The Constitutional Status of The

The Constitutional Status of the
Reasonable Doubt Rule

Donald A. Drippst
The Supreme Court has held that due processforbids convicting an

individual of a crime unless the government proves the elements of the
charged offense beyond a reasonabledoubt. The Court, however, permits
the government to convict even though the defendant bears the burden to
prove any affirmative defenses. The leading academic approach would
requirethe government toprove beyond a reasonabledoubt only thosefacts
sufficient to expose the defendant to criminalliability.
In this Article, ProfessorDripps argues that due process requires the
government to establish every fact that under the applicablestatutes gives
rise to a distinct range of criminalpunishments, regardlessof whether a
fact negates an element ofthe offense or establishesan affirmative defense.
He contends that due process includes the principle of legality-that no
person may be punished except for conduct that violates apreexistingpro-
vision of positive law. Because the Constitution establishes this guarantee
against the states,federal courts are responsiblefor enforcing the legality
principle. Federalism allows federal courts to prescribe the reasonable
doubt safeguardas a method ofpreventing the conviction ofpersons whose
conduct has not broken state law.
Working from this analysis, Professor Dripps argues that the Court
and prior commentators have failed to focus on the federal constitutional
right protected by the reasonable doubt standard.Prudential considera-
tions concerning the need for legislative compromises in adoptingprogres-
sive criminal law reforms do not justify underenforcement of this
constitutionalright against unlawful punishment
INTRODUCTION
Justice Frankfurter once ranked the rule requiring proof beyond a
reasonable doubt for a criminal conviction among "the boasts of a free
society."' Today, however, judicial doctrine has hedged that boast by
arbitrarily distinguishing elements of the charged offense, which the gov-
t Assistant Professor of Law, University of Illinois. B.A. 1980, Northwestern University;

J.D. 1983, University of Michigan.
I appreciate the valuable comments and advice of Wayne LaFave, R. J. Allen, Gerry Bradley,
Tony Honore, Kit Kinports, Lewis Liman, and of the participants in a faculty seminar at the
University of Illinois College of Law, in April, 1986, at which I presented a preliminary version of
this Article. Errors of thought and expression are mine, solely.
1. Leland v. Oregon, 343 U.S. 790, 803 (1952) (Frankfurter, J., dissenting).
1665
1666 CALIFORNIA LAW REVIEW [Vol. 75:1665
ernment must establish beyond a reasonable doubt, from affirmative

defenses, which the legislature may require the defendant to prove by a
preponderance of the evidence.2 "Positivist" legal scholars have attacked
this distinction because it severs legislative authority to define substantive
grounds of criminal liability from the included power to define proce-
dures for determining when those grounds exist.3 My thesis is that, con-
trary to both the Supreme Court and its positivist critics, due process of
law imposes on the government the burden of proving the defendant's
guilt beyond a reasonable doubt, whether the legislature has classified, or
might constitutionally classify, the exculpatory facts at issue as relating
to an element of the crime or to an affirmative defense.
Part I recounts the judicial development of the distinction between
elements and defenses with respect to the government's burden of proof.
Part I also describes the positivist critique of current doctrine. Using
these positions as points of departure, Part II argues that imposing crimi-
nal punishment absent conduct that has been clearly and prospectively
declared criminal is inconsistent with due process of law. Given the
imperfections of the factfinding process and the extraordinary hardship
inflicted by unjust conviction, a preponderance of the evidence standard
in criminal cases creates a constitutionally unjustifiable risk of unauthor-
ized punishment. The reasonable doubt rule guards against this risk by
resolving ambiguous cases in favor of the accused. Part II argues further
2. See Martin v. Ohio, 107 S. Ct. 1098 (1987); Patterson v. New York, 432 U.S. 197 (1977); 1
W. LAFAVE & A. ScoTr, SUBSTANTIVE CRIMINAL LAw 67-68 (1986); C. MCCORMICK,
MCCORMICK ON EVIDENCE 987 (3d ed. 1984).
3. See. eg., Allen, Mullaney v. Wilbur, the Supreme Court, and the Substantive Criminal
Law-An Examination of the Limits of Legitimate Intervention, 55 TEx. L. REV. 269 (1977)
[hereinafter Allen, Limits of Legitimate Intervention]; Allen, The Restoration of In re Winship: A
Comment on Burdens of Persuasion in Criminal Cases After Patterson v. New York, 76 MICH. L.
REv. 30 (1977) [hereinafter Allen, Burdens of Persuasion];Allen, StructuringJury Decisionmaking
in CriminalCases: A Unified ConstitutionalApproach to EvidentiaryDevices, 94 HARV. L. REV. 321
(1980) [hereinafter Allen, StructuringJury Decisionmaking]; Allen & DeGrazia, The Constitutional
Requirement of Proof Beyond Reasonable Doubt in Criminal Cases: A Comment Upon Incipient
Chaos in the Lower Courts, 20 AM. CRIM. L. REv. 1 (1982); Christie & Pye, Presumptions and
Assumptions in the CriminalLaw: Another View, 1970 DUKE L.J. 919; Jeffries & Stephan, Defenses,
Presumptions, and Burden of Proof in the Criminal Law, 88 YALE L.J. 1325 (1979). These
commentators are identified as "positivists" because they believe that substantive liability rules are
indistinguishable from procedural rules; both matter only to the extent that they may affect the
material consequences for the parties. Cf Holmes, The Path of the Law, 10 HARV. L. REV. 457, 461
(1897) ("The prophecies of what the courts will do in fact, and nothing more pretentious, are what I
mean by the law.").
For contrary views, see Nesson, Rationality, Presumptions,and Judicial Comment: A Response
to ProfessorAllen, 94 HARV. L. REV. 1574 (1981); Saltzburg, Burdens of Persuasionin Criminal
Cases: Harmonizingthe Views of the Justices, 20 AM. CRIM. L. REV. 393 (1983); Underwood, The
Thumb on the Scales ofJustice. Burdens of Persuasionin CriminalCases, 86 YALE L.J. 1299 (1977).
For an elaboration of how the argument presented here differs in important respects from those
offered by Professors Nesson, Underwood, and Saltzburg, see infra notes 201-17 and accompanying
text.
1987] REASONABLE DOUBT RULE 1667
that even when the substantive ground for exculpation claimed by the
accused is not itself required by the Constitution, due process requires
the government to define prospectively, and prove convincingly, the facts
that give rise to criminal liability. Thus, with respect to any fact that
determines criminal liability, the safeguard provided by the reasonable
doubt rule against unjust conviction is constitutionally compelled.
With the constitutional status of the reasonable doubt requirement
established by Part II, Part III returns to the case law and the commen-
tary. In light of the connection between the reasonable doubt rule and
the due process legality principle, the distinction between elements and
affirmative defenses appears untenable; the risk of unjust conviction is no
less urgent in the context of affirmative defenses. Neither deference to
historical practice nor concern for legislative flexibility can justify the
judicial distinction. Part III also suggests that the Court's critics have
failed to focus on the underlying constitutional right claimed by a
defendant invoking the reasonable doubt rule. Due process requires the
government to prove every element of the charged offense beyond a rea-
sonable doubt-and also to disprove claimed affirmative defenses beyond
a reasonable doubt when the evidence raises a reasonable doubt about the
defendant's guilt.4
I
POINTS OF DEPARTURE: THE SUPREME COURT
AND ITS CRITICS
The Supreme Court has held that due process requires the govern-
ment to bear the burden of proving beyond reasonable doubt some of the
facts necessary to establish criminal liability. But the Court also has held
that other facts necessary to establish criminal liability need not be
proved by the government. These latter cases permit shifting to the
defendant the burden of proving facts that establish affirmative defenses.
Legal scholars have criticized the Court for inconsistently forbid-
ding one method of allocating burdens of proof but upholding another
method, when the two may have indistinguishable effects. The leading
alternative approach would forbid easing the government's proof burden
only with respect to facts necessary to make a conviction constitutional.
This Part describes the conflicting doctrines endorsed by the
Supreme Court and the leading commentators. Subsequently, Part II
proposes an alternative constitutional analysis.
4. The approach I claim to be constitutionally required was followed by the federal courts, as
a matter of policy, for most of this century. See Patterson,432 U.S. at 202-03; Fletcher, Two Kinds
of Legal Rules: A Comparative Study ofBurden-of-PersuasionPracticesin CriminalCases, 77 YALE
L.J. 880, 917-18 (1968).
A. The Supreme Court and the Reasonable Doubt Rule

The Supreme Court's decisions regarding the government's burden
of proof in criminal trials fall into two basic categories. One line of cases
regulates the use of presumptions in the government's proof.' The other
line of cases regulates jury instructions that shift to the defense the bur-
den of proof with respect to particular issues.6
1. Presumptions and the Burden of Proof

A presumption-is simply an instruction that permits or requires the
jury to conclude that direct proof of one fact (the "basic fact") also pro-
vides evidence of another fact (the "presumed fact").7 Prior to 1979, it
was unclear whether due process required that the basic fact establish the
presumed fact beyond a reasonable doubt. The Court dispelled this
uncertainty in County Court v. Allen.'
In Allen, the State of New York relied on a presumption that illegal
firearms in an automobile are in the possession of each of the occupants. 9
Justice Stevens, writing for the Court, drew a distinction between
mandatory and permissive presumptions. An instruction requiring the
jury to infer the presumed fact from the basic fact-a mandatory pre-
sumption-is unconstitutional unless the connection between the basic
and presumed facts satisfies the reasonable doubt test without reference
to other evidence.10 Otherwise, the jury might convict the defendant
solely on the presumption, even though the basic fact alone does not
establish the presumed fact beyond a reasonable doubt. 1
By contrast, a permissive presumption-an instruction that merely
advises the jury that the basic fact may prove the presumed fact beyond a
reasonable doubt, but that the presumption is to be accorded just such
weight as the jury believes is justified-may survive constitutional scru-
tiny even if the connection between basic and presumed facts does not
satisfy the reasonable doubt test. Instead, a rational connection must
5. See, e.g., County Court v. Allen, 442 U.S. 140 (1979); Sandstrom v. Montana, 442 U.S. 510
(1979); Barnes v. United States, 412 U.S. 837 (1973); Turner v. United States, 396 U.S. 398 (1970);
Leary v. United States, 395 U.S. 6 (1969); United States v. Romano, 382 U.S. 136 (1965); Roviaro v.
United States, 353 U.S. 53 (1957); Tot v. United States, 319 U.S. 463 (1943).
6. See, eg., Patterson v. New York, 432 U.S. 197 (1977); Mullaney v. Wilbur, 421 U.S. 684
(1975); In re Winship, 397 U.S. 358 (1970); Morrison v. California, 291 U.S. 82 (1934).
7. See C. MCCORMICK, supranote 2, at 965, 988. For example, the jury might be allowed to
infer the defendant's intent to kill from her use of a deadly weapon.
8. 442 U.S. 140 (1979).
9. Id. at 143-45.
10. Id. at 167.
11. Id. at 157 ("Because this permissive presumption leaves the trier of fact free to credit or
reject the inference and does not shift the burden of proof, it affects the application of the 'beyond a
reasonable doubt' standard only if, under the facts of the case, there is no rational way the trier could
make the connection permitted by the inference.").
1987] 9]REASONABLE DOUBT RULE 1669
exist between the basic and presumed facts, and the evidence in its total-
ity must establish guilt beyond a reasonable doubt.12 The legislatively or
judicially noticed connection between the basic and presumed facts
becomes part of the evidence. If the evidence proves the elements of the
charge beyond a reasonable doubt in a particular case, the failure of the
basic fact alone to prove the presumed fact beyond a reasonable doubt is
13
irrelevant.
If the reasonable doubt rule is constitutionally required, the Allen
approach makes good sense. But the presumption cases assume, rather
than announce, such a constitutional rule.14 If the government need only
prove a particular fact by a preponderance of the evidence, there is no
reason to inquire whether the presumption, with or without other evi-
dence, satisfies the reasonable doubt test. 5 The more basic question,
then, is whether the reasonable doubt rule applies to a particular issue.
The second line of Supreme Court cases, those involving direct attempts
to assign the burden of proof to the defense, confronts this question.
2. Shifting the Burden of Proof

The seminal case in this line of authority is In re Winship. 6 Samuel
Winship was a defendant in a delinquency proceeding in a New York
family court. The complainant, Rae Goldman, testified that she had seen
Winship run out of an employee bathroom at the store where she
17
worked, after a co-worker had discovered Winship's presence there.
Mrs. Goldman then checked the employee locker room for her handbag
and found that it was missing; the bag, less $112 in cash, was discovered
in the employee bathroom from which Winship allegedly fled. Two days
later, police arrested Winship on an unrelated charge. Mrs. Goldman,
who claimed to have seen Winship sneaking about the store where she
worked on several previous occasions, identified him at the police station.
No one saw Winship steal the money, and the defense offered alibi testi-
mony from Winship's mother and his uncle. When arrested, Winship
12. Id. at 164-67.

13. Ial at 167 ("There is no more reason to require a permissive statutory presumption to meet
a reasonable-doubt standard before it may be permitted to play any part in a trial than there is to
require that degree of probative force for other relevant evidence before it may be admitted.").
Applying, this analysis, the Allen Court upheld the defendants' convictions. Presence in the
automobile, the Court concluded, was rationally connected to possession of the weapons. Id. at 163-
65. Given the rational connection, the Court refused to disturb the lower court's holding that the
evidence in its entirety, including the presumption, could have led a rational jury to find guilt beyond
reasonable doubt. Id at 164.
14. See, eg., Walker v. Butterworth, 599 F.2d 1074, 1079-80 (Ist Cir. 1979).
15. See id.; Greider v. Duckworth, 701 F.2d 1228, 1235-36 (7th Cir. 1983) (Posner, J.,
concurring).
16. 397 U.S. 358 (1970).
17. These facts are drawn from In re Winship, 397 U.S. 358 app. at 3-34 (1970) (Record of the
New York Family Court) (microfiche).
was in possession of several rolls of dimes. His mother could not explain
his possession of the dimes, but only paper money had been stolen from
Mrs. Goldman.
The family court judge overruled a defense objection to the civil pre-
ponderance of the evidence standard applied in the proceeding. The
judge then found Winship guilty under the preponderance standard, but
conceded that the evidence did not convince him of Winship's guilt
beyond a reasonable doubt."8 The Appellate Division affirmed the judg-
ment, as did the Court of Appeals, over the dissent of Chief Judge
Fuld. 19
The Supreme Court reversed, holding that the juvenile court pro-
ceeding was criminal in character and that, in criminal proceedings, "the
Due Process Clause protects the accused against conviction except upon
proof beyond a reasonable doubt of every fact necessary to constitute the
crime with which he is charged." 2 0 Writing for the Court, Justice Bren-
nan based this interpretation of the due process clause on the widespread
acceptance of the reasonable doubt rule in common law jurisdictions, in
the history of American jurisprudence, and, at least implicitly, in
Supreme Court precedents.2 1
More fundamentally, the Winship Court recognized two characteris-
tics of the criminal process as supporting a constitutional reasonable
doubt rule. The first is the difficulty of defending against a charge of
crime, a "disadvantage" that would amount to a denial of" 'fundamental
fairness, if [the defendant] could be adjudged guilty and imprisoned for
years on the strength of the same evidence as would suffice in a civil
case.' "22 The reasonable doubt standard thus provides "a prime instru-
ment for reducing the risk of convictions resting on factual error."2 3
The second aspect of the criminal process on which Winship relied is
the difference in consequences attending erroneous acquittals and errone-
ous convictions.
Where one party has at stake an interest of transcending value-as a
criminal defendant his liberty-[the] margin of error is reduced as to him
by the process of placing on the other party the burden of... persuading
the factfinder
24
at the conclusion of the trial of his guilt beyond a reason-
able doubt.
18. Id., app. at 28.

19. In re Samuel W., 30 A.D.2d 781, 291 N.Y.S.2d 1005 (1968), aff'd, 24 N.Y.2d 196, 247
N.E.2d 253, 299 N.Y.S.2d 414 (1969), rev'd sub nom. In re Winship, 397 U.S. 358 (1970).
20. Winship, 397 U.S. at 364.
21. Id. at 361-64.
22. Id. at 363 (quoting Samuel W, 24 N.Y.2d at 205, 247 N.E.2d at 259, 299 N.Y.S.2d at 422
(Fuld, C.J., dissenting)).
23. Id.
24. Id. at 364 (quoting Speiser v. Randall, 357 U.S. 513, 525-26 (1958)).
1987] REASONABLE DOUBT RULE
In a concurring opinion, Justice Harlan stressed the connection between

the reasonable doubt standard and "a fundamental value determination
of our society that it is far worse to convict an innocent man than to let a
guilty man go free." 25 The Winship holding thus appears to rest in part
on the reasonable doubt rule's contribution to preventing unjust convic-
tions, and in part on the recognition of that contribution in the Anglo-
American legal tradition.
The Winship opinion, however, explicitly extends only to "every fact
necessary to constitute the crime" charged.2 6 The Court also cited
Leland v. Oregon,2 in which the Court had held that the state might
constitutionally shift to the accused the burden of proving insanity.
Winship thus clearly grounded the reasonable doubt rule on the due pro-
cess clause but left open the precise scope of the constitutional guarantee.
Since the factual issue in Winship was the identity of the defendant, the
Court did not address the constitutionality of easing the government's
burden of proof with respect to either affirmative defenses or facts that
affect the grade of the offense.
In Mullaney v. Wilbur,2 8 the Court interpreted Winship to bar a
murder conviction based on a state procedure that placed the burden of
proving provocation, which would have reduced the offense from murder
to manslaughter, on the accused. The defendant, Stilman Wilbur, admit-
ted killing one Claude Hebert, but claimed that Hebert had provoked the
killing by inviting Wilbur to participate in homosexual activity.29 The
Maine trial judge instructed the jury that once the government had
proved intentional killing beyond a reasonable doubt, the law would pre-
sume the presence of malice. To rebut the presumption of malice, the
defense had to satisfy the jury'by a preponderance of the evidence that
the accused had killed in the heat of passion.3 ° The jury twice requested
reinstruction on the provocation issue before finding Wilbur guilty of
murder.3
The Supreme Judicial Court of Maine rejected Wilbur's claim that
25. Id at 372 (Harlan, J., concurring).

