Pennsylvania v. Union Gas Co.

491 U.S. 1

SCOTUS

1989

Check Treatment

PENNSYLVANIA v. UNION GAS CO.

No. 87-1241

CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT

Argued October 31, 1988-Decided June 15, 1989

John G. Knorr III, Chief Deputy Attorney General of Pennsylvania, argued the cause for petitioner. With him on the briefs were LeRoy S. Zimmerman, Attorney General, and Gregory R. Neuhauser, Senior Deputy Attorney General.

Robert A. Swift argued the cause for respondent. With him on the brief were Marguerite R. Goodman and Lawrence Demase.*

*A brief of amici curiae urging reversal was filed for the State of New York et al. by Robert Abrams, Attorney General of New York, O. Peter Sherwood, Solicitor General, Elaine Gail Suchman, Assistant Attorney General, John K. Van de Kamp, Attorney General of California, Clifford L. Rechtschaffen and J. Matthew Rodriquez, Deputy Attorneys General, Joseph I. Lieberman, Attorney General of Connecticut, Kenneth N. Tedford, Assistant Attorney General, Michael Bowers, Attorney General of Georgia, Barbara H. Gallo, Assistant Attorney General, Neil F. Hartigan, Attorney General of Illinois, Rosalyn Kaplan, Linley E. Pearson, Attorney General of Indiana, Harry John Watson III, Thomas J. Miller, Attorney General of Iowa, John P. Sarcone, Assistant Attorney General, Arthur L. Williams, J. Joseph Curran, Jr., Attorney General of Maryland, Andrew H. Baida, Richard M. Hall, and Michael C. Powell, Assistant Attorneys General, Hubert H. Humphrey III, Attorney General of Minnesota, John R. Tunheim, Chief Deputy Attorney General, William L. Webster, Attorney General of Missouri, Shelley A. Woods, Assistant Attorney General, W. Cary Edwards, Attorney General of New Jersey, John J. Maiorana, Deputy Attorney General, Hal Stratton, Attorney General of New Mexico, Alicia Mason, Assistant Attorney General, Lacy H. Thornburg, Attorney General of North Carolina, Gayl M. Manthei, Assistant Attorney General, Robert H. Henry, Attorney General of Oklahoma, Sara J. Drake, Assistant Attorney General, T. Travis Medlock, Attorney General of South Carolina, Walton J. McLeod III, Jacquelyn S. Dickman, W. J. Michael Cody, Attorney General of Tennessee, Michael W. Catalano, Deputy Attorney General, David L. Wilkinson, Attorney General of Utah, Fred G. Nelson, Assistant Attorney General, Jeffrey L. Amestoy, Attorney General of Vermont, Conrad W. Smith, Assistant Attorney General, Charles G. Brown, Attorney General of West Virginia, C. William Ullrich, First Deputy Attorney General, Donald J. Hanaway, Attorney General of Wisconsin, and Charles D. Hoornstra, Assistant Attorney General.

Briefs of amici curiae urging affirmance were filed for the American Federation of Labor and Congress of Industrial Organizations by Robert M. Weinberg and Laurence Gold; for the Association of American Publishers, Inc., et al. by Jon Baumgarten, Christopher A. Meyer, and Charles S. Sims; for the Chemical Manufacturers Association by Neil J. King and Carol F. Lee; and for the National Music Publishers’ Association, Inc., et al. by Steven B. Rosenfeld.

Ronald A. Zumbrun and Robin L. Rivett filed a brief for the Pacific Legal Foundation as amicus curiae.

JUSTICE BRENNAN announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I and II, and an opinion with respect to Part III, in which JUSTICE MARSHALL, JUSTICE BLACKMUN, and JUSTICE STEVENS join.

This case presents the questions whether the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 42 U.S. C. §9601 et seq., as amended by the Superfund Amendments and Reauthorization Act of 1986 (SARA), Pub. L. 99-499, 100 Stat. 1613, permits a suit for monetary damages against a State in federal court and, if so, whether Congress has the authority to create such a cause of action when legislating pursuant to the Commerce Clause. The answer to both questions is “yes.”

I

For about 50 years, the predecessors of respondent Union Gas Co. operated a coal gasification plant near Brodhead Creek in Stroudsburg, Pennsylvania, which produced coal tar as a by-product. The plant was dismantled around 1950. A few years later, Pennsylvania took part in major flood-control efforts along the creek. In 1980, shortly after acquiring easements to the property along the creek, the Commonwealth struck a large deposit of coal tar while excavating the creek. The coal tar began to seep into the creek, and the Environmental Protection Agency determined that the tar was a hazardous substance and declared the site the Nation‘s first emergency Superfund site. Working together, Pennsylvania and the Federal Government cleaned up the area, and the Federal Government reimbursed the State for clean-up costs of $720,000.

To recoup these costs, the United States sued Union Gas under §§ 104 and 106 of CERCLA, 42 U. S. C. §§ 9604 and 9606, claiming that Union Gas was liable for such costs because the company and its predecessors had deposited coal tar into the ground near Brodhead Creek. Union Gas filed a third-party complaint against Pennsylvania, asserting that the Commonwealth was responsible for at least a portion of the costs because it was an “owner or operator” of the hazardous-waste site, 42 U. S. C. § 9607(a), and because its flood-control efforts had negligently caused or contributed to the release of the coal tar into the creek. The District Court dismissed the complaint, accepting Pennsylvania‘s claim that its Eleventh Amendment immunity barred the suit. A divided panel of the Court of Appeals for the Third Circuit affirmed, finding no clear expression of congressional intent to hold States liable in monetary damages under CERCLA.

United States v. Union Gas Co., 792 F. 2d 372 (1986).

While Union Gas’ petition for certiorari was pending, Congress amended CERCLA by passing SARA. We granted certiorari, vacated the Court of Appeals’ opinion, and remanded for reconsideration in light of these amendments.

479 U. S. 1025 (1987). On remand, the Court of Appeals held that the language of CERCLA, as amended, clearly rendered States liable for monetary damages and that Congress had the power to do so when legislating pursuant to the Commerce Clause. . We granted certiorari, , and now affirm.

II

Hans v. Louisiana, 134 U. S. 1 (1890), this Court held that the principle of sovereign immunity reflected in the Eleventh Amendment rendered the States immune from suits for monetary damages in federal court even where jurisdiction was premised on the presence of a federal question. Congress may override this immunity when it acts pursuant to the power granted it under §5 of the Fourteenth Amendment, but it must make its intent to do so “unmistakably clear.” See . Before turning to the question whether Congress possesses the same power of abrogation under the Commerce Clause, we must first decide whether CERCLA, as amended by SARA, clearly expresses an intent to hold States liable in damages for conduct described in the statute. If we decide that it does not, then we need not consider the constitutional question.

CERCLA both provides a mechanism for cleaning up hazardous-waste sites, 42 U. S. C. §§ 9604, 9606 (1982 ed. and Supp. IV), and imposes the costs of the cleanup on those responsible for the contamination, § 9607. Two general terms, among others, describe those who may be liable under CERCLA for the costs of remedial action: “persons” and “owners or operators.” § 9607(a). “States” are explicitly included within the statute‘s definition of “persons.” § 9601(21). The term “owner or operator” is defined by reference to certain activities that a “person” may undertake. § 9601(20)(A).

Section 101(20)(D) of SARA excludes from the category of “owners or operators” States that “acquired ownership or control involuntarily through bankruptcy, tax delinquency, abandonment, or other circumstances in which the government involuntarily acquires title by virtue of its function as sovereign.” § 9601(20)(D).¹ However, §101(20)(D) continues, “[t]he exclusion provided under this paragraph shall not apply to any State or local government which has caused or contributed to the release or threatened release of a hazardous substance from the facility, and such a State or local government shall be subject to the provisions of this chapter in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under section 9607 of this title.” Ibid. The express inclusion of States within the statute‘s definition of “persons,” and the plain statement that States are to be considered “owners or operators” in all but very narrow circumstances, together convey a message of unmistakable clarity: Congress intended that States be liable along with everyone else for cleanup costs recoverable under CERCLA. Section 101(20)(D) is an express acknowledgment of Congress’ background understanding-evidenced first in its inclusion of States as “persons“-that States would be liable in any circumstance described in § 107(a) from which they were not expressly excluded. The “exclusion” furnished to the States in § 101(20)(D) would be unnecessary unless such a background understanding were at work.²

The plain language of another section of the statute reinforces this conclusion. Section 107(d)(2) of CERCLA, as set forth in 42 U. S. C. §9607(d) (2) (1982 ed., Supp. IV), headed “State and local governments,” provides: “No State or local government shall be liable under this subchapter for costs or damages as a result of actions taken in response to an emergency created by the release or threatened release of a hazardous substance generated by or from a facility owned by another person. This paragraph shall not preclude liability for costs or damages as a result of gross negligence or inten-tional misconduct by the State or local government.” This section is, needless to say, an explicit recognition of the potential liability of States under this statute; Congress need not exempt States from liability unless they would otherwise be liable. Similarly, unless suits against the States were elsewhere permitted, Congress would have had no reason to specify that citizen suits-as opposed to the kind of lawsuit involved here-could be brought “against any person (including the United States and any other governmental instrumentality or agency, to the extent permitted by the eleventh amendment to the Constitution).” 42 U. S. C. §9659(a)(1). The reservation of States’ rights under the Eleventh Amendment would be unnecessary if Congress had not elsewhere in the statute overridden the States’ immunity from suit.

It is also highly significant that, in §101(20)(D), Congress used language virtually identical to that it chose in waiving the Federal Government‘s immunity from suits for damages under CERCLA. Section 120(a)(1) of CERCLA, as set forth in 42 U. S. C. §9620(a)(1), provides: “Each department, agency, and instrumentality of the United States (including the executive, legislative, and judicial branches of government) shall be subject to, and comply with, this chapter in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under section 9607 of this title.” This is doubtless an “unequivoca[l] express[ion]” of the Federal Government‘s waiver of its own sovereign immunity,

United States v. Testan, 424 U. S. 392, 399 (1976), quoting , since we cannot imagine any other plausible explanation for this unqualified language. It can be no coincidence that in describing the potential liability of the States in §101(20)(D), Congress chose language mirroring that of § 120(a)(1). In choosing this mirroring language in §101(20)(D), therefore, Congress must have intended to override the States’ immunity from suit, just as it waived the Federal Government‘s immunity in § 120(a)(1).

