Federal Mens Rea Interpretation and the Limits of Culpability’s Relevance
FEDERAL MENS REA INTERPRETATION AND THE LIMITS OF CULPABILITY'S RELEVANCE
DARRYL K. BROWN
between the actor and the risk that the offense targets.3 As a category, those
offenses generally have low (misdemeanor-level) punishments, and they are not
my focus here.4 The more challenging and significant complications regarding
the fault requirements arise in serious offenses carrying significant prison terms.
As a practical matter, these difficulties concern which elements of felony
offenses carry mens rea requirements. For this question, the principle that
“criminal liability requires proof of a culpable mental state” is insufficient.
Strict liability as to some elements of offenses is widespread and often
noncontroversial. Yet courts lack reliable interpretive tools for distinguishing which
statutory elements will carry strict liability, surely in part because legislatures do
not seem to follow consistent principles when designating strict liability
elements and employ drafting conventions that make their intentions on such
Courts’ interpretive choices regarding mens rea present a choice of two
accounts of culpability, which are also bases for sentencing. These choices
address ideas of what degree of culpability is required to justify criminal
punishment, which is the same choice underlying legislatures’ design of offense
definitions. One account is proportionate culpability. This principle states that
punishment must be in accord with or in proportion to culpability, and it is
sometimes endorsed by courts6 and overwhelmingly by scholars.7 Mens rea must
3. See, e.g., Staples v. United States, 511 U.S. 600, 607 n.3 (1994) (reiterating that public welfare
statutes may dispense with a “mental element”); Liparota v. United States, 471 U.S. 419, 433 (1985)
(stating that a public welfare offense is “a type of conduct that a reasonable person should know is
subject to stringent public regulation and may seriously threaten the community’s health or safety”);
United States v. Int’l Minerals & Chem. Corp., 402 U.S. 558, 565 (1971) (reasoning that when
dangerous products or obnoxious waste materials are involved, the probability of regulation is so great
that anyone who is aware of possessing them is presumed to also be aware of the regulation).
4. Some of the classic literature includes GLANVILLE WILLIAMS, CRIMINAL LAW: THE
GENERAL PART ch. 6 (2d ed. 1961); JEROME HALL, GENERAL PRINCIPLES OF CRIMINAL LAW 325–51
(2d ed. 1960); Richard Wasserstrom, Strict Liability in Criminal Law, 12 STAN. L. REV. 731 (1960)
(offering justifications for some versions of strict felony liability). For an account of the tide turning
against strict liability in the late 20th century, see Richard G. Singer, The Resurgence of Mens Rea: III—
The Rise and Fall of Strict Criminal Liability, 30 B.C. L. REV. 337 (1989).
5. There is very little constitutional law restraint on legislatures’ creation of strict criminal
liability offenses, but, notwithstanding a couple of decisions implying the possibility of a constitutional
requirement of mens rea, such as Lambert v. California, 355 U.S. 225 (1957). For an example of a
statute held to violate due process for insufficient mental state requirements, see Illinois v. Madrigal,
948 N.E. 2d 591, 597–98 (Ill. 2011) and cases cited therein. For a leading account of the (minimal)
constitutional law on strict liability, see Alan C. Michaels, Constitutional Innocence, 112 HARV. L. REV.
6. Morissette is often considered a canonical case on this point, stating
[t]he 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.
342 U.S. at 250. For a more recent, though implicit statement of the idea, see United States v.
FloresFigueroa, 129 S. Ct. 1886, 1890 (2009) (reasoning that not applying a statute’s requirement of
knowledge to all elements is inconsistent with the heavy penalty). State courts routinely require a
mental element as well. See, e.g., Illinois v. Valley Steel Prods. Co., 375 N.E. 2d 1297, 1305 (1978)
(stating that “[i]t would be unthinkable to subject a person to a long term of imprisonment for an
attach to every normatively significant element of the offense as a means to
measure culpability; one who intends a bad result demonstrates more fault than
one who intends conduct but not the result it causes. The Model Penal Code
(MPC), the American Law Institute’s mid-century recommendation of how
American criminal law should be designed, strongly endorsed proportionate
culpability.8 This idea, however, characterizes neither American criminal law
nor courts’ interpretation of mens rea requirements.
The contrasting principle one might call threshold culpability and seems to
be the dominant view of American courts and legislatures. On this view, proof
of mens rea is needed only to determine whether one is innocent or
blameworthy for some offense. Culpability merely defines eligibility for
punishment; it plays no necessary role in setting the magnitude of a sentence
nor places an upper limit on punishment. Other considerations, such as
deterrence goals, perform these functions. I argue that limiting mens rea to this
threshold role explains how federal courts (like their state counterparts)
routinely interpret statutes to impose strict liability on elements that are both
central to a statute’s definition of wrongdoing and the sole basis for authorizing
significant sentence enhancements.9 The implication for statutory interpretation
is that mens rea need not attach to every element. Thus, culpability as to
conduct can shift one’s status from innocent to blameworthy, after which strict
liability may apply to circumstances or results, even when large distinctions in
liability and sentence may rest on their proof.10
This eligibility-defining function of mens rea coupled with denial of its
offense he might commit unknowingly”).
7. See, e.g., JOHN GARDNER, OFFENCES AND DEFENCES: SELECTED ESSAYS IN THE
PHILOSOPHY OF CRIMINAL LAW 213–38 (2007); ANDREW VON HIRSCH & ANDREW ASHWORTH,
PROPORTIONATE SENTENCING 15, 29 (2005); H.L.A. HART, PUNISHMENT AND RESPONSIBILITY 8–12
(2d ed. 2008); Gardner, supra note 2, at 51.
8. The MPC codified the commitment to the fault requirement in three related provisions
designed expressly as canons of statutory construction. One dictates that courts should imply a mens
rea of at least recklessness whenever a statute lacks a mental state requirement. MODEL PENAL CODE
§ 2.02(3) (1962). A second dictates a presumption that any mental state requirement in a statute applies
to all elements of the offense (so that no part of the offense is a strict liability element). Id. § 2.02(4).
The third defines a category of minor, non-criminal “violations” to which mens rea presumptions do
not apply. Id. § 2.02(5). The MPC has inspired more than half of the states to adopt varying degrees of
9. For an account of U.S. Supreme Court mens rea doctrine criticizing threshold culpability’s
inadequacy in assuring proportional liability, see Stephen F. Smith, Proportional Mens Rea, 46 AM.
CRIM. L. REV. 127, 137–42 (2009).
10. This differs from the public welfare offense category, in which mens rea may not attach to any
element of the offense. Strict felony liability usually requires some initial proof of an offender’s mens
rea for at least one element of an offense. Antony Duff describes this “substantive” strict liability:
Liability is strict if it requires no proof of fault as to an aspect of the offence: while mens rea
must be proved as to some elements . . ., it need not be proved as to every fact, consequence
or circumstance . . . . Liability is substantively strict if it does not depend on proof of some
appropriate moral culpability as to some aspect of the offence—proof of some fault that
would justify condemning the defendant for committing the offence.
R.A. Duff, Strict Liability, Legal Presumptions and the Presumption of Innocence, in APPRAISING
STRICT LIABILITY 125, 125–26
(A.P. Simester ed., 2005)
relevance to setting the quantum of punishment is pervasive in criminal
statutory interpretation, although the proportionality principle continues to
characterize some offense definitions and judicial decisions. Moreover, some
strict liability decisions, which formally follow a threshold liability model, in fact
require sufficient proof to infer some culpability even for the strict-liability
element. Courts note this, suggesting the appeal of proportionate culpability or
incomplete commitment to strict liability without any regard for culpability in
explaining the full magnitude of liability and punishment.
This ambivalence about the role of culpability and its proof through mens
rea requirements complicates courts’ design and adherence to clear rules of
statutory interpretation. In particular, rules that guide and simplify inquiries
into statutes’ meaning and legislative intent are problematic. Canons of
construction that lead to strict liability readings often conflict with those that
dictate an inference of mens rea to serve the value of proportionate liability
(conflicting interpretive canons are a familiar problem11). It is hard to read
contemporary decisions on mens rea and discover principled (or even
consistent) grounds for choosing one sort of canon over another.
To develop these arguments, I focus on federal law, although the story is
much the same in state jurisdictions, including even those in which legislatures
have adopted MPC-inspired statutes favoring the presumption of mens rea.
