SEEKING CLARITY IN THE FEDERAL HABEAS
FOG: DETERMINING WHAT CONSTITUTES
“CLEARLY ESTABLISHED” LAW UNDER THE
ANTITERRORISM AND EFFECTIVE DEATH
PENALTY ACT
Melissa M. Berry

+

“Clear[ly]: adj. . . . plain . . . free from obscurity or ambiguity: easily
1
understood: unmistakable.”
I. INTRODUCTION
“Clearly, Your Honor, . . . .” “It is clear that . . . .” Attorneys
2
frequently use these words to bolster a point. We all do it. Most of the
time, however, “clearly” is superfluous; the argument should speak for
itself. As a consequence of this overuse, “clearly” has lost much of its
significance in everyday speech and writing. But this crisp seven-letter
word now plays a significant role in the federal habeas corpus arena.
This Article argues that, in the federal habeas corpus context, “clearly” is
not superfluous. “Clearly” can mean the difference between freedom
and prison.
In 1996, Congress deliberately included the word “clearly” in the
3
Antiterrorism and Effective Death Penalty Act (AEDPA), a statute that
4
dramatically altered the federal writ of habeas corpus. The “Great
+

Assistant Professor of Law, Chapman University School of Law. J.D., Northwestern
University School of Law; B.A., Tulane University. Many thanks to Diane AtkinsonSanford, Steve Berry, Scott Howe, and Celestine McConville for their thoughtful
comments on earlier versions of this Article, and to Christopher Hodson, Amy Martinez,
and Amy Oakden for their diligent research assistance. This Article is dedicated to my
mother, Mary McGonigal, whose courage and strength are an inspiration.
1. MERRIAM-WEBSTER’S COLLEGIATE DICTIONARY 212 (10th ed. 1997).
2. Outside of this Article, I try to use “clear” and “clearly” sparingly. As a young
law firm associate, I worked with a partner who had a pet peeve about the use of “clearly”
in legal briefs and memoranda. “If you have to say it,” he advised, “then your assertion is
probably not clear.” Many thanks to Ronald Berenstain of Perkins Coie LLP. This forum
does not allow me to follow another critical piece of writing advice that I learned when
serving as a law clerk to the Honorable David B. Sentelle of the U.S. Court of Appeals for
the District of Columbia: use footnotes sparingly.
3. 28 U.S.C. § 2254(d)(1) (2000).
4. This Article does not attempt to describe all of the changes made by Congress in
the Antiterrorism and Effective Death Penalty Act (AEDPA). For a comprehensive
overview of AEDPA, see Larry W. Yackle, A Primer on the New Habeas Corpus Statute,
44 BUFF. L. REV. 381 (1996).

747

748

Catholic University Law Review

[Vol. 54:747

Writ” of habeas corpus allows federal courts to free state court prisoners
5
who have been unconstitutionally imprisoned. AEDPA altered many
aspects of federal habeas corpus, but perhaps the most prominent change
was to the method by which federal habeas courts decide legal claims
6
that state courts have denied on the merits.
Section 2254(d) of AEDPA limits a federal court’s ability to grant a
state prisoner’s habeas application:
An application for a writ of habeas corpus on behalf of a person
in custody pursuant to the judgment of a State court shall not be
granted with respect to any claim that was adjudicated on the
merits in State court proceedings unless the adjudication of the
claim—
(1) resulted in a decision that was contrary to, or involved an
unreasonable application of, clearly established Federal law, as
determined by the Supreme Court of the United States; or
(2) resulted in a decision that was based on an unreasonable
determination of the facts in light of the evidence presented in
7
the State court proceeding.
This provision operates as a “constraint on the power of a federal habeas
8
court to grant . . . [the] writ” where constitutional error has occurred.
5. The American writ of habeas corpus has deep roots in the writ of habeas corpus
ad subjiciendum, often deemed the “Great Writ.” The writ was employed by courts in the
colonies and new states before the adoption of the U.S. Constitution. In 1867, Congress
enacted a statute mandating that federal courts “shall have power to grant writs of habeas
corpus in all cases where any person may be restrained of his or her liberty in violation of
the constitution, or of any treaty or law of the United States.” Act of Feb. 5, 1867, ch. 28,
§ 1, 14 Stat. 385, 385. For a history of the Great Writ, see WILLIAM DUKER, A
CONSTITUTIONAL HISTORY OF HABEAS CORPUS (1980); Paul M. Bator, Finality in
Criminal Law and Federal Habeas Corpus for State Prisoners, 76 HARV. L. REV. 441, 463
(1963); Clarke D. Forsythe, The Historical Origins of Broad Federal Habeas Review
Reconsidered, 70 NOTRE DAME L. REV. 1079 (1995); Dallin Oaks, Habeas Corpus in the
States—1776-1865, 32 U. CHI. L. REV. 243 (1965); Dallin Oaks, Legal History in the High
Court—Habeas Corpus, 64 MICH. L. REV. 451 (1966); and Michael O’Neill, On Reforming
the Federal Writ of Habeas Corpus, 26 SETON HALL L. REV. 1493 (1996).
6. Some commentators have argued that Congress followed the Court’s lead in
paring down the availability of the writ, primarily by adding procedural hurdles. See, e.g.,
A. Christopher Bryant, Retroactive Application of New Rules and the Antiterrorism and
Effective Death Penalty Act, 70 GEO. WASH. L. REV. 1, 5-15 (2002) (urging clarification of
retroactive application of U.S. Supreme Court criminal procedure decisions under
AEDPA); Melissa L. Koehn, A Line in the Sand: The Supreme Court and the Writ of
Habeas Corpus, 32 TULSA L.J. 389, 390 (1997) (noting that approximately two decades
ago, the Supreme Court began reducing availability of habeas writs, particularly through
creation of technical procedures for petitions); David Blumberg, Note, Habeas Leaps from
the Pan and into the Fire: Jacobs v. Scott and the Antiterrorism and Effective Death Penalty
Act of 1996, 61 ALB. L. REV. 557, 559-60 (1997) (discussing different treatment under
Warren and Rehnquist Courts).
7. 28 U.S.C. § 2254(d) (2000).

2005]

Seeking Clarity in the Federal Habeas Fog

749

This constraint, or “standard of review” as it is commonly called,
changed the pre-AEDPA standard of review from de novo to one that is
9
more deferential to state courts. Federal courts and commentators have
10
struggled to understand the significance of this change.
Generally,
however, § 2254(d)(1) is viewed as addressing the appropriate standard
of review for questions of law and mixed questions of law and fact, while
§ 2254(d)(2) is viewed as addressing the appropriate standard of review
11
for questions of fact.
Section 2254 is also AEDPA’s most controversial section. Debates
over the meaning of “contrary to” and “unreasonable application” have
12
consumed hundreds of pages in law reviews and in the federal reporters.
8. Williams v. Taylor, 529 U.S. 362, 412 (2000).
9. See 28 U.S.C. § 2254(d). Some commentators have noted that § 2254(d) is more
accurately described as a “limitation on relief” rather than a “standard of review.” See
JAMES S. LIEBMAN & RANDY HERTZ, 2 FEDERAL HABEAS CORPUS PRACTICE AND
PROCEDURE § 32.1, at 1419-21 (4th ed. 2001); id. at 1421 (“[S]ection 2254(d)(1) operates
as a ‘constraint on the power of a federal habeas court to grant . . . the writ’ . . . .” (first
omission in original) (quoting Williams, 529 U.S. at 412)); see also ERWIN CHEMERINSKY,
FEDERAL JURISDICTION § 15.1, at 862 (4th ed. 2003) (“Technically, federal court
consideration of the habeas corpus petition is not considered a direct review of the state
court decision; rather, the petition constitutes a separate civil suit filed in federal court and
is termed collateral relief.”). This author agrees but employs the standard of review
terminology because it is most commonly used.
10. LIEBMAN & HERTZ, supra note 9, § 32.2, at 1421-28 (citing cases); see also infra
note 12.
11. Allan Ides, Habeas Standards of Review Under 28 U.S.C. § 2254(d)(1): A
Commentary of Statutory Text and Supreme Court Precedent, 60 WASH. & LEE L. REV.
677, 681 (2003).
12. Commentators immediately took to analyzing whether AEDPA had changed the
standard of review from de novo to one of deference. See, e.g., Allan K. Chen, Shadow
Law: Reasonable Unreasonableness, Habeas Theory, and the Nature of Legal Rules, 2
BUFF. CRIM. L. REV. 535 (1999); Marshall J. Hartman & Jeanette Nyden, Habeas Corpus
and the New Federalism After the Anti-Terrorism and Effective Death Penalty Act of 1996,
30 J. MARSHALL L. REV. 337 (1997); Ides, supra note 11; Evan T. Lee, Section 2254(d) of
the New Habeas Statute: An (Opinionated) User’s Manual, 51 VAND. L. REV. 103 (1998);
James S. Liebman & William F. Ryan, “Some Effectual Power”: The Quantity and Quality
of Decisionmaking Required of Article III Courts, 98 COLUM. L. REV. 696, 864-84 (1998);
Todd Pettys, Federal Habeas Relief and the New Tolerance for “Reasonably Erroneous”
Applications of Federal Law, 63 OHIO ST. L.J. 731 (2002); Kent S. Scheidegger, Habeas
Corpus, Relitigation, and the Legislative Power, 98 COLUM. L. REV. 888 (1998); Jordan
Steiker, Habeas Exceptionalism, 78 TEX. L. REV. 1703 (2000); Adam Steinman,
Reconceptualizing Federal Habeas Corpus for State Prisoners: How Should AEDPA’s
Standard of Review Operate After Williams v. Taylor, 2001 WIS. L. REV. 1493; Mark
Tushnet & Larry Yackle, Symbolic Statutes and Real Laws: The Pathologies of the
Antiterrorism and Effective Death Penalty Act and the Prison Litigation Reform Act, 47
DUKE L.J. 1 (1997); Yackle, supra note 4; Larry W. Yackle, The Figure in the Carpet, 78
TEX. L. REV. 1731 (2000) [hereinafter Yackle, Figure]; Sharad S. Khandelwal, Note, The
Path to Habeas Corpus Narrows: Interpreting 28 U.S.C. § 2254(d)(1), 96 MICH. L. REV.
434 (1997); Andrea A. Kochan, Note, The Antiterrorism and Effective Death Penalty Act

750

Catholic University Law Review

[Vol. 54:747

One issue generally glossed over in this debate has recently emerged as
an issue to be reckoned with: what constitutes “clearly established
Federal law, as determined by the Supreme Court of the United
13
States”? This Article focuses on that question.
Because AEDPA requires that clearly established law must exist for a
court to grant a writ of habeas corpus, a habeas court’s determination of
what constitutes clearly established law has grave consequences. If
federal courts deny habeas relief on the ground that there is no clearly
14
established law in a particular case—and they regularly do —a common
understanding of what constitutes clearly established law is imperative.
Not defined in AEDPA itself, the phrase “clearly established” was first
interpreted by the Supreme Court four years after AEDPA’s
15
16
enactment.
In Williams v. Taylor, the Court defined “clearly
established Federal law, as determined by the Supreme Court of the
United States” as meaning the “holdings, as opposed to the dicta” of
17
Supreme Court decisions. This common sense definition, however, has
not provided enough guidance to the federal district and appellate courts
which regularly grapple with determining what constitutes clearly
established law.
of 1996: Habeas Corpus Reform?, 52 WASH. U. J. URB. & CONTEMP. L. 399 (1997); Note,
Rewriting the Great Writ: Standards of Review for Habeas Corpus Under the New 28 U.S.C.
§ 2254, 110 HARV. L. REV. 1868 (1997); Kimberly Woolley, Note, Constitutional
Interpretations of the Antiterrorism Act’s Habeas Corpus Provisions, 66 GEO. WASH. L.
REV. 414 (1998); see also Lindh v. Murphy, 96 F.3d 856 (7th Cir. 1996) (examining the
meaning of “Federal law” and “as determined by the Supreme Court of the United States”
in AEDPA), rev’d on other grounds, 521 U.S. 320 (1997).
Some commentators have focused on AEDPA’s limitation that § 2254(d)(1) applies
only to claims that were “adjudicated on the merits” by the state court. See, e.g., Monique
Anne Gaylor, Postcards from the Bench: Federal Habeas Review of Unarticulated State
Court Decisions, 31 HOFSTRA L. REV. 1263 (2003); Brittany Glidden, When the State Is
Silent: An Analysis of AEDPA’s Adjudication Requirement, 27 N.Y.U. REV. L. & SOC.
CHANGE 177 (2001); Robert D. Sloane, AEDPA’s “Adjudication on the Merits”
Requirement: Collateral Review, Federalism, and Comity, 78 ST. JOHN’S L. REV. 615
(2004), WL 78 STJLR 615; Margery I. Miller, Note, A Different View of Habeas:
Interpreting AEDPA’s “Adjudicated on the Merits” Clause When Habeas Corpus Is
Understood as an Appellate Function of the Federal Courts, 72 FORDHAM L. REV. 2593
(2004); Claudia Wilner, Note, “We Would Not Defer to that Which Did Not Exist”:
AEDPA Meets the Silent State Court Opinion, 77 N.Y.U. L. REV. 1442 (2002).
13. 28 U.S.C. § 2254(d)(1) (2000).
14. See infra Part III.A (citing cases where the court disposed of the case due to the
lack of clearly established law). Moreover, federal judges are not the only judges who
must make this determination. State judges would be well-advised to make this
determination. Even though they are reviewing the issues on direct appeal rather than
habeas, they need to know what law must be followed.
15. Williams v. Taylor, 529 U.S. 362, 377 (2000) (opinion of Stevens, J.).
16. 529 U.S. 362 (2000).
17. Id. at 412 (O’Connor, J., writing for the majority).

2005]

Seeking Clarity in the Federal Habeas Fog

751

A key question about AEDPA’s “clearly established” limitation is
whether the law must simply be “established”—meaning that there is
some precedent on point—or whether the law also must be “clear.”
Grammatical analysis suggests that “clearly established” is an adjectival
phrase modifying the noun phrase “Federal law,” and that “clearly”
18
modifies “established.” Under this reading, “the status of the law’s
19
establishment must be readily and perhaps unmistakably discernable.”
The few commentators who have offered definitions of clearly
established law have suggested that it means legal directives that are
20
“clearly anchored in existing case law” or “embodied within Supreme
21
Court precedent.” While useful, these definitions (like the Supreme
Court’s definition) lack the necessary precision to guide the lower federal
courts, which currently do not share a common understanding of what
constitutes clearly established law.
Such confusion reflects a larger, more fundamental question in a
22
common law system: what constitutes precedent for a current decision?

18. Ides, supra note 11, at 682.
19. Id.
20. Lee, supra note 12, at 123 (stating that “the rule must blend effortlessly into the
mosaic of existing decisions”).
21. Ides, supra note 11, at 684. Professor Ides notes that “the majority view of the
[Supreme] Court expressed within the holding and rationale of a decided case represents
the clearly established rule of law.” Id. at 683. He concludes that this definition comports
with H.L.A. Hart’s “rule of recognition” in the statutory context. Id. at 682-83 (citing
H.L.A. HART, THE CONCEPT OF LAW 97-107 (1961)); see also Allan Ides, Judicial
Supremacy and the Law of the Constitution, 47 UCLA L. REV. 491, 491-93 (1999) (arguing
that the Supreme Court creates constitutional law with its decisions).
22. As Judge Ruggero Aldisert has acknowledged, “There are precedents, and there
are precedents . . . All . . . do not have the same bite.” Ruggero J. Aldisert, Precedent:
What It Is and What It Isn’t; When Do We Kiss It and When Do We Kill It?, 17 PEPP. L.
REV. 605, 630-31 (1990). Although it is beyond the scope of this Article to examine the
theoretical underpinning of the use of precedent in the common-law system, a number of
scholars have tackled this broader issue. See, e.g., Larry Alexander, Constrained by
Precedent, 63 S. CAL. L. REV. 1 (1989); James Hardistry, Reflections on Stare Decisis, 55
IND. L.J. 41 (1979); Oona A. Hathaway, Path Dependence in the Law: The Course and
Pattern of Legal Change in a Common Law System, 86 IOWA L. REV. 601 (2001); Gary
Lawson, Controlling Precedent: Congressional Regulation of Judicial Decision Making, 18
CONST. COMMENT. 191 (2001) [hereinafter Lawson, Controlling Precedent]; Gary Lawson,
The Constitutional Case Against Precedent, 17 HARV. J. L. & PUB. POL’Y 23 (1994); Henry
P. Monaghan, Stare Decisis and Constitutional Adjudication, 88 COLUM. L. REV. 723
(1988); Robert S. Summers, Two Types of Substantive Reasons: The Core of a Theory of
Common-Law Justification, 63 CORNELL L. REV. 707 (1978); see also LARRY
ALEXANDER & KEN KRESS, Against Legal Principles, Law and Interpretation, in LEGAL
RULES AND LEGAL REASONING 249 (2000); RONALD DWORKIN, TAKING RIGHTS
SERIOUSLY (1977); PRECEDENT IN LAW (Laurence Goldstein ed., 1987). In addition to
the habeas context, questions about the meaning of precedent have arisen recently
regarding the significance of unpublished opinions. See, e.g., Lauren Robel, The Practice

752

Catholic University Law Review

[Vol. 54:747

The core concept of precedent is “a norm limiting the decisionmaker’s
23
flexibility.”
Through AEDPA, Congress has limited federal judges’
flexibility in determining what constitutes precedent for purposes of
habeas relief. The Supreme Court has interpreted § 2254(d)(1) to limit
the source and timing of clearly established law to Supreme Court
precedent existing at the time of the state court decision on the
24
petitioner’s claims. Any other limitations on what constitutes clearly
established law, however, are not presently well defined. This Article
seeks to define these limits and to provide a framework for analyzing
them.
Section 2254(d)(1) purports to follow the rule model of precedential
25
constraint, which is the general approach of our common law system.
“According to this model, precedent courts . . . promulgate rules of law
26
[when deciding cases].” The “rule” of the precedent case either refers
to some “canonical formulation of a rule that appears in the opinion” or,
when no canonical rule formulation appears in the opinion, to the
27
rationales discernable from the opinion. The model’s primary problem
28
is the identification of the precedent’s rule. Three main questions that
center on the scope of the precedent typically arise when attempting to
identify the precedent’s rule: First, what is the breadth of the legal issue
the precedent case has settled? Second, how many cases, i.e., factual
situations, does the precedent case control? Third, how does one
29
differentiate between the holding and dictum in the precedent case?
Determining precedent under the rule model also raises questions
about the strength of the precedential constraint. Such questions tend to
of Precedent: Anastoff, Noncitation Rules, and the Meaning of Precedent in an Interpretive
Community, 35 IND. L. REV. 399 (2002).
23. Frederick Schauer, Precedent, 39 STAN. L. REV. 571, 575 (1987). This classic
article describes the law’s reliance on precedent and how precedent operates in the legal
world.
24. Williams v. Taylor, 529 U.S. 362, 412 (2000) (O’Connor, J., writing for the
majority).
25. See LARRY ALEXANDER, Precedent, in LEGAL RULES AND LEGAL REASONING,
supra note 22, at 167, 169-71 [hereinafter ALEXANDER, Precedent]. Professor Alexander
describes three models of precedential constraint: the rule model (described above); the
natural model (past decisions naturally generate reasons, namely equality and reliance, for
deciding present cases the same way); and the result model (the result is what binds the
court, rather than a rule articulated by the court). Id. at 169-74; see also Alexander, supra
note 22.
26. ALEXANDER, Precedent, supra note 25, at 170.
27. Id.
28. Id. Professor Alexander notes that “some might object to the rule model because
it explicitly recognizes and endorses judicial legislation. If one’s view of the proper role of
courts excludes their legislating, even interstitially, then the rule model of precedential
constraint would be objectionable.” Id. (citation omitted).
29. Id.

2005]

Seeking Clarity in the Federal Habeas Fog

753

30

focus on the reasons why a court may disregard precedent. While these
questions typically address whether the precedent is from a higher court
(vertical) or from the same court (horizontal), in the habeas context
questions of strength focus on whether a precedent’s rule has been
applied consistently and whether that rule has been endorsed by a
majority of the Supreme Court. In general, however, the questions
raised by the rule model are the same types of questions regularly
encountered in the habeas context.
This Article argues that § 2254(d)(1)’s restriction of habeas relief to
cases where there is “clearly established Federal law, as determined by
31
the Supreme Court of the United States” maintains the general
operation of the rule model of precedential constraint, but more
narrowly constrains the source and timing of the precedent. It also
argues that while the source, number, and timing of the precedent are
dispositive issues for the habeas court, the issues relating to the scope or
strength of the precedent should not be dispositive. Instead, the court
should analyze issues relating to the scope of the precedent under the
“contrary to” and “unreasonable application” prongs of § 2254(d)(1).
Part II of this Article briefly reviews pre-AEDPA law before
examining the text of § 2254(d)(1). It then explores how “clearly
established law” has been interpreted by the Supreme Court in three
32
33
major cases: Williams v. Taylor, Lockyer v. Andrade, and Yarborough
34
v. Alvarado.
Part II notes that Williams resolved the issues of the
appropriate source and timing of clearly established law under §
2254(d)(1). The Court’s decisions in Andrade and Alvarado, on the
other hand, raised, but did not resolve, questions about determining the
scope and strength of the precedent. Furthermore, they illustrate that
35
the clarity of the precedent is relevant under § 2254(d)(1). This Part
also observes that while the Supreme Court has instructed that the
36
clearly established law inquiry should be a “threshold” issue, it has not
specified whether that simply means the issue should be addressed first,
or whether it also means the issue should be dispositive and thus the
basis for denying relief. Part II concludes that the Court’s definition of
clearly established law is broad and works in tandem with the standards
of review under § 2254(d)(1).
30. Id. at 175-76.
31. 28 U.S.C. § 2254(d)(1) (2000).
32. 529 U.S. 362 (2000).
33. 538 U.S. 63 (2003).
34. 124 S. Ct. 2140 (2004).
35. One commentator has argued that the clarity of precedent should not be relevant
to the “clearly established” law inquiry. See Ides, supra note 11, at 761-62.
36. Andrade, 538 U.S. at 71; Williams, 529 U.S. at 390; see infra Part II.

754

Catholic University Law Review

[Vol. 54:747

Part III examines life in the lower federal courts after Williams,
Andrade, and Alvarado and identifies recurring situations in which
questions about clearly established law emerge. The types of questions
that the lower courts have wrestled with generally fall into three
categories. The first category comprises cases where courts find that
there is no “clearly established law” applicable to the habeas petitioner’s
case. Some courts treat this determination as a dispositive issue and deny
relief in the absence of clearly established law. Other courts, however,
go on to analyze issues relating to the scope of the precedent under the
contrary to or unreasonable application prongs of § 2254(d)(1). The
second category consists of cases in which the court must decide when a
rule applies to the petitioner’s factual situation. Courts tend to analyze
these scope of precedent issues under the contrary to or unreasonable
application prongs of § 2254(d)(1). The third category comprises cases
that inquire into the clarity of the precedent. These cases involve the
application of general principles, old or otherwise questionable
precedent, and splintered Supreme Court decisions. Courts view these
issues of clarity as impacting the reasonableness of the state court’s
decision.
Part IV recognizes the need for a common understanding of AEDPA’s
clearly established law requirement and offers an analytical framework
for this determination. It advocates that the determination of clearly
established law should be a threshold question in the sense that the court
will attempt to ascertain the relevant law at the beginning of its analysis.
37
Part IV identifies five “analytic touchstones” to which courts should
refer when making this determination: the number of cases required, the
source of the precedent, the timing of the precedent, the scope of the
precedent, and the strength of the precedent. The proposed analytical
framework recommends that the first three touchstones—number of
cases, source of the precedent, and the timing of the precedent—be
dispositive issues requiring the court to deny habeas relief if they are not
met. If, however, a habeas court encounters questions relating to the
final two touchstones—the scope of the precedent and the strength of the
precedent—the court should continue with the analysis under §
2254(d)(1) and address those questions under the contrary to or
unreasonable application prongs. Part IV also recognizes a role for
federal court of appeals precedent in the clearly established law
determination.
This Article concludes that the determination of clearly established
law works in tandem with the standards of review in § 2254(d)(1). It

37. This phrase is borrowed from Professor Todd E. Pettys of the University of Iowa
College of Law. See Pettys, supra note 12, at 788.

2005]

Seeking Clarity in the Federal Habeas Fog

755

urges that, where the law appears to be clearly established because it
passes the threshold test of the first three touchstones, the approach most
consistent with AEDPA and Supreme Court case law is one in which the
habeas court continues with its analysis under § 2254(d)(1), deferring to
the state court’s reasonable interpretation and application of that law on
a sliding scale based on the clarity of the law, which is determined by its
scope and strength, the final two touchstones.
II. DEFINING “CLEARLY ESTABLISHED” LAW UNDER § 2254(D)(1)
Although our legal system does not have an agreed-upon approach to
statutory interpretation, most lawyers and judges agree that the text of
38
the statute is the starting point. Lacking guidance from the statute’s
text and sparse legislative history, the federal district courts and courts of
appeals have struggled with the meaning of § 2254(d)(1). Commentators
immediately took to analyzing whether AEDPA had changed the
39
standard of review from de novo to one of deference. Another area of
concern was whether § 2254(d)(1) limited the source of clearly
established law to only Supreme Court precedent (thereby proscribing
the use of circuit court precedent), and, if so, whether that limitation was
40
constitutional. In general, however, the meaning of clearly established

38. See, e.g., W. Va. Univ. Hosp., Inc. v. Casey, 499 U.S. 83, 98 (1991) (“The best
evidence of that purpose is the statutory text adopted by both Houses of Congress and
submitted to the President.”); Kaiser Aluminum & Chem. Corp. v. Bonjorno, 494 U.S.
827, 835 (1990) (“The starting point for interpretation of a statute is the language of the
statute itself.”); United States v. Am. Trucking Ass’ns, Inc., 310 U.S. 534, 543 (1940)
(“There is, of course, no more persuasive evidence of the purpose of a statute than the
words by which the legislature undertook to give expression to its wishes.”). However,
even established canons of statutory construction have long been criticized. See Karl L.
Llewellyn, Remarks on the Theory of Appellate Decision and the Rules of Canons About
How Statutes Are To Be Construed, 3 VAND. L. REV. 395, 401 (1950). This classic article
recognizes that “there are two opposing canons on almost every point.” Id.
39. See supra note 12.
40. See, e.g., Bocain v. Godinez, 101 F.3d 465, 471 (7th Cir. 1996) (construing §
2254(d) as requiring that “[f]ederal courts are no longer permitted to apply their own
jurisprudence, but must look exclusively to Supreme Court caselaw”); Fern v. Gramley, 99
F.3d 255, 260 (7th Cir. 1996); Lindh v. Murphy, 96 F.3d 856, 869 (7th Cir. 1996) (examining
the meaning of “Federal law” and “as determined by the Supreme Court of the United
States” in § 2254(d)(1)), rev’d on other grounds, 521 U.S. 320 (1997). In Lindh, the
Seventh Circuit asserted that AEDPA’s new scope of review provision in § 2254(d)(1)
“significantly interfere[s] with the judicial role and to a great extent prevents the judicial
department from accomplishing its ‘constitutionally assigned functions.’ Simply put, the
statute, as amended, deprives a federal court of the right to adjudicate the case.” 96 F.3d
at 890. Although this issue was not resolved by the Supreme Court, its review of the
Seventh Circuit’s decision in Lindh clarified that AEDPA limits the source of clearly
established law to Supreme Court precedent in Williams. The Court has never directly
addressed the constitutional argument.

