Rethinking the Presumption of Mens Rea

By: Eric A. Johnson*

Introduction

Substantive background principles play a critical role in the courts’ interpretation of criminal statutes, particularly where the subject of mens rea is concerned.[1]  As Professor Dan Kahan has said, “criminal statutes typically emerge from the legislature only half-formed.”[2]  The effect of these “incompletely specified criminal statutes” is a tacit delegation of lawmaking authority from the legislature to the courts.[3]  A delegation of this sort occurs, for example, in connection with the question of causation.  By enacting statutes that require causation but leave the required causal relationship undefined, legislatures effectively have “left to judicial development”[4] the meaning of the statutory causation requirement.[5]  A similar, though more complex, delegation occurs in connection with mens rea.  Legislatures routinely fail to identify the culpable mental states associated with particular objective elements of crimes.[6]  And so the task of deciding what mental states, if any, to assign these elements falls to the courts.  To guide their exercise of this delegated power, courts have developed a rich—if somewhat untidy—body of substantive background principles.[7]

The most important of these substantive background principles is the presumption of mens rea—or the “mens rea principle,” as it sometimes is known.[8]  The origins of this principle are usually traced to Morissette v. United States,[9] where the Supreme Court famously said:

The contention that an injury can amount to a crime only when inflicted by intention is no provincial or transient notion.  It is as universal and persistent in mature systems of law as belief in freedom of the human will and a consequent ability and duty of the normal individual to choose between good and evil.[10]

In service of this universal notion, the Court read a requirement of intent into the federal conversion statute under which Morissette had been prosecuted.[11]  More broadly, the Court recognized a general presumption that every criminal statute requires proof of “some mental element.”[12]  This presumption, the Court said, could be overcome only by a “clear expression” of legislative intent to impose liability without fault.[13]

Nowadays, the mens rea question is more complicated than whether a crime requires proof of just “some mental element.”  Though “[t]he common law and older codes often defined an offense to require only a single mental state,”[14] the publication of the Model Penal Code in 1962 led to “a general rethinking of traditionalmens-rea analysis.”[15]  Among the components of this rethinking was a recognition that the question of mens rea must “be faced separately with respect to each material element of the crime.”[16]  In other words, the Model Penal Code showed that the question whether to require proof of “some mental element” must be addressed not in relation to the crime as a whole but rather in relation to each individual objective element of the crime.  And so it also showed that the mens rea principle must operate, somehow, at the level of individual material elements.

Unfortunately, nobody seems to know which material elements are subject to the mens rea presumption.  Students in the traditional first-year Criminal Law course learn two very different versions of the presumption.  The first is the Model Penal Code version, which requires proof of some mental state—purposely, knowingly, recklessly, or negligently—with respect to every material element of the offense,[17] unless the offense is a mere “violation.”[18]  The second is the judge-made version, which requires proof of some mental state only with respect to those “statutory elements that criminalize otherwise innocent conduct.”[19]  Justice (then Judge) Sotomayor precisely, if somewhat awkwardly, summarized this judge-made version of the presumption in her very first opinion as a judge of the Second Circuit.[20]  “Absent clear congressional intent to the contrary,” she said, “statutes defining federal crimes are . . . normally read to contain a mens rea requirement that attaches to enough elements of the crime that together would be sufficient to constitute an act in violation of the law.”[21]

Neither the Model Penal Code’s nor the courts’ version of the mens rea presumption is entirely right.  In Part I, I will argue that the Code’s drafters were wrong in assuming that elements designed to measure the harm from an offense invariably require the assignment of a mental state.  In Part II, I will argue—drawing on a recent dissenting opinion by Justice Stevens—that the courts are wrong in assuming that elements designed to do something other than measure the harm often do notrequire mental states.  In Part III, I will use these two criticisms—of the Model Penal Code and of the courts—as the basis for constructing an alternative version of the mens rea presumption, in which the mens rea presumption is reconceptualized as a kind of actus reus presumption.  Finally, in Part IV, I will show that this alternative version of the mens rea presumption is consistent with what the courts say about the confusing topic of general and specific intent.

I.  The Shortcomings of the Model Penal Code Approach

The Model Penal Code’s version of the mens rea requirement appears in section 2.02, which provides that “a person is not guilty of an offense unless he acted purposely, knowingly, recklessly or negligently, as the law may require, with respect to each material element of the offense.”[22]  This rule would require the courts to assign some mental state to every objective element of every offense—even elements whose function is to distinguish between more and less serious versions of the same offense.  As applied to the crime of aggravated theft, for example, this rule would require the courts to assign some mental state—”recklessly,” perhaps—even to the value of the stolen property.[23]  Thus, a defendant charged with aggravated theft could defend the case by asserting that he had not realized that the stolen property’s value might exceed the statutory threshold.[24]

This expansive version of the mens rea presumption undoubtedly is based in part on the uncontroversial proposition that culpability is a matter of degree.[25]  There is a difference, of course, between a thief who hopes or expects to obtain property valued at a million dollars and a thief who hopes or expects to obtain property valued at five dollars.  And one reason for differentiating aggravated from simple theft is to take this difference into account.  But the expansive version of the mens rea presumption also appears to be based on two more controversial assumptions, neither of which has won a broad following among courts.