Recent federal court decisions show that judges routinely interpret statutes to
impose severe sentence enhancements based solely on proof of a strict-liability
element. These decisions represent examples of how culpability serves only to
define eligibility for punishment rather than the magnitude of the sentence. Yet
the U.S. Supreme Court recently used a different interpretive approach. It
reaffirmed an approach to explicit mens rea requirements that uses fault as a
proportionate limit on sentencing enhancements and that restates an
interpretive rule favoring mens rea requirements for all significant offense
The federal criminal code lacks any general principles regarding mens rea
interpretation, but the Supreme Court has established several interpretive
canons to guide inferences of mental state requirements. Public welfare offenses
and regulatory crimes concerned with “dangerous devices” are marked as a
special category for which no mens rea will be presumed.12 Otherwise, “far more
than the simple omission of the appropriate phrase from the statutory definition
is necessary to justify dispensing with an intent requirement.”13 “[T]he failure of
No. 2 2012]
Congress explicitly and unambiguously to indicate whether mens rea is required
does not signal a departure from this background assumption of our criminal
law.”14 The distinctive nature of federal criminal law, in which offenses often
include elements specifying the basis for federal jurisdiction, has led to a
presumption of strict liability for jurisdictional elements. These elements rarely
play a normative role in defining criminal wrongdoing.15 And statutory crimes
based in common-law offenses are presumed to carry common-law culpability
requirements even when the statute includes no such term.16
The key components of the vast case law on statutory interpretation—and a
large body of scholarly literature focused on it—are familiar.17 A court’s first
step is to read the statute’s text for a plain or ordinary meaning, which can be
the best guide to legislative intent.18 This may require resorting to dictionary
definitions and—often in mens rea analysis—inferences of typical grammar or
syntax usage, sometimes drawing from writing style manuals as guides on
common usage—and therefore—textual meaning.19 If the text’s meaning is
unambiguous, the question is settled. If not, courts resort to other strategies,
including canons of construction, such as those just noted, and other evidence of
statutory purpose and legislative intent.20 They may also seek guidance from a
predecessor statute21 or from the title of the statute. Furthermore, under the
doctrine of in pari materia,22 the construction of related statutes or those
employing similar language can guide courts. The priority and influence of
these secondary tools vary according to the interpreter’s methodology.
All these strategies are means of determining the legislature’s intent for the
statute. Yet it is easy to see how these different interpretive methods can and do
lead to different results. Interpretation of a text built on assumptions about
common syntax and grammar usage, and the meanings they ordinarily generate
may yield an understanding different from one built on a normative
presumption that mens rea terms apply to all non-jurisdictional elements. A
statutory text’s ordinary meaning does not always yield the mens rea
requirement that we would infer from relying on a presumption for culpability
requirements. That disconnect, in fact, turns out to be a notable occurrence in
federal criminal law interpretation.23 And it has led multiple federal courts to
adopt interpretive strategies that minimize culpability requirements on the basis
of plain or ordinary meaning.
Consider United States v. Jones.24 The court’s task was to interpret the Mann
Act provision in 18 U.S.C. § 2423(a), which reads:
A person who knowingly transports an individual who has not attained the age of 18
years in interstate or foreign commerce, or in any commonwealth, territory or
possession of the United States, with intent that the individual engage in prostitution,
or in any sexual activity for which any person can be charged with a criminal offense,
shall be fined under this title and imprisoned not less than 10 years or for life.25
The defendant argued the mental-state term “knowingly” required the
government to prove not only that he knew he “transported an individual,” but
also that he knew she was under the age of eighteen.26 The government argued
the statute required no proof of knowledge regarding age. The age element is
critical to the severe sentencing range of ten years to life; an adjacent code
section criminalizes the same conduct but does not include an age element and
sets ten years as the maximum sentence, with no mandatory minimum.27 Thus,
whether mens rea attaches to the element is a question of culpability’s relation
to the gravity of punishment. Requiring mens rea only for the conduct follows
the principle of threshold culpability; once the defendant’s conduct makes him
culpable, the magnitude of his punishment can vary tremendously (decades in
prison) with no relation to his degree of culpability. Extending mens rea to the
age element, in contrast, links punishment proportionately to culpability.
The Jones court followed the federal courts’ consistent pattern, holding that
no mens rea requirement attached to the age element. The opinion relied on
two reasons for the conclusion. First, it first focused on the statute’s
“grammatical structure” and asserted that a standard grammar rule is that
“[a]dverbs generally modify verbs,”28 a premise that became a significant
interpretive canon in a series of cases. Given this view of the adverb’s role in
defining the sentence’s meaning, the court concluded that “the adverb
23. See United States v. Meek, 366 F.3d 705, 718–19 (9th Cir. 2004) (opting for the presumption of
scienter over the inference from the plain meaning).
24. 471 F.3d 535 (4th Cir. 2006).
25. 18 U.S.C. § 2423(a) (2006). The statute was amended after Jones’s offense; the statute as it
applied to him had a punishment range of five to thirty years. Jones, 471 F.3d at 538 & n.1.
26. Jones, 471 F.3d at 538.
27. 18 U.S.C. § 2421 (1998).
28. Jones, 471 F.3d at 539.
‘knowingly’ modifies the verb ‘transports’” but nothing else in the sentence.
Otherwise, “the thought that [adverbs] would typically modify the infinite
hereafters of statutory sentences would cause grammarians to recoil.”29 The
court found that “nothing on the face of this statute” suggested that Congress
intended to require defendant’s mens rea regarding the victim’s age.30
Evidence of Congress’s expressly instrumental purpose for the age element
confirmed the grammar-based reading. The statute “provid[es] extra
protection” to minor victims compared to adult victims, the court inferred,
making it unlikely that “Congress intended to make the evidentiary burdens” of
the crime harder for the prosecution to meet than for the lesser offense
protecting adult victims.31 Section 2423(a) “creates a more serious crime in
order to provide heightened protection against sexual exploitation of minors;” a
knowledge requirement would be “nonsensical” because it would “strip the
statute of its clear purpose: the protection of minors.”32
Every mens rea requirement makes offenses harder to prove; indeed, the
point is to ensure only the culpable are criminally punished and—if punishment
were calibrated to culpability—to ensure harsher punishment is based on
greater culpability. But in this Mann Act offense, once initial culpability is
demonstrated with proof of knowing transportation, culpability is normatively
irrelevant to justifying additional punishment; only instrumental goals control,
which is a view that contributes to the intuitive appeal of the grammar analysis.
Taken literally, Jones rejects the relevance of the blameworthiness of
knowingly transporting a minor to take part in sex crimes. But this
interpretation’s intuitive appeal rests, without acknowledgement in Jones, on a
recognition or presumption that most or nearly all offenders under § 2423(a)
are in fact culpable. Mr. Jones himself was driving a thirteen-year-old for three
days from one truck stop to another, where she engaged in commercial-sex
transactions.33 Proof of this conduct and of the minor’s age makes Mr. Jones’s
knowledge of her age easy to infer. Not all Mann Act cases make the inference
so easy. But the offense requires knowing transport of a person and intent for
that person to engage in sex crimes. These elements also make it likely (though
not certain) that the offender saw the victim, thus providing him some basis at
least to suspect or guess her age, in the context of intentional conduct (sex
crimes) that are especially harmful to minors. Even a legislature or court
committed to requiring culpability regarding the victim’s age might conclude
that proof of the offense elements, even with strict liability as to age, raises
29. Id. at 539; see also United States v. Jennings, 496 F.3d. 344, 354 (4th Cir. 2007) (quoting this
Jones passage and applying the same adverb analysis to hold that the statute’s knowledge requirement
in 18 U.S.C. § 2243(a) does not extend to the element defining a minor victim’s age).
30. Jones, 471 F.3d at 539 (because “the adverb ‘knowingly’ obviously modifies the verb,
‘transports,’” it “would be implausible to suggest that . . . [it] applies to both the noun and its dependent
32. Id. at 540.
33. Id. at 537.
sufficient inferential proof (in most cases) of minimal culpability—recklessness
or negligence—as to age. In other sex offenses with minor-age elements, such as
statutory rape and sexual assault, it is virtually assured that the offender sees
The plausibility of that inference depends in part on how reasonable it
seems that some offenders are non-negligent—for example, because a
seventeen-year-old appears older and has a false ID. Plus, the Mann Act’s
prohibited conduct, to transport, makes it possible that some offenders commit
the act without ever seeing the victim—although, the act plausibly provides a
reason to find negligence or recklessness given the nature of the intended
sexcrime activity. But by assuming that a large portion of cases raises a culpability
inference as to age, we can view the standard interpretation of the Mann Act
that Jones represents as resting on an expectation (in most cases) of a rough
proportionality between punishment and culpability, even without required
proof of mens rea and despite the court’s exclusively instrumental reasoning.