This cascade of plain language does not, however, impress Pennsylvania. In the face of such clarity, the Commonwealth bravely insists that CERCLA merely makes clear that States may be liable to the United States, not that they may be liable to private entities such as Union Gas. The Commonwealth relies principally on this Court‘s decision in

Employees v. Missouri Dept. of Public Health and Welfare, 411 U. S. 279 (1973). We held there that Congress had not abrogated the States’ immunity from suit in the Fair Labor Standards Act. Nevertheless, we found, the statute‘s explicit inclusion of state-run hospitals among those to whom the law would apply was not meaningless: since the statute allowed the United States to sue, the inclusion of States within the entities covered by the statute served to permit suits by the United States against the States. Id., at 285-286.

Although it is true that the inclusion of States within CERCLA‘s definition of “persons” would not be rendered meaningless if we held that CERCLA did not subject the States to suits brought by private citizens, it is equally certain that such a holding would deprive the last portion of §101(20)(D) of all meaning. Congress would have had no cause to stress that States would be liable “to the same extent... as any nongovernmental entity,” §101(20)(D), if it had meant only that they could be liable to the United States. In

United States v. Mississippi, 380 U. S. 128, 140-141 (1965), we recognized that the Constitution presents no barrier to lawsuits brought by the United States against a State. For purposes of such lawsuits, States are naturally just like “any nongovernmental entity“; there are no special rules dictating when they may be sued by the Federal Government, nor is there a stringent interpretive principle guiding construction of statutes that appear to authorize such suits. Indeed, this Court has gone so far as to hold that no explicit statutory authorization is necessary before the Federal Government may sue a State. See . Unless Congress intended to permit suits brought by private citizens against the States, therefore, the highly specific language of §101(20)(D) was unnecessary.

The same can be said about the clause of § 101(20)(D) specifying that States would be subject to CERCLA‘s provisions, “including liability under section 9607 of this title.” Section 9607 provides for liability in damages, and liability in damages is considered a special remedy, requiring special statutory language, only where the States’ immunity from suits by private citizens is involved. In light of § 101(20)(D)‘s very precise language, it would be exceedingly odd to interpret this provision as merely a signal that the United States—rather than private citizens—could sue the States for damages under CERCLA.³

Moreover, § 101(20)(D) does not, as Pennsylvania suggests, render States liable only if they acquire property involuntarily and then contribute to a release of harmful substances at that property. Section 101(20)(D) obviously explains and qualifies the entire definition of “owner or operator“-not

just that part of the definition applicable to involuntary owners.

Nor can it be decisive that § 101(20)(D) mentions local governments as well as States. The Commonwealth argues that, because local governments do not enjoy immunity from suit, § 101(20)(D)‘s reference to local governments means that the section shows no intent to abrogate States’ immunity. It was natural, however, for Congress to describe the potential liability of States and local governments in the same breath, since both are governmental entities and both enjoy special exemptions from liability under CERCLA. See §§ 101(20)(D), 107(d)(2). Pennsylvania also argues that § 101(20)(D) demonstrates no intent to hold the States liable because this provision limits the States’ liability. It is true that this section rescues the States from liability where they obtained ownership of cleanup sites involuntarily. The Commonwealth fails to grasp, however, that a limitation of liability is nonsensical unless liability existed in the first place.

We thus hold that the language of CERCLA as amended by SARA clearly evinces an intent to hold States liable in damages in federal court.¹

III

Our conclusion that CERCLA clearly permits suits for money damages against States in federal court requires us to decide whether the Commerce Clause grants Congress the power to enact such a statute. Pennsylvania argues that the principle of sovereign immunity found in the Eleventh

Amendment precludes such congressional authority. We do not agree.

A

Though we have never squarely resolved this issue of congressional power, our decisions mark a trail unmistakably leading to the conclusion that Congress may permit suits against the States for money damages. The trail begins with

Parden v. Terminal Railway of Alabama Docks Dept., 377 U. S. 184 (1964). There, in responding to a state-owned railway‘s argument that Congress had no authority to subject the railway to suit, we concluded that “the States surrendered a portion of their sovereignty when they granted Congress the power to regulate commerce,” id., at 191, and that “[b]y empowering Congress to regulate commerce, ... the States necessarily surrendered any portion of their sovereignty that would stand in the way of such regulation,” id., at 192. Although it is true that we have referred to Parden as a case involving a waiver of immunity, , the statements quoted above lay a firm foundation for the argument that Congress’ authority to regulate commerce includes the authority directly to abrogate States’ immunity from suit.

The path continues in

Employees v. Missouri Dept. of Public Health and Welfare, 411 U. S., at 286, in which we again acknowledged, quoting Parden, that “the States surrendered a portion of their sovereignty when they granted Congress the power to regulate commerce.” Although we declined “to extend Parden to cover every exercise by Congress of its commerce power,” we did so in Employees itself only because “the purpose of Congress to give force to the Supremacy Clause by lifting the sovereignty of the States and putting the States on the same footing as other employers [was] not clear.” . Employees’ message is plain: the power to regulate commerce includes the power to override States’ immunity from suit, but we will not conclude that Congress has overridden this immunity unless it does so clearly.

Since Employees, we have twice assumed that Congress has the authority to abrogate States’ immunity when acting pursuant to the Commerce Clause. See

Welch v. Texas Dept. of Highways and Public Transportation, 483 U. S. 468, 475-476, and n. 5 (1987); . See also (“States may not be sued in federal court... unless Congress, pursuant to a valid exercise of power, unequivocally expresses its intent to abrogate the immunity“); (referring to congressional power recognized in Employees as power “to abrogate Eleventh Amendment immunity“).

It is no accident, therefore, that every Court of Appeals to have reached this issue has concluded that Congress has the authority to abrogate States’ immunity from suit when legislating pursuant to the plenary powers granted it by the Constitution. See, e. g.,

United States v. Union Gas Co., 832 F. 2d 1343 (CA3 1987) (case below); , cert. denied, ; , cert. denied, ; ; .

Even if we never before had discussed the specific connection between Congress’ authority under the Commerce Clause and States’ immunity from suit, careful regard for precedent still would mandate the conclusion that Congress has the power to abrogate immunity when exercising its plenary authority to regulate interstate commerce. In

Fitzpatrick v. Bitzer, supra, we held that Congress may subject States to suits for money damages in federal court when legislating under §5 of the Fourteenth Amendment, and further held that Congress had done so in the 1972 Amendments to Title VII of the Civil Rights Act of 1964. Subsequent cases hold firmly to the principle that Congress can override States’ immunity under §5. See, e. g., ; ; ; .

Fitzpatrick‘s rationale is straightforward: “When Congress acts pursuant to §5, not only is it exercising legislative authority that is plenary within the terms of the constitutional grant, it is exercising that authority under one section of a constitutional Amendment whose other sections by their own terms embody limitations on state authority.”

427 U. S., at 456. In so reasoning, we emphasized the “shift in the federal-state balance” occasioned by the Civil War Amendments, id., at 455, and in particular quoted extensively from . The following passage from Ex parte Virginia is worth quoting here as well:

“Such enforcement [of the prohibitions of the Fourteenth Amendment] is no invasion of State sovereignty. No law can be, which the people of the States have, by the Constitution of the United States, empowered Congress to enact. [I]n exercising her rights, a State cannot disregard the limitations which the Federal Constitution has applied to her power. Her rights do not reach to that extent. Nor can she deny to the general government the right to exercise all its granted powers, though they may interfere with the full enjoyment of rights she would have if those powers had not been thus granted. Indeed, every addition of power to the general government involves a corresponding diminution of the governmental powers of the States. It is carved out of them.” Id., at 346, quoted in Fitzpatrick, supra, at 454-455.

Each of these points is as applicable to the Commerce Clause as it is to the Fourteenth Amendment. Like the Fourteenth Amendment, the Commerce Clause with one hand gives power to Congress while, with the other, it takes power away from the States. It cannot be relevant that the Four-teenth Amendment accomplishes this exchange in two steps (§§ 1-4, plus §5), while the Commerce Clause does it in one. The important point, rather, is that the provision both expands federal power and contracts state power; that is the meaning, in fact, of a “plenary” grant of authority, and the lower courts have rightly concluded that it makes no sense to conceive of §5 as somehow being an “ultraplenary” grant of authority. See, e. g.,

In re McVey Trucking, supra, at 316. See also (distinguishing Employees (Commerce Clause) from Fitzpatrick (§5) only by reference to the clarity of the congressional intent expressed in the relevant statutes).

Pennsylvania attempts to bring this case outside Fitzpatrick by asserting that “[t]he Fourteenth Amendment... alters what would otherwise be the proper constitutional balance between federal and state governments.” Brief for Petitioner 39. The Commonwealth believes, apparently, that the “constitutional balance” existing prior to the Fourteenth Amendment did not permit Congress to override the States’ immunity from suit. This claim, of course, begs the very question we face.

For its part, JUSTICE SCALIA‘s opinion casually announces: “Nothing in [Fitzpatrick‘s] reasoning justifies limitation of the principle embodied in the Eleventh Amendment through appeal to antecedent provisions of the Constitution.” Post, at 42. The operative word here is, it would appear, “antecedent“; and it is important to emphasize that, according to JUSTICE SCALIA, the Commerce Clause is antecedent, not to the Eleventh Amendment, but to “the principle embodied in the Eleventh Amendment.” But, according to Part II of JUSTICE SCALIA‘s opinion, this “principle” has been with us since the days before the Constitution was ratified—since the days, in other words, before the Commerce Clause. In describing the “consensus that the doctrine of sovereign immunity... was part of the understood background against which the Constitution was adopted, and which its jurisdic-tional provisions did not mean to sweep away,” post, at 31-32, JUSTICE SCALIA clearly refers to a state of affairs that existed well before the States ratified the Constitution. JUSTICE SCALIA, therefore, has things backwards: it is not the Commerce Clause that came first, but “the principle embodied in the Eleventh Amendment” that did so. Antecedence takes this case closer to, not further from, Fitzpatrick.