The first of these two assumptions is that harm has no independent bearing on a crime’s gravity and, accordingly, that the harm elements in criminal statutes really function only as markers—to tell the jurors in relation to what harm they are to measure the unjustifiability of risk and the culpability of the actor’s conduct.  From this assumption that harm elements are present in criminal statutes only as markers, it appears to follow that harm elements can fulfill their statutory function only if they are assigned mental states.  The commentaries to the Model Penal Code make roughly this point in connection with the aggravated-theft example: “The amount involved in a theft has criminological significance only if it corresponds with what the thief expected or hoped to get.  To punish on the basis of actual harm rather than on the basis of foreseen or desired harm is to measure the extent of criminality by fortuity.”[26]

This first apparent assumption—that harm has no bearing on the crime’s gravity except as a marker—is belied even by the Model Penal Code’s own special part.  Under the Code, a person who “recklessly engages in conduct which places or may place another person in danger of death” is guilty only of a misdemeanor—reckless endangerment—if the risk of death is not realized.[27]  On the other hand, if the risk of death is realized, the defendant is guilty of reckless manslaughter, a felony punishable by up to ten years in prison.[28]  The risk required by these two crimes is exactly the same, as is the degree of culpability.  What distinguishes the two crimes is just the harm.  In this setting, then, the Code obviously assigns independent significance to the degree of harm inflicted by the crime.[29]

State legislatures, too, have assigned independent significance to harm in a wide array of criminal statutes.  Consider, for example, two Iowa statutes on the subject of drag racing.  The first, section 321.278 of the Iowa Code, defines “drag racing” as a “motor vehicle speed contest . . . on any street or highway” and classifies it as a simple misdemeanor.[30]  The second, section 707.6A of the Iowa Code, provides that “[a] person commits a class ‘D’ felony when the person unintentionally causes the death of another while drag racing, in violation of section 321.278.”[31]  The second of these statutes appears to require nothing by way of risk or culpability that is not required by the first.  It says nothing about any requirement of recklessness or negligence with respect to the death, for example.[32]  And so it appears simply to require intentional or knowing participation in a drag race, as does the misdemeanor statute.[33]  The relationship between these two statutes is the same, then, as the relationship between reckless endangerment and reckless manslaughter.  What distinguishes the two drag-racing crimes is just the harm caused by the defendant.  In this and other like statutes, harm matters.

There is more behind the Model Penal Code’s expansive version of the mens rea requirement, though, than the drafters’ apparent assumption that harm lacks any independent bearing on a crime’s gravity.  There also is a second, distinct assumption, namely, that only by assigning a mental state to the social harm that is the statute’s target can the statute adequately answer the two normative questions on which criminal liability ought to hinge: (1) whether the risk posed by the defendant’s conduct was unjustifiable; and (2) whether the defendant’s disregard of the risk, or his failure to perceive the risk, “justifies condemnation.”[34]  The Code’s drafters assumed, in effect, that every criminal statute must operate on the same model as the Code’s reckless-manslaughter provision, which—by assigning a mental state of “recklessly” to the “death of another” element[35]—requires the jury to decide for itself whether the conduct posed an “unjustifiable” risk of death and whether the defendant was culpable in relation to this risk.[36]

This second of the drafters’ assumptions is, like the first, belied by statutes like Iowa’s drag-racing homicide statute, which measure the unjustifiability and culpability of the risk-taking without assigning a mental state to the social-harm element.  As Professor Mark Kelman has said, offenses like drag-racing homicide are related to reckless and criminally negligent homicide in much the same way that tort negligence per se is related to ordinary tort negligence.[37]  Statutes defining offenses like drag-racing homicide embody antecedent legislative judgments of unjustifiability and culpability per se.[38]  These antecedent legislative judgments—though made in relation to the social harm that is the statute’s target—are based on the statute’s other elements and on the mental states associated with those other elements.[39]  In the crime of drag-racing homicide, for example, the antecedent legislative judgment hinges on proof that the actor knowingly or intentionally participated in a motor vehicle “speed contest” on a public “highway.”  It would be redundant, then, to assign a mental state to the harm; it would be redundant, that is, to put to the jury directly the questions whether the conduct posed an unjustifiable risk of death and whether the actor was culpable in relation to this risk.  In effect, the legislature already has answered these questions on the basis of the statute’s circumstance and conduct elements and their accompanying mental states.