Yet courts’ reasonings matter. Pressed by limited time, judges rely on the
interpretive conventions and stated reasons of past decisions to provide quick
and clear guidance.35 Both of Jones’s rationales undermine a commitment to
culpability. Instrumental goals of maximizing harm prevention by more easily
imposing punishment simply reject the value of culpability. Grammar analysis
built on the “adverb canon” has the consistent effect of limiting mens rea terms
(usually expressed as adverbs) exclusively to conduct elements (verbs). If
federal common law includes a canon like “courts ordinarily read a phrase in a
criminal statute that introduces the elements of a crime with the word
‘knowingly’ as applying that word to each element”36—a canon that tracks the
MPC’s presumption that a mens rea term applies to all elements37—then Jones’s
adverb canon directly contradicts it.
These two different interpretive approaches are evident in a series of federal
court decisions interpreting the frequently prosecuted fraud offense defined in
18 U.S.C. § 1028A, which bars certain fraudulent uses of another’s identification
information. The statute reads: “Whoever, during and in relation to any felony
violation enumerated in subsection (c), knowingly transfers, possesses, or uses,
without lawful authority, a means of identification of another person shall, in
addition to the punishment for such felony, be sentenced to a term of
imprisonment of 2 years.”38
Prosecutors frequently use this statute to charge undocumented-immigrant
defendants who obtain fraudulent Social Security or Alien Registration cards
and similar documents to obtain employment in the United States. Defendants
usually admit that they knew their documents were fraudulent. (The fraud is
often obvious; many offenders use their correct names and photos, but a
fabricated ID number.)39 But they frequently and plausibly assert that they did
not know the identity number on the documents had been lawfully assigned to
other people. The question of statutory meaning is the familiar one: does
“knowingly” apply only to the verbs denoting conduct such as “uses,” or does it
apply also to “a means of identification” and most importantly to “identification
of another”? The interpretive choices distinguish two offenses: one that
punishes merely the knowing use of false documents by which one might
unknowingly harm a victim (though often not),40 and one that punishes only a
form of deliberate identity theft.
The Fourth Circuit faced a typical case in United States v. Montejo.41 The
court accepted that Mr. Montejo did not know the false numbers on his Social
Security and Alien Registration cards (which had his correct name and photo)
belonged to other people. But it found the statute unambiguous42 in not
requiring proof that he knew the numbers were “means of identification of
another.” Grammar analysis, backed by Congress’s instrumental purpose for the
statute, provided the key:
We begin with grammar. The word “knowingly” in this case is an adverb that modifies
the verbs “transfers, possesses, [and] uses.” . . . The direct object of these transitive
verbs is “a means of identification,” a nominal phrase that is further modified by the
adjectival prepositional phrase “of another person.” . . .
We think that, as a matter of common usage, “knowingly” does not modify the entire
lengthy predicate that follows it. Simply placing “knowingly” at the start of this long
predicate does not transform it into a modifier of all the words that follow. Good
usage requires that the limiting modifier, the adverb “knowingly,” be as close as
possible to the words which it modifies, here, “transfers, possesses, or uses.”43
40. A range of injuries is possible, but some false document use may not result in negative effects,
such as FICA reporting under another’s number. And that the offender’s income is reported to and
taxed by the Social Security Administration does not necessarily cause injury to the person assigned to
that number. The same is true with Alien Registration numbers, especially if, as in Montejo, the
number’s lawful assignee has subsequently obtained citizenship.
41. 442 F.3d 213 (4th Cir. 2006).
42. Id. at 217 (stating that “the statute is not ambiguous”).
43. Id. at 215 (citation omitted). The startling point of this analysis, however, is the judges
expected “good usage.” Id. Instead, given the judges’ familiarity with legislative drafting, one would
expect the weary observation that a given statute is “not a model of meticulous drafting.” See, e.g.,
The reading produced by that grammar analysis accords with the
interpretation dictated by the statute’s singular purpose of preventing injury:
“legislative history shows that Congress was concerned with aggravated identity
theft . . . [from] entirely innocent people.”44 Other appellate courts that also
reached this holding employed the same grammar analysis and assumed that
Congress’s instrumental goals took priority over more specific proof of
culpability.45 For the Eleventh and Eighth Circuits, “[t]he fact that the word
‘knowingly’—an adverb—is placed before the verbs ‘transfers, possesses, or
uses’ indicates that ‘knowingly’ modifies those verbs, not the later language in
In all these decisions, the interpretive tools undermine any commitment to
requiring mens rea for every element of the offense and to distinguishing
between more and less culpable offenders in the service of proportionate
liability and punishment. The courts frequently build analysis on the premise
(inferred to be Congress’s premise) of threshold culpability—that mens rea has
no role other than to separate innocent actors from knowing wrongdoers.
Hence Montejo understood Supreme Court precedents to apply “knowingly”
only to elements beyond the conduct verb when it was necessary to prevent the
liability of an innocent actor, rather than to further distinguish between greater
and lesser liability. Thus, “Liparota’s discussion of the scope of ‘knowingly’
should not be understood apart from the Court’s primary stated concern:
avoiding criminalization of otherwise non-culpable conduct.”47
Without this firm premise of mens rea’s limited role and of culpability’s
irrelevance to the severity (or proportionality) of punishment, it would be
puzzling that courts find a grammar analysis dictated by this adverb canon such
a compelling indication of plain meaning as to remove all ambiguity. For one,
precisely this question of what elements of a criminal statute an explicit mental
state term applies to is a longstanding and routine ambiguity in many offense
definitions.48 And a long line of federal court decisions interpret mens rea terms
United States v. Chin, 981 F.2d 1275, 1279 (D.C. Cir. 1992) (describing 21 U.S.C. § 861(a)(2)).
44. Montejo, 442 F.3d at 217. But see Villanueva-Sotelo, 515 F.3d 1234, 1243–44 (D.C. Cir. 2008)
(finding a contrary purpose in the legislation).
45. United States v. Mendoza-Gonzalez, 520 F.3d 912, 915–16 (8th Cir. 2008):
It is preposterous to think the same Congress that so plainly and firmly intended to increase
the penalty . . . would then so limit its imposition as to require the Government to prove that
the defendant knows he wrongfully possesses the identity “of another person” . . . [because it]
would place on the prosecution an often impossible burden.
(citations and internal quotation marks omitted).
46. United States v. Hurtado, 508 F.3d 603, 609 (11th Cir. 2007) (citing United States v. Jones, 471
F.3d 535, 539 (4th Cir. 2006), for the rule “[a]dverbs generally modify verbs”); Hurtado, 508 F.3d at 610
n.8 (stating that “the plain meaning of the knowledge requirement in § 1028A(a)(1) is unambiguous”);
Mendoza-Gonzales, 520 F.3d at 915 (8th Cir. 2008) (adopting and quoting the grammar-based analyses
in Hurtado and Montejo to reach the same holding on § 1028A(a)).
47. Montejo, 442 F.3d at 216 (referring to Liparota v. United States, 471 U.S. 419 (1985)); see also
Hurtado, 508 F.3d at 610 (quoting and adopting this view).
48. See W. LAFAVE & A. SCOTT, CRIMINAL LAW § 27 (1972) (describing “the ambiguity which
frequently exists concerning what the words or phrases in question modify” and offering the example of
to apply to statutory elements beyond the conduct term they are placed next to
in the sentence.49
Further, any plausible account of the ordinary meaning of a sentence like
the one in § 1028A—even as interpreted in Montejo—understands “knowingly”
to apply to more than the verb, despite denials to the contrary. In the sentence,
“whoever knowingly uses a means of identification of another [is guilty],”
knowingly surely applies to more than just the verb uses, simply because one
must know what one is using. Some verbs produce meaningful sentences
without direct objects; one can “knowingly walk” or “knowingly shout” because
those verbs need no referent to produce a meaningful sentence. But to
“knowingly engage” or to “knowingly transport” requires knowing what one
engages in or transports.50 Nonetheless, courts declare that “[t]he only
reasonable reading of [18 U.S.C.] § 2421 is one under which the adverb
‘knowingly’ acts only on the verb ‘transports’ and not on the noun
‘individual,’”51 or that in a statute that punishes one who “knowingly engages in
a sexual act” with a minor under sixteen, “[t]he efficacy of the adverb
‘knowingly’ extends no further than the verb ‘engages.’”52 Yet those courts
presumably expect proof that the defendant knew that what he transported was
a person, and knew the conduct he engaged in was sexual in nature.
