demanding standard of proof carried by the prosecution, such as proof beyond reasonable doubt, is much more apt to produce false acquittals than false convictions (assuming that defendants are as likely to be innocent as guilty). We accept the idea that such a high standard be imposed on the prosecution because we believe, in the classic and graphic metaphor of William Blackstone, that it is better that ten guilty men go free than that one innocent man is condemned to the gallows. Unless we think that the social costs of a false conviction are roughly ten times greater than the costs of a false conviction, then we have no business setting the standard of proof as high as we do (supposing proof beyond a reasonable doubt to be in the neighborhood of 90-95 percent confidence of guilt). If, for instance, we regarded the two sorts of mistakes as roughly equally costly, a preponderance standard would obviously be the appropriate one since it shows no bias towards one sort of error over the other.
It is a given in the burgeoning literature on the logic of the standard of proof that such a standard must be set at a level that reflects the social costs associated respectively with false
acquittals and false convictions. Essentially, the standard of proof is set sufficiently high (or low) to capture our shared social judgments about the respective costs of these errors. In still more rigorous treatments of this topic, the social benefits of true convictions and true acquittals
are likewise factored into the utility calculation. Whether one grounds the standard of proof simply on the respective costs of potential errors or on a more complex expected utility calculation, one conclusion is salient and inevitable: any modification in the criminal standard of proof that moves it significantly away from that point of confluence between the costs and probabilities of respective errors implies a drastic revision of assumptions about the costliness, and thereby the acceptability, of false convictions and false acquittals. Lowering the standard of proof imposed on the prosecution or imposing a significant standard of proof on the
defendant would entail precisely such a revision of the social contract about the relative costs 11of the two sorts of error.
It is not only legal scholars with a penchant for quantification who think of the standard of proof as a mechanism for distributing errors. When the Supreme Court in Winship settled that
reasonable doubt was to be the constitutional standard of proof, the justices had explicitly in mind the idea that acquittals of the guilty were to be strongly preferred over convictions of the 12innocent. Even before reasonable doubt acquired constitutional status, its near universal
11 For a detailed articulation of such arguments, see: Richard Bell, Decision Theory and Due Process,
78 J. Crim. L. & Crim., 557 (1987); Michael DeKay, The Difference between Blackstone-Like Error
Ratios and Probabilistic Standards of Proof, 21 Law & So. Inquiry 95 (1996), and Erik Lillquist,
Recasting Reasonable Doubt: Decision Theory and the Virtues of Variability, 36 U.C. Davis L. Rev. 85
(2002). 12 As Justice Harlan (concurring) insisted in Winship, ―In a criminal case, on the other hand, we
do not view the social disutility of convicting an innocent man as equivalent to the disutility of acquitting someone who is guilty. ― He went on to quote Justice Brennan’s earlier insistence
that ―the requirement of proof beyond a reasonable doubt in a criminal case [is] bottomed on a fundamental value determination of our society that it is far worse to convict an innocent man
occurrence in the common law likewise reflected an unequivocal judgment that, in criminal trials, false acquittals were preferable to false convictions.
The question before us involves asking what this ―social contract‖ implies for affirmative defenses that require the defendant to establish more than a reasonable doubt about his innocence. The answer is clear: If a state requires that the defendant establish a certain defense (say self-defense or consent) to a preponderance of the evidence or even higher level, that state is saying that erroneously convicting someone who genuinely acted in self defense or with the consent of the victim is no more egregious an injustice than acquitting someone who falsely alleges self-defense or consent. If a state says that a defendant claiming insanity must prove his insanity by clear and convincing evidence (let alone beyond a reasonable doubt), that state is saying that erroneously convicting someone who was truly insane is a vastly lesser injustice than
erroneously acquitting someone who was sane when he committed the crime.
These, I will argue, are curious and ultimately incoherent judgments of value. It is bad enough that they fly in the face of the Blackstonian thesis that false acquittals are less costly than false convictions. They add insult to that injury by undermining the presumption of innocence and the prosecutorial burden of proof when they insist that a defendant can win an acquittal only if he can prove his innocence to a relatively high standard.