Statutes that embody these sorts of antecedent determinations of unjustifiability and culpability per se are commonplace.  For example, most state criminal codes have drunk-driving-homicide statutes, in which the driver’s liability hinges exclusively on his or her intoxication at the time of the fatal accident.[40]  These statutes do not require the jury to make a determination that the defendant was reckless or negligent with respect to the result element—that is, the death of a person.[41]  Instead, the statutes’ only mental states pertain to the conduct and attendant-circumstance elements.[42]  They usually require, first, that the defendant act purposely with respect to the conduct element—namely, driving a motor vehicle—and, second, that the defendant act knowingly with respect to an attendant circumstance element—namely, the fact that the defendant had consumed an intoxicant.[43]

Likewise, a substantial minority of states have specific “drug-induced homicide” statutes.[44]  These statutes generally require, first, that the defendant deliver one of several specified controlled substances—for example, heroin, methamphetamine, or cocaine—and, second, that another person die as the result of ingesting the controlled substance.[45]  The statutes do not require the government to prove that the defendant was reckless or criminally negligent with respect to the social harm that is the target of the statute.[46]  Instead, by way of mens rea, they typically require the government to prove only that the defendant knew that he or she was delivering the controlled substance.[47]

There is room for disagreement about whether statutes like these are desirable[48]—about whether society is better served by rules embodying antecedent legislative judgments of unjustifiability and culpability per se,[49] or instead is better served by statutes that delegate to the finder of fact the responsibility for making ad hoc, case-specific judgments of unjustifiability and culpability.[50]  What is not subject to disagreement, though—and what is critical to the argument here—is just that legislatures traditionally have made extensive use of both kinds of criminal statutes.[51]  From the fact that legislatures traditionally have made extensive use of both kinds of criminal statutes, it follows that the courts ought not to adopt a version of the mens rea presumption that wishes away statutes embodying antecedent legislative judgments of unjustifiability and culpability per se.  In exercising their delegated power to develop substantive criminal-law background principles,[52]after all, the courts are merely “partners in the enterprise of lawmaking.”[53]  They do not dictate to the legislature.

II.  The Shortcomings of the Supreme Court’s Approach

This criticism of the expansive Model Penal Code version of the mens rea presumption seems to point toward a particular limitation on the presumption.  One of the defining features of crimes like drunk-driving homicide is the fact that the result element does not define the boundary between lawful and unlawful conduct.  The underlying conduct in drunk-driving homicide—driving while drunk—is criminal even when it does not cause death, injury, or property damage.  And the same is true of drag racing and drug trafficking and of the felonies that provide the bases for felony-murder prosecutions.  This feature of the homicide statutes suggests a possible shorthand formula for identifying elements that do not require the assignment of a mental state.  We could say an element does not require the assignment of a mental state if—like the element of death in drunk-driving homicide—it merely aggravates conduct that already is criminal.

This, as it happens, is the formula that the courts usually have used to define the scope of the mens rea presumption.  State and federal courts, when they have recognized that the question of mens rea must “be faced separately with respect to each material element of the crime,”[54] usually have held that the presumption of mens rea does not apply to elements that make a crime more serious; it only applies to elements that “make[] the conduct criminal.”[55]  In Staples v. United States,[56] for example, the Supreme Court said that the presumption requires the government to prove some mental state with respect to all “the facts that make [the] conduct illegal.”[57]  The Supreme Court spoke even more clearly in Carter v. United States,[58] where it said that “[t]he presumption in favor of scienter . . . requires a court to read into a statute only that mens rea which is necessary to separate wrongful conduct from ‘otherwise innocent conduct.’”[59]

The Court’s most recent application of this limiting principle came in Dean v. United States.[60]  The statute at issue in Dean was 18 U.S.C. § 924(c)(1)(A)(iii), which, in effect, defines an aggravated version of the offense of carrying a firearm during a crime of violence.[61]  Under this section, a person who uses or possesses a firearm during a crime of violence or a drug-trafficking crime will be subject to an enhanced minimum sentence of ten years “if the firearm is discharged.”[62]  InDean, both sides agreed that the defendant, Dean, had carried a firearm during a crime of violence—the robbery of a bank.[63]  And both sides agreed that the firearm had gone off.[64]  But the discharge appeared to have been accidental (since Dean cursed after the gun went off).[65]  So the question arose whether the government was required to prove some mental state with respect to the discharge.[66]  Dean argued that the government was required to prove that he had discharged the gun intentionally or knowingly.[67]  The Supreme Court concluded, though, that Congress had meant, by its omission of a mental state, not to require a mental state with respect to the discharge.[68]

In reaching this result, the Supreme Court said that Dean’s reliance on the presumption of mens rea was misplaced.[69]  The Court explained that the presumption did not apply to the discharge element, since the defendant’s conduct in cases prosecuted under 18 U.S.C. § 924(c)(1)(A)(iii) is unlawful even apart from the discharge of the firearm.  “It is unusual to impose criminal punishment for the consequences of purely accidental conduct,” the Court said.[70]  “But it is not unusual to punish individuals for the unintended consequences of their unlawful acts.”[71]  In effect the Court applied in Dean the same limiting principle it had applied in cases likeStaples and Carter, namely, that the presumption of mens rea applies only to “the facts that make [the] conduct criminal.”[72]