Similarly, every court including Montejo implicitly applies “knowingly” in §
1028A beyond the verb, at least to “a means of identification.”53 The debate is
not whether the adverb modifies more than the verb; it is whether it applies
statute with a mens rea term “knowingly”); Liparota, 471 U.S. at 424–25, 425 n.7 (1985) (citing LaFave
& Scott and discussing the general problem of criminal statute ambiguity because “it is not at all clear
how far down the sentence the word ‘knowingly’ is intended to travel” and recognizing that
“knowingly” sometimes applies to several statutory elements).
merely to “a means of identification” or to “a means of identification of
another.” Yet the commitment to interpreting a statute to require mens rea only
for “knowingly using a means of identification” without lawful authority while
applying strict liability to the result that the identity information belongs to
another, leads courts to mis-describe their own reading of the language in an
effort to strengthen its justification.
The disregard for mens rea’s role in distinguishing between more and less
culpable offenders has more significant implications under § 1028A than under
the Mann Act’s § 2423. Under the Mann Act, as suggested above, Congress
could plausibly conclude some evidence of culpability regarding age will be
demonstrated from the statute’s other proof requirements even without
attaching mens rea to the age element. (The exceptions—one can speculate how
large this group is—are offenders who were reasonable in wrongly believing a
victim was at least eighteen.) The same inference of culpability for the
strictliability element does not hold for Aggravated Identity Theft under § 1028A.
Some offenders will knowingly use another’s name and identity information;
their knowing use of another’s identity will be easy to prove.54 Others will attach
randomly fabricated identity numbers to their true names and photos; some of
those fabricated numbers will be (unknown to the offender) assigned by law to
other people. Without requiring mens rea to “means of identification of
another,” no other proof required for the offense will provide a basis to infer
culpability on that point. More starkly than the Mann Act, strict liability here
groups together two sets of offenders of notably different culpability—knowing
thieves of others’ identities and unknowing (though arguably negligent)55
fabricators of identity numbers.
That distinction could be the reason why the U.S. Supreme Court has never
taken up federal appellate courts’ consensus that the Mann Act imposes strict
liability, although the Court did recently reject (nine to zero) in Flores-Figueroa
v. United States56 the appellate courts’ strict liability reading of § 1028A. Yet if
that was a reason, the Court left it unstated. Instead, the Court in
FloresFigueroa explained its rationale largely through ordinary grammar and textual
analysis. In Justice Breyer’s opinion for the Court, the plain meaning of the
statute required knowledge for all the critical elements:
54. See, e.g., United States v. Hurtado, 508 F.3d 603, 604, 609 n.7 (11th Cir. 2007) (Hurtado used
the birth certificate, a social security card, driver’s license (with Hurtado’s photo), bank debit card,
library card, and vehicle registration and insurance with the name, identity numbers, and other
information of Marcos Colon).
55. The argument from negligence might go as follows: one billion numbers are possible for
ninedigit Social Security numbers. Roughly 300 million are assigned now to living Americans. If we assume
reasonable people know the deceaseds’ numbers are not reassigned, the total assigned numbers rise to
approximately 450 million. Randomly taking a number, then, risks about a forty-five percent chance
that the number is assigned—a plausible basis for arguing one is negligent in doing so. (Thanks to my
colleague Rich Hines for this point.) Accepting that argument, one still must conclude the culpability of
such negligence does not meaningfully differ from the culpability of the knowing theft of another’s
identity number so that both acts receive the mandatory additional sentence of § 1028A.
56. Flores-Figueroa v. United States, 129 S. Ct. 1886, 1890 (2009).
As a matter of ordinary English grammar, it seems natural to read the statute’s word
“knowingly” as applying to all the subsequently listed elements of the crime. . . . In
ordinary English, where a transitive verb has an object, listeners in most contexts
assume that an adverb (such as knowingly) that modifies the transitive verb tells the
listener how the subject performed the entire action, including the object as set forth
in the sentence.57
The Court found these grammatical observations to hold more generally for
judicial construction of criminal offenses: “The manner in which the courts
ordinarily interpret criminal statutes is fully consistent with this ordinary
English usage. That is to say courts ordinarily read a phrase in a criminal statute
that introduces the elements of a crime with the word ‘knowingly’ as applying
that word to each element.”58
If Breyer’s claim is taken as a canon of construction,59 it should work much
like the MPC’s provision that dictates a culpability term should apply to all
material elements.60 At least for statutes that contain the word “knowingly” and
have a similar structure to § 1028A, mens rea disputes may be reduced. But
unless the Supreme Court recognizes the canon’s substantive implications for
the role of mens rea, it seems unlikely that courts will give much weight to the
canon’s effect. Although unnoted in Flores-Figueroa, the effect is to reserve
additional punishment only for knowing users of false documents who are also
knowing thieves of others’ identities, and lesser punishment for knowing users
who unknowingly—and thus less culpably—used another’s identity. In other
words, it seems unlikely that a grammar-based canon will be taken to signal that
mens rea interpretation in offense definitions should be informed partly by a
commitment to link punishment proportionately to liability, even though that is
Without judicial (or legislative) endorsement of that proportionate
culpability and punishment, Flores-Figueroa likely will meet resistance. A large
body of case law explicitly justifies strict liability interpretations by endorsing
the contrasting idea that mens rea serves only to shield innocent actors from
liability and not also to inform allocation of different magnitudes of
punishment. As a rationale for mens rea interpretation, this latter position,
“threshold culpability,” is the dominant one and seems to hold wide intuitive
appeal. That appeal is manifest in decisions endorsing strict-liability-based
sentence enhancements, and it is implicit in the lack of ambiguity many judges
find in the statutes they interpret to impose strict liability for aggravating
circumstance or result elements. Perhaps the adverb canon was attractive
57. Id. at 1890.
58. Id. at 1891.
59. Justices Scalia, Thomas, and Alito, at least, doubt that Justice Breyer’s claim is descriptively
accurate and expressly deny it has any normative validity as a canon of construction (that is, that
“knowingly” normally should apply to all elements). They join the holding only on grounds that the
statute’s plain meaning is clear. Id. at 1894 (Scalia, J., concurring); id. at 1895 (Alito, J., concurring).
60. MODEL PENAL CODE § 2.02(4) (1962). This provision adds the limitation “unless a contrary
purpose plainly appears.” One would expect the interpretive rule of Flores-Figueroa to be similarly
limited by distinctions in the structure and syntax of statutory language.
because it confirmed some judges’ misunderstanding of how grammar affects
ordinary meaning. However, with Flores-Figueroa’s rejection of it (for one type
of statute), courts will no longer be misled. Furthermore, one might suspect that
the adverb canon’s attraction was partly the normative appeal of the outcomes
it generated, which were fewer evidentiary barriers for punishments serving
instrumental goals. Linking degrees of culpability to degrees of punishment
creates just such a barrier. As it happened, another Supreme Court decision
handed down a week before Flores-Figueroa reiterated the principle of
In United States v. Dean,61 the Court held that a firearm statute imposes a
significant prison term solely upon proof of a strict-liability element. Mr. Dean,
who had been convicted of conspiracy to commit bank robbery and of
discharging a firearm during the robbery, argued that the firearm offense
required proof that he intentionally discharged the weapon. There was no
dispute that he fired the gun accidentally, hitting a wall of the bank. The statute
[A]ny person who, during and in relation to any crime of violence . . . uses or carries a
firearm, or who, in furtherance of any such crime, possesses a firearm, shall, in
addition to the punishment provided for such crime of violence . . . —
(i) be sentenced to a term of imprisonment of not less than 5 years;
(ii) if the firearm is brandished, be sentenced to a term of imprisonment of not less
than 7 years; and
t(hiiai)ni1f0thyeeafrirse.6a2rm is discharged, be sentenced to a term of imprisonment of not less
Dean was sentenced to the mandatory minimum of ten years under the final
clause for accidentally discharging the firearm, a term that ran consecutively to
his sentence of eight-and-a-third years for the robbery.63
Chief Justice Roberts’s opinion for the Court, like both lower courts in
Dean, concluded that Congress intended the final clause to apply even to
offenders who discharged guns accidentally. Part of the Court’s rationale was
textual: The passive voice syntax in the final clause implied absence of a mens
rea requirement (though not in the preceding clause). And structuring the
sentence into numbered clauses suggested Congress did not mean for the
Court’s implied intent requirement in the initial clause64 to extend to the final
clause. The “most natural reading” is that the initial “adverbial phrases . . .