756

Catholic University Law Review

[Vol. 54:747

41

law was not a focus. The time has come to focus the spotlight on this
issue.
After briefly reviewing pre-AEDPA law, this Part examines the text
and relevant legislative history of § 2254(d)(1), neither of which directly
provides a definition for “clearly established” law. It then examines the
three major cases in which the Supreme Court has given shape to the
42
clearly established law requirement: Williams v. Taylor, Lockyer v.
43
44
Andrade, and Yarborough v. Alvarado.
These decisions provide a

It is beyond the scope of this Article to address the constitutional argument in detail.
The basic argument is that AEDPA’s scope of review limitations interfere with the
exercise of Article III powers because Congress’s authority to tell the federal courts how
to use precedent is constitutionally questionable. Some scholars argue that AEDPA
unconstitutionally limits the stare decisis effect of lower federal court decisions.
Narrowing the role of lower federal courts coincides with and complements both the
Court’s and Congress’s obvious effort to narrow and curtail constitutional rights,
particularly in the criminal context. The counter-argument is that there is no
constitutionally enshrined right to mount a collateral attack on a state court’s judgment in
the inferior Article III courts and, therefore, no mandate that state court judgments
embracing questionable (or even erroneous) interpretations of the Federal Constitution
be reviewed by the inferior Article III courts. Thus, collateral review of judgments is
subject to legislative control. For a discussion of these constitutional arguments, see
Lawson, Controlling Precedent, supra note 22, at 194 (rejecting Congress’s power to
legislate when the courts’ decision-making power has been affected and making no specific
reference to AEDPA); Liebman & Ryan, supra note 12 (explaining why one reading of §
2254(d)(1) is constitutional and why the Fifth, Seventh, and Eleventh’s Circuit’s reading of
§ 2254(d)(1) is unconstitutional); Daniel Meltzer, Congress, Courts, and Constitutional
Remedies, 86 GEO. L.J. 2537, 2537-38 (1998) (discussing the scope of congressional
authority to define and limit judicial redress); Note, Powers of Congress and the Court
Regarding the Availability and Scope of Review, 114 HARV. L. REV. 1551 (2001)
(examining whether Congress can require courts to defer on direct review). But see
Scheidegger, supra note 12, at 891 (arguing that AEDPA’s limitation is constitutional).
41. For the most part, before the Supreme Court’s decision in Williams, courts
assumed the AEDPA codified Teague v. Lane’s, 489 U.S. 288 (1989), anti-retroactivity
doctrine. Accordingly, most lower courts declared that to be “clearly established,” a rule
must be compelled by existing precedent of the Supreme Court, which was required under
Teague. See, e.g., Morgan v. Krenke, 232 F.3d 562, 567 (7th Cir. 2000); Lindh, 96 F.3d at
889; Walker v. McCaughtry, 72 F. Supp. 2d 1025, 1031 (E.D. Wis. 1999); Breedlove v.
Moore, 74 F. Supp. 2d 1226, 1231 (S.D. Fla. 1999). In Williams, however, a majority of the
Supreme Court did not agree that AEDPA simply codified Teague. 529 U.S. at 379-80
(opinion of Stevens, J.). See infra Part II.C.1, for a discussion of Williams and infra Part
II.A, for a discussion of Teague.
This is not to say that commentators did not discuss the language or contemplate its
meaning. But they did so in conjunction with a broader focus on the statute. See, e.g.,
Chen, supra note 12, at 552-53; Lee, supra note 12, at 117-35; Khandelwal, supra note 12,
at 439-45.
42. 529 U.S. 362 (2000).
43. 538 U.S. 63 (2003).
44. 124 S. Ct. 2140 (2004).

2005]

Seeking Clarity in the Federal Habeas Fog

757

45

“baseline interpretation” and illustrate how the clearly established law
analysis works in tandem with the standards of review under §
2254(d)(1).
A. The Pre-AEDPA World: Teague v. Lane’s Anti-retroactivity Rule
The habeas fog rolled in before AEDPA; its roots are in the anti46
retroactivity rule of Teague v. Lane and its progeny. The Teague
doctrine (as it has come to be called) is relevant to understanding
AEDPA’s clearly established law limitation because both Teague’s antiretroactivity rule and AEDPA § 2254(d)(1) premise habeas relief on the
existence of precedent at a particular point in time. Their focus on
timing reflects a critical distinction between direct appellate review and
habeas review. On direct review, courts apply the principles of federal
47
law at the time of appellate decision. On habeas review, a question
arises when the habeas claim proposes an expansion of federal rights
from the precedent as to whether the Court should apply the right to
48
pending cases.
49
Described as an “intellectual disaster area,” the history of the Teague
50
doctrine has been chronicled comprehensively by others. This Article
45. I borrow the term “baseline interpretation” from Professor Allan Ides. See Ides,
supra note 11, at 679. Professor Ides asserts that the Supreme Court provided a baseline
interpretation of § 2254(d)(1) in Williams. My use of the term is narrower in the sense
that I am referring only to the “clearly established Federal law, as determined by the
Supreme Court of the United States” language, and it is broader in the sense that the
baseline definition for this language is synthesized from several Supreme Court cases,
including but not limited to Williams.
46. 489 U.S. 288 (1989) (plurality opinion).
47. See Yackle, supra note 4, at 414. Professor Yackle notes that
[t]his is true even if the Court chooses a particular case as the occasion for
announcing a “new” rule of constitutional law. When that happens, the “new”
rule is fully applicable to the case at bar, as well as to any other case that has not
yet reached final disposition on direct review.
Id.
48. See Kaufman v. United States, 394 U.S. 217, 226 (1969); Brown v. Allen, 344 U.S.
443, 462-64 (1953).
49. Yackle, Figure, supra note 12, at 1756.
50. This Article does not attempt to provide a comprehensive review of the Teague
doctrine and its progeny, which continue today; it has been explored in-depth by other
commentators. The primary reaction to Teague was criticism, which provoked a great deal
of scholarship. See, e.g., Marc M. Arkin, The Prisoner’s Dilemma: Life in the Lower
Federal Courts After Teague v. Lane, 69 N.C. L. REV. 371, 380-90 (1991); Susan Bandes,
Taking Justice to Its Logical Extreme: A Comment on Teague v. Lane, 66 S. CAL. L. REV.
2453, 2462-66 (1993); Vivian Berger, Justice Delayed or Justice Denied?—A Comment on
Recent Proposals To Reform Death Penalty Habeas Corpus, 90 COLUM. L. REV. 1665,
1703 (1990); John Blume & William Pratt, Understanding Teague v. Lane, 18 N.Y.U. REV.
L. & SOC. CHANGE 325, 343-56 (1990-1991); David R. Dow, Teague and Death: The
Impact of Current Retroactivity Doctrine on Capital Defendants, 19 HASTINGS CONST.

758

Catholic University Law Review

[Vol. 54:747

does not attempt to duplicate those efforts, but a brief overview of the
doctrine is necessary to understand its relevance to AEDPA.
51
In Teague, the Supreme Court divided on a state prisoner’s habeas
challenge that was based on the racial composition of his petit jury and
the prosecutor’s use of preemptory strikes to exclude African-American
52
53
jurors.
The prisoner relied on the rule from Batson v. Kentucky
54
prohibiting the use of preemptory strikes on the basis of race. The
55
Court decided the case after the prisoner’s conviction became final.
56
The petitioner also sought to extend Taylor v. Louisiana’s “fair cross
section” requirement to the composition of his petit jury, even though
the Taylor Court had expressly limited its holding to the composition of
57
the jury venire.
A plurality of the Court, led by Justice O’Connor, explained that a
decision would be considered to announce a new rule if it required a
court to “break[] new ground or impose[] a new obligation on the States
L.Q. 23, 31-41 (1991); Markus Dirk Dubber, Prudence and Substance: How the Supreme
Court’s New Habeas Retroactivity Doctrine Mirrors and Affects Substantive Constitutional
Law, 30 AM. CRIM. L. REV. 1, 3-9 (1992); Richard H. Fallon, Jr. & Daniel J. Meltzer, New
Law, Non-Retroactivity, and Constitutional Remedies, 104 HARV. L. REV. 1731, 1746-49
(1991); Stephen M. Feldman, Diagnosing Power: Postmodernism in Legal Scholarship and
Judicial Practice, 88 NW. U. L. REV. 1046 (1994); Timothy Finley, Habeas Corpus—
Retroactivity of Post-Conviction Rulings: Finality at the Expense of Justice, 84 J. CRIM. L. &
CRIMINOLOGY 975, 982-88 (1994); Barry Friedman, Habeas and Hubris, 45 VAND. L.
REV. 797, 802-14 (1992); Barry Friedman, Pas De Deux: The Supreme Court and the
Habeas Courts, 66 S. CAL. L. REV. 2467, 2496-501 (1993); Mary C. Hutton, Retroactivity in
the States: The Impact of Teague v. Lane on State Postconviction Remedies, 44 ALA. L.
REV. 421, 427-32 (1993); James S. Liebman, More than “Slightly Retro”: The Rehnquist
Court’s Rout of Habeas Corpus Jurisdiction in Teague v. Lane, 18 N.Y.U. REV. L. & SOC.
CHANGE 537 (1990-1991); Linda Meyer, “Nothing We Say Matters”: Teague and New
Rules, 61 U. CHI. L. REV. 423 (1994); Jordan Steiker, Habeas Exceptionalism, 78 TEX. L.
REV. 1703 (2000); Ann Woolhandler, Demodeling Habeas, 45 STAN. L. REV. 575, 638-41
(1993); Tung Yin, A Better Mousetrap: Procedural Default as a Retroactivity Alternative to
Teague v. Lane and the Antiterrorism and Effective Death Penalty Act of 1996, 25 AM. J.
CRIM. L. 203, 255-81 (1998); Roger D. Branigin III, Note, Sixth Amendment—The
Evolution of the Supreme Court’s Retroactivity Doctrine: A Futile Search for Theoretical
Clarity, 80 J. CRIM. L. & CRIMINOLOGY 1128 (1990); Khandelwal, supra note 12, at 43940; The Supreme Court, 1989 Term—Leading Cases, 104 HARV. L. REV. 308, 309-14
(1990).
51. Although Teague was a plurality decision, a majority of the Court adopted the
rule announced by the Teague plurality in Penry v. Lynaugh, 492 U.S. 302, 330 (1989).
52. Teague v. Lane, 489 U.S. 288, 292 (1989).
53. 476 U.S. 79 (1986).
54. Id. at 80, 96. A lower federal court recently faced the issue of whether to extend
the Batson rule to Italian-Americans. See infra Part III.B (discussing Rico v. LeftridgeByrd, 340 F.3d 178, 187 (3d Cir. 2003)).
55. Teague, 489 U.S. at 295-96.
56. 419 U.S. 522 (1975).
57. Teague, 489 U.S. at 299.

2005]

Seeking Clarity in the Federal Habeas Fog

759

or the Federal Government”—that is, when it would require a result “not
dictated by precedent existing at the time the defendant’s conviction
58
became final.” The plurality found that the petitioner could neither rely
on Batson nor extend Taylor, because to do so would allow him to
invoke a “new” rule of constitutional law that had not been established
59
at the time his conviction became final. Batson was a good example of a
60
new rule because the rule it announced overturned existing precedent.
The plurality interpreted the Habeas Corpus Act (as it then existed) to
prohibit the announcement or application of “new” rules of
61
constitutional law, except in extraordinary circumstances.
While at first Teague’s prohibition appears to be procedural—a timing
inquiry that bars retroactive application of new rules—it draws on
fundamental substantive issues about the nature of our legal system. The
rule announced by the plurality constrained the scope of the habeas
remedy; the consequence of a new rule is no habeas relief. Because rules
can be viewed at varying levels of generality, the Teague Court
recognized that it is not always easy to determine when a new rule has
62
been announced.

58. Id. at 301 (plurality opinion).
59. See id. at 294-96, 315-16.
60. Id. at 295.
61. Yackle, supra note 4, at 414. The Court recognized two exceptions where new
rules could be retroactively applied. The first exception was “if it places ‘certain kinds of
primary, private individual conduct beyond the power of the criminal law-making
authority to proscribe.’” Teague, 489 U.S. at 311 (plurality opinion) (quoting Mackey v.
United States, 401 U.S. 667, 692 (1971) (Harlan, J., concurring)). The second exception
applied “if it requires the observance of ‘those procedures that . . . are implicit in the
concept of ordered liberty.’” Id. (plurality opinion) (omission in original) (quoting
Mackey, 401 U.S. at 693 (Harlan, J., concurring)).
62. Id. at 301 (plurality opinion) (“It is admittedly often difficult to determine when a
case announces a new rule, and we do not attempt to define the spectrum of what may or
may not constitute a new rule for retroactivity purposes.”). Professor Evan Tsen Lee
provided this example:
[I]n Edwards v. Arizona, [451 U.S. 477, 484-85 (1981),] the Court held that a
suspect who had expressed his desire with the police only through counsel may
not be interrogated further until counsel had been made available, unless the
accused himself initiates further communication with the police. Then, in
Arizona v. Roberson, [486 U.S. 675, 682 (1988),] the Court held that a suspect in
this situation may not be interrogated further even as to an offense unrelated to
the subject of the original interrogation. Did Roberson create a new rule or did it
simply flesh out a more general rule . . . ? In Butler v. McKellar, [494 U.S. 407,
414-15 (1990),] the Court held that Roberson did establish a new rule, despite
language in the Roberson opinion that the result therein was logically compelled
by Edwards. Because Roberson was handed down after the petitioner’s
conviction became final, it could not form the basis for habeas relief.
Lee, supra note 12, at 118-19 (footnotes omitted).

760

Catholic University Law Review

[Vol. 54:747

63

Several years later, the Court again divided in Wright v. West. A
majority of the Court agreed that a new rule would be one that the state
court did not feel “compelled by existing precedent to conclude . . . was
64
required by the Constitution.” The Court split, however, on whether
habeas courts should validate “reasonable, good-faith interpretations of
existing precedents made by the state courts even though they are shown
65
to be contrary to later decisions.” Justice O’Connor was joined by three
other justices who rejected deferring to the state court; federal courts
66
must review questions of federal law de novo. Although he did not join
her opinion in full, Justice Kennedy noted in his concurring opinion that
Teague does bear on applications of law to fact which result in
the announcement of a new rule. . . . If the rule in question is
one which of necessity requires a case-by-case examination of
the evidence, then we can tolerate a number of specific
applications without saying that those applications themselves
67
create a new rule.
Three justices, led by Justice Thomas, favored an approach where
federal habeas courts “‘must defer to the state court’s decision rejecting
68
the [petitioner’s] claim unless that decision is patently unreasonable.’”
Justice Thomas’s test for a new rule would be based on this deference:
“[I]f a state court has reasonably rejected the legal claim asserted by a
habeas petitioner under existing law, then the claim seeks the benefit of a
69
‘new’ rule . . . and is therefore not cognizable on habeas under Teague.”
This reading was grounded in a respect for the states’ strong interest in
the finality of criminal convictions and the recognition that a state court
should not be penalized for relying on “‘the constitutional standards that
70
prevailed at the time the original proceedings took place.’”
Accordingly, a rule is “new” if, at the time the prisoner’s conviction
71
became final, the rule was one about which reasonable jurists disagree.
Although not joining Justice Thomas’s opinion, Justice Souter read
Teague to hold that the “unlawfulness [of the conviction] must be
63. 505 U.S. 277 (1992) (plurality opinion).
64. Id. at 310 (Souter, J., concurring in the judgment).
65. Id. at 311 (Souter, J., concurring in the judgment).
66. Id. at 300 (O’Connor, J., concurring). Justice Kennedy cast a fourth vote in
support of this narrow reading, but wrote a concurring opinion. Id. at 306 (Kennedy, J.,
concurring in the judgment).
67. Id. at 308 (citation omitted) (Kennedy, J., concurring in the judgment).
68. Id. at 291 (plurality opinion) (quoting Butler v. McKellar, 494 U.S. 407, 422
(1990)).
69. Id. (plurality opinion).
70. Teague v. Lane, 489 U.S. 288, 307 (1989) (plurality opinion) (quoting Desist v.
United States, 394 U.S. 244, 263 (1969) (Harlan, J., dissenting)).
71. Wright, 505 U.S. at 291 (plurality opinion).

2005]

Seeking Clarity in the Federal Habeas Fog

761

72

apparent” for a prisoner to be entitled to habeas relief. He reasoned
that
our cases have recognized that “[t]he interests in finality,
predictability, and comity underlying our new rule
jurisprudence may be undetermined to an equal degree by the
invocation of a rule that was not dictated by precedent as by the
application of an old rule in a manner that was not dictated by
73
precedent.”
Following Wright, the Supreme Court and lower federal courts used
language consistent with Justice Thomas’s approach. But the evolution
of the Teague doctrine (and the resolution of confusion surrounding it)
74
was cut short by the enactment of AEDPA in 1996.
B. Statutory Text and Relevant Legislative History
Section 2254(d)(1) sets out a two-pronged approach: a federal court
may grant habeas relief if either the state court decision was “contrary to
clearly established Federal law, as determined by the Supreme Court” or
the state court decision “involved an unreasonable application of clearly
75
established Federal law, as determined by the Supreme Court.” The
existence of “clearly established” law is a requirement for relief under
either prong. Understanding what qualifies as “clearly established” law
is therefore critical to the operation of the statutory provision as a whole.
None of the key phrases in § 2254(d)(1) are defined in AEDPA.
76
While the legislative history does shed some light on how the standard
77
of review provisions were intended to operate, it does not specifically
72. Id. at 313 (Souter, J., concurring in the judgment).
73. Id. (Souter, J., concurring in the judgment) (alteration in original) (quoting
Stringer v. Black, 503 U.S. 222, 228 (1992)).
74. Despite speculation to the contrary, the Supreme Court has recently recognized
that the Teague doctrine survived AEDPA’s enactment. See Horn v. Banks, 536 U.S. 266
(2002) (per curiam) (applying Teague but not clarifying the meaning of new rules); see also
Bryant, supra note 6, at 18 n.113.
75. 28 U.S.C. § 2254(d)(1) (2000).
76. I recognize that the use of legislative history in statutory interpretation is
controversial in some circles. At least two Supreme Court justices, Justice Thomas and
Justice Scalia, eschew the use of legislative history. See Bradford C. Mank, Legal Context:
Reading Statutes in Light of Prevailing Legal Precedent, 34 ARIZ. ST. L.J. 815, 825, 826 &
n.60; Antonin Scalia, COMMON-LAW COURTS IN A CIVIL-LAW SYSTEM: The Role of
United States Federal Courts in Interpreting the Constitution and Laws, in A MATTER OF
INTERPRETATION 3, 17 (Amy Gutmann ed., 1997). A majority of the Court, however,
appears to find legislative history helpful in statutory interpretation when the statute’s
language is ambiguous. See, e.g., Williams v. Taylor, 529 U.S. 362, 378 n.10, 408 (2000)
(using legislative history to interpret AEDPA).
77. The most controversial issue debated with respect to § 2254 was the meaning of
the “unreasonable application” standard and whether it required federal courts to defer to

762

Catholic University Law Review

[Vol. 54:747

address the meaning of “clearly established Federal law as determined by
78
the Supreme Court.”
The closest reference to the clearly established language was a
reference to the governmental qualified immunity doctrine in tort law
when Senator Orrin Hatch, the primary sponsor of § 2254, referred to
79
Harlow v. Fitzgerald.
Arguably, that reference was not made as a

state court decisions. See Carrie M. Bowden, Note, The Need for Comity: A Proposal for
Federal Court Review of Suppression Issues in the Dual Sovereignty Context After the
Antiterrorism and Effective Death Penalty Act of 1996, 60 WASH. & LEE L. REV. 185, 22526 (2003) (describing in detail floor debates on the issue); see also Bryant, supra note 6, at
27-28; Ides, supra note 11, at 693-97 (discussing the legislative history of the “unreasonable
application” language and arguing the congruency between the legislative history and the
text); Tushnet & Yackle, supra note 12, at 5-21 (exploring political pressures leading to
AEDPA’s enactment); Yackle, supra note 4, at 436-38. Professor Yackle argues that the
legislative history demonstrates that AEDPA was the product of three significant
compromises. These compromises abandoned previous legislative efforts to remove
federal court jurisdiction over habeas petitions from state convicts, to give state court
decisions on the merits preclusive effect in federal court, and to require federal courts to
defer to “reasonable” state court decisions on questions of law and mixed questions of law
and fact. Id. at 436-37.
78. 28 U.S.C. § 2254(d)(1); see Ides, supra note 11, at 693-97; Yackle, supra note 4, at
437-39; Bowden, supra note 77, at 225-26.
79. Yackle, supra note 4, at 406 & n.89, 407 n.90 (citing 141 CONG. REC. S7848 (daily
ed. June 7, 1995) (statement of Sen. Hatch)). The qualified immunity doctrine requires
federal courts to engage in an analysis similar to that required by Teague and AEDPA §
2254(d)(1): to determine whether a constitutional right had been “clearly established” by a
particular point in time. Cases involving the affirmative defense of qualified immunity
arise out of claims of constitutional violations under 42 U.S.C. § 1983, which allows a
plaintiff to recover civil damages from a defendant government official. In the seminal
decision of Harlow v. Fitzgerald, 457 U.S. 800 (1982), the Court announced that a
government official performing a discretionary function is entitled to qualified immunity
from a claim under 42 U.S.C. § 1983 if there was no “clearly established” law at the time
the official violated the constitutional rights of the plaintiff, id. at 818.
Since Harlow, the Court has intermittently attempted to clarify the meaning of “clearly
established” law. See Anderson v. Creighton, 483 U.S. 635 (1987); Davis v. Scherer, 468
U.S. 183 (1984). In doing so, the Court has stressed the competing policy concerns in
qualified immunity:
When government officials abuse their offices, “action[s] for damages may
offer the only realistic avenue for vindication of constitutional guarantees.” On
the other hand, permitting damages suits against government officials can entail
substantial social costs, including the risk that fear of personal monetary liability
and harassing litigation will unduly inhibit officials in the discharge of their
duties.
Anderson, 483 U.S. at 638 (alteration in original) (quoting Harlow, 457 U.S. at 814).
Recognizing the need to balance the right of a plaintiff to bring constitutional violation
claims under 42 U.S.C. § 1983 with the right of an official to perform job duties without
the fear of possible litigation, the Court announced that “[t]he contours of the right must
be sufficiently clear that a reasonable official would understand that what he is doing
violates that right.” Id. at 640 (emphasis added). However, at the time AEDPA was

2005]

Seeking Clarity in the Federal Habeas Fog

763

enacted, the Court had failed to elaborate on what exactly created a “sufficiently clear”
right.
Two recent decisions, United States v. Lanier, 520 U.S. 259 (1997), and Hope v. Pelzer,
536 U.S. 730 (2002), elaborate on the meaning of clearly established law in the qualified
immunity context. In Lanier, the defendant was convicted under 18 U.S.C. § 242 for
criminally violating the constitutional rights of five women by sexually assaulting them
while he served as a state judge. 520 U.S. at 261. The Sixth Circuit dismissed the
indictment, finding that criminal liability may be imposed under 18 U.S.C. § 242, only if
“the constitutional right said to have been violated is first identified in a decision of this
Court (not any other federal, or state, court), and only when the right has been held to
apply in a ‘factual situation fundamentally similar to the one at bar.’” Id. at 263 (quoting
United States v. Lanier, 73 F.3d 1380, 1393 (6th Cir. 1996), vacated by 520 U.S. 259
(1997)). The Sixth Circuit, following a narrow view, found “no decision of this Court
applying a right to be free from unjustified assault or invasions of bodily integrity in a
situation ‘fundamentally similar’ to those charged” and thus dismissed the charges against
the defendant. Id. Applying the clearly established law standard for qualified immunity
under 42 U.S.C. § 1983, the Supreme Court reversed. Id. at 270. The Court first found
that nothing in its precedent required the narrow “fundamentally similar” situation for
determining whether the conduct at issue violated constitutional rights. Id. at 269. The
Court rejected the Sixth Circuit’s use of the “fundamentally similar” standard because it
would “lead trial judges to demand a degree of certainty at once unnecessarily high” and
found that claims under § 242 needed nothing more than the “clearly established” or “fair
warning” standard. Id. at 270. The Court, announced:
[A]s with civil liability under § 1983 . . . all that can usefully be said about
criminal liability under § 242 is that it may be imposed for deprivation of a
constitutional right if, but only if, “in the light of pre-existing law the
unlawfulness [. . . is] apparent.” Where it is, the constitutional requirement of
fair warning is satisfied.
Id. at 271-72 (second alteration in original) (quoting Anderson, 483 U.S. at 640). The
Court also held that prior Supreme Court precedent was not necessary to give “fair
warning” to an established right. Id. at 271.
The Court further clarified the clearly established law definition standard in Hope. The
plaintiff, a prison inmate, brought a claim under 42 U.S.C. § 1983 for violation of his
Eighth Amendment rights against cruel and unusual punishment because he was
handcuffed by prison guards to a hitching post for seven hours in the hot sun. Hope, 536
U.S. at 734-35. The Eleventh Circuit found that the use of the hitching post violated the
plaintiff’s Eighth Amendment rights. Id. at 736. However, it granted the defendant prison
guards qualified immunity after stating that “‘the federal law by which the government
official’s conduct should be evaluated must be preexisting, obvious and mandatory,’ and
established, not by ‘abstractions,’ but by cases that are ‘materially similar’ to the facts in
the case in front of us.” Id. (emphasis added) (internal quotation marks omitted) (quoting
Hope v. Pelzer, 240 F.3d 975, 981 (11th Cir. 2001), rev’d, 536 U.S. 730 (2002)). The
Supreme Court rejected the Eleventh Circuit’s “materially similar” standard. Id. at 741.
The Court recognized that it is “clear that officials can still be on notice that their conduct
violates established law even in novel factual circumstances.” Id. The Court found ample
evidence that the law was clearly established enough to put the defendant prison guards
on notice. It determined that the law was clearly established based on binding Eleventh
Circuit precedent, an Alabama Department of Corrections regulation, and a Department
of Justice report, all of which concluded that the use of handcuffing to a fence or hitching
post was unconstitutional. Id. at 744.
For a comprehensive review of the qualified immunity doctrine, see Alan K. Chen, The
Burdens of Qualified Immunity: Summary Judgment and the Role of Facts in Constitutional

764

Catholic University Law Review

[Vol. 54:747

suggestion to adopt the same “clearly established” standard from the
qualified immunity context, but rather to “compare[] the deference
provisions in the habeas reforms to the preference for reasonableness
80
that courts give in other areas of the law.” Accordingly, the Court had
to shed light on the meaning of “clearly established” law in the habeas
context and it did so in a series of decisions beginning in 2000.
C. The Supreme Court Speaks
The Supreme Court addressed the proper construction of § 2254(d)(1)
81
for the first time in Williams, four years after AEDPA’s enactment.
While the determination of clearly established law was not a primary
issue in Williams, the Court took the opportunity to define this phrase
along with the other language in the section. The Williams Court also
clarified § 2254(d)(1)’s relationship to the Teague doctrine. Although a
majority of the Court did not accept either as a complete analogue, it did
declare an explicit connection to Teague’s concept of “old” rules and
82
clearly established law. It was not until 2003 in Lockyer v. Andrade that
the Court directly confronted the issue of what constitutes clearly
83
established law. Most recently, in Yarborough v. Alvarado, the Court
explored the relationship between clearly established law and the
84
unreasonable application of such law. These cases illustrate that the
clarity of the law is relevant under § 2254(d)(1), and that the Supreme
Court has interpreted this provision relatively broadly.