Even if “the principle embodied in the Eleventh Amendment” made its first appearance at the same moment as the Commerce Clause, and not before, JUSTICE SCALIA could no longer rely on chronology in distinguishing Fitzpatrick. Only if it were the Eleventh Amendment itself that introduced the principle of sovereign immunity into the Constitution would the Commerce Clause have preceded this principle. Even then, the order of events would matter only if the Amendment changed things; that is, it would matter only if, before the Eleventh Amendment, the Commerce Clause did authorize Congress to abrogate sovereign immunity. But if Congress enjoyed such power prior to the enactment of this Amendment, we would require a showing far more powerful than JUSTICE SCALIA can muster that the Amendment was intended to obliterate that authority. The language of the Eleventh Amendment gives us no hint that it limits congressional authority; it refers only to “the judicial power” and forbids “constru[ing]” that power to extend to the enumerated suits—language plainly intended to rein in the Judiciary, not Congress. It would be a fragile Constitution indeed if subsequent amendments could, without express reference, be interpreted to wipe out the original understanding of congressional power.

JUSTICE SCALIA attempts to avoid the pull of our prior decisions by claiming that Hans answered this constitutional question over 100 years ago. Because Hans was brought into federal court via the Judiciary Act of 1875 and because the Court there held that the suit was barred by the Eleventh Amendment, JUSTICE SCALIA argues, that case disposed of the question whether Congress has the authority to abrogate States’ immunity when legislating pursuant to the powers granted it by the Constitution. See post, at 36-37. This argument depends on the notion that, in passing the Judiciary Act, “Congress... sought to eliminate [the] state sovereign immunity” that Article III had not eliminated. Post, at 36 (emphasis in original). As JUSTICE SCALIA is well aware, however, the Judiciary Act merely gave effect to the grant of federal-question jurisdiction under Article III, which was not self-executing. Thus, if Article III did not “automatically eliminate” sovereign immunity, see post, at 33, then neither did the Judiciary Act of 1875. That unsurprising conclusion does not begin to address the question whether other congressional enactments, not designed simply to implement Article III‘s grants of jurisdiction, may override States’ immunity. When one recalls, in addition, our conclusion that “Art[icle] III ‘arising under’ jurisdiction is broader than federal-question jurisdiction under § 1331,”

Verlinden B. V. v. Central Bank of Nigeria, 461 U. S. 480, 495 (1983), JUSTICE SCALIA‘s conception of Hans’ holding looks particularly exaggerated.

Our prior cases thus indicate that Congress has the authority to override States’ immunity when legislating pursuant to the Commerce Clause. This conclusion is confirmed by a consideration of the special nature of the power conferred by that Clause.

B

We have recognized that the States enjoy no immunity where there has been “a surrender of this immunity in the plan of the convention.”

Monaco v. Mississippi, 292 U. S. 313, 322-323 (1934), quoting The Federalist No. 81, p. 657 (H. Dawson ed. 1876) (A. Hamilton). Because the Commerce Clause withholds power from the States at the same time as it confers it on Congress, and because the congressional power thus conferred would be incomplete without the authority to render States liable in damages, it must be that, to the ex-tent that the States gave Congress the authority to regulate commerce, they also relinquished their immunity where Congress found it necessary, in exercising this authority, to render them liable. The States held liable under such a congressional enactment are thus not “unconsenting“; they gave their consent all at once, in ratifying the Constitution containing the Commerce Clause, rather than on a case-by-case basis.

It would be difficult to overstate the breadth and depth of the commerce power. See, e. g.,

NLRB v. Jones & Laughlin Steel Corp., 301 U. S. 1 (1937); ; . It is not the vastness of this power, however, that is so important here: it is its effect on the power of the States. The Commerce Clause, we long have held, displaces state authority even where Congress has chosen not to act, see ; ; , and it sometimes precludes state regulation even though existing federal law does not pre-empt it, see ; . Since the States may not legislate at all in these last two situations, a conclusion that Congress may not create a cause of action for money damages against the States would mean that no one could do so. And in many situations, it is only money damages that will carry out Congress’ legitimate objectives under the Commerce Clause.

The case before us brilliantly illuminates these points. The general problem of environmental harm is often not susceptible of a local solution. See

Illinois v. Milwaukee, 406 U. S. 91 (1972) (recognizing authority of federal courts to create federal “common law” of nuisance to apply to interstate water pollution, displacing state nuisance laws). We have, in fact, invalidated one State‘s effort to deal with the problem

of waste disposal on a local level. See

Philadelphia v. New Jersey, supra. A New Jersey statute prohibited the treatment and disposal, within the State, of any solid or liquid wastes generated outside the State. Indicating that a law applicable to all wastes would have survived under the Commerce Clause, , we held that the exemption of locally produced wastes doomed the statute, . As a practical matter, however, it is difficult to imagine that a State could forbid the disposal of all wastes. Hence, the Commerce Clause as interpreted in ensures that we often must look to the Federal Government for environmental solutions. And often those solutions, to be satisfactory, must include a cause of action for money damages.

The cause of action under consideration, for example, came about only after Congress had tried to solve the problem posed by hazardous substances through other means. Prior statutes such as the Resource Conservation and Recovery Act of 1976, 90 Stat. 2796, as amended, 42 U. S. C. § 6901 et seq., had failed in large part because they focused on preventive measures to the exclusion of remedial ones. See Note, Superfund and California‘s Implementation: Potential Conflict, 19 C. W. L. R. 373, 376, n. 23 (1983). The remedy that Congress felt it needed in CERCLA is sweeping: everyone who is potentially responsible for hazardous-waste contamination may be forced to contribute to the costs of cleanup. See, e. g., 42 U. S. C. § 9613(f)(1) (1986 ed., Supp. IV). Congress did not think it enough, moreover, to permit only the Federal Government to recoup the costs of its own cleanups of hazardous-waste sites; the Government‘s resources being finite, it could neither pay up front for all necessary cleanups nor undertake many different projects at the same time. Some help was needed, and Congress sought to encourage that help by allowing private parties who voluntarily cleaned up hazardous-waste sites to recover a proportionate amount of the costs of cleanup from the other potentially responsible parties. See ibid.;

Mardan Corp. v. C. G. C. Music, Ltd., 804 F. 2d 1454, 1457, n. 3 (CA9 1986); . If States, which comprise a significant class of owners and operators of hazardous-waste sites, see Brief for Respondent 8, need not pay for the costs of cleanup, the overall effect on voluntary cleanups will be substantial. This case thus shows why the space carved out for federal legislation under the commerce power must include the power to hold States financially accountable not only to the Federal Government, but to private citizens as well.

It does not follow that Congress, pursuant to its authority under the Commerce Clause, could authorize suits in federal court that the bare terms of Article III would not permit. No one suggests that if the Commerce Clause confers on Congress the power of abrogation, it must also confer the power to direct that certain state-law suits (not falling under the diversity jurisdiction) be brought in federal court.

According to Pennsylvania, however, to decide that Congress may permit suits against States for money damages in federal court is equivalent to holding that Congress may expand the jurisdiction of the federal courts beyond the bounds of Article III. Pennsylvania argues that the federal judicial power as set forth in Article III does not extend to any suits for damages brought by private citizens against unconsenting States. See Brief for Petitioner 35-36, quoting

Ex parte New York, 256 U. S. 490, 497 (1921) (“[T]he entire judicial power granted by the Constitution does not embrace authority to entertain a suit brought by private parties against a State without consent given“). We never have held, however, that Article III does not permit such suits where the States have consented to them. Pennsylvania‘s argument thus is answered by our conclusion that, in approving the commerce power, the States consented to suits against them based on congressionally created causes of action. Its claim also is answered by (1976). The Fourteenth Amendment does not purport to expand or even change the scope of Article III. If Pennsylvania were right about the limitations on Article III, then our holding in would mean that the Fourteenth Amendment, though silent on the subject, expanded the judicial power as originally conceived. We do not share that view of .³

IV

We hold that CERCLA renders States liable in money damages in federal court, and that Congress has the authority to render them so liable when legislating pursuant to the Commerce Clause. Given our ruling in favor of Union Gas, we need not reach its argument that

Hans v. Louisiana, 134 U. S. 1 (1890), should be overruled. We affirm the judgment of the Court of Appeals for the Third Circuit and remand the case for further proceedings consistent with this opinion.

It is so ordered.

JUSTICE STEVENS, concurring.

It is important to emphasize the distinction between our two Eleventh Amendments. There is first the correct and literal interpretation of the plain language of the Eleventh Amendment that is fully explained in JUSTICE BRENNAN‘S dissenting opinion in

Atascadero State Hospital v. Scanlon, 473 U. S. 234, 247 (1985). In addition, there is the defense of sovereign immunity that the Court has added to the text of the Amendment in cases like . With respect to the former—the legitimate scope of the Eleventh Amendment limitation on federal judicial power—I do not believe Congress has the power under the Commerce Clause, or under any other provision of the Constitution, to abrogate the States’ immunity. A statute cannot amend the Constitution. With respect to the latter—the judicially created doctrine of state immunity even from suits alleging violation of federally protected rights—I agree that Congress has plenary power to subject the States to suit in federal court.

Because JUSTICE BRENNAN‘s opinion in

Atascadero and the works of numerous scholars¹ have exhaustively and conclusively refuted the contention that the Eleventh Amendment embodies a general grant of sovereign immunity to the States, further explication on this point is unnecessary. Suffice it to say that the Eleventh Amendment carefully mirrors the language of the citizen-state and alien-state diversity clauses of Article III and only provides that “[t]he Judicial power of the United States shall not be construed to extend” to these cases. There is absolutely nothing in the text of the Amendment that in any way affects the other grants of “judicial Power” contained in Article III.² Plainer language is seldom, if ever, found in constitutional law.