61. 129 S. Ct. 1849 (2009).
62. 18. U.S.C. § 924(c)(1)(A) (2006).
63. United States v. Dean, 517 F.3d 1224, 1227 (11th Cir. 2008).
64. Dean, 129 S. Ct. at 1854 (citing United States v. Smith, 508 U.S. 228, 238 (1993), as holding the
statute’s phrases “during and in relation to” and “in furtherance of” imply an intent requirement
because each phrase “clarifies that the firearm must have some purpose or effect with respect to the
drug trafficking crime; its presence or involvement cannot be the result of accident or coincidence”)
(internal quotation marks omitted).
modify [only] their respective nearby verbs.”65 Unlike in Flores-Figueroa, the
Dean Court was explicit about the substantive implications of this textual
reading. The sentencing policy resulting from its reading of the statute reflects a
widely accepted norm of punishing unintended consequences more severely
than (different) intentional conduct or consequences: “it is not unusual to
punish individuals for the unintended consequences of their unlawful acts.”66
Thus, Dean makes a straightforward observation of the punishment tradition
under strict-liability elements instantiating the threshold culpability principle.
From Dean and Flores-Figueroa, one can draw the following interpretive
[I]t seems natural to read . . . “knowingly” as applying to all the subsequently listed
elements of the crime. . . . In ordinary English, where a transitive verb has an object,
listeners in most contexts assume that an adverb . . . that modifies the transitive verb
tells the listener how the subject performed the entire action, including the object as
set forth in the sentence.67
“The most natural . . . [and] better reading of the statute is that the adverbial
phrases in the opening . . . modify their respective nearby verbs, and that
neither phrase extends to the sentencing factors.”68
One view could be that courts now have another iteration of Llewellyn’s
dueling canons.69 The first supports reading a culpability requirement
throughout the statute and facilitates proof of mens rea as a basis for enhanced
punishment. The other stands as authority for finding no legislative intent to do
Another view, of course, is that the statutes are distinguishable, and so are
the Court’s interpretive rules. One statute has the express term “knowingly,”
and the Flores-Figueroa canon taken literally applies only to statutes with that
term; the other statute had an intent requirement inferred from the phrases
“during and in relation to” and “in furtherance of.”70 Also, Flores-Figueroa
addressed a statute with a straightforward sentence structure; Dean’s
interpretation arose from a statute broken into distinct clauses defining
sentencing factors and minimums.
Those with confidence in courts’ fidelity to such canons and ability to
recognize proper occasions for their application can be optimistic that—
although lower courts divided sharply in reading both statutes—this new
interpretive guidance should prevent future disagreements in analogous
statutes. Even if that confidence is well founded, however, the statutes as now
defined by the Court embody contrasting approaches to culpability and
punishment. That distinction in outcomes—and arguably in underlying
punishment principle—can mislead courts’ decisions of when and how each
interpretive rule applies.
Assume that each decision correctly inferred Congress’s intent. The result is
that Congress enacted statutes that dictate different relationships between
culpability and sentencing. The firearm statute imposes more punishment on
the already-guilty based on unintended consequences, and Dean reaffirms this
threshold culpability idea. The identity-theft offense as interpreted in
FloresFigueroa, on the other hand, links increased punishment to proof of increased
culpability, although the Court does not acknowledge this
proportionalculpability outcome. Those conflicting views of whether culpability has a role in
authorizing and justifying additional punishment for actors already culpable for
initial wrongdoing make it more difficult for courts to consistently choose
between and correctly apply the Dean and Flores-Figueroa canons. But
misapplication is more likely to spring from courts overlooking the
FloresFigueroa presumption that “knowingly” applies to all elements in statutes of
simple sentence structure, because judicial recognition of Dean’s outcome and
the threshold-culpability principle is more explicitly articulated and endorsed.
Consider that strict liability readings are well established for offenses with
language closely tracking § 1028A’s simple sentence structure.71 The Mann Act
is one example: “A person who knowingly transports an individual who has not
attained the age of 18 years in interstate or foreign commerce . . . with intent
that the individual engage in prostitution . . . shall be . . . imprisoned not less
than 10 years or for life.”72
The statute contains the word “knowingly,” and Flores-Figueroa says “it
seems natural to read . . . ‘knowingly’ as applying to all the subsequently listed
elements of the crime” because ordinarily we “assume that an adverb . . . tells
the listener how the subject performed the entire action, including the object as
set forth in the sentence.”73 No sentence structure distinguishes the Mann Act
from the identity-theft offense,74 such as the Dean firearm statute’s separately
71. See, e.g., 18 U.S.C. § 2243(a) (2006) (criminalizing “knowingly engaging in a sexual act” with
youth under the age of sixteen); United States v. Jennings, 496 F.3d. 344, 355 (4th Cir. 2007) (holding
no mens rea applies to § 2243(a)’s age element).
72. 18 U.S.C. § 2423(a) (2006).
73. United States v. Flores-Figueroa, 129 S. Ct. 1886, 1890 (2009).
74. 18 U.S.C. § 1028A(a)(1) (2006) (“Whoever, during and in relation to any felony violation
numbered clauses. Yet courts, taking Flores-Figueroa’s signal that “special
context” might justify a different reading of a similar statute, have held that
Flores-Figueroa’s interpretation does not change the established strict-liability
reading of the Mann Act.75
Courts have so far made no effort to explain what that “special context”
might be. But recall the earlier suggestion of a distinction between the Mann
Act and Aggravated Identity Theft: Culpability as to the strict-liability element
in the Mann Act will often be apparent through proving the statute’s
requirements; most offenders are similarly situated, with practical notice of the
victim’s age.76 That is not true for many offenders under the identity-theft
statute. Two very different types of actors would face the same enhanced
punishment if the Court had not rejected the strict liability interpretation
“identification of another”: deliberate identity thieves who knowingly used
another’s identity, and knowing users of false but randomly-generated identity
numbers that might belong to no one.77 The implication is that proportionate
culpability does matter, despite courts’ frequent statements to the contrary. The
“special context” that makes strict liability normatively acceptable is the type
that characterizes the Mann Act offenses—offenders whose culpability
regarding the strict liability element is typically apparent even without a mens
rea requirement, thereby justifying the enhanced punishment triggered by that
Two other examples suggest the same implicit rationale. The first is Dean. In
enumerated in subsection (c), knowingly transfers, possesses, or uses, without lawful authority, a means
of identification of another person shall, in addition to the punishment for such felony, be sentenced to
a term of imprisonment of 2 years.”).
75. See United States v. Cox, 577 F.3d 833, 838 (7th Cir. 2009) (“Flores–Figueroa does not compel
a [different] interpretation of § 2423(a) . . . . The Flores–Figueroa Court made clear . . . that ‘the
inquiry into a sentence’s meaning is a contextual one,’ and that a ‘special context’ might call for a
different statutory interpretation.”)
(citing Flores-Figueroa, 129 S. Ct. at 1891)
; United States v.
Daniels, 653 F.3d 399 (6th Cir. 2011) (presenting the same holding and rationale as in Cox); United
States v. Murray, 663 F. Supp. 2d 709, 713 (W.D. Wis. 2009) (rejecting defendant’s argument that 18
U.S.C. § 922(g) requires proof that he knew it was unlawful to possess a firearm because “how far the
adverb ‘knowingly’ reaches into the elements of an offense depends on the circumstances”).
76. United State v. Jones, 471 F.3d 535, 541 (4th Cir. 2006) (“Because an individual ‘is already on
notice that he is committing a crime when he transports an individual of any age in interstate commerce
for the purpose of prostitution,’ . . . it is both reasonable and just to conclude that ‘the transporter
assumes the risk that the victim is a minor.’”) (citation omitted).
As an inference about all offenders under a particular statute, this conclusion may be weaker in
the Mann Act than other offenses concerned with minors and sexual conduct because the Mann Act’s
age limit of eighteen includes older teenagers and the core conduct is “transporting” rather than direct
sexual contact. The inference is surely strong in statutes such as 18 U.S.C. § 2243(a), which prohibits
“knowingly engaging in a sexual act” with youth who are under sixteen. See, e.g., United States v.
Jennings, 496 F.3d. 344, 354 (4th Cir. 2007) (holding no mens rea applies to the age element).