It might be helpful to consider this hypothetical pair of examples: Jones and Smith are both on trial for first-degree murder. Jones offers an alibi (which is not an affirmative defense), presenting witnesses who claim to have been with him elsewhere at the time of the crime. The judge will instruct the jury that, to win an acquittal, Jones has no burden of proof and the prosecutor must prove beyond a reasonable doubt (among other things) that Jones was at the crime scene. Smith, by contrast, claims to have been acting in self-defense. In many jurisdictions (significantly, not in federal courts), Jones must present enough evidence to make it more likely than not that he was indeed acting in self defense. Unless he does so, his evidence may be excluded or, even if admitted, the judge may elect not to give the jury a self-defense instruction. Jones and Smith, charged with the same crime, obviously enjoy quite different prospects. Jones will be acquitted if he can raise a reasonable doubt about his guilt. Smith, by contrast, is bound to be convicted if all he can do is to raise a reasonable doubt; he must prove it more likely than not that he acted in self defense.
The crucial issue for us involves the message being sent by the justice system with these two cases. In Jones’ trial, the message –implicit in the standard of proof beyond a reasonable
doubt-- is that it is far better to acquit the guilty than to convict the innocent. In Smith’s trial, by contrast, the inescapable message is that convicting the innocent and acquitting the guilty are equally undesirable.
The salient question is a simple one: What is the principled difference between these cases
that would justify such discrepant assessments of the relative costs of errors? Jones, it is true, is denying that he committed an act that the state is obliged to prove, while Smith is conceding that he committed the act but insisting that his behavior was justified and can point to a state statute on self-defense that stipulates self-defense to be a full justification for acting as he did. If Smith’s action was genuinely one of self defense, then he is every bit as innocent of the
crime as Jones is, if his alibi is true. But when the state insists that Smith must prove it to be
than to let a guilty man go free.‖ (In re Winship, 397 US 358, at 369-70 (1970) Similar
sentiments have been voiced by the Supreme Court numerous times since Winship.
more likely than not that he acted in self-defense, it is saying that convicting an innocent Smith would be a much less egregious error than convicting an innocent Jones. This is nonsense. We must hew to the line that convicting a person innocent of a given crime brings the same costs,
independently of the specific attributes that render innocent those who are wrongly convicted. Likewise, acquitting a guilty person arguably generally brings the same costs. To hold that
convicting the innocent is sometimes much worse than acquitting the guilty while other times saying that convicting the innocent is no worse (or perhaps even better) than acquitting the guilty is to fall into babbling incoherence, especially when we are making reference to the same generic crime in the two cases. Before we can assert that conclusion, however, we need to consider briefly the plethora of arguments that have been proffered for putting in place a system that permits very different standards to apply in different criminal trials. (I might add in passing that there is virtually no legal system in the world that does not use such a combination of discrepant standards.)
Spurious Justifications of the Double Standard
Different legal systems offer different reasons for imposing a probatory burden on the defendant who adopts an affirmative defense, even while those systems invariably and explicitly subscribe to the doctrine that, in a criminal trial, the burden of proof falls officially on the prosecution. Here are the more common arguments for this policy:
1). The „Can‟ Implies „Ought‟ Thesis--In the US, it is frequently argued that states are free to set
whatever standard of proof they like for such defenses on the grounds that nothing obliges a state to make available the mechanism of an affirmative defense. [LL: give a quote] It is (just) conceivable that a state might include no defenses at all in its penal code, not even self-defense.