26. Id. at 364.
27. 343 U.S. 790 (1952), cited at Winship, 397 U.S. at 362.
28. 421 U.S. 684 (1975).
29. These facts are derived from the federal court proceeding for habeas corpus relief. See
Wilbur v. Robbins, 349 F. Supp. 149, 151 (D. Me. 1972), aff'd sub nom. Wilbur v. Mullaney, 473
F.2d 943 (1st Cir. 1973), vacated, 414 U.S. 1139, aff'd on remand, 496 F.2d 1303 (1st Cir. 1974),
aff'd, 421 U.S. 684 (1975).
30. See id. at 151 (reproducing trial judge's charge).
31. See id. ("The trial judge's initial charge and two supplementary charges, given when the
jury returned to ask for further instructions, included repeated references to the mandatory nature of
the presumption of malice and the burden on the defendant to rebut the presumption in order to
'reduce' the homicide from murder to manslaughter.").
Winship required reversal of the conviction.32 The court based its hold-
ing in part on the theory that Winship did not apply retroactively, a the-
ory subsequently rejected by the Supreme Court. 33 The state court also
relied on the structure of the applicable law of homicide. For more than
a century, Maine had classified murder and manslaughter as different
degrees of the single offense of felonious homicide.34 The provocation
defense raised by Wilbur thus implicated the appropriate sentence rather
than the adjudication of guilt or innocence. Since sentencing determina-
tions are not subjected to the reasonable doubt test, the Maine court con-
cluded that Winship did not apply. According to the Maine court, the
reasonable doubt standard governs the proof required to place the
defendant in a formal statutory category, but does not extend to facts
affecting the scope of liability within such categories.3
On Wilbur's petition for federal habeas corpus relief, the district
court and the First Circuit rejected the Maine court's construction of the
state homicide statute.36 The federal courts found that malice was an
element of a distinct crime of murder and vacated Wilbur's conviction.
While the state's petition for certiorari was pending, the Maine Supreme
Judicial Court in State v. Lafferty37 reaffirmed the reading it had given
the homicide statute in Wilbur. 8 In light of the new Maine ruling, the
United States Supreme Court remanded Wilbur.39 The First Circuit
again concluded that the writ of habeas corpus should issue; whether
provocation implicated an element of the crime or affected only the
applicable penalty was immaterial. The government could not shift to
the defense the burden of proof on an issue with such serious
consequences. 4°
A unanimous Supreme Court affirmed, in an opinion by Justice
Powell. 4 ' In holding that due process does not permit shifting to the
defense the burden of proving the absence of malice, the Court observed
that a murder conviction results in a more serious deprivation of liberty
and of reputation than does a manslaughter conviction. Since Winship
32. State v. Wilbur, 278 A.2d 139, 146 (Me. 1971).

33. See Ivan V. v. City of New York, 407 U.S. 203, 204-05 (1972).
34. See, eg., State v. Rollins, 295 A.2d 914, 918 (Me. 1972); State v. Knight, 43 Me. 11, 33,
137 (1857), quoted in Brine v. State, 264 A.2d 530, 533-34 (Me. 1970).
35. See Wilbur, 278 A.2d at 146; see also Allen, Limits ofLegitimate Intervention, supranote 3,
at 272-73.
36. Wilbur v. Robbins, 349 F. Supp. 149, 151 (D. Me. 1972), aff'd sub nom. Wilbur v.
Mullaney, 473 F.2d 943 (Ist Cir. 1973), vacated,414 U.S. 1139, aff'd on remand, 496 F.2d 1303 (1st
Cir. 1974), aff'd, 421 U.S. 684 (1975).
37. 309 A.2d 647 (Me. 1973).
38. Id. at 661-62.
39. 414 U.S. at 1139.
40. 496 F.2d at 1307.
41. Mullaney v. Wilbur, 421 U.S. 684 (1975).
19871 REASONABLE DOUBT RULE 1673
involved these same interests, the Court rejected the state's contention
that due process requires proof beyond a reasonable doubt only of the
elements of a unified "felonious homicide" offense. "Winship," the Court
declared, "is concerned with substance rather than this kind of
formalism."'4 2
The Court also rejected the argument that applying Winship to the
Maine statutory scheme would unduly interfere with the state's power to
define the substantive grounds of criminal liability. Accepting this argu-
ment, Justice Powell wrote, would invite the states to circumvent
Winship by redefining crimes in the broadest possible terms, shifting the
burden of proof to the defendant with respect to a host of newly denomi-
nated "mitigating factors."4 3 On the other hand, the Wilbur opinion rec-
ognized that inflexibly requiring proof beyond a reasonable doubt might
discourage the states from establishing new defenses. To guard against
this danger and to limit federal intervention to more fundamental
grounds of exculpation, the reasonable doubt standard was to govern
only the proof of historically established defenses or mitigating circum-
stances. Wilbur permitted the states to define crimes as they wished but
forbade the states from circumventing Winship by recasting essential and
well-understood elements of offenses carried over from the common
law. 44
The Wilbur approach thus involves both substantive and procedural
components. As a procedural matter, the government must prove
beyond a reasonable doubt every fact necessary to establish a charged
offense. As a substantive matter, the facts necessary to establish an
offense are not wholly subject to legislative modification. When a fact
has played a central role in the historical development of a crime's defini-
tion, the burden of proving that fact or its absence cannot be shifted to
the defense.
Wilbur fostered a great deal of confusion in the lower courts. 4 This
confusion resulted in part from ambiguities in the Court's opinion. But
the unhappy history of the Wilbur doctrine is in large part due to the
inherent difficulties in determining when a legal concept has become suf-
ficiently entrenched in the Anglo-American legal tradition. The defer-
ence due to history is more easily determined by a supreme court in
a single case than by trial courts in hundreds of diverse cases. 46 For
42. Id. at 699.

43. Id. at 698 ("It would only be necessary to redefine the elements that constitute different
crimes, characterizing them as factors that bear solely on the extent of punishment.").
44. Id. at 692-96.
45. See 1 W. LAFAVE & A. ScoTr, supra note 2, at 73; Allen, Limits of Legitimate
Intervention, supra note 3, at 274-76.
46. As an illustration, Justice Powell's historical analysis in Wilbur is difficult to follow.
Initially, the Court noted that "almost from the inception of the common law of homicide,"
example, the concepts of intent and insanity often, but not always,
exclude each other.4 7 The Wilbur doctrine unquestionably would pre-
clude a state from convicting a defendant because she failed to establish
absence of intent. Yet by holding that Wilbur did not forbid shifting to
the accused the burden of proving insanity,4" the lower courts endorsed
the continued vitality of Leland v. Oregon.4 9 When one such case
reached the Supreme Court, the appeal was dismissed for failure to pres-
ent a substantial federal question.5
The emphasis on history, moreover, ran the risk of freezing the
criminal law in familiar forms.5 " Wilbur appeared to imply that the
states could not constitutionally abolish the manslaughter classifica-
tion." Yet this implication was not justified by reliance on the eighth
amendment's, cruel and unusual punishment clause or on a substantive
interpretation of the due process clause. Perhaps most disturbingly,
Wilbur appeared to prohibit legislative compromises recognizing new
provocation has been the "single most important factor in determining the degree of culpability
attaching to an unlawful homicide." Wilbur, 421 U.S. at 696. But at common law, the defense bore
the burden of proof with respect to provocation. Id. at 693. The historical support for holding the
reasonable doubt rule applicable to provocation was that "the clear trend" in twentieth century state
practice was toward this result. Id at 696. Justice Powell did not explain why common law practice
informs the constitutional inquiry as to the defenses for which the federal courts may properly
formulate rules of proof, but contemporary state practice governs the content of the rules which the
federal courts impose. Nor did the opinion suggest how lower courts should apply common law and
modern practice in subsequent cases. Moreover, Wilbur presented a relatively pristine variation on
Winship, namely, a traditional definition of substantive offenses and an aberrant allocation of proof
burdens. But what of a jurisdiction with a subjective view of provocation or self-defense? The
accused in such jurisdictions asserts a claim that was not cognizable at common law. Fidelity to the
Wilbur approach would seem to require the prosecution to disprove objectively reasonable self-
defense or provocation beyond a reasonable doubt, but allow the state to shift the burden with
respect to parallel subjective claims-claims that might often be plausible in the same case. A trial
judge forced to apply a historical constitutional analysis to a statute that shifts the burden of proof
without discriminating on a historical basis among substantive claims would have to devise a jury
instruction similar to a Chinese box.
47. As Judge Posner has astutely noted, if the defendant is "under the delusion that he was
shooting two gerbils rather than two human beings, he could not be guilty of murder, but if his
delusion took the form of thinking that he had a sacred duty to reduce the human population by two,
he could be guilty of murder, at least guilty prima facie, though he might have a defense of insanity."
Greider v. Duckworth, 701 F.2d 1228, 1236 (7th Cir. 1983) (Posner, I., concurring).
48. See, eg., Buzynski v. Oliver, 538 F.2d 6, 10 (1st Cir.), cert. denied, 429 U.S. 984 (1976);
United States v. Caldwell, 543 F.2d 1333, 1368-71 (D.C. Cir. 1974), cert. denied, 423 U.S. 1087
(1976).
49. 343 U.S. 790 (1952); see supra text accompanying note 27.
50. Rivera v. Delaware, 429 U.S. 877 (1976).
51. See Jeffries & Stephan, supra note 3, at 1363-64.
52. See Tushnet, ConstitutionalLimitation of Substantive Criminal Law: An Examination of
the Meaning of Mullaney v. Wilbur, 55 B.U.L. REV. 775, 776 (1975) (Wilbur "stands for the
proposition that the due process clause limits governmental power to define crimes in whatever
manner desired"). The Court's reliance on the central role of provocation in the historical definition
of murder would be relevant only if historical importance qualifies an issue for enhanced
constitutional protection.
grounds of exculpation and mitigation on which the defense would bear

the burden of proof.
Notwithstanding the Court's unanimity, Wilbur did more to raise
than to settle these difficult questions about the scope of the Winship
doctrine. Two years later, Patterson v. New York 53 raised these same
questions in a slightly different context and provided some very different
answers.
Gordon Patterson, Jr. killed John Northrup after discovering him in
the company of Patterson's partially undressed wife. 54 Patterson con-
fessed to the killing and relied at trial on the defense of provocation.
New York had recently revised its criminal code, adopting the Model
Penal Code formulation of this defense. The New York statute pre-
cluded a conviction for murder, but not for manslaughter, if the defend-
ant "acted under the influence of extreme emotional disturbance for
which there was a reasonable explanation or excuse."5 The state, how-
ever, provided this defense to a murder charge only when the defendant
established explicable emotional disturbance by a preponderance of the
evidence.5 6 The trial court instructed the jury according to the statute,
and Patterson was found guilty of murder.
Affirming the New York Court of Appeals, which upheld the con-
viction despite Patterson's reliance on Wilbur,57 the Supreme Court held
that due process does not forbid placing on the accused the burden of
proving affirmative defenses by a preponderance of the evidence. The
Court distinguished Wilbur as a case in which the defendant bore the
burden of proving the absence of malice aforethought, an element of the
crime of murder.5 8 In contrast, the New York statute did not require
proof of malice but defined murder as any intentional killing. Since the
prosecution had proved the intentional killing beyond a reasonable
doubt, the government had sustained its constitutional duty, under Win-
ship, of proving " 'beyond a reasonable doubt.., every fact necessary to
constitute the crime with which [Patterson was] charged.' ,
The majority buttressed this formal distinction by emphasizing the
primary role of the states in defining crimes,' and by affirming the state's
authority to recognize some mitigating circumstances without assuming
the burden of disproving their presence beyond a reasonable doubt.6 1 As
53. 432 U.S. 197 (1977).

54. See id. at 198.
55. Id.; see also id. at 198 n.2, 199 n.3 (quoting sections of the statute).
56. Id. at 200.
57. See People v. Patterson, 39 N.Y.2d 288, 347 N.E.2d 898, 383 N.Y.S.2d 573 (1976).
58. 432 U.S. at 212-16.
59. Id. at 204 (quoting In re Winship, 397 U.S. 358, 364 (1970).
60. See icL at 207-08.
61. Id. at 210 ("We thus decline to adopt as a constitutional imperative, operative
for the risk of convicting for murder those in fact guilty only of man-
slaughter, the Court explained that "[d]ue process does not require that
every conceivable step be taken, at whatever cost, to eliminate the possi-
bility of convicting an innocent person."'62
Justice Powell, the author of the Wilbur opinion, dissented in an
opinion joined by two other Justices. 61 In his view, Wilbur had estab-
lished a substantive constraint on state power that precluded defining
crimes so as to circumvent the rule of Winship. He urged the Court to
forbid shifting the burden of proof by rearranging historically recognized
fundamental concepts but not to forbid defining new crimes or
defenses."
Patterson suggests that the constitutionality of burden shifting turns
solely on legislative form and not on the moral or legal significance of the
defendant's claim. 6' The Court's most recent pronouncement on the
question, Martin v. Ohio,6 6 reaffirms, and indeed extends, the formalistic
approach taken in Patterson. In Martin, the Court held that the state
may shift the burden of proving self-defense to one accused of homi-
cide. Self-defense is the defense most deeply rooted in our history and
shared morality. If due process or the eighth amendment requires state
law to recognize any exculpatory doctrine at all, self-defense would
surely rank among those required. The Martin holding, however, indi-
cated that the legislature, by classifying an exculpatory doctrine as an
affirmative defense, may shift the burden of proof even with respect to a
constitutionally required defense firmly rooted in the Anglo-American
legal tradition.
countrywide, that a State must disprove beyond a reasonable doubt every fact constituting any and
all affirmative defenses related to the culpability of an accused.").
62. Id. at 208.
63. Id. at 216 (Powell, J., dissenting).
64. Id. at 226-27 (Powell, J., dissenting) ("[A] substantial difference in punishment alone is not
enough. It also must be shown that in the Anglo-American legal tradition the factor in question
historically has held that level of importance. If either branch of the test is not met, then the
legislature retains its traditional authority over matters of proof.") (footnotes omitted).
65. See Jeffries & Stephan, supranote 3, at 1331 ("The trouble, of course, is that the distinction
is essentially arbitrary."). The authors elaborate:
Traditionally, the only functional difference between a "crime" and a "defense" has
been precisely the issue under consideration-allocation of the burden of proof.... To
make the scope of that doctrine depend on the legislative allocation of the burden of proof
is to assume the point in issue and thus to reduce Winship to a circularity.
Id at 1332. Justice Rehnquist has embraced this formalistic analysis with gusto. In his view,
Winship applies, under Patterson, only to elements; classification of elements and defenses depends
on legislative intent, and the surest guide to legislative intent with respect to classification as an
affirmative defense is a legislative directive shifting the burden of proof. See McElroy v. Holloway,
451 U.S. 1028, 1028-29 (1981) (Rehnquist, J., dissenting from denial of certiorari). This approach
would require overruling Wilbur explicitly and limiting Winship to its facts as a substantive holding
that the state must prove identity beyond a reasonable doubt in juvenile delinquency proceedings.
66. 107 S. Ct. 1098 (1987).
67. Id. at 1103.
19871 REASONABLE DOUBT RULE 1677
Thus the Martin majority did not merely reject the Wilbur
approach, which Justice Powell defended again in dissent. 68 It also
rejected the possibility that the reasonable doubt rule might apply to
constitutionally required defenses without regard to legislative
classification.6 9
If the due process clause speaks about individual rights against the
state, in individual terms, this distinction between elements of a charged
offense and affirmative defenses is plainly arbitrary. But the legislative
classification of exculpatory facts as elemental or defensive corresponds
with one substantive factor, and that is legislative deliberation itself. The
clarity with which a legislature identifies exculpatory conditions as
affirmative defenses will likely vary with the legislative purpose to shift
the burden of proof. The Court's rule thus has the virtue of permitting
deliberate legislative compromises that balance substantive expansion of
exculpatory conditions against procedural concessions to the prosecu-
tion. Put another way, the Patterson holding treats Wilbur and Winship
as permitting the legislature to shift the burden of proof as long as it does
so knowingly. Such a requirement strikes a coherent, if contestable, bal-
ance between the risk of convicting innocent individuals and the risk that
broader grounds of exculpation will go unrecognized because of per-
ceived proof problems for the prosecution.7"
B. The Positivist Commentary

Wilbur was criticized vigorously for unduly interfering with state
authority to define criminal offenses. The basic argument, made by sev-
eral eminent commentators,7 1 is that the legislature's power to define
substantive grounds of criminal liability includes the power to specify the
procedures that determine when those grounds exist. The positivists
claim that if the Constitution permits the legislature to abolish a defense,
68. Id. at 1106 (Powell, J., dissenting).

69. Although Martin clearly pressed the eighth amendment argument, see Brief for Petitioner
at 26-29, id (No. 85-646), the Court's opinion does not address the issue, see Martin, 107 S. Ct. at
1106 (Powell, J., dissenting) ("Even Patterson ... recognized that 'there are obviously constitutional
limits beyond which the States may not go [in labeling elements of a crime as an affirmative
defense].' Today, however, the Court simply asserts that Ohio law properly allocates the burdens,
without giving any indication of where those limits lie.") (citation and footnote omitted). Cf Moran
v. Ohio, 469 U.S. 948, 953-56 (1984) (Brennan, J., dissenting from denial of certiorari) (due process
and Eighth Amendment require exculpation for pure self-defense).
70. Martin suggests that this concern operates regardless of whether the defense is defined by
legislation or adjudication, inasmuch as self-defense in Ohio is a matter of common law. See State v.
Martin, 21 Ohio St. 3d 91, 94, 488 N.E.2d 166, 168 (1986), aff'd, 107 S. Ct. 1098 (1987). For my
critique of the legislative compromise argument, see infra text accompanying notes 189-200.
71. See sources cited supra note 3. In a recent opinion, Judge Easterbrook presented a
thorough defense of the positivist position. Cole v. Young, 817 F.2d 412, 428-40 (7th Cir. 1987)
(Easterbrook, J., dissenting).
such as provocation or insanity, then the legislature may also take the
more modest step of shifting the burden of proof.
This analysis parallels the familiar positivist position regarding pro-
cedural due process cases involving nonconstitutional entitlements.7 2
The greater power of eliminating the defense is seen as including the
lesser power of shifting the burden of proof. It is no surprise that Justice
Holmes, author of McAuliffe v. Mayor of New Bedford,73 and Justice
Rehnquist, author of Arnett v. Kennedy,74 have found this position
attractive.
According to the positivists, federal courts properly may apply con-
stitutional scrutiny to procedural rules only when federal law also gov-
erns the underlying substantive issue.75 Punishing as a murderer one who
killed only after extreme provocation, or in objectively reasonable self-
defense, may violate substantive due process or the eighth amendment's
cruel and unusual punishment clause.76 Where the Constitution requires
72. See, eg., Goldberg v. Kelly, 397 U.S. 254, 274-79 (1970) (Black, J., dissenting);
Easterbrook, Substance and Due Process, 1982 Sup. Cr. REv. 85; Smolla, The Right-Privilege
Distinction in ConstitutionalLaw: The Price of Protesting Too Much, 35 STAN. L. REV. 69 (1982).
Cf Reich, The New Property, 73 YALE L.J. 733 (1964); Van Alstyne, The Demise of the Right-
Privilege Distinctionin ConstitutionalLaw, 81 HARv. L. REV. 1439 (1968). It would be tempting to
apply the rejection of the right-privilege distinction in the entitlement cases to the problem of proof
burdens regarding gratuitous defenses. There are, nonetheless, two reasons to avoid this approach.
The first is the Court's insistence, made explicit in Gerstein v. Pugh, 420 U.S. 103, 125 n.27 (1975),
on treating the criminal process differently from civil entitlements, even when that means extending
"less procedural protection to an imprisoned human being than is required to test... the custody of
a refrigerator," id. at 127 (Stewart, J., dissenting) (citation omitted). The second reason is that,
although the Court unambiguously rejected the right-privilege distinction in Cleveland Bd. of Educ.
v. Loudermill, 470 U.S. 532, 538-41 (1985), that rejection rested more on fiat than on persuasive
reasoning. The Court has neither expressly adopted a dignitary theory of due process nor explained
why it rejects the right-privilege distinction. Grounding a constitutional case for the reasonable
doubt rule on the procedural due process cases would therefore overcome the positivist argument
only by reference to its irrational rejection in another context--scarcely an enduring foundation for a
rule of constitutional law. Should the Court ever articulate a principled reason for rejecting the
positivist account in the entitlement cases and look more favorably on applying the entitlement cases
to the criminal context, the argument presented here, infra text accompanying notes 113-25, would
justify a general reasonable doubt rule under the analysis prescribed by Mathews v. Eldridge, 424
U.S. 319 (1976). Cf. Santosky v. Kramer, 455 U.S. 745 (1982) (applying Eldridge to require clear
and convincing evidence in child custody termination proceedings).
73. 155 Mass. 216, 220, 29 N.E. 517, 517 (1892) ("The petitioner may have a constitutional
right to talk politics, but he has no constitutional right to be a policeman."); see also Ferry v.
Ramsey, 277 U.S. 88, 94 (1928) (Holmes, J.) ("The statute in short imposed a liability that was less
than might have been imposed, and that being so, the thing to be considered is the result reached,
not the possibly inartificial or clumsy way of reaching it.").
74. 416 U.S. 134, 153-54 (1974) (plurality opinion); see also McElroy v. Holloway, 451 U.S.
1028, 1028 (1981) (Rehnquist, J., dissenting from denial of certiorari).
75. See Allen, Structuring Jury Decisionmaking, supra note 3, at 342-48; Jeffries & Stephan,
supra note 3, at 1365.
76. See Moran v. Ohio, 469 U.S. 948, 953-56 (1984) (Brennan, J., dissenting from denial of
certiorari); Engle v. Isaac, 456 U.S. 107, 122 n.22 (1982); Jeffries & Stephan, supra note 3, at 1366-
79.
the state to provide a particular defense, the positivists would also