The academic commentary has been broadly critical of this limitation on the mens rea presumption.[73]  What interests me, however, is the somewhat more focused criticism offered by Justice Stevens in his dissenting opinion in Dean.[74]  In arguing that the discharge element in 18 U.S.C. § 924(c)(1)(A)(iii) required a mental state, Justice Stevens relied in part on the mens rea presumption.[75]  To the Dean majority’s reliance on the distinction between aggravating elements and elements that make conduct criminal, Justice Stevens responded by proposing a refinement of the distinction.  He said, in substance, that the “aggravating-element” limitation on the mens rea principle really only applies to aggravating elements that measure the degree of harm inflicted by the defendant:

The Court cites the felony-murder rule . . . and Sentencing Guidelines provisions that permit increased punishment based on the seriousness of the harm caused by the predicate act . . . .  These examples have in common the provision of enhanced penalties for the infliction of some additional harm.  By contrast, § 924(c)(1)(A)(iii) punishes discharges whether or not any harm is realized. . . .  For [this and other] reasons, § 924(c)(1)(A)(iii) is readily distinguishable from the provisions the majority cites.[76]

These four sentences are brief to a fault.  But the twofold gist of the sentences can be summarized as follows.  First, elements that are designed to measure the degree of harm inflicted by the defendant—that, in Justice Stevens’s words, go to “the seriousness of the harm caused by the predicate act”—sometimes can justify increased punishment quite apart from whether the government is required to prove any mental state with respect to the harm.  Second, elements designed to do something other than measure the harm—like the discharge of a firearm under § 924(c)(1)(A)(iii)—usually cannot justify increased punishment unless the government is required to prove some mental state with respect to them.[77]

This second point is the controversial one.  Why did Justice Stevens suppose that elements that are designed to measure something other than the harm usually cannot justify increased punishment absent proof of an accompanying mental state?  The only explanation appears in Justice Stevens’s enigmatic statement that 18 U.S.C. § 924(c)(1)(A)(iii) was intended “to serve a different purpose [than provisions that impose increased punishment on the basis of the seriousness of the harm]—namely, to punish the more culpable act of intentional discharge.”[78]  The implication of this remark is that factors other than harm are significant only to the degree that they signal enhanced culpability.

This explanation seems wrong, though.  It is at least arguable that, as the majority said in Dean,[79] the discharge element in § 924(c)(1)(A)(iii) was designed to do something other than measure the degree of the defendant’s culpability.  The majority thought the discharge element mattered not because it signified enhanced culpability but because it signified enhanced risk.  The discharge of a firearm during a bank robbery, the Court said, “increases the risk that others will be injured, that people will panic, or that violence (with its own danger to those nearby) will be used in response.”[80]  According to the majority, then, the discharge element might have been designed not to measure the defendant’s culpability—not, that is, to measure the defendant’s subjective perception of risk—but rather to measure the degree of objective risk posed by his conduct.

The Dean majority appears to have been correct in thinking that the degree of risk posed by an actor’s conduct sometimes has significance that is independent of the actor’s perception of the risk.  The Model Penal Code’s definition of reckless endangerment, for example, requires proof not only of culpability but of actual risk; it is satisfied only when the actor’s conduct “places or may place another person in danger of death or serious bodily injury.”[81]  A person who believes without any basis that he is driving ninety miles per hour is not guilty of reckless endangerment if he really is driving within the speed limit.[82]  The same is true of criminally negligent homicide, reckless manslaughter, and even depraved-indifference homicide.  In all these offenses, the actor’s liability depends not only on the actor’s culpability but also on the degree of objective risk posed by his conduct.[83]  There is no reason in principle, then, why increased risk should never be significant in its own right.  And indeed some criminal codes assign—or purport to assign—significance to the risks created by a defendant’s conduct without requiring proof of enhanced culpability.[84]

The Dean majority’s reliance on the objective risk posed by Dean’s conduct—as a basis for the enhanced punishment imposed under § 924(c)(1)(A)(iii)—suggests a powerful alternative basis for Justice Stevens’s implied criticism of the traditional judge-made version of the mens rea presumption, however.  The magnitude of even an “objective” risk, and indeed the very existence of the risk, is always tied to the defendant’s perspective—to what the defendant knew about his conduct and about the surrounding facts and circumstances.  Strictly speaking, purely objective probabilities don’t exist outside the world of indeterministic microphysics.[85]  At the macroscopic level, probabilities are just a reflection of the incompleteness of our knowledge of the world.[86]  If we knew everything there was to know about the objective facts—“all the forces by which nature is animated and the respective situation of the beings who compose it”[87]—probability would give way to certainty.[88]  The very notion of probability, then, presupposes “a perspective that is defined by possession of certain information but not other information.”[89]