Since the defining and offering of defenses is purely at the option of the state legislature and since such defenses are not governed by Winship’s insistence on proof beyond a reasonable
doubt for the key elements of the crime, some jurists, legislators, and academics --not to mention the US Supreme Court itself—have argued that each state may impose, on either the
prosecution or the defendant (though it will typically be the latter), whatever sort of burden of proof or production it deems appropriate for such defenses. Since, the argument goes, an affirmative defense is a creature of the state legislature --in effect a gratuitous sop to the defendant--, that body is at liberty both to provide it (or not) and, if provided, to specify the quantum of proof required to satisfy it. I will not dispute this assertion insofar as it concerns the thesis that states have constitutional authority to proceed in this fashion. But constitutional
authority is one thing; good reasons are quite another. Our question must be not whether courts and legislators have the legal authority to proceed in this fashion but rather whether any right-minded person would choose to do so. The fact that states have it within their constitutional authority to impose a demanding burden of poof on the defendant in cases involving an affirmative defense patently does not in itself justify the imposition of such a burden of proof. The fact that doing so is arguably constitutional does not mean that it makes for coherent policy. That a state can do this is no reason to believe that it should. 2). The Initiator Thesis – If the defendant makes a positive assertion about the events
surrounding the crime (as opposed to merely denying the prosecutor’s allegations), it is only
natural that he is obliged to establish that hypothesis. Putting the doctrine in its traditional
13guise: ―He who alleges a fact must prove it.‖ This doctrine is, of course, traditional in civil
trials, where we expect the party affirming some fact to carry the burden of presenting enough evidence to make that allegation plausible. The question is whether its transposition into criminal proceedings makes any sense.
The salient issue, of course, is who is doing the alleging and why. Imagine the following
situation: Sally accuses John of rape. Rape is clearly a criminal offense. On the other hand, consensual relations between adults don’t count as rape because both morality and the law
hold that consent annuls culpability. So, a crucial issue in John’s trial will turn on the question
of consent. Suppose that the jurisdiction in question passes a statute saying that the crime of rape occurs when one party has sexual intercourse with another without the latter’s consent. If the law is so constructed, then it will fall to the prosecutor to prove the two elements of the crime, sexual relations and lack of consent. John will not have to prove anything nor present any evidence. Suppose, on the other hand, that the legislature in question defines rape as sexual union and then specifies consent as a legitimate affirmative defense. If that jurisdiction is likewise one in which the defendant is required to prove his AD by a preponderance of the evidence, then John will have some serious work to do if he hopes to win an acquittal. My point is that our decision as to whether the defendant is alleging a fact (as opposed to simply denying an allegation of wrong-doing) is nothing more than an artifact of how the relevant law in question is framed. If rape is defined by statute as nonconsensual sex, then a defendant accused of rape who believes that he had the consent of his alleged victim is not asserting a fact but denying one element of the prosecution’s case. On the other hand, if rape is defined as sexual intercourse by a man with a woman who is not his wife --with a side statute stipulating that consent is an affirmative defense-- then the defendant in a rape case is suddenly alleging an affirmative fact and carries a burden of proof. In other words, it is in principle open to the drafters of criminal legislation to stuff all the justificatory and exculpatory clauses into an affirmative defense statute, thereby burdening the defendant with having to prove what it would otherwise fall to the state to have to disprove: his innocence. Accordingly, the distinction between what the defendant denies and what he alleges becomes hostage to the vicissitudes of how the legislation was drafted. If society’s considered view is that certain circumstances are genuinely exculpatory or justificatory (as consent most surely is in the case of sexual relations), then it should fall to the state to have to prove beyond a reasonable doubt that those circumstances did not obtain in the case in question. The argument that the defendant should have to prove that those circumstances applied because he is alleging that they did is nothing but the result of a cynical legislative shell game obviously designed to undermine the presumption of innocence and to circumvent the state’s obligation to prove the
defendant’s wrongdoing. Dressing up this sham by invoking the principle that he who alleges X must prove X is a disgrace.
3). The thesis of the impossibility of “proving a negative” – For several centuries, jurists have been
enamored of the silly idea that it is impossible to prove a negative. Indeed, for many of them, that is why the burden of proof in a criminal trial usually falls on the prosecution, since if the defendant had to prove that he did not commit the crime, the state would be asking him to do something which is difficult or impossible to do. This is, of course, sheer nonsense, which should never have been allowed to escape from Logic 101. Having to prove a negative, as legal vernacular conceives it, is no different epistemically from having to prove a positive and
13 Reus excipiendo fit actor.