Pao Law Irr

Berkeley Law
Berkeley Law Scholarship Repository

Faculty Scholarship
1-1-1989
Statutory Interpretation and Legislative Supremacy

Daniel A. Farber
Berkeley Law
Follow this and additional works at: http://scholarship.law.berkeley.edu/facpubs

Part of the Law Commons
Recommended Citation
Daniel A. Farber, Statutory Interpretation and Legislative Supremacy, 78 Geo. L.J. 281 (1989)
This Article is brought to you for free and open access by Berkeley Law Scholarship Repository. It has been accepted for inclusion in Faculty
Scholarship by an authorized administrator of Berkeley Law Scholarship Repository. For more information, please contact [email protected].
Statutory Interpretation and
Legislative Supremacy
DANIEL A. FARBER*
"The general contemporary American view of statutory interpretation is

that there is not a great deal to say about the subject. As a result, nothing
else as important in the law receives so little attention.",
In the six years since that statement was written, there has been something
of a renaissance of scholarship about statutory interpretation. 2 Nevertheless,
our understanding of statutory interpretation remains quite limited. The tre-
mendous variety of statutes, as well as the broad range of contexts in which
they must be interpreted, suggests that a unified theory of statutory interpre-
tation may be unattainable. Still, theory can at least illuminate some of the
recurring problems of statutory interpretation.
This article is an attempt to clarify one particular interpretation question:
to what extent are judges constrained by statutory language and legislative
intent? It is a commonplace that, apart from constitutional issues, judges are
subordinate to legislatures in the making of public policy. 3 If this
* Henry J. Fletcher Professor of Law, University of Minnesota. Doug Baird, Carol Chomsky,
Bill Eskridge, Phil Frickey, Walter Gellhorn, Hank Greely, Roger Park, Richard Posner, and Fred
Schauer contributed valuable comments on earlier drafts, as did participants in faculty workshops
at Columbia University and the University of Minnesota.
1. Weisberg, The CalabresianJudicialArtist: Statutesand the New Legal Process, 35 STAN. L.
REV. 213, 213 (1983); see also Froomkin, Climbing the Most DangerousBranch:Legisprudence and
the New Legal Process (Book Review), 66 TEX. L. REv. 1071, 1071 (1988) (reviewing W. ESKRIDGE
& P. FRICKEY, CASES AND MATERIALS ON LEGISLATION: STATUTES AND THE CREATION OF
PUBLIC POLICY (1988)) ("Legislatures, and the statutes they produce, are the poor cousins of legal
education.").
2. Much of the "new learning" is collected in W. ESKRIDGE & P. FRICKEY, CASES AND MATER-
IALS ON LEGISLATION: STATUTES AND THE CREATION OF PUBLIC POLICY ch. 7 (1988). For a
thoughtful survey of the scholarly literature, see Eskridge & Frickey, Legislation Scholarship and
Pedagogy in the Post-LegalProcess Era, 48 U. Prrr. L. REv. 691 (1987).
3. See, e.g., California v. Sierra Club, 451 U.S. 287, 297-98 (1981) (refusing to imply a private
cause of action in the face of contrary statutory language and legislative history); S. BURTON, AN
INTRODUCTION TO LAW AND LEGAL REASONING 41-42 (1985) (judiciary subordinate to legisla-
ture in statutory application because judges are less accountable to the electorate); R. POSNER, LAW
AND LITERATURE: A MISUNDERSTOOD RELATION 240, 252-53 (1988) (subordinate judiciary must
ferret out the intent of the legislature in statutory interpretation); Aleinikoff, Updating Statutory
Interpretation,87 MICH. L. REV. 20, 22 (1988) (describing as "archaeological" statutory interpreta-
tion seeks to fulfill the intent of the drafting legislature); Langevoort, Statutory Obsolescence andthe
JudicialProcess: The Revisionist Role of the Courts in Banking Regulation, 85 MICH. L. REV. 672,
672 n.1 (1987) (citing both Easterbrook and Posner as proponents of the dominant jurisprudential
tradition that statutory obsolescence is a matter of concern for the legislature only, not their faithful
servants, the courts); Maltz, StatutoryInterpretationandLegislative Power: The Casefor a Modified
Intentionalist Approach, 63 TUL. L. REv. 1, 9 (1988) (doctrine of legislative supremacy rests on
HeinOnline -- 78 Geo. L. J. 281 1989-1990

THE GEORGETOWN LAW JOURNAL [Vol. 78:281
subordinate role means anything at all, it must somehow constrain judges

who interpret statutes from implementing their own notions of public policy.
Although this much is clear, the extent of the constraint is far from obvious.
At the one extreme, judges may be free to implement virtually any policies
they want; at'the other, they may be forbidden even to consider their own
views of public policy in deciding statutory issues.
Part I of this article attempts to pinpoint this constraint on judicial
policymaking. It argues that subordinating courts to the legislature in all
nonconstitutional areas of public policymaking (the legislative supremacy
principle) precludes judicial policymaking only when a statutory directive 4 is
clear. 5 Even then, this "weak conception" of the supremacy principle may
allow courts to go beyond (but not against) clear legislative intent. Thus, the
supremacy principle leaves courts considerable, although subordinate,
6
policymaking authority.
Part II demonstrates that the supremacy principle, although limited in
scope, nevertheless significantly constrains judicial decisions. On occasion,
the legislative supremacy principle has led judges to embrace results obvi-
ously contrary to their own views of public policy. On other, less creditable
occasions, courts have failed to live up to the demands of the supremacy
principle.
Part III considers some possible qualifications to the principle of legislative
supremacy. Under some circumstances, rational legislators would not want
judges to adhere strictly to their statutory mandates. For example, blind
adherence to a statute may be contrary to the drafters' intent because of an
unforeseeable development or because a mistaken judicial interpretation has
developed independent significance as precedent. These types of post-enact-
deeply embedded premise of the American political system that legislature may prescribe rules of
law that bind all other governmental actors within constitutional limits); Redish, Abstention, Sepa-
ration of Powers, and the Limits of the JudicialFunction, 94 YALE L.. 71, 76 (1984) (judiciary may
openly ignore a legislative judgment only on the grounds of unconstitutionality).
As Professor Aleinikoff points out, most legislative interpretation is actually done by administra-
tive agencies rather than courts. Aleinikoff, supra, at 42-43. Chevron, Inc. v. NRDC, 467 U.S. 837,
842-43 (1984), suggests a view of statutory interpretation by agencies that is strikingly similar to
that espoused in this article regarding judges.
4. A statute may be purely descriptive or declaratory of the common law, rather than being a
command. For example, Congress has enacted recommendations about the proper ceremonial
treatment of the American flag. See Texas v. Johnson, 109 S. Ct. 2533, 2547 (1989); see also MINN.
STAT. § 1.135 subdiv. 5 (Supp. 1989) (explaining the symbolism of the state seal). Such a statute,
which does not purport to implement a legislative mandate, is outside the scope of the supremacy
principle. Also, this article is only concerned with cases in which a statute is considered to be
directly applicable, rather than with the use of statutes by analogy in common law cases.
5. The idea of statutory clarity is itself somewhat unclear, in part because it could refer to either
the text or legislative history of the statute. One of the tasks in Part I will be to clarify this concept.
See infra text accompanying notes 35-53.
6. There may well exist other constraints on judicial policymaking, but these do not derive from
the court's subordination to the legislature.

1989] LEGISLATIVE SUPREMACY
ment events may justify the court in disobeying a clear statutory directive,
because such disobedience is actually consistent with the legislature's pre-
sumed "meta-intent." In contrast, I argue that legislators generally do not
want to condition the continued effectiveness of a statute on the state of pub-
lic opinion. Consequently, even radical changes in public opinion do not, in
7
themselves, justify deviations from statutory directives.
The institutional structures and normative commitments needed to sustain
democratic government must include judicial respect for legislative decisions.
It should come as no shock, however, that the proper judicial role in a
democracy is not readily reducible to a formula. The supremacy principle
defines only part of the judicial role in statutory cases. Other norms, such as
legal stability and fairness to individual litigants, also play legitimate roles in
judicial decisions. Although the principle of legislative supremacy is but one
piece of the interpretation puzzle, finding that principle can provide a point
of reference as we struggle with the rest of the puzzle.
I. THE SCOPE OF THE SUPREMACY PRINCIPLE
Legislative supremacy, as a doctrine of statutory interpretation, is

grounded in the notion that, except when exercising the power of judicial
review, courts are subordinate to legislatures.8 This is a fairly commonplace
notion, but it turns out to be surprisingly difficult to pin down the precise
nature of that subordinate relationship. What exactly do we mean when we
say that courts are subordinate to legislatures in making public policy?
In analyzing this issue, we must be careful to distinguish between legisla-
tive supremacy and legislative exclusivity in the making of public policy. In
general, supremacy and exclusivity are quite separate concepts. 9 Legislatures
are typically the supreme, but not exclusive, lawmakers. 10 State courts have
the authority to create common law doctrines embodying their own views of
public policy, subject to legislative modification. Although the federal courts
have narrower common law powers, they also have this power in areas of
particular federal concern. 11 Given the existence of a legitimate judicial role
7. The supremacy principle does not, however, prohibit courts from giving effect to the current
state of public opinion when construing ambiguous statutes. See infra notes 19-21 and accompany-
ing text.
8. See supra note 3.
9. See Westen & Lehman, Is There Life for Erie After the Death of Diversity?, 78 MICH. L. REV.
322, 326, 331 (1980) (distinguishing between areas in which Congress' power is exclusive and areas
in which other branches have concurrent legislative power but Congress' authority is final). For
example, Congress has supreme lawmaking power over interstate commerce, but its authority is not
exclusive. States are free to implement their own policies, as long as they do not thwart congres-
sional goals. Id. at 322-24.
10. See id. at 325-26 (the Constitution gives Congress final authority, but not exclusive power, to
establish the majority of legal norms).
11. See Eskridge, DynamicStatutory Interpretation,135 U. PA. L. REV. 1479, 1499-1501 (1987).

in policymaking, the question becomes whether, and to what extent, that role
extends to statutory cases.
A. THE STRONG CONCEPTION OF SUPREMACY
Perhaps the most obvious understanding of legislative supremacy is that

courts must follow legislative directives: "Judges must be honest agents of
12
the political branches. They carry out decisions they do not make."
Hence, the "judges' role is to decipher and enforce" the statute.13 This is a
"strong," or "formalist," conception of legislative supremacy. The strong
conception views the supremacy principle as an all-encompassing prescrip-
tion of the judicial role in statutory cases rather than as a mere constraint on
judicial behavior. This conception does not merely rule out judicial views on
controversial matters of public policy. It also precludes judges from consid-
ering relatively noncontroversial policies like stare decisis, judicial adminis-
tration, or other "legal process" values, unless the enacting legislature has
endorsed those values. 14
The fundamental flaw in the strong conception is that it oversimplifies the
agency model. The judicial role in statutory cases is indeed something like
that of an agent of the legislature, and this analogy can be quite illuminating.
But the analogy must be understood in its full complexity. First, agents do
not simply execute orders that their principals give. Rather, agents act on
their principals' behalf, carrying out orders that may be subject to multiple
interpretations in light of their understanding of the principals' overarching
goals. Second, even in statutory cases, federal judges are not the agents of
Congress. Their employer is not the Congress but the United States, and
their ultimate allegiance is to the Constitution rather than to the House and
Senate. Hence, it is the courts' role to carry out congressional directives in
light of their understanding of the Constitution.
In determining the appropriate methods of statutory interpretation, then,
it is not enough to consider only how best to implement the purposes of
Congress. We must also determine what approach is most consistent with
our scheme of constitutional government.1 5 Many presumptions and clear
A notable example is the Court's recent creation of a federal common law rule limiting the liability
of government suppliers for defective products. See Boyle v. United Technologies Corp., 108 S. Ct.
2510 (1988).
12. Easterbrook, Foreword,The Court and the Economic System, 98 HARV. L. REV. 4, 60 (1984).
13. United States v. 9/1 KG. Containers, 854 F.2d 173, 179 (7th Cir. 1988) (Easterbrook, J.),
cert. denied, 109 S. Ct. 118 (1989).
14. Some objections to this view of the judicial role are sketched in Farber, StatutoryInterpreta-
tion, Legislative Inaction, and Civil Rights, 87 MICH. L. REV. 2, 6-7, 14-19 (1988) (reviewing argu-
ments against originalism but in favor of judges' reliance both on personal values in "hard cases"
and on consideration of public interest in stare decisis analysis).
15. See, eg., Eskridge, supra note 11, at 1523-38 (cautious dynamic statutory interpretation is

statement rules are designed to promote constitutionally rooted goals like fair
notice to criminal defendants, federalism, and fairness to "discrete and insu-
lar minorities." 16 These rules are faithful to the constitutional scheme, and
their use in statutory interpretation serves to advance by means of interstitial
policymaking the constitutional goal of democratic, just, and effective
government.
Judges have been compared to military officers attempting to obey unclear
orders from headquarters. 17 This military analogy suggests some of the diffi-
culties with the strong conception of supremacy. Within the confines of his
orders, even a subordinate officer is expected to exercise his discretion in
accordance with sound military judgment. In the absence of orders, the of-
ficer is not expected to sit on his hands regardless of circumstances.", Again,
the agency model can be misleading if oversimplified. The captain must fol-
low the general's orders, but to do so the captain may have to interpret an
occasionally unclear order in light of his primary allegiance to the United
States Army. Similarly, legislative supremacy does not proscribe statutory
interpretation that may result in the elimination of judicial policymaking.
This is not to say, of course, that statutory interpretation is an exercise in raw
policymaking, but only that on occasion, the judiciary may properly consider
nonlegislative sources of public policy.
A number of well-accepted judicial practices, such as the use of "clear
statement" rules and other canons of construction, allow judges to include
policy considerations in their statutory decisions. 19 For example, statutory
not significantly countermajoritarian and is therefore consistent with the constitutional scheme of
government).
16. See W. ESKRIDGE & P. FRICKEY, supra note 2, at 655-59 (discussing the application of sub-
stantive policy presumptions in statutory interpretation). McNally v. United States, 483 U.S. 350
(1987), is a recent notable example. In McNally the Court read the mail fraud statute narrowly
because of concerns about federalism and fair notice. Id. at 359-60. Both of these concerns have
obvious constitutional roots. Courts may also find sources of policy in statutes that do not directly
apply to the case at hand. See Gellhorn, Contractsand PublicPolicy, 35 COLUM. L. REv.679, 690-
93 (1935) (advocating the application of statutes by analogy in complex cases that, although they do
not fall squarely within the statute, are related to the subject matter of the statute). By seeking
guidance in statutory and constitutional provisions, courts can blunt the charge that they have
assumed unbridled legislative authority.
17. Judge Posner, who originated the military analogy, explains that "[i]n our system of govern-
ment the framers of statutes and constitutions are the superiors of the judges. The framers commu-
nicate orders to the judges through legislative texts.... If the orders are clear, the judges must obey
them." Posner, Legal Formalism, Legal Realism and the Interpretation of Statutes and the Consti-
tution, 37 CASE W. REs. 179, 189 (1986-87).
18. See id. at 189-90 (military officers who receive orders by radio that begin "go-," but the rest
of the message is unclear, should use their own judgment in going forward rather than remaining
stationary until complete orders arrive, because it is clear that the commanding officer intended
some action).
19. For an extensive discussion of these judicial practices, see W. ESKRIDGE & P. FRICKEY,
supra note 2, at 655-95.