Dean illustrates this.  It is possible now, after the fact, to reconstruct the objective facts surrounding the discharge of Dean’s gun—the position and orientation of the gun, the trajectory of the bullet, the location of the bank’s employees and customers, and so on.  And so it is possible now to say that, when the gun discharged, the purely “objective” probability that the bullet would injure one of the bank’s employee or customers was zero.[90]  The bullet was bound to travel through the partition separating the two bank tellers, ricochet off a computer, and come to rest harmlessly on the teller counter.[91]  Thus, when the majority in Dean says that the discharge of Dean’s firearm “increase[d] the risk that others [would] be injured,”[92] it cannot mean the agent-independent risk.  It must, rather, mean the risk or probability as calculated from some “perspective that is defined by possession of certain information but not other information.”[93]

In criminal law, objective probabilities are calculated from the defendant’s perspective.  More precisely, the probabilities of interest to the criminal law are calculated on the basis of a factual setup defined by what the defendant knows of the background facts and circumstances.[94]  (In the words of the Model Penal Code’s definitions of recklessness and negligence, the probabilities are measured on the basis of “the circumstances known to [the actor].”)[95]  This is true where the finder of fact bears the responsibility for making a case-specific assessment of the nature and degree of risk, as he does in, say, a prosecution for reckless homicide.[96]  But it is true as well where the legislature uses specific, factual elements—such as the discharge of a firearm—to mark the existence of a risk that is unjustifiable per se.[97]  After all, the probabilities that are the subject of the antecedent legislative judgment of unjustifiability per se are the same probabilities that are the subject of a fact-finder’s case-specific judgment.  And so, for example, in a prosecution for drunk driving, the antecedent legislative determination of unjustifiability per se hinges on proof that the defendant knew he was driving and knew that he had consumed an intoxicant.[98]  And in a prosecution for drag-racing homicide, the antecedent legislative judgment of unjustifiability per se hinges on proof that the defendant knew he was participating in a speed contest on a public highway.

This account of objective risk, though curious sounding, is utterly uncontroversial.  Consider a typical Fourth Amendment case, for example.  The lawfulness of a warrantless search or seizure usually depends on whether the evidence available to the officer satisfied one of two probability thresholds: the probable cause standard or the reasonable suspicion standard.  In applying these two probability thresholds, the courts insist that the probabilities at work are “objective,” rather than “subjective.”[99]  Still, the courts measure these objective probabilities just as the Model Penal Code requires the fact finder to do in criminal cases: on the basis of “the facts and circumstances known to the officer.”[100]  Here and elsewhere, then, courts measure even objective probabilities according to what the actor himself knew of the background facts and circumstances.

This is to say, the real trouble with the traditional judge-made version of the mens rea presumption is not, as Justice Stevens supposed, that the moral significance of risk depends on the defendant’s mental state.  The real trouble is that the very existence of risk depends on the defendant’s mental state.  Offense elements like the discharge of a firearm—elements that are designed to measure the objective risk posed by the actor’s conduct, rather than the harm inflicted by his conduct—can perform their assigned function only if they are tied somehow to what the actor knew about the underlying facts.  Therefore, elements designed to measure the risk ordinarily require the assignment either of a “knowingly” mental state or of a mental state like negligence or recklessness, whose existence turns on an assessment of the underlying “circumstances known to [the actor].”[101]

III.  The Mens Rea Presumption as an Actus Reus Presumption

From these criticisms—of the Model Penal Code version of the mens rea presumption, on the one hand, and of the judge-made version, on the other—it is possible to construct a new version of the presumption.  This new version of the presumption would not apply to elements whose exclusive function is to measure the degree of harm inflicted by the crime.  Instead it would apply to every other kind of element: to elements that define risk-enhancing attendant circumstances, like the intoxication of the actor in drunk-driving homicide; to elements that define the nature of the required conduct, like the “driving” element in drag-racing homicide; and to elements that define risk-manifesting intermediate results, like the discharge of the firearm in 18 U.S.C. § 924.  Moreover, the presumption would apply even to harm elements when the statute’s remaining elements—the circumstances, the conduct, and the intermediate results—do not clearly embody an antecedent legislative judgment of unjustifiability and culpability per se.

This might sound, at first hearing, like a relatively modest change in the mens rea presumption.  But it really works a fundamental change in the presumption’s underpinnings.  The new version is grounded not on concerns about fine-tuned assessments of subjective moral blameworthiness but rather on concerns about whether the defendant’s conduct even was wrong.  It is grounded, in other words, on concerns about the existence of the crime’s actus reus, not on concerns about culpability.