require the government to disprove the defense beyond a reasonable
77
doubt.
The positivist approach plainly conflicts with Patterson. Many issues
defined as elements of an offense are not constitutionally compulsory,
and due process or the eighth amendment may require at least some
affirmative defenses. 78 The positivists make a strong claim that their
approach offers a more coherent doctrine than does the Court's formalis-
tic distinction between elements and defenses. Yet the positivists do not
overcome the fundamental problem with Patterson,which is the inability
to explain Winship.
Neither the Patterson majority nor the positivist commentators have
offered a convincing reason to hold that due process ever requires proof
of guilt beyond a reasonable doubt. Justice White was surely right in
saying that the due process clause does not require every conceivable step
to prevent convicting the innocent, 79 for only abolition of the criminal
law could satisfy that demand. But the Patterson opinion did not explain
why the Constitution requires one conceivable check on convicting the
innocent-proof beyond a reasonable doubt for elements-but not
another conceivable check-similar disproof of affirmative defenses. Nor
have the positivists explained why due process requires the reasonable
doubt standard when the Constitution requires the state to recognize an
exculpatory condition. In analogous situations, the Constitution affirma-
tively and unambiguously forbids unreasonable searches and self-incrimi-
nation, but the preponderance standard is sufficient to determine whether
a search was unreasonable ° or a confession involuntary. 1
If Winship was wrongly decided, the issue is far simpler than the
Court and the positivists have made it. Absent Winship, the states would
never face a constitutional requirement of proof beyond a reasonable
doubt. But if Winship was correctly decided, then both the courts and
the leading commentators have gone seriously astray.
In Part II, I offer a constitutional justification for the Winship hold-
ing. Part III then considers how this justification bears on the prevailing
judicial and academic conceptions of the reasonable doubt rule.

78. See, e.g., Martin v. Ohio, 107 S. Ct. 1098 (1987); Adkins v. Bordenkircher, 674 F.2d 279,
282 (4th Cir.) (state must disprove alibi beyond a reasonable doubt), cert. denied, 459 U.S. 853
(1982).
79. Patterson v. New York, 432 U.S. 197, 210 (1977); see supra note 61.
80. United States v. Matlock, 415 U.S. 164, 177 n.14 (1974).
81. Colorado v. Connelly, 107 S. Ct. 515, 523 (1986); Lego v. Twomey, 404 U.S. 477 (1972).
II
WINSHIP REVISITED: THE REASONABLE DOUBT RULE
AND THE LEGALITY PRINCIPLE
This Part argues in favor of a constitutional reasonable doubt rule at

criminal trials. My premise is that due process incorporates a legality
principle, according to which the government may not punish one for
crime unless that person's conduct violates a contemporary provision of
positive law. Since the Constitution provides this protection against state
governments, the federal courts must determine appropriate means of
enforcement. The reasonable doubt rule, in both substantive and proce-
dural forms, serves as a legitimate safeguard against unjust conviction.
A. The FederalRight Against Unlawful State Convictions

The due process clause of the fourteenth amendment is the most
controversial provision in the Constitution. The reasonable doubt rule's
constitutional status does not depend, however, on any strained or open-
ended reading of that amendment. Indeed, if due process has any techni-
cal content narrower than a roving judicial commission to war on injus-
tice, that technical content has no clearer component than the legality
principle. The text and history of the due process clause, the provision of
other procedural safeguards in the Bill of Rights, and several long lines
of Supreme Court precedents support this interpretation.
The framers of the fourteenth amendment transplanted the due pro-
cess language from the fifth amendment. 2 The framers of the fifth
amendment took the language from the Magna Charta's Chapter 39, as
interpreted by Coke and Blackstone.8 3 Whatever significance the Great
82. The use of identical language suggests as much, and those who ratified the amendment
could not have understood otherwise. See R. Mo'r, DUE PROCESS OF LAW 163 (1926). Whether
those who drafted the amendment entertained a clear idea of what "due process of law" entailed is
less clear. See J. JAMES, THE FRAMING OF THE FOURTEENTH AMENDMENT 86-87 (1956)
(discussing Rep. Bingham's confusion); R. Ma-rr,supra, at 165 ("[Tihe members of Congress, with
the exception of a very few first-rate constitutional lawyers in that body, really had no definite
conception of what the phrase meant.").
83. See A. HOWARD, THE ROAD FROM RUNNYMEDE 298-315 (1968); H. TAYLOR, THE
ORIGIN AND GROWTH OF THE AMERICAN CONSTITUTION 236-37 (1911). Dean Pound translated
Chapter 39 of Magna Charta as follows:
No freeman shall be arrested, or detained in prison, or deprived of his freehold, or
outlawed, or banished, or in any way molested; and we will not set forth against him, nor
send against him, unless by the lawful judgment of his peers and by the law of the land.
R. POUND, THE DEVELOPMENT OF CONSTITUTIONAL GUARANTEES OF LIBERTY 123 (1957). One
of the abuses inspiring Chapter 39 was disregard of the legality principle. See W. McKECHNIE,
MAGNA CHARTA 376-77 (1905); R. MOT, supranote 82, at 33 ("particular abuse which the section
aims to curb is that of execution before judgment").
Coke, in his Second Institute, published in 1642, equated "by the law of the land" with "due
process of law." 2 E. COKE, INSTITUTES OF THE LAW OF ENGLAND *50. Habeas corpus, available
at equity if the law courts were unavailable, both effected the guarantee and confirmed its nature.
"By these writs it manifestly appeareth, that no man ought to be imprisoned, but for some certain
Charter had in 1215, the American colonists revered its guarantees as

they were reaffirmed during the English Revolution of 1688. The con-
temporary meaning of Chapter 39 was that ascribed to it by Coke and
Blackstone, according
84
to which individuals could not be punished except
as provided by law.
The early American authorities confirm this reading. Kent 85 and
Story8 6 followed Blackstone in reading due process of law to forbid
detention that was not authorized by law, and to permit conviction only
after an indictment and trial by jury. The Supreme Court's first full-
fledged due process case, Murray'sLessee v. Hoboken Land andImprove-
ment Co.,87 approved this interpretation; the legislature could not alter
cause: and these words, ad sujiciend. et recipiend. &c. prove that cause must be shewed: for other-
wise how can the court take order therein according to law?" Id at *53. The requirement of legal
cause could be satisfied by a bill of attainder (a concept repugnant to the founders, see U.S. CONST.
art. I §§ 9-10, but not unknown in post-revolutionary America) but Coke took pains to note that no
penalty could attach before the attainder. Id at *48. By Blackstone's time the procedural guaran-
tees against unjust punishment had been expanded, but the central principle that a person's liberty
could be terminated only for violations of law established by regular procedures remained. See IW.
BLACKSTONE, COMMENTARIES *133-34, *136.
84. See R. POUND, supra note 83, at 109-10 [arguing for open-ended interpretation]:
But the terms "liberty" and "due process" were legal terms with well-understood meanings
known to lawyers; and when one notes the names of the great lawyers who signed the
original Constitution-one future Chief Justice and two future Justices of the Supreme
Court of the United States, a Chief Justice of New Jersey, and many who were then
accounted leaders of the profession, it is idle to assume that they did not know the
significance of the words they used.
H. TAYLOR, supra note 83, at 79-80, 237, objected to undue emphasis on the weight accorded Coke's
views by the founders, noting that Coke's Second Institute was written during the reign of James I,
when the Court of Star Chamber was still unchecked. Taylor argued that the broader conception of
due process articulated by Blackstone more closely approximates the framers' vision. Id But Coke
and Blackstone did not disagree about the incorporation of the legality principle in Magna Charta.
See supra note 83.
85. 2 J. KENT, COMMENTARIES ON AMERICAN LAW *12-13.
86. 2 J. STORY, COMMENTARIES ON THE CONSTITUTION OF THE UNITED STATES 565-67 (5th
ed. Boston 1891) (1st ed. Boston 1833).
87. 59 U.S. (18 How.) 272 (1855). Murray'sLessee involved the constitutionality of a distress
warrant for the seizure of property owned by a delinquent taxpayer. The Court, per Justice Curtis,
rejected the notion that the legislature could make "any process 'due process of law,' by its mere
will." Id. at 276. Instead, due process incorporated those procedures recognized at common law;
since the distress warrant had analogues in English common law, its use did not violate due process.
Id. at 277-80. The reasoning applied in Murray's Lessee could neither protect individuals from new
forms of oppression nor permit benign departures from common law procedures. For this reason,
the historical approach gave way to less restrictive interpretive techniques. See Hurtado v.
California, 110 U.S. 516 (1884) (upholding for the first time an indictment on information rather
than by the traditional grand jury); see also, Twining v. New Jersey, 211 U.S. 78, 101 (1908) ("It
does not follow, however, that a procedure settled in English law at the time of the emigration ... is
an essential element of due process of law."). Departure from historical standards in favor of natural
law interpretations invites the objection that the judges are indulging contemporary, idiosyncratic,
political preferences rather than applying an authoritative legal text. Cf.Lochner v. New York, 198
U.S. 45 (1905). This choice between inherited and contemporary arbitrariness remains at the heart
of most due process cases. See generally Kadish, Methodology and Criteria in Due Process
Adjudication-A Survey and Criticism, 66 YALE L.J. 319 (1957). I am not here proposing a general
by statute the settled procedures of the common law. The fourteenth

amendment could not have been adopted without carrying over to the
states at least the legality requirement.
All of the due process safeguards against unjust conviction confirm
the centrality of legality to due process of law. The undisputed, clearly
intended requirements of due process are notice of the charge and protec-
tion against unjust conviction by the grand and petit jury procedures. 8
Contemporary doctrine has added counsel, 89 confrontation, 90 compul-
sory process, 91 identification safeguards, 9 and immunity from double
jeopardy9 to the procedures due process requires. These procedural
aspects of due process, however, are entirely nugatory unless due process
incorporates the legality principle. What is the point of providing notice
of the charge, trial by jury, counsel, confrontation, or compulsory pro-
cess, if innocence of the charge would not bar the defendant's punish-
ment? It might make sense for a state to accept the positivist position,
and provide that guilt of crime is equivalent to conviction according to
specified procedures. But it does not make sense for the federal Constitu-
tion to impose procedures designed to prevent convicting people who
have not broken state law unless the same constitution forbids punishing
such people. The framers did not entertain a purely procedural concep-
tion of criminal liability. In their day, just as in ours, those whose con-
duct clearly had not broken a contemporary law could not be put on trial
no matter how "fair" the procedures.94 Guilt is a particular conjunction
approach to due process cases; on the contrary, my thesis is that even on the narrower understanding
of due process as a technical provision governing criminal prosecutions' according to common law
standards, legality is the central element of due process.
88. Even Judge Easterbrook admits this much. See Easterbrook, supra note 72, at 98-100.
89. Gideon v. Wainwright, 372 U.S. 335 (1963).
90. Pointer v. Texas, 380 U.S. 400 (1965).
91. Webb v. Texas, 409 U.S. 95 (1972).
92. United States v. Wade, 388 U.S. 218 (1967).
93. Benton v. Maryland, 395 U.S. 784 (1969).
94. It might be objected that the jury system of the late 18th and 19th century, whether in
England or America, gave jurors power over law as well as fact. This system could be seen as
concerned less with measuring conduct according to legal criteria than with assessing the
defendant's moral fitness for punishment. The objection would be misplaced, however, because the
jury system has never been understood as authorizing extra-legal punishment. Even the most
vigorous defenders of the jury's power over questions of law understood the jurors to exercise solely
the power of reducing, not of increasing, the scope and severity of criminal liability. See T. GREEN,
VERDICT ACCORDING TO CONSCIENCE 261 (1985) (according to opponents of Stuart rule who
defended the jury's power over issues of law, "Right to jury was right to judgment by two lay bodies,
mainly to provide a double check on fact but also to prevent prosecutions for activity that was not
truly criminal."). Blackstone, who strongly defended the legality principle, see 1 W. BLACKSTONE,
supra note 83, at *92, pointed out that although the court could set aside jury verdicts of guilty, none
that acquitted the prisoner had been set aside, 4 id at *355. See, eg., The King & Smith, 84 Eng.
Rep. 1197 (K.B. 1681) (setting aside conviction against the evidence); The King & Primate v.
Fenwicke & Holt, 83 Eng. Rep. 1104 (K.B. 1663) (denying motion for new trial even though
acquittal resulted when defendant's servants kidnapped witnesses); The King v. Read, 83 Eng. Rep.
1987] RFSONABLE DOUBT RULE 1683
of law and fact, which exists independently of the procedures employed

to identify it. Due process safeguards express a commitment to regulat-
ing punishment by law rather than to providing a relatively fair govern-
ment of functionaries.9"
271 (K.B. 1660) (forbidding new trial after acquittal); see also W. HAWKINS, PLEAS OF THE CROWN
442 (2d ed. London 1724). Illegal acquittals were thus distinguished from illegal convictions which
were beyond the jury's power.
The court's authority to prevent an unjust guilty verdict was not, at common law, limited to
insufficiency of the evidence. Before trial, on a plea of demurrer or a plea of the general issue, the
judges would dismiss an indictment that charged the wrong crime or that failed to charge any crime.
4 W. BLACKSTONE, supra note 83, at *327-28. After a jury verdict of guilty, the defense could move
in arrest ofjudgment on the same ground. Id. at *368-69 ("a defective indictment is not aided by a
verdict"). If unsuccessful, review was available in the Court of King's Bench on writs of error and
certiorari. Id. at *262, *384.
Early American practice involved a similar refusal to permit juries to convict except according
to the law. See, ag., United States v. Villato, 28 F. Cas. 377 (C.C.D. Pa. 1797) (No. 16,622) (alien
cannot be guilty of treason); Commonwealth v. Hearsey, I Mass. *137 (1804) (plea of guilty to
indictment that failed to charge crime arrested by court acting sua sponte); Commonwealth v. Catlin,
I Mass. *8 (1804) (evidence of private sexual activity insufficient to support indictment for public
lewdness); State v. Boon, 1 N.C. (Tray.) 191 (1801) (jury verdict of guilty set aside because statute
forbidding killing slave was "too uncertain" to enforce); Respublica v. Richards, 2 Dall. 224 (Pa.
1795) (verdict of not guilty directed under statute forbidding enslavement of blacks because
defendant owned slave under laws of another state); Respublica v. Roberts, 2 DalI. 124 (Pa. 1791)
(unmarried man cannot be convicted of adultery, but only of fornication); State v. S.S., I Tyl. 180
(Vt. 1801) (indictment charging breach of the peace quashed because statute forbidding
"challenging" another did not apply to a written challenge to duel). Even when the indictment
charged a common law crime for conduct not previously condemned, the court retained the
authority to determine the scope of the common law of crimes. See, eg, Respublica v. Teischer, I
Dall. 335 (Pa. 1788) (malicious killing of horse is criminal at common law).
Aggregate data confirm that early American juries were not authorized to convict except
according to the law. Of 2718 federal indictments filed between 1801 and 1828, 902 ended in nolle
prosequ, 661 had no recorded disposition, and 80 were abated, quashed, discharged, discontinued, or
terminated by escape. Only 1075, fewer than half, ended in jury verdicts. D. HENDERSON,
CONGRESS, COURTS, AND CRIMINALS 46-47 (1985). Against this background, the Bill of Rights can
scarcely be seen as designed to assist the jury in a lawless inquiry about the defendant's moral fitness
for punishment. Rather, it establishes a fundamental defense against unlawful punishment.
95. Professor Kadish has observed:
The various procedural safeguards traditionally demanded in the name of due process
appear to be directed to two objectives. One is the goal of insuring the reliability of the
guilt-determining process-reducing to a minimum the possibility that any innocent
individual will be punished. It is not of crucial importance whether the individual tried is
in fact guilty or innocent, but it is of crucial concern that the integrity of the process of
ascertaining guilt or innocence never be impaired. If in this effort to insure that none but
those guilty be convicted, many guilty go free, the price is not too great in the long view of
democratic government. If there is any consideration basic to all civilized procedures it is
this, no matter how disparate the means chosen to give it effect.
This consideration, often expressed in terms of "fairness," gives meaning to the great
bulk of procedures that have become part of the due process of law: that the accused be
put on fair notice of the nature of the prohibited acts; that he be given an adequate
opportunity to present his side through counsel before a fair and impartial tribunal free
from prejudicial influences; that he be entitled to be continuously present at the trial, and
to confront and cross-examine his accusers; that he have the right to be free of the
damaging and untrustworthy influence of coerced confessions and testimony knowingly
perjured.
The relation between restrictions on process directed to this objective and the ultimate
values of human freedom is not obscure.... Due process thus serves the same end as a
All of the Supreme Court cases incorporating within the fourteenth