286 THE GEORGETOWN LAW JOURNAL [Vol. 78:281
directives are often extremely unclear, and one does not have to be a legal
realist to know that the resulting judicial interpretation often arises out of the
court's view of public policy as much as any attempt to fathom the statutory
text. 20 Not surprisingly, judicial opinions that "apply" open-ended or vague
21
statutes often read like common law decisions.
This judicial policymaking is not inherently objectionable. After all, if a
state court can legitimately make public policy when the legislature has said
nothing at all about a subject, 22 the same court should also be able to make
policy when the legislature has spoken, but spoken so unclearly that the
court cannot confidently decipher its directions. In such a case, the fact that
the legislature has spoken makes it clear that the legislature intends for some-
thing to be done about the particular issue. Filling such gaps, "even if this
calls for creation as well as re-creation, is an essential part of government,
lest statutes become brittle and fail of their essential purposes. '23 Both de-
scriptively and normatively, then, the line between common law policymak-
ing and statutory construction is not so sharp as it might seem. 24
Despite well-accepted justifications for judges to engage in interstitial
policymaking, a supporter of the strong conception might question its legiti-
macy. 2 5 If the court is not simply implementing a policy judgment by the
legislature, where does it derive its authority to act? This question may seem
particularly pressing with respect to the federal courts, which have only lim-
ited common law powers. 26 Once a court is given jurisdiction to decide cases
arising under a statute, however, by implication the court is empowered to
use normal methods of judicial decision. 27 The nature of these judicial meth-
20. See Pierce, Chevron and Its Aftermath: JudicialReview of Agency Interpretationsof Statu-
tory Provisions,41 VAND. L. REV. 301, 305 (1988) (statutory ambiguity results when the legislature
declines to make certain policy decisions; hence, the courts' "interpretation" of an ambiguous stat-
ute necessitates judicial policymaking). In the context of statutory interpretation by agencies rather
than courts, an argument that policymaking is illegitimate would be even more tenuous. See Chev-
ron, 467 U.S. at 843-44 (Congress intends for administrative agencies to fill explicit statutory gaps).
21. See, eg., Patrick v. Buhget, 486 U.S. 94 (1988) (construing Sherman Antitrust Act); Aspen
Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985) (same); United States v. Basye,
410 U.S. 441 (1973) (discussing "assignment of income" doctrine in income tax cases); Helvering v.
Horst, 311 U.S. 112 (1940) (same).
22. See supra text accompanying notes 9-11.
23. Easterbrook, Statutes' Domain, 50 U. CHI. L. REV. 533, 551 (1983).
24. See Merrill, The Common Law Powers of Federal Courts, 52 U. CHI. L. REV. 1, 3-7 (1985);
Westen & Lehman, supra note 9, at 33 1.
25. The strong conception seems to rest largely on concerns about legitimacy. See Merrill, supra
note 24, at 3, 7-12 (analyzing the legitimacy of federal common law in light of concerns for federal-
ism, separation of powers, and political accountability); see also Froomkin, supra note 1, at 1091-94
(allowing courts to become involved in policymaking raises grave separation of powers issues).
26. See City of Milwaukee v. Illinois, 451 U.S. 304, 312-13 (1981) (federal courts do not have
general power to develop their own rules of decision, except when Congress has not spoken on a
particular issue).
27. For this reason, it does not seem necessary to search within a particular statute for specific

ods is partly a matter of convention and history and partly a matter of the
judiciary's current sense of its role in the overall scheme of government. Our
legal traditions do not countenance freewheeling judicial activism in statu-
tory cases. They do, however, contemplate the use of a variety of interpreta-
tive techniques that allow interstitial policymaking. 2 8
The essence of the strong conception is that judges are "just following or-
ders" without regard to their own views of public policy. When directives
are clear, this is a sound justification. But when the directives are reasonably
open to conflicting interpretations (as is often the case), we should be reluc-
tant to excuse judges from responsibility for their own actions on the ground
that they were just implementing directives from above. Judges should not
be encouraged to duck the responsibility for their choices. Since Holmes, it
has been a commonplace that judges engage in interstitial policymaking; 29
the rhetoric of legislative supremacy should not be abused as a means of
concealing the true judicial role. 30
B. THE WEAK CONCEPTION OF SUPREMACY
Under the weak conception of legislative supremacy, a judge may not con-
travene statutory directives. 31 Thus, unlike the strong conception, the weak
conception views supremacy as a constraint on judicial action, rather than as
a complete specification of the judicial role in statutory cases.
The military analogy is helpful in clarifying the nature of this constraint.
We cannot readily specify just what it means to tell an officer to follow or-
ders, because to do so would require a complete philosophical account of
interpretation. Inverting the prescription may be more useful: whatever else
an officer may do, he cannot disobey lawful orders.
One advantage of the weak conception is that it demands relatively little
consensus about what sources of public policy, if any, a judge may bring to
evidence of a congressional delegation of authority to the courts, as some writers have suggested.
See P. BATOR, P. MISHKIN, D. SHAPIRO & H. WECHSLER, HART AND WECHSLER'S THE FED-
ERAL COURTS AND THE FEDERAL SYSTEM 863 (3d Ed. 1988) (congressional delegation of lawmak-
ing powerto the federal courts must be either express or implied within a statute); Merrill, supra
note 24, at 41-44 (same).
28. See Blatt, The History of Statutory Interpretation: 4 Study in Form and Substance, 6 CAR-
DOZO L. REV. 799, 800-05, 813-15, 833-34 (1985) (discussing the role of equity in statutory analysis
throughout history); Eskridge, supra note 11, at 1502-03 (discussing Blackstone's theory that
"judges should enforce the textual commands of statutes, though not to the detriment of the stat-
ute's overall purposes and the current demands of equity").
29. See, e.g., Southern Pac. Co. v. Jensen, 244 U.S. 205, 221 (1917) (Holmes, J., dissenting)
(recognizing that judges must legislate, but that they can only do so interstitially).
30. See Posner, Statutory Interpretation-Inthe Classroom and in the Courtroom, 50 U. CHI. L.
REV. 800, 816-17 (1983) (courts' pretense that statutory interpretation merely involves application
of canons of construction conceals extent to which courts create new law).
31. This article presumes that all statutes are constitutional.

bear in interpreting statutes. It does not assume that statutes are determinate
in the sense of having a single correct interpretation. It assumes only that
methods of interpretation are not completely "up for grabs," and that a stat-
ute's language and legislative history together preclude at least some inter-
pretations. Although the requirement that the President be at least thirty-
five years old 32 seems perfectly clear to most people, 33 there are some who
argue that it is potentially indeterminate. 34 Even they do not argue, how-
ever, that courts could interpret the provision to allow a toddler to become
President. The weak supremacy principle requires only this low degree of
determinacy.
Because the weak conception analyzes supremacy by analogy to the con-
cept of disobedience, it remains only to define what it means to disobey an
order.35 Again, what seems initially obvious turns out to be unexpectedly
elusive. If obedience may be construed as complying with the intent of a
superior, then disobedience may be defined as a failure to comply with the
intent of a superior. An officer, however, is not obliged to carry out the
unexpressed wishes of superiors, even if he can somehow divine them. Nor
can the intentions of legislators have any binding effect unless they are ex-
pressed in a law. 36 The Constitution contains procedural barriers intended to
hinder the transformation of legislators' preferences into law. 37 Hence, to
give legal effect to legislative intentions in the absence of any relevant statu-
tory text would undermine the constitutional scheme. 38
Disobedience, therefore, must relate to a text rather than merely to an
32. U.S. CONST. art. II, § 1.

33. See Schauer, Easy Cases, 58 S. CAL. L. REv. 399, 414 (1985) (noting obvious legal result of
29-year-old presidential candidate).
34. See Tushnet, A Note on the Revival of Textualism in Constitutional Theory, 58 S. CAL. L.
REV. 683, 686-88 (1985) (interpreting age requirement as "shorthand" expression of Framers' poli-
cies, and arguing that the requirement could be changed without impairing the Framers' scheme of
constitutional government). But see R. POSNER, supra note 3, at 219 (criticizing Tushnet's
argument).
35. This analogy, it must be realized, is only that. Statutes do not, in general, take the form of
orders to judges. See generally H.L.A. HART, THE CONCEPT OF LAW 18-26 (1961) (discussing
general forms that law may take to communicate orders from government to society).
36. Thus, it is a mistake to see the statutory text as "a kind of second-best," as if "we would
prefer to take the top off the heads of authors and framers-like soft-boiled eggs-to look inside for
the truest account of their brain states at the moment that the texts were created." Fried, Sonnet
LXV and the "Black Ink" of the Framers'Intention,100 HARV. L. REv. 751, 758-59 (1987); see
also Tribe, JudicialInterpretationsof Statutes: Three Axioms, 11 HARV. J.L. & PUB. POL'Y 51, 51
(1988) ("our search in statutory interpretation.., must be not for a subjective, unenacted intent but
for an objective, enacted meaning of a legal text").
37. See U.S. CONST. art. I, §§ 1 (bicameralism requirement), 7 (presentment requirement); see
generally INS v. Chadha, 462 U.S. 919 (1983) (bicameralism and presentment "serve essential con-
stitutional functions").
38. See Easterbrook, The Role of Original Intent in Statutory Construction, 11 HARV. J.L. &
PUB. PoL'Y 59, 64-65 (1988).

1989] LEGISLATIVE SUPREMACY 289
unexpressed intention. The requisite relationship with the text is less clear.
Perhaps one naive view would be that disobeying an order means violating its
literal terms. 3 9 After all, it is said, "[t]he words of the statute, and not the
intent of the drafters, are the 'law.' "4 This sanctification of the statutory
text, however, goes too far. If the directive contains a typographical error,
correcting the error can hardly be considered disobedience.4 1 Moreover, lit-
eral application of a directive might be senseless or contrary to its obvious
purpose. For example, virtually no one doubts the correctness of the ancient
decision that a statute prohibiting "letting blood in the streets" did not ban
emergency surgery.4 2 In these examples, extrinsic evidence about the
speaker's intentions is not needed to tell the ordinary legal reader that some
literal applications of a statute are unintended. Rejecting the literal applica-
tion in these situations can hardly be considered disobedience.43 Today,
some courts go even farther and refuse to follow the "plain meaning" of a
statute not because its literal meaning is untenable on its face, but because its
39. Recently, literalism has been revived as an approach to statutory interpretation, with Justice
Scalia as one of its leading advocates. For recent statements of Scalia's view, see Blanchard v.
Bergeron, 109 S. Ct. 939, 946-47 (1989) (Scalia, J., concurring in part); Green v. Bock Laundry
Machine Co., 109 S. Ct. 1981, 1994-95 (1989) (Scalia, J., concurring in the judgment); see also
Breger, IntroductoryRemarks, 1987 DUKE L.J. 362, 367-69 (citing additional Scalia opinions). At
least one commentator discerned a trend toward literalism even before Justice Scalia's appointment.
See Comment, Intent, Clear Statements, and the Common Law: Statutory Interpretation in the
Supreme Court, 95 HARV. L. REV. 892, 894-98 (1982).
40. Easterbrook, supra note 38, at 60.
41. For statements of this principle, see Shine v. Shine, 802 F.2d 583, 588 (Ist Cir. 1986); H.
HART & A. SACKS, THE LEGAL PROCESS: BASIC PROBLEMS IN THE MAKING AND APPLICATION
OF LAW 1412 (tent. ed. 1958).
42. See K Mart Corp. v. Cartier, Inc., 108 S.Ct. 1811, 1834 n.2 (1988) (Scalia, J., concurring in
part). The most commonly cited authority for rejecting literal meaning when applying it would
lead to an "absurd" result is Rector of Holy Trinity Church v. United States, 143 U.S. 457 (1892).
For a recent application of this principle, see Public Citizen v. United States Dep't of Justice, 109 S.
Ct. 2558, 2565-67 (1989).
43. Judge Easterbrook suggests that the proper approach to interpretation is to ask how the
"objectively reasonable person" would understand the statute. Easterbrook, supra note 38, at 65.
But this begs the question. Does the objectively reasonable person know about the political context
of the statute, or the historical period in which it was passed, or the practical problems in imple-
menting it? If so, this method of interpretation differs little from that currently used. But, if the
"objectively reasonable person" is intended to be the "Person from Pluto" (the gender-neutral suc-
cessor to the famous Man from Mars), who knows nothing about American society and history, we
must ask why judges should adopt such a peculiar persona. The vantage point more properly
should be that of the statute's actual audience-typically, lawyers applying the conventional ap-
proach to statutory interpretation, which currently includes considering legislative history. A more
fundamental flaw in Easterbrook's literalist approach, however, is its assumption that statutes have
meaning before they are interpreted. Given such an assumption, literalism appears to be the only
possible legitimate form of statutory interpretation, rather than one among several. For analysis of
a similar problem with the method of interpretation proposed by Professors Hart and Sacks, see
Fuller, Positivism and Fidelity to Law-A Reply to ProfessorHart,71 HARV. L. REV. 630, 661-69
(1958).