To explain, criminal liability is usually thought to hinge on the answers to two different questions.  The first is the question of “legality” or “wrongdoing,” which in effect asks whether the actor’s conduct violated an external, objective rule of conduct.[102]  The second is the question of “culpability,” which in effect asks whether the actor, despite having violated a rule of conduct, nevertheless lacks “the minimum blameworthiness required to be held criminally liable for the violation.”[103]  Courts and scholars both have assumed that the mens rea presumption really speaks to the second of these questions, and not without justification.  After all, the Supreme Court in Morissette identified the presumption of mens rea not with the requirement of “an evil-doing hand” but rather with the apparently distinct requirement of “an evil-meaning mind.”[104]

Still, it is the requirement of wrongdoing—of “an evil-doing hand”—on which the so-called mens rea presumption mostly bears.  This was one of Holmes’s central insights in The Common Law.  Holmes was concerned with establishing, in criminal law as elsewhere, “tests of liability [that] are external, and independent of the degree of evil in a particular person’s motives or intentions.”[105]  But he recognized that the objective risk posed by an actor’s conduct could not be measured except according to “the circumstances known to him.”[106]  And so he recognized that “[s]o far . . . as criminal liability is founded upon wrong-doing in any sense, . . . [it] must be confined to cases where circumstances making the conduct dangerous were known [to the actor].”[107]  He dismissed, moreover, the possibility that the requirement of mens rea is meant to accomplish more than this: “[T]he mens rea, or actual wickedness of the party, is wholly unnecessary, and all reference to the state of his consciousness is misleading if it means anything more than that the circumstances in connection with which the tendency of his act is judged are the circumstances known to him.”[108]

If Holmes was wrong in expressing doubt about whether “the actual degree of personal guilt involved in any particular transgression . . . is an element at all,”[109] he was right in thinking that the most important function of mental-state requirements is to tell us what the actor knew of the surrounding circumstances and thereby to tell us what the objective risk posed by the actor’s conduct was.  It is this critical function that the mens rea presumption, as reconfigured by Justice Stevens in Dean, really is designed to serve.  The mens rea presumption serves this critical function by requiring that mental states be assigned to elements whose purpose is, at least in part, to measure the risk associated with the actor’s conduct.  Paradoxically, then, the mens rea presumption really is an actus reus presumption; it requires the courts to presume that the legislature meant to require proof of an indispensable mental component of the actus reus—knowledge of the “circumstances making the conduct dangerous.”[110]

Finally, there is nothing conceptually problematic in the recognition that the actus reus has an indispensible mental component.  Courts long have recognized that the actus reus includes a requirement of a voluntary act and that this voluntary-act requirement has a mental component.  The Washington Court of Appeals explained this point nicely in State v. Utter[111]:

There are two components of every crime.  One is objective—the actus reus; the other subjective—the mens rea.  The actus reus is the culpable act itself, the mens rea is the criminal intent with which one performs the criminal act.  However, the mens rea does not encompass the entire mental process of one accused of a crime.  There is a certain minimal mental element required in order to establish the actus reus itself.  This is the element of volition.[112]

The effect of Justice Stevens’s reconceptualization of the mens rea presumption is just to show that another facet of the actus reus—the objective risk—has a mental component as well.[113]

IV.  The Actus Reus Presumption as a Presumption of General Intent

What I have said so far would provide, at best, a thin basis for urging a state or federal court to adopt the proposed limitation on the scope of the mens rea presumption.  Thankfully, though, the proposed limitation is grounded in more than Justice Stevens’s remark in his dissent in Dean and on more than my own theoretical excursus into the nature of objective probability.  The proposed limitation also has a strong grounding in what the courts say about the difficult subject of general and specific intent.

Federal and state courts often have said that what the presumption of mens rea really presumes is that the legislature meant to require “general criminal intent,” as opposed to “specific intent.”  This version of the mens rea requirement appears to have originated in United States v. Lewis,[114] where the Fourth Circuit observed that courts applying the mens rea requirement usually wound up concluding—of the statute being interpreted—that “only general intent is needed.”[115]  In the intervening years, the Fourth Circuit’s observation has become a kind of formula.  The Supreme Court invoked this formula in Carter v. United States,[116] where it said  that “the presumption in favor of scienter demands only that we read [18 U.S.C. § 2113(a), which defines the federal bank-robbery offense] as requiring proof of general intent.”[117]  Federal courts of appeals, too, now frequently say of the mens rea requirement that “absent any express reference to intent, [courts] . . . generally presume that proof only of ‘general’ rather than of ‘specific’ intent is required.”[118]

At first glance, this proposition—that the presumption of mens rea requires only general intent, not specific intent—appears to have little bearing on the question addressed in this Article.  After all, the proposition appears to speak only to the kind of mental state required by the presumption of mens rea, rather than to the question of which elements require mental states.  But this first glance is deceiving.  When the courts say that the presumption of mens rea requires only general intent, not specific intent, they are not just saying something about what kind of mental state is required.[119]  They are also saying something about which objective elements the mental state attaches to.  And what they are saying, as it turns out, revolves around exactly the same distinction that formed the basis for Justice Stevens’s argument in Dean, namely, the distinction between elements that measure harm and elements that measure risk.