amendment the protections of the Bill of Rights against unjust conviction
recognize that due process includes the prohibition of punishment.
Other cases stand for the same proposition. The void-for-vagueness doc-
trine is a conspicuous example. 9 6 In most of the vagueness cases, the
defendant engaged in conduct that he or she probably thought was illegal
and that the state legitimately could have criminalized.9 7 But unless the
state identifies the forbidden conduct with reasonable clarity, a convic-
tion for violating the statute contravenes due process.
In part, such decisions may reflect notions of fair notice, but in most
of the cases such an explanation would be attenuated to the point of fic-
tion. Rather, the vagueness cases provide an excellent illustration of the
substantive value served by the legality principle. Their underlying
rationale is that in the absence of a prior definition of criminal conduct,
punishment is more likely to further some ulterior discriminatory or
oppressive purpose than it is to further legitimate public interests. 98
Put another way, the legality principle ensures a degree of neutrality
among persons in the administration of justice. Rules made in advance
cannot as easily be directed toward despised individuals. Even absent
oppressive motives, the legality principle helps to prevent punishment
that is merely gratuitous and arbitrary. The legality principle expresses
the judgment that punishment must be justified by some public purpose
important enough to be articulated generally and prospectively. Without
positive law, as contrasted with a law residing in the will of a ruling power-the attainment
of legal security and certainty.
Kadish, supra note 87, at 346.
96. See, eg., Kolender v. Lawson, 461 U.S. 352 (1983) (striking down a "stop and identify"
statute for vagueness because it encouraged arbitrary enforcement); Papachristou v. City of
Jacksonville, 405 U.S. 156 (1972) (finding criminal vagrancy statute unconstitutionally vague);
Lanzetta v. New Jersey, 306 U.S. 451, 458 (1939) ("The challenged provision . . . [is] so vague,
indefinite and uncertain that it [is] repugnant to the due process clause of the Fourteenth
Amendment.").
97. See, e.g., Bouie v. City of Columbia, 378 U.S. 347, 355 n.5 (1964) ("We think it irrelevant
that petitioners at one point testified that they had intended to be arrested."). But see Rose v. Locke,
423 U.S. 48 (1975) ("crime against nature" statute not unconstitutionally vague as applied to forcible
cunnilingus). This pattern, involving a defendant who knows his conduct to be criminal but who is
prosecuted on an unforeseeable ground, should be distinguished from that involving defendants
whose conduct is clearly within a statute whose applicability to other conduct is uncertain; typically
such a defendant has no vagueness claim. See Parker v. Levy, 417 U.S. 733 (1974).
98. Explanations of the vagueness doctrine typically supplement the fair notice rationale with
the prevention of discriminatory enforcement, especially when enforcement might trench upon other
constitutional rights. See, eg., Papachristou,405 U.S. at 168-71; 1 W. LAFAVE & A. ScoTr, supra
note 2, § 2.3(c), at 132; P. Low, J. JEFcRIEs & R. BONNIE, CRIMINAL LAW 67-79 (1982). In my
view this account, although on the right track, is somewhat incomplete. Whether prosecution under
a vague statute is discriminatory or simply arbitrary, it risks making individuals suffer for reasons
insufficiently important to be articulated prospectively. Thus, the penalties imposed are
objectionable even when not inspired by official animus; gratuitous punishment is wrong even when
it is based neither on racial prejudice nor on a program of political oppression.
that declaration, the state may not argue that the general interest
requires the suffering of a particular person. The chance that such an
argument is meritorious is too remote, the chance that the argument is
feigned too immediate, to justify its temptation in every case.
Several other arguments support the legality principle. It may pre-
vent political oppression9 9 and contribute to individual security."° More
categorically, it may simply be wrong to punish those who have not
offended an expressed rule. 01° But none of these arguments can explain
why the legality principle enjoys nearly complete priority over the public
interest in punishing wrongdoers-why punishment unauthorized by law
is, in civilized communities, simply beyond the pale. The priority of the
legality principle derives, I believe, from profound skepticism about
assertions that punishing legal conduct will contribute to the public inter-
est. In the face of the undoubted pain that accompanies any punishment,
this skepticism fully justifies the priority of the legality principle over the
general goals of punishment. 10 2
Many other Supreme Court precedents depend on the central role of
the legality principle in due process adjudication. The cases can be sepa-
rated into two categories, procedural and substantive. In the first cate-
gory, the Court has required state courts to follow specific procedures
designed to minimize the risk of erroneous conviction. These cases
include selective incorporation decisions in which the Court has iead the
due process clause as specifying procedures necessary to enforce the
legality principle and therefore implied or incorporated by the basic
guarantee of due process.10 3 In other cases in this category, the Court
has imposed procedures on the state courts on the basis of the due pro-
cess clause's independent potency. 1 4
The second category of cases involve direct, substantive efforts by
99. See Rawls, Two Concepts of Rules, 1955 PHIL. REV. 7-8.
100. See H.L.A. HART, PUNISHMENT AND RESPONSIBILITY 11-13, 17-24 (1968); Kadish, supra
note 87, at 346; Raz, The Rule of Law and Its Virtue, 93 LAW Q. REv. 195, 202-05 (1977).
101. See H.L.A. HART, supra note 100, at 22.
102. Legality, then, is an ultimately utilitarian principle. Indeed, frugality of punishment had a
prominent place in Jeremy Bentham's thought. See I J. BENTHAM, WORKS 398-99 (Edinburgh and
London 1859). The notion that retribution for immoral but legal actions might justify official
punishment is not implausible, but the cases in which such action might justifiably be taken are
always isolated, because the community can always reform the law prospectively. Allowing
punishment in one such case necessarily hazards many instances of unjustified suffering for every,
albeit outrageous, instance of lawful wrongdoing.
103. See cases cited supra notes 89-93. See generally Kadish, supra note 87. In a strained
attempt to avoid some unsavory implications of their theory that state power to define grounds of
liability entails state power to define the applicable procedures, the positivists argue that burden of
proof rules serve different functions than other aspects of due process. On the viability of this
distinction, see infra text accompanying notes 209-18.
104. See, eg., Betts v. Brady, 316 U.S. 455 (1942) (due process required state to provide counsel
to indigents in noncapital cases only in special circumstances).
the Supreme Court to prevent erroneous convictions in particular trials.

In the vagueness cases, for example, state courts have entered facially
valid judgments of conviction-judgments that formally satisfied the
legality principle. But simple acceptance of such formalism could com-
pletely subvert the legality principle, as the "trials" that occur in certain
totalitarian countries evidence. Limiting criminal liability to specified
categories of conduct would not limit state power to punish if a mere
judicial announcement that an individual has violated a valid law would
satisfy the legality principle. The Court has refused to accept any such
procedural definition of crime, and this refusal has fostered a large body
of case law.
In Moore v. Dempsey,105 for example, the Court ruled that a mob-
dominated show trial violated due process, even though the state appel-
late courts found that the prosecution had complied with state proce-
dural requirements. In Mooney v. Holohan,1 06 the Court held that the
prosecution's knowing use of false testimony would violate due process;
due process could not be satisfied by "mere notice and hearing if a State
has contrived a conviction through the pretense of a trial which in truth
is but used as a means of depriving a defendant of liberty through a delib-
erate deception of court and jury by the presentation of testimony known
to be perjured." ' 7 And in Thompson v. City of Louisville,10 8 the Court-
per Justice Black, no advocate of an elastic due process clause-struck
down a procedurally proper conviction because it was supported by "no
evidence whatever." The Thompson Court equated a conviction based on
no evidence with a conviction for "a charge not made" at all. °9 Both
practices amount to "sheer denial of due process."110
105. 261 U.S. 86 (1923).

106. 294 U.S. 103 (1935).
107. IAt at 112.
108. 362 U.S. 199, 206 (1960).
109. Id
110. Id In a recent decision upholding the preventive detention provisions of the Bail Reform
Act of 1984, the Court recognized the constitutional role of the legality principle, but held that a
"compelling" government interest justified "narrowly focuse[d]" pretrial detention. United States v.
Salerno, 107 S. Ct. 2095, 2103 (1987). The majority freely conceded "the 'general rule' of
substantive due process that the government may not detain a person prior to a judgment of guilt in
a criminal trial." Id. at 2102. But the Court found that the Act "narrowly focuses on a particularly
acute problem in which the government interests are overwhelming." Id. at 2103. Accordingly, the
statute fell within the exception to the general rule recognized for such practices as arrest and civil
commitment. Id. at 2102.
The Salerno majority's analysis is unconvincing because the Act authorizes detention only of
indicted persons. The statute therefore authorizes detention on the basis of an accusation that is
practically ex parte. Detention more restrictive than necessary to adjudicate the charge (arrest and
pretrial detention to ensure appearance) cannot be squared with the legality principle, unless the
detention is not punitive. Limiting preventive detention to those accused of crime looks very much
like punishment for offenses charged but not yet proved. See id. at 2112. (Stevens, J., dissenting).
The Salerno majority, nevertheless, does not question the centrality of the legality principle to
In re Winship Ill belongs in the procedural category because it held

that a state conviction is invalid unless the trier of fact applies the reason-
able doubt test to the evidence. Winship's substantive counterpart is
Jackson v. Virginia,112 which held that a federal habeas corpus petition
alleging insufficient evidence (under the Winship standard) to support a
state conviction required the federal court to conduct an independent
examination of the record. I next consider whether Winship and Jackson
represent justifiable efforts to enforce the legality principle by preventing
unauthorized punishment.
B. Enforcing the Legality Principle

Federal imposition of the reasonable doubt standard on state crimi-
nal trials may be justified on the grounds that federal courts have a duty
to prevent violations of due process. This duty is qualified, of course, by
principles of federalism, which limit the degree of federal judicial intru-
sion into the state criminal process. This section will demonstrate that
the reasonable doubt rule offers an important safeguard against the risk
of unjust conviction without derogating the constitutional value of feder-
alism. It follows that Winship was decided correctly.
1. Federalism and Due Process

Whenever state litigation implicates federal rights, federal law gov-
erns the standards and procedures affecting the federal claim.113 On
due process. "In our society liberty is the norm, and detention prior to trial or without trial is the
carefully limited exception." Id at 2105. The decision casts no doubt on the legality principle's
application to the adjudication of guilt.
111. 397 U.S. 358 (1970).
112. 443 U.S. 307 (1979).
113. When a given procedure forms an integral part of the federal right or when the absence of
the procedure would render the federal right nugatory, the supremacy clause requires at least this
much. See Monaghan, The Supreme Court, 1974 Term-Foreword:ConstitutionalCommon Law, 89
HARV. L. REV. 1, 22-23 (1975). The federal courts also have crafted various prophylactic rules-
that is, rules designed to enforce constitutional provisions even if they override state authority in the
absence of a constitutional violation. The Miranda rules are the classic example. See Miranda v.
Arizona, 384 U.S. 436 (1966). In Professor Monaghan's view, rules of this type are legitimate,
though subject to legislative modification. Monaghan, supra, at 27-30. His position has been
criticized for endorsing judicial interference with political decisions in the absence of constitutional
authority, see Shrock & Welsh, Reconsideringthe ConstitutionalCommon Law, 91 HARV. L. REv.
1117 (1978), and for approving constitutional adjudication in the absence of a genuine case or
controversy, see Grano, Prophylactic Rules in Criminal Procedure. A Question of Article III
Legitimacy, 80 Nw. U.L. REv. 100 (1985).
At least with respect to the fourth amendment exclusionary rule and the Mirandawarnings,
Professor Monaghan's critics are wide of the mark. The problem in both situations is that the
outcome of a particular case may determine the constitutional rights of persons not before the court.
It is not the purpose of the fourth amendment to protect crime, but unless some criminals are
protected, the lawful enjoyment of privacy might be lost. See Dipps, Living with Leon, 95 YALE
L.J. 906, 918-22 (1986). Similarly, the Miranda rules may require exclusion of some voluntary
statements, but without such rules courts would admit a great many involuntary statements
review, the federal courts may make an independent determination of

facts as well as law.114 Just as federal courts are responsible for enforcing
the fourth amendment's proscription of unreasonable searches and the
fifth amendment's privilege against self-incrimination against state viola-
tions, they are also responsible for enforcing the legality principle.
Enforcing the privilege against self-incrimination or the immunity
from unreasonable searches, however, poses a more modest threat to
state autonomy than does enforcing the legality principle. Every aspect
of trial procedure is potentially subject to attack as violating the legality
principle because every aspect of trial procedure can affect the likelihood
of erroneous conviction.
extracted by the police. See Schulhofer, The Fifth Amendment at Justice: A Reply, 54 U. Cai. L.
REv. 950, 956 (1987). The objections to the Monaghan thesis assume that the risk of future
violations of constitutional rights can never justify the reversal of a conviction that did not result
from a constitutional violation. This assumption is, in my view, mistaken.
With respect to the objection that constitutional authority is lacking, the refusal to reverse a
constitutionally rendered conviction, when reversal is the only way to prevent future constitutional
violations, amounts to treating the Constitution, not as an authority, but as a mere "code of ethics
under an honor system." Stewart, The Road to Mapp v. Ohio and Beyond, 83 COLUM. L. REv.
1365, 1383-84 (1983); see Dripps, supra, at 934-39. For illustration, consider Tennessee v. Garner,
471 U.S. 1 (1985), which holds that the fatal shooting of a nondangerous fleeing suspect amounts to
an unreasonable seizure under the fourth amendment. If the substantive holding is correct, any
enforcement of the suspect's fourth amendment rights would require an action by survivors -
individuals not themselves victims of the constitutional violation. There is thus no doctrinal
difference between the action allowed in Garner and the reversal of convictions under Mapp and
Miranda;in all three cases, the courts strike down state action because to do otherwise would foster
violations of the Constitution. In such situations, the article III objection confronts Chief Justice
Marshall's commonsense observation that the Court has "no more right to decline the exercise of
jurisdiction which is given, than to usurp that which is not given. The one or the other would be
treason to the constitution." Cohens v. Virginia, 19 U.S. (6 Wheat.) 264, 404 (1821); see also Moses
H. Cone Memorial Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 15 (1983) ("[Federal courts have a
'virtually unflagging obligation to exercise the jurisdiction given them.'" (quoting Colorado River
Water Conserv. Dist. v. United States, 424 U.S. 800, 817 (1976)); Currie, The Supreme Court and
FederalJurisdiction:1975 Term, 1976 Sup. Cr. REv. 183, 213-15; Redish, Abstention, Separation of
Powers, and the Limits of the JudicialFunction, 94 YALE L.J. 71 (1984) (criticizing the abstention
doctrine as abdication of mandatory jurisdiction).
My point is not to condemn prudential abstention doctrines, which may be defensible on other
grounds. See Shapiro, Jurisdictionand Discretion, 60 N.Y.U. L. REv. 543 (1985). Rather, I suggest
that the Supreme Court has at least the authority, if not the obligation, to choose constitutional
overenforcement over constitutional underenforcement. Cf Sosna v. Iowa, 419 U.S. 393 (1975)
(expiration of claim of class representative did not moot case for the entire class). If this is so, then
Professor Monaghan's characterization of Mapp and Miranda is unfortunate, for these decisions are
in no way subconstitutional, although they do depend on empirical causal relationships that might
be altered by legislation. Cases announcing prophylactic rules are constitutional decisions insofar as
they recognize criminal defendants' standing to enforce the constitutional rights of other parties.
The rules announced in Mapp and Miranda, therefore, could not be modified without infringing on
the constitutional rights of individuals, albeit different individuals than those convicted as a result of
the modifications.
114. See Norris v. Alabama, 294 U.S. 587, 589-90 (1935); Monaghan, Constitutional Fact
Review, 85 COLUM. L. REv. 229, 271-76 (1985). Of course, federal courts may-and most
frequently do-decline to review the facts independently.
Nonetheless, as the Supreme Court has consistently maintained,

enforcing the legality principle does not amount to exercising supervisory
power over state criminal trials. In devising procedures to enforce the
guarantees of the Bill of Rights, the federal courts must acknowledge the
primary role of the states in the administration of criminal justice.1 1
Acknowledging the primacy of the states does not derogate a constitu-
tional right in order to accommodate a contemporary policy preference,
because the Constitution itself provides for this state responsibility.
Two arguments confirm the constitutional character of federalism as
a constraint on enforcement of criminal procedure guarantees. First, the
structure of the government established by the Constitution transfers to
the federal government only some of the sovereignty otherwise confided
generally to the states. The fourteenth amendment's due process clause
diminishes the degree of sovereignty states may exercise over individuals,
but that limitation confirms the state's primary role. By limiting state
power to deprive individuals of life, liberty, or property, the fourteenth
amendment implicitly recognizes the continued primacy of the states in
enforcing criminal laws-laws characterized by the deprivation of life or
liberty. The fourteenth amendment plainly does not require, indeed can-
not justify, treating the Supreme Court's certiorari jurisdiction or the fed-
eral habeas statute as authorizing trial de novo of all state criminal
prosecutions.
Such an arrangement might well be outrageously wasteful or largely
futile in vindicating federal rights. It is also constitutionally indefensible
because the very authority relied on to justify federal review affirms the
central role of the states. The fourteenth amendment would not incorpo-
rate any of the procedural protections in the Bill of Rights if trial de novo
in federal courts with those safeguards were to follow in any event. The
fourteenth amendment regulates state criminal prosecutions precisely
because state criminal prosecutions matter-because state prosecutions
can deprive individuals of life and liberty.
Federalism, then, is an important constraint on federal courts con-
structing rules for enforcing the due process clause. The Constitution
does not, however, clearly determine the point at which the competing
constitutional concerns of federalism and due process statiate. The text
does no more than identify the relevant values, both of which must be
honored to some degree.
The constitutional tension between the legality principle and feder-
alism therefore defines a range of principled choices respecting enforce-
ment of the due process clause. At the very least, the states must facially
115. See, e.g., Patterson v. New York, 432 U.S. 197, 201-02 (1977) ("[Wle should not lightly
construe the Constitution so as to intrude upon the administration of justice by the individual
States.").
comply with the legality principle, that is, punish no one for crime with-
1 16 If
out a formal adjudication of conduct made criminal by positive law.
this much is legitimate, then cases like Mooney 117 and Thompson 18 are
also correct in requiring that the adjudication of guilt amount to some-
thing more than a palpable sham. At the other extreme, if federalism is
of any constitutional moment, then trial de novo in federal courts would
usurp the state responsibility that establishes the constitutional premise
of federal review.
Between preventing patently unauthorized punishment on the one
hand and avoiding trial de novo on the other, the Constitution provides
little guidance. Reasonable judges have endorsed at least three
approaches: (1) requiring nothing more than minimal legality-that is,
formal judgment, subject to federal review solely to prevent state use of
show trials designed to circumvent the adjudication requirement; (2)
requiring state trial courts to apply proof burdens and rules of evidence
that bias the outcome against erroneous conviction and toward erroneous
acquittal; and (3) requiring federal review of state judgments to ensure
that the judgment rendered satisfies these proof burdens and rules of evi-
dence, with substantial deference, in the interest of federalism, to the
state court judgment. Winship follows the second approach, and Jackson
the third. Patterson and Martin, in effect, apply the first approach to any
issue the state chooses to classify as an affirmative defense. The choice
among these competing approaches depends on the method used to
derive procedures and remedies from substantive constitutional
provisions.
2. ConstitutionalMethod
To choose among enforcement options for protecting constitutional
rights, the Court generally follows an instrumental approach, which
focuses on the purpose of the constitutional provision at issue. The
Court requires the states to follow procedures calculated to enforce effec-
tively the federal right, but does not require the most effective remedy or
most painstaking procedures. What procedures and remedies suffice
depends on the Court's perception of both how difficult it will be to
enforce the particular right in question and how costly additional proce-
dures would be.
Gerstein v. Pugh119 illustrates the Court's instrumental approach.
The Court held that the fourth amendment right against arbitrary deten-
116. Again, even Judge Easterbrook appears to admit this much. See Easterbrook, supra note
72, at 98.
117. 294 U.S. 103 (1935); see supra text accompanying notes 106-07.
119. 420 U.S. 103 (1975).
tion requires a prompt post-arrest judicial determination of probable

cause. Enforcement of this federal right does not require representation
by counsel at a trial-type hearing, however, because probable cause "can
be determined reliably without an adversary hearing." 120
The Court applied this instrumental approach to the assignment of
proof burdens in Lego v. Twomey,' 2 1 where it held that the preponder-
ance standard is constitutionally sufficient to establish the voluntariness
of a confession. According to the Court,
the exclusionary rules are very much aimed at deterring lawless conduct
by police and prosecution and it is very doubtful that escalating the pros-
ecution's burden of proof in Fourth and Fifth Amendment suppression
hearings would be sufficiently productive in this respect to outweigh the
public interest in placing probative evidence before juries for
122
the purpose
of arriving at truthful decisions about guilt or innocence.
The Lego Court distinguished Winship as requiring a higher burden of
proof regarding the elements of a crime "to ensure against unjust convic-
'
tions by giving substance to the presumption of innocence. "123
The Court's instrumental inquiry necessarily leaves a great deal of
room for judicial maneuver. But the Court could avoid such an
approach only by tolerating violations of unambiguous constitutional
value judgments. As an alternative to the instrumental inquiry, the
Court occasionally has invoked either historical or contemporary state
practice as a limit to its discretion, but neither offers a principled guide to
constitutional interpretation.
History may clarify the values underlying certain constitutional pro-
visions, but when the core of the constitutional value judgment is unam-
biguous, implementation requires the judiciary to fashion measures
informed by contemporary reality. 24 Just as importing extrinsic values
into the document abandons constitutionalism, refusing to give effect to
clear constitutional judgments would deny the Constitution's political
authority. 125 The key concern is not, after all, what jury instructions
were commonplace at the time of the Civil War, but rather what political
value judgments are expressed by the due process clause.
120. Id. at 120.