legislative history shows that the legislature had contrary intentions. 44

This may or may not be a wise method for interpreting statutes, but it does
not violate the supremacy principle. The concept of following instructions
does not necessarily entail literalism. Indeed, wooden literalism, at the ex-
pense of the speaker's known purposes, is often used as a subtle form of diso-
bedience, Whether by obstructive bureaucrats or rebellious workers. A
faithful agent will not give an order an entirely context-free or literal inter-
pretation. 4 5 Similarly, the legislature could hardly complain that the courts
have been insubordinate when judges have stretched the literal meaning of a
46
statute to implement the legislative intent.
The idea of legislative intent, however, is notoriously slippery.4 7 If it is
taken to require that a majority of the legislators share the same subjective
view of the statute, the condition will rarely be met. Most legislators do not
have time actually to read and come to an independent understanding of the
statutes on which they vote. Rather, legislators depend on institutional ac-
tors (sponsors, committees, floor leaders, and staffers), who are charged with
drafting statutes and moving them to enactment, to explain the meaning and
import of the statutes under consideration. 48 Legislators normally-quite le-
gitimately-accept the statements of these actors as commitments about the
meaning of the enactments. 4 9
44. For judicial statements rejecting the plain meaning approach, see INS v. Cardoza Fonseca,
480 U.S. 421,432 n.12 (1987); United States v. American Trucking Ass'n, 310 U.S. 534, 543 (1940);
Escobar Ruiz v. INS, 838 F.2d 1020, 1023 (9th Cir. 1988); see also Wald, Some Observationson the
Use of Legislative History in the 1981 Supreme Court Term, 68 IowA L. REv. 195, 197-99 (1982);
see generally Aleinikoff, supra note 3, at 31-32 (literalism is "no favor to the legislature" when it
exalts plain meaning over legislative intent); Langevoort, supra note 3, at 729-30 (developing similar
arguments against literalism).
45. For a general critique of strict literalism, see Brest, The Misconceived Questfor the Original
Understanding,60 B.U.L. REV. 204, 205-09 (1980).
46. As Professor Merrill has stated: "If the legislative power belongs primarily to Congress, then
it should not be regarded as a usurpation of that power for courts to interpret legislation to reach
results that were specifically intended by Congress." Merrill, supra note 24, at 23. This statement is
especially noteworthy because Merrill takes a very narrow view of the permissible scope of
policymaking by the federal courts.
47. For a summary of the recent debate about the existence and significance of legislative intent,
see Farber & Frickey, Legislative Intent and Public Choice, 74 VA. L. REV. 423, 425-37, 453-61
(1988).
48. See W. OLESZEK, CONGRESSIONAL PROCEDURES AND THE POLICY PROCESS 94-95 (2d ed.
1984) (legislators generally defer to committee decisions); R. POSNER, THE FEDERAL COURTS:
CRISIS AND REFORM 269-70 (1985) (arguing that committee reports and statements of sponsors are
valid indicia of legislative intent, but statements of witnesses at hearings and of nonsponsoring
members are not); Maltz, supra note 3, at 26-27 (legislators rely on committee reports and state-
ments of floor managers to form their positions on bills).
49. Thus, these actors have an incentive to make their statements accurate. If the meaning of a
statute is not consistent with the committee report or the sponsor's statement on which legislators
have relied, the legislators will not likely cooperate in future endeavors with the committee or the
sponsors. See R. POSNER, supra note 48, at 269.

Even assuming that the legislature possesses an ascertainable institutional

understanding of a statute, one could ask why judges should be barred from
adopting a contrary interpretation if the statutory language allows it. This
question can be answered at two levels.
First, ignoring the legislature's understanding of statutes burdens the pro-
cess of enactment with additional uncertainties. This, in turn, makes the pro-
cess even more cumbersome and difficult.50 A recent example can be found.
in the Senate debates over an arms control treaty. Approval of the treaty was
delayed because the Reagan Administration refused to abide by a prior ad-
ministration's commitments to the Senate regarding the meaning of a previ-
ous treaty, on the basis of which the Senate had approved that previous
treaty.5 1 If the legislature cannot trust the implementing authorities to re-
spect its collective understanding, it may be reluctant to legislate at all; and if
it does legislate, it will be forced to make substantial efforts to ensure that its
intended meaning cannot be willfully ignored. An interpretative rule that
ignores legislative intent will impose undue burdens on the legislative pro-
cess, hindering the ability of the democratic branches to function effectively.
The second reason for respecting the legislature's collective understanding
is more fundamental. The legislature's directives are entitled to the force of
law because of their origins: "Statutory authority derives from its political
source ... .- "52 Once a statute is enacted, its meaning may not be limited to
the intentions of its authors because the text has acquired a legal status of its
own. But to construe the text squarely contrary to the collective understand-
53
ing of the legislature weakens its claim to legitimacy.
Thus, both the text and other evidence of the legislative intent seem rele-
vant to determining judicial disobedience. Neither clear evidence of intent
nor clear language, however, is alone sufficient to render contrary actions
disobedient. There sometimes may be doubts about whether the intent was
ever enacted or about whether the clear language should be taken literally.
Thus, it is hard to state a mechanical rule to determine disobedience. The
ultimate question is whether genuine doubt exists about the meaning of the
legislative command. If a judge does have genuine doubts about the legisla-
50. For a discussion of some of these increased transaction costs, see Farber & Frickey, supra
note 47, at 458-59.
51. See Chayes & Chayes, Testing andDevelopment of "Exotic" Systems Under the ABM Treaty:
The Great Reinterpretation Caper, 99 HARv. L. REV. 1956, 1963-68 (1986); Gordon, Within the
Arms Debate, a 2d Debate, N.Y. Times, Feb. 9, 1988, at A3, col. 1.
52. Weisberg, supra note 1, at 226. A literary work derives its authority from its own text. The
identity of the author is of interest only because the text itself is significant. We care about Shake-
speare because he wrote Hamlet, rather than vice versa. Commands are different. A general's
orders are entitled to obedience because their source is the general, not because of their merits as
texts. In this respect, statutes are more like orders than they are like literature.
53. See R. POSNER, supra note 3, at 229, 242-44, 256.

tion's meaning, there can be no question of obedience or disobedience.5 4
C. DEMOCRACY AND DISOBEDIENCE
The foregoing analysis suggests this formulation of the supremacy princi-

ple: When statutory language and legislative intent are unambiguous, courts
may not take action to the contrary. In other words, when legislation clearly
embodies a collective legislative understanding, the court must give way,
even if its own view of public policy is quite different.
This is not to suggest that the court should ignore statutory language or
evidence of legislative intent in other situations. The same considerations
that make language and intent binding when they are clear entitle them to
judicial attention when they are unclear. When military orders are unclear,
the officer must exercise his own judgment, but this should not mean that he
ignores whatever guidance can be gleaned from the orders themselves. He
simply may consider other factors as well.
Using this modest interpretation, the supremacy principle may seem obvi-
ous and noncontroversial. It is, nevertheless, worth pausing to inquire about
its foundation. There are two reasons why courts should follow the
supremacy principle.
The first is that the principle is a basic part of the traditional judicial meth-
odology in our legal culture. We could conceive of a society in which courts
were paramount, and legislatures merely proposed possible changes in the
common law for judicial consideration. But that is not the world in which
we live. Because the supremacy principle is fundamental to our institutional
framework, violations of the principle defeat justified expectations and im-
pair legal stability.
Second, in breaching the supremacy principle, a court is declining to im-
plement clear decisions by the democratic branches of government. In our
legal system, unlike some others, this is permissible if the court is willing to
declare the statute unconstitutional. If the court actually believes the statute
is unconstitutional, adopting a completely implausible "interpretation" to
54. This assumes, of course, that it is possible for a text to communicate a clear meaning. In the
context of legislative supremacy analysis, this assumption does not seem to be especially problem-
atic, particularly because the contrary "indeterminacy" thesis is vulnerable to attack on several
grounds. See generally Solum, On the Indeterminacy Crisis: Critiquing CriticalDogma, 54 U. CH.
L. REv. 462 (1987). As often stated, the indeterminacy thesis contends that "a text lacks a single,
objectively determinable meaning. A text's only meaning is the one given by an interpreter, who in
turn always reads the text against a particular social and political backdrop." Carter, Constitutional
Adjudication and the Indeterminate Text: A PreliminaryDefense of an Imperfect Muddle, 94 YALE
L.J. 821, 823 (1985). Congress can reasonably anticipate that the judges and officials applying
statutes will be somewhere within the political mainstream. If the meaning of a text is clear to
mainstream judges, then that must be the meaning Congress desired. After all, Congress is also
(perhaps by definition) rather firmly within the political mainstream.

reach the same result is intellectually dishonest and a breach of the duty of
judicial candor.5 5 On the other hand, if the court does not think the statute
is unconstitutional, this necessarily means that under the court's own view of
the Constitution, the legislature is entitled to have its way. In that event,
refusing to implement the statutory command is simply lawless.
Violations of the supremacy principle are particularly serious because they
impair the basic social norm of democratic self-government. When a court
refuses to follow an admittedly constitutional statute, it arrogates the ulti-
mate power to make public policy. 5 6 When the court concedes that a statute
is within the legislature's prerogatives under the existing scheme of constitu-
tional government, refusal to implement that statute is in a sense a minor
57
assault on the constitutional structure itself.
Although the imperative of the supremacy principle is quite strong, its
scope is limited. It has a decisive impact only when the text and statutory
history preclude genuine doubt about the statute's meaning. Moreover,
although courts may not violate clearly enacted legislative intent, the
supremacy principle does not prevent them from going beyond such intent in
implementing statutory language when there are gaps in the legislative
scheme. The language of a statute may also plausibly extend to situations
beyond the legislators' specific intentions. Congress may have failed to con-
sider broader applications of the statutory language, or may have failed to
agree on the matter, leaving it to the courts to resolve. In deciding whether
to adopt a broader reading, the court must look beyond the supremacy prin-
ciple for guidance.
The supremacy principle is an incomplete guide to judicial decision-
58
making, leaving much to be settled by recourse to other principles. Its lim-
ited scope does not mean, of course, that beyond its mandate "anything
55. On the importance of judicial candor, see Shapiro, In Defense ofJudicialCandor, 100 HARV.
L. REv. 731 (1987); see also Gunther, The Subtle Vices of the "Passive Virtues"--A Comment on
Principleand Expediency in JudicialReview, 64 COLUM. L. REV. 1, 10-22 (1964) (appellate judges
must give candid reasons for accepting or dismissing an appeal). But see Greenawalt, The Enduring
Significance ofNeutral Principles,78 COLUM. L. REv. 982, 1006-13 (1978) (candid statement of all
reasons for decision may be inappropriate if the court needs to hand down a single majority opinion
or if the actual grounds for decision would be legally suspect); Zeppos, JudicialCandorand Statu-
tory Interpretation,78 GEo. L.J. 353, 411 (1989) (complexity of judicial decisionmaking process
makes it impossible for judges to articulate their reasoning).
56. See Frankfurter, Some Reflections on the Reading of Statutes, 47 COLUM. L. REV. 527, 532,
545-46 (1947) (legislation should come from legislators, who are chosen to legislate, and not from
the courts).
57. See Radin, A Short Way With Statutes, 56 HARV. L. REV.388, 395-96 (1942). At the federal
level, the supremacy principle gains added authority from the supremacy clause, which makes valid
statutes "the supreme law of the land." U.S. CONST. art. VI, § 2. The clause appears to assume
that courts lack any power to disregard federal statutes on nonconstitutional grounds.
58. The scope of the supremacy principle should not be exaggerated, because to do so would only
hamper the search for other principled constraints on statutory interpretation.

goes." On the contrary, judicial decisionmaking may be tightly constrained

in other ways, even when the principle of legislative supremacy is not
involved.
Given its limited scope and relatively noncontroversial status, the
supremacy principle may seem inconsequential. Nevertheless, as Part II will
demonstrate, the principle is far from toothless.
II. JUDICIAL IMPLEMENTATION OF THE SUPREMACY PRINCIPLE
A. THE PARADIGM CASE
Chief Justice Burger's opinion for the Court in Tennessee Valley Authority
v. Hill 59 is perhaps the most notable modem example of conscious judicial
adherence to the supremacy principle. The facts of the case are familiar: the
Court blocked completion of the Tellico Dam to protect an unremarkable
but endangered species of fish known as the snail darter. Although the Court
seemed somewhat uncomfortable with this result, it was compelled to reach
60
its decision by the statute.
6
The relevant provision, section 7 of the Endangered Species Act (ESA), 1
commanded all federal agencies "to insure that actions authorized, funded,
or carried out by them do not jeopardize the continued existence" of an en-
dangered species, or "result in the destruction or modification of habitat of
such species .... -"62 As Chief Justice Burger's opinion made clear, Congress
had repeatedly rejected efforts to qualify this language with references to the
impairment of an agency's primary mission or to practicability. 63 Moreover,
the legislative history contains repeated references to the mandatory nature
64
and critical importance of an agency's duty to protect endangered species.
The statute and legislative history did not limit the agencies' duty to situa-
tions in which protection of endangered species is "reasonable." Justice
59. 437 U.S. 153 (1978).

60. The Court's discussion of the legal issues opens with the following observation:
It may seem curious to some that the survival of a relatively small number of three-inch
fish among all the countless millions of species extant would require the permanent halting
of a virtually completed dam for which Congress has expended more than $100 million.
The paradox is not minimized by the fact that Congress continued to appropriate large
sums of public money for the project, even after congressional Appropriations Commit-
tees were apprised of its apparent impact upon the survival of the snail darter. We con-
clude, however, that the explicit provisions of the Endangered Species Act require
precisely that result.
437 U.S. at 172-73.
61. Pub. L. No. 93-205, § 7, 87 Stat. 884, 892 (1973) (codified as amended at 16 U.S.C.
§ 1536(a)(2) (1982)).
62. 16 U.S.C. § 1536 (1976) (amended 1979).
63. 437 U.S. at 181-83.
64. Id. at 177-78, 182-84.

Powell, in dissent, suggested that Congress really only "intended to require

governmental agencies to take endangered species into account in the plan-
ning and execution of their programs. ' 65 This was simply wishful thinking,
in support of which Justice, Powell was able to marshall no convincing
66
arguments.
In criticizing the majority opinion, both Justice Powell and, later, Profes-
sor Dworkin relied heavily- on the fact that Congress continued to appropri-
ate money to build the Tellico Dam even after it knew of the threat to the
68
snail darter. 67 The Court carefully considered and rejected this argument.
The legislators who voted on the appropriations measures that included
funds for the dam 69 were never put on notice that they were being asked to
modify the Endangered Species Act. 70 Not only did the appropriations bills
65. Id. at 207-08 (Powell, J., dissenting).