To explain, the terms “general intent” and “specific intent” do not describe mental states, or at least they do not describe mental states in the way that terms like intentionally, purposely, knowingly, recklessly, negligently, willfully, and maliciously do.  When a legislature defines the mental state for an element, it uses terms like purposely, knowingly, recklessly, and so on.  It never uses the terms general intent and specific intent.  Nor, in most places, do judges use the terms general intent and specific intent in instructing juries.  Rather, they use terms like purposely, knowingly, recklessly, and so on.[120]

If general and specific intent are not the names of mental states, though, what are they?  The answer is that whether a particular mental state counts as a general intent or a specific intent will depend not just on the nature of the mental state itself but also on the kind of objective element to which it is attached.  The mental state of “intentionally,” for example, sometimes will count as a general intent and sometimes will count as a specific intent, depending on what objective element the mental state attaches to.  When the mental state of intentionally attaches to an element that is designed to measure the harm from the offense—say, the element of serious bodily injury in the crime of aggravated assault—the mental state of intentionally will usually be classified as a specific intent.[121]  When the mental state of intentionally attaches instead to an element that is designed to measure the risk posed by an offense—say, the element of discharge of a firearm—it will be classified as a general intent.

Granted, this isn’t what the courts actually say when they articulate the distinction between general and specific intent.  What the courts typically say is that a crime is a general-intent offense if it requires the government to prove only that “the defendant intended to do the proscribed act,”[122] and that, by contrast, a crime is a specific-intent offense if it requires the government to prove that the defendant also intended to “achieve some additional consequence.”[123]  But the only way to make sense of this distinction between an “additional consequence” and “the proscribed conduct” is to differentiate the (1) the social harm that is the statute’s ultimate target from (2) earlier events in the causal sequence leading up to the social harm, whose significance lies in their contribution to the risk.

To illustrate, imagine a case where the defendant uses a firearm to kill another person.  The event can be broken down into several steps: first, the shooter squeezes the trigger of the firearm; second, the firearm goes off, sending a bullet in the direction of the victim; third, the bullet strikes the victim’s body; and fourth, the damage inflicted by the bullet causes the victim’s death.  The act of squeezing the trigger clearly seems to be part of the “act,” rather than an “additional consequence.”[124]  And the last event in the causal sequence—the death of the victim—is clearly an “additional consequence.”  (Courts uniformly classify intent-to-kill homicide as a specific intent crime.)[125]  But what of the two events that mediate the causal connection between the squeezing of the trigger and the death of the victim?  Are they “additional consequences” or just part of “the proscribed act”?

At first glance, the discharge of the firearm might appear to be an “additional consequence.”  In causal terms, the discharge of the firearm is a consequence of squeezing the trigger.  What is more, it appears to be a truly separate or “additional” event.  After all, sometimes pulling the trigger of a gun causes a gun to discharge, and sometimes it does not.[126]

But courts have said that the discharge of a firearm does not qualify as an “additional consequence.”  Consider, for example, California decisions interpreting a state statute that prohibits “discharg[ing] a firearm in a grossly negligent manner.”[127]  The California courts have held that this statute requires proof that the defendant actually intended that the firearm go off; it is not enough that he intended to squeeze the trigger.[128]  Nevertheless, the courts have said that this statute defines a “general intent crime, because . . . its mental state consists of an intent to do the act that causes the harm.”[129]  Thus, the discharge of the firearm cannot be an “additional consequence” for purposes of the definition of specific intent.

Nor, in our original illustration, is the bullet’s initial contact with the victim’s body “an additional consequence.”  Granted, in purely causal terms, the bullet’s contact with the victim’s body plainly is a consequence both of the squeezing of the trigger and the firearm’s discharge.  What is more, this initial contact appears to be a truly separate event; the discharge of a firearm sometimes causes a bullet to strike another person’s body, and it sometimes does not.  Nevertheless, courts uniformly have held—in interpreting statutes that define the crime of battery—that an intent to bring about physical contact with another person’s body is a form of general intent, not specific intent.[130]  This means that the bullet’s initial contact with the other person’s body cannot be considered an “additional consequence” for purposes of our definition of specific intent.

So what’s going on here?  All three of the events that followed the squeezing of the trigger—the discharge of the firearm, the bullet’s initial contact with the victim’s body, and the death of the victim—appear to be consequences of the conduct.  Why is only one of these events—the death of the victim—treated as an “additional consequence” for purposes of the definition of specific intent?  The answer, as I have said, lies in the distinction between (1) the ultimate harm at which the statute is targeted and (2) the intermediate events that contribute to the risk of that harm occurring.  In our hypothetical shooting, only the death of the victim is the kind of harm at which criminal statutes are targeted.  Statutes that proscribe, say, the intentional discharge of a firearm are not ultimately targeted at the discharge of firearms.  These statutes proscribe the discharge of firearms not because the discharge of a firearm is harmful in itself but because the discharge of a firearm creates or enhances a risk of death or physical injury.