77. See Flores-Figueroa, 129 S. Ct. at 1896 (Alito, J., concurring) (reasoning that without extending
the knowledge requirement, “if a defendant uses a made-up Social Security number without having any
reason to know whether it belongs to a real person, the defendant’s liability under § 1028A(a)(1)
depends on chance”) (emphasis added). Thus, to avoid enhanced punishment based on chance rather
than a culpable mental state, even for offenders already culpable for using false identification, Congress
must have intended a mens rea requirement.
addition to citing the statute’s ordinary meaning and its seeming endorsement
of threshold culpability (via its observation that sentences routinely increase
based on unintended consequences), Roberts’s opinion argued that strict
liability for discharging a gun was normatively appropriate (and therefore more
likely to be the intent of Congress) because the statute’s violation yields
evidence of blameworthiness even without the mens rea requirement. Roberts
emphasized the apparent fault of even the accidental gun discharge. Proof of
mens rea is not the only way for a court to find blameworthiness. For the
conduct this statute governs, “the actual discharge of a gun”—even when
unintended—“does not mean that the defendant is blameless. The sentencing
enhancement in subsection (iii) accounts for the risk of harm resulting from the
manner in which the crime is carried out, for which the defendant is
responsible.”78 One who “brings a loaded weapon to commit a crime runs the
risk that the gun will discharge accidentally.”79 Anyone who violates the statute,
in other words, is in all likelihood reckless,80 which is the same assumption
Congress might make (perhaps less convincingly) about those who knowingly
transport a youth under eighteen years old for criminalized sexual conduct. The
elements of this firearm offense—culpable commission of a predicate crime and
knowing possession of a gun—make the gun’s discharge inevitably culpable
(reckless), even if accidental.
The second example that strict liability tends to apply only when proof of
mens rea is unnecessary to reveal culpability is the large and well-established
case law holding that jurisdictional elements in federal crimes require no mens
rea. Jurisdictional elements are often not normatively significant and so do little
to distinguish levels of culpability. For example, one demonstrates fault by
knowingly making false statements to another, but one is not necessarily more
at fault when making them to a federal agent or agency.81 One demonstrates
fault in assaulting another, but not notably more fault by knowing the victim is a
federal employee.82 One can be culpable transporting firearms in some
circumstances, but not more culpable for knowing the firearms crossed state
lines.83 And so on.
Together, these interpretations of statutes suggest some commitment to the
idea that culpability should play a role in justifying degrees of and eligibility for
punishment. Strict-liability elements appear only when proof of mens rea is not
78. United States v. Dean, 129 S. Ct. 1849, 1855 (2009) (emphasis added).
79. Id. at 1855–56 (emphasis added).
80. More precisely, it is possible, if unlikely, that one might be merely negligent. That would
require a robber who possessed a loaded gun during a robbery but did not recognize (and then
disregard) the risk of discharge that reasonable people would recognize in that setting.
81. United States v. Yermian, 468 U.S. 63, 69–70 (1984). Concededly, some false statements—such
as those to friends about having “other plans”—are worse than others. But as in Yermian, statements
about a significant matter—one’s criminal history in a job application requiring security clearance—is
not notably worse for being made to a federal agency than merely to a corporate employer.
82. United States v. Feola, 420 U.S. 671, 684 (1975).
83. United States v. Dancy, 861 F.2d 77, 81–82 (5th Cir. 1988) (no mens rea in firearm offense 18
U.S.C. § 922(g) for the element “in interstate or foreign commerce”).
needed to calibrate culpability because either the element does not matter or
the evidence of culpability (for the strict-liability element) comes from proof of
other elements. One problem with this suggestion is that it does not track
courts’ stated rationales; it is hard to have confidence in a principle that courts
not only do not mention but often deny. Dean rejected the defendant’s
argument that the firearm statute ratcheted up sentences stepwise with
consistent culpability requirements tied to various conduct: intentional
possession (five years), then intentional brandishing (seven), then intentional
discharge (ten). The Court saw no reason to be confident that Congress
connected sentence severity so closely to mens rea requirements. That is hardly
a heartening commitment to proportionate culpability as a punishment
Another problem with this suggestion is that it may not be reliably served by
the interpretive canons. Some canons work well: The rule against implying
mens rea for jurisdictional elements fits this account, as does the
FloresFigueroa canon that “knowingly” ordinarily applies to all elements. Its “special
context” exception could serve the goal of proportionate culpability, if courts
specify and limit its application along these lines.84 But it is not clear that Dean’s
interpretive model, which rejects mens rea for a statute’s enumerated clauses,
will lead to proportionate-culpability readings of other, similarly drafted
statutes. For that to be so, the legislature would have to reserve such drafting
for elements with no normative import or—as in Dean—elements for which the
proof will typically make culpability evident even without a mens rea
requirement. It seems clear that Congress does not do that. One statutory
example among many is 21 U.S.C. § 841, a frequently enforced felony drug
offense that specifies sentencing ranges depending on the quantity and type of
drug involved. Courts uniformly interpret those drug-quantity clauses to lack
mens rea requirements,85 and in many cases proof of the other elements allows
no easy inference that defendants likely knew, or were reckless as to, the drug’s
weight or quantity. And though § 841(a)(1) limits liability to offenders who
“knowingly or intentionally . . . distribute . . . a controlled substance,” its
adjacent penalty-enhancement sections do not. Two provisions authorize a
doubling of a § 841 sentence upon proof of an additional fact: if the offender
“violates section 841(a)(1) of this title by distributing a controlled substance to
a person under twenty-one years of age”86 or by doing so “within one thousand
84. Thus far, United States v. Cox, 577 F.3d 833 (7th Cir. 2009), best explains why the Mann Act
provision of 18 U.S.C. § 2423(a), a statute similar in structure to § 1028A, is distinguishable by context.
The explanation is notably thin. It recognizes that Section 2423(a) has long been interpreted to contain
no mens rea for the minor-age element and notes that the element is not essential to distinguishing
culpable from innocent conduct—although that is true for § 1028A as well. Id. at 837–38.
85. See, e.g., United States v. Gonzalez-Rodriguez, 621 F.3d 354, 362 (5th Cir. 2010); see also
United States v. Frazier, 213 F.3d 409, 418–19 (7th Cir. 2000) (holding that 21 U.S.C. § 861(a), which
makes it an offense to use a minor in a drug operation, does not require knowledge of the minor’s age);
United States v. Jennings, 496 F.3d. 344, 354–55 (4th Cir. 2007) (holding that U.S.C. §2243(a)’s
knowledge requirement does not extend to the element defining a minor victim’s age).
86. 21 U.S.C. § 859(a) (2006).
feet of . . . public or private elementary, vocational, or secondary school.”87
Neither section repeats § 841’s “knowingly” requirement,88 and both are
consistently interpreted not to require proof of mens rea for the element that
triggers aggravation of the sentence—the minor’s age or proximity to the
One might posit the same story about drug sales to minors for the Mann
Act: offenses that nearly always require face-to-face dealings with a minor will
yield proof of negligence or recklessness as to age even without a requirement
of proof. But that is a much less plausible assumption regarding proximity to
schools, especially in “urban areas where schools are not clearly visible from
points within the 1,000-foot zone or are not readily identifiable.”90 In these
cases, courts explain strict liability without the suggestion present in Dean of
evident recklessness regarding the strict-liability element.
Usually, the rationale for strict liability is, as with other statutes, purely
instrumental: the statute aims not to enhance punishment based on greater
culpability but in order to increase deterrence. To that end, offenders “bear the
burden of ascertaining where schools are located and removing their operations
from those areas.”91 With the school-zone drug offense as with other statutes,
courts have reasonable evidence that legislatures in fact drafted an offense with
prevention goals in mind that supersede any commitment to punishment in
proportion to culpability.92 Unless courts have wildly misread congressional
intent—unless Congress in fact wanted mens rea requirements implied into
grading and penalty enhancement statutes that lack such terms, despite its long
acquiescence to strict-liability interpretations—the evidence is clear that
legislators often intend culpability to serve only the threshold function of
triggering punishment eligibility. Sentences can then be adjusted with regard to
result and circumstance elements unconnected to an actor’s culpability.