66. The only legislative history Justice Powell offered in support of his view that a reasonableness
test should be imputed to § 7 is a statement by Senator Tunney that § 7 would not prohibit the
Army Corps of Engineers from building a particular road if they deemed it necessary to do so. Id.
at 208 n.17 (Powell, J., dissenting). As Chief Justice Burger noted, this statement was made during
debate on an early version of § 7 that contained a practicability limitation later dropped by Con-
gress. Id. at 186 n.31 (Burger, C.J., majority op.).
Justice Powell also argued that Congress intended that the statute apply only to future projects
rather than to those in which large commitments of resources had already been made. Id. at 202-06
(Powell, J., dissenting). This argument is only plausible if we ignore the high priority that Congress
placed on saving endangered species. If Congress was willing to forego unlimited future benefits to
preserve endangered species, there is no reason to assume that it was any less willing to forego
receiving a return from past expenditures. Whether Congress intends for a statute to be applied to
existing projects may depend in part upon the legislative priority attached to that issue. For exam-
ple, if Congress passed a statute absolutely forbidding the federal government from endangering the
lives of children, it would be perverse to imply a grandfather clause allowing the government to
complete any homicidal projects it already had underway.
In response to Hill, Congress amended the Endangered Species Act, creating a special Endan-
gered Species Committee empowered to grant exemptions from the Act. Such exemptions may be
granted if no reasonable alternatives are available, the costs of compliance outweigh the benefits of
allowing the action, and the action is in the public interest and of at least regional significance. See
Endangered Species Amendment of 1978, Pub. L. No. 95-632, § 3, 92 Stat. 3751, 3752 (codified as
amended at 16 U.S.C. § 1636(e)-(h) (1982)). The Committee denied an exception to the Tellico
Dam, see 9 ENVTL. L. REP. 10,033 (1979), but Congress authorized its completion nonetheless. See
id. at 10,183 (citing Pub. L. No. 96-69, tit. IV, 93 Stat. 437, 449 (1979)).
That Congress amended the ESA to allow exemptions on reasonableness grounds, and ultimately
authorized the Tellico Dam, did not immunize other projects; nor did it authorize courts to make
reasonableness judgments. Congress' action merely reaffirmed its institutional prerogative to make
law that the courts are obligated to enforce.- See infra notes 78 & 81. Indeed, very few projects have
ever been exempted from the statute, see Plater, In the Wake of the Snail Darter:An Environmental
Law Paradigmand its Consequences, 19 U. MICH. J.L. REF. 805, 829 (1986), and Congress has been
content to allow the Act to have an extremely broad sweep. See infra note 81.
67. See Tennessee Valley Authority v. Hill, 437 U.S. at 209-10 (Powell, J.,dissenting); R.
DWORKIN, LAW'S EMPIRE 348-49 (1986).
68. 437 U.S. at 189-93.
69. See Pub. L. No. 94-355, 90 Stat. 889 (1976); Pub. L. No. 95-96, 91 Stat. 797 (1977).
70. See 437 U.S. at 192. Congress customarily continues to appropriate funds while litigation is
pending, without prejudice to the results of the litigation. For examples ofjudicial discussion of this

fail to refer to the earlier statute, they also failed to refer specifically to the
dam itself.7 1 A legislator would have had to consult the Appropriations
Committee report to have known that passage of the appropriations bill
might affect the ESA.72 As the Court pointed out, to expect members to
scrutinize appropriations committee reports would have undesirable effects
on the congressional process:
When voting on appropriations measures, legislators are entitled to operate
under the assumption that the funds will be devoted to purposes which are
lawful and not for any purpose forbidden. Without such an assurance,
every appropriations measure would be pregnant with prospects of altering
substantive legislation, repealing by implication any prior statute which
might prohibit the expenditure. Not only would this lead to the absurd
result of requiring Members to review exhaustively the background of
every authorization before voting on an appropriation, but it would flout
73
the very rules the Congress carefully adopted to avoid this need.
Thus, passage of the appropriations bills did not itself carry any weight. The
views of the Appropriations Committee, considered apart from any possible
ratification of Congress as a whole, were also entitled to little weight. The
Appropriations Committee apparently believed that the Tellico Dam could
be lawfully completed, 74 but it conducted no hearings on the subject and
lacked any particular expertise on the ESA. 75 Giving weight to such action
on appropriations bills would merely invite special interests to abuse the ap-
76
propriations process as a means of undercutting substantive legislation.
For these reasons, the Court was correct to conclude that the statutory
mandate was clear. The Court was also correct in its closing statement of the
practice prior to Hill, see Environmental Defense Fund v. Froelke, 473 F.2d 346, 352-55 (8th Cir.
1972); Committee for Nuclear Responsibility, Inc. v. Seaborg, 463 F.2d 783, 785-86 (D.C. Cir.
1971); Environmental Defense Fund v. TVA, 468 F.2d 1164, 1182 (6th Cir. 1972); see also Plater,
supra note 66, at 842-44 (appropriations committees often continue funding despite contrary stat-
utes or policies).
71. See 437 U.S. at 189 n.14; 90 Stat. at 899 (TVA appropriations); 91 Stat. at 808 (same).
72. See 437 U.S. at 189.
73. Id. at 190-91.
74. Id. at 189.
75. As the Court observed:
We venture to suggest that the House Committee on Merchant Marine and Fisheries and
the Senate Committee on Commerce would be somewhat surprised to learn that their
careful work on the substantive legislation had been undone by the simple - and brief-
insertion of some inconsistent language in Appropriations Committees' Reports.
Id. Moreover, the committee reports may have been based on a factual judgment about the possibil-
ity of transplanting the snail darter rather than on any specific interpretation of the statute. See id.
at 192-93.
76. See Sunstein, Beyond the Republican Revival, 97 YALE L.J. 1539, 1582 (1988) (appropria-
tions process lacks visibility and is "comparatively likely to be dominated by well-organized private
groups").

supremacy principle: "Our individual appraisal of the wisdom or unwisdom

of a particular course consciously selected by the Congress is to be put aside
in the process of interpreting a statute. Once the meaning of an enactment is
discerned and its constitutionality determined, the judicial process comes to
77
an end."
In short, Hill is a judicial performance exemplifying obedience to the legis-
lature's commands at the expense of the Justices' own political preferences. 78
Ironically, it is just this aspect of the opinion that has been most harshly
criticized as formalistic. For example, Professor Dworkin faults the Justices
for failing to implement their own policy views. Hercules, his mythical ideal
judge, has no difficulty in reaching a contrary result because Hercules thinks
that "reading the statute to save the dam would make it better from the point
of view of sound policy."'79 According to Dworkin, Hercules has "no reason
of textual integrity arguing against that reading." Hercules' notion of textual
integrity, one gathers, has an extraordinary degree of elasticity.
Professor Dworkin is not wrong to stress the creative aspects of the judge's
role in statutory cases, nor is he wrong to suggest that the judge's own views
of public policy can properly enter into the process. Where he goes wrong, at
least if his discussion of Hill is to be taken seriously,8 0 is in his apparent
inability to perceive the difference between creatively interpreting a statute
and rewriting the statute to suit one's desires. This is the distinction that the
supremacy principle respects.
What may make Hill seem formalistic is that the Court takes Congress at
its word, even though it is plausible to assume that many legislators really
would have preferred that the Tellico Dam be built. Many of those who
voted for the measure probably did not have strong feelings about endan-
gered species. But the question is not what they privately felt, but rather
what they voted on and what they understood it to mean. It seems clear that
they did not understand the statute to embody a reasonableness test.
77. 437 U.S. at 194. The opinion then ends with a quotation ascribed to Sir Thomas More in a
play about the crucial importance of the rule of law. Id. at 195 (citing R. BOLT, A Man for All
Seasons, in THREE PLAYS 147 (Heinemann ed. 1967)).
78. The Court was correct to refuse an invitation to "balance the equities" before issuing an
injunction. 437 U.S. at 193-95. In essence, balancing the equities in this context would simply have
meant displacing Congress' decision about the importance of protecting endangered species. See
Farber, Equitable Discretion, Legal Duties, and Environmental Injunctions, 45 U. PITT. L. REV.
513, 518-19 (1984) (arguing that "a court cannot use equitable discretion as a pretext for reordering
priorities that Congress already has set"); Plater, Statutory Violations and EquitableDiscretion, 70
CALIF. L. REV. 524, 591 (1982) ("When a court is asked to exempt a defendant from a legislative
enactment, it is being invited to take over a core function of the legislature.").
79. R. DWORKIN, supra note 67, at 347.
80. Professor Dworkin has previously been faulted for his lapses in analyzing particular cases.
See Christie, Dworkin's "Empire" (Book Review), 1987 DUKE L.J. 157, 182 n.186 (reviewing R.
DWORKIN, LAW'S EMPIRE (1986)). Although Professor Dworkin cites actual cases, perhaps he
only means that they should be taken as the basis for his own hypotheticals.

Dworkin's argument is plausible because of suspicions that the legislators

were insincere in voting for the Endangered Species Act and even less honest
in their posturing while the bill was being considered. To give effect to this
suspicion would legitimate such legislative hypocrisy. If judges want the leg-
islature to act with integrity, they must hold legislators to their public posi-
tions. Judges must not allow legislators to use statutes to strike poses,
knowing that courts will bail them out later. Not only does the supremacy
principle act as a constraint on courts, it also, indirectly, disciplines the
legislature.81
B. VIOLATIONS OF THE SUPREMACY PRINCIPLE
Tennessee Valley Authority v. Hill is the paradigm of judicial adherence to

the supremacy principle. Unfortunately, it is not hard to find cases in which
the courts have been less obedient toward legislative mandates.8 2 The cases
in this section are not instances of simple judicial mistakes.8 3 Rather, they
resemble "jury nullification" because the judges were motivated by their disa-
greement with the legislature's view of public policy.
The first case, InternationalHarvester Co. v. Ruckelshaus, 4 is a leading
81. Congress responded to Hill by amending the Endangered Species Act to create a cabinet-
level committee to review agency requests for exemptions from the Act. See supra note 66. The
review committee rejected TVA's request for an exemption because it found that the benefits of
completing the Tellico Dam did not clearly outweigh the benefits of other available alternatives.
Summary and Comments: The 1978 Amendments to the EndangeredSpecies Act: Evaluating the
New Exemption Process Under § 7, 9 Envtl. L. Rep. (Envtl. L. Inst.) 10,031, 10,033 (1979). In
response, Congress inserted a provision into the Fiscal Year 1980 energy and water development
appropriations bill, authorizing completion of the Tellico Dam "notwithstanding the provisions of
16 U.S.C., chapter 35 or any other law." Pub. L. No. 96-69, tit. IV, 93 Stat. 437, 449 (1979). This
provision, a prime example of a pork-barrel measure, had been strongly supported by the Tennessee
congressional delegation. See Plater, supra note 66, at 857 ("linkage of Tennessee politicians to the
TVA" was a critical factor in passage of legislation). President Carter signed the provision into law
as part of a political deal to obtain a three-year authorization of the Endangered Species Act. See
96th Congress, 1st Session: Environmental Issues in Limbo, 10 Envtl. L. Rep. (Envtl. L. Inst.)
10,009, 10,013 n.54 (1980). Thus, the Supreme Court's decision forced Congress directly to con-
front the problem posed by Hill. Given the inherently political nature of the problem-a trade-off
between protecting endangered species and promoting energy generation and economic growth, the
cabinet review committee was probably a better solution than judicial exercise of equitable discre-
tion. It would have been quite difficult for a court candidly to weigh the value of the dam against
the value of saving the snail darter. To do so accurately, it would have had to announce the eco-
nomic worthlessness of the congressionally authorized dam. For a detailed discussion of the litiga-
tion and its aftermath, see Plater, supra note 66. The amendment adding the cabinet-level
committee may actually have strengthened the statute because exemptions became so difficult to
obtain. Id. at 828.
82. This is not to say that such violations are common. The point is merely that the supremacy
principle has more "bite" than might first appear.
83. These cases are drawn from the author's teaching in areas of environmental law and civil
rights. Presumably, cases in these areas are not unique.
84. 478 F.2d 615 (D.C. Cir. 1973).

case in environmental law. Under the Clean Air Act,8 5 car manufacturers
were required to meet stringent air pollution standards by 1975. The
Environmental Protection Agency (EPA) was authorized to grant a one-year
suspension from the standards imposed by the Act if, among other things,
"the applicant has established that effective control technology... [is] not
available or [has] not been available for a sufficient period of time to achieve
compliance prior to the effective date ....-86 In 1972, the EPA denied
applications for a one-year suspension filed by several car manufacturers.8 7
In tests the manufacturers had performed on catalytic converters, none of the
test cars had met the EPA's standards. After the EPA had made various
adjustments to the test results, the agency concluded that the applicants had
failed to demonstrate that the technology would be unavailable in time to
88
meet the statute's mandate.
In an opinion by Judge Levanthal, the D.C. Circuit reversed. As a matter
of policy, Judge Levanthal concluded, the agency should not take a "hard-
nosed" approach to the suspension issue.8 9 After all, he observed, the envi-
ronmental harm caused by a suspension would be relatively minor and the
economic consequences of denying the suspension would be grave. 90 Based
on this risk balancing, Judge Levanthal determined that once the companies
reported their test results, the burden of proof shifted to the EPA to support
the reliability of its methodology for adjusting these results. 91 After con-
ducting a detailed examination of the technical issues, Judge Levanthal con-
cluded that the EPA had failed to carry its burden:
The number of unexplained assumptions used by [the EPA], the variance
in the [EPA's] methodology from that of the Report of the National Acad-
85. Pub. L. No. 89-272, 79 Stat. 992 (1955) (codified as amended at 42 U.S.C. §§ 7401-7642
(1982)).
86. 42 U.S.C. § 1857f-l(b)(5)(D)(iii) (1970) (current version at 42 U.S.C. § 7521(b)(5)(C)(iii)
(1982)). Under this provision, the EPA also must determine that the suspension is essential to the
public interest, that all good faith efforts have been made to meet the standards, and that the Na-
tional Academy of Sciences agrees that the technology is unavailable. Id. at § 7521(b)(5)(C).
87. 478 F.2d at 624.
88. Id. at 626.
89. Id. at 649.
90. Judge Levanthal reasoned thus:
This case inevitably presents, to the court as to the Administrator, the need for a perspec-
tive on the suspension [issue] that is informed by an analysis which balances the costs of a
"wrong decision" on feasibility against the gains of a correct one. These costs include the
risks of grave maladjustments for the technological leader from the eleventh-hour grant of
a suspension, and the impact on jobs and the economy from a decision which is only
partially accurate, allowing companies to produce cars but at a significantly reduced level
of output. Against this must be weighed the environmental savings from denial of suspen-
sion. The record indicates that these will be relatively modest.
478 F.2d at 641.
91. Id. at 641, 643.

emy of Sciences, and the absence of an indication of the statistical reliabil-

ity of the [EPA's] prediction, combine to generate grave doubts as to
whether technology is available to meet the 1975 statutory standards. 92
Judge Levanthal's reasoning, boiled down, was that it was better policy to
err in the direction of granting the suspension, so doubtful cases should be
decided in favor of the industry. This meant, of course, that the burden was
effectively on the EPA to establish technological feasibility. There is only
one problem with this result: the statute required precisely the opposite.9 3
Thus, if the availability of technology was in doubt, as the court apparently
believed it was in InternationalHarvester, the statute required a decision in
favor of the EPA. Obviously, Judge Levanthal believed-perhaps cor-
rectly-that such a decision would be poor public policy. 94 But equally obvi-
ous is that Congress had taken (to use Judge Levanthal's expression) more of
a "hardnosed" approach to achieving air pollution standards. The principle
of legislative supremacy required Judge Levanthal to defer to Congress' ex-
95
pressed views.
Another environmental law decision that violates the supremacy principle
is Ruckelshaus v. Sierra Club.9 6 At issue in Ruckelshaus was whether the
Sierra Club was entitled to attorneys' fees for its participation in an impor-
tant air pollution case. The underlying litigation was an extremely complex,
multi-party challenge to the EPA regulation governing sulfur dioxide emis-
sions by coal-burning power plants. None of the Sierra Club's attacks on the
regulation were successful, but the D.C. Circuit found the Sierra Club's
briefs helpful in understanding the complicated issues in the case. Addition-
ally, the Sierra Club's arguments may have redounded to the benefit of the
EPA as the agency defended the regulations from industry challenges.
Under the circumstances, the court concluded that a fee award was
92. Id. at 648.