121. 404 U.S. 477 (1972).
122. Id. at 489.
123. Id. at 486-87.
124. See infra text accompanying notes 176-88. It is worth noting here, however, that the
Framers understood the Constitution to be a public document, subject to public interpretation, see
Powell, The Original Understandingof OriginalIntent, 98 HARV. L. REv. 885 (1985); that they
vested interpretative authority in life-tenured judges, see THE FEDERALIsT No. 78 (A. Hamilton);
and that they included the due process language in the Constitution knowing that its meaning was
not immutable. If to this we add more than a century of judicial interpretation which has expanded
the contours of due process, the assertion in the text is surely a mild one.
125. See Dripps, supra note 113, at 934-39; Stewart, supra note 113, at 1383-84.
Reliance on contemporary state practice as a guide to constitutional

interpretation is equally problematic. While the Court should respect the
weight of arguments that have persuaded legislatures of the constitution-
ality of a particular statute, contemporary state practice cannot deter-
mine constitutional analysis. Indeed, process-oriented theories of judicial
review 1 6 would suggest that unconstitutional legislation will rarely be
confined to a single jurisdiction; the political pressures that produce its
adoption in one place are likely to operate more broadly. Contemporary
practice is therefore likely to approve the most dangerous unconstitu-
tional trends-those posing the most sweeping threats to individual lib-
12 7
erty and those least amenable to political remedies.
That leaves the instrumental approach, whose great defect is its
open-ended character. While the Court's use of history and practice
reflects a laudable search for authority, authority alone may not resolve
difficult constitutional questions. At that point, treating as authoritative
factors whose relationship to the case is completely arbitrary confines
judicial discretion at the expense of rationality. Moreover, resort to his-
tory or state practice may not confine judicial discretion at all. Rarely
will a judge read history or state practice to require upholding a statute
that the judge's own instrumental analysis condemns. 2
While some sort of instrumental inquiry seems ultimately unavoida-
ble, the Court's discretion is not entirely open-ended. That reasonable
minds disagree does not mean that there are no wrong answers among
the plausible options. The criteria for discriminating among the plausible
options should not be wholly extrinsic to the document being interpreted.
Resolving hard cases well requires, first, fairly exhausting the direct
126. See, eg., J. ELY, DEMOCRACY AND DismusT 73-104 (1980) (defining and defending
"representation-reinforcing" approach to constitutional interpretation which focuses on the political
process). Cf United States v. Carolene Products Co., 304 U.S. 144, 152 n.4 (1938) (reserving
question of the level of scrutiny to be applied to legislation restricting the political processes that
might lead to repeal of undesirable legislation).
127. Justice Rehnquist's opinion for the Court in Schall v. Martin, 467 U.S. 253 (1984), offers
an excellent example. There, the Court upheld preventive detention of juveniles, in large part
because laws in every state provided for such incarceration. Id. at 267 n.16. Since juveniles are not
allowed to vote, process-based theory suggests that they are vulnerable to political abuse. The
existence of laws authorizing preventive detention of juveniles (based on predictions of future
criminal conduct) offers more empirical support for the process-based argument for
unconstitutionality than it offers for the reasoned-legislative-judgment argument for
constitutionality. On the implications of preventive detention for the legality principle, see supra
note 110.
The Court recently has disavowed any reliance on state practice in deciding whether due
process requires proof of crime beyond reasonable doubt. See Martin v. Ohio, 107 S. Ct. 1098, 1102-
03 (1987).
128. See Powell, Rules for Orfiginalists,73 VA. L. REv. 659, 677 (1987) ("Justices Hugo Black
and William Rehnquist, perhaps the two most consistent originalists in the Supreme Court's history,
have been equally consistent in their claims that the founders' views coincided with their own,
despite historial evidence to the contrary.").
authority of the constitutional provision and, second, approaching the

problem with as constitutional an attitude as possible. The assumption
should be that the Constitution is right as well as controlling; the judge's
duty is not to make the best constitution he can, but to resolve particular
ambiguities as if the rest of the Constitution were the best possible and
thus a guide to value judgments required by the process of
interpretation. 12 9
The instrumental approach need not degenerate into unrestrained
interest balancing. Interests must be balanced, but only constitutional
interests count, and the weight accorded them ought to reflect constitu-
tional rather than judicial preferences. The task is not easy, but it is nec-
essary and capable of being done either well or poorly.
To recapitulate, the legality principle is a key element of due process
of law. Federal courts therefore have a duty to enforce that principle
against the states. Federalism is a constraining constitutional principle,
but federalism and due process are like apples and oranges. There is no
reason to suppose that they are equally important merely because they
are both of constitutional stature. Resolving the competition between
them requires an assessment of their purposes and the degree to which
the reasonable doubt rule might protect the legality principle or infringe
on legitimate state authority.
3. The Reasonable Doubt Rule as a ProceduralSafeguardfor the

Legality Principle
The legality principle protects individuals against gratuitous or
oppressive punishment. The effects of an erroneous conviction on an
individual are potentially overwhelming. Indeed, the consequences of
any conviction include damage to reputation and some form of official
supervision. Incarceration, which is imposed for some offenses, is a
dreadful hardship, even in our best prisons. One measure of the magni-
tude of this hardship is the severity with which we punish private wrong-
ful imprisonment, also known as kidnapping.
This assessment of the legality principle's importance is not
129. But cf R. DwoRmIN, LAW'S EMPIRE (1986) (arguing that judges should interpret legal
texts so as to make them the best possible). Dworkin's approach runs the same risk that textual
indeterminism runs-namely, defeating the purpose of constitutions. Written constitutions have the
purpose of organizing and limiting government according to a dispositive authority. Once political
actors, including judges, have unfettered discretion to do as they please, the constitution is no longer
dispositive. Professor Dworkin thinks the constitution is determinate because he thinks political
morality is determinate; in other words, he thinks contemporary values will point to a right answer
from among the many allowed by the constitutional text. I think the point of a constitution is to
prevent contemporary values from operating politically except according to constitutional authority.
His view, I believe, would have effects indistinguishable from candid indeterminism. See Monaghan,
Our Perfect Constitution, 56 N.Y.U. L. Rnv. 353 (1981).
extraconstitutional. The framers of both the fifth and fourteenth amend-

ments shared the common law's concern for the legality principle,
expressed by an emphatic preference that the innocent not suffer even if
the guilty must go free. 130 The due process clause and the panoply of
procedural protections for the accused contained in the Bill of Rights
reflect the intensity of that preference.
The reasonable doubt rule makes an important contribution to hon-
oring the constitutional preference, with respect to either elements or
affirmative defenses. In practical terms, defending an unfounded crimi-
nal charge can be extremely difficult. The witnesses are frequently diffi-
cult to locate, uncooperative or incompetent, or forgetful of key facts. 131
Many witnesses, and frequently the defendant, have criminal records or
132
are for some other reason not entirely credible.
A first-rate criminal defense requires a thorough investigation, but
the public defenders who represent most criminal defendants have
neither the time nor the funding for such an investigation. 133 In many
cases, therefore, exculpatory evidence may exist, but never come to the
attention of the court. The defendant's inability to present exculpatory
evidence is more likely to lead to erroneous conviction if guilt can be
proved by a mere preponderance of the evidence.
This is not to say that the government always enjoys an overwhelm-
ing investigative advantage. While the government can rely on profes-
sional investigators, police investigation is geared primarily toward
arrests rather than convictions.1 34 Thus, police investigation all too often
leaves unresolved key uncertainties affecting the suspect's guilt. 135 In a
major case, no stone may go unturned, but in ordinary cases, the quality
of police investigation is often quite poor. Nearly half of all cases initi-
130. The common law preference was to acquit ten guilty defendants rather than to convict one
innocent defendant. 4 W. BLACKSTONE, supra note 83, at *558. The framers clearly subscribed to
this view. See supra notes 82-84 and accompanying text. The reasonable doubt rule apparently
represented a concession to law enforcement, for its predecessor at common law demanded acquittal
if there were any doubt, reasonable or otherwise. See Morano, A Reexamination of the Development
of the Reasonable Doubt Rule, 55 B.U.L. REv. 507, 510-11 (1975).
131. See 1 A. AMSTERDAM, TRIAL MANUAL FOR THE DEFENSE OF CRIMINAL CASES § 108, at
1-116 (4th ed. 1984).
132. Id, § 390, at 1-525 to -527.
133. See eg., How Public Defenders Deal with the Pressure of the Crowded Courts, Wall St. J.,
July 5, 1985, at I, col. 1.
134. See, eg., Terry v. Ohio, 392 U.S. 1, 14 (1968).
135. See B. FORST, J. LUCiANOviC & S. Cox, WHAT HAPPENS AFTER ARREST? A COURT
PERSPECTIVE OF POLICE OPERATIONS IN THE DISTRICT OF COLUMBIA 64-66 (1977). One study
compared samples of police reports from two jurisdictions with a list of questions proposed by an
experienced prosecutor. The questions were those the prosecutor thought the police investigation
should address tofacilitate the prosecution. In one jurisdiction, each of the 39 questions was on the
average covered in only 45% of the cases sampled; in the other, the figure was as low as 26%. See
NAT'L INSTrUTT OF LAW ENFORCEMENT AND CRIMINAL JUSTICE, THE CRIMINAL
INVESTIGATION PROCESS 20-22 (1978).
ated by arrest are dismissed, largely for evidentiary reasons.13 6 To the

extent police conduct an investigation, it is unlikely to focus on exculpa-
tory details. Accordingly, even generous provisions for defense discovery
would do little to bring exculpatory evidence to light.
Even if our system provided better investigation and produced more
evidence, defendants would still need the protection of the reasonable
doubt rule because the nature of a criminal prosecution imposes on the
defense an important handicap not borne by the prosecution. The charge
itself, with its enormous consequences, imposes on the defendant a pow-
erful incentive to offer exculpatory testimony. Unlike a civil case, in
which the plaintiff's gain is the defendant's loss and vice versa, the
defendant who offers evidence in a criminal case does so with the unilat-
137
eral taint of an obvious personal interest in the outcome of the dispute.
How many accused rapists, for example, would hesitate to offer perjured
testimony about consent if they thought it might help them avoid convic-
tion? Because of this incentive, the trier of fact is likely to discount any
exculpatory testimony given by the accused.
In a criminal case, the unilateral consequences of a conviction there-
fore undercut the credibility of the defendant's testimony. The credibil-
ity problem explains why access to evidence cannot justify shifting the
burden of proof to a criminal defendant. Although the privilege against
self-incrimination or the risk of impeachment with prejudicial prior con-
victions also distinguish criminal from civil trials, the credibility problem
alone provides sufficient reason to reject the access-to-evidence analysis
in criminal cases. The defendant has access to evidence about intent, for
example, but is the jury likely to believe it? The classic access-to-evi-
dence scenario, in which the defendant is the sole surviving witness to a
fatal encounter or the only one who can offer evidence of his mental
state, is precisely the situation in which the preponderance standard
poses the greatest risk of convicting the innocent.
The reasonable doubt standard also provides an important safeguard
in the majority of cases that do not go to trial. The terms of plea agree-
ments reflect the strength of the opposing cases. 13 An individual
accused of homicide who in fact killed in reasonable self-defense is far
more likely to plead to a manslaughter conviction if the government need
not disprove the affirmative defense beyond a reasonable doubt. The rea-
sonable doubt rule therefore contributes to preventing unjust convictions
136. See H. ZEISEL, THE LIMITS OF LAW ENFORCEMENT 19-21, 25-26 (1982).
137. Cf A. AMsTERDAm , supra note 131, § 280, at 1-402 ("That a criminal defendant is
presumed to be innocent until proved guilty is a canard. In reality, most criminal defendants must
prove their innocence .... ").
138. See, eg., Alschuler, The Prosecutor'sRole in Plea Bargaining 36 U. CHI. L. REv. 50, 60
(1968) ("The universal rule is that the sentence differential between guilty-plea and trial defendants
increases in direct proportion to the likelihood of acquittal.").
by reducing the pressures to plead to an unfounded but plausible

charge.139 Whether or not a criminal charge results in trial, the reason-
able doubt rule provides an important procedural safeguard for the legal-
ity principle.
4. Impact of the Reasonable Doubt Rule on Federalism

While protecting the legality principle, the reasonable doubt stan-
dard would not infringe on the constitutional value of federalism. The
federal structure established by the Constitution has the object of
preventing tyranny by decentralizing political power." Vigorously
enforcing individual constitutional rights does not conflict with that pur-
pose. As Justice Goldberg observed, any practices that are recognized as
beyond state power because of due process requirements are necessarily
beyond federal power as well. 4 Thus, requiring state courts to apply
the reasonable doubt standard would not-indeed, has not-enhanced
federal power at the expense of the states.
This is not to say that federalism imposes no limits on federal review
of state convictions. Trial de novo in federal courts would seriously
endanger the constitutional structure. The fourteenth amendment implic-
itly provides that the states retain primary authority to deprive individu-
als of life or liberty for violations of the criminal law. Reserving to the
states the responsibility to define, prosecute, and adjudicate general crim-
inal liability serves as an important hedge against the federal power a
national ministry of justice would represent. While it may now be possi-
ble to visualize federal adjudication of state law prosecutions, such a
vision is alien to the constitutional scheme of autonomous states. 142
Since such a scheme would likely make only a marginal or illusory con-
tribution to enforcing the due process clause, principled constitutional
interpretation cannot justify any such radical extension of federal judicial
power.
The Court struck a balance between requiring the reasonable doubt
safeguard at state trials and trial de novo in federal court in Jackson v.
139. See Christie & Pye, supra note 3, at 938-39.

140. See THE FEDERALisr No. 10 (J. Madison).
141. Pointer v. Texas, 380 U.S. 400, 414 (1965) (Goldberg, J.,concurring) ("mo deny to the
States the power to impair a fundamental constitutional right is not to increase federal power, but,
rather, to limit the power of both federal and state governments in favor of safeguarding the
fundamental rights and liberties of the individual.... [Tihis promotes rather than undermines the
basic policy of avoiding excess concentration of power in government, federal or state, which
underlies our concepts of federalism.").
142. See Malloy v. Hogan, 378 U.S. 1, 28 (1965) (Harlan, J.,dissenting) ("If the power of the
States to deal with local crime is unduly restricted, the likely consequence is a shift of responsibility
in this area to the Federal Government, with its vastly greater resources. Such a shift, if it occurs,
may in the end serve to weaken the very liberties which the Fourteenth Amendment safeguards by
bringing us closer to the monolithic society which our federalism rejects.").
Virginia,14 3 requiring limited federal court review of the sufficiency of the

evidence in state criminal cases. The Court has recognized that some
scrutiny of the facts in particular cases is ultimately necessary to prevent
states from resorting to show trials to circumvent the legality principle.
Mooney v. Holohan 14 and Thompson v. City of Louisville145 illustrate
this concern. Under the Jackson standard, moreover, federal court
review of state convictions is as deferential as their review of federal con-
victions. Federal courts are scarcely in the habit of reversing state con-
victions out of mere whim or caprice; it probably takes a case close to the
Thompson "no evidence" standard before reversal under Jackson
becomes appropriate. Nevertheless, Jackson does intrude on state
authority more seriously than In re Winship,"' while doing less to
enforce the legality principle. In my view, Jackson is defensible, but
scarcely compelled by the argument set out here.
C. The Scope of the Legality Principle: Legislative and Judicial