The evidence for this limited consideration of culpability goes, in fact, much
deeper than statutes with strict-liability sentencing provisions, such as those we
find in drug offenses. Adjusting punishment based on either unintended (or
No. 2 2012]
FEDERAL MENS REA INTERPRETATION
unforeseen) resulting consequences or circumstances unknown to the offenders
is a standard feature of sentencing policy, independent of the punishment
components built in to offense definitions.93 More significantly, punishment
without a proportional connection to culpability has deep roots in the criminal
law tradition; it characterizes definition of core criminal offenses. Many familiar
crimes are defined and distinguished from lesser offenses solely by results that
defendants did not intend or foresee. Defendants who do identical conduct with
identical intent and identical knowledge of surrounding circumstances and risks
can commit different offenses carrying very different penalties. Core examples
include homicide, assault, and criminal attempt offenses. These offenses impose
different degrees of liability on offenders whose conduct, intentions, and
awareness are the same, depending on whether their identical actions caused a
death, merely injured, or caused no injury at all.94 Theft offenses also commonly
adjust sentences according to the value of property taken, rather than the value
offenders intended to take or believed they took.95 The car driver whose
drunken or otherwise negligent driving happens, by good fortune, to cause no
injury faces less liability than an identically situated driver whose identical
negligent conduct results in a fatality. This is, of course, a well-known puzzle, “a
deep, unresolved issue in the criminal liability,”96 at least if one is strongly
committed to treating like cases alike based on individual culpability. It is part
of the large literature on “moral luck,” which addresses the challenge of
explaining why we hold actors responsible based on outcomes partly beyond
93. See U.S. SENTENCING GUIDELINES MANUAL § 2A2.2(b)(3) (2008), available at
http://www.ussc.gov/Guidelines/2008_guidelines/2008_manual.cfm (increasing offense level for
aggravated assault according to the seriousness of the injury); id. § 2D2.3 (increasing offense level for
operating or directing the operation of a common carrier under the influence of alcohol or drugs if
death or serious bodily injury results); see also United States v. Dean, 129 S. Ct. 1849, 1855 (2009)
(citing the U.S. Sentencing Guidelines).
94. The lack of any relation between culpability and punishment based on results is built in to
several traditional criminal law statutes and doctrines, including lesser punishments for attempts than
completed offenses and the availability of the impossibility defense (now rejected in many
jurisdictions). See Sanford H. Kadish, Supreme Court Review, Foreword: The Criminal Law and the
Luck of the Draw, 84 J. CRIM. L. & CRIMINOLOGY 679, 680–84 (1994).
95. Theft offenses, at least, probably share the rationale that the Dean opinion identified for the
weapon-discharge offense: negligence or recklessness as to property value may fairly be imputed to
most offenders as a justification for the greater penalty. The story is more complicated for
homicideassault-harmless-attempt distinctions, for which the state of mind regarding the specific risk is the same
for offenders in each instance. Furthermore, separate from the definition of the offense, sentencing
policies adjust punishments according to actual magnitudes of harm without regard to offenders’
awareness or intent. See U.S. SENTENCING GUIDELINES MANUAL § 2A2.2(b)(3) (2008) (increasing
offense level for aggravated assault according to the seriousness of the injury); id. § 2D2.3 (increasing
offense level for operating or directing the operation of a common carrier under the influence of
alcohol or drugs if death or serious bodily injury results); see also Dean, 129 S. Ct. at 1855.
96. GEORGE FLETCHER, RETHINKING CRIMINAL LAW 473–74 (1978).
97. See, e.g., BERNARD WILLIAMS, MORAL LUCK (1981); Dana K. Nelkin, Moral Luck, in
STANFORD ENCYCLOPEDIA OF PHILOSOPHY (Edward N. Zalta ed., 2008), available at
http://plato.stanford.edu/archives/fall2008/entries/moral-luck/ (“Moral luck occurs when an agent can
be correctly treated as an object of moral judgment despite the fact that a significant aspect of what she
Montejo’s construction of the identity-theft statute notwithstanding, all of
this suggests that courts are often correct to infer that legislators are not
committed to a principle of proportionate culpability. They often do not intend
punishment to be proportionately connected to culpability, which means they
do not intend mens rea terms to attach to facts that adjust sentences even when
these facts can impact sentences quite dramatically. Criminal law’s central
commitment—evident in its core offenses and doctrines and in courts’ routine
construction of every felony crime definition—is only proof of threshold
culpability. Proof of fault shifts one from the status of innocent individual, who
is protected from official punishment, to the status of culpable wrongdoer, on
whom a wide range of punishments can be imposed for a wide range of reasons
bearing no relation to the wrongdoer’s fault other than some causal link.
Proportionate culpability retains sufficient normative appeal to characterize
some offense definitions and their attendant allocations of punishment.
FloresFigueroa is one example. Justice Alito explained the need for proof of
knowledge for the statute’s elements so that imposing additional prison time
(for offenders already culpable for using false IDs) would not “depend on
chance.”98 Chance refers to the probability that “a made-up Social Security
number” in fact belonged to another although the offender was “without having
any reason to know” that.99 Perhaps Justice Roberts’s account of recklessness
for weapon discharges in Dean is some indication of its appeal as well. But too
many offenses and penalty-enhancement statutes reject any role for culpability
in apportioning sentences to support a conclusion that mens rea consistently
does anything more than define eligibility for punishment.
Felony strict liability will endure as long as that is true. Taking that as a
given, the remaining concern is whether the interpretive rules will serve courts
well in their construction of criminal statutes. Courts show institutional
inclination to devise canons that strengthen the presumption of mens rea to a
degree that they would force Congress to be exceedingly clear about its intent
to create a strict-liability element. Yet existing canons do not do so. Canons are
instead designed simply to yield interpretations that fulfill legislative intent
regarding mens rea and culpability, which is no easy goal given the
is assessed for depends on factors beyond her control. . . . The problem of moral luck arises because we
seem to be committed to the general principle that we are morally assessable only to the extent that
what we are assessed for depends on factors under our control . . . At the same time, when it comes to
countless particular cases, we morally assess agents for things that depend on factors that are not in
their control. . . . [I]f we accept the Control Principle in unqualified form, and deny the existence of
moral circumstantial, character, and causal luck, then it seems that no actual punishment could be
justified on the basis of moral desert.”).
98. United States v. Flores-Figueroa, 129 S. Ct. 1886, 1896 (2009) (Alito, J., concurring).
99. Id. Under a strict liability interpretation, “if a defendant uses a made-up Social Security
number without having any reason to know whether it belongs to a real person, the defendant’s liability
under § 1028A(a)(1) depends on chance.” Id. To avoid enhanced punishment based on chance rather
than a culpable mental state, even for offenders already culpable for using false identification, Congress
must have intended a mens rea requirement.
FEDERAL MENS REA INTERPRETATION
congressional track record of employing a range of statutory drafting
conventions. Perhaps Flores-Figueroa will reduce courts’ susceptibility to
illconceived grammar assumptions and bring some clarity at least to the
construction of statutes comparable in their language to § 1028A. But by so far
not applying the Flores-Figueroa presumption to the Mann Act, courts’
decisions might be evidence of the challenge of even this aspiration. Canons,
and their exceptions, must be interpreted as well. To do so, courts refer to
criminal law’s underlying values and traditions. The tradition courts routinely
cite for mens rea questions (though not by this label) is threshold culpability—
that mens rea’s task is only to distinguish the innocent from the culpable.100 The
utility of interpretive canons is reduced to the degree courts need to invoke that
tradition to inform their application of canons. Although the mens rea tradition
of threshold culpability is criminal law’s dominant account of the role of mens
rea, it is not, as Flores-Figueroa demonstrates, the exclusive one.
11. See Karl N. Llewellyn , Remarks on the Theory of Appellate Decision and the Rules or Canons about How Statutes are to be Construed, 3 VAND . L. REV. 395 ( 1950 ) (famously juxtaposing a series of seemingly conflicting canons of construction).
12. See Staples v. United States , 511 U.S. 600 , 606 - 07 ( 1994 ); Morissette v . United States , 342 U.S. 246 , 254 ( 1952 ).
13. United States v. U. S. Gypsum Co., 438 U.S. 422 , 438 ( 1978 ).
14. Liparota v. United States , 471 U.S. 419 , 426 ( 1985 ).
15. See , e.g., United States v . Yermian , 468 U.S. 63 , 68 - 71 ( 1984 ) ; United States v . Feola , 420 U.S. 671 , 676 - 77 ( 1975 ).
16. Morissette , 342 U.S. at 263 (discussing the presumption of mens rea for common-law offense analogues). Furthermore, the rule of lenity dictates narrow construction in the face of statutory ambiguity, supporting the notion that common-law offenses require mens rea .