93. 42 U.S.C. § 7521(b)(5)(C)(iii) (1982). The statute explicitly required the waiver applicant to
establish that the necessary technology was not available. See supra note 86 and accompanying
text. The disparity between the statute and the court's decision is noted in Stewart, Regulation,
Innovation and Administrative Law: A ConceptualFramework, 69 CALIF. L. REv. 1256, 1304-05
(1981).
94. Judge Levanthal's opinion gains plausibility from the suspicion that Congress did not really
mean what it said. See Elliott, Ackerman & Millian, Toward a Theory of Statutory Evolution: The
FederalizationofEnvironmentalLaw, 1 J.L. ECON. & ORG. 313, 336 (1985) (arguing that Congress
knew pollution standards were unrealistic). As in Hill, this suspicion, although perhaps well-
founded, should not carry much weight in statutory interpretation. See supra text accompanying
note 78.
95. Had Judge Levanthal deferred to Congress' expressed views, Congress would have been
forced to intervene, which in the long run seems institutionally healthier than a judicial "bail-out."
Congress might well have been able to devise a more creative remedy, such as a differential emis-
sions tax that would have prevented an industry shut-down while rewarding Ford, the most diligent
of the car companies.
96. 463 U.S. 680 (1983).

LEGISLATIVE SUPREMACY
appropriate. 97
In an opinion by Justice Rehnquist, the Supreme Court reversed. If its
liberal use of italics is any guide, the Court found the idea of awarding fees to
a nonprevailing party to be almost scandalous:
Our basic point of reference is the "American Rule" . . . under which
even "theprevailinglitigant is ordinarily not entitled to collect a reasonable
attorneys' fee from the loser." It is clear that generations of American
judges, lawyers, and legislators, with this rule as the point of departure,
would regard it as quite "inappropriate" to award the "loser" an attorney's
fee from the "prevailing litigant." Similarly, when Congress has chosen to
depart from the American Rule by statute, virtually every one of the more
than 150 federal fee-shifting provisions predicates fee awards on some suc-
cess by the claimant .... 98
This consistent rule, the Court continued, reflects ordinary notions of justice
that "reject the idea that a party who wrongly charges someone with viola-
tions of the law should be able to force that defendant to pay the costs of the
wholly unsuccessful suit against it." 99
Congress, however, is apparently less absolutely attached to the American
Rule, as illustrated by the routine provision for attorneys' fees in modem
statutes.' °° Unlike the statutes cited by the Court, 0 1 the fee-shifting statute
applied in Ruckelshaus quite pointedly avoided the usual reference to pre-
vailing parties. Instead, it provided for a fee award whenever the court deter-
mines "such award is appropriate."' 102
This phrasing was not inadvertent. The House Report on the provision
explains that the purpose of fee awards is "to encourage litigation which will
assure proper implementation and administration of the act or otherwise
serve the public interest."' 0 3 Consequently, the report continues, "[t]he com-
mittee did not intend that the court's discretion to award fees under this
provision should be restricted to cases in which the party seeking fees was the
97. Sierra Club v. Gorsuch, 672 F.2d 33, 41 (D.C. Cir. 1982) (per curiam), rev'd sub nom. Ruck-
elshaus v. Sierra Club, 463 U.S. 680 (1983).
98. 463 U.S. at 683-84 (citation omitted) (emphasis in original).
99. Id. at 685.
100. See, eg., 42 U.S.C. § 1988 (1982) (court granted discretion to award prevailing party rea-
sonable attorneys' fees).
101. 463 U.S. at 684 nn.3-5 (citing the Voting Rights Act of 1965, the Freedom of Information
Act, and the Real Estate Settlement Procedure Act, among others).
102. 42 U.S.C. § 7607(f) (1982). Such reference to granting an award when "appropriate" nor-
mally would be quite open-ended, but is less so in this context. Given the fact that fee-shifting
provisions in other statutes routinely refer to prevailing parties, the use of different language here
suggests strongly that Congress meant something other than a prevailing party standard. Whatever
"appropriate" means here, it seems least likely to mean "prevailing party." Yet the Court adopted
precisely the latter reading. See Ruckelshaus, 463 U.S. at 686.
103. 463 U.S. at 687 (citing H.R. REP. No. 294, 95th Cong., 1st Sess. 337, reprintedin 1978 U.S.
CODE CONG. & ADMIN. NEWS 1077, 1416 [hereinafter H.R. REP. No. 294]).

'prevailing party.' "4 The committee then cited a court of appeals opinion
rejecting the prevailing party standard because of the important role of pri-
vate watchdogs in implementing "novel and complex" legislation. 10 5
Despite the Court's half-hearted effort to explain away the language and
legislative history of the statute, 10 6 the meaning of the statute seems unmis-
takable. The Supreme Court approached the case with a deep suspicion of
fee-shifting. As a result, the Court apparently was either unwilling or unable
to comply with the clear statutory directive. Justice Stevens was quite right
to argue in dissent that the crux of the case was the Court's duty to defer to
10 7
Congress' policy decision.
These cases are not subject to criticism because they violate some supposed
rule against judges consulting their own views of public policy when inter-
preting statutes. Rather, they should be criticized for crossing the line be-
tween creative interpretation made necessary by statutory ambiguity and
creative rewriting made possible by the judges' "sly refusal" to follow a rea-
sonably clear statutory command.108
A similar problem is found in a more dramatic setting in United Steelwork-
09 which involved one of the most "starkly divi-
ers of America v. Weber, 1
sive"11 0 issues in American politics. The Weber Court held that title VII I
allows employers to engage voluntarily in affirmative action favoring minor-
104. H.R. REP. No. 294, supra note 103, at 337.

105. Id. (citing NRDC v. EPA, 484 F.2d 1331, 1338 (1st Cir. 1973)).
106. The Court argued that Congress understood the "prevailing party" standard to be quite
narrow, excluding parties that had succeeded only on some issues. Hence, the Court reasoned,
Congress intended by rejecting the prevailing party standard to "expand the class of parties eligible
for fee awards from prevailing parties to partiallyprevailing parties." 463 U.S. at 687-88 (emphasis
in original). As Justice Stevens explained in dissent, this was historically inaccurate. Id. at 708-11
(Stevens, J., dissenting).
107. Justice Stevens said:
Regardless of our views about the wisdom of the choice Congress made, we have a plain
duty to accept it. Congress consciously selected a particular course: that a party who
seeks judicial review of an EPA regulation may be entitled to compensation from the
Government, when the court deems it "appropriate," even if the reviewing court deter-
mines that there is no ground for disturbing the agency's conclusions. I would construe
this category of "appropriate" cases to be narrow; it is wrong, however, to read it out of
the statute altogether. It is not the function of the courts to "sit as a committee of review,
nor are we vested with the power of veto."
Id. at 711-12 (footnote and citations omitted).
108. Dean Calabresi used the term "sly refusal" to characterize willful judicial ignorance of dis-
cernible ttatutory meaning. See G. CALABRESI, A COMMON LAW FOR THE AGE OF STATUTES 38
(1982).
109. 443 U.S. 193 (1979).
110. See Sullivan, Sins of Discrimination: Last Term's Affirmative Action Cases, 100 HARV. L.
REV. 78, 78 (1986).
111. Civil Rights Act of 1964, Pub. L. No. 88-352, tit. VII, 78 Stat. 241, 253 (1964) (codified as
amended at 42 U.S.C. § 2000e-e17 (1982)).

ity groups, when a "manifest racial imbalance" is present, provided that the
affirmative action plan meets certain standards of reasonableness.' 112 Nota-
bly, Weber does not require that the threshold racial imbalance be the result
of even arguably illegal conduct by the employer, or of illegal discrimination
by third parties.' 13 In Weber, because blacks were a large percentage of the
local work force, but only a tiny percentage of the employer's skilled
craftworkers, the Court upheld a crafts skills training program that reserved
for blacks half of all openings in the program.114
In discussing Weber, it is helpful to distinguish between two conceptually
distinct types of affirmative action. Remedial affirmative action is aimed at
correcting possible illegal discrimination by an employer."15 It need not rest
on actual proof of a violation, but it 'does require at least some evidence of
unlawful discrimination, such as a "conspicuous... imbalance in tradition-
ally segregated job categories."' " 6 The Court has recently made clear that
only remedial affirmative action is permissible in situations involving state
governments." 7 Nonremedial programs are aimed at correcting racial im-
balances regardless of their cause, even when the employer has a spotless
record. For example, even though an employer has actually hired a greater
percentage of minority applicants than white applicants, if a shortage of mi-
nority applicants has caused a racially imbalanced workforce the employer
might institute a nonremedial program. In Weber, at least as construed in
later opinions, the Court upheld nonremedial affirmative action.1 8
The problem with the Weber holding, as Justice Rehnquist pointed out in
dissent, is that it contradicts both the plain language and the legislative his-
tory of title VII. 119 The statute itself clearly establishes a general policy
112. See Johnson v. Transportation Agency, 480 U.S. 616, 625-26, 630 (1987) (explaining
Weber).
113. Id. at 624-25.
114. Weber, 443 U.S. at 197-200.
115. For particularly clear examples of the use of affirmative action as a remedial tool, see United
States v. Paradise, 480 U.S. 149 (1987), and Sheet Metal Workers v. EEOC, 478 U.S. 421 (1986).
116. Johnson, 480 U.S. at 630 (quoting Weber, 443 U.S. at 209 (Blackmun, J., concurring)).
117. See City of Richmond v. J.A. Croson Co., 109 S. Ct. 706, 727 (1989) (under fourteenth
amendment, classifications based on race are valid as remedial measures).
118. E.g., Wygant v. Jackson Board of Education, 476 U.S. 267, 280-82 (1987) (plurality op.);
Johnson, 480 U.S. at 630.
119. Weber, 443 U.S. at 252-53 (Rehnquist, J., dissenting). Arguments that the statute is ambigu-
ous focus on the use of the word "discriminate" in some provisions, claiming that this term might
not apply to affirmative action because any discrimination resulting from affirmative action would
not be invidious. See R. DWORKIN, A MATTER OF PRINCIPLE 318 (1985) ("discriminate" may be
used in a neutral way or in an evaluative way); Eskridge, supra note 11, at 1489-90 ("discriminate"
has at least two plausible interpretations). This indeterminacy is the basis for Dworkin's argument
that because Congress had "no pertinent collective understanding" about affirmative action, judges
must use their own political judgment about its merits. See Dworkin, supra, at 329. In context,
however, there seems to be no reason to read "discriminate" as meaning anything more subtle than
"disadvantage."

against considering race in employment decisions. Section 703(a)(2), for ex-

ample, prohibits employers from classifying employees "in any way which
would deprive or tend to deprive any individual of employment opportunities
or otherwise adversely affect his status as an employee, because of such indi-
vidual's race, color, religion, sex or national origin."' 120 Furthermore, the
statute's legislative history demonstrates that Congress was uneasy about
preferential treatment, even as a remedy for past violations. The Senate floor
captain in charge of the bill said that even if an employer had discriminated
in the past, his only "obligation would be simply to fill future vacancies on a
nondiscriminatory basis. He would not be obliged - or indeedpermitted...
to prefer Negroes for future vacancies."' 121 In sum, allowing employers to
use racial preferences to remedy "imbalance" cannot be reconciled with the
statutory language or legislative history. 122
In response to these criticisms, Justice Brennan, writing for the Court,
made two arguments. First, he pointed out that the supporters of the Civil
Rights Act intended to improve the economic lot of minority groups. 1 23
Congress chose to pursue this goal, however, through an antidiscrimination
policy. The Court was not free to displace that congressional choice merely
because the Justices thought Congress had picked the wrong strategy. To
return to the military analogy, Justice Brennan was in somewhat the same
position as a captain who has violated a general's order. The captain argues
that he did not really disobey, because his action was a better way of attain-
ing the general's ultimate purpose of winning the war. If such a defense were
recognized, a military order would become little more than a recommenda-
tion. Similarly, Justice Brennan's argument would reduce statutory man-
dates to the status of policy proposals.
Justice Brennan's fallback argument was that section 7030) of the statute
introduces some ambiguity about congressional intent regarding remedial
programs that employers enter into voluntarily. Section 7030) states that
nothing in the statute shall be construed to "require" any employer to give
preferential treatment to remedy a racial imbalance. 124 Justice Brennan ar-
gued that this provision authorizes voluntary preferential treatment. Other-
wise, he suggested, Congress would have disclaimed any intent to "require or
120. 42 U.S.C. § 2000e-2(a)(2) (1982). Similar language is found in subsection (a)(1) and in
section (d).
121. 110 CONG. REC. 7213 (1964) (emphasis added). As Justice Rehnquist demonstrates in his
dissent, the record is replete with similar statements in favor of "color-blindness." 443 U.S. at 235-
51.
122. For an extensive discussion of this point, see Meltzer, The Weber Case: The JudicialAbro-
gation of the Antidiscrimination Standard in Employment, 47 U. CHi. L. REV. 423 (1980).
123. Weber, 443 U.S. at 202-04. For a general discussion regarding the risks of allowing such
appeals to statutory purpose to undermine statutory rules, see Schauer, Formalism, 97 YALE L.J.
509, 532-35 (1988).
124. 42 U.S.C. § 2000e-2(j) (1982).

permit" preferential treatment.1 25 As even those who support Weber con-

cede, this is less than a compelling argument. 2 6 If Congress meant to
authorize preferential hiring, it almost certainly would have used less subtle
means.
Against Justice Brennan's arguments must be weighed the relatively clear
statutory language and legislative history stressing the inappropriateness of
racial considerations in employment decisions. It is difficult to escape the
conclusion that the Court went "not merely beyond, but directly against title
' 27
VII's language and legislative history."'
In displacing congressional policy to allow preferential hiring, the Court
appears to have violated the supremacy principle. It is not certain, however,
that all types of affirmative action invariably violate this principle with re-
spect to title VII, because some form of affirmative action may be necessary
to remedy violations of the statute. Vigorously pursuing an antidiscrimina-
tion policy may require employers, the Equal Employment Opportunity
Commission (EEOC), or courts to use hiring targets as the only effective way
of ensuring compliance with title VII's mandate.' 28 For example, it would be
enormously inefficient for General Motors to monitor company-wide compli-
ance with an antidiscrimination directive by reviewing every minority job
application submitted to each of the company's regional offices and plants.
Instead, GM could ensure compliance with the directive by instituting mi-
nority hiring targets at each office and plant. Only if a plant failed to meet its
target would headquarters actually need to audit applications. Similarly, it
might not be efficient or even possible for the EEOC to monitor discrimina-
125. 443 U.S. at 205-06.