Torts Law, 47 AM. U. L. REV. 1 (1997); Kit Kinports, Habeas Corpus, Qualified Immunity,
and Crystal Balls: Predicting the Course of Constitutional Law, 33 ARIZ. L. REV. 115
(1991); Heather Meeker, “Clearly Established” Law in Qualified Immunity Analysis for
Civil Rights Actions in the Tenth Circuit, 35 WASHBURN L.J. 79 (1995); Recent Case,
Jenkins v. Talladega City Board of Education, 115 F.3d 821 (11th Cir.) (en banc), cert.
denied, 118 S. Ct. 412 (1997), 111 HARV. L. REV. 1341 (1998); see also infra Part II.C.1.a.
80. Bowden, supra note 77, at 226. This conclusion was generally accepted by
commentators. See Yackle, supra note 4, at 406; Khandelwal, supra note 12, at 440 (noting
that “[a]ssessing whether precedent ‘dictates’ a rule is a reasonableness inquiry”); see also
infra Part II.C.1.a.
81. Williams v. Taylor, 529 U.S. 362, 367 (2000).
82. 538 U.S. 63 (2003).
83. 124 S. Ct. 2140 (2004).
84. The Court will be deciding Goughnour v. Payton, 124 S. Ct. 2388 (2004) (mem.),
during the 2004-2005 Term. The case is an appeal from a Ninth Circuit decision holding
that the California Supreme Court unreasonably applied clearly established precedent that
California’s “catch-all” mitigation instruction in capital cases is constitutional as applied to
post-crime evidence of mitigation. Payton v. Woodford, 346 F.3d 1204, 1206-07 (9th Cir.
2003), cert. granted, 124 S. Ct. 2388 (2004). This case should add to the understanding of
the relationship between § 2254(d)(1)’s clearly established law limitation and the
unreasonable application prong.

2005]

Seeking Clarity in the Federal Habeas Fog

765

1. Williams v. Taylor
In Williams, the Supreme Court for the first time defined the terms of
85
§ 2254(d)(1) in the context of a habeas appeal. Terence Williams was
86
convicted by a Virginia jury of robbery and capital murder. Sentenced
to die, he filed a state habeas corpus petition alleging that he received
87
ineffective assistance of counsel. The Virginia circuit court granted his
habeas claim with regard to the sentencing phase of the trial and
88
recommended granting a rehearing for the sentencing phase.
The
Virginia Supreme Court, however, rejected the circuit court’s
89
recommendation, thus rejecting Williams’s habeas claim. Williams then
90
filed a habeas corpus petition in federal district court.
On federal
habeas review, Williams contended that he had been deprived of his
Sixth Amendment right to effective assistance of counsel, under the
91
standard established in Strickland v. Washington, during his sentencing
92
93
proceedings. The district court agreed and ordered the writ granted.
The Fourth Circuit reversed, finding that the Virginia Supreme Court’s
decision was not “contrary to” and did not involve an “unreasonable
94
application” of clearly established law under § 2254(d)(1).
95
The Supreme Court reversed the Fourth Circuit. In doing so, the
Court divided into three opinions with a shifting majority. Justice
O’Connor wrote the majority opinion regarding the proper construction
96
of § 2254(d)(1). Justice Stevens wrote separately to address how §
2254(d)(1)’s scope of review applied to the decision issued by the
97
Virginia Supreme Court in Williams’s case. A majority of the justices
98
joined him in concluding that Williams satisfied § 2254(d)(1)’s standard.
The third opinion was that of Chief Justice Rehnquist concurring in part
99
and dissenting in part, joined by Justices Scalia and Thomas.

85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.

Williams, 529 U.S. at 370-72.
Id. at 367-68.
Id. at 370.
Id. at 371.
Id.
Id. at 372.
466 U.S. 668 (1984).
Williams, 529 U.S. at 373-74.
Id.
Id. at 374.
Id.
Id. at 402-13.
Id. at 390-99.
Id. at 367, 390-91.
Id. at 416-19 (Rehnquist, C.J., concurring in part and dissenting in part).

766

Catholic University Law Review

[Vol. 54:747

a. “Clearly Established Law as Determined by the Supreme Court of
the United States”
Even though Williams did not turn on the question of whether there
was “clearly established” law, the Court took this opportunity to provide
a baseline definition. The Court announced that whether there was
100
relevant clearly established law was a “threshold question.”
Beyond
that statement, however, the Court was not in full agreement on how to
interpret the statutory language. A majority of the Court agreed on
three criteria for “clearly established” law under § 2254(d)(1). First, a
majority agreed with Justice O’Connor’s definition of the phrase as
referring “to the holdings, as opposed to the dicta, of [the Supreme]
101
Court’s decisions.” Second, a majority agreed that the time limitation
102
was the date of the relevant state court decision.
Third, a majority
agreed that the clause “as determined by the United States Supreme
Court” limited the source of “clearly established Federal law” to
103
Supreme Court precedent.
These last two points differed significantly from pre-AEDPA law.
Before AEDPA, the timing of the relevant precedent and the source of
that precedent were governed by the anti-retroactivity rule articulated in
104
Teague. In Teague, the Court held that the habeas petitioner was not
entitled to relief because he relied on a rule that was announced after his
105
conviction became final.
Likewise, the source of the precedent that
could be used by the habeas petitioner was broader; it included federal
106
court of appeals decisions as well as Supreme Court decisions.
Although a majority of the Court recognized only a “slight connection”
between its Teague jurisprudence and § 2254(d)(1), it rejected the view
that Congress’s sole intention in enacting § 2254(d)(1) was to codify
107
Teague.
Rather than finding that the two were equivalent, Justice
O’Connor explained that at least “whatever would qualify as an old rule
100. Id. at 390.
101. Id. at 412. The Court elsewhere has defined a holding as “not only the result but
also those portions of the opinion necessary to that result.” Seminole Tribe of Fla. v.
Florida, 517 U.S. 44, 67 (1996); accord Cent. Green Co. v. United States, 531 U.S. 425, 431
(2001) (stating that a sentence is dictum when it was not essential to the disposition of the
contested issues); Humphrey’s Ex’r v. United States, 295 U.S. 602, 626 (1935) (noting that
an expression by the Court that goes beyond the point actually decided is not a holding).
102. Williams, 529 U.S. at 412 (O’Connor, J., writing for the majority).
103. Id.
104. Teague v. Lane, 489 U.S. 288, 299 (1989); see supra Part II.A.
105. Teague, 489 U.S. at 311 (plurality opinion). Under Teague, the timing of the
relevant precedent that a habeas petitioner could invoke was arguably slightly longer
because it was based on the state conviction becoming final. Id. (plurality opinion).
106. Williams, 529 U.S. at 412 (O’Connor, J., writing for the majority).
107. Id. (O’Connor, J., writing for the majority).

2005]

Seeking Clarity in the Federal Habeas Fog

767

under [the Court’s] Teague jurisprudence will constitute ‘clearly
established Federal law, as determined by the Supreme Court of the
108
United States.’”
She placed a caveat on that analogy, limiting the
109
source of clearly established law to Supreme Court precedent.
Justice Stevens, in the portion of his opinion that was not joined by a
majority, found that § 2254(d)(1)’s clearly established law requirement
110
was the “functional equivalent” of the Teague anti-retroactivity rule.
He declared that Teague’s “new rule” doctrine was the “conceptual twin”
111
of § 2254(d)(1)’s clearly established law requirement.
He further
explicated that clearly established law may be “expressed in terms of a
112
generalized standard rather than as a bright-line rule.”
Under this
definition, a general rule that requires a case-by-case application can
constitute clearly established law.
Justice Stevens also noted that his definition was consistent with the
view articulated by Justice Kennedy in the pre-AEDPA Teague doctrine
113
case of Wright v. West. In Wright, Justice Kennedy explained that some
rules are general principles requiring case-by-case application to facts:
“‘Where the beginning point is a rule of this general application, a rule
designed for the specific purpose of evaluating a myriad of factual
contexts, it will be the infrequent case that yields a result so novel that it

108. Id.
109. Id.
110. Id. at 379 (opinion of Stevens, J.). Justice Stevens noted that “[i]t is not unusual
for Congress to codify earlier precedent in the habeas context.” Id. at 380 n.11 (opinion of
Stevens, J.). Simply because Congress used the phrase “clearly established” law rather
than “new rule” does not mean that Congress did not intend to codify Teague’s antiretroactivity rule. Id. n.12 (opinion of Stevens, J.). Moreover, Justice Stevens rejected the
respondent’s argument that Congress meant to import an aspect of governmental qualified
immunity doctrine because that “doctrinally distinct” area employs a “clearly established
law” test. Id. (opinion of Stevens, J.). In doing so, Justice Stevens agreed with the
majority of commentators who considered the relationship between qualified immunity
and § 2254(d)(1) before Williams was decided. Most commentators argued that the
underlying policy rationales for § 2254(d)(1) and qualified immunity were too disparate to
treat them equally. See, e.g., Lee, supra note 12, at 123-24 (rejecting the possibility that
clearly established under AEDPA has the same meaning as in the qualified immunity
context); Yackle, supra note 4, at 406 (arguing that “there is not a stitch of evidence that . .
. Congress proceeded with the immunity cases in view”); Khandelwal, supra note 12, at
441 n.48 (arguing that AEDPA borrows from qualified immunity doctrine the “clearly
established” law language and operating mechanism, but not its objective). But see
Scheidegger, supra note 12, at 925 (arguing that qualified immunity is closely analogous to
§ 2254(d)(1)’s clearly established law requirement). Justice O’Connor and the Chief
Justice simply ignored the qualified immunity argument.
111. Williams, 529 U.S. at 380 n.12 (opinion of Stevens, J.).
112. Id. at 382 (opinion of Stevens, J.).
113. Id. (opinion of Stevens, J.); Wright v. West, 505 U.S. 277 (1992); see also supra
Part II.A.

768

Catholic University Law Review

[Vol. 54:747

114

forges a new rule, one not dictated by precedent.’”
Thus, it seems,
even though not all rules of law are articulated with the same degree of
precision, broader rules may still be sufficiently clear for habeas
115
purposes.
b. The Standards of Review: The Meaning of “Contrary to” and
“Unreasonable Application”
Justice O’Connor, joined by five justices, found that the construction
of § 2254(d) “places a new constraint” on federal habeas courts’ ability to
116
review state courts’ applications of law to fact.
In the portion of her
opinion joined by the majority, she concluded that the “contrary to” and
“unreasonable application” clauses in § 2254(d)(1) have independent
117
meanings.
She then explained each clause in turn. A state court’s
ruling can be “contrary to” clearly established Supreme Court precedent
in two different ways. The first is when the state court “applies a rule
118
that contradicts the governing law set forth in [Supreme Court] cases.”
The second is when the state court “confronts a set of facts that are
materially indistinguishable from a decision of this Court and
nevertheless arrives at a result different from [Supreme Court]
119
precedent.”
Justice O’Connor next defined the “unreasonable application” clause
to mean that a state court’s application of Supreme Court precedent is
unreasonable where the state court’s decision “identifies the correct
governing legal rule . . . but unreasonably applies it to the facts of the
120
particular state prisoner’s case.”
The standard, she stressed, is
121
unreasonable, not erroneous or incorrect.
She clarified: “[A] federal
habeas court may not issue the writ simply because that court concludes
in its independent judgment that the relevant state-court decision applied
114. Williams, 529 U.S. at 382 (opinion of Stevens, J.) (quoting Wright, 505 U.S. at 30809).
115. I intentionally use the phrase “sufficiently clear” to suggest that, although §
2254(d)(1)’s requirement is broader than the clearly established law requirement in the
qualified immunity context, it shares the concept that there is a spectrum of clearly
established law. In the qualified immunity context, the law must be “sufficiently clear” to
provide “fair warning” to a reasonable state official. See supra note 79. In the habeas
context, the audience shifts to judges, who are better equipped to ascertain the clarity of
the law.
116. Williams, 529 U.S. at 412 (O’Connor, J., writing for the majority).
117. Id.
118. Id. at 405-06 (O’Connor, J., writing for the majority).
119. Id. at 406 (O’Connor, J., writing for the majority).
120. Id. at 407 (O’Connor, J., writing for the majority).
121. Id. at 411-12 (O’Connor, J., writing for the majority) (stating that “an
unreasonable application of federal law is different from an incorrect . . . application of
federal law”).

2005]

Seeking Clarity in the Federal Habeas Fog

769

clearly established federal law erroneously or incorrectly. Rather, that
122
application must also be unreasonable.”
Thus, with respect to this prong of § 2254(d)(1), the Court declared
that a federal habeas court could no longer reject a state court’s
interpretation or application of federal constitutional law under a de
novo standard of review. In dicta, the Court declined to adopt the rest of
the Fourth Circuit’s definition of “unreasonable application,” which
included extending a legal principle from Supreme Court precedent to a
123
new context where it should not apply.
Deciding it did not need to
reach that issue, the Court recognized that this “classification does have
124
some problems of precision.” Shortly after the Court decided Williams,
a plurality of the Court adopted this second way for a state court to
125
unreasonably apply the law to facts under § 2254(d)(1).
c. Application of § 2254(d)(1) to Williams’s Case
Justice Stevens authored the majority opinion applying § 2254(d)(1) to
126
the Virginia Supreme Court’s decision to the facts of Williams’s case.
A different six-justice majority of the Court, which included Justice
O’Connor, concluded that in Williams’s case, the state court’s denial of
Williams’s ineffective assistance of counsel claim was both contrary to
127
and an unreasonable application of the clearly established law.
Justice Stevens began the analysis by identifying the clearly established
law relevant to Williams’s claim. This “threshold question” was “easily
answered” because Williams’s claim of ineffective assistance of counsel
was “squarely governed” by the Court’s holding in Strickland v.
128
Washington. Although the test articulated in Strickland was a standard
that required a case-by-case application, that “obviate[d] neither the
clarity of the rule nor the extent to which the rule must be seen as
122. Id.
123. Id. at 408 (O’Connor, J., writing for the majority).
124. Id. at 408-09 (O’Connor, J., writing for the majority).
125. Ramdass v. Angelone, 530 U.S. 156 (2000) (plurality opinion). In Ramdass, the
petitioner had argued that the state court should have applied a Supreme Court case
(Simmons v. South Carolina, 512 U.S. 154 (1994)) to his case, Ramdass, 530 U.S. at 164-65
(plurality opinion). While a plurality of the court found that a federal court may grant
habeas relief if, “under clearly established federal law, the state court was unreasonable in
refusing to extend the governing legal principle to a context in which the principle should
have controlled,” in petitioner’s case the state court had not acted unreasonably when it
refused to extend the precedent to the facts of his case. Id. at 166.
126. Williams, 529 U.S. at 390-91.
127. Id. at 411-13.
128. Id. at 390; Strickland v. Washington, 466 U.S. 668 (1984). Justice Stevens also
noted that “[i]t is past question that the rule set forth in Strickland qualifies as ‘clearly
established Federal law, as determined by the Supreme Court of the United States.’”
Williams, 529 U.S. at 391.

770

Catholic University Law Review

[Vol. 54:747

129

‘established’ by [the] Court.”
Using language from Teague, Justice
Stevens noted that the Court’s precedent “dictated” that the state court
apply Strickland to Williams’s ineffective assistance claim, and that
recognizing the right to counsel did not “‘break[] new ground or impose[]
130
a new obligation on the States.’”
Concluding that the state court’s decision was contrary to clearly
established federal law, the Court found that the state court had
“mischaracterized” the Strickland rule by incorrectly interpreting a
subsequent Supreme Court case as imposing an additional prejudice
131
requirement for the second part of the Strickland standard.
Because
this case did not modify the Strickland standard, the Court held that the
132
state court’s decision was contrary to clearly established law.
The Court also held that the state court’s application of Strickland’s
two-part standard was unreasonable because “it failed to evaluate the
133
totality of the available mitigation evidence” in assessing prejudice.
The Court found that Williams’s attorneys had fallen short of
professional standards in their handling of his sentencing proceedings in
a number of ways. In particular, they failed to examine numerous
records concerning Williams’s “nightmarish childhood,” to introduce
evidence of Williams’s borderline mental retardation, and to obtain
134
evidence of Williams’s favorable conduct in prison.
Mitigating
evidence, the Court noted, even if irrelevant to future dangerousness,
“might [still] have influenced the jury’s appraisal of Williams’s moral
culpability” and may have altered the jury’s selection of penalty, but “it
135
does not undermine or rebut the prosecution’s death-eligibility case.”
The Court found that the state court did not “accord appropriate weight
136
to the body of mitigation evidence available to trial counsel.”
It
therefore held that the state court’s decision that trial counsel’s failures
had not prejudiced Williams was an unreasonable application of clearly
137
established law.
129. Williams, 529 U.S. at 391.
130. Id. (quoting Teague v. Lane, 489 U.S. 288, 301 (1989) (plurality opinion)).
131. Id. at 391, 393, 397. The state court had viewed Lockhart v. Fretwell, 506 U.S. 364
(1993), as modifying the Strickland standard, Williams, 529 U.S. at 391. The Court found
that Lockhart and Nixon v. Whiteside, 475 U.S. 157 (1986), did not “justify a departure
from a straightforward application of Strickland when the ineffectiveness of counsel does
deprive the defendant of a substantive or procedural right to which the law entitles him,”
Williams, 529 U.S. at 393 (Stevens, J., writing for the majority).
132. Williams, 529 U.S. at 397 (Stevens, J., writing for the majority).
133. Id. at 397-98 (Stevens, J., writing for the majority).
134. Id. at 395-96 (Stevens, J., writing for the majority).
135. Id. at 398 (Stevens, J., writing for the majority).
136. Id.
137. Id. at 396-97.

2005]

Seeking Clarity in the Federal Habeas Fog

771

Chief Justice Rehnquist, joined by Justices Scalia and Thomas,
concurred in part and dissented in part. Noting that Strickland was the
clearly established law governing Williams’s claim, he argued that the
state court had reasonably concluded that even if a jury had heard all of
the available mitigating evidence, it would have sentenced Williams to
death because there was strong evidence of Williams’s future
138
dangerousness.
d. The Aftermath
In Williams, the Court resolved several big questions about §
2254(d)(1), including the relationship of the “clearly established” law
requirement to Teague. The majority’s rejection of Teague as a complete
analogue was critical given the similarities between Teague and the
Court’s interpretation of § 2254(d)(1). Both are rooted in policies
encouraging finality and fairness to the state. Teague’s anti-retroactivity
rule prevents innovation: a habeas court cannot announce or apply new
139
doctrines on habeas review.
Section 2254(d)(1) is grounded in the
same policy, perhaps even more obviously so. By limiting the source of
the clearly established law to Supreme Court precedent, Congress
intended to curb the development of constitutional law on habeas
140
review. Consequently, under both Teague and AEDPA, new law can
141
only be developed on direct review.
The Williams Court recognized this connection but stopped short of
142
saying that all “new” rules are not clearly established law.
It instead
carefully declared that clearly established law included at least “whatever
143
would qualify as an old rule” under Teague.
This deliberate
phraseology left the door open to widen the scope of “clearly
established” law beyond rules that would qualify as “new” under Teague.
Notwithstanding this distinction (which I argue is not subtle), the
Court employed the same analytical framework for interpreting what
constitutes “clearly established” law under § 2254(d)(1) as it did under
Teague. Both § 2254(d)(1) and Teague require that to be available to the
habeas petitioner, a rule must have been articulated by the appropriate
source and by a specific point in time. Under Teague, the issue about
138. Id. at 416-19 (Rehnquist, C.J., concurring in part and dissenting in part).
139. See Lee, supra note 12, at 118-19.
140. Id. at 131-32.
141. Cf. Durden v. California, 531 U.S. 1184 (2001) (mem.) (Souter, J., joined by
Breyer, J., dissenting from denial of certiorari) (urging the Court to hear the three-strikes
gross disproportionality issue on direct review because of the “potential for disagreement
over application of” Teague’s review standards).
142. See Williams, 529 U.S. at 412 (O’Connor, J., writing for the majority).
143. Id.

772

Catholic University Law Review

[Vol. 54:747

whether the legal rule the habeas petitioner seeks to apply is a “new”
rule is a threshold question to be addressed before the underlying
144
claim.
Under Williams, the Court indicated § 2254(d)(1) requires a
similar approach: the determination of “clearly established” law is a
145
“threshold question.”
The Court did not, however, define what it
meant by “threshold” other than indicating that the question should be
the first addressed.
The Court left unresolved several other questions about the operation
of § 2254(d)(1), particularly concerning the meaning of “unreasonable
146
application.”
With regard to clearly established law, however, all the
justices agreed that Strickland governed in Williams’s case. Although
Justice O’Connor and Justice Stevens provided definitions that are not
fully reconcilable, particularly with respect to the statutory language’s
relationship to the Court’s Teague jurisprudence, it is instructive to note
how the Court indirectly defined clearly established law in relation to
other language in § 2254(d)(1). Namely, the Court explained that a state
court decision is “contrary to” clearly established law where a state court
issues a decision that is “substantially different from the relevant
precedent of this Court” or uses a “rule that contradicts the governing
147
law set forth in our cases.”
The Court did not use “clearly” or any
other modifier when referring to “precedent of this Court” or “our
148
cases.”
Similarly, when defining the meaning of “unreasonable
application,” the Court announced that the state court may use the
“correct governing legal rule from this Court’s cases” but not reach the
149
same conclusion reached by the Court.
The Court essentially viewed
clearly established law as nothing more than a reasonable application of
the Court’s own precedent.
2. Post-Williams Enlightenment
The Williams Court’s definition of “clearly established” law was a
common-sense, uncontroversial definition centering on the Court’s
holdings in its own precedent. Justice Stevens, however, suggested that
144. See, e.g., Caspari v. Bohlen, 510 U.S. 383, 389 (1994) (stating that a federal habeas
court must ask whether a petitioner’s claim implicates a “new” rule—and is therefore
barred by Teague—before addressing the claim on the merits); see also Kinports, supra
note 79, at 175 n.296; Scheidegger, supra note 12, at 935-36, 940 & n.370, 959 n.500.
145. Williams, 529 U.S. at 390 (Stevens, J., writing for the majority); see also Lockyer
v. Andrade, 538 U.S. 63, 71 (2003).
146. The meaning of unreasonable application has been the focus of several
commentators. See, e.g., Chen, supra note 12 passim; Pettys, supra note 12 passim;
Steinman, supra note 12 passim.
147. Williams, 529 U.S. at 405 (O’Connor, J., writing for the majority).
148. Id. at 406.
149. See id. at 407-08.

2005]

Seeking Clarity in the Federal Habeas Fog

773

150

the clarity of the law, in addition to its establishment, was relevant.
Not until 2003 did the Court have to decide whether the law must be
151
“clear” as well as “established.” In Lockyer v. Andrade, the Court
appeared to answer “yes” to that question. Most recently, in Yarborough
152
v. Alvarado, the Court drew a connection between the clarity of the law
and the reasonableness of the state court’s application of that law.
a. Lockyer v. Andrade
The most significant decision since Williams for defining what
constitutes clearly established law under § 2254(d)(1) is the Supreme
Court’s decision in Andrade. For the first time, the Court faced a body
of law that was considerably less settled than those in other AEDPA
153
cases that had come before the Court.
In 1996, a California state court sentenced Leandro Andrade to a
prison term of two consecutive terms of twenty-five years to life for
154
shoplifting videotapes from two K-Mart stores on two occasions. The
155
merchandise was worth a total of $153.54.
Andrade was sentenced
under the “three strikes” law, which does not allow for the possibility of
parole until after the term of years is served—fifty years in this case.
Andrade will not be eligible for parole until 2046, when he will be eightyseven years old.
Under California law, Andrade’s crimes would normally be classified
as petty theft, which is generally treated as a misdemeanor offense with a
maximum sentence of up to six months in a county jail and up to a $1000
156
fine.
Under the three strikes law, however, which imposes harsher
penalties on repeat offenders, Andrade’s past history of criminal
157
convictions led to the enhancement of his sentence. Because of a prior
misdemeanor conviction, each of his thefts of the videotapes was
150. See id. at 390-91 (Stevens, J., writing for the majority) (noting that the
requirement of applying Strickland on a case-by-case basis did not preclude “the clarity of
the rule nor the extent to which the rule must be seen as ‘established’”).
151. 538 U.S. 63 (2003).
152. 124 S. Ct. 2140 (2004).
153. For instance, in Williams, Strickland’s ineffective assistance of counsel standard
was easily identified as the relevant law. Williams, 529 U.S. at 390-91 (Stevens, J., writing
for the majority). Similarly, in Penry v. Johnson, 532 U.S. 782 (2001), analysis of the
petitioner’s Fifth Amendment claim centered on whether Estelle v. Smith, 451 U.S. 454
(1981), could be distinguished, see Penry, 532 U.S. at 793-95. The analysis of the
petitioner’s Eighth Amendment claim centered on Texas’s compliance with the first Penry
case, Penry v. Lynaugh, 492 U.S. 302 (1989). Penry v. Johnson, 532 U.S. 782, 786 (2001).
154. Andrade, 538 U.S. at 66.
155. Id.
156. Id. at 67; see CAL. PENAL CODE ANN. § 490 (West 1999).
157. Andrade, 538 U.S. at 67; cf. Ewing v. California, 538 U.S. 11 (2003) (reviewing
sentencing under California’s three strikes law on direct review).

774

Catholic University Law Review

[Vol. 54:747

158

classified as a petty theft with a prior.
Such offenses are called
“wobbler” offenses because the prosecutor had discretion whether to
159
charge him with a misdemeanor or a felony.
In Andrade’s case, the
prosecutor opted to charge the two petty thefts with a prior as felonies,
160
triggering the three strikes law.
Under the three strikes law, only those offenses defined as “violent” or
“serious” by statute qualify as prior strikes, but the third strike may be
161
any felony. Andrade’s earlier convictions for residential burglary were
162
counted as his first and second strikes.
Consequently, Andrade’s two
convictions for shoplifting $153.54 worth of videotapes counted as his
163
third and fourth strikes.
For a defendant with two prior strikes, the
three strikes law mandates an indeterminate life sentence with eligibility
164
for parole only after serving at least twenty-five years.
Andrade
165
received two such sentences, to be served consecutively.
Andrade
appealed his sentence and conviction to the California Court of Appeal,
contending that his sentence violated the Eighth Amendment prohibition
166
of cruel and unusual punishment.
The Andrade appeal raised the question of whether, in noncapital
cases, sentences that are grossly disproportionate to the crime committed
violate the Eighth Amendment’s prohibition on cruel and unusual
167
punishment.
Before describing the federal courts’ decisions on direct
appeal and habeas review, this Part reviews some background on the
Supreme Court’s Eighth Amendment jurisprudence. It then traces
Andrade’s appeals through the California and federal court systems.

158. Andrade, 538 U.S. at 67. In a single proceeding, Andrade pled guilty to three
counts of first-degree residential burglary. Id. at 66.
159. Id. at 67.
160. Id. at 67-68. The trial judge also had discretion to change the classification of the
offense at sentencing. Id.
161. Id.; CAL. PENAL CODE ANN. § 667(e)(2)(A) (West 1999). On November 2, 2004,
California voters rejected Proposition 66, which would have changed the three-strikes law
to eliminate non-violent crimes as triggering a third strike. Shawn Steel, Not Just Bush
Scored a Victory on Nov. 2 Schwarzenegger, Businesses and Parties Walked Away Winners,
L.A. DAILY NEWS, Nov. 10, 2004, at N19, 2004 WLNR 7591588.
162. Andrade, 538 U.S. at 68. The jury made a special finding regarding Andrade’s
convictions for three counts of first-degree burglary. Id. Under the three-strikes law, a
first-degree burglary conviction qualifies as a serious or violent felony, which counts as a
strike. Id. (citing CAL. PENAL CODE ANN. §§ 667.5, 1192.7 (West 1999)); see also Ewing,
538 U.S. at 19.
163. Andrade, 538 U.S. at 68.
164. Id. at 67-68 (citing CAL. PENAL CODE ANN. § 667(c)(6), (e)(2)(B) (West 1999)).
165. Id. at 68.
166. Id.
167. Id. at 68-69.