17. For a brief overview, see FREDERICK SCHAUER , THINKING LIKE A LAWYER 148-70 ( 2009 ). For a thorough overview of the literature and an insightful account of the textualist-intentionalist distinction, see Caleb Nelson , What Is Textualism?, 91 VA. L. REV. 347 ( 2005 ).
18. See , e.g., United States v. Fisher, 289 F.3d 1329 , 1338 ( 11th Cir . 2002 ) (“The plain language of a statute is presumed to express congressional intent .”).
19. Nelson , supra note 17, at 376 ( noting “all mainstream interpreters” begin with these strategies); see , e.g., Carter v . United States , 530 U.S. 255 , 271 ( 2000 ) (“In analyzing a statute, we begin by analyzing the text .”); United States v . Montejo , 442 F.3d 213 , 215 ( 4th Cir . 2006 ).
20. Larry Alexander & Saikrishna Prakash, “Is That English You're Speaking?” Why Intention Free Interpretation Is an Impossibility , 41 SAN DIEGO L. REV. 967 , 968 - 74 ( 2004 ) ; Kent Greenawalt, Are Mental States Relevant for Statutory and Constitutional Interpretation?, 85 CORNELL L . REV. 1609 , 1619 - 23 ( 2000 ) (both noting assumptions about a writer-speaker's purpose or intent is critical to inferring meaning); Nelson, supra note 17, at 355 (noting that textualist judges use these strategies). Textualists in particular would likely look to the statute and surrounding statutes alone, rather than to other evidence of legislative intent . Id. at 355-60; SCHAUER, supra note 17 , at 158-63.
21. See , e.g., Morissette , 342 U.S. at 266 n.28.
22. Statutes “in pari materia” -dealing with the same subject matter-should be construed consistently .
34. Strict liability as to age is the general rule for these offenses also . See Richard Singer, The Resurgence of Mens Rea: II-Honest But Unreasonable Mistake of Fact in Self Defense , 28 B.C. L. REV . 459 , 469 - 70 ( 1987 ) (noting this and tracing strict liability regarding a minor's age to Regina v. Prince and examining its wide adoption by American courts even after the British Parliament rejected the Prince rule) . Strict liability regarding a minor's age is common in non-sex-related offenses as well. The Fourth Circuit had previously held the age element in a drug statute to lack a mens rea requirement in a drug statute. See United States v . Cook , 76 F.3d 596 , 602 ( 4th Cir . 1996 ).
35. There are also normative reasons to constrain judges with rules, interpretive and otherwise . See Frederick Schauer, Rules and the Rule of Law , 14 HARV. J. L. & PUB. POL 'Y 645 , 679 - 94 ( 1991 ) (arguing that rules, even if over- or under-inclusive in some applications, are desirable means of constraining officials' power).
36. Flores-Figueroa v . United States , 129 S. Ct . 1886 , 1891 ( 2009 ).
37. MODEL PENAL CODE § 2 . 02 ( 3 ) ( 1962 ).
38. 18 U.S.C. § 1028A(a)(1) ( 2004 ).
39. But see United States v . Hurtado , 508 F.3d 603 , 604 , 609 n. 7 ( 11th Cir . 2007 ) (stating that the defendant used several documents with another's name and identity numbers), abrogated by FloresFigueroa v . United States , 129 S. Ct . 1886 , 1894 ( 2009 ). Yet when a defendant uses another's name and ID number, circumstantial evidence makes it easy to prove that he knew he was using the identity “of another .” Id.
49. Again , a canonical example is Morissette v . United States , 342 U.S. 246 , 271 ( 1952 ) (applying “knowingly” to “property of the United States” as well as to verbs defining prohibited conduct); see also Arthur Andersen LLP v . United States , 544 U.S. 696 , 704 - 05 ( 2005 ) ; United States v . X-Citement Video , Inc., 513 U.S. 64 , 68 - 70 ( 1994 ).
50. Academic linguists -who courts call grammarians-make this point compellingly in an amicus brief:
“[b]eginning January 1 , 2007 , every Colorado employer will be required to affirm . . . that . . .
it has not knowingly hired an unauthorized alien .” . . . If it were correct to say that adverbs modify only verbs, th[is] sentence would be nonsensical . [It] would mean that Colorado employers must affirm that they have not knowingly hired (anyone). Brief of Professors of Linguistics as Amici Curiae in Support of Neither Party at 4, Flores-Figueroa v . United States , 129 S. Ct . 1886 ( 2009 ) (No. 08 - 108 ).
51. United States v. Cox , 577 F.3d 833 , 837 ( 7th Cir . 2009 ); see also Jones, 471 F.3d at 539 (interpreting 18 U.S.C. §§ 2421 , 2423 and concluding “[i] t would be implausible to suggest that 'knowingly' . . . is intended to modify the noun 'individual .'”).
52. United States v. Jennings , 496 F.3d 344 , 354 ( 4th Cir . 2007 ).
53. See Flores-Figueroa, 129 S. Ct . at 1890 ( 2009 ) (“It makes little sense to read the provision's language as heavily penalizing a person who 'transfers, possesses, or uses, without lawful authority' a something, but does not know, at the very least, that the 'something' (perhaps inside a box) is a 'means of identification .'”); id. at 1894 ( Scalia, J., concurring) (making the same point and citing United States v . Villanueva-Sotelo , 515 F.3d 1234 , 1237 (D.C. Cir . 2008 )).
65. Id . at 1854 . Justice Breyer dissented from the holding and, one assumes, this generalization about adverbs' typical effects. A week later, he wrote something very close to the opposite in his FloresFigueroa majority opinion . Chief Justice Roberts, despite authoring Dean's majority opinion, joined Justice Breyer's Flores-Figueroa opinion and not Justice Scalia's Flores-Figueroa concurrence, 129 S. Ct . at 1894 , which pointedly rejected Breyer's generalization about adverb effects .
66. Id . at 1855 (emphasis in original).
67. United States v. Flores-Figueroa , 129 S. Ct . 1886 , 1890 ( 2009 ).
68. Dean , 129 S. Ct at 1854 . Dean alone, in fact, nicely encapsulates opposing canons of mens rea interpretation. One the one hand, because “[s]ome indication of congressional intent, express or implied, is required to dispense with mens rea as an element of a crime,” the Court “on a number of occasions read a state-of-mind component into an offense even when the statutory definition did not in terms so provide .” Id. at 1855 (internal quotation marks omitted). On the other hand, “[w]e start, as always, with the language of the statute” and “we ordinarily resist reading words or elements into a statute that do not appear on its face .” Id. at 1853 (internal quotation marks omitted).
69. See Llewellyn, supra note 11 , at 395-96 ( famously juxtaposing a series of seemingly conflicting canons of construction).
70. For a distinction between mens rea interpretation of explicit and implicit requirements, see , e.g., United States v . Holmes , 607 F.3d 332 , 337 ( 3d Cir . 2010 ) (interpretations of statutes with “knowingly” term are “inapposite” to interpretation of an implicit mens rea requirement).
87. Id . § 860 (a).
88. A third section doubles the sentence for those who “knowingly and intentionally employ, hire [or] use . . . a person under eighteen years of age” in a drug operation . 21 U.S.C. § 861 (a) ( 2006 ). Courts require knowledge that “employs, hires or uses” another but not that the person employed is under eighteen . United States v. Frazier , 213 F.3d 409 , 418 - 19 ( 7th Cir . 2000 ) (holding that proof of defendant's knowledge of employee's age is not required); see also United States v . Cook , 76 F.3d 596 , 600 - 01 ( 4th Cir . 1996 ); United States v . Chin , 981 F.2d 1275 , 1280 (D.C. Cir . 1992 ).
89. See , e.g., United States v . Jackson , 443 F.3d 293 , 299 ( 3d Cir . 2006 ) (no knowledge of proximity to school required); United States v . Falu , 776 F.2d 46 , 49 - 50 ( 2d Cir . 1985 ) (same).
90. Falu , 776 F.2d at 49.
91. Id . at 50; see also United States v . Nieves , 608 F. Supp . 1147 , 1149 -50 (S.D.N .Y. 1985 ) (“It is difficult to imagine a more rational way of keeping drug traffickers out of areas where children are more likely to come into contact with them than to subject them to a risk of stiffer penalties for doing business near school property .”).
92. Falu , 776 F.2d at 49 -50 (citing legislative history and other arguments suggesting Congress's predominant goal was to prevent the school-zone offense).