126. See Edwards, Affirmative Action or Reverse Discrimination:The Head or Tail of Weber, 13
CREIGHTON L. REV. 713, 744-45 (1980) (stating that any voluntary preferential treatment poten-
tially allowed by § 7030) is still subject to §§ 703(a) and (d), which clearly "control the validity" of
such treatment); Eskridge, supra note 11, at 1490 n.41 (suggesting that this argument is not wholly
invalid, but rests upon a "highly unreliable" canon of construction and provides an "implausible
view" of the purpose of § 7030)). By analogy to Brennan's argument in Weber, one could read the
eleventh amendment as an affirmative grant of federal jurisdiction over states in federal question
cases.
127. 443 U.S. at 255 (Rehnquist, J., dissenting) (emphasis in original).
128. The need for hiring targets has probably become more acute since the statute has been
construed to prohibit not only intentional discrimination but also facially neutral employment prac-
tices that have an unintended disparate impact on minorities. See Watson v. Fort Worth Bank &
Trust, 108 S. Ct. 2777, 2785 (1988) (O'Connor, J., plurality op.). Several Justices have commented
on the pressure such a construction of title VII creates toward affirmative action, id. at 2787-88,
including Justice Blackmun in hisWeber concurrence. 443 U.S. at 209-11 (Blackmun, J., concur-
ring). However, this pressure alone is not a sufficient basis for allowing affirmative action. Since
Weber, the Court has retreated from the disparate impact doctrine. See Wards Cove Packing Co. v.
Atonio, 109 S. Ct. 2115, 2124-25 (1989) (cutting back on "disparate impact" claims under title
VII). If the Court were to rely solely on the pressure for affirmative action created by disparate
impact analysis, it appears that the legality of affirmative action under title VII could change every
decade.

tion on a case-by-case basis, whereas using minority hiring targets as a trig-

ger for case-by-case investigation would be both affordable and effective.
An analogy to Hill may be illuminating. Congress' high priority on saving
endangered species prohibits a court from weighing other social policies
against that congressional goal. On the other hand, if endangering one spe-
cies was the only way to save another, either choice would be arguably con-
sistent with the congressional mandate. Allowing employers to use
affirmative action to remedy unlawful discrimination,129 like threatening one
endangered species to save another, does not violate the supremacy principle.
The congressional mandate is perhaps bent, but not flouted by remedial af-
firmative action. In contrast, nonremedial affirmative action cannot be
squared with the statute. 130
Title VII can, and perhaps should, be read to allow affirmative action to
correct unlawful discrimination. The Weber Court violated the congres-
sional mandate, however, by allowing broad nonremedial use of affirmative
action. The most interesting arguments in favor of upholding Weber concede
this point, but argue that events since 1964 justify the result.' 3' Before con-
sidering those arguments, we first must consider the broader question of how
the passage of time affects the obligation to obey legislative mandates.
III. THE TEMPORAL ELEMENT IN LEGISLATIVE SUPREMACY
Several writers have suggested that because a statute's legal meaning is not
entirely fixed when enacted, post-enactment events can be relevant to inter-
preting that statute. 132 A corollary of this theory is that courts should have
more flexibility in interpreting old statutes than they have with respect to
recent enactments. 133 The supremacy principle might suggest a contrary
conclusion. If the text and original legislative understanding are clear, a
129. In Justice O'Connor's view, title VII permits this use of affirmative action. See Johnson v.
Transportation Agency, 480 U.S. 616, 650 (1987) (O'Connor, J., concurring).
130. See Watson, 108 S. Ct. at 2788 (congressional intent is "clear and emphatic" that title VII
not lead to nonremedial affirmative action).
131. See infra text accompanying notes 175-80.
132. See, eg., R. DWORKIN, LAW'S EMPIRE 347-48 (1986) (suggesting that political decisions
subsequent to enactment may be relevant to statutory interpretation); Farber, supra note 14, at 6-7
(arguing that alternative interpretive approaches are justified when originalist analysis becomes un-
tenable); Popkin, The CollaborativeModel ofStatutory Interpretation,61 S. CAL. L. REV. 541, 620-
26 (1987) (providing examples of statutes interpreted in light of later developments).
133. See R. DWORKIN, supra note 132, at 350 (as statute ages, court may pay less attention to
formal statements made at time of enactment); Simon, The Authority of the Framersof the Constitu-
tion: Can OriginalistInterpretationBe Justified?,73 CALIF. L. REv. 1482, 1537 (1985) (arguing that
the Court has more flexibility when interpreting older constitutional amendments than when inter-
preting more recently enacted amendments). Legislators may be indifferent to the long-term effects
of statutes because these effects are minimal after being discounted to present value. This observa-
tion may sometimes be correct, but the argument is slippery because legislators may expect fairly
significant long-term benefits that, even when discounted to present value, would be far from mini-

1989] LEGISLATIVE SUPREMACY 307
court would appear to have an obligation to apply the statute regardless of

later developments. In fact, the supremacy principle may allow courts to
disobey clear statutory language when, because of post-enactment events, the
legislature's intent would be undermined by strict adherence to the statutory
formula.
A. CHANGING VIEWS OF PUBLIC POLICY
Statutes are difficult to pass, but they are also hard to repeal. Conse-
quently, a statute may remain on the books indefinitely, even though it no
longer represents the dominant societal view of public policy. If a court be-
lieves that a statute is no longer consistent with contemporary mores, enforc-
ing the legislative command may be a rather distasteful exercise.
Under certain limited circumstances, accepted methods are available to
deal with this problem.13 4 For example, if later legislation reflects a more
palatable view of public policy, the court can declare that the earlier statute
has suffered repeal by implication, or, less drastically, can reinterpret the old
statute in light of later enactments.135 Alternatively, if the statute touches
upon a constitutionally sensitive area like gender-based classification, the vin-
tage of the statute may be relevant to its constitutionality. 136 But often
neither of these approaches is applicable, leaving the court with no apparent
way to deal with the problem of the outmoded statute.
A somewhat radical solution is to declare the outmoded statute defunct.
In support of this solution, Dean Guido Calabresi argues that the legitimacy
of an old statute rests only on the facts that "it has gone unrepealed; and it
once commanded a majoritarian bias."' 137 The first fact, -he observes, is
mal. Moreover, even if legislators themselves have short time perspectives, the interest groups in-
volved in the legislative process may take a longer view.
134. These conventional approaches are discussed in W. ESKRIDGE & P. FRICKEY, supra note 2,
at 869-71, 879-82; see also Bonfield, The Abrogation ofPenalStatutes By Nonenforcement, 49 IOWA
L. REV. 389, 392-95 (1964) (outlining judicial approaches and arguing for civil doctrine of desue-
tude in which prolonged failure to enforce results in abrogation or repeal).
135. See United States v. Fausto, 108 S. Ct. 668, 676-77 (1988) (Civil Service Reform Act re-
quired repeal of judicial interpretation provision of the Back Pay Act, but not of Back Pay Act
itself).
136. See Mississippi Univ. for Women v. Hogan, 458 U.S. 717, 725, 730 n.16 (1982) (statute's
gender-based classification constitutes "archaic" characterizations of women and denies equal pro-
tection); Califano v. Goldfarb, 430 U.S. 199, 211 (1977) (plurality op.) (gender-based classification
deriving from "archaic and overbroad" generalizations deny equal protection); id. at 223-24 (Ste-
vens, J., concurring) (arguing that discrimination at issue is accidental byproduct of traditional way
of thinking about women, and hence is not a substantial basis, under fifth amendment analysis, for
disparate treatment).
137. G. CALABRESI, supra note 108, at 101-02.- Calabresi's view of legislative legitimacy is criti-
qued in Hutchinson & Morgan, Calabresian Sunset: Statutes in the Shade (Book Review), 82
COLUM. L. REV. 1752, 1762-66 (1982) (reviewing G. CALABRESI, A COMMON LAW FOR THE AGE
OF STATUTES (1982)).

equally true of old common law decisions that the legislature has left stand-
ing, and the second fact lacks significance if majority support for the statute
has evaporated. 138 Hence, an old statute is no more (although, perhaps, no
less) entitled to the court's respect than an old judicial precedent.
The major flaw in this argument is that it treats inertia as simply an unfor-
tunate side effect of the political system-as if an ideal democratic system
would instantly reflect changes of majority sentiment. On the contrary, the
agenda rules and institutional structures that create legislative inertia are
themselves fundamental to the workings of legislatures. 139 Without these
constraints and constructs, legislatures would be plagued by instability and
would be unable to function as deliberative bodies.14°
Our legislative process is designed so that laws will outlive the political
coalitions that enact them. 14 1 It is not at all clear that a democratic system
could function otherwise. What is clear is that pure majoritarianism has
138. G. CALABRESI, supra note 108, at 102-03. Calabresi's thesis is presaged in G. GILMORE &
C. BLACK, THE LAW OF ADMIRALTY 447-48 (2d ed. 1975).
139. See generally Farber & Frickey, The Jurisprudenceof Public Choice, 65 TEX. L. REv. 873,
903-06 (1987) (discussing how structural and rule-based constraints on legislative process promote
legislative stability).
140. This is one of the fundamental insights of modem public choice theory. See W. RIKER,
LIBERALISM AGAINST POPULISM 169-95 (1982) (arguing that the outcome of a legislative vote is
the product of manipulation of legislative choices by a means of agenda control); Shepsle & Wein-
gast, Structure-Induced Equilibrium and Legislative Choice, 37 PUB. CHOICE 503, 507-11 (1981)
(institutional arrangements that restrict excessive legislative exchange enhance stability in pure
majoritarian rule legislatures). For a detailed summary of the literature, see Farber & Frickey,
supra note 47, at 425-35.
141. The Framers were keenly aware of the inherent instability of majority rule. In support of
his argument for an extended republic, Madison criticized pure democracies as "spectacles of turbu-
lence and contention." THE FEDERALIST No. 10, at 81 (J. Madison) (C. Rossiter ed. 1961). One of
Madison's major arguments for bicameralism (and in particular, for the Senate) was "the propensity
of all single and numerous assemblies to yield to the impulse of sudden and violent passions .... "
THE FEDERALIST No. 62, at 379 (J. Madison) (C. Rossiter ed. 1961). Federalist No. 62 also speaks
at length about the costs of instability:
The internal effects of a mutable policy are still more calamitous. It poisons the blessing
of liberty itself. It will be of little avail to the people, that the laws are made by men of
their own choice, if the laws be so voluminous that they cannot be read, or so incoherent
that they cannot be understood; if they be repealed or revised before they are promul-
gated, or undergo such incessant changes that no man, who knows what the law is to-day,
can guess what it will be to-morrow. Law is defined to be a rule of action; but how can
that be a rule, which is little known, and less fixed?
In another point of view, great injury results from an unstable government. The want
of confidence in the public councils damps every useful undertaking, the success and profit
of which may depend on a continuance of existing arrangements.... In a word, no great
improvement or laudable enterprise can go forward which requires the auspices of a
steady system of national policy.
But the most deplorable effect of all is that diminution of attachment and reverence
which steals into the hearts of the people towards a political system which betrays so
many marks of infirmity, and disappoints so many of their flattering hopes. No govern-

never been our system of government. It is a basic institutional requirement

of our system of government that the legitimacy of a statute be independent
of the current state of public opinion. Thus, it would be a mistake to add a
proviso to the supremacy principle that "courts must enforce statutes, unless
the latest Gallup poll shows that a statute has clearly lost majority support."
A less radical way to deal with outmoded statutes is for courts to interpret
them consistent with current public mores.1 42 When the statutory command
is unclear, the supremacy principle does not preclude courts from bringing
contemporary values to bear. But when the statutory language and legisla-
tive history make the statute's meaning unmistakable, such "interpretation"
is more in the nature of a partial repeal. Under such circumstances, dynamic
interpretation collapses into the Calabresian argument, 14 3 and it should be
rejected for the same reason.
B. LEGISLATIVE META-INTENT AND DYNAMIC INTERPRETATION
The previous section considered Dean Calabresi's proposal that courts

should have the power to nullify statutory directives that have lost majority
support. This proposal would not be favored by enacting legislators, because
it would substantially decrease the value of legislation to its supporters.
Knowing that majority coalitions are unstable, the members of a coalition
have good reason to want their legislation to survive it.144 Consequently,
Calabresi's proposal would bring courts into conflict not only with the statu-
tory directives, but also with the legislature's own "meta-intent" concerning
how courts should apply statutes.
A rational legislator might, however, favor a rule that would allow courts
to disregard statutory directives under other circumstances.1 45 Our hypo-
thetical military officer, for example, might not want his order carried out if
unforeseen circumstances make the order obviously futile or counterproduc-
ment, any more than an individual, will long be respected without being truly respectable;
nor be truly respectable without possessing a certain portion of order and stability.
Id. at 381-82; see also THE FEDERALIST No. 71, at 431 (A. Hamilton) (C. Rossiter ed. 1961) (strong
executive adds to stability and reduces government's vulnerability to capricious shifts in public
opinion).
142. This is the tack Professor Dworkin takes. See R. DWORKIN, supra note 132, at 340-41, 348-
49 (arguing that political fairness requires that statutory interpretation be based on current public
opinion, not public opinion at the time of enactment). Related arguments for dynamic interpreta-
tion can be found in Aleinikoff, supra note 3, at 46-66; Eskridge, supra note 11, passim; and, in a
more limited form, in Farber & Frickey, supra note 47, at 461-68 (proposing model in which subse-
quent legislative events are relevant factors in interpreting statutes).
143. Calabresi himself suggests that this form of interpretation is undesirable. See G. CALA-
BREST, supra note 108, at 32-39; supra notes 137-38 and accompanying text.
144. See McCubbins, Noll & Weingast, Administrative Procedures as Instruments of Political
Control, 3 J.L. ECON. & ORG. 243, 255 (1987).
145. See generally Aleinikoff, supra note 3, at 56-57.