Hans v. Louisiana, supra, however, the Court departed from the plain language, purpose, and history of the Eleventh Amendment, extending to the States immunity from suits premised on the “arising under” jurisdictional grant of Article III. Later adjustments to this rule, as well as the Court‘s inability to develop a coherent doctrine of Eleventh Amendment immunity, make clear that this expansion of state immunity is not a matter of Eleventh Amendment law at all, but rather is based on a prudential interest in federal-state comity and a concern for “Our Federalism.” The Eleventh Amendment, as does Article III, speaks in terms of “judicial power.” The question that must therefore animate the inquiry in any actual Eleventh Amendment case is whether the federal court has power to entertain the suit. In cases in which there is no such power, Congress cannot provide it—even through a “clear statement.” Many of this Court‘s decisions, however, purporting to apply the Eleventh Amendment, do not deal with judicial power at all. Instead, the issue of immunity is treated as a question of the proper role of the federal courts in the amalgam of federal-state relations. It is in these cases that congressional abrogation is appropriate.

Several of this Court‘s decisions make clear that much of our state immunity doctrine has absolutely nothing to do with the limit on judicial power contained in the Eleventh Amendment. For example, it is well established that a State may waive its immunity, subjecting itself to possible suit in federal court. See

Atascadero, 473 U. S., at 238; ; ; . Yet, the cases are legion holding that a party may not waive a defect in subject-matter jurisdiction or invoke federal jurisdiction simply by consent. See, e. g., ; ; ; ; ; . This must be particularly so in cases in which the federal courts are entirely without Article III power to entertain the suit. Our willingness to allow States to waive their immunity thus demonstrates that this immunity is not a product of the limitation of judicial power contained in the Eleventh Amendment.

Another striking example of the application of prudential—rather than true jurisdictional—concerns is found in our decision in

Edelman v. Jordan, 415 U. S. 651 (1974). There, the Court inexplicably limited the fiction established in , which permits suits against state officials in their official capacities for ultra vires acts, and concluded that the Young fiction only applies to prospective grants of relief. If simply involved an application of the limitation on judicial power contained in the Eleventh Amendment, once judicial power was found to exist to award prospective relief (even at some monetary cost to the State, see, e. g., ), it is difficult to understand why that same judicial power would not extend to award other forms of relief. See (STEVENS, J., concurring in judgment). In , the Court made explicit what was implicit in : the Young fiction “rests on the need to promote the vindication of federal rights,” while represents an attempt to “accommodate” this protection to the “competing inter-est” in “the constitutional immunity of the States.” Similarly, in , the Court explained:

“Both prospective and retrospective relief implicate Eleventh Amendment concerns, but the availability of prospective relief of the sort awarded in Ex parte Young gives life to the Supremacy Clause. Remedies designed to end a continuing violation of federal law are necessary to vindicate the federal interest in assuring the supremacy of that law. But compensatory or deterrence interests are insufficient to overcome the dictates of the Eleventh Amendment.” (Citations omitted.)

The theme that thus emerges from cases such as

Edelman, , and is one of balancing of state and federal interests. This sort of balancing, however, like waiver, is antithetical to traditional understandings of Article III subject-matter jurisdiction—either the judicial power extends to a suit brought against a State or it does not. See (Frankfurter, J., dissenting). As a result, these cases are better understood as simply invoking the comity and federalism concerns discussed in our abstention cases, see, e. g., ; ; ; ; ; , although admittedly in a slightly different voice.³ In my view, federal courts “have a primary obligation to protect the rights of the individual that are embodied in the Federal Constitution” and laws, (STEVENS, J., concurring), and generally should not eschew this responsibility based on some diffuse, instrumental concern for state autonomy. Yet, even if I were convinced otherwise, I would think it readily apparent that congressional abrogation is entirely appropriate. Cf. . Congress is not superseding a constitutional provision in these cases, but rather is setting aside the Court‘s assessment of the extent to which the use of constitutionally prescribed federal authority is prudent.

Because Congress has decided that the federal interest in protecting the environment outweighs any countervailing interest in not subjecting States to the possible award of monetary damages in a federal court, and because the “judicial power” of the United States plainly extends to such suits, I join JUSTICE BRENNAN‘s opinion. Even if a majority of this Court might have reached a different assessment of the proper balance of state and federal interests as an original matter, once Congress has spoken, we may not disregard its express decision to subject the States to liability under federal law.

JUSTICE SCALIA, with whom THE CHIEF JUSTICE, JUSTICE O‘CONNOR, and JUSTICE KENNEDY join as to Parts II, III, and IV, concurring in part and dissenting in part.

I

I join Part II of JUSTICE BRENNAN‘s opinion holding that the text of the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 42 U. S. C. § 9601 et seq., as amended by the Superfund Amendments and Reauthorization Act of 1986 (SARA), Pub. L. 99-499, 100 Stat. 1613, clearly renders States liable for money damages in private suits. JUSTICE WHITE‘s contention that there is no clear statement is given plausibility only by his methodology of considering CERCLA and SARA separately, finding that first the one and then the other does not necessarily import monetary liability to private individuals—CERCLA because, as we held in

Employees v. Missouri Dept. of Public Health and Welfare, 411 U. S. 279 (1973), the inclusion of States within defined terms is not alone enough to evince clear intent to abrogate Eleventh Amendment immunity, post, at 48-49 (opinion concurring in judgment in part and dissenting in part); and SARA because there the unquestionable reference to liability coextensive with the liability of private persons was set forth in a section dealing with limitation of liability, thus not assuring the intent of the Congress which enacted that provision to extend liability to the States, post, at 51-52.

That methodology is appropriate, and JUSTICE WHITE‘S conclusion is perhaps correct, if one assumes that the task of a court of law is to plumb the intent of the particular Congress that enacted a particular provision. That methodology is not mine nor, I think, the one that courts have traditionally followed. It is our task, as I see it, not to enter the minds of the Members of Congress—who need have nothing in mind in order for their votes to be both lawful and effective—but rather to give fair and reasonable meaning to the text of the United States Code, adopted by various Congresses at various times. See

United States v. Fausto, 484 U. S. 439, 454-455 (1988). CERCLA, as amended by SARA, clearly holds the States liable for damages in private suits. The inclusion of States, apparently for all purposes, within the definition of “person,” reinforced by the language of the limitation that assumes state liability equivalent to the liability of private individuals, leaves no fair doubt that States are liable to private persons for money damages. Whether it was the CERCLA Congress that envisioned this, or the SARA Congress, is to me irrelevant. The law does.

Finding that the statute renders the States liable in private suits for money damages, I must consider the continuing validity of

Hans v. Louisiana, 134 U. S. 1 (1890), which held that the Eleventh Amendment precludes individuals from bringing damages suits against States in federal court even where the asserted basis of jurisdiction is not diversity of citizenship but the existence of a federal question.

II

Eight Members of the Court addressed the question whether to overrule

Hans only two Terms ago—but inconclusively, since they were evenly divided. See . Since the substantive issue was addressed so extensively by the plurality opinion announcing the judgment of the Court in that case (which I will refer to as the “plurality opinion“), and by the dissent, I will only sketch its outlines here.

The Eleventh Amendment states:

“The Judicial power of the United States shall not be construed to extend to any suit in law or equity, com-menced or prosecuted against one of the United States by Citizens of another State, or by Citizens or Subjects of any Foreign State.”

If this text were intended as a comprehensive description of state sovereign immunity in federal courts—that is, if there were no state sovereign immunity beyond its precise terms—then it would unquestionably be most reasonable to interpret it as providing immunity only when the sole basis of federal jurisdiction is the diversity of citizenship that it describes (which of course tracks some of the diversity jurisdictional grants in U. S. Const., Art. III, § 2). For there is no plausible reason why one would wish to protect a State from being sued in federal court for violation of federal law (a suit falling within the jurisdictional grant over cases “arising under ... the Laws of the United States“) when the plaintiff is a citizen of another State or country, but to permit a State to be sued there when the plaintiff is citizen of the State itself. Thus, unless some other constitutional principle beyond the immediate text of the Eleventh Amendment confers immunity in the latter situation—that is to say, unless the text of the Eleventh Amendment is not comprehensive—even if the parties to a suit fell within its precise terms (for example, a State and the citizen of another State) sovereign immunity would not exist so long as one of the other, nondiversity grounds of jurisdiction existed.

About a century ago, in the landmark case of

Hans v. Louisiana, the Court unanimously rejected this “comprehensive” approach to the Amendment, finding sovereign immunity where not only a nondiversity basis of jurisdiction was present, but even where the parties did not fit the description of the Eleventh Amendment, the plaintiff being a citizen not of another State or country, but of Louisiana itself. What we said in was, essentially, that the Eleventh Amendment was important not merely for what it said but for what it reflected: a consensus that the doctrine of sovereign immunity, for States as well as for the Federal Govern-ment, was part of the understood background against which the Constitution was adopted, and which its jurisdictional provisions did not mean to sweep away. “[T]he cognizance of suits and actions [against unconsenting States] was not contemplated by the Constitution when establishing the judicial power of the United States.” . We noted that the decision of this Court that prompted the Eleventh Amendment, , permitting a South Carolina citizen to bring an assumpsit action for damages against the State of Georgia in federal court, had “created a shock of surprise throughout the country,” ; and we concluded that the Amendment which by its precise terms repudiated that decision reflected as well a repudiation of the premise upon which that decision was based, namely, that Article III‘s jurisdictional grants over the States are unlimited by the doctrine of sovereign immunity. “The letter [of Article III and the Eleventh Amendment] is appealed to now,” we said, “as [the letter of Article III] was then, as a ground for sustaining a suit brought by an individual against a State.” . We rejected that appeal. The rationale of and of the many cases that have followed it was concisely expressed, again for a unanimous Court, by Chief Justice Hughes in a case which held that, despite Article III‘s express grant of jurisdiction over suits “between a State ... and foreign States,” and despite the absence of express grant of sovereign immunity in the Eleventh Amendment, a State could not be sued by a foreign State in federal court:

“Manifestly, we cannot rest with a mere literal application of the words of § 2 of Article III, or assume that the letter of the Eleventh Amendment exhausts the restrictions upon suits against non-consenting States. Behind the words of the constitutional provisions are postulates which limit and control. There is the essential postulate that the controversies, as contemplated, shall be found to be of a justiciable character. There is also the postulate that States of the Union, still possessing attributes of sovereignty, shall be immune from suits, without their consent, save where there has been ‘a surrender of this immunity in the plan of the convention.’ The Federalist, No. 81.”
Monaco v. Mississippi, 292 U. S. 313, 322-323 (1934) (footnote omitted).