2005]

Seeking Clarity in the Federal Habeas Fog

775

i. The Supreme Court’s Eighth Amendment Jurisprudence
The universe of Supreme Court law dealing with disproportionality in
168
noncapital cases consists of three cases, Rummel v. Estelle, Solem v.
169
170
Helm, and Harmelin v. Michigan, all of which were decided well
before Andrade’s conviction and sentence became final.
The
disagreement about the Court’s jurisprudence in Andrade was about the
state of the law derived from this precedent: how “clearly” the law was
established at the time of Andrade’s direct appeal. Accordingly, this
section examines these three cases in the order of decision.
In Rummel, the Court split five-four and held that a life sentence
imposed under the recidivist statute was not unconstitutionally
disproportionate but acknowledged a general principle of proportionality
171
between crimes and sentences. The petitioner was sentenced under the
Texas mandatory recidivist statute for three felonies involving small
amounts of money, including fraudulent use of a credit card for eighty
dollars, check forgery for $28.36, and obtaining $120.75 by false
172
pretenses.
Without the recidivist statute, the sentence for his third
173
felony would have been ten years.
In deciding that the sentence was
constitutional, the Court reasoned that the legislature should be given
174
great deference in sentencing schemes.
Justice Powell, joined by
Justices Brennan, Marshall, and Stevens, dissented. They urged an
extension of the Court’s proportionality principle from capital cases to
175
noncapital cases.
They maintained that the disproportionality of
Rummel’s sentence should be determined by weighing three factors: (1)
the nature of the offense, (2) the penalties imposed in Texas for similar
offenses, and (3) the penalties imposed in other jurisdictions for the same
176
offense.
The principles outlined in Justice Powell’s dissent in Rummel were
applied three years later when the Court decided Solem, another case
177
involving a nonviolent recidivist.
The Court again split five-four, but
this time it granted relief to a petitioner sentenced to life without parole

168. 445 U.S. 263 (1980).
169. 463 U.S. 277 (1983).
170. 501 U.S. 957 (1991).
171. Rummel, 445 U.S. at 271-73, 285. The majority consisted of Justices Rehnquist,
Burger, White, Stewart, and Blackmun. Id. at 264.
172. Id. at 264-66.
173. Id. at 266.
174. Id. at 283-84.
175. Id. at 306-07 (Powell, J., dissenting).
176. Id. at 295 (Powell, J., dissenting).
177. See Solem v. Helm, 463 U.S. 277, 279, 290-92 (1983).

776

Catholic University Law Review

[Vol. 54:747

178

under South Dakota’s recidivist statute.
The petitioner in Solem
received this life sentence after a conviction for writing a $100 “no
179
account” check.
Without the recidivist statute, this offense was
180
punishable by a maximum of five years and a $5000 fine.
The
petitioner, however, had convictions for six prior nonviolent felonies,
181
including three for third-degree burglary.
Writing for the Court, Justice Powell articulated a three-factor test to
determine if a penalty is grossly disproportionate as requiring the court
to weigh: (1) the “harshness of the penalty” against “the gravity of the
offense,” (2) “sentences imposed on other offenders in the same
jurisdiction” (intra-jurisdictional comparison), and (3) “sentences
imposed for [the] commission of the same offense in other jurisdictions”
182
(inter-jurisdictional comparison). The Court held that the petitioner’s
183
life sentence was excessive under each factor. While specifically noting
that Rummel was still good law, the Court distinguished it based on the
petitioner’s possible eligibility for parole within twelve years in that case
184
and limited it to its facts. The four dissenting justices asserted that this
185
case could not be reconciled with the Court’s decision in Rummel.
The third relevant Eighth Amendment disproportionality decision
came in 1991. In Harmelin, the petitioner was sentenced to life in prison
without parole on a first felony conviction for possession of a substantial
186
amount of cocaine.
A five-justice majority agreed to uphold the
sentence under a Michigan statute mandating life sentences for certain
drug offenses, but it failed to achieve a majority for any single
187
rationale.
Three of the justices, Justices Kennedy, Souter, and O’Connor, joined
a separate opinion written by Justice Kennedy. Justice Kennedy asserted
that the Eighth Amendment and principles of stare decisis “encompass[]
a narrow proportionality principle” that applies to noncapital

178. Id. at 279, 303.
179. Id. at 281-82.
180. Id. at 281 (citing S.D. CODIFIED LAWS § 22-6-1(7) (Michie Supp. 1982)).
181. Id. at 279-80.
182. Id. at 292.
183. Id. at 296-300.
184. Id. at 300-04.
185. Id. at 304 (Burger, C.J., dissenting). Chief Justice Burger authored the dissenting
opinion, which was joined by Justices White, Rehnquist, and O’Connor. Id.
186. Harmelin v. Michigan, 501 U.S. 957, 961 (1991). Petitioner was convicted of
possessing 672 grams of cocaine. Id. Michigan law required a mandatory life sentence for
possession of more than 650 grams of cocaine. Id. n.1 (citing MICH. COMP. LAWS ANN. §
333.7403(2)(a)(i) (West Supp. 1990-1991)).
187. Id. at 961 n.1, 996.

2005]

Seeking Clarity in the Federal Habeas Fog

777

188

sentences.
He noted, however, that the “‘precise contours’” of the
189
principle recognized in the Court’s decisions are “unclear.”
Justice
Kennedy suggested a modification, but not an overruling, of prior case
190
law, particularly Solem.
The modification would require a threshold
inquiry to determine whether, if in light of the gravity of petitioner’s
offense, a comparison of his crime with his sentence gives rise to “an
inference of gross disproportionality” between the harshness of the
191
sentence and the severity of the offense. Only if the inference is drawn
should the reviewing court compare the petitioner’s sentence to other
sentences imposed by the same jurisdiction and to other jurisdictions for
192
similar offenses.
Comprising the rest of the majority, Justices Scalia and Rehnquist
upheld the sentence on the ground that the Eighth Amendment
embodies no proportionality requirement for noncapital sentences
193
imposed pursuant to valid state legislation.
They further argued that
the Court’s prior decisions applying the proportionality principle to
194
noncapital cases were wrong and that Solem should be overruled.
Justice White, joined by Justices Blackmun and Stevens dissented, as
did Justice Marshall separately. All four justices disagreed with Justice
Scalia’s opinion and further insisted that the Eighth Amendment
embodies a stricter proportionality requirement than Justice Kennedy’s
195
plurality opinion acknowledged. They urged keeping the same three196
factor test articulated in Solem.
A splintered decision and the Court’s last word on the subject,
Harmelin reflects the fissures in the Court’s Eighth Amendment
188. Id. at 996-97 (Kennedy, J., concurring in part and concurring in the judgment).
189. Id. at 998 (Kennedy, J., concurring in part and concurring in the judgment).
190. Id. at 1005 (Kennedy, J., concurring in part and concurring in the judgment).
191. Id. at 1004-05 (Kennedy, J., concurring in part and concurring in the judgment).
192. Id. at 1005 (Kennedy, J., concurring in part and concurring in the judgment).
Justice Kennedy’s opinion also outlined four “principles” for determining “gross
disproportionality”: (1) federal courts should defer in general to state legislatures fixing of
prison terms for specific crimes; (2) the Eighth Amendment does not mandate any one
penological theory (such as retribution, deterrence, incapacitation, and rehabilitation); (3)
marked divergences in penological theories and in sentences are inevitable, and sometimes
beneficial, results of the federal system; and (4) federal judicial review should be informed
to the maximum extent possible by “objective factors.” Id. at 998-1001 (Kennedy, J.,
concurring in part and concurring in the judgment).
193. Id. at 994-96.
194. Id. at 962-65. They specifically noted that Rummel would still be good law
because the principle of disproportionality it applied was treated as an aspect of the
Court’s death penalty jurisprudence and not a more generalized Eighth Amendment
principle. Id. at 994.
195. Id. at 1009, 1018-21 (White, J., dissenting); id. at 1027 (Marshall, J., dissenting).
196. Id. at 1009, 1018-21 (White, J., dissenting).

778

Catholic University Law Review

[Vol. 54:747

proportionality requirement for noncapital cases. Left to discern the
holding of this decision, many lower federal courts noted the confused
197
state of the law in this area.
Most courts, however, accepted Justice
198
Kennedy’s concurrence as reflecting the holding of the case.
ii. California State Court Decision
Andrade appealed his sentence and conviction to the California Court
of Appeal, contending that his sentence violated the Eighth Amendment
199
prohibition of cruel and unusual punishment.
That court held that
Andrade’s sentence was not disproportionate to his offense and affirmed
200
the conviction.
It noted that the ongoing validity of Solem’s
proportionality analysis “‘is questionable in light of’ Harmelin v.
197. See, e.g., Hawkins v. Hargett, 200 F.3d 1279, 1281 (10th Cir. 1999) (stating that
Harmelin “fractured” and left the significance of Solem “less than clear”); United States v.
Sarbello, 985 F.2d 716, 723 (3d Cir. 1993) (noting a “lack of clear directive from the
Supreme Court”); Neal v. Grammer, 975 F.2d 463, 465 (8th Cir. 1992) (“[T]he future of
the proportionality test is uncertain.”); McCullough v. Singletary, 967 F.2d 530, 535 (11th
Cir. 1992) (stating that Solem’s continuing viability is called into doubt by Harmelin). A
number of state courts also expressed confusion about the state of the law post-Harmelin.
See, e.g., People v. Gibson, 90 Cal. App. 4th 371, 388 (Cal. Ct. App. 2001) (noting
proportionality review questionable after Harmelin); State v. Brown, 825 P.2d 482, 490
(Idaho 1992) (“Harmelin [is] “fractured . . . .”); State v. Oliver, 745 A.2d 1165, 1169 (N.J.
2000) (“We have generally avoided entering the debate among the several members of the
Supreme Court concerning the Eighth Amendment’s proscription against cruel and
unusual punishment.”); State v. Thorp, 2 P.3d 903, 906 (Or. Ct. App. 2000) (stating that
Harmelin was “severely fractured”); State v. Jones, 543 S.E.2d 541, 545 n.11 (S.C. 2001)
(questioning whether Solem test was still required after Harmelin); State v. Bonner, 577
N.W.2d 575, 579 (S.D. 1998) (noting lack of majority opinion in Harmelin); State v. Harris,
844 S.W.2d 601, 602 (Tenn. 1992) (noting that the “precise contours” of the
proportionality principle were unclear); State v. Bacon, 702 A.2d 116, 122 n.7 (Vt. 1997)
(stating that Harmelin was a “fractured decision” casting doubt on Solem’s viability).
198. Justice Kennedy’s concurrence articulates the test that most circuits regard as the
rule to determine the constitutionality of a sentence under the Eighth Amendment. See,
e.g., Hawkins, 200 F.3d at 1281-82; United States v. Bucuvalas, 970 F.2d 937, 946 n.15 (1st
Cir. 1992), overruled on other grounds by Cleveland v. United Sates, 531 U.S. 12 (2000);
see also People v. Hindson, 703 N.E.2d 956, 965 (Ill. App. Ct. 1998) (noting that Harmelin
“narrowed” Solem); State v. Lee, 841 S.W.2d 648, 654 (Mo. 1992) (stating that Harmelin
“altered” Solem’s test). In doing so, these courts were applying a doctrine articulated in
Marks v. United States, 430 U.S 188 (1977), that, in a fragmented opinion, the holding of
the case is the narrowest grounds on which the concurrences agree, id. at 193. Although
this doctrine does not work with all fragmented opinions, it can be applied to Harmelin’s
opinions. See infra Part III.C.3, for a discussion of the Marks doctrine. Two circuits
followed Solem’s articulation of the principle. E.g., United States v. Rice, 77 Fed. Appx.
692, 698 & n.3 (4th Cir. 2003); Henry v. Page, 223 F.3d 477, 482 (7th Cir. 2000). But cf.
Smallwood v. Johnson, 73 F.3d 1346, 1246 n.4 (5th Cir. 1996) (“Solem was overruled to the
extent that it found in the Eighth Amendment a guarantee of proportionality.”).
199. People v. Andrade, No. F040587, 2003 Cal. App. Unpub. LEXIS 8320, at *44
(Cal. Ct. App. Aug. 28, 2003).
200. Lockyer v. Andrade, 538 U.S. 63, 68-69 (2003).

2005]

Seeking Clarity in the Federal Habeas Fog

779

Michigan,” the high Court’s most recent statement on whether the
Eighth Amendment to the U.S. Constitution includes a proportionality
201
guarantee in a noncapital case.
It then proceeded with its analysis
applying only Rummel; the court did not refer again to Solem or
202
Harmelin.
The state court compared Andrade’s crimes, criminal
history, and sentence with the facts of Rummel and concluded that
203
Andrade’s sentence of fifty years to life was not disproportionate.
Andrade’s subsequent petition for review to the California Supreme
204
Court was summarily denied.
iii. Federal Habeas Petition
205

Andrade’s bid for habeas relief failed at the district court level. The
Ninth Circuit Court of Appeals reversed the district court’s denial and
206
ordered the writ to be issued. After conducting a review of the three
relevant Supreme Court cases, the Ninth Circuit concluded that
207
Andrade’s case was most similar to Solem. Accordingly, the court held
that the state court’s discounting of Solem resulted in a decision that
208
unreasonably applied clearly established law.
In a five-four decision, the Supreme Court reversed the Ninth Circuit,
holding that the state court’s decision affirming petitioner’s two
consecutive terms of twenty-five years to life was not “contrary to” or an
209
“unreasonable application” of clearly established law. Writing for the
majority, Justice O’Connor began the § 2254(d)(1) analysis with the

201. Id. at 68.
202. See id. at 68-69.
203. Id. at 69. The state court did, however, conduct a more thorough analysis under
the California State Constitution, which through case law had been found to proscribe
cruel and unusual punishment when a penalty is “so disproportionate to the crime for
which it is inflicted that it shocks the conscience and offends fundamental notions of
human dignity.” Andrade, 2003 Cal. App. Unpub. LEXIS 8320, at *46 (internal quotation
marks omitted) (citation omitted). It followed a tripartite test that is very similar to the
Solem test. Id.
204. Andrade, 538 U.S. at 69.
205. Id.
206. Id.
207. Id. at 69-70.
208. Id. at 70. Judge Sneed dissented. He would have upheld Andrade’s sentence
because Andrade’s sentence was “‘not one of the “exceedingly rare” terms of
imprisonment prohibited by the Eighth Amendment’s proscription against cruel and
unusual punishment.’” Id. (quoting Andrade v. Attorney Gen. of State of Cal., 270 F.3d
743, 767 (9th Cir. 2002) (Sneed, J., dissenting) (quoting Harmelin v. Michigan, 501 U.S.
957, 1001 (1991) (Kennedy, J., concurring in part and concurring in the judgment)).
209. Id. at 73-74, 77. The Supreme Court consolidated oral arguments in this case with
Ewing v. California, 538 U.S. 11 (2003), which was a direct appeal challenging a California
three-strikes law life sentence on Eighth Amendment grounds, id. at 14.

780

Catholic University Law Review

[Vol. 54:747

210

“threshold matter” of what constitutes clearly established law.
The
Court reiterated the language from Williams, defining clearly established
law as the holdings of Supreme Court precedents, namely, the
“governing legal principle or principles set forth by the Supreme Court at
211
the time the state court renders its decision.”
The Court accurately
noted that, in most cases, the clearly established law inquiry “will be
212
straightforward.”
But in this case, the Court stressed, its Eighth
213
Amendment precedents “have not been a model of clarity.” Here, the
Court recognized a potential problem for both appeals and habeas
review: the Court has “not established a clear or consistent path for
214
courts to follow.”
Continuing its self-critical analysis, the Court extracted the proverbial
pearl from the oyster: “Through this thicket of Eighth Amendment
jurisprudence, one governing legal principle emerges as ‘clearly
established’ under § 2254(d)(1): A gross disproportionality principle is
215
applicable to sentences for terms of years.” The Court noted, however,
that this principle was somewhat vague because the precedents “exhibit a
lack of clarity” about what factors may be relevant to determining a
216
grossly disproportionate sentence.
The Court then quoted Justice
Scalia’s and Justice Kennedy’s concurring opinions in Harmelin, stressing
217
the lack of clarity in the Court’s jurisprudence in this area. Here, the
Court concluded that the only law relevant to the § 2254(d)(1) analysis
was the gross disproportionality principle, the “precise contours of which
are unclear, applicable only in the ‘exceedingly rare’ and ‘extreme’
218
case.”
This was, however, enough to pass through the “clearly
established” law threshold.
The Court next analyzed whether the state court decision was contrary
to or an unreasonable application of the gross disproportionality
principle. The Court found that it was not contrary to clearly established
law for the state court to compare Andrade’s case only to Rummel
because in Solem and Harmelin the Court expressly declined to overrule

210. Andrade, 538 U.S. at 66, 71.
211. Id. at 71-72 (citing Williams v. Taylor, 529 U.S. 362, 405, 412-413 (2000); Bell v.
Cone, 535 U.S. 685, 698 (2002)).
212. Id. at 72.
213. Id.
214. Id. (citing Ewing v. California, 538 U.S. 11, 20-23 (2003)).
215. Id.
216. Id.
217. Id. at 72-73.
218. Id. at 73 (quoting Harmelin v. Michigan, 501 U.S. 957, 1001 (1991) (Kennedy, J.,
concurring in part and concurring in the judgment)); see also Solem v. Helm, 463 U.S. 277,
290 (1983); Rummel v. Estelle, 445 U.S. 263, 272 (1980)).

2005]

Seeking Clarity in the Federal Habeas Fog

781

219

Rummel.
The state court decision also was not contrary to clearly
established law because Andrade’s case was not materially
indistinguishable from the facts in Solem or any other Supreme Court
220
case.
Moving to the next prong of the analysis under § 2254(d)(1), the Court
held that the state court did not unreasonably apply the gross
221
disproportionality principle. Because the Court viewed the governing
legal principle in this case to be broad and lacking “precise contours,” the
Court found that the proportionality principle afforded state legislatures
222
broad discretion to determine constitutional sentences.
After
highlighting that several justices had expressed “uncertainty” about the
application of the proportionality principle to three strikes laws, the
Court concluded that the “contours” of the principle meant that the state
court’s decision to uphold Andrade’s sentence was not objectively
223
unreasonable.
Justice Souter dissented, with Justices Breyer, Ginsburg, and Stevens
joining in his opinion. He asserted that, “[i]f Andrade’s sentence is not
224
grossly disproportionate, the principle has no meaning.” Justice Souter
argued that, despite the breadth of the principle, Andrade was entitled to
225
habeas relief for two independent reasons. First, he argued that Solem
was controlling precedent that required the same result in Andrade’s
226
case. Second, he contended that the state court unreasonably applied
227
the proportionality principle in Andrade’s case. Justice Souter viewed
Andrade’s sentence as a fifty-year sentence rather than as two separate,

219. Andrade, 538 U.S. at 73.
220. Id. at 74. The Court found that Andrade’s case fell between the facts of two
cases: Rummel and Solem. Id. In his dissent, Justice Souter disputed this point and argued
that the similarity between Andrade’s case and the facts in Solem required habeas relief
under § 2254(d)(1) under the contrary to prong. Id. at 78-79 (Souter, J., dissenting).
221. Id. at 77. The Court first rejected the Ninth Circuit’s approach of reviewing the
state court decision de novo before applying the AEDPA standard of review. Id. at 75-76.
222. Id. at 76 (quoting Harmelin, 501 U.S. at 998) (Kennedy, J., concurring in part and
concurring in the judgment)).
223. Id.; cf. Durden v. California, 531 U.S. 1184 (2001) (Souter, J., dissenting from
denial of certiorari).
224. Andrade, 538 U.S. at 83 (Souter, J., dissenting).
225. Id. at 77-78 (Souter, J., dissenting).
226. Id. at 78-79 (Souter, J., dissenting). Although Justice Souter framed the issue as
resulting in the state court’s unreasonable application of “clear law,” see id. (Souter, J.,
dissenting), he appears to be applying the “materially indistinguishable” test under the
“contrary to” prong of § 2254(d)(1). Note Justice Souter’s deliberate use of “clear law” to
rebut the majority’s claim that the law in this area is unclear. Id. at 78 (Souter, J.,
dissenting). Justice Souter also pointed out that the state court’s questioning of Solem’s
authority was wrong as a matter of law. Id. (Souter, J., dissenting).
227. Id. at 79-83 (Souter, J., dissenting).

782

Catholic University Law Review

228

[Vol. 54:747

consecutive sentences of twenty-five years.
Although he noted
legislatures are institutionally better equipped than courts to determine
what the penalty should be for a particular crime, he urged that, in
229
Andrade’s case, the state’s penological policy is unjustified. Doubling
the sentence for a second minor crime committed less than a month after
the first minor crime is “irrational” and “does not raise a seriously
230
debatable point on which judgments might reasonably differ.”
Accordingly, Justice Souter concluded that Andrade’s was “the rare
231
sentence of demonstrable gross disproportionality.”
The Court’s decision in Andrade is important for several reasons.
First, the Court finally faced directly the question of what constitutes
clearly established law. While the Court recognized that its noncapital
Eighth Amendment proportionality decisions were not a “model of
232
clarity,” the Court was able to glean a general principle. This lack of
clarity, therefore, did not mean that there was no clearly established law;
it was not a dispositive issue for the Court. The Court instead shifted the
concern about the clarity of the law to its analysis under the
unreasonable application prong of § 2254(d)(1). By doing so, the Court
suggested that clearly established law is a spectrum running from very
general to very specific principles. As long as the petitioner identifies a
principle in Supreme Court precedent, the petitioner will pass through
the threshold determination, but questions about the clarity and
specificity of the principle will affect the reasonableness of the state
court’s decision applying that principle.
The Court’s decision in Andrade’s companion case, Ewing v.
233
California, highlights the distinction between habeas review and direct
review. In Andrade, the Supreme Court was highly deferential to the
state court because the Court viewed its own decisions as unclear. By
contrast, in Ewing, the Court was forced to reconcile Rummel, Solem,
and Harmelin. In Ewing, the Court went into much more depth about
234
the facts of these cases and the rules articulated in them. Clarification,
therefore, is a job for the Court only on direct review, not habeas review.

228.
229.
230.
231.
232.
233.
234.

Id. at 79-82 (Souter, J., dissenting).
Id. at 80-82 (Souter, J., dissenting).
Id. at 82 (Souter, J., dissenting).
Id. at 83 (Souter, J., dissenting).
Id. at 72.
538 U.S. 11 (2003).
See id. at 21-24.

2005]

Seeking Clarity in the Federal Habeas Fog

783

b. Yarborough v. Alvarado
The most recent case to shed light on the meaning of clearly
235
established law is Yarborough v. Alvarado.
On June 1, 2004, the
Supreme Court issued a five-four decision denying habeas relief to a
seventeen-year-old who was questioned at a police station by police
236
concerning an attempted robbery and murder. During the questioning,
the young man eventually confessed to being present during the murder
237
and helping to hide the murder weapon.
On direct appeal, the
petitioner argued that his confession should have been suppressed
because he was not given a Miranda warning until after the questioning
238
concluded. The state court determined that the petitioner was not “in
custody” at the time of the questioning, and, thus, no Miranda warning
239
was required.
The petitioner then brought a federal habeas claim, which was denied
240
by the federal district court. The Ninth Circuit reversed, holding that
the state court failed to account for petitioner’s “youth and
inexperience” in evaluating whether he was in custody, and that, in light
241
of Supreme Court precedent, these factors were required. Although no
Supreme Court case was directly on point, the Ninth Circuit found that
AEDPA did not bar relief “because the relevance of juvenile status in
Supreme Court case law as a whole compelled the ‘extension of the
principle that juvenile status is relevant’ to the context of Miranda
242
custody determinations.”
The Ninth Circuit therefore concluded that
the state court had unreasonably applied clearly established law because
it had failed to “‘extend a clearly established legal principle . . . to a new
243
context.’”
244
Justice Kennedy,
The Court reversed the Ninth Circuit’s decision.
writing for the majority, began the analysis by determining the relevant
clearly established law. First, he noted that the Court’s “more recent
cases instruct that custody must be determined based on . . . how a
235. 124 S. Ct. 2140 (2004).
236. Id. at 2144-45.
237. Id. at 2146.
238. Id.
239. Id. In doing so, the state court relied on the custody test articulated in Thompson
v. Keohane, 516 U.S. 99, 112 (1995). Under that test, a court must “consider the
circumstances surrounding the interrogation and then determine whether a reasonable
person would have felt at liberty to leave.” Id.
240. Alvarado, 124 S. Ct. at 2146.
241. Id. at 2147.
242. Id.
243. Id. at 2150 (quoting Alvarado v. Hickman, 316 F.3d 841, 853 (9th Cir. 2002), rev’d,
124 S. Ct. 2140 (2004)).
244. Id. at 2144, 2152.

784

Catholic University Law Review

[Vol. 54:747

reasonable person in the suspect’s situation would perceive his
245
circumstances.”
After describing the relevant precedent as requiring
an objective inquiry using a “reasonable person” standard, the Court
addressed whether the state court’s decision was an unreasonable
246
application of that precedent.
The Court rejected petitioner’s claim that the state court’s decision was
objectively unreasonable for refusing to extend a general principle about
juvenile status from the Court’s precedent in other contexts to include
demanding consideration of the petitioner’s age and inexperience in
247
Miranda warnings. After comparing the facts suggesting the petitioner
was in custody with those against such a finding, the Court concluded
that the state court’s application of the Court’s clearly established law
was reasonable because “fair-minded jurists could disagree over whether
248
[the petitioner] was in custody.”
Rebuking the Ninth Circuit as “nowhere close to the mark when it
249
concluded otherwise,” Justice Kennedy explained that “the [Miranda]
custody test is general, and the state court’s application of [the test] fits
250
within the matrix of our prior decisions.”
Justice Kennedy expressed
concern about the Ninth Circuit’s conclusion that “the state court’s
failure to ‘extend a clearly established legal principle [of the relevance of
251
juvenile status] to a new context’ [was] objectively unreasonable.”
Recognizing the strength of the state’s argument that “if a habeas court
must extend a rationale before it can apply to the facts at hand then the
rationale cannot be clearly established at the time of the state-court
decision,” Justice Kennedy reasoned that “[s]ection 2254(d)(1) would be
undermined if habeas courts introduced rules not clearly established
252
under the guise of extensions to existing law.”

245. Id. at 2148. In Alvarado, Justice Kennedy described the Court’s prior decision in
Berkemer v. McCarty, 468 U.S. 420, 442 (1984), as follows: “[A] traffic stop [was]
noncustodial despite the officer’s intent to arrest because [the officer] had not
communicated that intent to the [suspect;] . . . ‘[T]he . . . [proper] inquiry is how a
reasonable man in the suspect’s position would have understood his [position],’” 124 S. Ct.
at 2148 (citation omitted); see also Thompson v. Keohane, 516 U.S. 99, 112 (1995)
(describing Miranda test as requiring two discrete inquiries); Stansbury v. California, 511
U.S. 318, 323 (1994) (per curiam) (explaining that custody determination depends on
objective circumstances, not subjective views).
246. Alvarado, 124 S. Ct. at 2149.
247. Id. at 2149-50.
248. Id.
249. Id. at 2150.
250. Id.
251. Id. at 2150-51.
252. Id. In further support, Justice Kennedy cited Hawkins v. Alabama, 318 F.3d 1302,
1306 n.3 (11th Cir. 2003), and a “cf.” citation to Teague v. Lane, 489 U.S. 288 (1989).