GradingDecisions
The standards announced in Winship and Jackson apply to each of
the elements of a charged offense, even when the defendant is guilty of a
lesser offense if the government fails to prove a particular element.14 7
Judicial sentencing decisions, however, are generally subject to far less
rigorous due process safeguards. Under Williams v. New York 14'8 consti-
tutional rules governing proof of incriminating facts do not apply to the
sentencing stage.
Positivists argue that legislative grading decisions about offenses and
potential penalties are indistinguishable from judicial grading decisions
of individual sentences.149 Since both legislative and judicial grading
decisions matter because they affect actual sentences, these commenta-
tors contend that similar burdens of proof should govern the determina-
tion of lesser-included offenses and sentences.
If the constitutional reasonable doubt requirement is grounded on
the legality principle, however, this argument loses most of its force.
First, equating judicial sentencing discretion with legislative grading of
offenses justifies neither the current judicial approach nor that of the pos-
itivist commentators. Such an equation would restrict the application of
143. 443 U.S. 307 (1979); see supra text accompanying note 112.
147. See Jackson v. Virginia, 443 U.S. 307, 313 (1979) (notwithstanding sufficiency of evidence
to establish second degree murder, issue remained whether state had proved premeditation element
of first degree murder beyond a reasonable doubt).
148. 337 U.S. 241 (1949).
149. See eg., Jeffries & Stephan, supra note 3, at 1352-53 & n.79.
the legality principle to facts that establish the existence, as opposed to

the seriousness, of a criminal violation. Once the state established that
an offense had been committed, the seriousness of the offense could be
measured free from constitutional restrictions. While the government
would have to establish intent and the absence of self-defense beyond a
reasonable doubt, it could prove facts establishing the grade of the
offense, such as provocation or the use of a weapon, by a mere prepon-
derance of the evidence. Whether the fact at issue related to a statutory
offense or defense would make no difference. The analogy to sentencing
holds up only if the defendant is properly convicted of some crime.
Second, the positivist's equation of legislative and judicial grading
incorrectly assumes that the legality principle's function is complete
when the government justifies some degree of punishment for the
accused-that is, when a violation of law has been proved. The legality
principle, however, limits unnecessary punishment with respect to the
potential seriousness, as well as the existence, of criminal liability. Prior
to conviction, the accused enjoys a constitutional liberty interest termina-
ble only by compliance with constitutional standards-including the rea-
sonable doubt rule. A valid conviction extinguishes that interest, but
only as against the punishment prescribed. Within the prescribed range
of penalties the convict has no constitutional objection founded on the
legality principle. 15 0
The distinction between legislative and judicial grading decisions
transcends constitutional formalism. Fundamental principles of criminal
law support setting the boundaries of the constitutional liberty interest
according to legislative grading decisions. These principles include the
basic purposes of institutional punishment-what H.L.A. Hart has called
the "general justifying aim" of punishment." I The basic framework of
punishment also includes certain side constraints on the pursuit of the
general justifying aim. These constraints, which Hart calls distributive
principles of punishment, preclude certain punishments even though they
might further the general justifying aim. 152 Terrorism, for example,
might be countered by executing hostages taken from among the popula-
tion on whose behalf the terrorism is committed. Such measures, how-
ever, would inflict a significant amount of pain, not only on those
punished but also on all whose security is threatened and who cannot
guarantee their safety even by obeying the law. This pain, moreover,
falls on individuals who have not engaged in immoral conduct or
150. A convict can make a legality-based objection if the judge enhances the sentence because of
the convict's lawful conduct. See North Carolina v. Pearce, 395 U.S. 711 (1969) (after conviction on
retrial, trial judge may impose a higher sentence if the higher sentence is within the statutorily
defined range and is not designed to punish the defendant for exercising a lawful right to appeal).
151. H.L.A. HART, supra note 100, at 8-11.
152. Id. at 11-13.
1987] 9]REASONABLE DOUBT RULE 1699
breached an obligation to the state. Thus, based either on utilitarian or

deontological ethical premises, the contribution to the general justifying
aim made by punishing the innocent does not overcome the contrary dis-
tributive3 principle limiting punishment to the legally and factually
15
guilty.
Recognition that the liberty protected by the due process clause
includes the freedom from punishment except for conduct violative of
law accords with these basic principles. The difficulty is explaining the
difference between a twenty-year sentence that is the judicially imposed
maximum for what a jury finds was manslaughter and a twenty-year sen-
tence that is the judicially imposed minimum for what the jury finds was
murder. The obvious formal difference is legality itself; to punish as a
murderer a defendant actually guilty only of manslaughter contravenes
the legislative accommodation of the general justifying aim and distribu-
tive principles. Does this formal difference correspond to any articulable
value judgment?
The formal difference, I believe, corresponds precisely with the pri-
mary value served by the legality principle-the minimization of needless
punishment. Sentencing discretion runs the very risks that the legality
principle combats. The nearly unreviewable discretion of sentencing
judges may be exercised for malicious reasons or for no reason at all. But
this risk is largely unavoidable. Not only must any system rely to some
extent on intelligent discretion to accommodate contending principles,
but, in a given case, reasonable people will differ over what those princi-
ples require. 154 The common sense of the system is that the convict has
qualified himself as a fit object of punishment within the discretionary
range, and may not complain that he might have fared better before a
different judge.
In contrast, legislative grading systems specify ranges of penalties
from which judges may not depart. The precept that the punishment
should fit the crime is surely a vaguer imperative than that some crime
must precede any punishment.' 55 The relative weakness of the propor-
tionality claim helps explain why within statutory ranges judges have dis-
cretion to consider proportionality along with instrumental aspects of
punishment' 5 6 and why eighth amendment proportionality review is so
153. See supra notes 98-103 and accompanying text.

154. Se4 eg., United States v. Bergman, 416 F. Supp. 496 (S.D.N.Y. 1976) (64-year old rabbi
should serve four month prison term because sentencing decision encompasses general deterrence
and respect for gravity of offense as well as rehabilitation and specific deterrence).
155. Cf H.L.A. HART, supra note 100, at 25 (discussing vagueness of proportionality principle).
156. On using instrumental considerations in sentencing, see id. at 24-25 (sudden increase in
sentences for a particular, rapidly increasing offense are unfair to those made examples of, but may
nonetheless be justifiable); see, eg., Bergman, 416 F. Supp. at 501-02.
1700 CALIFORNIA LAW REVIEW [V/ol. 75:1665
deferential to legislative judgments. 157 But when the legislature itself has
concluded that punishment above the maximum for the lesser-included
offense would be disproportionate, the convict's claim is stronger. Those
who act in a certain way deserve punishment, and up to the maximum
there might be social advantage in punishing some of them. Beyond the
maximum, however, the legislature cannot foresee any community
advantage that might justify more severe penalties.
The constitutional liberty interest in freedom from unauthorized
punishment thus requires not only prospective definition of criminal con-
duct but also prospective definition of the potential punishment. Just as
the state may not imprison an individual for, say, possession of a hand-
gun unless some state law prohibits handgun possession, so too the state
may not sentence the defendant to ten years in prison for handgun pos-
session if the law provides a maximum sentence of one year for that
offense. The prescription made in advance has the imprimatur of neu-
trality; to depart from it once the convict's identity is known risks pun-
ishing the defendant for something she didn't do.
The Supreme Court has recognized the legality principle's place in
limiting the scope, as well as the existence, of criminal liability. In Specht
v. Patterson,15 8 the Court struck down a Colorado statute that permitted
the sentencing court, upon a conviction for a predicate sex offense, to
classify the defendant as a sexual psychopath and impose an indetermi-
nate sentence of from one day to life in prison.15 9 In effect, the statute
created a separate offense-liability to a distinct range of penalties upon a
finding of a specified fact-and permitted adjudication of the separate
charge without the procedural safeguards of a criminal trial."6
The federal courts, however, have retreated from the principle, if
not the explicit holding, of Specht. The circuit courts upheld the federal
dangerous special offender statute (repealed in 1985) under which the
trial court could increase the sentence of a "dangerous special offender"
to a maximum of twenty-five years. Classification as a dangerous special
offender was made by the trial court after a sentencing hearing governed
157. Cf Rummel v. Estelle, 445 U.S. 263, 304-06 (1980) (Powell, J., dissenting) (Fourth
Circuit's application of the proportionality analysis, eventually adopted by the Court in Solem v.
Helm, 463 U.S. 277 (1983), resulted in reversal of only three sentences in ten years).
158. 386 U.S. 605 (1967).
159. Id. at 607-08.
160. The majority opinion stated:
The Sex Offenders Act does not make the commission of a specified crime the basis for
sentencing. It makes one conviction the basis for commencing another proceeding under
another Act to determine whether a person constitutes a threat of bodily harm to the
public, or is an habitual offender and mentally ill. That is a new finding of fact that was not
an ingredient of the offense charged.
Id. at 608 (citation omitted).
19871 REASONABLE DOUBT RULE
by a "preponderance of the information" standard.16

Relying in part on this trend in the circuit courts, the Supreme
Court in McMillan v. Pennsylvania 6' 2 upheld a Pennsylvania statute that
creates a minimum sentence for certain predicate felonies upon a show-
ing, by a preponderance of the evidence adduced at trial and at the sen-
tencing hearing, that the felony was accompanied by the visible
possession of a firearm. The statute does not alter the maximum penalty
for any of the predicate felonies, but, upon the required showing, it man-
dates a five-year minimum sentence, even if a lower minimum exists for
the felony proved at trial. Writing for the Court, Justice Rehnquist char-
acterized Specht as applying only when the statute might confront the
defendant with a "'radically different situation' from the usual sentenc-
ing proceeding."' 6 3 Furthermore, applying Patterson, the Court found
that Pennsylvania's use of the preponderance standard was constitutional
because the Pennsylvania legislature had expressly declined to denomi-
nate the visible possession of a firearm as an element of the offense or to
increase the maximum penalty for the predicate felony because of such
possession.'"6
The McMillan analysis deprives Specht and Winship of most of their
vitality. If an "element of the offense" does not denote a fact that defines
161. The special offender statute, 18 U.S.C. § 3575 (1982), was repealed in 1985. Pub. L. No.
98-473, tit. II, ch. II, § 212(a)(2), 98 Stat. 1987 (1984). For cases upholding the statute, see, e.g.,
United States v. Schell, 692 F.2d 672, 676-79 (10th Cir. 1982); United States v. Bowdach, 561 F.2d
1160, 1172-75 (5th Cir. 1977); United States v. Stewart, 531 F.2d 326, 332-35 (6th Cir.), cert denied,
426 U.S. 922 (1976). But see United States v. Duardi, 384 F. Supp. 874, 882-83 (W.D. Mo. 1974),
aff'd on other grounds, 529 F.2d 123 (8th Cir. 1975). Specht antedates Winship, and the federal
courts have limited the procedural protections guaranteed by Specht to those constitutionally
established at the time of that opinion. Cf.Schell, 692 F.2d at 677. The statute nonetheless created a
separate offense, just as in Specht, for once the defendant was shown to be a dangerous special
offender (defined in section 3575 (e) & (f) as a person who has committed two prior offenses of
specified types and who requires additional confinement for the protection of the public), the
maximum sentence for any of the predicate felonies could be increased to as much as twenty-five
years. Whether the trial court actually increased the sentence depended on the sentencing judge's
assessment of the defendant's dangerousness. Sentencing judges routinely make precisely this sort of
determination but usually within a range predetermined by the legislature. To my mind, the
government should have been required to prove the predicate felonies and dangerous proclivities
beyond a reasonable doubt. Otherwise, the defendant would be exposed to a range of liability far in
excess of that prescribed prior to the conduct. As Judge McKay noted in a thoughtful opinion:
A statute that imposes an additional sentence through a separate post-conviction
proceeding results in a reclassification of the defendant for the purpose of enhanced
punishment. This reclassification is essentially identical to an additional criminal
conviction. Consequently, the defendant's liberty interest that survived the original
conviction is entitled to procedural protections, similar to those provided at trial, in the
post-conviction proceeding.
Schell, 692 F.2d at 682 n.1 (McKay, J., concurring and dissenting) (citing Specht, 386 U.S. at 608-
10).
162. 106 S. Ct. 2411, 2420 (1986).
163. Id. at 2418.
164. Id. at 2416-18.
a range of judicial sentencing discretion, what does it mean? Both the

federal special dangerous offender statute and the Pennsylvania statute
upheld in McMillan created the same risk of unlawful punishment as any
other circumvention of the reasonable doubt standard. The felon who
did not in fact visibly possess a firearm or the defendant who is not in
fact a dangerous recidivist is likely to be punished more severely than is
justified by the prospective assessment of community purposes.
McMillan, decided by a bare majority of the Court, accordingly
deserves serious reconsideration. The Court's analysis is inconsistent
with any fair reading of Specht and is surely not essential to rational
sentencing procedures. State courts, for example, have uniformly
required proof beyond a reasonable doubt of prior convictions where
such evidence triggers a distinct range of judicial sentencing discre-
tion.16 In upholding the state statutes, the Supreme Court has insisted
on the defendant's right to a trial of the elements of the separate offense
created by recidivist legislation. 166 This is the only approach consistent
with Specht and Winship. The McMillan Court failed to expressly con-
sider the legality principle's relevance to legislative grading decisions.
Had the Court considered the issue, it might have recognized the funda-
mental inconsistency between the requirement of a formal trial of the
elements of a charged offense and deference to legislative nominalism
regarding the elements of the charge.
The nature of the right against unlawful punishment is thus essen-
tially the same in the context of legislative grading decisions as in the
context of defining elements of an offense. Punishment beyond the legis-
lative maximum was thought by the community to make no foreseeable
contribution to the general aims of punishment sufficient to justify the
incremental suffering of those convicted. The legislative maximum
165. See Annotation, Evidence of IdentityforPurposesof Statute as to Enhanced Punishmentin
Case of PriorConviction, I1 A.L.R.2d 870, 879 (1950). For representative cases, see In re Yurko, 10
Cal. 3d 857, 519 P.2d 561, 112 Cal. Rptr. 513 (1974); People v. Casey, 399 Ill. 374, 77 N.E.2d 812
(1948); People v. Reese, 258 N.Y. 89, 179 N.E. 305 (1932) (superseded by statute). The Seventh
Circuit suggested that the federal dangerous special offender statute may have been constitutionally
infirm because prior convictions could be proved by a preponderance. United States v. Neary, 552
F.2d 1184, 1193-94 (7th Cir.), cert denied, 434 U.S. 864 (1977). The circuit court contrasted this
infirmity with the finding of dangerousness. The court found the preponderance standard adequate
for the traditional sentencing determination of dangerousness because the standard provides the
accused with substantially more protection than Williams requires. Id. (citing Williams v. New
York, 337 U.S. 241 (1944)); see supra text accompanying note 148. This analysis is especially
tempting because the statute defined a convict as "dangerous" whenever an enhanced sentence was
required for the protection of the public-that is, when the sentencing judge, in his discretion, so
concluded. 18 U.S.C. § 3575(f). Nonetheless, under the statute, prior convictions themselves did
not justify a higher maximum. As indicated in the text, applying the preponderance standard to the
dangerousness element risked punishing the nondangerous recidivist with gratuitous severity.
166. See Graham v. West Virginia, 224 U.S. 616, 625 (1912). Indeed, in Spencer v. Texas, 385
U.S. 554 (1967), the Court upheld a procedure that required the defendant to try the existence of
prior convictions to the jury.
creates the same presumption that the absence of a criminal provision

creates-that punishment not authorized in advance is gratuitous or
oppressive. In both situations, the legality principle places a shield of
neutrality between the accused and the community. In the sentencing
context, however, the legislature has provided, with prospective neutral-
ity, that in certain circumstances any punishment within the statutory
range may be justified by its contribution to the general justifying aim of
punishment.
Finally, if,
as some suggest, the constitutional rules governing trial
and sentencing are hopelessly in conflict, 16 7 the discrepancy need not be
resolved against the accused. Consistency could be achieved by requiring
the reasonable doubt standard with respect to issues of fact at the sen-
tencing stage. This is not a terribly radical suggestion. A similar rule is
apparently applied in certain Commonwealth jurisdictions.16 The pro-
posal is also consistent with a Supreme Court trend toward greater scru-
tiny of the sentencing process.169 The Court has never suggested that
trials on issues of lesser included offenses could be conducted under the
free-wheeling sentencing standards allowed in Williams.17 It follows
that inconsistency will remain until due process makes further inroads on
the sentencing process. Pending the culmination of that trend, there is
no reason to impose arbitrary restrictions on the application of the rea-
sonable doubt test at the trial stage.
III
POINTS OF DEPARTURE REVISITED
Part II explained why due process prohibits punishment not author-

ized by the state's positive law. The reasonable doubt rule provides a
critical safeguard against the deprivation of this federal right. This Part
167. See eg., Note, The Constitutionality of Statutes Permitting Increased Sentences for
Habitual or DangerousCriminals, 89 HARV. L. REv. 356, 385-86 (1975).
168. See eg., R. v. Gardiner, 140 D.L.R.3d 612 (Can. 1983) (Supreme Court of Canada held
that sentencing constitutes a part of the trial process and therefore the Crown bears the burden to
prove all disputed facts); see also Wasik, Rules ofEvidence in the Sentencing Process, 38 CURRENT
LEGAL PROBS. 187, 194-200 (1985).
169. See Gardner.v. Florida, 430 U.S. 349 (1977) (plurality opinion) (due process violated by
death sentence imposed on the basis of information that defendant did not have opportunity to refute
or explain); Weissman, Sentencing Due Process: Evolving ConstitutionalPrinciples, 18 WAKE FoREST
L. REV. 523 (1982).
170. In rejecting the claim that the elements test would permit the states to recast traditional
criteria of liability in the form of mitigating circumstances to be considered at sentencing, the
Pattersonmajority stated that "there are obviously constitutional limits beyond which the States may
not go in this regard." Patterson v. New York, 432 U.S. 197, 210 (1977). This appears to indicate
that the Williams approach may not be extended to the trial of legislative grading determinations,
and that the inconsistency between the procedures required at sentencing and those required at trial
may be resolved only by gradually increasing the scrutiny of the sentencing process.
reexamines the case law and the commentary from the standpoint of the
legality principle.
The risk of unauthorized punishment is no less grgent in the context
of affirmative defenses than in the context of proving elements of the
offense. Neither history nor the need to facilitate legislative reform can
justify altering the constitutional protection against unauthorized punish-
ment according to legislative classifications. The positivist commentators
have rightly condemned the Court for failing to provide a neutral princi-
ple that justifies requiring the reasonable doubt standard sometimes, but
not always. But the positivists have also misunderstood the nature of the
constitutional right at issue. As a result, they have urged an approach
that would logically entail abandoning constitutional restrictions on state
criminal procedure with respect to a very large proportion of the issues
raised in typical litigation. Yet ironically, the positivists have failed to
offer a coherent justification for the reasonable doubt standard. Neither
the positivists nor other commentators have grounded the reasonable
doubt rule on a convincing constitutional warrant.
Basing the reasonable doubt rule on the legality principle therefore
offers important advantages over the positions thus far advanced in the
case law and the legal literature. I first consider the judicial distinction
and then turn to the positivist commentary.
A. The JudicialDistinction Between Elements and Defenses

Since legislative classification of a particular factual issue as an ele-
ment or as a defense is alterable by the legislature, reliance on the distinc-
tion for allocating burdens of proof is utterly arbitrary. The irrationality
of this approach is evident in Justice Rehnquist's contention that, since
the only difference between elements and defenses is the constitutionality
of shifting the burden of proof, legislative intention to shift the burden of
proof conclusively evidences an issue's status as an affirmative defense.171
Moreover, even if legislative inertia is supposed, there is no discernible
distinction between elements and defenses. Some defenses are hard to
disprove, but some elements are just as hard to prove.172 Some defenses
171. See McElroy v. Holloway, 451 U.S. 1028, 1028-29 (1981) (Rehnquist, J., dissenting from
denial of certiorari).
172. Insanity is frequently mentioned in this regard, with some support from Addington v.
Texas, 441 U.S. 418 (1979) (standard appropriate for civil commitment is clear and convincing
evidence of danger to self or others). One reason for theAddington result is that "[g]iven the lack of
certainty and the fallibility of psychiatric diagnosis, there is a serious question as to whether a state
could ever prove beyond a reasonable doubt that an individual is both mentally ill and likely to be
dangerous." Id. at 429. But this is surely just as true of the element of intent in cases in which intent
and insanity legally exclude each other. See supra note 47. Indeed the only likely explanation for
treating insanity differently from other exculpatory conditions with respect to the government's
burden of proof is the notion that even the patently insane deserve to be punished if they intended
their crimes. Since the probable consequence of an insanity acquittal is permanent commitment to a
173
are morally ambiguous, but so too are certain elements.
The only way to justify the distinction is to assign constitutional
significance to legislative classification. This might be done by one of two
arguments. The first, lurking in Justice Powell's Patterson dissent, takes
the historical classification of issues as elements or defenses as controlling
the constitutional question. 74 The second, on which the Patterson
majority explicitly relies, asserts that progressive criminal law reform
depends on legislative ability to shift the burden of proof.175 Neither
argument is persuasive.
1. History and FundamentalFairness