The evidence is strong that the jurisdictional grants in Article III of the Constitution did not automatically eliminate underlying state sovereign immunity, and even stronger that that assumption was implicit in the Eleventh Amendment. What is subject to greater dispute, however, is how much sovereign immunity was implicitly eliminated by what Hamilton called the “plan of the convention.” We have already held that “inherent in the constitutional plan,”

Monaco v. Mississippi, supra, at 329, are a waiver of immunity against suits by the United States itself, see ; , and a waiver of immunity against suits by other States, see . The foremost argument urged in favor of overruling is that a waiver of immunity against suits presenting federal questions is also implicit in the constitutional scheme. On this single point I add a few words to what was so recently said in .

The inherent necessity of a tribunal for peaceful resolution of disputes between the Union and the individual States, and between the individual States themselves, is incomparably greater, in my view, than the need for a tribunal to resolve disputes on federal questions between individuals and the States. Undoubtedly the Constitution envisions the necessary judicial means to assure compliance with the Constitution and laws. But since the Constitution does not deem this to require that private individuals be able to bring claims against the Federal Government for violation of the Constitution or laws, see

United States v. Testan, 424 U. S. 392, 399-402 (1976); U. S. Const., Art. I, § 9, cl. 7 (“No Money shall be drawn from the Treasury, but in Consequence of Appropriations made by Law“), it is difficult to see why it must be interpreted to require that private individuals be able to bring such claims against the States. If private initiation of suit against the offending sovereign as such is essential to preservation of the structure, it is difficult to see why it would not be essential at both levels. Indeed if anything it would seem more important at the federal level, since suits against the States for violation of the Constitution or laws can at least be brought by the Federal Government itself, see . In providing federal immunity from private suit, therefore, the Constitution strongly suggests that state immunity exists as well. Of course federal law can give, and has given, the private suitor many means short of actions against the State to assure compliance with federal law. He may obtain a federal injunction against the state officer, which will effectively stop the unlawful action, see , and may obtain money damages against state officers, and even local governments, under 42 U. S. C. § 1983; see . I think it impossible to find in the scheme of the Constitution a necessity that private remedies be expanded beyond this, to include a remedy not available, for a similar infraction, against the United States itself.

Even if I were wrong, however, about the original meaning of the Constitution, or the assumption adopted by the Eleventh Amendment, or the structural necessity for federal-question suits against the States, it cannot possibly be denied that the question is at least close. In that situation, the mere venerability of an answer consistently adhered to for almost a century, and the difficulty of changing, or even clearly identifying, the intervening law that has been based on that answer, strongly argue against a change. As noted by the

Welch plurality, ” has been reaffirmed in case after case, often unanimously and by exceptionally strong Courts“; its reversal “would overrule at least 17 cases, in addition to Hans itself” and cast doubt on “a variety of other cases that were concerned with this Court‘s traditional treatment of sovereign immunity.” . Moreover, unlike the vast majority of judicial decisions, has had a pervasive effect upon statutory law, automatically assuring that private damages actions created by federal law do not extend against the States. Forty-nine Congresses since have legislated under that assurance. It is impossible to say how many extant statutes would have included an explicit preclusion of suits against States if it had not been thought that such suits were automatically barred. Indeed, it is not even possible to say that, without , all constitutional amendments would have taken the form they did. The Seventeenth Amendment, eliminating the election of Senators by state legislatures, was ratified in 1913, 23 years after . If it had been known at that time that the Federal Government could confer upon private individuals federal causes of action reaching state treasuries; and if the state legislatures had had the experience of urging the Senators they chose to protect them against the proposed creation of such liability; it is not inconceivable, especially at a time when voluntary state waiver of sovereign immunity was rare, that the Amendment (which had to be ratified by three-quarters of the same state legislatures) would have contained a proviso protecting against such incursions upon state sovereignty.

I would therefore decline respondent‘s invitation to overrule

Hans v. Louisiana.

III

JUSTICE BRENNAN‘s plurality opinion purports to assume the validity of

Hans, and yet reaches the result that CERCLA‘s imposition of monetary liability is constitutional because Congress has the power to abrogate state sovereign immunity in the exercise of its Commerce Clause power. JUSTICE WHITE, who not merely assumes the validity of but actually believes in it, agrees with that disposition. Better to overrule , I should think, than to perpetuate the complexities that it creates, see (BRENNAN, J., dissenting), but eliminate all its benefits to the federal system. If means only that federal-question suits for money damages against the States cannot be brought in federal court unless Congress clearly says so, it means nothing at all. We do not need for the “clear statement” rule—just as we do not need to rely on any constitutional prohibition of suits against the Federal Government to require a similar rule for elimination of the sovereign immunity of the United States. See ; . As far as I can discern, the course the Court today pursues—preserving but permitting Congress to overrule it—achieves the worst of both worlds. And it is a course no more justified by text than by consequences.

To begin with,

Hans did not merely hold that Article III failed to eliminate state sovereign immunity of its own force, without any congressional action to that end. In , as here, there was a congressional statute that could be pointed to as eliminating state sovereign immunity—namely, the Judiciary Act of 1875, ch. 137, § 1, 18 Stat. 470, which gave United States courts jurisdiction over cases involving federal questions. (The Court was unquestionably aware of that refinement, because it was the statutory ground of interpretation of the Judiciary Act of 1789, ch. 20, § 13, 1 Stat. 80, rather than the constitutional ground, that Justice Iredell had relied upon in his dissent in , which the Court discussed at some length.) Thus, the distinction that the Court must rely upon is not one between cases in which Congress has assertedly sought to eliminate state sovereign immunity and cases in which in no such assertion is available; but rather the much more gossamer distinction between cases in which Congress has assertedly sought to eliminate state sovereign immunity pursuant to its powers to create and organize courts, and cases in which it has assertedly sought to do so pursuant to some of its other powers.

I think it plain that the position adopted by the Court contradicts the rationale of

Hans, if not its narrow holding. was not expressing some narrow objection to the particular federal power by which Louisiana had been haled into court, but was rather enunciating a fundamental principle of federalism, evidenced by the Eleventh Amendment, that the States retained their sovereign prerogative of immunity. That is clear throughout the opinion, but particularly in the following passage:

“Suppose that Congress, when proposing the Eleventh Amendment, had appended to it a proviso that nothing therein contained should prevent a State from being sued by its own citizens in cases arising under the Constitution or laws of the United States: can we imagine that it would have been adopted by the States? The supposition that it would is almost an absurdity on its face.

“The truth is, that the cognizance of suits and actions unknown to the law, and forbidden by the law, was not contemplated by the Constitution when establishing the judicial power of the United States.”
134 U. S., at 15.

This rationale is also evident from

Hans’ reliance upon the dissenting opinion of Justice Iredell in —whose views, the Court said, “were clearly right,—as the people of the United States in their sovereign capacity [by ratifying the Eleventh Amendment] subsequently decided.” . Iredell‘s only words addressed precisely to the constitutional issue were as follows:

“So much, however, has been said on the Constitution, that it may not be improper to intimate that my present opinion is strongly against any construction of it, which will admit, under any circumstances, a compulsive suit against a State for the recovery of money. I think every word in the Constitution may have its full effect without involving this consequence, and that nothing but express words, or an insurmountable implication (neither of which I consider, can be found in this case) would authorise the deduction of so high a power.”
2 Dall., at 449-450.

Our later cases are similarly clear that state immunity from suit in federal courts is a structural component of federalism, and not merely a default disposition that can be altered by action of Congress pursuant to its Article I powers. As we unanimously explained in

Ex parte New York, 256 U. S. 490, 497 (1921):

“That a State may not be sued without its consent is a fundamental rule of jurisprudence having so important a bearing upon the construction of the Constitution of the United States that it has become established by repeated decisions of this court that the entire judicial power granted by the Constitution does not embrace authority to entertain a suit brought by private parties against a State without consent given: not one brought by citizens of another State, or by citizens or subjects of a foreign State, because of the Eleventh Amendment; and not even one brought by its own citizens, because of the fundamental rule of which the Amendment is but an exemplification.”

Great Northern Ins. Co. v. Read, 322 U. S. 47, 51 (1944), we said:

“A state‘s freedom from litigation was established as a constitutional right through the Eleventh Amendment. The inherent nature of sovereignty prevents actions against a state by its own citizens without its consent.”

Atascadero, 473 U. S., at 242, we identified this principle as an essential element of the constitutional checks and balances:

“The ‘constitutionally mandated balance of power’ between the States and the Federal Government was adopted by the Framers to ensure the protection of ‘our fundamental liberties.’ [
Garcia v. San Antonio Metropolitan Transit Authority, 469 U. S. 528, 572 (1985) (Powell, J., dissenting)]. By guaranteeing the sovereign immunity of the States against suit in federal court, the Eleventh Amendment serves to maintain this balance.”

And in recently refusing to overrule

Hans in —an opinion joined by JUSTICE WHITE—the plurality opinion observed that “established that the Eleventh Amendment embodies a broad constitutional principle of sovereign immunity“; that “‘a suit directly against a State by one of its own citizens is not one to which the judicial power of the United States extends, unless the State itself consents to be sued.‘” , quoting (Harlan, J. concurring). The only attempt by either the plurality or JUSTICE WHITE to reconcile today‘s holding with the “broad constitutional principle of sovereign immunity” established by these precedents is the plurality‘s facile assertion that “in approving the commerce power, the States consented to suits against them based on congressionally created causes of action,” ante, at 22. The suggestion that this is the kind of consent our cases had in mind when reciting the familiar phrase, “the States may not be sued without their consent,” does not warrant response.