There are other facets to the complex distinction between general and specific intent.[131]  But this facet defines the real content of the distinction.  When courts say that an offense will qualify as a specific-intent offense if it requires proof that the defendant intended to “achieve some additional consequence” beyond “the proscribed act,” what they really mean (usually) is that an offense will qualify as a specific-intent offense if it requires proof that the defendant intended to bring about the social harm at which the statute is targeted.  And when the courts say that an offense will qualify as general-intent offense if it requires only proof that the defendant intended to do “the prohibited act,” what they really mean is that an offense will qualify as a general-intent offense if it requires only proof that the defendant intended to do something, or to cause something to exist or occur, that creates a risk or increases the magnitude of the risk.[132]

What all this means, finally, is that the limiting principle grounded in the distinction between general and specific intent often will operate very much like the limiting principle suggested by Justice Stevens in his dissent in Dean.  It will favor the assignment of mental states to those elements that are designed to measure the risk associated with the conduct (that are part of “the proscribed act,” in other words) but not to those elements that are designed instead to measure the social harm from the offense (that qualify as “additional consequences,” in other words).

Conclusion

Stuart Taylor said in 1990 that “[t]he careful case-by-case distinctions of [Justice] Stevens do not lend themselves to pigeonholing and do not attract much attention.”[133]  Justice Stevens’s dissenting opinion in Dean v. United States provides further evidence both of Justice Stevens’s tendency to articulate “careful case-by-case distinctions” and of the unfortunate fact that these distinctions only rarely “attract much attention.” [134]  In Dean, Justice Stevens recognized that what underlies the courts’ intuitions about the limits of the mens rea presumption is not a distinction between aggravating elements and elements that “criminalize otherwise innocent conduct,” but rather a distinction between elements that measure harm and elements that measure risk.[135]  Despite the novelty and force of this insight, though, no member of the Court joined Justice Stevens’s dissent, and no commentator has paid this insight any attention.[136]

Justice Stevens’s insight deserves attention, and not just because it happens to be right.  Legislatures routinely fail to specify the mental states associated with objective elements, and so courts frequently face the question whether a particular element requires a mental state.  Justice Stevens’s proposed refinement of the mens rea principle would make itself felt in a substantial number of these cases.  It would have made itself felt in Dean itself, of course, if the majority had heeded it.  It would have suggested that the discharge element in 18 U.S.C. § 924(c)(1)(A)(iii) required a mental state—if not “knowingly” or “intentionally,” as Dean’s attorneys hoped,[137] then perhaps “with criminal negligence” or “recklessly.”[138]  But the import of this distinction is not remotely limited to Dean.

Take, for example, the question addressed by the Second Circuit in United States v. Falu.[139]  Falu was convicted of aiding and abetting the distribution of heroin within 1000 feet of a school and accordingly was subject to the sentence enhancement imposed by 21 U.S.C. § 860(a).[140]  Section 860(a), like many state statutes,[141] enhances the penalties for drug dealers whose offenses occur near schools.[142]  But § 860(a) is silent on the question of whether this proximity element has an associated mental state.[143]  On appeal, Falu argued that § 860(a) “does not apply unless a defendant had specific knowledge of the proximity of a school.”[144]  But the Second Circuit rejected this claim, relying primarily on the traditional version of the mens rea presumption:

[The proximity element in] section [860(a)] does not criminalize otherwise innocent activity, since the statute incorporates section 841(a)(1), which already contains amens rea requirement—one must ‘knowingly or intentionally . . .
distribute . . . a controlled substance.’ . . . Anyone who violates section [860(a)] knows that distribution of narcotics is illegal, although the violator may not know that the distribution occurred within 1,000 feet of a school.[145]

Application of Justice Stevens’s more refined version of the mens rea principle might have led to a different result in Falu.  The proximity element in 21 U.S.C. § 860(a) is designed to measure not the harm associated with the defendant’s drug dealing but the risk.  In other words, the sentence enhancement triggered by the proximity element is based not on the assumption that narcotics sales in the vicinity of an elementary or secondary school somehow harm the students, but rather on the assumption “that narcotics sales in the vicinity of an elementary or secondary school endanger the students” by increasing the risk that the drugs will fall into the students’ hands.[146]  Because the proximity element in § 860(a) is designed to measure risk, rather than harm, and because the existence even of objective risk depends on the background facts and circumstances known to the actor, the court ought at least to have presumed that the proximity element required the assignment of a mental state—if not “knowingly,” as Falu’s attorneys hoped,[147] then perhaps the mental state of “recklessly” or the mental state of “with criminal negligence,” both of which judge the risk of an attendant circumstance’s existence on the basis of “the [underlying] circumstances known to [the actor].”[148]

Finally, if Justice Stevens’s approach to this issue lacks the ideological purity of the Model Penal Code’s approach,[149] it also has the potential to succeed where the Code’s approach failed: in actually getting itself accepted by the courts.[150]  In recent years, courts increasingly have framed the mens rea as a presumption of general intent and so have taken pains to distinguish general from specific intent.[151]  The distinction between general and specific intent—between mental states attached to the “prohibited act” and mental states attached to “additional consequences”—closely parallels the distinction made by Justice Stevens between elements that measure risk and elements that measure harm.  The courts, then, already are stumbling toward the very rule proposed by Justice Stevens.