2005]

Seeking Clarity in the Federal Habeas Fog

785

Justice Kennedy then made a significant point about the nature of
legal rules:
[T]he range of reasonable judgment can depend in part on the
nature of the relevant rule. If a legal rule is specific, the range
may be narrow. . . . Other rules are more general, and their
meaning must emerge in application over the course of time. . . .
The more general the rule, the more leeway the courts have in
253
reaching outcomes in case by case determinations.
He noted that “the difference between applying a rule and extending it is
254
not always clear.” For example, “[c]ertain principles are fundamental
enough that when new factual permutations arise, the necessity to apply
255
the earlier rule will be beyond doubt.”
He concluded, however, that
256
petitioner’s was not such a case.
While the custody inquiry is an
“objective rule designed to give clear guidance to the police,”
consideration of a suspect’s age and inexperience with the police could
257
create a subjective inquiry.
Justice O’Connor concurred with the majority decision, but focused on
258
the petitioner’s age. Because the petitioner in this case was “almost 18
259
years old[,]” his age was not relevant to the Miranda custody inquiry.
She acknowledged, however, “There may be cases in which a suspect’s
260
age will be relevant to the . . . inquiry.”
Justice Breyer, joined by Justices Souter, Stevens, and Ginsburg,
261
dissented. To these justices, “the law in this case [was] clear”; it “asks
judges to apply, not arcane or complex legal directives, but ordinary
262
common sense.” Justice Breyer focused on the application of the law
to the facts. He began by presenting the question “in terms of federal
law’s well-established standards: Would a ‘reasonable person’ in
Alvarado’s ‘position’ have felt he was ‘at liberty to terminate the

Alvarado, 124 S. Ct. at 2150-51. See supra Part II.A, for a discussion of Teague, and infra
text accompanying notes 362-92, for a discussion of Hawkins.
253. Alvarado, 124 S. Ct. at 2149. In support of this statement, Justice Kennedy cited
his concurring opinion in the pre-AEDPA case of Wright v. West, 505 U.S. 277, 308-09
(1992), which was also cited by Justice Stevens in Williams, see Williams v. Taylor, 529
U.S. 362, 379 (2000) (opinion of Stevens, J.).
254. Alvarado, 124 S. Ct. at 2151.
255. Id.
256. Id.
257. Id. at 2151-52.
258. Id. at 2152 (O’Connor, J., concurring).
259. Id. (O’Connor, J., concurring).
260. Id. (O’Connor, J., concurring).
261. Id. at 2156 (Breyer, J., dissenting).
262. Id. at 2153 (Breyer, J., dissenting).

786

Catholic University Law Review

[Vol. 54:747

263

interrogation and leave’?”
Noting that “a court must answer [the]
question in light of ‘all the circumstances surrounding the
264
interrogation,’” he concluded that answer was “no.”
Justice Breyer
thus found that the petitioner “was [clearly] . . . ‘in custody’ when the
265
police questioned him.” Accordingly, the state court’s conclusion that
Alvarado was not in custody when the police questioned him without
266
Miranda warnings was objectively unreasonable.
Justice Breyer reasoned that the suspect’s youth was an objective
267
circumstance known to the police in this case. He did not take on the
Ninth Circuit’s extension argument from the juvenile status cases. He
did, however, make an analogy to the “reasonable person” in tort law as
268
taking into account personal characteristics.
He also noted that the
majority did not cite any case suggesting a limitation that courts should
269
ignore objective characteristics like age.
The Court’s decision in Alvarado is significant because it illustrates the
relationship between clearly established law and unreasonable
application under § 2254(d)(1). In denying relief under the unreasonable
application prong, the Court recognized a connection between the scope
of the precedent and whether a state court’s application of that precedent
is reasonable.
Two legal principles were at play in Alvarado: the custody standard
(how a reasonable person in the suspect’s situation would perceive his
circumstances) and the “reasonable person” standard. The Court
identified the custody standard without difficulty. The Court also
seemed to agree that the “reasonable person” standard embedded in the
custody standard is objective. The justices disagreed, however, about
whether the suspect’s age and inexperience were factors that should have
been considered under this objective standard. The majority rejected
270
these as subjective factors.
Analogizing to the reasonable person
standard in tort law, which takes a child’s age, intelligence, and
experience into consideration, Justice Breyer insisted that Alvarado’s age
271
and inexperience should have been considered.
Justice O’Connor’s
263. Id. at 2152 (Breyer, J., dissenting) (quoting Thompson v. Keohane, 516 U.S. 99,
112 (1995); Stansbury v. California, 511 U.S. 318, 325 (1994) (per curiam)).
264. Id. (Breyer, J., dissenting) (quoting Stansbury, 511 U.S. at 322).
265. Id. (Breyer, J., dissenting).
266. Id. at 2153 (Breyer, J., dissenting).
267. Id. at 2155 (Breyer, J., dissenting).
268. See id. (Breyer, J., dissenting).
269. See id. at 2154-55 (Breyer, J., dissenting).
270. See id. at 2151.
271. See id. at 2155-56 (Breyer, J., dissenting) (noting that “all American jurisdictions
count a person’s childhood as a ‘relevant circumstance’ in negligence determinations”
(citing and quoting RESTATEMENT (THIRD) OF TORTS § 10 cmt. b, at 128-130 (Tentative

2005]

Seeking Clarity in the Federal Habeas Fog

787

concurring opinion is less straightforward. She seemed to agree that age
may be part of the reasonable person inquiry in certain cases, but not in
Alvarado’s case. By doing so, she left for another case the clarification of
whether age is an appropriate consideration under the reasonable person
standard.
The Court, however, framed the issue not as a disagreement about the
contours of the reasonable person standard but as a disagreement about
the application of that standard to Alvarado’s facts. Because the
reasonable person standard is embedded in the Miranda custody
standard, application of the custody standard permits a wide range of
discretion by the state court. Where the facts of the habeas petitioner’s
case are not materially indistinguishable from a Supreme Court
precedent, § 2254(d)(1) requires deference to the state court’s reasonable
application of the standard.
The reasonableness of the state court’s application of the standard
depends on the breadth of that standard. The breadth of any legal
principle can be viewed on a spectrum from broad to narrow: from very
general principle to very fact-specific rule. The more general the
principle, the more the state court needs to invoke its discretion to apply
it; on the other hand, bright-line rules require less discretion to apply.
The Court’s decisions in Andrade and Alvarado demonstrate that it will
not find a lack of clearly established law where it can discern a legal
principle in the Court’s precedent, even if that principle is general or
otherwise indeterminate. Instead of analyzing the scope of the precedent
at a threshold level, the Court analyzed questions of scope under the
unreasonable application prong. Because a general principle requires
more discretion for a court to apply it to a specific factual situation, the
Court deferred to the state court’s reasonable application of that
principle.
In Andrade and Alvarado, the Court seemed to employ a sliding scale
of deference to the state court’s application of the legal principles
depending on how broad or narrow the principles were. In Andrade, the
disproportionality principle was viewed as very general, and the Court
found that the state court’s application was not objectively
272
unreasonable.
Likewise, in Alvarado, it was not clear from the
precedent that a suspect’s age and inexperience should be considered
under the reasonable person standard, and the Court found that the state
court’s decision to ignore the suspect’s age and inexperience was not
Draft No. 1, 2001))); see also RESTATEMENT (SECOND) OF TORTS § 283A (1965) (“If the
actor is a child, the standard of conduct to which he must conform to avoid being negligent
is that of a reasonable person of like age, intelligence, and experience under like
circumstances.”).
272. See Lockyer v. Andrade, 538 U.S. 63, 76-77 (2003).

788

Catholic University Law Review

[Vol. 54:747

objectively unreasonable. A reasonable application, therefore, includes a
spectrum of outcomes that are linked to the scope of the precedent. In
the deference environment of AEDPA, when a large degree of discretion
is required for the state court to apply the relevant precedent, the state
court’s application will likely be reasonable.
AEDPA therefore limits the scope of review by federal courts in a
number of ways. While the statute does not define the key terms in §
2254(d)(1), the Supreme Court’s decisions in Williams, Andrade, and
Alvarado provide a context for understanding the meaning of clearly
established law. First, they define clearly established law as the Court’s
holdings and (revealing a broad reading of the requirement) reasonable
applications of the Court’s precedent. Second, they instruct that the
determination of clearly established law should be a threshold issue.
Where the Court’s precedents were not “clear,” it did not deny relief on
that basis alone. Instead, it analyzed the clarity of the law under the
unreasonable application prong. Third, these decisions confirm that
broad principles can qualify as clearly established law, but they also
reveal that the broader the principle, the more likely the state court’s
decision will be reasonable. Although one commentator has argued that
the clarity of the law should not be part of the analysis under
273
§ 2254(d)(1), these decisions contrast that view. This Article argues
that clarity does matter and that it is an appropriate part of the analysis.
The next Part will demonstrate that lower courts are considering clarity,
but they are not doing so in a uniform fashion.
III. LIFE IN THE LOWER FEDERAL COURTS
The dearth of Supreme Court guidance about what constitutes clearly
274
established law has left the lower federal courts to fill in the details.
Cases in which the lower courts have struggled with clearly established
law questions fall into three categories. The first occurs when courts find
that there is no clearly established law applicable to the habeas
petitioner’s case. Some courts treat this determination as a dispositive
issue and deny relief. Other courts go on to analyze clearly established
law issues under the contrary to or unreasonable application prongs of §
2254(d)(1). The second category relates to the scope of precedent. In
273. See Ides, supra note 11, at 761-62.
274. To the extent that state courts must ascertain U.S. Supreme Court precedent for a
particular issue, these situations will likely require a similar analysis. Although this Article
does not address how state courts should analyze Supreme Court precedent on direct
review, many of the same principles apply. For a discussion of how state courts should
determine federal law, see Donald H. Zeigler, Gazing into the Crystal Ball: Reflections on
the Standards State Judges Should Use To Ascertain Federal Law, 40 WM. & MARY L.
REV. 1143 (1999).

2005]

Seeking Clarity in the Federal Habeas Fog

789

these cases, the courts decide whether a rule applies to the habeas
petitioner’s factual situation. The third category concerns inquiries into
the clarity of the precedent. These cases involve the application of
general principles, old or otherwise questionable precedent, and
splintered Supreme Court decisions. This Part examines each of these
categories in turn.
A. Lack of Precedent
The first type of issue that habeas courts encounter is the lack of any
Supreme Court precedent for the petitioner’s claim. Given the small
number of Supreme Court cases decided each year, this is hardly a
surprise. Although the Supreme Court has not yet had to deny habeas
relief under AEDPA because there was no clearly established law at the
time of the state court decision, lower federal courts have done so.
In this sense, the Court’s instruction that the determination of clearly
275
established law is a “threshold question” in § 2254(d)(1) cases is
ambiguous. By “threshold,” did the Court simply mean that habeas
courts should ask this question at the beginning of the analysis, or did it
intend for this question to be dispositive? If it is a dispositive question
and the court finds that there is no clearly established law, then habeas
relief must be denied.
One seemingly easy case would be when the habeas court faces an
issue of first impression. If the Supreme Court has not addressed the
issue by the time the state court decision is rendered, then the habeas
court should deny relief outright. The rationale supporting this outcome
is logical as well as practical: if the state court has no Supreme Court
precedent to guide it, the state court’s decision cannot be contrary to or
an unreasonable application of law that does not exist. It would
practically be very difficult for the habeas court to meet AEDPA’s
276
standard under these circumstances.
This approach also makes sense
from a policy perspective, and it is consistent with legislative intent, the
Court’s approach in Teague, and qualified immunity jurisprudence. The
state court should not be in a position to speculate as to how the
Supreme Court—or any federal court—would decide the issue when and
if it ever faced it. State court decisions on issues of first impression
should be decided on direct review, not habeas review.
Unfortunately, the clearly established law inquiry is seldom so simple.
Unless the Supreme Court has expressly declined to address a particular
275. Andrade, 538 U.S. at 71; Williams v. Taylor, 529 U.S. 362, 390 (2000) (Stevens, J.,
writing for the majority).
276. See Ides, supra note 11, at 761 (arguing that it would be “nigh impossible for a
state court decision to be contrary to or an unreasonable application of law [if the law is]
stated at its most abstract level”).

790

Catholic University Law Review

[Vol. 54:747

issue, the questions quickly become more tangled. For instance, must a
court of appeals ignore its own circuit’s precedent? Must the district
court ignore what would otherwise be binding circuit precedent?
AEDPA’s apparent disregard for traditional rules of vertical stare decisis
in the federal courts can lead to dramatic results. In the absence of at
least one Supreme Court case on a given issue, AEDPA effectively ties
the hands of the habeas court: lower federal courts are prohibited from
granting relief even if that court has directly addressed the issue in its
own precedent. This curtails the development of constitutional law in
the lower courts on habeas review, a congressional goal in enacting
AEDPA. While AEDPA’s source limitation for clearly established law
277
has been the subject of debate, the lower federal courts are dealing
with this limitation fairly consistently by requiring the petitioner to cite at
278
least one Supreme Court case in support of the petitioner’s position.
More complicated questions arise when courts need to discern the
scope of the Supreme Court precedent. This is a two-fold question: (1)
how broad or narrow is the rule articulated in the precedent, and (2) how
similar are the facts of the precedent case to the habeas petitioner’s case.
These same questions are at the heart of our common-law system’s
reliance on precedent and practice of reasoning by analogy. In the
habeas context, however, some courts deny relief based on their answers
to these questions.
1. Dispositive Issue
Most lower federal courts begin their analysis under § 2254(d)(1) by
determining as a threshold issue whether there is Supreme Court
precedent “on point” before proceeding with § 2254(d)(1)’s two-pronged
inquiry about whether the state court’s decision is contrary to or an
279
unreasonable application of clearly established law.
From a practical
standpoint, if there is no law to apply, the habeas court seemingly could
277. See supra note 40.
278. See, e.g., Gonzalez v. Fischer, No. 01-CV-8523 (JBW), 03-MISC-0066 (JBW),
2003 U.S. Dist. LEXIS 23874, at *38 (E.D.N.Y. Oct. 16, 2003) (stating that “[a habeas]
petitioner must ‘identify a clearly established Supreme Court precedent that bears on his
claim’” (quoting Loliscio v. Goord, 263 F.3d 178, 191 (2d Cir. 2001); Sellan v. Kuhlman,
261 F.3d 303 (2d Cir. 2001))). The district court need not proceed with the analysis if the
petitioner fails to cite a Supreme Court case. Id.; see Estrada v. Jones, No. 03 C 3092, 2004
U.S. Dist. LEXIS 6502, at *4-5 (N.D. Ill. Apr. 14, 2004).
[A] “petitioner first must show that the Supreme Court has ‘clearly established’
the propositions essential to [his] position.” . . . “[He] must have a Supreme
Court case to support his claim, and that Supreme Court decision must have
clearly established the relevant principle as of the time of his direct appeal.”
Id. (second alteration in original) (quoting Schaff v. Snyder, 190 F.3d 513, 522 (7th Cir.
1999)).
279. See, e.g., Chin v. Runnels, 343 F. Supp. 2d 891, 902 (N.D. Cal. 2004).

2005]

Seeking Clarity in the Federal Habeas Fog

791

not analyze whether the state court decision was contrary to or an
280
unreasonable application of clearly established law.
The most straightforward case is where the Supreme Court has
expressly declined to decide an issue. If the issue was still unresolved
when the state court decided the habeas petitioner’s case, then there is
no clearly established law. For example, the Supreme Court expressly
left open the question of whether, in a noncapital case, the Due Process
Clause requires that a lesser-included offense be charged when
281
supported by the evidence.
Lower courts facing this issue on habeas
review have denied relief because the law governing that issue is not
282
clearly established.
The slightly harder cases are those where the line between holding and
dicta is not as clear as Justice O’Connor suggested in her statement in
283
284
Williams.
For example, in Dallio v. Spitzer, a habeas petitioner
280. The lack of law for the state court to apply is somewhat analogous to the
“committed to agency discretion by law” doctrine in administrative law. The Federal
Administrative Procedure Act excludes agency action from judicial review “to the extent
that—the agency action is committed to agency discretion by law.” 5 U.S.C. § 701(a)(2)
(2000). The Supreme Court has interpreted this exception narrowly to apply “in those
rare instances where ‘statutes are drawn in such broad terms that in a given case there is
no law to apply.’” See Citizens To Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 410
(1971) (quoting S. REP. NO. 79-752, at 26 (1945)). In other words, judicial review would
be precluded in those situations where “the statute is drawn so that a court would have no
meaningful standard against which to judge the agency’s discretion.” Heckler v. Chaney,
470 U.S. 821, 830 (1985). In the habeas context, a further question arises whether a
principle might be too broad to constitute “clearly established” law.
281. Beck v. Alabama, 447 U.S. 625, 638 & n.14 (1980).
282. See, e.g., House v. Miller, No. 02-CV-5379, 03-MISC-0066, 2003 U.S. Dist. LEXIS
24380, at *46-47 (E.D.N.Y. Nov. 13, 2003) (“[T]here is no Supreme Court precedent that
addresses the issue of whether a lesser-included offense must be charged when supported
by the evidence.”); Davis v. Herbert, No. 02-CV-04908, 03-MISC 0066 (JBW), 2003 U.S.
Dist. LEXIS 24121 (E.D.N.Y. Oct. 24, 2003). But see Peakes v. Spitzer, No. 04 Civ. 1342
(RMB) (AJP), 2004 U.S. Dist. LEXIS 10905, at *34-44 (S.D.N.Y. June 16, 2004) (finding
lack of Supreme Court precedent but also looking to federal circuit court cases before
concluding that there is no clearly established law). For an example in another context,
see Gonzalez, 2003 U.S. Dist. LEXIS 23874, at *39 (holding that while the Supreme Court
has held that indigent defendants are entitled to a free set of trial transcripts to prepare for
direct appeal, there is no clearly established law entitling petitioner to a free set of trial
transcripts to prepare for collateral attacks because the Court expressly left that question
open). See supra note 278.
283. See Williams v. Taylor, 529 U.S. 362, 412 (2000) (O’Connor, J., writing for the
majority) (stating that the Court’s holdings, not dicta, constitute clearly established law).
First-year law students should be comforted that this distinction occasionally eludes even
some judges. See, e.g., Johnson v. Carroll, 369 F.3d 253, 263 (3d Cir. 2004) (noting that
“the Supreme Court[] . . . has not held, not even in dicta, let alone ‘clearly established,’
that the mere appearance of bias on the part of a state trial judge, without more, violates
the Due Process Clause”).
284. 343 F.3d 553 (2d Cir. 2003).

792

Catholic University Law Review

[Vol. 54:747

claimed that his Sixth Amendment right to counsel and corollary right to
self-representation were violated when the trial court failed to warn him
285
of the dangers of self-representation. He relied on one Supreme Court
286
case, Faretta v. California, where the defendant’s request to proceed
287
pro se was denied.
The Second Circuit distinguished Faretta and
determined that the only relevant portion of the case—that a defendant
waiving his right to counsel “‘should be made aware of the dangers and
disadvantages of self-representation’”—was dictum and not part of the
288
holding. The court further noted that the word “should” in this context
289
reflects a preference rather than a mandate.
Finally, the court noted
that “the Supreme Court has repeatedly emphasized that a knowing and
intelligent waiver depends on the totality of ‘facts and circumstances’ in a
290
given case.”
Citing a lack of clearly established Supreme Court
291
precedent, the Second Circuit denied habeas relief.
In a concurring opinion, Judge Katzmann disagreed with the majority’s
characterization of the holding in Faretta.
In his view, Faretta
292
“establishe[d] clear federal law [that] requir[ed]” at least “some basis in
the record . . . to conclude that the defendant was aware of the adverse
293
consequences of proceeding pro se.”
He therefore would have found
294
that the state court unreasonably applied the rule from Faretta.
He
agreed with the district court, however, that the error was harmless, and
295
thus would still deny relief. While the panel ultimately agreed on the
same result (no relief), Judge Katzmann’s concurrence demonstrates a
fundamental disagreement about the scope of the holding of the relevant
Supreme Court precedent.
Disagreement over the scope of the precedent has significant
consequences when courts treat the clearly established law inquiry as a

285. Id. at 555.
286. 422 U.S. 806 (1975).
287. Id. at 807. On direct appeal, the petitioner also cited a number of New York
Court of Appeals decisions. Id.
288. Dallio, 343 F.3d at 561 (quoting Faretta, 422 U.S. at 835).
289. Id. at 562.
290. Id. at 563 (quoting Edwards v. Arizona, 451 U.S. 477, 482 (1981)).
291. Id. at 564-65.
292. Id. at 565 (Katzmann, J., concurring).
293. Id. at 567 (Katzmann, J., concurring).
294. Id. at 568 (Katzmann, J., concurring).
295. Id. at 565 (Katzmann, J., concurring). The district court found that there was a
constitutional violation and that the state court decision was an unreasonable application
of clearly established federal law, but concluded that the error was harmless and thus
denied relief. Dallio v. Spitzer, 170 F. Supp. 2d 327, 336-37, 338 (E.D.N.Y. 2001), aff’d,
343 F.3d 553 (2d Cir. 2003).

2005]

Seeking Clarity in the Federal Habeas Fog

793

296

dispositive issue. For example, in Estrada v. Jones, the court read the
Supreme Court precedent narrowly.
The petitioner in Estrada
“claim[ed] . . . that the trial court erred in denying his motion for funds to
obtain expert psychiatric assistance” to determine his fitness at the time
297
of his plea.
The court declared that the petitioner failed to cite any
298
authority in support of his claim and denied relief. The petitioner did
299
cite one Supreme Court case, Ake v. Oklahoma; however, the court
distinguished Ake because it involved “the appointment of psychiatric
experts in the context of an insanity defense . . . [relating to] defendant’s
300
sanity at the time of the offense.” Because the court treated the lack of
clearly established law as a dispositive issue, it did not continue with the
analysis under § 2254(d)(1) to determine whether it was objectively
unreasonable for the state court not to extend Ake’s rule to the facts of
petitioner’s case.
By contrast, a habeas court can view a general legal principle as being
too broad to apply to petitioner’s specific factual situation. In Johnson v.
301
Carroll, the district court and the appellate court disagreed about the
scope of Supreme Court precedent. The case raised the issue of whether
there was clearly established law that the appearance of bias on the part
302
of a state judge violates the Due Process Clause. After concluding that
the judge’s actions did have an appearance of bias, the district court
303
conditionally granted the petition.
Citing the AEDPA’s “stringent
304
provisions” the Third Circuit reversed.
In Johnson, the district court and the habeas petitioner relied on three
Supreme Court cases as establishing a general principle that the
appearance of bias on the part of a state judge violates the Due Process
305
Clause. The Third Circuit did not view these three cases, or any other

296. No. 03 C 3092, 2004 U.S. Dist. LEXIS 6502 (N.D. Ill. Apr. 14, 2004).
297. Id. at *9.
298. Id.
299. 470 U.S. 68 (1985).
300. Estrada, 2004 U.S. Dist. LEXIS 6502, at *9-10.
301. 369 F.3d 253 (3d Cir. 2004).
302. Id. at 259.
303. Id. at 255. The defendant did not file any postconviction motions, instead filing a
petition for a writ of habeas corpus in federal court. Id.
304. Id.
305. Id. at 258. The three Supreme Court cases relied on were In re Murchison, 349
U.S. 133 (1955); Liljeberg v. Health Services Acquisition Corp., 486 U.S. 847 (1988); and
Liteky v. United States, 510 U.S. 540 (1994). Johnson, 369 F.3d at 258. The district court
also relied on a Delaware state court decision, but because AEDPA limits the relevant law
to Supreme Court decisions, the Third Circuit disregarded the state court case in its
review. See id. n.2. The Third Circuit also noted that the district court had not used the
phrase “clearly established” in its opinion. Id. at 259.

794

Catholic University Law Review

[Vol. 54:747

306

Supreme Court case, as standing for the legal principle at issue.
The
Third Circuit noted that the district court seemed to acknowledge that it
made an overly broad reading. In a “decision to grant the state’s motion
for an enlargement of the stay . . . pending . . . resolution of [the] appeal,”
the district court stated that “its earlier decision . . . ‘was based on an
analogy to Supreme Court cases related to the issue of recusal under 28
U.S.C. § 455 [for federal judges] and not on direct precedent related to the
307
trial judge’s appearance of bias under the Due Process Clause.’”
As
further support for its conclusion, the Third Circuit cited three other
circuits’ pre-AEDPA decisions rejecting arguments similar to those
308
made by the petitioner.
The Third Circuit concluded that “the
Supreme Court[] . . . has not held, not even in dicta, let alone ‘clearly
established,’ that the mere appearance of bias on the part of a state trial
judge, without more, violates the Due Process Clause,” and it reversed
309
and remanded the case with directions to dismiss the habeas petition.
Because this case involved a general principle, the Third Circuit should
have examined the scope of that principle under the unreasonable
application prong of § 2254(d)(1). Even if the court ultimately would
have denied relief under that prong, continuing with the analysis to
evaluate the reasonableness of the state court’s decision with existing
Supreme Court precedent would have been more in line with AEDPA’s
goals. By treating the determination of clearly established law as a
dispositive issue when there is a disagreement about the scope of the
precedent, courts are not fulfilling their duty under AEDPA to review
310
the state court’s decision.
306. Johnson, 369 F.3d at 262.
307. Id. at 259 (second alteration in original) (quoting Johnson v. Carroll, No. 02-562JJF, 2003 WL 22136302, at *1 (D. Del. Sept. 10, 2003)).
308. Id. at 262. The court cited Hardy v. United States, 878 F.2d 94, 97 (2d Cir. 1989)
(concluding that the Due Process Clause does not mandate § 455’s “‘appearance of
impropriety standard’” (citation omitted)); United States v. Couch, 896 F.2d 78, 81 (5th
Cir. 1990) (“‘[C]onduct violative of section 455 may not [necessarily] constitute a due
process deficiency.’” (alterations in original) (citation omitted)); and Del Vecchio v.
Illinois Department of Corrections, 31 F.3d 1363, 1371 (7th Cir. 1994) (en banc) (rejecting
the view that two of the cases cited by petitioner, Murchinson and Aetna, held that due
process “‘requires judges to recuse themselves based solely on appearances’” (citation
omitted)).
309. Johnson, 369 F.3d at 263.
310. A related but different problem is when a court completely fails to analyze
whether there is clearly established law. For example, the Ninth Circuit recently issued an
amended opinion in Shaw v. Terhune, 380 F.3d 473 (9th Cir. 2004). In the original
opinion, the panel disagreed about the scope of Supreme Court precedent. Shaw v.
Terhune, 353 F.3d 697, 706 (9th Cir. 2003) (Wallace, J., dissenting), withdrawn, 380 F.3d
473 (9th Cir. 2003). The Ninth Circuit reviewed two issues and found there was no clearly
established law for either issue. The first issue was whether prosecutors must give
consistent interpretations of the same evidence in different trials. Id. at 701. The court

2005]

Seeking Clarity in the Federal Habeas Fog

795

2. Non-dispositive Issue
While the dispositive approach makes sense in theory, the previous
Subpart has demonstrated that this can be complicated in practice. This
is because the analysis often depends on the court’s determination of
whether the Supreme Court precedent is “on point,” which is another
way of assessing the scope of the precedent. To determine how similar
the relevant Supreme Court precedent is to the case before the habeas
court, habeas courts usually need to reference the factual context; they
generally cannot ascertain clearly established law in the abstract.
Accordingly, many courts address clearly established law as a threshold
question (in the sense of asking at the beginning of the analysis what the
relevant law is) but consider issues about the scope of the precedent
under the contrary to and unreasonable application prongs of
§ 2254(d)(1).

found that while prosecutors are prohibited from knowingly presenting false evidence, no
clearly established federal law precluded a prosecutor from arguing two inconsistent
theories that were equally supported by ambiguous evidence. Id. at 704-05. The second
issue was whether inconsistent convictions violated due process. Id. at 701. The majority
noted that “[t]he [Supreme] Court has . . . expressly rejected the proposition that due
process always requires consistent convictions.” Id. at 705. Treating the “absence of
‘clearly established federal law’” as a dispositive issue, the majority denied relief. Id. at
706.
Judge Wallace dissented. While the majority viewed the Supreme Court as not having
directly addressed the issue, he insisted that to find clearly established law, the identical
factual circumstances at issue need not be previously addressed by the Supreme Court. Id.
at 707 (Wallace, J., dissenting). He relied instead on the broad principle derived from
several Supreme Court cases that “[t]he prosecutor’s duty to seek the truth and vindicate
the demands of justice distinguishes his role from that of ordinary trial counsel,” and that
“prosecutorial foul play” is not acceptable. Id. (Wallace, J., dissenting). Judge Wallace
also found that the prosecutor acted in bad faith, and that prosecutorial bad faith is
proscribed by Supreme Court precedent. Id. at 710 (Wallace, J., dissenting); see also
Smith v. Groose, 205 F.3d 1045, 1051 (8th Cir. 2001). His approach was similar to the
Ninth Circuit’s approach in Alvarado v. Hickman, 316 F.3d 841 (9th Cir. 2002), rev’d sub
nom. Yarborough v. Alvarado, 124 S. Ct. 2140 (2004), where the court extracted a general
principle about recognizing juvenile status as a special circumstance to be considered. See
supra Part II.C.2.b.
Almost eight months later, the Ninth Circuit withdrew the original opinion and issued
an amended opinion. The amended opinion is odd. In the amended opinion, a unanimous
panel simply notes § 2254(d)(1)’s standard and skips over the clearly established law
question. Shaw, 380 F.3d at 477. After acknowledging the canon of avoiding
constitutional questions if an issue may be resolved on narrower grounds, the court stated
that it was “not reach[ing] the . . . merits of [petitioner’s] due process claim [because it]
conclude[ed] that, even if [petitioner’s] due process rights were infringed by the
prosecutor[] . . . , such error was harmless.” Id. at 748. It is not appropriate for a court to
skip over the threshold question of clearly established law. Theoretically, the canon of
avoidance of constitutional questions could be invoked in every habeas case. Such a result
would be antithetical to habeas review under § 2254.