tive. For instance, it would hardly be an act of insubordination to disregard

an order to shell a hill, if in the meantime the hill had been taken by friendly
forces. Similarly, the legislature might not appreciate having its intended
beneficiaries wiped out by the courts' "friendly fire."
Several private law analogies exist. In contract law, for example, it is well
established that performance of a contract may be excused if performance
becomes impractical, or if changed circumstances frustrate the basic purpose
of the contract. 146 Similarly, under the cy pres doctrine, provisions of a trust
may be nullified or even rewritten by a court when changed circumstances
prevent the original provisions from attaining the donor's goals. 147 To a lit-
eralist, these doctrines might seem to involve judicial disobedience to the di-
rections of the contracting parties or the donor. But this criticism misfires
because these rules are, in fact, those that rational parties would have chosen
to incorporate in their directives if they had considered them. 48
For this reason, a statutory cy pres doctrine would only superficially con-
flict with the supremacy principle. 149 Because the doctrine is one that ra-
146. See U.C.C. § 2-615 (1988) (excuse by failure of presupposed conditions); E. FARNSWORTH,
CONTRACTS § 9.1, at 647, § 9.6, at 677, § 9.7, at 689 (1982) (discussing mistake, impracticability,
and frustration doctrines). Also, there are growing arguments in favor of allowing judicial modifi-
cation of long-term contracts because of changed circumstances. See generally Hillman, CourtAd-
justment of Long-Term Contracts: An Analysis UnderModern Contract Law, 1987 DUKE L.J. 1.
147. According to the Second Restatement of Trusts:
If property is given in trust to be applied to a particular charitable purpose, and it is or
becomes impossible or impracticable or illegal to carry out the particular purpose, and if
the settlor manifested a more general intention to devote the property to charitable pur-
poses, the trust will not fail but the court will direct the application of the property to
some charitable purpose which falls within the general charitable intention of the settlor.
RESTATEMENT (SECOND) OF TRUSTS § 399 (1959). For an interesting discussion of this doctrine,
see Comment, Phantom Selves: The Search for a General CharitableIntent in the Application of the
Cy PresDoctrine, 40 STAN. L. REv. 973 (1988) (by V. Laird).
148. This is the standard "law and economics" view of state supplied contract terms. See R.
POSNER, ECONOMIC ANALYSIS OF LAW 82 (3d ed. 1986); Goetz & Scott, The Limits of Expanded
Choice:An Analysis of the InteractionsBetween Express and Implied Contract Terms, 73 CALIF. L.
REv. 261, 262, 266-67 (1985) [hereinafter Goetz & Scott, Limits of Expanded Choice]; Goetz &
Scott, Principles of Relational Contracts,67 VA. L. REV. 1089, 1089-90 (1981).
149. Although the concept of statutory cy pres has not found explicit recognition in American
law, there are cases that seem to reach similar results. For example, in Lesbian/Gay Freedom Day
Comm., Inc. v. INS, 541 F. Supp. 569 (N.D. Cal. 1982), aff'd sub nom. Hill v. INS, 714 F.2d 1470
(9th Cir. 1983), the court declined to enforce a clear congressional mandate that homosexuals be
excluded from entry into the country because Congress had acted on the outmoded understanding
that homosexuality was a psychiatric disorder. Id. at 584-85. In Li v. Yellow Cab, 13 Cal. 3d 804,
532 P.2d 1226, 119 Cal. Rptr. 858 (1975), the California Supreme Court abrogated a long-codified
doctrine of contributory negligence, replacing it with comparative negligence. The court justified
this apparent step into the legislative sphere by demonstrating that the California statute embodying
contributory negligence was not intended to freeze the development of the negligence doctrine in
the courts, but merely to clarify the extent of that development at the time the statute was enacted.
Id. at 1232-39. Id. at 812-23, 532 P.2d at 1232-39, 119 Cal. Rptr. at 864-71. Li is not a typical
statutory cypres case, because the disputed statute was intended from its inception to be amendable

tional legislators themselves would favor, courts that engage in this type of
dynamic interpretation can properly claim to be implementing rather than
frustrating the legislators' design. For example, technological change might
turn a statute passed for safety reasons into an actual source of increased
danger. Under such circumstances, serious doubt will arise whether the en-
acting legislators intended the statute to apply; thus, disregarding the statute
would not be considered an act of insubordination.
A recent case prompted an intriguing debate among the Justices about the
effect of changed circumstances on statutory directives. K Mart Corp. v.
Cartier,Inc.150 involved an obscure customs regulation governing the impor-
tation of foreign-manufactured, trademarked goods. Section 526 of the Tariff
Act of 1930151 prohibits importing such goods if the trademark is owned by a
United States citizen or corporation.' 52 K Mart involved "gray-market"
goods-that is, goods bearing United States trademarks, but imported with-
out the consent of the trademark holder. The Customs regulation challenged
in K Mart allowed gray-market goods to be imported under certain circum-
stances, 153 such as when the United States owner had authorized the use but
not the importation of the trademark. 154 Justice Kennedy, in his opinion for
the Court, rather brusquely held this exception to be inconsistent with the
155
plain statutory language.
For present purposes, the more interesting discussion is contained in the
separate opinions of Justices Brennan and Scalia, each of whom represented
a faction of four Justices. Justice Brennan, in the dissenting portion of his
opinion, essentially conceded that the plain language of the statute was in-
consistent with an exception for imported goods bearing a trademark that
was authorized by the trademark's United States owner, 156 but argued that
57
Congress never intended section 526's protections to cover such goods.'
Justice Brennan reasoned that trademark law had changed radically in the
fifty years since the statute was passed. When Congress imposed the import
restriction, he suggested, trademarks served only to identify the source of
to future jurisprudential developments. The Li approach is, however, essentially similar to cy pres
in that it involves updating legislative pronouncements in light of developments unforeseen by the
enacting legislators, while still abiding by the legislators' original purpose in enacting the statute.
Cf Langevoort, supra note 3, at 732-33 (courts have engaged in covert statutory updating in bank
regulation cases).
150. 108 S. Ct. 1811 (1988).
151. Tariff Act of 1930, ch. 497, tit. IV, § 526, 46 Stat. 590, 741 (1930) (codified as amended at
19 U.S.C. § 1526 (1982)).
152. 19 U.S.C. § 1526(a).
153. 19 C.F.R. § 133.21(c) (1987).
154. Id. § 133.21(c)(3).
155. 108 S. Ct. at 1817-18.
156. Id. at 1828 (Brennan, J., concurring in part).
157. Id. at 1830.

goods. 158 Under this regime, Manufacturer Y could not legally license an-
other manufacturer to use the "X" trademark; to do so would be "philosoph-
ically impossible since licensing meant that the mark was being used by
persons not associated with the real manufacturing 'source' in a strict, physi-
cal sense of the word."' 59 Hence, the Congress that enacted section 526
could not have imagined a foreign manufacturer's goods bearing United
States trademarks under license, and could not have intended section 526 to
reach such goods. 160 Justice Brennan drew an analogy to a nineteenth-cen-
tury statute requiring ovens to be inspected for their propensity to spew
flames; such a statute, he maintained, was never intended to apply to, and
need not be applied to, microwave or electric ovens. 16 1
Justice Scalia vigorously attacked Brennan's analysis. He argued that sec-
tion 526 applied unambiguously to the case at bar, 162 and that courts should
not "rewrite the United States Code to accord with the unenacted purposes
of Congresses long since called home."' 163 Rather, he argued, it is "the pre-
rogative of each currently elected Congress to allow those laws which change
has rendered nugatory to die an unobserved death if it no longer thinks their
purposes worthwhile; and to allow those laws whose effects have been ex-
panded by change to remain alive if it favors the new effects."1 64
Justice Scalia's argument has several weaknesses. First, Congress does not
have the luxury of "allowing statutes to die an unobserved death." Rather,
Congress must engage in active euthanasia by repealing them. As Justice
Scalia himself has argued elsewhere, the fact that a statute has not been re-
pealed often reflects nothing more than inertia. 165 In K Mart, Scalia's argu-
ment is particularly weak, because Congress might well have assumed that
the agency's regulations were valid, making legislative action to update the
statute unnecessary.
Secondly, Justice Scalia's reference to rewriting statutes to accord with the
"unenacted purposes" of past Congresses is puzzling, because it is precisely
the enacted purpose of the statute that concerned Justice Brennan. Here, it
is important to distinguish Justice Brennan's argument in K Mart from his
argument in Weber. In Weber, Justice Brennan argued that the statute
passed by Congress was a poor way to achieve its antidiscrimination goals.
158. Id. at 1828.

159. Id. (quoting 1 J. MCCARTHY, TRADEMARKS AND UNFAIR COMPETITION 826 (2d ed.
1984)).
160. Id. at 1830.
161. Id.
162. Id. at 1834 (Scalia, J., concurring in part).
163. Id.
164. Id.
165. See, e.g., Johnson v. Transportation Agency, 480 U.S. 616, 671-72 (1987) (Scalia, J.,
dissent-
ing) ("vindication by congressional inaction is a canard").

Justice Scalia's criticism strikes home in this case. But in K Mart, Justice
Brennan argued that Congress had never made any decision about how to
apply the statute to imported goods bearing a United States trademark under
authorization of the United States owner, because such licensing arrange-
ments did not exist at the time of enactment. This argument, based on
changed circumstances, respects rather than flouts the legislature's
prerogatives.
Thirdly, Justice Scalia asserts that section 526's command is unambiguous.
But the very fact that there was doubt whether the legislature intended the
statute to apply in the particular circumstances suggests that the statute is
not entirely unambiguous. Of course, the legislature could have included an
express provision about unforeseen circumstances. In the absence of such an
express provision, however, it seems reasonable for Justice Brennan to have
implied one. It is a commonplace that in private law, courts serve a useful,
legitimate function by implying contractual terms that rational parties them-
selves would agree to if they had unlimited time and resources to draft and
negotiate contract terms. By implying these contractual terms, courts reduce
the transaction costs of contract formation. 166 The courts' provision of such
"off-the-rack" terms is equally appropriate in the statutory setting, in which
Congress, like contracting parties, lacks the time to consider all fact permu-
tations when memorializing its policy decisions.
Justice Scalia is on stronger ground when he suggests that what I have
called the cy pres approach should be implemented only in "clear cases." 167
Between the time a statute is passed and the time it is construed by the
Supreme Court, some, perhaps unforeseen, societal changes will almost al-
ways have occurred. If applied too broadly, statutory cy pres would give
judges carte blanche to rewrite statutes, a result that enacting legislators
would reject. As Justice Scalia suggests, statutory cy pres must be limited to
cases in which "(1) it is clear that the alleged changed circumstances were
unknown to, and unenvisioned by, the enacting legislature, and (2) it is clear
that [the changed circumstances] cause the challenged application of the stat-
ute to exceed its original purpose." 168 A similar analysis should apply, of
course, when changed circumstances cause the statute to fall short of, rather
than exceed, its purpose.
Even when cy pres is permissible, it sometimes may be preferable for a
166. See Goetz & Scott, Limits of Expanded Choice, supra note 148, at 266; D. Baird, Mitigation
Rules and the Limits of Contract (1988) (unpublished manuscript) (copy on file at The Georgetown
Law Journal). Some of the so-called "canons of construction" may, in fact, be such "off-the-rack"
rules.
167. 108 S. Ct. at 1835 (Scalia, J., concurring in part).
168. Id. Sometimes, as may have been true in Weber, legislators may foresee, but prefer not to
think about, later developments. Such legislative avoidance should neither prevent judicial updat-
ing when the statute is unclear nor justify the court's ignoring a statutory directive.

court to decline to update the statute. For example, if the value judgment
involved is particularly controversial, or when the court lacks remedial
power to implement an adequate substitute for the old statute, legislative
updating would be preferable to judicial updating.169 Against this preference
must be considered the likelihood that legislative inertia will prevent any re-
sponse, thereby resulting in clear harm to the public interest or injustice to a
particular litigant. Given the threshold requirements for cy pres and these
prudential concerns, statutory cy pres should not be invoked lightly. Never-
theless, on occasion, it could be a valuable doctrine.
C. META-INTENT AND STARE DECISIS
Sometimes the changed circumstances relevant to statutory interpretation

are not extrinsic to the legal process, but rather consist of actions taken by
judges themselves. As demonstrated in Part II, sometimes courts have been
less than perfect in their fidelity to legislative supremacy. Once the infidelity
has been committed, however, considerations of stare decisis may make the
court reluctant to reverse itself. The result may be that the court in later
cases will continue to interpret a statute in violation of the legislature's
clearly enacted collective understanding. At first blush, this seems to be a
violation of the supremacy principle. If the court is not authorized to amend
the statute, how can a previous erroneous decision suspend the court's obli-
gation to obey the legislature's command?
Recently, several writers have argued against adherence to stare decisis
under these circumstances. 170 One has written: "If courts become instru-
ments by which packages are undone, laws will be harder to pass. Bargains
must be kept to be believed .... As a general matter, this observation is
"171
169. The possibility of administrative updating also weighs against judicial use of statutory cy
pres,because agencies have several advantages over courts in bringing policy considerations to bear
on statutory issues. See Chevron, Inc. v. NRDC, 467 U.S. 837, 865-66 (1984) (court deferred to
agency expertise to define "stationary source" when Congress failed to provide definition); Pierce,
supra note 20, at 307 (agency preferred over courts to make policy because of greater accountability
to electorate).
170. See Cooper, Stare Decisis: Precedentand Principlein ConstitutionalAdjudication, 73 COR-
NELL L. REV. 401, 404-05 (1988) (criticizing the Court in Johnson v. Transportation Agency, 480
U.S. 616 (1980), for effectively repealing title VII under cover of deference to Weber); Maltz, The
Nature of Precedent, 66 N.C.L. REV. 367, 388-92 (1988) (criticizing stare decisis in statutory con-
text; arguing that courts erroneously rely on legislature to correct courts' mistaken interpretations);
Rees, Cathedrals Without Walls: A View from the Outside (Book Review), 61 TEx. L. REV. 347,
374 (1982) (reviewing G.CALABRESI, A COMMON LAW FOR THE AGE OF STATUTES (1982)) (argu-
ing that stare decisis is less appropriate in statutory than common law decisions, because people rely
on what statutes say, not on courts' interpretations of what statutes say).
171. Easterbrook, Stability and Reliability in JudicialDecisions, 73 CORNELL L. REV.422, 429
(1988). The idea of viewing statutes as legislative deals originated in Landes & Posner, The In-
dependent Judiciaryin an Interest Group Perspective, 18 J.L. & ECON. 875, 877-85 (1976); see also
Morrison-Knudsen Constr. Co. v. Director of Workers' Compensation Programs, 461 U.S. 624,