The reasonable doubt rule has a long history in our law, but the
standard has not always been applied to affirmative defenses. At com-
mon law, the defense had the burden of proving defensive matter to the
satisfaction of the jury. 1 76 This view prevailed in the United States until
the end of the nineteenth century, when the Supreme Court's decision in
Davis v. United States 177 initiated a major movement toward imposing on
the government the burden of persuasion, beyond a reasonable doubt, for
affirmative defenses whenever the defendant's evidence raised a question
for the jury. 17 Thus, it is arguable that deference to tradition in inter-
mental hospital, see Jones v. United States, 463 U.S. 354, 381 n.16 (1983) (Brennan, J., dissenting),
there is far less to fear from an erroneous insanity acquittal than from an erroneous acquittal based
on entrapment or absence of intent. The bete noirfor the foes of the insanity defense is the Hinckley
case, which indeed led to the passage of a federal statute shifting the burden of proof to the defense
on this issue. See 18 U.S.C. § 20 (Supp. III 1985). Yet few would disagree with the proposition that
Hinckley was insane at the time of his attack on the President. Apparently, the point of shifting the
burden of proof is to reduce arbitrarily the number of successful assertions of the defense. If this
rationale is correct, the defense should be abolished, not rationed according to the defendant's ability
to hire psychiatrists and lawyers. Cf. N. MORRIS, MADNESS AND THE CRIMINAL LAW 53-76 (1985)
(proposing abolition of special defense of insanity, returning to common law principles of actus reus
and mens rea). On the other hand, since the difference between conviction and commitment is
relatively slight, an erroneous rejection of an insanity claim less significantly implicates the core
concern of the legality principle-minimizing useless punishment-than does the erroneous rejection
of other defenses. If the Court believes an exception for insanity is indispensable, this distinction
might rationalize such an exception. In any event, if certain issues, such as sanity, are so clearly
distinct from the typical issue of historical fact that an exception to the government's burden must be
recognized, the elements test remains indefensible, for it looks to legislative classification rather than
to susceptibility to proof in determining the certainty required for conviction.
173. Withdrawal from conspiracy or attempt is typically an affirmative defense, yet the overt act
element of conspiracy, and the dangerous proximity element of attempt, are typically classified as
elements. These elements and defenses govern the same ambiguous issue of culpability, namely
whether unfulfilled criminal intent deserves to be punished.
174. See 432 U.S. at 226-30 (Powell, J., dissenting); see also supra text accompanying notes 63-
64.
175. See 432 U.S. at 209-11 & nn.ll-13; see also supra notes 57-62 and accompanying text.
176. See Fletcher, supra note 4, at 902-03.
177. 160 U.S. 469 (1895) (federal jury cannot convict where it has a reasonable doubt whether
defendant was mentally competent at the time of the killing).
178. See Fletcher, supra note 4, at 917-18; see also Patterson, 432 U.S. at 202-03:
preting the due process clause could justify a distinction between ele-
ments and defenses.
Nevertheless, the constitutional values served by the reasonable
doubt rule are implicated equally whether the legislature has classified a
fact as an element or as a defense. Whatever else it may mean, due pro-
cess of law surely means that individuals whose conduct has not violated
the law may not be subjected to criminal punishment. This value judg-
ment speaks to reality and not to procedural appearances. In terms of
substantive legality, one who kills on reasonable provocation is not guilty
of murder, whether state law deems provocation a defense or malice
aforethought an element. Similarly, one who kills in the reasonable belief
that lethal violence is immediately necessary to save her own life has
committed no crime, whether self-defense constitutes an affirmative
defense or negates the element of unlawfulness.17 9
Extending the reasonable doubt rule to affirmative defenses would
therefore not involve an assertion of judicial values not expressed in the
due process clause itself. Extending the rule would only impose a proce-
dural safeguard to protect a constitutionally recognized value. The
Court has often, and justifiably, extended contemporary procedures to
protect constitutional values in the area of criminal justice.
Two well-known examples suffice to illustrate this extension process.
In Katz v. United States,18 the Court held that electronic surveillance
that exposes private communibations may yield admissible evidence only
if the techniques conform with fourth amendment standards. Wiretaps
and bugs were unknown to the founders; the framers, therefore, could
not have intended to subject them to fourth amendment standards. The
Court, however, did not depart from a principled understanding of the
fourth amendment in its holding. The Court did no more than extend a
familiar and accepted procedure to protect the central value judgment of
1 81
the amendment.
Davis... had wide impact on the practice in the federal courts with respect to the burden
of proving various affirmative defenses, and the prosecution in a majority of jurisdictions in
this country sooner or later came to shoulder the burden of proving the sanity of the
accused and of disproving the facts constituting other affirmative defenses, including
provocation.
179. That certain crimes depend on elements that amount to moral judgments, in addition to
historical facts, is immaterial. The inclusion of such elements-recklessness, for example-expresses
the judgment that absent a jury's judgment of unreasonableness, punishment is unjustified. The risk
of gratuitous punishment is, as with predictions of dangerousness, at least as high in this situation as
it is when guilt hinges on historical facts. Of course, defining guilt solely in terms of the jury's moral
sense is a different matter entirely. See Giaccio v. Pennsylvania, 382 U.S. 399 (1966) (statutes that
allow a jury to impose penalties on the basis of the jury's moral sense of the defendant's conduct are
unconstitutional).
180. 389 U.S. 347 (1967).
181. See Monaghan, supra note 113, at 22-23.
Similarly, in Miranda v. Arizona 182 the Court created a set of proce-

dural protections against uncounseled or involuntary self-incrimination.
The Miranda rules do result in the exclusion of some voluntary confes-
sions, but without such rules, 8many
3
confessions might be coerced in the
confines of the station house.1
The Court has not disavowed either Katz or Miranda; both deci-
sions have become accepted constitutional precedents. 184 Like the exten-
sion of the reasonable doubt rule to affirmative defenses, both represent
legitimate exercises in constitutional interpretation. In each case the
Court recognized the necessity of appropriate procedural safeguards to
honor an unambiguous constitutional value judgment.
The argument for subordinating formal historical concerns to effec-
tive enforcement of the substantive constitutional values is not limited to
cases involving police practices. The Court has rejected traditional prac-
tices when necessary to effectuate the presumption of innocence. The
retroactive application of Winship and Wilbur surely evinced the judg-
ment that traditional practice must yield to that basic safeguard against
unjust conviction.' 85 Traditidnal practice, moreover, would permit the
government to rely on presumptions regarding elements, as well as to
shift the burden of proof with respect to defenses-yet the Court con-
tinues to scrutinize such presumptions on constitutional grounds. In
Sandstrom v. Montana,186 for example, the Court held that an instruction
that the law presumes that people intend the ordinary consequences of
their voluntary acts-a presumption with an impeccable common-law
pedigreeX8 ---ran afoul of Winship. The Court recently reaffirmed the
182. 384 U.S. 436 (1966).

183. See Schulhofer, supra note 113, at 956.
184. On the vitality of Katz see, e.g., United States v. Karo, 468 U.S. 705 (1984) (electronic
beeper concealed in a metal drum containing chemicals constitutes a search when it informs
government agents that the drum remains inside suspect's residence; effect is the same as physical
search of premises). Although the Court has refused to expand the scope of Miranda, the
procedures established in that case continue to be required in criminal investigations. See Colorado
v. Connelly, 107 S. Ct. 515, 523 (1986); see also New York v. Quarles, 467 U.S. 649, 660 (1984)
(O'Connor, J., concurring) (expressing unwillingness to overrule or extend Miranda);Rhode Island
v. Innis, 446 U.S. 291, 305 (1980) (Burger, C.J., concurring) (same).
185. See Hankerson v. North Carolina, 432 U.S. 233 (1977) (applying Wilbur retroactively);
Ivan V. v. City of New York, 407 U.S. 203 (1972) (applying Winship retroactively). In both cases,
the Court treated the reasonable doubt standard as a device for preventing (or "reducing the number
of") unjust convictions. The standard was therefore subject to retroactive application regardless of
the impact on judicial administration or the reasonableness of state reliance on previous doctrine.
See Hankerson, 432 U.S. at 241-44. These decisions provide unambiguous authority for the
proposition that the legality principle animates the constitutional status of the reasonable doubt rule,
and trumps prevailing state practice in cases of conflict.
186. 442 U.S. 510 (1979).
187. See, eg., R. v. Sheppard, 168 Eng. Rep. 742 (K.B. 1810) (forging a stock receipt is
sufficient evidence of intent to defraud).
Sandstrom holding."' 8
Due regard for the role of history thus does not compel a distinction
between elements and defenses. The Court has, in parallel contexts,
found a contemporary need for new procedures to safeguard undisputed
constitutional rights, and it has followed this practice with respect to
elements in the context of the Winship rule itself.
Finally, even if history is taken both as countenancing burden-shift-
ing and as determining the constitutional inquiry, the current distinction
between elements and defenses remains arbitrary. Classification as an
element is, according to the Patterson and Martin majorities, terminable
at will by the legislature. Justice Powell's Patterson dissent offers the
only reasonable basis for historically distinguishing between elements
and defenses. That basis is history itself, not latter-day legislative classifi-
cations that conflict rather than conform with the Anglo-American legal
tradition.
2. The Need for Legislative Compromise

Justice White's opinion for the Court in Patterson places considera-
ble emphasis on the risk that imposing the reasonable doubt rule might
stifle the process of legislative compromise.1 8 9 New York had recently
recodifled its criminal law, borrowing heavily from the much-respected
ALI Model Penal Code. In many instances, the legislature had broad-
ened the statutory grounds of exculpation or mitigation, but also had
shifted the burden of proof to the defense with respect to these
grounds. 190 The Court reasoned that unless the legislature had the abil-
ity to compromise substance and procedure, important advances in crim-
inal law would become politically impractical. Principled application of
the reasonable doubt standard appears to conflict with encouraging the
liberal reform of the substantive criminal law.
This appeal to the need for compromise should be rejected for three
important reasons. First, constitutional interpretation should not be
determined by the Court's perception of desirable legislative policy. Sec-
ond, the value of the legislative compromise is problematic. Third, the
distinction between elements and defenses is not necessary to enable pro-
gressive compromises.
First and fundamentally, judicial reaction to the wisdom of legisla-
tive choices has no legitimate place in constitutional interpretation.
Nothing in the due process clause requires or forbids particular defini-
188. See Francis v. Franklin, 471 U.S. 307 (1985).

189. See Patterson v. New York, 432 U.S. 197, 207-11 (1977), especially nn.10 & 13.
190. See Jeffries & Stephan, supra note 3, at 1398-1407 (identifying states, including New York,
that shift the burden of persuasion on particular issues to the defense); see also supra text
accompanying notes 55-56.
tions of provocation or self-defense. Yet the Patterson majority suggests

that because the Court prefers one constitutional legislative option over
another, an unambiguous constitutional value-namely the legality prin-
ciple-may be put at risk. The Court's critics rightly have objected to
this sort of free-wheeling interest balancing. As Professor Allen astutely
characterizes the appeal to compromise: "Is not the real point, in other
words, that affirmative defenses are unconstitutional but that such a con-
clusion may result in an unfortunate legislative choice, and thus the bet-
ter tack is to permit an unconstitutional choice as an expedient?" 1 9 ' Put
another way, principled interpretation necessarily hazards enforcement
of a constitutional rule that is unwise, and the Court has no business
taking one path because it would have written a different amendment.
Certainly the Court would refuse to entertain the argument from
compromise in other contexts. The regulation of the death penalty under
the eighth amendment precludes a compromise that would abolish felony
murder but provide a general capital penalty for premeditated murder.
The fourth amendment exclusionary rule prevents the legislature from
decriminalizing marijuana possession in "exchange" for allowing the use
of tainted evidence in major narcotics cases.
One arguable distinction is that a burden-shifting device sometimes
offers a more obvious compromise than a substantive concession on some
other issue. But allowing otherwise unconstitutional legislation because
of its obviousness as a bargaining chip would convert constitutional
interpretation into hair splitting, and tawdry hair splitting at that. The
Court would need to ask whether a particular procedural safeguard is so
clearly related to an area of potential progress that the legislature would
certainly perceive the connection. This type of inquiry would make the
Court the judge of legislative progress and make the Court's perception
of the legislature's acuity the "neutral principle" separating defendants
invoking one procedural safeguard from those invoking another.
Moreover, the elimination of other constitutional restrictions on
state power would offer far more obvious connections to progressive
compromises. If the Court were to disincorporate the privilege against
self-incrimination from the due process clause of the fourteenth amend-
ment, subjective definitions of explicable provocation, reasonable self-
defense, and entrapment, or the abolition of strict criminal liability,
would become more attractive options. Does not the right to confront
adverse witnesses interfere with developing a rational law of sexual
assault?
But the Court has never ascribed constitutional significance to such
consequences. If the privilege against self-incrimination is fundamental
191. Alien, Burdens of Persuasion, supra note 3, at 51.

to the Anglo-American concept of ordered liberty, then what the legisla-

ture might do absent a constitutional privilege is simply irrelevant. The
compromise argument has no greater force in urging an unprincipled
application of the reasonable doubt rule.
The second reason for rejecting the compromise argument is that
the Court's approach does not clearly redound to the greatest good of the
greatest number, even on the assumption that the Court has correctly
equated broader exculpatory categories with wise policy. The new or
modified defenses of concern to the Patterson majority tend to provide
for cases that occurred with relative infrequency under the older
codes.1 92 The revised codes shift the burden of proof to all defendants,
not just to those who would not have qualified for a jury instruction
under prior law. The compromise provisions thus benefit a few defend-
ants by bringing their cases within exculpatory categories but injure
many others who now bear the burden of proof that the government pre-
viously was required to sustain.
If this argument has an abstract quality about it, consider Gordon
Patterson's case itself. Patterson, a Viet Nam veteran seriously disturbed
by his experience in combat, killed Northrup after discovering him in
apparent adultery with Patterson's wife.193 Patterson therefore qualified
for a manslaughter instruction, even under the common law's restrictive
categories of allowable provocation.1 94 The jury twice sought reinstruc-
tion on provocation but could not conclude that he had met his burden
on that issue.1 95 As a result, Patterson was convicted of murder. The
compromise position would celebrate this result because the rule that
allowed Patterson's murder conviction-a conviction doubtful even on
traditional principles-gives killers at the fringe of the manslaughter cat-
egory a chance at mitigation. At the very least, this conclusion is highly
speculative.
The third reason for rejecting the appeal to compromise is that the
actual connection between legislative progress and the reasonable doubt
standard is far more problematic than the Court recognizes. In some
states, but not in others, adoption of the Model Penal Code approach to
certain defenses was accompanied by shifting the burden of proving these
defenses to the accused.1 9 6 This fact, however, does not of itself prove
that burden-shifting devices are essential to such compromises. Criminal
law recodifications involve a host of issues, each of which invites a
192. A glance at the compilation in Jeffries & Stephan, supra note 3, at 1398-1407, proves the
point.
193. See Patterson, app. at A-7 to A-8 (No. 75-1861) (psychiatrist's testimony) (microfiche).
194. See J. KAPLAN & R. WEISBERG, CRIMINAL LAW: CASES AND MATERIALS 263 (1986).
195. Patterson, app. at A-84 & A-92 (trial court charge).
defense position and a prosecution position. To say that protecting indi-

viduals against unjust conviction by fully constitutionalizing the reason-
able doubt standard would frustrate liberalizing compromises seems to
neglect the entire prosecutorial agenda except as it pertains to defenses
and proof burdens. If burden-shifting were off-limits, the legislature's
law-and-order faction might be convinced to accept a new affirmative
defense if offered another concession, such as a restriction of some other
197
substantive defense that has outlived its usefulness.
The logic of the compromise argument thoroughly undercuts the
apparent liberality of those who proffer this rationale. If criminal law
reform is an auction among legislative factions, the balance of forces in
the legislature cannot be changed by judicial decision.19 8 The implicit
assumption is that the aggregate benefits ceded to the class of criminal
defendants by the legislature are fixed; those benefits may take different
forms, but the total remains unchanged over the range of issues that are
subject to legislative discretion.
It follows, then, according to the auction theory, that refusing to
constitutionalize burden of proof rules cannot affect the net gains for law
reform. If a legislature declines to recognize an affirmative defense
because it cannot shift the burden of proof to the accused, the same polit-
ical currency held by the due process faction of the legislature could
purchase some other benefit. If unable to broaden the provocation
defense, the reform faction might insist on alternative measures, such as
a subjective view of entrapment, supraconstitutional procedural safe-
guards in death penalty cases, or better identification procedures.
This reasoning applies to individual legislators as well as to compet-
ing factions. A lawmaker with doubts about the substantive attractive-
ness of a defense might be persuaded to vote for it if the burden of proof
were shifted with respect to the new defense, but such a legislator might
also be persuaded by rejection of some other defense entirely, by stiffer
sentences for those convicted, or by anything else in the legislative
agenda-including a new state office building in her district. Thus, from
the standpoint of criminal defendants as a class, there is no theoretical
reason to believe that a principled application of Winship would discour-
age intelligent criminal law reform. The available empirical evidence
confirms this analysis. 99
197. For example, a subjective view of provocation and self-defense might be adopted in
exchange for abolition of the insanity defense.
198. See Easterbrook, supra note 72, at 112 ("There will be terms of trade between process and
procedure for almost any statute.").
199. It is easy to point out new grounds of exculpation that in fact have been linked to a shift in
the burden of proof, but this does not demonstrate a causal connection. If the compromise
hypothesis were correct, we would expect conservative states to retain the older rules in the absence
of a compromise. However, this does not appear to be the case. For example, several very
Every item in the reform agenda that is constitutionalized advances

the cause of law reform. Unless, as seems unlikely, a general reasonable
doubt rule would strengthen the law-and-order faction, the political capi-
tal of the reformers can buy as much as before; the reformers, however,
would start out with a shorter shopping list. The compromise argument
is therefore indefensible even from the standpoint of criminal defendants
considered as a class.2 "°
Deference to legislative compromise thus exchanges judicial enforce-
ment of the legality principle for an illusory greasing of the democratic
machinery. If the Court is serious about the primary role of state legisla-
tures in defining substantive grounds of criminal liability, it will abandon
openly favoring certain quite properly political trends. If the Court is
serious about preventing the unjust convictions of individuals, it can
scarcely do otherwise.
The Court cannot extricate itself from the contradictions of its pres-
ent position without overruling either Patterson or Winship. Martin
clearly indicates the Court's current refusal to reconsider Patterson, even
in an especially attractive case for doing so. Neither has the Court over-
ruled Winship, Specht, Wilbur, or Sandstrom. The Court would there-
fore not disturb its institutional commitment to stare decisis any more by
reaffirming the legality principle (rejecting Patterson) than it would by
enabling unrestrained state discretion over criminal procedure (repudiat-
ing Winship). The argument presented here suggests that when a new
Supreme Court majority takes the opportunity to reconsider the burden
of proof in criminal cases, the legality principle requires that consistency
should be achieved by reaffirming Winship and repudiating Patterson.
B. The Positivist Commentary Reconsidered
Before explaining in detail my disagreement with the positivist