796

Catholic University Law Review

[Vol. 54:747

311

For instance, in Lopez v. Wilson, the Sixth Circuit affirmed the
district court’s denial of habeas relief to a petitioner claiming that his
right to effective assistance of counsel was violated because he was
312
denied appointment of counsel to file a motion to reopen the appeal.
Rather than treating the existence of clearly established law as a
threshold inquiry, the court asked that question in conjunction with
determining whether the state court decision was contrary to Supreme
313
Court precedent.
The Sixth Circuit held that the state court decision
was not contrary to Supreme Court precedent because the two Supreme
Court cases relied on by the petitioner were materially distinguishable
314
from his case.
315
The first case the petitioner relied on was Douglas v. California. In
Douglas, the Supreme Court held that a defendant is entitled to counsel
for purposes of an appeal, but specifically stated that the Court was
316
“‘dealing only with the first appeal.’” In the second case the petitioner
cited, the Supreme Court held that due process guarantees a defendant
effective assistance of counsel where his appointed counsel failed to
317
properly file an appeal. The Sixth Circuit found that neither of these
cases constituted clearly established Supreme Court precedent on the
issue of denying the appointment of counsel to file a motion to reopen an
318
appeal.
Although the Sixth Circuit itself had directly addressed the issue itself
several years ago and had determined that a defendant did have a
319
constitutional right to counsel in such a situation, the court announced
that because its decision predated Williams, the decision was not relevant
320
for purposes of AEDPA.
Accordingly, it held that the state court’s
321
decision was not contrary to clearly established law.
Judge Cole
concurred in the judgment, but disagreed with the majority’s analytical

311. 355 F.3d 931 (6th Cir.), vacated, reh’g en banc granted, 366 F.3d 430 (6th Cir.
2004). At the time this Article went to press, no en banc decision had been issued.
312. Id. at 933-34.
313. Id. at 937.
314. Id. at 939-41.
315. Id. at 940; Douglas v. California, 372 U.S. 353 (1963).
316. Lopez, 355 F.3d at 940-41 (quoting Douglas, 372 U.S. at 356).
317. Id. at 941; see Evitts v. Lucey, 469 U.S. 387, 393 (1985).
318. Lopez, 355 F.3d at 941.
319. White v. Schotten, 201 F.3d 743, 752-53 (6th Cir. 2000).
320. Lopez, 355 F.3d at 936. This statement highlights the importance of AEDPA’s
source limitation, which is more restrictive than under Teague or in the qualified immunity
context. See supra Part II.A.
321. Lopez, 355 F.3d at 941.

2005]

Seeking Clarity in the Federal Habeas Fog

797

approach. He would have treated the absence of Supreme Court
322
precedent as a dispositive issue and denied relief on that basis alone.
The majority’s view in Lopez reflects an approach the lower federal
courts often take. Where the scope of the precedent is not clear, these
courts tend to continue with the § 2254(d)(1) analysis rather than treat
323
the clearly established inquiry as dispositive. For example, in Gilchrist
324
v. O’Keefe, the Second Circuit reviewed under AEDPA whether the
state court had unconstitutionally denied the petitioner’s right to counsel
when it refused to appoint new counsel at sentencing after petitioner’s
previous appointed counsel withdrew because the petitioner punched
325
him, puncturing his eardrum.
The State argued that “petitioner
326
forfeited rather than waived his right to counsel.”
Although there was not a Supreme Court case directly on point, the
Second Circuit found clearly established law through the following
analysis. First, Supreme Court precedent recognizes a distinction
327
between waiver and forfeiture of constitutional rights in general.
Second, the Supreme Court has not held that an indigent defendant may
328
not forfeit the right to counsel by misconduct.
Third, the Supreme
Court has not held that a defendant may not forfeit a constitutional
329
right. The court then held that the state court ruling that petitioner had
forfeited his right to counsel was not contrary to clearly established
330
law.
The court next considered whether the state court’s ruling that the
petitioner forfeited his right to counsel by this single, violent incident was
an “‘unreasonable application’ of the more general principles in Supreme
Court cases such as Gideon [v. Wainwright] emphasizing the tremendous
331
importance of the right to counsel.”
The court emphasized that this

322. Id. at 942 (Cole, J., concurring in the judgment).
323. For instance, the court and parties could agree that a particular case stands for a
particular rule or standard, but disagree on the facts of the case at bar.
324. 260 F.3d 87 (2d Cir. 2001).
325. Id. at 90.
326. Id. at 94.
327. Id. at 95-97. The court examined the following cases: Johnson v. Zerbst, 304 U.S.
458, 464 (1958); Illinois v. Allen, 397 U.S. 337, 343 (1970) (holding that, after a judge has
warned a defendant, a defendant can lose the right to be present at his trial if he continues
disruptive behavior in the courtroom); and Taylor v. United States, 414 U.S. 17, 20 (1973)
(rejecting defendant’s constitutional claim where defendant had only attended the
beginning of trial and failed to appear for the remainder, even absent a specific warning by
the judge).
328. Gilchrist, 260 F.3d at 97.
329. Id.
330. Id.
331. Id.

798

Catholic University Law Review

[Vol. 54:747

step was “necessary because the lack of Supreme Court precedent
specifically addressing forfeiture of the right to counsel does not mean
that any determination that such a fundamental right has been forfeited,
even if based on an utterly trivial ground, would survive habeas
332
review.”
Because the Supreme Court had recognized that other
important constitutional rights may be forfeited by serious misconduct,
the Second Circuit held that the state court did not unreasonably apply
333
clearly established law.
As support, the Second Circuit cited other
334
circuits that had reached similar results.
In the absence of clearly
established law on the precise issue, the habeas court was unwilling to
find fault in either the state court’s analysis of or its application of the
335
more general existing law.
The divergence in the lower courts’ approaches regarding whether
clearly established law is a dispositive issue is not easily explained.
Currently some courts are denying habeas relief without analyzing the
questions of scope under the contrary to or unreasonable application
prongs of § 2254(d)(1). Other courts have not been comfortable with this
approach. In close cases, courts should be reluctant to draw the line at
denying relief for lack of clearly established law, because liberty interests
are at stake.
B. Failure to Extend a Supreme Court Precedent to New Facts
The second recurring issue habeas courts face arises when they must
determine whether a rule from Supreme Court precedent should apply to
the facts of the petitioner’s case. Cases in this category essentially
require the court to determine the scope of the precedent, which in turn
asks two questions: how broad or narrow is the precedent and how many
cases does the precedent control? This category requires courts to
reason by analogy and directly invokes § 2254(d)(1)’s contrary to and
unreasonable application prongs.
Courts must first ascertain the scope of the rule articulated in the
336
precedent. For example, in Rico v. Leftridge-Byrd the Third Circuit
affirmed the denial of habeas relief to a petitioner who argued that the
332. Id.
333. Id.
334. Id. at 97-99. The court cited United States v. McLeod, 53 F.3d 322 (11th Cir. 1995)
(affirming denial of defendant’s request for a new attorney to represent him at a motion
for a new trial after prior attorney withdrew on the basis that defendant had threatened to
harm the attorney and had verbally abused him), and United States v. Leggett, 162 F.3d 237
(3d Cir. 1998) (affirming district court’s ruling at sentencing that defendant had forfeited
right to counsel when he physically attacked his second court-appointed attorney in the
courtroom on the day he was to be sentenced).
335. Gilchrist, 260 F.3d at 100.
336. 340 F.3d 178 (3d Cir. 2003).

2005]

Seeking Clarity in the Federal Habeas Fog

799

337

rule from Batson v. Kentucky should apply to exercising peremptory
338
challenges to eliminate Italian-American jurors.
The state supreme
court found the Batson rule applied to Italian-Americans but rejected the
petitioner’s Batson claim because the trial court found that the
prosecutor’s strikes were ethnically-neutral and not purposeful
339
discrimination.
The Third Circuit began its analysis by examining the scope of the
Batson rule. In Batson, the Supreme Court held that to establish a prima
facie case of discrimination for a prosecutor’s exercise of peremptory
challenges, a black “defendant must ‘show that he is a member of a
cognizable racial group and that the prosecutor has exercised peremptory
challenges to remove from the venire members of the defendant’s
340
race.’”
In subsequent cases, the Supreme Court extended the rule to
341
cover discrimination against bilingual Latino jurors, with the Court
342
referring to “Latino” as both a race and an ethnicity, and to cover
343
discrimination on the basis of gender. The Third Circuit stressed that
the Supreme Court had not at the time of the state court decision, or
subsequently, extended the Batson rule to any European-American
344
ethnicity or national origin.
Accordingly, the Third Circuit quickly
concluded that the state court’s decision, which applied Batson to strikes
against Italian-American jurors, was not contrary to Supreme Court
345
precedent.
The next portion of the Third Circuit’s analysis—whether the state
court unreasonably applied Batson—was more detailed. The court
supported its conclusion that the state court did not unreasonably apply
346
Batson by emphasizing the uncertain state of the law. The court first
noted that the Supreme Court’s Batson jurisprudence was not consistent

337. 476 U.S. 19 (1986).
338. Rico, 340 F.3d at 186-87. The court noted that the habeas petitioner, Joseph Rico,
was not Italian-American, but had changed his surname to “Rico” from “Gavel.” Id. at
180.
339. Id. at 181-82. “The [Pennsylvania Supreme] Court assumed without deciding that
Italian-Americans were a cognizable group subject to the Batson rule,” id. at 182, which
was applicable to an ethnic group that is “‘a cognizable group that has been or is currently
subjected to discriminatory treatment’ in the community,” id. at 181-82 (quoting
Commonwealth v. Rico, 711 A.2d 990, 994 (Pa. 1998)).
340. Id. at 182 (quoting Batson, 476 U.S. at 80, 96).
341. Id. (citing Hernandez v. New York, 500 U.S. 352 (1991)).
342. Id. n.3 (citing Hernandez, 500 U.S. at 357 n.2).
343. Id. at 182.
344. Id.
345. Id. at 182-83.
346. Id.

800

Catholic University Law Review

[Vol. 54:747

347

in distinguishing between race and ethnicity or ethnic origin.
For
348
example, in Hernandez v. New York, where the Court extended Batson
349
to Latino jurors, the Court used both “ethnicity” and “race,” but in her
350
concurrence, Justice O’Connor used only the term “race.”
This
351
“uncertainty,” the Third Circuit acknowledged, had not been resolved.
The Third Circuit further noted that the definition of “race” had changed
352
over time.
Finally, the court recognized that the federal courts of
appeals had been restrictive in extending Batson beyond the
classifications explicitly recognized by the Supreme Court—race, gender,
353
and ethnicity of Latinos. In light of the uncertain state of the law, the
court held that “[i]t was . . . not objectively unreasonable for the state
courts to consider challenges to Italian-American prospective jurors
under Batson and that, when they did so, they did not unreasonably
354
apply Supreme Court precedent.”
The Third Circuit’s analysis in Rico is instructive on several levels.
First, it did not treat clearly established law as a dispositive issue. Had
the court done so, petitioner’s writ could have been denied for lack of
Supreme Court precedent regarding Batson’s application to ItalianAmericans, an issue the Supreme Court has not directly faced. Second,
the court found that the state court’s extension of Batson to ItalianAmericans was neither contrary to Supreme Court precedent nor
objectively unreasonable. Third, the court’s emphasis on the uncertain
state of the law allowed the state court more leeway to be incorrect
before becoming objectively unreasonable under § 2254(d)(1). The court

347. Id. at 183.
348. 500 U.S. 352 (1991).
349. Rico, 340 F.3d at 183.
350. Id. at 183 n.5 (“‘[A] peremptory strike will constitute a Batson violation only if
the prosecutor struck a juror because of the juror’s race.’” (quoting Hernandez, 500 U.S. at
373 (O’Connor, J., concurring))).
351. Id. at 183. In support, the court cited United States v. Martinez-Salazar, 528 U.S.
304, 315 (2000), in which the Supreme Court stated that “under the Equal Protection
Clause, ‘a defendant may not exercise a peremptory challenge to remove a potential juror
solely on the basis of the juror’s gender, ethnic origin, or race.’” Rico, 340 F.3d at 183
(quoting Martinez-Salazar, 528 U.S. at 315).
352. Rico, 340 F.3d at 183.
353. Id. n.6. The court cited the following cases as extending Batson: Brewer v.
Marshall, 119 F.3d 993 (1st Cir. 1997); United States v. Taylor, 92 F.3d 1313 (2d Cir. 1996);
United States v. Krout, 66 F.3d 1420 (5th Cir. 1995); and United States v. Canoy, 38 F.3d
893, 897-98 (7th Cir. 1994). The court cited the following cases that refused to extend
Batson: Murchu v. United States, 926 F.2d 50 (1st Cir. 1991) (per curiam), and United States
v. Campione, 942 F.2d 429 (7th Cir. 1991). Rico, 340 F.3d at 184. The Third Circuit itself
had been exposed to, but had not decided, the question of whether Batson extended to
Italian-Americans. Id. at 183.
354. Rico, 340 F.3d at 184.

2005]

Seeking Clarity in the Federal Habeas Fog

801

established the uncertainty in the law through the Supreme Court’s own
language as well as the trepidation expressed by circuit courts.
355
In another example, Rockwell v. Yukins, the Sixth Circuit, sitting en
356
banc, clashed over the scope of a Supreme Court precedent.
The
petitioner, a mother convicted of conspiring with her sons to kill their
father, argued that the state court’s decision to exclude evidence of the
father’s alleged sexual abuse of the sons violated her Sixth Amendment
357
right to present a complete defense. A majority of the court rejected
the petitioner’s attempt to extract a general rule that a criminal
defendant must be permitted to present any evidence deemed critical to
the defense and held that the state court decision was not
358
unreasonable. Although the majority did not explicitly frame the issue
as a clearly established law inquiry, it found that the Supreme Court
cases that the petitioner relied on did not stand for as broad a principle
359
as the petitioner argued.
Four judges joined in a strong dissent. They argued that not only was
there clearly established Supreme Court precedent on the issue as
petitioner claimed, but that the state court’s decision excluding the
360
evidence of the father’s abuse was objectively unreasonable.
The
dissenters noted that the majority’s conclusion that one of the Supreme
Court cases relied on by petitioner was “readily distinguishable” from
the precedent flew “in the face of habeas review under the unreasonable
application prong of § 2254(d)(1) inasmuch as under this prong, relief
may be granted ‘based on an application of a governing legal principle to
a set of facts different from those of the case in which the principle was
361
announced.’”
The dissenters understood that the unreasonable application prong
stretches the concept of clearly established law to include principles that
355. 341 F.3d 507 (6th Cir. 2003) (en banc).
356. Id. at 509, 514; see also Swiger v. Brown, 86 Fed. Appx. 877, 880-81 (6th Cir.
2004).
357. Rockwell, 341 F.3d at 509-10.
358. Id. at 514.
359. Id. at 512-13.
360. Id. at 514, 523 (Clay, J., dissenting).
361. Id. at 522 n.1 (quoting Lockyer v. Andrade, 538 U.S. 63, 76 (2003)). For an
example of another case where the court analyzed whether the state court decision
unreasonably applied a general principle from Supreme Court precedent recognizing that
“the opportunity to present a defense is one of the constitutional requirements of a fair
trial,” see Jones v. Stinson, 229 F.3d 112, 119-20 (2d Cir. 2000) (finding that the Supreme
Court “[has] not decided the specific circumstances under which a criminal defendant
must be allowed to introduce evidence of prior non-criminal conduct to demonstrate that
he did not commit the crime at issue” and holding that the state court’s decision was not
contrary to or an unreasonable application of a more general principle from Supreme
Court precedent).

802

Catholic University Law Review

[Vol. 54:747

can be reasonably applied to different factual situations. The majority,
on the other hand, had a narrower view of clearly established law as
governing only those cases that are materially indistinguishable.
Although that standard is relevant under the contrary to prong of §
2254(d)(1), the majority’s view is too limiting to accommodate the
unreasonable application prong.
Sometimes courts can readily identify the relevant precedent and the
only question is whether the state court should have extended the rule to
362
a new factual situation. For example, in Hawkins v. Alabama, the
petitioner asked the Eleventh Circuit to find that the state court
unreasonably applied clearly established federal law by refusing to
363
extend a legal principle to his case.
The legal principle at issue,
364
articulated by the Supreme Court in Oregon v. Kennedy, covered
365
prosecutorial misconduct that was intended to and did cause a mistrial.
The petitioner argued for an extension of the Kennedy rule, and cited
366
four cases from other circuits in support of his argument.
In the cases relied on by the petitioner, the Second and Seventh
367
Circuits indicated that they might extend the Kennedy rule.
The
Eleventh Circuit found these cases unhelpful for two reasons. First, the
368
supporting cases were not habeas cases. This assertion is weak, as cases
analyzing the extension of Supreme Court precedent on direct appeal are
still relevant. Second, and more persuasively, the court asserted that
clearly established law inquiry must focus on Supreme Court decisions,
369
not those of lower federal courts, even of the same circuit. The court
observed that the decisions of other federal courts and its own decisions
are relevant to the AEDPA inquiry in a limited sense; they are helpful
“only to the extent that the decisions demonstrate that the Supreme
Court’s pre-existing, clearly established law compelled the circuit courts

362. 318 F.3d 1302 (11th Cir. 2003).
363. Id. at 1306. Hawkins was cited with approval by the Court in Yarborough v.
Alvarado, 124 S. Ct. 2140, 2150-51 (2004).
364. 456 U.S. 667 (1982).
365. Id. at 676 (“[O]nly where the governmental conduct in question is intended to
‘goad’ the defendant into moving for mistrial may a defendant raise the bar of double
jeopardy to a second trial after having succeeded in aborting the first on his own
motion.”).
366. Hawkins, 318 F.3d at 1308 (citing United States v. Catton, 130 F.3d 805 (7th Cir.
1997); United States v. Doyle, 121 F.3d 1078 (7th Cir. 1997); United States v. Pavloyianis,
996 F.2d 1467 (2d Cir. 1993); United States v. Wallach, 979 F.2d 912 (2d Cir. 1992)).
367. Id. at 1309 & n.5.
368. Id. at 1308-09.
369. Id. at 1309.

2005]

Seeking Clarity in the Federal Habeas Fog

803

(and by implication would compel a state court) to decide in a definite
370
way the case before them.”
The Eleventh Circuit rejected the Second and Seventh Circuits’ cases
because an extension of the Kennedy rule was not “truly compelled” by
371
the Supreme Court’s rule. Rather, “[t]he inclination to widen Kennedy
represents an independent judgment by these circuits on what the law
ought to be in the circumstances contemplated by those circuit courts in
372
those cases.”
The court acknowledged the difficulty of applying the unreasonable
application standard. Even in light of the Supreme Court’s instruction
that a state court unreasonably applies clearly established federal law if
“the state court was unreasonable in refusing to extend the governing
legal principle to a context in which the principle should have
373
controlled,”
the Eleventh Circuit contended that it lacked the
necessary guidance to resolve the issue because it was “uncertain about
374
what is meant by ‘extend’ and by ‘context.’”
It further expressed
frustration that these critical but undefined words “appear in a plurality
opinion which did not receive the votes of a majority of the Court, and
375
the other opinions did not elaborate on the concept of extension.”
The Eleventh Circuit then attempted to define “extend,” noting that
376
the word “can mean different things at different times.”
Specifically,
the court conjectured:
Extend might only mean to apply the ratio decidendi of
Supreme Court decisions fully and completely (and not in some
crabbed way) so that the rule of law covers new and different
facts and circumstances as long as the new facts and
circumstances—objectively reasonably viewed—are materially
or, put differently, substantially the same that were in the mind
377
of the Supreme Court when it laid down the rule.
370. Id.
371. Id. n.5. Note the court’s use of Teague’s language of results that are “compelled”
and “dictated”: “The Seventh and Second Circuit cases cited by Hawkins do not reach this
kind of conclusion about the Supreme Court’s preexisting precedent dictating the circuit’s
decision.” Id. at 1309. This Article argues that while AEDPA limits the source of clearly
established law to Supreme Court precedent, lower federal court decisions do have a role
to play. See infra Part IV.
372. Hawkins, 318 F.3d at 1309 n.5.
373. See Ramdass v. Angelone, 530 U.S. 156, 166 (2000) (plurality opinion).
374. Hawkins, 318 F.3d at 1306 n.2.
375. Id. See infra Parts III.C, IV, for a discussion of the problem of plurality decisions.
376. Hawkins, 318 F.3d at 1306 n.3.
377. Id. This author does not believe that the Supreme Court has a collective “mind”
from which we can glean a rationale, other than that which is articulated in the written
opinion itself.

804

Catholic University Law Review

[Vol. 54:747

Using this definition, the Eleventh Circuit announced that it agreed
378
AEDPA requires state courts to “extend” a Supreme Court rule.
The Eleventh Circuit also noted that “extend” has a different meaning:
“to ‘widen the range, scope, area of application of (a law, operation,
dominion, state of things, etc.); to enlarge the scope or meaning of (a
379
word).’” The court rejected this definition as contrary to congressional
intent and proclaimed that “[t]o widen the scope of or to enlarge
Supreme Court rules is not to follow ‘clearly established’ Supreme Court
380
law, but is to innovate.”
In its attempt to define “extend,” the Eleventh Circuit made an
important connection between the clearly established law requirement
and the unreasonable application prong. The Eleventh Circuit declared
that state courts are not required to predict whether the Supreme Court
381
might widen the scope of a rule. In the Eleventh Circuit’s view, a state
court’s failure to widen the scope of “a Supreme Court rule can never be
382
an ‘unreasonable application of clearly established law.’”
The
Eleventh Circuit reasoned that, in such situations, “the state court’s
decision is filling a gap for which the Supreme Court (the state court can
383
reasonably believe) has not yet determined the law.”
While it seemed concerned with judicial activism by the federal habeas
courts in developing constitutional law on habeas review, the Eleventh
Circuit stressed that “not every factual difference between a Supreme
Court precedent and the case before a state court would carry the state
384
case beyond the borders of the Supreme Court’s decisional principle.”
It proposed that the cases must have “substantially different
circumstances,” meaning “differences which objectively reasonable
judges could believe might make a difference on whether or not the
Supreme Court’s preexisting decisional principles laid down in one set of
385
circumstances would apply to the new circumstances.”
The Eleventh
Circuit then drew the final connection:
[A] state court is not obliged to predict more or less accurately
what the Supreme Court might do in circumstances, the likes of
which objectively reasonable judges can believe the Supreme
378. Id. at 1307 n.3.
379. Id. (quoting 5 THE OXFORD ENGLISH DICTIONARY 595 (2d ed. 1989)).
380. Id.
381. Id.
382. Id.
383. Id.
384. Id.
385. Id. The court’s use of a range of terms for precedent, including “precedent,”
“case,” and “preexisting decisional principle,” reflects a common terminology problem
that is not limited to the habeas context.

2005]

Seeking Clarity in the Federal Habeas Fog

805

Court has never faced: in such state cases, no clearly established
federal law exists on point that might be unreasonably applied
386
for AEDPA purposes.
In other words, state courts are prohibited from narrowing Supreme
387
Court rules, but they are not required to widen them.
The Eleventh Circuit ultimately denied relief because it found the facts
388
of petitioner’s case substantially different from those in Kennedy. The
court further concluded that “[t]he Supreme Court [had] not
389
contemplated extending Kennedy to cases like Hawkins’s case.”
Although Hawkins’s facts involved purposeful prosecutorial misconduct
that resulted in a mistrial, it was not the same “goaded-mistrial context”
390
as in Kennedy. The court therefore held that the state court’s decision
was not an unreasonable application of clearly established law and
391
vacated the district court’s order.
The Eleventh Circuit’s decision in Hawkins confuses the contrary to
and the unreasonable application analyses. By insisting that new facts
are “materially” or “substantially the same” as those in the case where
the Supreme Court articulated the rule, the court is describing only one
of two tests under the contrary to prong: a state court decision is contrary
to clearly established law where the habeas case is “materially
indistinguishable” from the Supreme Court precedent, and the state
court has made the opposite conclusion of the Supreme Court. Under
the unreasonable application prong, clearly established law includes rules
that can be reasonably extended to new factual situations that are not
materially indistinguishable. Clearly established law, therefore, is a
broader category than those cases in which the facts are materially or
392
substantially similar to the facts of the habeas case.
The range of factual situations to which a rule may apply in each
context is a question about the scope of the precedent. These questions
cannot be answered at a threshold or abstract level; they must be
addressed under the contrary to or unreasonable application prongs of §
2254(d)(1).

386. Id.
387. Id.
388. Id. at 1309, 1310.
389. Id. at 1309.
390. Id. at 1310.
391. Id.
392. Even in the narrower qualified immunity context, the Supreme Court rejected a
standard that limited clearly established law to those cases with “fundamentally similar”
facts. The Court adopted a standard of “fair warning” instead. See supra note 79.