The Court‘s conclusion is not only contrary to the clear understanding of a century of cases regarding the Eleventh Amendment, but it contradicts our unvarying approach to Article III as setting forth the exclusive catalog of permissible federal-court jurisdiction. When we have turned to consider whether “a surrender of [state] immunity [is inherent] in the plan of the convention,” we have discussed that issue under the rubric of the various grants of jurisdiction in Article III, seeking to determine which of those grants must reasonably be thought to include suits against the States. See, e. g.,

Monaco, 292 U. S., at 328-330. We have never gone thumbing through the Constitution, to see what other original grants of authority—as opposed to Amendments adopted after the Eleventh Amendment—might justify elimination of state sovereign immunity. If private suits against States, though not permitted under Article III (by virtue of the understanding represented by the Eleventh Amendment), are nonetheless permitted under the Commerce Clause, or under some other Article I grant of federal power, then there is no reason why the other limitations of Article III cannot be similarly exceeded. That Article would be transformed from a comprehensive description of the permissible scope of federal judicial authority to a mere default disposition, applicable unless and until Congress prescribes more expansive authority in the exercise of one of its Article I powers. That is not the regime the Constitution establishes.

The Court‘s error is clear enough from the embarrassing frailty of the case support to which the plurality opinion appeals. JUSTICE BRENNAN refers to “statements ... [that] lay a firm foundation,” ante, at 14, a “path [that] continues,” ibid., and a “message [that] is plain,” ibid. What he notably does not cite is a single Supreme Court case, over the past 200 years upholding (in absence of a waiver) the congressional exercise of the asserted power—or even a single Supreme Court case finding that such an exercise has occurred. How strange that such a useful power—one that the plurality finds essential to the achievement of congressional objectives, ante, at 20-22—should never have been approved and rarely (if ever) have been asserted. Even the “message-sending” dicta that the plurality describes cannot be taken at face value. When the plurality states, for example, that “we have twice assumed that Congress has the authority to abrogate States’ immunity when acting pursuant to the Com-

merce Clause,”

ante, at 15, it means not that we have assumed it to be true, but that we have assumed it for the sake of argument. See (specifically refraining from even “intimating a view of the question“); . And of the two cases cited as referring to existence of a congressional power “to abrogate . . . immunity,” , one is plainly discussing abrogation not pursuant to Article I but pursuant to the Fourteenth Amendment, see , and the other is ambiguous but surely susceptible of that interpretation, see . In fact the only dicta even suggesting the position the Court today adopts were contained in , and (because it quoted Parden) in . As our later cases have made plain, see , Parden‘s holding was based upon the State‘s waiver of its sovereign immunity. One aspect of the case has already been overruled, and another cast in doubt, see infra, at 43; its dicta, and the dicta of a later case quoting its dicta, are hardly substantial support for the new constitutional principle the Court adopts.

Finally, the plurality opinion errs in relying on

Fitzpatrick v. Bitzer, supra, which upheld a money award against a State under Title VII of the Civil Rights Act of 1964, 78 Stat. 253, as amended, 42 U.S.C. § 2000e et seq. The distinction, as we carefully explained in that opinion, is that the Civil Rights Act was enacted pursuant to § 5 of the Fourteenth Amendment. We held that “the Eleventh Amendment, and the principle of state sovereignty which it embodies, see , . . . are necessarily limited” by the later Amendment, , whose substantive provisions were “by express terms directed at the States,” , and “were intended to be, what they really are, limitations of the power of the States and enlargements of the power of Congress,” , quoting . Nothing in this reasoning justifies limitation of the principle embodied in the Eleventh Amendment through appeal to antecedent provisions of the Constitution. The plurality asserts that it is no more impossible for provisions of the Constitution adopted concurrently with Article III to permit abrogation of state sovereign immunity than it is for provisions adopted subsequently. We do not dispute that that is possible, but only that it happened. As suggested above, if the Article I commerce power enables abrogation of state sovereign immunity, so do all the other Article I powers.

An interpretation of the original Constitution which permits Congress to eliminate sovereign immunity only if it wants to renders the doctrine a practical nullity and is therefore unreasonable. The Fourteenth Amendment, on the other hand, was avowedly directed against the power of the States, and permits abrogation of their sovereign immunity only for a limited purpose.

IV

It remains for me to consider whether the doctrine of waiver applies here. The basis for application of a waiver theory would be that, subsequent to enactment of CERCLA, Pennsylvania acted as the “owner and operator of . . . a facility,” 42 U.S.C. § 9607(a)(1), which latter term includes a “site or area where a hazardous substance has been deposited, stored, disposed of, or placed, or otherwise come to be located,” § 9601(9)(B); and that, by so acting, Pennsylvania voluntarily assumed the state liability for private suit that the legislation (assertedly) contains.

Parden is the only case in which we have held that the Federal Government can demand, as a condition to its permission of state action regulable under the Commerce Clause, the waiver of state sovereign immunity.¹ Two Terms ago, in , we overruled insofar as that case spoke to the clarity of language necessary to constitute such a demand. See (plurality opinion); (SCALIA, J., concurring in part and concurring in judgment). We explicitly declined to address, however, the continuing validity of ‘s holding that the Commerce Clause provided the constitutional power to make such a demand, . I would drop the other shoe.

There are obvious and fatal difficulties in acknowledging such a power if no Commerce Clause power to abrogate state sovereign immunity exists. All congressional creations of private rights of action attach recovery to the defendant‘s commission of some act, or possession of some status, in a field where Congress has authority to regulate conduct. Thus, all federal prescriptions are, insofar as their prospective application is concerned, in a sense conditional, and—to the extent that the objects of the prescriptions consciously engage in the activity or hold the status that produces liability—can be redescribed as invitations to “waiver.” For example, one is not liable for damages to private parties under the federal securities laws, see the Securities Exchange Act of 1934, § 10(b), 48 Stat. 891, 15 U.S.C. § 78j(b), unless one participates in the activity of purchasing or selling securities affecting interstate commerce; and it is possible to describe that liability as not having been categorically imposed, but rather as being the result of a “waiver” of one‘s immunity, in

exchange for federal permission to engage in that activity. At bottom, then, to acknowledge that the Federal Government can make the waiver of state sovereign immunity a condition to the State‘s action in a field that Congress has authority to regulate is substantially the same as acknowledging that the Federal Government can eliminate state sovereign immunity in the exercise of its Article I powers²—that is, to adopt the very principle I have just rejected. There is little more than a verbal distinction between saying that Congress can make the Commonwealth of Pennsylvania liable to private parties for hazardous-waste cleanup costs on sites that the Commonwealth owns and operates, and saying the same thing but adding at the end “if the Commonwealth chooses to own and operate them.” If state sovereign immunity has any reality, it must mean more than this.

* * *

The Court‘s holding today can be applauded only by those who think state sovereign immunity so constitutionally insignificant that

Hans itself might as well be abandoned. It is only the Court‘s steadfast refusal to accept the fundamental structural importance of that doctrine, reflected in and the other cases discussed above, that permits it to regard abrogation through Article I as an open question, and enables the plurality to fight the - battle all over again—but this time to win it—on the field of the Commerce Clause. It is a particularly unhappy victory, since instead of cleaning up the allegedly muddled Eleventh Amendment jurisprudence produced by , the Court leaves that in

place, and adds to the clutter the astounding principle that Article III limitations can be overcome by simply exercising Article I powers. It is an unstable victory as well, since that principle is too much at war with itself to endure. We shall either overrule

Hans in form as well as in fact, or return to its genuine meaning.

I would reverse the judgment of the Court of Appeals on the ground that federal courts have no power to entertain the present suit against the Commonwealth of Pennsylvania.

JUSTICE WHITE, with whom THE CHIEF JUSTICE, JUSTICE O‘CONNOR, and JUSTICE KENNEDY join as to Part I, concurring in the judgment in part and dissenting in part.

I find no “unmistakably clear language,”

Welch v. Texas Dept. of Highways and Public Transportation, 483 U.S. 468, 478 (1987), in either CERCLA or SARA that expresses Congress’ intent to abrogate the States’ Eleventh Amendment immunity. However, a majority of the Court concludes otherwise, and therefore I reach the constitutional issue presented here. On that question, I concur in JUSTICE BRENNAN‘s conclusion, but not his reasoning.

I

Our cases make it plain that only the most direct expression of Congress’ intent to make the States subject to suit will suffice to abrogate their sovereign immunity as recognized in the Eleventh Amendment. Thus, we have said that Congress must “explicitly and by clear language indicate on [the] face [of an enactment] an intent to sweep away the immunity of the States“; and that any such law must “have a history which focuses directly on the question of state liability and which shows that Congress considered and firmly decided to abrogate the Eleventh Amendment immunity of the States.”

Quern v. Jordan, 440 U.S. 332, 345 (1979). As we put it more recently: “Congress must express its intention to abrogate the Eleventh Amendment in unmistakable language in the statute itself.” .

Two statutes are offered by the Court as providing the “unmistakable language” required by our cases to abrogate the States’ Eleventh Amendment immunity: the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 42 U.S.C. § 9601 et seq. (1982 ed. and Supp. IV), and the 1986 Amendments to CERCLA, found in the Superfund Amendments and Reauthorization Act of 1986 (SARA), Pub. L. 99-499, 100 Stat. 1613. I consider both of these statutes in turn.

A

I begin by examining CERCLA, in the form in which Congress originally adopted it in 1980. In its initial consideration of this case—under CERCLA before the SARA amendments were added in 1986—the Third Circuit concluded that the statute did not contain an “unmistakable” abrogation of the Eleventh Amendment.

United States v. Union Gas Co., 792 F.2d 372, 378-382 (1986). The Court disagrees, however, suggesting that because CERCLA includes “States” within its definition of “persons,” 42 U.S.C. § 9601(21), and because the statute makes “persons” who are “owners or operators,” 42 U.S.C. § 9601(20) (1982 ed., Supp. IV), liable under § 9607, Congress expressed in CERCLA an “unmistakably” clear intent to make the States liable to suit by private parties in federal court. Ante, at 7-8. I reject this conclusion for several reasons.