quite correct. It may not have as much force, however, when the issue is
framed as whether courts should enforce the statutory "bargain" even at the
cost of overruling precedent.
The crucial question with respect to the supremacy principle is the enact-
ing legislators' meta-intent regarding the priority of stare decisis. If a ra-
tional enacting legislator would favor stare decisis even when an initial
decision violates the supremacy principle, then a rule of stare decisis is con-
sistent with the supremacy principle. The court cannot be accused of insub-
ordination if stare decisis is an appropriate "implied term" in statutes.172
In analyzing this problem, it is important to recall that we are considering
what rule legislators would approve at the time of enactment. At that time,
legislators know that judges may make serious mistakes, but they do not
know the direction of the mistakes. Supporters of the legislation have no
way of knowing whether judicial mistakes will favor them (by giving them
more than their original "bargain"), or injure them (by giving them less than
they bargained for). This is not to say that legislators would be indifferent to
the possibility of serious judicial errors. It does mean, however, that their
interest in having these errors corrected by subsequent decisions is somewhat
reduced by the possibility that honest judicial mistakes might redound to
their benefit.
Against this interest, legislators also must balance the potentially serious
social costs of legal instability, including the potential damage to reliance
interests.7 3 Sometimes, the balance will weigh in favor of stare decisis. In
other cases, the legislators' interest in legal stability may be outweighed by
the importance of fidelity to their original intentions. Thus, as a matter of
meta-intent, legislators presumably would approve of a rule giving some
weight to stare decisis, even when the result might be to entrench a case that
violates the supremacy principle. 174
635-36 (1983) (updating statute in accordance with "current trends" would upset original statutory
deal).
172. Legislatures may sometimes wish to opt out of this implied term. One way of doing so
would be to vest interpretive authority in an administrative agency, because agencies are typically
not bound by stare decisis and may change their positions in light of new developments or even
political shifts. See Garland, Deregulation and JudicialReview, 98 HARV. L. REV. 507, 584-85 &
nn. 456-57 (1985) (change of administrations may justify change in agency's emphasis).
173. Legislators have several reasons to be concerned regarding the prospect of public uncer-
tainty about judicial interpretations. If enacting legislators are risk-averse, they may be reluctant to
gamble on the direction of judicial error. Social costs, such as increased litigation and difficulty in
planning transactions, may be associated with uncertainty about how statutes will be construed.
Jettisoning stare decisis, however, would do relatively little to reduce these various uncertainty
costs. Moreover, a rule allowing ready judicial correction of prior mistaken opinions would create a
variety of deadweight social costs. In general, then, rational enacting legislators would probably
prefer that courts give strong weight to stare decisis in statutory cases, even at the expense of fidelity
to the original legislative deal.
174. This argument is developed at greater length in Farber, supra note 14, at 11-14. An alterna-

D. LEGISLATIVE META-INTENT AND WEBER
As discussed earlier, the most interesting arguments in favor of Weber rely

on post-enactment events. The first such argument is that informed opinion
on affirmative action has shifted to the extent that title VII requires updating.
However, enacting legislators have little reason to want statutory directives
to be made contingent on the state of public opinion. Such changes might
well be relevant if the meaning of the statute were subject to genuine doubt,
but they cannot release the court from compliance with a clear statutory
17 5
command.
The second argument is that events after 1964 demonstrated that the clear
congressional mandate to eradicate racial discrimination could not be exe-
cuted without permitting remedial affirmative action. 176 This argument,
however, does not justify the kind of radical surgery performed by the Weber
Court to legitimize nonremedial affirmative action under title VII. In Weber,
Justice Brennan replaced Congress' solution to the problem of racial inequal-
ity (eliminating discrimination) with his own solution (preferential hiring
even in the absence of any evidence whatsoever of prior illegal discrimina-
tion). Justice Brennan's approach to title VII cannot be justified even under
the statutory cy pres doctrine, because events after 1964 were not such as to
cause enforcement of title VII to work against the purposes of its drafters,
nor did these events result in extending title VII's prohibitions to an area that
Congress had never contemplated. In fact, post-1964 events merely revealed
that perhaps Congress did not, by enacting title VII, choose the best strata-
gem to achieve its goals. But, as argued throughout this article, the
supremacy principle does not allow courts to rechoose strategies for
Congress.
The third argument is based on stare decisis and is actually an argument in
177
favor of Weber's progeny rather than of the original decision in Weber.
tive formulation of the argument would classify stare decisis as one of several "background under-
standings" that Congress takes for granted when it legislates. See Stewart & Sunstein, Public
ProgramsandPrivateRights, 95 HARV. L. REv. 1195, 1200 (1982) (discussing various understand-
ings among government entities that promote effective governance). For another perspective on
stare decisis in statutory cases, see Eskridge, Overruling Statutory Precedents, 76 GEO. L.J. 1361,
1385 (1988) (arguing that statutory precedents not be given stronger presumption against over-
ruling than other precedents).
175. See supra notes 133-42.
176. See supra notes 128-31 and accompanying text; see also Eskridge, supra note 11, at 1492
(noting that current equal opportunities for minorities do not remedy previous effects of discrimina-
tion). Additionally, it is unclear whether affirmative action is an effective way of raising the eco-
nomic status of blacks. For a survey of the relevant literature, see Smith & Welch, Black Economic
Progress after Myrdal, 27 J. EcON. LIT. 519, 552-55 (1989) (affirmative action reallocated blacks
between firms and caused a short-term rise in wages but no long-term wage increase).
177. See Johnson v. Transportation Agency, 480 U.S. 616, 644 (1987) (Stevens, J., concurring)
(agreeing with Weber's interpretation of title VII on grounds of stare decisis).

Following Weber, not a single bill was introduced in Congress to overturn

the decision, and it has now been the law for over ten years. 178 In addition,
Weber has been the subject of substantial reliance by employers and unions.
Even if the Court were to immunize those who have relied on Weber from
damage suits based on past affirmative action plans, abolishing the plans
would upset many careers, and granting remedies to white male "victims"
would disrupt seniority rights and pension plans. 179 Moreover, judicial waf-
fling on such a crucial issue could undermine public confidence in the stabil-
ity of the legal system. Under the circumstances, then, although the broad
holding of Weber violated the supremacy principle, that principle does not
require that the decision be overruled. 180
IV. CONCLUSION
Judges must, except when exercising the power of judicial review, defer to
the decisions of legislatures. This principle of legislative supremacy seems
obvious on its face, but its meaning is surprisingly elusive.
The supremacy principle is actually quite narrow. Under the principle,
courts engaged in statutory interpretation must consider both the language
and legislative history of a statute. If, when taken together, they leave no
room for doubt about the statute's meaning, courts may not interpret the
statute contrary to that meaning. This principle does not prevent courts
from going beyond (as opposed to "against") even clear statutory language
and legislative intent. Nor does it seriously constrain methods of statutory
interpretation when a statute's language and legislative intent are unclear. In
such cases, the supremacy principle allows a court to consider any additional
factors it deems appropriate, including its own view of public policy.' 81
The supremacy principle is further qualified by considerations of legisla-
178. Id. at 629 n.7. On the subject of legislative silence, see Eskridge, InterpretingLegislative
Inaction, 87 MICH. L. REV. 67 passim (1988); Farber, supra note 14, at 8-14. Although these two
articles advocate different theories, their practical implications appear quite similar: congressional
silence can be a factor in favor of retaining a Supreme Court ruling.
179. More limited, remedial forms of affirmative action might be permissible under the statute.
See supra text accompanying notes 128-30. Hence, if Weber were reversed, the result would not be
simply to invalidate all affirmative action programs; rather it would be necessary to invest substan-
tial legal resources into investigating each individual program that passed the Weber test to see if it
conformed with the new requirements for permissible remedial affirmative action under title VII.
180. Even some strong opponents of affirmative action are now convinced that Weber is too well-
entrenched to uproot. See Glazer, The Affirmative Action Stalemate, 90 PUB. INT. 99, 105-11
(1988). For a more complete discussion of why it would be undesirable to overrule Weber, see
Eskridge, supra note 174, at 1409-14. On the other hand, although the stare decisis argument for
allowing nonremedial affirmative action is fairly strong, it may be weaker than the argument for
remedial affirmative action. Thus, stare decisis might allow but not require adherence to Weber's
endorsement of nonremedial plans.
181. See supra text accompanying notes 9-30. This conception of the judicial role need not entail
judicial activism. See Posner, The Jurisprudence of Skepticism, 86 MICH. L. REv. 827, 862-64

tive meta-intent. Courts can read "off-the-rack" rules of interpretation into

congressional enactments, provided those rules would be favored by rational
enacting legislators. These rules can operate to allow courts to violate what
otherwise appear to be clear statutory directives. For example, the occur-
rence of significant events unforeseen at the time of a statute's enactment
may make enforcement of that statute counterproductive or pointless. If so,
the court can apply a statutory version of the cy pres doctrine without violat-
ing the supremacy principle.1 82 In another situation, a court may refuse to
overrule an earlier precedent that violates the supremacy principle because
rational legislators would favor stare decisis in statutory cases, even though it
may entail some sacrifice of their original goals in passing the statute.1 83
This article has focused on how statutory interpretation is constrained by
the subordinate position of courts with respect to legislatures. The argument
has been that the constraint is limited, but, within its bounds, very powerful.
Once outside those bounds, we must look elsewhere to understand judicial
deference to legislatures. Deference is not, after all, necessarily a matter of
subordination.1 84 Indeed, preferences about the locus of decision can run
counter to supremacy relationships, as when the Supreme Court defers to the
discretion of a trial judge who is in every sense subordinate to the Justices.
Similarly, quite apart from the supremacy principle, on occasion courts may
have good reasons to defer to legislatures.
Like all metaphors, that of the court as the legislature's agent has both its
uses and its limitations. The agent metaphor is useful to illuminate the ways
in which statutes constrain courts. It does little, however, to illuminate other
aspects of the relationship between courts and legislatures, such as the extent
to which courts should leave to the legislature the initiative in creating public
policy.1 85 The metaphor of supremacy can only take us so far. But within its
limits, it can shed light on some crucial aspects of the judicial role.
(1988) (framers of statutory and constitutional provisions know and expect that judges will bring
personal policy preferences to bear on interpretation of open-ended language).
182. The fact that statutory cypres is consistent with the supremacy principle does not necessar-
ily establish that it is a desirable policy. The benefits of a statutory cy pres doctrine may be out-
weighed by its costs as an additional source of legal unpredictability, or by the fact that the doctrine
may be too susceptible to abuse. If the doctrine is limited in the ways suggested earlier, however,
see supra text accompanying notes 167-68, these drawbacks do not seem substantial.
183. See supra text accompanying notes 172-74.
184. When the Supreme Court says a problem should be resolved by state or local governments
rather than the federal courts, it does not necessarily mean that it considers itself subordinate to the
state legislatures or city councils.
185. For purposes other than supremacy principle analysis, it may be better to think of the court
as the legislature's partner, each having its own area of primary responsibility. The partnership
metaphor suggests, for example, that courts should feel free to make public policy in traditionally
common law areas. For an insightful attempt to explore the complex judicial role in statutory
interpretation, see Eskridge & Frickey, PracticalReason in Statutory Interpretation,42 STAN. L,
REv. - (forthcoming 1990).

Pao Law Irr

Uploaded by

Copyright:

Available Formats

Pao Law Irr

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pao Law Irr

Uploaded by

Copyright:

Available Formats

Berkeley Law

Berkeley Law Scholarship Repository

Statutory Interpretation and Legislative Supremacy

Follow this and additional works at: http://scholarship.law.berkeley.edu/facpubs

"The general contemporary American view of statutory interpretation is

HeinOnline -- 78 Geo. L. J. 281 1989-1990

subordinate role means anything at all, it must somehow constrain judges

HeinOnline -- 78 Geo. L. J. 282 1989-1990

I. THE SCOPE OF THE SUPREMACY PRINCIPLE

Legislative supremacy, as a doctrine of statutory interpretation, is

HeinOnline -- 78 Geo. L. J. 283 1989-1990

A. THE STRONG CONCEPTION OF SUPREMACY

Perhaps the most obvious understanding of legislative supremacy is that

HeinOnline -- 78 Geo. L. J. 284 1989-1990

HeinOnline -- 78 Geo. L. J. 285 1989-1990

HeinOnline -- 78 Geo. L. J. 286 1989-1990

B. THE WEAK CONCEPTION OF SUPREMACY

HeinOnline -- 78 Geo. L. J. 287 1989-1990

32. U.S. CONST. art. II, § 1.

HeinOnline -- 78 Geo. L. J. 288 1989-1990

HeinOnline -- 78 Geo. L. J. 289 1989-1990

legislative history shows that the legislature had contrary intentions. 44

HeinOnline -- 78 Geo. L. J. 290 1989-1990

Even assuming that the legislature possesses an ascertainable institutional

HeinOnline -- 78 Geo. L. J. 291 1989-1990

tion's meaning, there can be no question of obedience or disobedience.5 4

C. DEMOCRACY AND DISOBEDIENCE

The foregoing analysis suggests this formulation of the supremacy princi-

HeinOnline -- 78 Geo. L. J. 292 1989-1990

HeinOnline -- 78 Geo. L. J. 293 1989-1990

goes." On the contrary, judicial decisionmaking may be tightly constrained

II. JUDICIAL IMPLEMENTATION OF THE SUPREMACY PRINCIPLE

A. THE PARADIGM CASE

59. 437 U.S. 153 (1978).

HeinOnline -- 78 Geo. L. J. 294 1989-1990

Powell, in dissent, suggested that Congress really only "intended to require

65. Id. at 207-08 (Powell, J., dissenting).

HeinOnline -- 78 Geo. L. J. 295 1989-1990

HeinOnline -- 78 Geo. L. J. 296 1989-1990

supremacy principle: "Our individual appraisal of the wisdom or unwisdom

HeinOnline -- 78 Geo. L. J. 297 1989-1990

Dworkin's argument is plausible because of suspicions that the legislators

B. VIOLATIONS OF THE SUPREMACY PRINCIPLE

Tennessee Valley Authority v. Hill is the paradigm of judicial adherence to

HeinOnline -- 78 Geo. L. J. 298 1989-1990

HeinOnline -- 78 Geo. L. J. 299 1989-1990

emy of Sciences, and the absence of an indication of the statistical reliabil-

92. Id. at 648.

HeinOnline -- 78 Geo. L. J. 300 1989-1990

HeinOnline -- 78 Geo. L. J. 301 1989-1990

allows employers to engage voluntarily in affirmative action favoring minor-

104. H.R. REP. No. 294, supra note 103, at 337.

HeinOnline -- 78 Geo. L. J. 302 1989-1990

HeinOnline -- 78 Geo. L. J. 303 1989-1990

against considering race in employment decisions. Section 703(a)(2), for ex-

HeinOnline -- 78 Geo. L. J. 304 1989-1990

permit" preferential treatment.1 25 As even those who support Weber con-

125. 443 U.S. at 205-06.