806

Catholic University Law Review

[Vol. 54:747

C. Lack of Clarity
The third category raises questions about the uncertain state of the
law. A lack of clarity can occur for several reasons. First, the Court may
not have articulated a very specific rule, such that the “precise contours”
393
of the rule are unknown.
Second, the Supreme Court may have
articulated a rule that has been called into question by later decisions or
confusing language, as highlighted in Rico v. Leftridge-Byrd with the
394
Batson rule. Third, the Supreme Court may have addressed the issue
in a nonmajority or splintered decision. This problem raises the question
of whether a majority of the Court must endorse the rule articulated by a
plurality for it to be considered clearly established law “as determined by
395
the Supreme Court of the United States.” All three of these problems
were present in Andrade, where the Supreme Court noted its
396
jurisprudence was “not a model of clarity.”
The lower courts have
followed the Supreme Court’s lead in dealing with these problems under
§ 2254(d)(1)’s unreasonable application prong.
1. General Principles
As seen with the first two categories of recurring issues, some courts
have had trouble applying § 2254(d)(1) where the principle articulated by
the Supreme Court is a general principle. For example, in Cotto v.
397
Herbert, the Second Circuit refused to find unreasonable an extension
of the principle that a defendant can forfeit his constitutional right to
398
confrontation through misconduct.
In Cotto, the petitioner was a
criminal defendant who was found by the trial court to have intimidated
399
the eyewitness of a murder. The eyewitness refused to testify, and the
400
trial court allowed the witness’s hearsay statements into evidence. The
Second Circuit rejected petitioner’s claim that the state court
unreasonably extended the “forfeiture-by-misconduct principle” because
petitioner’s facts were not a new context; rather, the court said, his was
the “paradigmatic example of the type of ‘misconduct’ that can lead to
401
the forfeiture of confrontation rights.”
The Second Circuit then paused at petitioner’s argument that the state
court unreasonably applied the law regarding the sufficiency of the
393.
394.
395.
396.
397.
398.
399.
400.
401.

Lockyer v. Andrade, 538 U.S. 63, 72-73 (2003).
See supra Part III.B.
See 28 U.S.C. § 2254(d)(1) (2000).
Andrade, 538 U.S. at 72.
331 F.3d 217 (2d Cir. 2003).
Id. at 234.
Id. at 227.
Id. at 224, 227.
Id. at 234.

2005]

Seeking Clarity in the Federal Habeas Fog

807

402

evidence necessary for forfeiture. It ultimately denied habeas relief on
this issue, though, because there was no Supreme Court precedent
“definitively establishing” the circumstances or the standard of proof for
403
forfeiture by misconduct.
Throughout its analysis of this issue, the
Second Circuit referred to its own precedent and that of other circuit
courts as evidence of “extensive federal precedent” permitting out-of404
court statements when a defendant has intimidated a witness. But the
lack of Supreme Court precedent limiting the circumstances that
constitute forfeiture by misconduct prohibited the court from finding that
the state court decision involved an unreasonable application of clearly
405
established law.
The Cotto court considered a second, and related, issue of whether
there was clearly established law on the waiver of the right to cross406
examination through misconduct.
The Second Circuit noted that the
Supreme Court had addressed this issue directly only once, in an 1878
407
case, Reynolds v. United States. Reynolds, the court recognized, had a
narrow holding that did not speak to the precise question of the scope of
forfeiture-by-misconduct of the right to cross-examine a witness when
408
the witness testifies at trial. And even though the law in this area had
developed primarily in the federal courts of appeals, none have faced this
precise issue.
Given other similar decisions, one would expect that the Second
Circuit would have denied habeas relief on the grounds that the state
court decision was not an unreasonable application of clearly established
law. The Second Circuit did the opposite: it granted habeas relief
409
because the state court decision was objectively unreasonable.
The
court gave three reasons for its decision. First, the court reasoned that
the lack of any federal precedent—Supreme Court or courts of appeals—
supported the state court’s determination that the petitioner had waived
410
his right to cross-examine a witness at trial.
In doing so, the court
402. Id. at 234-35.
403. Id. at 234.
404. Id. at 235.
405. Id.
406. Id. at 249.
407. Id.; Reynolds v. United States, 98 U.S. 145 (1878). The court also noted that even
though the circuits are split on the proper application of the rule from Reynolds, the
Supreme Court has repeatedly declined to resolve the split. Cotto, 331 F.3d at 249.
408. Cotto, 331 F.3d at 249-50.
409. Id. at 251.
410. Id. The court noted that in Williams v. Taylor, 529 U.S. 362 (2000), the Supreme
Court held that the Virginia Supreme Court had unreasonably applied the law, even
though several federal courts of appeals and other state courts had applied the law in a
similar fashion, Cotto, 331 F.3d at 251.

808

Catholic University Law Review

[Vol. 54:747

seemed to be placing the burden of citing appropriate Supreme Court
precedent on the state: “[R]espondent has pointed to no other case—and
411
our research reveals none—in which this occurred.” Second, the court
saw a lack of specific reasons to extend the preclusion of cross412
examination to this extreme, i.e., a complete ban.
Third, the court
considered the preclusion of cross-examination under these
circumstances to be inconsistent with the underlying purpose of the
413
broader forfeiture-by-misconduct rule.
The Second Circuit’s decision has several controversial aspects. The
first is the burden of persuasion. It is inappropriate for the court to shift
the burden of citing clearly established law to the state; the habeas
petitioner bears the burden on federal habeas review to prove the state
414
courts were in error.
Furthermore, many courts would agree that
“[t]he Supreme Court’s silence on a particular issue cannot constitute
415
‘clearly established’ Federal law.”
In Cotto, the court seemed to find
the opposite. But the court was not operating in a complete absence of
precedent; it was dealing with a broad principle articulated in one
century-old Supreme Court case. Although the principle may have been
broad, it was enough to pass through the threshold question of §
2254(d)(1). Perhaps the most controversial aspect of the case is the
determination that the state court unreasonably applied this broad
principle because it extended the principle to a new set of facts. While
rare, such a determination is consistent with AEDPA as interpreted by
the Supreme Court. The Second Circuit supported its holding with
reasons grounded in the fundamental purposes of the Reynolds rule and
of cross-examination.
2. Old or Otherwise Questionable Precedent
As in Cotto, lower courts sometimes must determine whether a
Supreme Court case is still “good law” even though it is old or has been
called into question by subsequent Supreme Court decisions. These
questions go to how “clearly” the law has been “established.” Like in
Andrade, they often involve broad principles, the “precise contours” of
416
which are not settled.
The Seventh Circuit encountered one such broad principle in Jackson
417
v. Frank. Jackson involved a defendant who told the detective during
411.
412.
413.
414.
415.
416.
417.

Cotto, 331 F.3d at 251.
Id. at 251-52.
Id. at 252.
See Hill v. Hofbauer, 337 F.3d 706, 712 n.3 (6th Cir. 2003).
Id.
Lockyer v. Andrade, 538 U.S. 63, 72-73 (2003).
348 F.3d 658 (7th Cir. 2003), cert. denied, 124 S. Ct. 1723 (2004).

2005]

Seeking Clarity in the Federal Habeas Fog

809

418

interrogation that he wanted a lawyer “right now.”
The detective
misstated the state law and told the defendant that he was unable to get
him a lawyer and would have to end the interview unless the defendant
419
agreed to continue without a lawyer.
The defendant then waived his
420
Miranda rights and confessed to the crime. He later argued the waiver
421
was not voluntary and moved to suppress the confession.
On habeas
review, the defendant claimed that his Miranda rights were violated
because the detective’s statements clouded his understanding of his Fifth
422
Amendment right as recognized in Miranda.
The confusion about this claim centered on the holding from
423
Duckworth v. Eagan.
“In Duckworth, the Court held that when a
suspect was informed that he would be provided an attorney ‘if and when
[he] went to court,’ his subsequent waiver under Miranda was
424
voluntary.”
After the Court decided Duckworth, however, it found
misrepresentations by police to violate defendants’ Miranda rights in
425
some circumstances.
The Seventh Circuit’s uncertainty as to how to
reconcile Duckworth with these cases led the court to conclude that there
426
was no clearly established law. The Seventh Circuit held that, in light
418. Id. at 660.
419. Id. at 660-61. The court found that the detective’s statement was misleading
because state law made public defenders available on an emergency basis to suspects in
custody. Id. at 661.
420. Id.
421. Id.
422. Id. at 663. He sought relief on two grounds. His first claim was that his Miranda
rights were violated because the detective’s statement was contrary to state law. Id. This
argument failed because state law is more generous than federal law; Miranda does not
require that a lawyer be made immediately available to a defendant. Id.
423. 492 U.S. 195 (1989).
424. Jackson, 348 F.3d at 663-64 (quoting Duckworth, 492 U.S. at 203-04). In Jackson,
the defendant otherwise received warnings that comported with Miranda. Id.
425. Id.
426. Id. In doing so, the Seventh Circuit relied heavily on a Fifth Circuit habeas
decision on a similar issue. Id. at 665. In Soffar v. Cockrell, 300 F.3d 588 (5th Cir. 2002)
(en banc), the Fifth Circuit denied habeas relief for a petitioner who claimed that he had
invoked his right to counsel during police interrogations, id. at 598. The court explored
several Supreme Court precedents and concluded that the line of cases did not apply to
petitioner’s situation, and, even if it did, petitioner would be barred from relying on those
cases under Teague’s anti-retroactivity principle. Id. at 596-97. Three of the en banc
judges dissented. Id. at 598. Judge DeMoss, joined by Judges Parker and Dennis, strongly
disagreed with both the majority’s interpretation of the relevant law and its application of
the law to the facts of the case. Id. One other member of the en banc court concurred in
part and in the judgment. Id. at 590. Two judges concurred in the judgment only. Id.
Teague’s anti-retroactivity rule was raised in another recent AEDPA case, Hill v.
Hofbauer, 337 F.3d 706, 713 (6th Cir. 2003). In Hill, one of the cases relied on by
petitioner was decided a year after his conviction was affirmed by the state court of
appeals. Id. at 712. The state argued that the case was therefore a “new rule” under

810

Catholic University Law Review

[Vol. 54:747

of Duckworth, the state court opinion was not objectively
427
unreasonable.
428
Similarly, in Wilson v. Superintendent of Attica Correctional Facility,
a federal magistrate judge had to evaluate the extent to which Supreme
Court precedent regarding how to review alleged error by the trial court
429
was clearly established.
The petitioner claimed, among other things,
that the state trial court judge had improperly ruled that photographs of
petitioner’s apartment and testimony about the condition of the
apartment could be used as evidence at trial even though the judge had
suppressed items seized from the apartment because they were products
430
of an illegal search.
The magistrate judge noted that the standard of
431
review articulated in Brecht v. Abrahamson for habeas review of a trial
court’s errors had been called into question by several federal circuit
432
courts. While some circuit courts questioned the continuing validity of
the Brecht test in light of AEDPA, other circuits held that the Brecht test
433
was still good law.
After noting that the Supreme Court had cited

Teague, and thus not clearly established law under AEDPA. Id. Although the Sixth
Circuit agreed that reliance on that case was improper, it held that the Supreme Court had
clearly established the relevant legal principle—“that a co-defendant’s custodial
confessions are unreliable and not within a ‘firmly rooted’ hearsay exception”—in three
prior cases that predated petitioner’s conviction. Id. at 712, 717. The court went on to
hold that the state court’s decision was both contrary to the law established in those three
cases and an unreasonable application of the legal principles espoused in those cases. Id.
427. Jackson, 348 F.3d at 659.
428. No. 9:00-CV-0767 (NAM/GLS), 2003 U.S. Dist. LEXIS 2111 (N.D.N.Y. Nov. 24,
2003).
429. Id. at *24-25.
430. Id. at *24.
431. 507 U.S. 619 (1993). The Brecht test examines whether the trial court’s error had
a “‘substantial and injurious effect or influence in determining the jury’s verdict.’” Id. at
637 (quoting Kotteakos v. United States, 328 U.S. 750, 776 (1946)).
432. Wilson, 2003 U.S. Dist. LEXIS 2111, at *24-25.
433. Id. The question is whether under AEDPA federal courts should review errors
found to be harmless by state courts using the Brecht test or using the “harmless beyond a
reasonable doubt” test articulated in Chapman v. California, 386 U.S. 18 (1967). Wilson,
2003 U.S. Dist. LEXIS 2111, at *25. The Seventh and Eighth Circuits have suggested that
the Chapman test apply rather than the Brecht test. See Anderson v. Cowan, 227 F.3d 893,
897 (7th Cir. 2000); Whitmore v. Kemna, 213 F.3d 431, 433-34 (8th Cir. 2000). The Fifth,
Sixth, and Tenth Circuits have held that Brecht is the appropriate test. See Hill v.
Hofbauer, 337 F.3d 706, 718 (6th Cir. 2003); Robertson v. Cain, 324 F.3d 297, 299 (5th Cir.
2003); Herrera v. Lemaster, 301 F.3d 1192, 1200 (10th Cir. 2002). While the Second
Circuit has noted this unresolved issue, it has not had the occasion to decide it. E.g., Cotto
v. Herbert, 331 F.3d 217, 253-54 (2d Cir. 2003) (holding that the error was not harmless
under both standards and thus declining to resolve the issue).

2005]

Seeking Clarity in the Federal Habeas Fog

811

434

Brecht on two different occasions since AEDPA was enacted, the
magistrate judge played it safe and applied both standards.
These cases reveal that courts determining the state of the law typically
do so under the unreasonable application prong of § 2254(d)(1) rather
than as a dispositive issue. A question about the clarity of the Supreme
Court’s precedent can affect the reasonableness of a state court’s
decision in applying that precedent. The more uncertain the precedent,
the more leeway the state court has in reasonably applying it.
3. Splintered Decisions
An added level of complexity manifests when the Supreme Court has
addressed a particular issue and articulated a rule but has done so in a
splintered decision. Since the precedential value of nonmajority opinions
435
is often uncertain even among legal scholars, one might anticipate a
high degree of confusion or inconsistency by habeas courts attempting to
determine whether a nonmajority opinion constitutes clearly established
law. However, several habeas courts have overcome this hurdle by
436
turning to Marks v. United States, in which the Supreme Court

434. Wilson, 2003 U.S. Dist. LEXIS 2111, at *25 n.11. The Supreme Court cited
Brecht in Penry v. Johnson, 532 U.S. 782, 795 (2001), and Early v. Packer, 537 U.S. 3, 10
(2002), but it has not specifically held that the Brecht test survived the enactment of
AEDPA, Wilson, 2003 U.S. Dist. LEXIS 2111, at *25 n.11. The two citations do, however,
post-date the circuit court decisions calling Brecht’s validity into question.
435. See generally Aldisert, supra note 22, at 605; John F. Davis & William L.
Reynolds, Juridical Cripples: Plurality Opinions in the Supreme Court, 1974 DUKE L.J. 59;
Ken Kimura, A Legitimacy Model for the Interpretation of Plurality Decisions, 77
CORNELL L. REV. 1593 (1992); G.P.J. McGinley, The Search for Unity: The Impact of
Consensus Seeking Procedures in Appellate Courts, 11 ADEL. L. REV. 203 (1987); Burt
Neuborne, The Binding Quality of Supreme Court Precedent, 61 TUL. L. REV. 991 (1987);
Laura Krugman Ray, The Justices Write Separately: Uses of the Concurrence by the
Rehnquist Court, 23 U.C. DAVIS L. REV. 777 (1990); David C. Bratz, Comment, Stare
Decisis in Lower Courts: Predicting the Demise of Supreme Court Precedent, 60 WASH. L.
REV. 87 (1984); Adam S. Hochschild, Note, The Modern Problem of Supreme Court
Plurality Decisions: Interpretation in Historical Perspective, 4 WASH. U.J.L. & POL’Y 261
(2000); Igor Kirman, Note, Standing Apart To Be a Part: The Precedential Value of
Supreme Court Concurring Opinions, 95 COLUM. L. REV. 2083 (1995); Linda Novak,
Note, The Precedential Value of Supreme Court Plurality Decisions, 80 COLUM. L. REV.
756 (1980); William G. Peterson, Note, Splintered Decisions, Implicit Reversal and Lower
Federal Courts: Planned Parenthood v. Casey, 1992 BYU L. REV. 289 (1992); Note,
Plurality Decisions and Judicial Decisionmaking, 94 HARV. L. REV. 1127 (1981);
Comment, Supreme Court No-Clear-Majority Decisions: A Study in Stare Decisis, 24 U.
CHI. L. REV. 99 (1956); Mark Alan Thurmon, Note, When the Court Divides:
Reconsidering the Precedential Value of Supreme Court Plurality Decisions, 42 DUKE L.J.
419 (1992); cf. Erwin Chemerinsky & Barry Friedman, The Fragmentation of Federal
Rules, 46 MERCER L. REV. 757 (1995) (discussing in the context of civil procedure).
436. 430 U.S. 188 (1977).

812

Catholic University Law Review

[Vol. 54:747

developed a doctrine for establishing controlling precedent from
437
nonmajority opinions.
Marks involved an appeal from a conviction for transportation of
438
obscene materials.
The Court had to examine two of its previous
fragmented cases in order to evaluate the then-current state of the law,
stating that “[w]hen a fragmented Court decides a case and no single
rationale explaining the result enjoys the assent of five Justices, ‘the
holding of the Court may be viewed as that position taken by those
439
Members who concurred in the judgments on the narrowest grounds.’”
Simply put, the controlling precedent is not necessarily the plurality
opinion, but the opinion that “represent[s] a common denominator of the
Court’s reasoning; it must embody a position implicitly approved by at
440
least five Justices who support the judgment.”
Although lower courts sometimes struggle with properly applying the
441
Marks doctrine, most habeas courts invoking the doctrine to identify
clearly established law have done so with relative ease. Murillo v.
442
Frank is a good example of Marks doctrine application in the habeas
corpus context. In Murillo, the basis of the petitioner’s habeas claim was
that his confrontation rights were violated when his brother’s statements
443
made during a police interrogation were admitted into evidence. The
petitioner’s brother, who was also a suspect in the homicide
investigation, implicated the defendant in the murder and subsequently
refused to take the stand at trial, thus denying the defendant an
444
opportunity to cross-examine his brother.
The district court noted that
the state court’s decision was contrary to current law because the
445
Supreme Court had recently held in Crawford v. Washington that the
Confrontation Clause proscribes the use of statements against the
446
defendant made by a non-testifying witness during a police interview.
But the Court decided Crawford after the defendant’s conviction became
447
final. Because Crawford overruled previous Supreme Court precedent

437. Id. at 193.
438. Id. at 189.
439. Id. at 193 (quoting Gregg v. Georgia, 428 U.S. 153, 169 n.15 (1976)).
440. King v. Palmer, 950 F.2d 771, 781 (D.C. Cir. 1991).
441. See, e.g., Hopwood v. Texas, 236 F.3d 256, 275 n.66 (5th Cir. 2000) (criticizing the
Ninth Circuit’s application of Marks in Smith v. University of Washington, 233 F.3d 1188
(9th Cir. 2000)).
442. 316 F. Supp. 2d 744 (E.D. Wis. 2004).
443. Id. at 746-47.
444. Id. at 747.
445. Crawford v. Washington, 124 S. Ct. 1354 (2004).
446. Id. at 1374.
447. Murillo, 316 F. Supp. 2d at 749.

2005]

Seeking Clarity in the Federal Habeas Fog

813

448

(Ohio v. Roberts ), the district court considered it to be a “new rule”
449
and consequently not clearly established law for purposes of AEDPA.
The defendant therefore relied on the nonmajority Supreme Court
450
opinion in Lilly v. Virginia.
The Lilly plurality stated that the Confrontation Clause allows
admission of statements “‘[w]hen a court can be confident—as in the
context of hearsay falling within a firmly rooted exception—that the
declarant’s truthfulness is so clear from the surrounding circumstances
451
that the test of cross-examination would be of marginal utility.’”
In
addition, the plurality pointed out that the Court has “‘over the years
spoken with one voice in declaring presumptively unreliable accomplices’
452
confessions that incriminate defendants.’” Justice Scalia concurred in
the judgment, phrasing the use of recorded statements without making
the witness available for cross-examination in Lilly as “‘a paradigmatic
453
Confrontation Clause violation.’”
The broad language in Justice Scalia’s concurrence led the district
court to consider the plurality the more narrow of the two opinions and
454
thus the controlling rationale under Marks. The district court granted
the petition, reasoning that the state court’s conviction was contrary to
clearly established law because the petitioner’s facts were materially
455
indistinguishable from those in Lilly.
456
In Richmond v. Polk, the habeas court utilized the Marks doctrine to
find that the state court had unreasonably applied clearly established
457
law. In Richmond, the petitioner was convicted of three counts of first448. 448 U.S. 56, 73 (1980) (upholding the admission of preliminary hearing testimony
of an unavailable witness because the circumstances under which the prior testimony was
given provided “sufficient indicia of its reliability”).
449. Murillo, 316 F. Supp. 2d at 749. The district court noted that it was a close
question whether Crawford overruled Roberts because the Supreme Court had not
explicitly applied Roberts to testimonial statements. Id. n.4.
450. 527 U.S. 116 (1999) (plurality opinion).
451. Murillo, 316 F. Supp. 2d at 752 (internal quotation marks omitted) (quoting Lilly,
527 U.S. at 136 (plurality opinion)).
452. Id. at 751 (internal quotation marks omitted) (quoting Lilly, 527 U.S. at 131
(plurality opinion)).
453. Id. at 753 (quoting Lilly, 527 U.S. at 143 (Scalia, J., concurring)).
454. Id. The district court acknowledged that there were some factual differences
between Lilly and Murillo’s case, but concluded that the differences were not legally
significant. Id. at 754. In doing so, the court noted that “it is not enough that there are
differences between the two cases. No two cases are ever exactly alike in all respects. To
warrant different conclusions, the differences must be material; they must be legally
significant.” Id.
455. Id. at 756.
456. 375 F.3d 309 (4th Cir. 2004).
457. Id. at 331 n.10.

814

Catholic University Law Review

[Vol. 54:747

degree murder and one count of first-degree rape and was sentenced to
458
death. On habeas review, he claimed, inter alia, that the state court’s
refusal to instruct the jury about his parole ineligibility for a prior federal
murder conviction was contrary to or an unreasonable application of the
459
Supreme Court’s decision in Simmons v. South Carolina. The Fourth
Circuit applied the Marks doctrine and found Justice O’Connor’s
concurring opinion in Simmons to be the narrowest grounds on which the
460
concurring justices agreed, and therefore controlling precedent.
In Simmons, a plurality of the Court concluded that due process
requires a sentencing jury be informed that a defendant is parole
ineligible if the defendant’s future dangerousness is at issue and state law
461
disallows parole. Justice O’Connor concurred in the judgment, stating
that “‘[w]here the State puts the defendant’s future dangerousness in
issue, and the only available alternative sentence to death is life
imprisonment without possibility of parole, due process entitles the
defendant to inform the capital sentencing jury . . . that he is parole
462
ineligible.’” Justice O’Connor also noted, however, that the defendant
does not have this right when the prosecution limits its dangerousness
463
argument to the defendant’s potential dangerousness in prison.
Disagreeing with the state court’s conclusion that the prosecution’s
dangerousness argument was limited to dangerousness in prison, the
Fourth Circuit held that the state court unreasonably applied clearly
464
established law.
465
Coe v. Bell is another example of a habeas court invoking the Marks
doctrine to find clearly established law. In Coe, the Sixth Circuit was
faced with applying the fragmented Supreme Court opinion of Ford v.
466
Wainwright
to determine whether the petitioner was mentally
467
competent to be executed.
In his concurring opinion in Ford, Justice
Powell stated that prisoners will be deemed insane for competency
458. Id. at 314.
459. Id. at 314, 331; Simmons v. South Carolina, 512 U.S. 154 (1994) (plurality
opinion).
460. Polk, 375 F.3d at 331 n.10; accord Smallwood v. Gibson, 191 F.3d 1257, 1280 n.15
(10th Cir. 1999) (acknowledging Justice O’Connor’s concurrence in Simmons as
controlling precedent).
461. Simmons, 512 U.S. at 156 (plurality opinion).
462. Polk, 375 F.3d at 331 (alteration in original) (quoting Simmons, 512 U.S. at 178
(O’Connor, J., concurring)).
463. Id.
464. Id. at 335. However, the Fourth Circuit denied the petition because it found the
state court’s error to be harmless. Id.
465. 209 F.3d 815 (6th Cir. 2000).
466. 477 U.S. 399 (1986).
467. Coe, 209 F.3d at 818.

2005]

Seeking Clarity in the Federal Habeas Fog

815

purposes when “unaware of the punishment they are about to suffer and
468
why they are about to suffer it.” Justice Powell also concluded that the
process of evaluating a prisoner’s competency must comport with due
469
process, and the prisoner must be afforded a “fair hearing.” The Sixth
Circuit found Justice Powell’s concurring opinion to be the narrowest
grounds on which all concurring justices agreed, since the plurality
opinion, penned by Justice Marshall, advocated a more rigorous
470
competency evaluation process.
To bolster its view that Justice
Powell’s concurring opinion served as the opinion of the Court, the Sixth
Circuit cited a later Supreme Court case, which seemed to support the
471
view that Justice Powell’s concurrence was controlling precedent. The
Sixth Circuit then denied the petition, holding that the competency
hearing met the minimum requirements as enunciated by Justice Powell’s
472
concurring opinion.
The Marks doctrine can be critical in identifying whether there is
clearly established law. In some situations, however, the doctrine cannot
473
be applied logically. The Supreme Court has agreed that Marks is not
always an appropriate method for gleaning precedent from fragmented
474
cases.
Despite its difficulties in application, the Marks doctrine is
available to habeas courts and, at least in certain contexts, should be used
when a nonmajority Supreme Court opinion would otherwise prevent the
475
lower court from identifying clearly established law.
The divergence in the lower courts’ approaches about how to analyze
clearly established law issues appears to stem from the type of issue the

468. Id. at 818-19 (quoting Ford, 477 U.S. at 422 (Powell, J., concurring)).
469. Id. at 819 (quoting Ford, 477 U.S. at 424 (Powell, J., concurring)).
470. Id. n.1.
471. Id. at 819 (citing Penry v. Lynaugh, 492 U.S. 302, 333 (1989)).
472. Id. at 827-28.
473. See, e.g., King v. Palmer, 950 F.2d 771, 782 (D.C. Cir. 1991) (noting that when
“one opinion supporting the judgment does not fit entirely within a broader circle drawn
by the others, Marks is problematic. If applied in situations where the various opinions
supporting the judgment are mutually exclusive, Marks will turn a single opinion that lacks
majority support into national law”).
474. See Nichols v. United States, 511 U.S. 738, 745-46 (1994) (stating that “[w]e think
it not useful to pursue the Marks inquiry to the utmost logical possibility when it has so
obviously baffled and divided the lower courts that have considered it”).
475. This Article does not take a position on whether the Marks doctrine is
normatively the best solution; it treats Marks as current available doctrine. Others have
addressed this normative question. See generally Hochschild, supra note 435; Lewis A.
Kornhauser & Lawrence G. Sager, The One and the Many: Adjudication in Collegial
Courts, 81 CAL. L. REV. 1 (1993); Richard L. Revesz & Pamela S. Karlan, Nonmajority
Rules and the Supreme Court, 136 U. PA. L. REV. 1067 (1988); Maxwell Stearns, The Case
for Including Marks v. United States in the Canon of Constitutional Law, 17 CONST.
COMMENT. 321 (2000); Thurmon, supra note 435; Kirman, supra note 435.