First, I note that of the four federal judges who examined this question under CERCLA, only one—Judge Higginbotham in dissent in the Third Circuit‘s initial consideration of this case,

792 F.2d, at 383-386—found in this statutory scheme the requisite clear statement of Congress’ intent to abrogate the States’ immunity. See n. 7, infra. While such a “judicial headcount” is, of course, not dispositive, it does suggest that whatever one can say about CERCLA, it did not include an “unmistakable” declaration of abrogation of state immunity. If we are going to be faithful to and as providing our standard for this sort of case, then the fact that experienced jurists could disagree about Congress’ intent under CERCLA is relevant, because the disagreement suggests that the statute‘s provisions about state liability were certainly not “unmistakably clear.”

Second, the significance that the Court draws from CERCLA‘s inclusion of States within its definition of persons is suspect for its impact on other portions of the statute. The definitional section the Court relies on also includes the “United States Government” within the term “person.” 42 U.S.C. § 9601(21). Yet Congress also adopted, in CERCLA, an entirely separate statutory provision rendering the Federal Government suable under the statute‘s liability provision, see § 9607(g). If the Court‘s views about the significance of including States within the definition of persons is correct, then § 9607(g) was wholly redundant, because—by including the United States Government within the definition of persons—Congress had already stripped the Federal Government of its sovereign immunity.¹

Rather than assuming that Congress wrote a wholly redundant subsection of § 9607, however, it seems more likely to conclude that Congress did not think that including the United States Government or the States within § 9601(21)‘s general definition of “persons” subject to CERCLA‘s regime was enough to abrogate the sovereign immunity of either for damages awards.² Cf.

United States v. Testan, 424 U.S. 392, 399 (1976). With respect to the Federal Government, Congress went on to enact a separate provision executing the requisite waiver of immunity, § 9607(g). However, with respect to the States, Congress made no such additional provision: the conclusion to be drawn is obvious.

Finally, and most importantly, the Court‘s reading of CERCLA employs the precise analytical approach we rejected in

Employees v. Missouri Dept. of Public Health and Welfare, 411 U.S. 279 (1973). There, as is true here, the relevant statutory term that described who was covered by the Act (in Employees, it was the term “employers” in the Fair Labor Standards Act (FLSA)), expressly included the state defendant (in Employees, it was the State as an employer of “employees of a State . . . hospital“); invoking these provisions, a private litigant sought to hold the State liable under the statute‘s damages remedy. . Nonetheless, in Employees, we held that Congress had not thereby abrogated the States’ Eleventh Amendment immu-nity; instead, we concluded, Congress had meant only to make the States subject to enforcement actions brought by the Federal Government. .

In all relevant respects, the portion of CERCLA on which the Court relies and the portion of the FLSA that was before us in Employees are indistinguishable, as are the arguments made for considering the statutes to have abrogated the States’ immunity. In Employees, we rejected these arguments; the same result should attach here. Instead, we should conclude, as we did in Employees, that Congress’ intent could have been to let the Act‘s policies be achieved through enforcement actions taken by the Federal Government against the States. As we observed in

Employees, supra, at 286: “The policy of the Act so far as the States are concerned is wholly served by allowing the delicate federal-state relationship to be managed through” enforcement actions directed by the Federal Executive Branch—and not through litigation by private parties against the States.

Nor is the Court‘s result supported by reference to the purposes of CERCLA. Respondent finds much significance in the fact that this statute was designed to be “comprehensive” in nature.

792 F.2d, at 381 (summarizing respondent‘s contention below). But surely the Federal Employers’ Liability Act (), the Rehabilitation Act (), and the FLSA (Employees) were all “comprehensive” statutes in their respective fields, and yet this was not enough to deem the Eleventh Amendment abrogated in those cases. Nor is it true that CERCLA‘s “comprehensiveness” will be substantially lessened by deeming the States’ immunity to have survived intact. The States remain subject to liability at the hands of the Federal Government; this provides a viable means of achieving CERCLA‘s ends. See Reply Brief for Petitioner 10.³

Above all, the entire purpose of our “clear statement” rule would be obliterated if this Court were to imply Eleventh Amendment abrogation from our sense of what would best serve the general policy ends Congress was trying to achieve in a statute. Such arguments based on the statute‘s general goals, whatever weight they might have under a normal exercise in statutory construction, have no bearing on our analysis of congressional abrogation. Cf.

Dellmuth v. Muth, post, at 230-231. If Congress believes that making the States liable to private parties is critical to the scheme it has created in CERCLA, it is up to Congress to say so in unmistakable language. Since it has not, I believe that our “clear statement” precedents bar us from implying such a policy choice—even if it is “latent” in the statutory scheme, or an advisable means of achieving the statute‘s ends.

B

The question then becomes whether, as the Court of Appeals found,

United States v. Union Gas Co., 832 F.2d 1343 (1987), the 1986 amendments to CERCLA (known as SARA) added such an “unmistakable” statement of abrogation to the statute.

The text of the relevant portion of SARA (now codified at 42 U.S.C. § 9601(20)(D) (1982 ed., Supp. IV)) states, in full:

“STATE OR LOCAL GOVERNMENT LIMITATION.—Paragraph (20) of [42 U.S.C. § 9601] (defining ‘owner or operator‘) is amended as follows:

“(1) Add the following new subparagraph at the end thereof:

“‘(D) The term “owner or operator” does not include a unit of State or local government which acquired ownership or control involuntarily through bankruptcy, tax delinquency, abandonment, or other circumstances in which the government involuntarily acquires title by virtue of its function as sovereign. The exclusion provided under this paragraph shall not apply to any State or local government which has caused or contributed to the release . . . of a hazardous substance from the facility, and such a State or local government shall be subject to the provisions of this Act in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under [42 U.S.C. § 9607].‘” Pub. L. 99-499, 100 Stat. 1615.

Although Congress entitled the amendment “STATE OR LOCAL GOVERNMENT LIMITATION,” the Court disparages the idea that § 9601(20)(D) was enacted solely as a limitation on governmental unit liability. The Court asserts that such a view ignores that § 101(20)(D) “would be unnecessary unless” the States could be liable under § 9607. Ante, at 8. But everyone agrees that States may be liable under § 9607: the liability of the Commonwealth of Pennsylvania to the United States. Section 9601(20)(D) provides a significant reduction of that potential liability, as it limits the circumstances under which state and local governments will be forced to pay the United States Government for cleanups at involuntarily acquired sites. Given this fact, § 9601(20)(D) makes perfectly good sense without any contortion of it to imply an intent of Congress to abrogate the Eleventh Amendment.⁴

There is a second fact about the relevant part of SARA that makes it an odd candidate for an Eleventh Amendment abrogation provision: it only applies to facilities acquired by state and local governments “involuntarily . . . by virtue of [their] function[s] as sovereign.” See § 9601(20)(D). If this amendment is the means by which Congress intended to make the States liable to suit, it did so only with respect to those properties which a State acquired involuntarily; States would remain immune for sites which they owned and operated by choice. A State would be immune from private suit under § 9607 for costs associated with the cleanup of a state-created, owned, and operated hazardous-waste dump, but it would be liable for discharges at sites it acquired when an owner abandoned his property. Surely if the two cases are to be distinguished, the logical distinction would be exactly the opposite one.

Recognizing that Congress could not have intended such a result, the Court avoids this conclusion by saying that this part of SARA “explains and qualifies the entire definition of ‘owner or operator‘—not just that part of the definition applicable to involuntary owners.” Ante, at 12-13. But this is plainly wrong: the portion of the sentence which the

Court says renders the States liable (“a State or local government shall be subject . . .“) is introduced by the words, “[t]he exclusion provided under this paragraph shall not apply. . . .” § 9601(20)(D). Thus, the liability-creating portion of § 9601(20)(D) exists only as a “limit” on the liability-limiting portion of § 9601(20)(D). Under the Court‘s reading of the statute, we are left with the paradox of Congress being tougher on States that find themselves involuntary operators of waste sites, than it was on those that had owned and operated such facilities on their own accord.⁵

The Court argues that the last clause of the last sentence of § 9601(20)(D)—making involuntary-owner state and local governments that cause the release of toxic chemicals “subject to the provisions of [CERCLA] in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity“—provides the clear statement of abrogation required by our cases. But like the Court‘s reliance on the inclusion of States within CERCLA‘s definition of “persons” subject to the Act (which I discussed above),

this method of analysis is directly contrary to the approach we took in

Employees v. Missouri Dept. of Public Health and Welfare, 411 U.S. 279 (1973). The Court insists that its reliance on this part of SARA is correct because, if the statute is interpreted to mean something other than abrogating state immunity, the provision is rendered redundant and meaningless. Ante, at 11-12.

The provision, however, has meaning as something less than an abrogation provision because, like the statute in question in Employees, it exists to make the States liable to the Federal Government. While the Court is surely correct when it observes that, under

United States v. California, 332 U.S. 19, 26-27 (1947), no statutory provision is required as a general matter to permit the United States to sue a State, here, the Congress forbade such actions in the first part of § 9601(20)(D) with respect to some States (i. e., involuntary owners of waste sites). Thus, the portion of § 9601(20)(D) on which the Court rests its case is precisely like the 1966 amendment to § 3(d) of the FLSA that was before us in Employees: it operates to put some States back into the class of entities that may be liable to the United States, after Congress had previously exempted them from such actions. See . As in Employees, the statute should be read as only authorizing suits by the United States against the States, absent a more clear statement of an authorization of private actions.⁶

Edelman v. Jordan, 415 U.S. 651, 673 (1974), we said of the related question of interpreting a state statute to find a waiver of Eleventh Amendment immunity, that such a waiver would only be found “where stated ‘by the most express language or by such overwhelming implications from the text as [will] leave no room for any other reasonable construction‘” of the statute in question. Here, there is room for a “reasonable construction” of SARA that does not entail an Eleventh Amendment abrogation; i. e., that Congress intended it as a modification of the liability of the States to the Federal Government. Even if the Court‘s interpretation of § 9601(20)(D) were itself “reasonable,” the existence of an alternative, non-abrogating “reasonable” interpretation of the section dictates rejection of its view.