approach, it is helpful to recall why some thoughtful commentators have
found it attractive, and why those opposing them have experienced diffi-
culty articulating a convincing rejoinder. As Professors Jeffries and Ste-
phan explain the attraction of their position:
conservative jurisdictions adopted the Model Penal Code definition of voluntary manslaughter
without shifting the burden of proof. See ARK.. STAT. ANN. §§ 41-110, -1504 (1981); Ky. REV.
STAT. ANN. §§ 500.070, 507.020 (Michie/Bobbs-Merrill 1985); MONT. CODE ANN. §§ 46-16-602,
45-5-103 (1983); N.D. CENT. CODE §§ 12.1-01-03, -16-01 (1985). Perhaps New York is indeed less
susceptible to progress than Arkansas, but I suspect proponents of the compromise approach would
be slow to come to that conclusion.
200. If gnly to confirm the cynicism of the compromise argument, it is worth pointing out that
the wave of recodification is now largely spent. Reviving Wilbu?s recourse to historical practice
therefore offers the Justices a fine opportunity to rook the state legislatures out of their conservative
procedures, given that legislative inertia is likely to maintain the liberal substance of "progressive"
compromises.
The logic of the greater-power-includes-the-lesser argument seems

compelling; it could be avoided only if one could identify an independent
rationale for proof beyond a reasonable doubt. In other words, a purely
procedural interpretation of Winship-one that is wholly illogical as a
statement of substantive policy-must find its justification in an exclu-
sively procedural concern that exists no matter what the content of the
underlying substantive issue. The case for reading Winship to disallow
every exception to proof beyond a reasonable doubt boils down to a
search for some such exclusively procedural justification.2z 1
The challenge, then, is to identify a constitutional proof requirement
independent from any constitutional requirement of what is to be proved.
Previous responses have failed to meet the positivist challenge. For
example, Professor Saltzburg has argued that one convicted of a greater
offense because of inability to prove a constitutionally optional defense
unfairly bears the stigma attached to the greater offense. 20 2 This argu-
ment assumes, without empirical support, that public perceptions of
stigma vary with statutory categories, but that the public is not aware of
the allocation of proof burdens. It seems far more likely that perceived
stigma turns on public understanding of the facts of a particular case.2 °3
I would be surprised if potential employers or old friends think of
Gordon Patterson simply as a murderer, no different from a Mafia hit
man; they more likely think of him as a man who killed his wife's lover in
the wake of a troubled marriage.
The more general point is that stigma is not within the government's
control. Certainly stigma does not correlate precisely with statutory cat-
egories. Conviction for tax evasion or insider trading, in the current
milieu, is more likely to ostracize the convict as a poltroon than as a
malefactor. On the other hand, a conviction for soliciting deviate sexual
intercourse, a misdemeanor unlikely to result in a sentence more serious
than a fine, can stigmatize quite seriously. Arrests and indictments stig-
matize routinely; they are often used for that very purpose. Even acquit-
tals, such as those won by John DeLorean and David Hinckley, may
stigmatize a defendant permanently. Yet the Winship rule does not apply
until the trial, and the rule applies at trial whether the offense is one
accompanied by public vilification or public adulation.
Professor Underwood, opposing the positivist approach, offers an
alternative argument based on public perceptions. 2° In her view,
201. Jeffries & Stephan, supra note 3, at 1345-46.
202. Saltzburg, supra note 3, at 405-08.
203. See Allen, Rationality and Accuracy in the Criminal Process: A Discordant Note on the
Harmonizing of the Justices' Views on Burdens of Persuasion in Criminal Cases, 74 J. CRIM. L. &
CRIMINOLOGY 1147, 1158-66 (1983) (criticizing Saltzburg's formulation of a stigma concern); see
also Allen, Limits of Legitimate Intervention, supra note 3, at 282 (arguing that "stigmatization
interest" is too indefinite to be given independent consideration).
204. Underwood, supra note 3, at 1324.
popular but mistaken beliefs in the applicability of the reasonable doubt

test will lead individuals to engage in unprovably innocent conduct. This
argument is even weaker than the claim based on stigma, for there is no
support for the notion that surprise as to the procedural rules applicable
to a potential trial violates any constitutional right.2 0 5 Moreover, it is
unlikely that many people, other than professional criminals, predicate
their conduct on the reasonable doubt rule. Common experience is that
of the naif, not the Holmesian bad man. The notion that even law stu-
dents take careful note of the available defenses, and then mistakenly
assume that the reasonable doubt rule applies, simply is not founded on
reality.
In contrast to these somewhat strained, if ingenious, arguments, the
legality principle meets the positivist challenge on its own terms. The
right asserted by a defendant invoking Winship is not the statutory right
to the exculpatory conditions recognized by state law, but the federal
constitutional right to freedom from punishment except as provided by
law. This federal constitutional right is inconsistent with purely proce-
dural definitions of criminal liability. 0 6 Legality requires more than an
official judgment of condemnation; it requires a prospective definition of
forbidden conduct and the potential consequences. Due process then
requires a formal adjudication designed to ensure that only those whose
conduct in fact falls within the prospective definition are punished, and
that the punishment does not exceed the specified limits.
We are so used to thinking of due process as a looking glass that the
most likely reason for rejecting my argument is skepticism about whether
legality is "really" part of due process, in the same way that the right
against coerced confessions is "really" part of due process. Yet, as I
argued earlier, legality is one aspect of due process whose pedigree dates
back to the original understanding. 0 7 Indeed, were it not for the legality
principle, the government could circumvent every other aspect of due
process in the criminal context. The positivist commentary makes this
point graphically clear, and should dispel the impression that recognizing
the constitutional role of the legality principle is either avoidable or
illegitimate.
205. In spite of Paul v. Davis, 424 U.S. 693 (1976) (reputational harm alone does not implicate
liberty or property interests), authority can still be found that emphasizes the importance of stigma
as an independent deprivation of constitutionally protected liberty. See Saltzburg, supra note 3, at
405.
206. Legality protects the substantive value of avoiding gratuitous or oppressive punishment,
and formal legality alone may not protect that fundamental value. See supra text accompanying
notes 98-110. Cf Tribe, The Puzzling Persistence of Process-Based ConstitutionalTheories, 89 YALE
L.J. 1063 (1980) (discussing the procedural norms mandated by substantive values underlying
specific constitutional provisions).
207. See supra notes 82-94 and accompanying text.
A defendant alleging denial of the right to trial by jury, to the assist-

ance of counsel, or to confront adverse witnesses, is ultimately invoking
the legality principle. Were it otherwise-if defendants in criminal pro-
cedure cases claimed merely a denial of state statutory entitlements-
then the state could deny any due process safeguard with respect to mat-
ters of legislative grace. The eighth amendment surely does not forbid
punishing larcenists and robbers to the same degree. According to the
greater-includes-the-lesser approach, then, the state could try the rob-
bery-or-larceny issue not only on the preponderance standard, but also
without a jury, without defense counsel, and without the right of con-
frontation, and if (inconceivably) the state lost such a "trial," it could
ignore the double jeopardy limitation and retry the issue, relentlessly,
until a conviction resulted.2 °8
Such a result would be inconsistent with every criminal procedure
decision the Court has ever rendered. The fourth amendment cases, the
confession cases, the right to counsel cases, the jury trial cases, the con-
frontation cases, the identification cases,-and the exculpatory evidence
cases all affirm rules that apply without regard to whether the state con-
stitutionally could dispense with the substantive statutory rule to which
the constitutionally infirm procedure relates.
The positivists have attempted to distinguish other procedural safe-
guards from the reasonable doubt standard on the theory that the former
protect the constitutional value of "rationality" by maximizing the likeli-
hood that the trier of fact will correctly assess the relevant facts.20 9 The
reasonable doubt rule, in contrast, is seen as protecting the separate con-
stitutional value of "liberty." The purported distinction, however, is
untenable for several reasons.
First, the Constitution does not affirm rationality as a value distinct
from that of freedom from unauthorized punishment. The procedures
incorporated by the due process clause are not animated by a scientific
preference for minimizing erroneous judgments, but by a political prefer-
ence for minimizing erroneous convictions. For example, the sixth
amendment guarantees "the accused"-not the court-the "Assistance
of Counsel for his defence." 2 1 Accordingly, the accused may waive
counsel, although rationality may suffer as a result," 1 and a claim of
ineffective assistance of counsel may prevail only on a showing of demon-
strable or presumptive prejudice.' 1 2 Similarly, the right to trial by jury in
208. See Saltzburg, supra note 3, at 400; Underwood, supra note 3, at 1329-30.
209. See, eg., Allen, supra note 203, at 1150-58; Allen, Burdens of Persuasion,supra note 3, at
45 n.60; Jeffries & Stephan, supra note 3, at 1351 & n.76.
210. U.S. CONsT. amend. VI; see United States v. Cronic, 466 U.S. 648, 654-55 (1984) (sixth
amendment requires not merely the formal provision of counsel but actual "assistance" of counsel).
211. Faretta v. California, 422 U.S. 806 (1975).
212. Strickland v. Washington, 466 U.S. 668, 691-96 (1984).
a criminal case includes the right to jury nullification, unchecked by

court authority to direct a verdict of guilty even when the government
presents overwhelming evidence.21 3 Confrontation, compulsory process,
speedy trial, and immunity from successive trial all are constitutionally
guaranteed only to the accused.2 14
In certain contexts, the constitutional preference for erroneous
acquittals rather than erroneous convictions renders the positivist's dis-
tinction between rationality and liberty functionally incoherent. Jury
trial hedges against unjust conviction, but does nothing to increase or
decrease the evidence before the court; it is essentially a device for allo-
cating errors, as is the reasonable doubt rule. 215 Defense counsel has a
duty to resist introduction of probative adverse evidence. In many cases,
counsel's primary role is to apply professional expertise to the exploita-
tion of the reasonable doubt rule itself.2 6
Second, even if rationality amounts to a distinct constitutional
value, the positivists have not explained why the legislature enjoys
greater power over liberty than over rationality. The framers imposed
far more rigorous restrictions on criminal cases, which implicate both
rationality and liberty, than they imposed on civil cases, which implicate
rationality alone. Legislatures are not likely to undervalue rationality,
but they may very well undervalue liberty. Why then do the positivists
argue that the courts should show greater deference to legislative judg-
ments that threaten liberty than to legislative judgments that threaten
rationality?
Finally, it is not clear that the reasonable doubt standard serves lib-
erty but not rationality. Indeed, Professor Nesson's defense of the rea-
sonable doubt standard is based on the independent constitutional status
of rationality.21 7 Similarly, Professor Saltzburg questions why rationality
is valued if not for accuracy, at least in the sense of avoiding false
218
positives.
The positivist approach depends on an arbitrary distinction to
explain established standards of minimal due process with respect to gra-
tuitous defenses.21 Viewed from the perspective of the case law, the pos-
213. United Bhd. of Carpenters v. United States, 330 U.S. 395, 408 (1947); Sparf & Hansen v.
United States, 156 U.S. 51, 105-06 (1895).
214. U.S. CONST. amends. V & VI.
215. See H. KALVEN & H. ZEMEL, THE AMERICAN JURY 190 (1966) ("Trial by jury is not an
instrument of getting at the truth; it is a process designed to make it as sure as possible that no
innocent man is convicted.") (quoting Lord Devlin, Jan. 1960).
216. See United States v. Cronic, 466 U.S. 648, 656 n.19 (1984).
217. Nesson, supra note 3, at 1582-84; Nesson, Reasonable Doubt and Permissive Inferences:
The Value of Complexity, 92 HARV. L. REv. 1187, 1201-02 (1979).
218. Saltzburg, supra note 3, at 400 n.45 ("mhe degree of accuracy required is intimately
connected to the question of who bears the burden of proof .... ).
219. I am surprised that Professor Allen would prefer to extend the greater-includes-the-lesser
itivist approach is illogical and indeed bizarre. It is credible only because

it is defended by scholars and jurists of the first rank.
Nonetheless, I cannot imagine that such astute observers would tol-
erate punishing an individual for, say, possessing marijuana, if the gov-
ernment had not made a law forbidding marijuana possession.2 2 ° The
state may surely make such a law, but until it does so individuals retain a
constitutional right against criminal punishment based on lawful con-
duct. Thus, the reasonable doubt rule is not a safeguard against the loss
of state statutory entitlements. Rather it is a safeguard against the loss of
constitutionally recognized liberty except as provided by law.
CONCLUSION
Winship, as it turns out, was on the right track. The purpose of the
reasonable doubt rule is to enforce the legality principle. The Supreme
Court, following a simplistic analysis of how legislatures might react to a
principled application of the reasonable doubt rule, has lost sight of the
principles underlying Winship. As a result, the Court has taken the para-
approach to counsel and jury trial rather than to abandon its application to the burden of proof. See
Allen, Burdens of Persuasion,supra note 3, at 45 n.60.
The positivists might respond by noting that the right to counsel extends to sentencing
proceedings, see, ag., Mempa v. Rhay, 389 U.S. 128 (1967), which I have distinguished as not
implicating the legality principle. Accordingly, the argument might go, the right to counsel must
serve some function independent of the legality principle.
This rejoinder, however, fails to extricate positivists from the reductionist tendency of their
position. Presumably they would deny the right to counsel at sentencing precisely because a valid
conviction extinguishes the constitutional liberty interest. See Greenholtz v. Nebraska Penal
Inmates, 442 U.S. 1, 7 (1979); Meachum v. Fano, 427 U.S. 215, 224 (1976). As at trial, the provision
of counsel at sentencing serves the interests of the client, not any general commitment to rationality.
Unless the positivists can articulate a justification for the right to counsel that applies generally at
trial, and on the same constitutional basis at sentencing, the objection based on sentencing does no
more than identify another line of precedents with which the positivist approach is inconsistent.
Rationality in sentencing cannot explain the role of counsel at sentencing hearings because the sixth
amendment does not extend to administrative proceedings that may have a more significant effect on
the sentence served by the convict. See Gagnon v. Scarpelli, 411 U.S. 778, 787-90 (1973) (states are
not constitutionally required to provide counsel for indigents in all probation or parole revocation
hearings).
As for my linking the right to counsel with the legality principle, it is entirely plausible that the
sixth amendment embodies a generalized solicitude for criminal defendants, at trial as well as at
sentencing. At trial, however, this solicitude extends no further than protecting the legality
principle, for the accused surely has no right to the assistance of counsel in urging jury nullification
or in presenting perjured testimony. See Nix v. Whiteside, 475 U.S. 157, 173 (1986) ("right to
counsel includes no right to have a lawyer who will cooperate with planned perjury"). But whatever
purpose counsel at sentencing may serve, the government could circumvent that function simply by
illegally imprisoning the convict for some reason other than violating the statute under which he is
sentenced.
220. See Johnson v. United States, 805 F.2d 1284, 1288 (7th Cir. 1986) (Posner, J.) ("To punish
a person criminally for an act that is not a crime would seem the quintessence of denying due process
of law.").
1718 CALIFORNIA LAW REVIEW
doxical position that constitutional limits on state power depend entirely

on legislative acquiescence.
The positivist commentators have mistakenly viewed the reasonable
doubt standard as requiring no more than that the government prove
convincingly a set of facts that might expose the defendant to criminal
liability if the legislature were to draft a statute for that very purpose.
The rule of law is to the contrary.
Ultimately, the reasonable doubt rule expresses a denial of formal-
ism. Requiring proof beyond a reasonable doubt guards against con-
demning people for crimes they did not commit. Rearranging the words
in the lawbooks does not change what the defendant did, and it ought not
alter the protection against unjust conviction. The government may
choose to broaden the grounds of criminal liability or to expand the
range of penalties applicable to particular offenses. But having made the
rules, the government must abide by them. It may not infer the judg-
ment that one among us has broken the law from a procedure that, in
light of the circumstances, is no more than a chance.

The Constitutional Status of The

Uploaded by

Copyright:

Available Formats

The Constitutional Status of The

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Constitutional Status of The

Uploaded by

Copyright:

Available Formats

The Constitutional Status of the

Reasonable Doubt Rule

The Supreme Court has held that due processforbids convicting an

t Assistant Professor of Law, University of Illinois. B.A. 1980, Northwestern University;

ernment must establish beyond a reasonable doubt, from affirmative

A. The Supreme Court and the Reasonable Doubt Rule

1. Presumptions and the Burden of Proof

2. Shifting the Burden of Proof

12. Id. at 164-67.

18. Id., app. at 28.

In a concurring opinion, Justice Harlan stressed the connection between

25. Id at 372 (Harlan, J., concurring).

32. State v. Wilbur, 278 A.2d 139, 146 (Me. 1971).

42. Id. at 699.

grounds of exculpation and mitigation on which the defense would bear

53. 432 U.S. 197 (1977).

B. The Positivist Commentary

68. Id. at 1106 (Powell, J., dissenting).

the state to provide a particular defense, the positivists would also

77. See Jeffries & Stephan, supra note 3, at 1365-66.

This Part argues in favor of a constitutional reasonable doubt rule at

A. The FederalRight Against Unlawful State Convictions

Charter had in 1215, the American colonists revered its guarantees as

by statute the settled procedures of the common law. The fourteenth

of law and fact, which exists independently of the procedures employed

All of the Supreme Court cases incorporating within the fourteenth

the Supreme Court to prevent erroneous convictions in particular trials.

105. 261 U.S. 86 (1923).

In re Winship Ill belongs in the procedural category because it held

B. Enforcing the Legality Principle

1. Federalism and Due Process

review, the federal courts may make an independent determination of

Nonetheless, as the Supreme Court has consistently maintained,

tion requires a prompt post-arrest judicial determination of probable

120. Id. at 120.

Reliance on contemporary state practice as a guide to constitutional

authority of the constitutional provision and, second, approaching the

3. The Reasonable Doubt Rule as a ProceduralSafeguardfor the

extraconstitutional. The framers of both the fifth and fourteenth amend-

ated by arrest are dismissed, largely for evidentiary reasons.13 6 To the

by reducing the pressures to plead to an unfounded but plausible

4. Impact of the Reasonable Doubt Rule on Federalism

139. See Christie & Pye, supra note 3, at 938-39.

Virginia,14 3 requiring limited federal court review of the sufficiency of the

C. The Scope of the Legality Principle: Legislative and Judicial

the legality principle to facts that establish the existence, as opposed to

breached an obligation to the state. Thus, based either on utilitarian or

153. See supra notes 98-103 and accompanying text.

by a "preponderance of the information" standard.16

a range of judicial sentencing discretion, what does it mean? Both the

creates the same presumption that the absence of a criminal provision

Part II explained why due process prohibits punishment not author-

A. The JudicialDistinction Between Elements and Defenses

1. History and FundamentalFairness

Similarly, in Miranda v. Arizona 182 the Court created a set of proce-

182. 384 U.S. 436 (1966).

2. The Need for Legislative Compromise

188. See Francis v. Franklin, 471 U.S. 307 (1985).

tions of provocation or self-defense. Yet the Patterson majority suggests