816

Catholic University Law Review

[Vol. 54:747

court faces. Viewing clearly established law as a dispositive issue, some
courts deny relief under § 2254(d)(1) when the petitioner fails to cite any
Supreme Court precedent at all or cites only one Supreme Court case
476
that was decided after the relevant state court decision was issued.
When courts address issues of the scope of the precedent, however, they
tend to do so under the contrary to and unreasonable application prongs
of § 2254(d)(1). Similarly, courts tend not to treat issues about the clarity
of the precedent as dispositive; instead, these are addressed as part of the
unreasonable application analysis.
IV. TOWARD A COMMON UNDERSTANDING
The struggles of the lower courts show that a common understanding
of both the definition of clearly established law and the appropriate
analytical framework is necessary. Although courts are regularly
encountering the same types of issues, they are not employing a common
framework for analyzing those issues.
This Part offers an analytical framework to assist federal habeas courts
with the often difficult task of determining what constitutes clearly
established law under § 2254(d)(1). The framework begins with the
threshold question of whether there is clearly established law. It then
identifies five analytic touchstones relevant to this question: (1) the
source of the precedent, (2) the number of cases, (3) the timing of the
precedent, (4) the scope of the precedent, (5) and the strength of the
precedent. The proposed framework also provides a role for federal
courts of appeals precedent to serve as evidence relating to these
touchstones.
While these touchstones are guideposts along the way for courts, their
use will not necessarily lead to identical results in all cases. Indeed, the
very nature of our common law system is that “holdings” is a flexible
concept—they are often arguable as broad or narrow. This Article does
not attempt to resolve these inherent problems—that task has been
477
taken on by others —but rather offers a framework to make the inquiry
more uniform. Courts should at least be asking the same questions, even
if they do not reach the same conclusions.
This Part then proposes a methodology based on the five touchstones.
It recommends that the first three touchstones— source of the precedent,
number of cases, and timing of the precedent—are dispositive issues
requiring the court to deny habeas relief if they are not met. If, however,
a habeas court encounters questions relating to the final two
476. This issue is more complex than it sounds because it raises a question under the
Teague doctrine.
477. See supra note 22.

2005]

Seeking Clarity in the Federal Habeas Fog

817

touchstones—the scope of the precedent or the strength of the
precedent—the court should continue with the analysis under §
2254(d)(1) and address those questions under the contrary to or
unreasonable application prongs. This Part concludes by advocating that
federal courts employ a sliding scale of deference based on the
reasonableness of the court’s assessment of the precedent’s scope and
strength, in other words, based on how “clearly” the law is established.
A. Analytic Touchstones for Determining Whether the Law Is Clearly
Established
Although these are not bright-line rules, these considerations provide
some parameters for understanding what constitutes clearly established
law under § 2254(d)(1).
1. Dispositive Questions: Source, Number, and Timing of Precedent
The first three touchstones are the source, number, and timing of the
precedent. These three touchstones are the most straightforward in the
sense that they are most easily measurable. They are interconnected
because each is a dispositive question. In other words, the petitioner
must cite at least one Supreme Court case that bears on petitioner’s
claim, and that case must have been decided before the relevant state
court decision on petitioner’s claim. If the petitioner fails to do so, the
petition should be denied under § 2254(d)(1).
The source of the relevant precedent is the first touchstone. The
Supreme Court clarified this issue in Williams when it interpreted “as
determined by the Supreme Court of the United States” to refer to
478
Supreme Court precedent only. This source limitation is a change from
pre-AEDPA law under Teague, which allowed the use of circuit court
479
precedent.
The purpose of this first touchstone is to determine whether the
petitioner’s case hinges on an issue that the Supreme Court has not
addressed by the time of the relevant state court decision. It requires the
habeas court to separate issues of first impression from issues regarding
the scope or strength of the precedent. The most straightforward
example of an issue of first impression is where the Court expressly
480
reserved the issue for another case.
Both Supreme Court and lower
federal court opinions may provide evidence that the issue was left open
by the Supreme Court. In true cases of first impression, the lack of
clearly established law should be dispositive.
478. Williams v. Taylor, 529 U.S. 362, 412 (2000).
479. See supra Part II.A.
480. For an example, see supra Part III.A.

818

Catholic University Law Review

[Vol. 54:747

The source limitation is related to the second touchstone: the number
of cases required. The habeas petitioner must cite at least one Supreme
481
Court case that bears on petitioner’s claim.
Timing is the third touchstone. In Williams, the Court announced that
the relevant timeframe for purposes of clearly established law is the date
the state court issued the decision rejecting the petitioner’s constitutional
482
claims.
2. Scope of Precedent
The fourth touchstone steps away from the more determinative issues
of source, number, and timing and focuses on the scope of the precedent.
The scope of the precedent refers to the breadth of the issue decided and
483
the number of possible cases that the precedent case controls.
Ascertaining the scope of a precedent is a two-step inquiry. The first
step is to determine the breadth of the legal directive articulated in the
precedent. The second step is to compare the factual similarity of the
precedent case and the petitioner’s case.
a. Breadth of the Legal Directive
In determining how broad or narrow the precedent is, the court should
first ask whether the Supreme Court articulated a rule or standard.
Although the differences between rules and standards have been well
documented by others, a brief overview is necessary here. Rules and
484
standards are both forms of legal directives. As Professor Pierre Schlag
has pointed out, “[D]irectives can be general or specific, conditional or

481. See, e.g., Murillo v. Frank, 316 F. Supp. 2d 744, 748 (E.D. Wis. 2004) (explaining
that habeas petitioners “‘must show that the Supreme Court has clearly established the
propositions essential to their position’” (internal quotation marks omitted) (quoting
Mueller v. Sullivan, 141 F.3d 1232, 1234 (7th Cir. 1998))); Gonzalez v. Fischer, No. 01-CV8523 (JBW), 03-MISC-0066 (JBW), 2003 U.S. Dist. LEXIS 23874, at *38 (E.D.N.Y. Oct.
16, 2003) (stating that a habeas petitioner must “‘identify a clearly established Supreme
Court precedent that bears on his claim’” (quoting Loliscio v. Goord, 263 F.3d 178, 191
(2d Cir. 2001); Sellan v. Kuhlman, 261 F.3d 303 (2d Cir. 2001))). The district court need
not proceed with the analysis if the petitioner fails to cite a Supreme Court case. Id.
[A] “petitioner first must show that the Supreme Court has ‘clearly established’
the proposition essential to [his] position.” . . . “[He] must have a Supreme
Court case to support his claim, and that Supreme Court decision must have
clearly established the relevant principle as of the time of his direct appeal.”
Estrada v. Jones, No. 03 C 3092, 2004 U.S. Dist. LEXIS 6502, at *4-5 (N.D. Ill. Apr. 14,
2004) (quoting Schaff v. Snyder, 190 F.3d 513, 521 (7th Cir. 1999)).
482. Williams, 529 U.S. at 412. This statement has generally been interpreted to mean
when the state court conviction becomes final. See Bryant, supra note 6, at 45. It does
leave open, however, the question of whether the Teague exceptions apply under § 2254.
483. ALEXANDER, Precedent, supra note 25, at 167.
484. Pierre Schlag, Rules and Standards, 33 UCLA L. REV. 379, 381-82 (1985).

2005]

Seeking Clarity in the Federal Habeas Fog

819

485

absolute, narrow or broad, weak or strong.”
Most law students,
lawyers, and judges are familiar with talking about rules and standards as
486
“‘bright line rule[s]’” and “‘[generalized] standard[s].’”
Although the terms “rules” and “standards” do not have fixed
487
meanings, they are generally understood in the following way.
Rules
are thought to provide guidance to those who must follow them and limit
the discretion of those who must apply them; they “establish legal
boundaries based on the presence or absence of well-specified triggering
488
489
facts.” A classic example of a rule is “[n]o dogs allowed.” The term
“bright-line rule” often refers to this type of legal norm.
490
Standards, by contrast, allow judges greater discretion in application.
Unlike rules, standards “incorporate into the legal pronouncement a
range of facts that are too broad, too variable, or too unpredictable to be
491
cobbled into a rule.”
A classic example here is “[n]o unreasonably
492
annoying animals allowed.”
The term “generalized standard” usually
refers to this type of legal norm.

485. Id. (footnotes omitted).
486. Id. at 379; see, e.g., Overton v. Newton, 295 F.3d 270, 278 (2d Cir. 2002)
(“[F]ederal law, as determined by the Supreme Court, may as much be a generalized
standard that must be followed, as a bright-line rule designed to effectuate such a standard
in a particular context.”)
487. Various definitions have been articulated in the scholarly literature over time.
See, e.g., Schlag, supra note 484, at 382 n.16; see also Isaac Ehrlich & Richard Posner, An
Economic Analysis of Legal Rulemaking, 3 J. LEGAL STUD. 257, 258 (1974) (distinguishing
rules from standards on the grounds of precision and generality); Roscoe Pound,
Hierarchy of Sources and Forms in Different Systems of Law, 7 TUL. U. L. REV. 475, 482,
485 (1933) (defining rules as prescribing “definite, detailed legal consequence[s] to a
definite, [set of] detailed . . . facts” and standards as specifying a general limit of
permissible conduct requiring application in view of the particular facts of the case).
488. Russell B. Korobkin, Behavioral Analysis and Legal Form: Rules v. Standards
Revisited, 79 OR. L. REV. 23, 25 (2000); see also Pettys, supra note 12, at 790.
489. See Chen, supra note 12, at 600-01.
490. See Pettys, supra note 12, at 790-91.
491. Korobkin, supra note 488, at 25-26.
492. See Chen, supra note 12, at 600-01.

820

Catholic University Law Review

[Vol. 54:747

493

Federal constitutional law has both rules and standards.
For
purposes of habeas relief, both rules and standards can qualify as clearly
494
established law under AEDPA.
As Justice Stevens recognized in
Williams, “[R]ules of law may be sufficiently clear for habeas purposes
even when they are expressed in terms of a generalized standard rather
495
than as a bright-line rule.”
But whether the relevant Supreme Court
precedent is a rule or standard can affect the scope of the precedent for
purposes of qualifying as clearly established law. Although the usage of
these terms by the courts is not precise, typically a standard will be
496
broader and will require case-by-case analysis.
Professor Todd Pettys has suggested that the rules-standards
distinction can play a role under § 2254(d)(1)’s unreasonable application
497
prong. He argues that “the more the governing legal directive appears
in the form of a standard, the wider the range of outcomes that may
reasonably be deemed permitted by that directive; the more the directive
resembles a rule, the narrower the range of outcomes that it may

493. Pettys, supra note 12, at 789. Professor Pettys notes that, in recent years,
standards have come to dominate in certain areas. Id. n.287. For example, standards are
abundant in the Fourth Amendment search and seizure area:
[W]hether a search or seizure has occurred, whether a search or seizure was
reasonable, whether an officer had probable cause to arrest, whether an officer
had sufficient cause to justify an investigative stop, and whether a warrant was
required for a search. . . . In contrast, [the Court’s decision in] Miranda v.
Arizona issued a directive that . . . is more in the nature of a rule: a prosecutor
may not introduce any evidence “stemming from custodial interrogation of the
defendant unless it demonstrates the use of procedural safeguards effective to
preserve the privilege against self-incrimination.”
Id. at 792 (footnotes omitted).
494. See Chen supra note 12, at 600-01; Pettys, supra note 12, at 792-93.
495. 529 U.S. 362, 382 (2000) (opinion of Stevens, J.). Justice Stevens cited Justice
Kennedy’s concurring opinion in Wright v. West, 505 U.S. 277, (1992), as support:
“If the rule in question is one which of necessity requires a case-by-case
examination of the evidence, then we can tolerate a number of specific
applications without saying that those applications themselves create a new rule.
. . . Where the beginning point is a rule of this general application, a rule
designed for the specific purpose of evaluating a myriad of factual contexts, it
will be the infrequent case that yields a result so novel that it forges a new rule,
one not dictated by precedent.”
Id. (omission in original) (quoting Wright, 505 U.S. at 308-09 (Kennedy, J., concurring in
the judgment)).
496. The Court has acknowledged the diversity of possible outcomes broadly framed
standards or rules may reasonably be interpreted to permit. See Yarborough v. Alvarado,
124 S. Ct. 2140, 2149 (2004); see also Overton v. Newton, 295 F.3d 270, 278 (2d Cir. 2002)
(“[F]ederal law, as determined by the Supreme Court, may as much be a generalized
standard that must be followed, as a bright-line rule designed to effectuate such a standard
in a particular context.”).
497. See Pettys, supra note 12, at 792-93.

2005]

Seeking Clarity in the Federal Habeas Fog

821

498

reasonably be interpreted to permit.”
I agree with Professor Pettys
that
the more general the terms appearing in the governing legal
directive, as construed in the controlling Supreme Court
precedent, the broader the range of outcomes a state court may
reasonably interpret that legal directive to permit, and the
greater the likelihood that the state court’s ruling should be left
499
undisturbed.
Thus, when the relevant precedent is a standard, the clearly established
law inquiry should not be dispositive; the court will need to analyze the
state court decision to see if it has reasonably applied the standard to the
habeas petitioner’s facts. Even when dealing with a rule, the breadth of
500
that rule may require case-by-case analysis, as it did in Andrade.
Reasonable “extensions” of Supreme Court precedent are expressly
permissible under the Court’s interpretation of AEDPA, as are
applications of general principles to specific factual situations. The more
general the directive, however, the more reasonable the state court
decision not to extend becomes.
b. Factual Similarity
To qualify as clearly established law, a Supreme Court case need not
be “on all fours” with the petitioner’s case. Identical facts are not
501
required.
While a federal court may grant habeas relief under the
contrary to prong of § 2254(d)(1) if the state court decision reached a
different result than a “materially indistinguishable” Supreme Court
precedent, relief is not limited to such cases. Indeed, the “materially
indistinguishable” test is but one of two possible ways in which a habeas

498. Id.
499. Id. at 793.
500. The dichotomy between rules and standards is not always so severe. Professor
Korobkin suggests that directives are
better understood as spanning a spectrum rather than as being dichotomous
variables. . . . At a certain point, rules can become so riddled with unpredictable
exceptions that they are as much standard as rule, and standards can become so
determinate that they are as much rule as standard; these composites reside in
the “gray area” at the center of the spectrum. In more extreme cases, standards
can become so determinate that they are transformed into rules, and rules so
unpredictable that they are transformed into standards.
Korobkin, supra note 488, at 30 (footnote omitted).
501. See, e.g., Murillo v. Frank, 316 F. Supp. 2d 744, 754 (E.D. Wis. 2004) (“[I]t is not
enough that there are differences between the two cases. No two cases are ever exactly
alike in all respects. To warrant different conclusions, the differences must be material;
they must be legally significant.”).

822

Catholic University Law Review

[Vol. 54:747

court may find that a state court’s decision was contrary to clearly
502
established law.
The Court’s interpretation of the unreasonable application prong,
however, demonstrates that the concept of clearly established law is
broader than materially indistinguishable cases. The unreasonable
application prong contemplates that clearly established law can
503
reasonably be extended to a new factual context. For a federal court to
grant relief under the unreasonable application prong of § 2254(d)(1),
therefore, clearly established law must have a broader meaning that
includes extending precedent to new factual contexts.
Such an
understanding comports with the normal operation of precedent in our
rule model system, which does not require identical factual circumstances
for precedents to apply.
Section § 2254(d)(1)’s clearly established law requirement, therefore,
encompasses a broader range of precedents than Teague’s “old” rules
504
category.
In Williams, a majority of the Court agreed that “at least”
whatever qualified as an “old rule” under Teague would constitute
505
clearly established law under § 2254(d)(1), rejecting Justice Stevens’s
506
stronger position that Congress had codified Teague.
This choice of
language is significant. Before the Court decided Williams, some of the
lower federal courts were using the language of Teague (“compelled by
existing precedent”) to determine whether a rule was clearly
507
established. The Court’s language in Williams was not as strong as that
502. See Lockyer v. Andrade, 538 U.S. 63, 77-78 (2003) (Souter, J., dissenting).
Because § 2254 also includes the “unreasonable application” prong, clearly established law
must encompass a broader category of precedents.
503. By contrast, under Teague, federal habeas courts are “typically barred from
applying a settled legal standard in a ‘novel setting,’ so that the reasoning undergirding the
rule is ‘extended.’” Yackle, supra note 4, at 414. “If, accordingly, a state court
entertaining a prisoner’s claim would not have felt ‘compelled’ by then-existing precedents
to find his claim meritorious, then a federal habeas court cannot do so without establishing
a ‘new’ rule of law.” Id.
504. See Yackle, Figure, supra note 12, at 1754 & n.131 (arguing that AEDPA’s clearly
established law requirement is broader than “the old rule category” under Teague); Lee,
supra note 12, at 129 (noting that the “argument is surely defensible, on balance, it is not
the better course”); see also Kinports, supra note 79, at 189 & n.378 (pointing out that
every Supreme Court decision could potentially be considered a new rule and arguing that
the definition of new rule is therefore too broad). In this sense, it has also narrowed the
“new” rules category.
505. Williams v. Taylor, 529 U.S. 362, 412 (2000). The Court put “one caveat” on this
connection: the source of “clearly established” law under § 2254 is limited to Supreme
Court precedent only. Id.
506. See id. at 379-80 (opinion of Stevens, J.).
507. See, e.g., Walker v. McCaughtry, 72 F. Supp. 2d 1025, 1031 (E.D. Wis. 1999)
(stating that “[a] rule [is] not clearly established unless it [is] compelled by existing
precedent”); Breedlove v. Moore, 74 F. Supp. 2d 1226, 1231 (S.D. Fl. 1999).

2005]

Seeking Clarity in the Federal Habeas Fog

823

of Teague. Moreover, Justice O’Connor’s “holdings, as opposed to
dictum” test is more flexible. Accordingly, I disagree with those
508
commentators who have argued that AEDPA simply codified Teague.
The Supreme Court’s interpretation of the “contrary to” prong is the
most similar to Teague’s requirements that precedent “compel” or
“dictate” a particular result in a case. The unreasonable application
prong, however, is not as narrow. As a result, clearly established law
under § 2254(d)(1) can include a rule or standard that may not have been
applied to the factual context of the habeas petitioner’s case. A habeas
court should not find that there is no clearly established law in such
509
cases.
Instead, as proposed below, the habeas court should continue
the analysis under the unreasonable application prong.
3. Strength of Precedent
The final touchstone, the strength of the precedent, typically refers to
the reasons that a court might use to justify not being bound by the
510
precedent.
Often we hear the strength analysis described as one of
“weight” of the precedent or the “binding nature” of the precedent.
Because in the AEDPA context courts are limited to applying Supreme
Court precedent on issues of federal law, the binding nature of the
precedent is more evident. This Article submits that the clarity of the
precedent is related to its strength. As with most scope issues, the
strength of the precedent typically should not result in a determination of
no clearly established law. Instead, the strength is relevant to the
reasonableness of the state court’s decision.
For purposes of determining clearly established law, therefore,
strength issues are limited to examining: (1) whether the rule or standard
articulated was approved by a majority of the Supreme Court justices;
and (2) whether the rule or standard has been consistently applied or
otherwise called into question. If only the Court’s holdings constitute
clearly established law, a rule issued in a plurality decision may not
qualify because it is not a holding. Although sometimes this problem is
insurmountable, the narrowest-grounds rule articulated in Marks allows
courts to distill a holding from fragmented opinions. Because the Marks
508. See, e.g., Khandelwal, supra note 12, at 440 & n.45 (arguing that in “us[ing] the
phrase ‘clearly established,’ [Congress] meant to codify the entire Teague doctrine,
including the exceptions”); Note, supra note 12, at 1883-85 (arguing that eliminating the
Teague exceptions might raise constitutional objections); see also Lee, supra note 12, at
119 (noting that Teague and § 2254’s “similarity . . . is striking”).
509. For example, the Fourth Circuit held that “the relevant Supreme Court precedent
need not be directly on point, but must provide a ‘governing legal principle’ and articulate
specific considerations for the lower courts to follow when applying the precedent.”
Quinn v. Hayes, 234 F.3d 837, 844 (4th Cir. 2000); see also supra Part III.B.
510. ALEXANDER, Precedent, supra note 25, at 167.

824

Catholic University Law Review

[Vol. 54:747

511

rule is “settled jurisprudence,” it should be employed whenever
feasible. As Part III of this Article illustrated, federal district courts and
courts of appeals are utilizing the Marks rule to find clearly established
law. This use is appropriate and should be encouraged.
The second strength issue arises where the Court has not consistently
applied a rule or standard in the same manner. This was the type of issue
encountered by the courts in Jackson and Wilson, discussed in Part III of
512
513
this Article. This issue was also raised in Lockyer v. Andrade, where
the Court stressed that it had “not established a clear or consistent path
514
for courts to follow.”
If it appears that the Supreme Court has not consistently applied a
rule, but it has not overruled the rule, then the rule is still precedent and
can qualify as clearly established law for purposes of § 2254(d)(1).
However, the lack of consistency is relevant to the reasonableness of the
state court’s application of that rule. The less consistent the Supreme
Court has been in applying the rule to different factual contexts, the
more flexibility the state court should have in applying the rule.
B. Role of Circuit Court Precedent
As a habeas court applies each of these analytic touchstones, it may
use as evidence whether the federal courts of appeals have agreed or
disagreed about the state of the law. That is, lower court agreement or
confusion about Supreme Court precedent is relevant to the clearly
established law inquiry and should be included as part of the analysis.
While federal circuit court decisions may not themselves serve as clearly
515
established law under AEDPA, they do have a role in providing
evidence of the clarity of the law. Recognizing the importance of federal
circuit court decisions also avoids the constitutional question of whether

511. City of Lakewood v. Plain Dealer Publ’g Co., 486 U.S. 750, 764 n.9 (1988) (writing
for a four-justice majority, Justice Brennan used the rule to support the decision that a
plurality decision from another case was controlling).
512. See supra Part III.C.2.
513. 538 U.S. 63, 72 (2003).
514. Id. at 72 (emphasis added).
515. See, e.g., Garcia v. Cockrell, No. 3:01-CV-2245-D, 2003 U.S. Dist. LEXIS 1874, at
*14 n.8 (N.D. Tex. Feb. 7, 2003) (recognizing that Supreme Court precedent controls the
AEDPA clearly established federal law inquiry but citing circuit court cases in the opinion
“only to the extent that they illuminate [the relevant Supreme Court precedent], not to
expand or extend [it]”), rev’d sub nom. Garcia v. Dretke, 388 F.3d 496 (5th Cir. 2004);
Kennedy v. Lockyer, 379 F.3d 1041, 1048 n.8 (9th Cir. 2004) (recognizing that own circuit’s
precedent is only persuasive authority under AEDPA).

2005]

Seeking Clarity in the Federal Habeas Fog

825

516

AEDPA’s limitation is constitutional and is consistent with the Teague
517
doctrine’s utilization of lower court cases.
Agreement among the lower courts about the rule articulated in a
Supreme Court case may provide evidence of whether a law is
“established” and, if so, how “clearly” established it is. For example, in
Andrade, the Supreme Court failed to acknowledge that most federal
circuits had been following Justice Kennedy’s articulation of the
disproportionality principle in his concurrence in Harmelin v.
518
Michigan.
Agreement among the lower courts should get the
petitioner over the threshold inquiry of clearly established law. And
such agreement is relevant to, but not determinative of, the
reasonableness of the state court’s decision. This is particularly true
where the Supreme Court precedent at issue is a general principle or
standard that may reasonably vary when applied to new factual contexts.
Confusion in the lower courts provides the opposite type of evidence;
it is a symptom of an unclear or conflicted body of law, i.e., law that is
not “clearly” established. As with agreement among the lower courts,
conflicting views of lower courts should not be determinative. They
should not preclude a court from finding the law established enough to
proceed with the analysis under the unreasonable application prong. In
sum, confusion in the lower courts is relevant to the clarity of that law
and its strength, which are factors for the habeas court to consider when
determining the reasonableness of the state court’s decision.
C. Methodology
This Article agrees with the Supreme Court’s pronouncement that the
determination of clearly established law should be a threshold inquiry.
But in doing so this Article defines “threshold” narrowly to simply mean
that the habeas court must be able to identify Supreme Court precedent
that meets the first three touchstones—number of cases, source, and
516. See supra note 40.
517. As Justice O’Connor explained in Williams, lower court opinions may be relevant
to the analysis under Teague:
[W]ith respect to the “reasonable jurist” standard in the Teague context, “[e]ven
though we have characterized the new rule inquiry as whether ‘reasonable
jurists’ could disagree as to whether a result is dictated by precedent, the
standard for determining when a case establishes a new rule is ‘objective,’ and
the mere existence of conflicting authority does not necessarily mean a rule is
new.”
Williams v. Taylor, 529 U.S. 362, 410 (alteration in original) (quoting Wright v. West, 505
U.S. 277, 304 (1992)). Contra Caspari v. Bohlen, 510 U.S. 383, 394-96 (1994) (stating that
conflicting, reasonable views of lower courts precluded the “old rule” status under
Teague).
518. See supra Part II.C.2; Harmelin v. Michigan, 501 U.S. 957 (1991).

826

Catholic University Law Review

[Vol. 54:747

timing—before proceeding with the rest of analysis under § 2254(d)(1).
If the petitioner has failed to cite in support of his petition one Supreme
Court case that was decided by the date of the state court decision, then
denial of relief for lack of clearly established law is warranted because
there is no “established” law.
Most cases, however, will get past this threshold determination. In
those cases, it is the scope and strength of the precedent that determines
the parameters of how “clearly” the law is established. These issues were
raised in Andrade, Alvarado, and many of the lower court cases discussed
in Part III. Although in some cases the court may analyze these
considerations under the “contrary to” prong of § 2254(d)(1), most of the
time these considerations go to the reasonableness of the state court’s
decision. Because § 2254(d)(1) requires federal courts to defer to
reasonable applications of clearly established federal law by state courts,
a lack of clarity in the law increases the reasonableness of the state
court’s application more than if it were applying an unambiguous rule.
Similarly, the amount of discretion necessary to apply a general principle
or standard to a new factual context is a relevant consideration to the
reasonableness of the state court’s decision.
This Article therefore proposes a sliding scale of deference to the state
court based on the complexity of the questions about the scope and
strength of the Supreme Court’s precedent. These will often be
debatable issues; such is the nature of precedent in our legal system.
Although this means that a habeas court is likely to deny relief where the
519
law is unclear, it is consistent with the underlying policies of AEDPA:
fairness, finality, and federalism. If the Supreme Court needs to clarify
520
the law, that is a function of direct review, not habeas review.
V. CONCLUSION
This Article urges a broad reading of AEDPA § 2254(d)(1)’s “clearly
established law” limitation on federal court habeas relief. AEDPA’s
limitation hinges on a concept vital to our common law system:
precedent. While determining what constitutes precedent in the habeas
context is complex, fundamentally it asks the same core questions about
the scope and strength of Supreme Court pronouncements that courts
handle everyday. Federal courts are already well equipped to handle
these questions, but currently do not share a common understanding of
what constitutes clearly established law in the habeas context.
519. See Ides, supra note 11, at 761-65 (noting that where the law is unclear, habeas
courts are much less likely to find that a state court decision was an unreasonable
application).
520. The risk is that the Supreme Court will not clarify the issue on direct review. See
Cotto v. Herbert, 331 F.3d 217, 240 n.11, 251 (2d Cir. 2003).

2005]

Seeking Clarity in the Federal Habeas Fog

827

The need for a common understanding is urgent, not just to ensure the
appropriate application of AEDPA, but because liberty is at stake. This
Article’s proposed analytical framework for determining what constitutes
clearly established law provides guidance in the habeas fog. Although
the clearly established law limitation does serve a gatekeeper function,
most of the time it should not be dispositive. The clearly established law
limitation works in tandem with the standards of review in § 2254(d)(1).
The interconnection of these constraints on federal court review means
that uncertainties about whether the law is “clearly” established are
related to the reasonableness of the state court’s decision. The approach
proposed in this Article minimizes federal intrusions and defers to
reasonable interpretations by the state courts when the law is not a
521
“model of clarity.”
Under AEDPA, “[l]iberty finds no refuge in a
522
jurisprudence of doubt.”

521. Lockyer v. Andrade, 538 U.S. 63, 72 (2003).
522. Planned Parenthood v. Casey, 505 U.S. 833, 844 (1992).