Consequently, I do not think that SARA‘s liability-limiting amendment to CERCLA contains an “unmistakably clear” statement by Congress that it wanted to abrogate the

States’ solemn immunity to private suit under the Eleventh Amendment.⁷

II

My view on the statutory issue has not prevailed, however; a majority of the Court has ruled that the statute, as amended, plainly intended to abrogate the immunity of the

States from suit in the federal courts. I accept that judgment. This brings me to the question whether Congress has the constitutional power to abrogate the States’ immunity.⁸ In that respect, I agree with the conclusion reached by JUSTICE BRENNAN in Part III of his opinion, that Congress has the authority under Article I to abrogate the Eleventh Amendment immunity of the States, although I do not agree with much of his reasoning.

Accordingly, I would affirm the judgment of the Court of Appeals.

JUSTICE O‘CONNOR, dissenting.

I agree with JUSTICE SCALIA that a faithful interpretation of the Eleventh Amendment embodies a concept of state sovereignty which limits the power of Congress to abrogate States’ immunity when acting pursuant to the Commerce Clause. But that view does not command a majority of the Court, thus necessitating an inquiry whether Congress intended in CERCLA, 42 U.S.C. § 9601 et seq., and SARA, Pub. L. 99-499, 100 Stat. 1613, to abrogate the States’ Eleventh Amendment immunity. On that question, I join Part I of JUSTICE WHITE‘s opinion. I also join Parts II, III, and IV of JUSTICE SCALIA‘s opinion concurring in part and dissenting in part.

Notes

Section 101(20)(D), as set forth in 42 U. S. C. 9601(20)(D), provides in full: “(D) The term ‘owner or operator’ does not include a unit of State or local government which acquired ownership or control involuntarily through bankruptcy, tax delinquency, abandonment, or other circumstances in which the government involuntarily acquires title by virtue of its function as sovereign. The exclusion provided under this paragraph shall not apply to any State or local government which has caused or contributed to the release or threatened release of a hazardous substance from the facility, and such a State or local government shall be subject to the provisions of this chapter in the same manner and to the same extent, both procedurally and substantively, as any nongovernmental entity, including liability under section 9607 of this title.” The language of the Rehabilitation Act Amendments of 1986, Pub. L. 99-506, 100 Stat. 1807, is indeed more pointed on the subject of abrogation than is CERCLA, since it mentions the Eleventh Amendment by name. See post, at 56, n. 7. It is surprising that JUSTICE WHITE‘s opinion lays so much stress on this difference in wording, however, because it expressly disclaims any intent to require that the words “Eleventh Amendment” appear in a statute in order to find abrogation. Ibid. If no magic words are required for abrogation, then each statute must be evaluated on its own terms, not defeated by reference to another statute that uses more specific language. See, e. g., Marshall, Fighting the Words of the Eleventh Amendment, 102 Harv. L. Rev. 1342 (1989); Jackson, The Supreme Court, the Eleventh Amendment, and State Sovereign Immunity, 98 Yale L. J. 1 (1988); Amar, Of Sovereignty and Federalism, 96 Yale L. J. 1425 (1987); Lee, Sovereign Immunity and the Eleventh Amendment: The Uses of History, 18 Urb. Law. 519 (1986); Shapiro, Wrong Turns: The Eleventh Amendment and the Pennhurst Case, 98 Harv. L. Rev. 61 (1984); Gibbons, The Eleventh Amendment and State Sovereign Immunity: A Reinterpretation, 83 Colum. L. Rev. 1889 (1983); Fletcher, A Historical Interpretation of the Eleventh Amendment: A Narrow Construction of an Affirmative Grant of Jurisdiction Rather than a Prohibition Against Jurisdiction, 35 Stan. L. Rev. 1033 (1983); Tribe, Intergovernmental Immunities in Litigation, Taxation, and Regulation: Separation of Powers Issues in Controversies About Federalism, 89 Harv. L. Rev. 682 (1976). In , we said that a condition of suability of the Bridge Commission, which we interpreted Congress to have attached to its approval of the interstate compact creating the Commission, was accepted by the States when they implemented the compact. That was an alternative holding, since we also found that the terms of the compact itself made the Commission suable. Obviously, moreover, what Congress may exact with respect to new entities created by compacts that the States have no constitutional power to make without its explicit consent, see U.S. Const., Art. I, § 10, cl. 3, may be much greater than what it may exact in other contexts. In an effort to avoid the force of this observation, the Court unleashes its oft-repeated statement that it relies on a “combination” of CERCLA and SARA to reach its conclusion. Ante, at 9, n. 2. The Court says that it is my “failure to recognize” this quality in its analysis that leads to my “confusion” about this case. Ibid.

I do not “fail to recognize” the Court‘s approach—I reject it outright. The search for an “unmistakable statement” of abrogation is the search for unmistakable proof that Congress purposefully intended to set aside the States’ immunity. It is, therefore, the search for a historical fact that either was or was not true at the time Congress legislated. The Court‘s “combination” analysis loses sight of this underlying theory behind our cases and, unfortunately, substantially undermines our precedents.

As I see it, the analysis must be this: either Congress abrogated the Eleventh Amendment when it enacted CERCLA—in which case, § 9607(g) was superfluous when adopted—or Congress did not do so until it adopted SARA—which is a peculiar view, for reasons I explain in Part I-B below—or Congress did not have an intent to abrogate in either instance. Blur-ring the choice among these possible historical facts by resting on a “combination” analysis is only an effort to make this difficult case artificially easier.

JUSTICE WHITE‘s attack on the notion that the definition of the word “persons,” standing alone, abrogates the States’ immunity from suit, see post, at 46-50, is directed at an argument that we do not make. We do not say that CERCLA‘s definition of “persons” alone overrides the States’ im-munity, but instead read CERCLA and SARA together, and argue that SARA‘s wording must inform our understanding of the other definitional sections of the statute. The failure to appreciate this point leads to four mistakes. First, in his “judicial headcount,” post, at 46-47, JUSTICE WHITE counts the votes as to the wrong statute. The judges who ruled that CERCLA did not render States liable did so when they considered the unamended version of CERCLA; as to CERCLA as amended by SARA, the three-judge panel unanimously agreed that it clearly abrogated the States’ immunity. (This headcounting approach is flawed for another, more fundamental reason: surely judges can disagree about the content and rigor of the standard of “unmistakable clarity,” and if they do, they are likely to reach different results on States’ amenability to suit for reasons having nothing to do with the statutory language itself.) Second, JUSTICE WHITE asserts that our reading of CERCLA is inconsistent with the Court‘s conclusion in , that a statute literally including the States as “persons” subject to the statute was not clear enough to abrogate the States’ immunity. Post, at 48-49. This claim ignores SARA‘s more specific language. Third, JUSTICE WHITE claims that our reading of CERCLA renders § 107(g)—which overrides the United States’ sovereign immunity from suit—redundant. Post, at 47. However, since we do not argue here that the inclusion of the States and the Federal Government in § 101(21)‘s definition of “persons,” standing alone, overrides these entities’ immunity, our position does not make § 107(g) superfluous. Finally, only a failure to recognize that we rely on § 101(21) and § 101(20)(D) in combination could lead to the suggestion that States would enjoy § 101(20)(D)‘s more favorable standard of liability even if they voluntarily acquired a site. Post, at 53, n. 5. The Eleventh Amendment asserts: “The Judicial power of the United States shall not be construed to extend to any suit in law or equity, commenced or prosecuted against one of the United States by Citizens of another State, or by Citizens or Subjects of any Foreign State.” This language parallels Article III, which provides in pertinent part: “The judicial Power shall extend ... to Controversies ... between a State and Citizens of another State ... and between a State ... and foreign ... Citizens or Subjects.” A “waiver” theory would not support retroactive imposition of liability—but that is rare in any event. Moreover, it could be held that waiver cannot occur when the State is unaware of the facts that trigger its liability, or of the law that imposes it. It is difficult to imagine how ignorance of the facts could ever be found, unless (as is most unlikely) we should decline to attribute the knowledge of the State‘s agents to the State itself. Our cases discussing waiver have displayed no interest in “actual” state knowledge of either facts or law. This conclusion is also supported by the fact that in two other places in § 9607, where Congress wished a particular provision to apply to private persons and the United States and the States, it used the phrases “[n]o person (including the United States or any State) . . .” and “any person (including the United States or any State).” See §§ 9607(i), (j). If Congress believed (as the Court contends that it did) that its inclusion of States within CERCLA‘s definition of “person” was adequate to bring the States fully within the operation of § 9607, then the parenthetical phrases I quote here would have been wholly redundant.

JUSTICE WHITE‘s response to this point is unconvincing. After claiming that our reading renders a part of the statute redundant—an accusation without merit, see n. 2, supra—JUSTICE WHITE resorts to a reading of § 101(20)(D) that, he admits, renders the phrase “as any nongovernmental entity” superfluous. Post, at 55, n. 6. To say that this phrase can be explained as a “statutory ‘exclamation point,‘” post, at 54-55, n. 6, is just another way of describing redundancy. Nor is it possible to explain this passage as an effort to pre-empt state-law immunity for local governments. See post, at 55, n. 6. Given our recognition that “there is no tradition of immunity for municipal corporations,” , and our refusal in the past to allow state-law immunities to define the scope of federal statutes, see, e. g., , Congress would see no need to use emphatic language to override this kind of immunity. Unless we conclude, therefore, that the phrase “as any nongovernmental entity” is superfluous, this clause demonstrates that §101(20)(D) was designed to do more than render the States liable in damages to the Federal Government. Since Union Gas itself eschews reliance on the theory of waiver we announced in , see Brief for Respondent 31, we neither discuss this theory here nor understand why JUSTICE SCALIA feels the need to do so. See post, at 42-44. This understanding of our state immunity cases explains an additional anomaly. Over the years, this Court has repeatedly exercised Article III power to review state-court judgments in cases involving claims that, under our post-Hans decisions, could not have been brought in federal district court. See, e. g., ; ; ; ; ; . See also ; . To the extent the Eleventh Amendment is broadly construed to have removed all federal power to adjudicate claims against the States regardless of whether or not the claim is one arising under federal law, it is difficult to justify our exercise of power in these cases. See (BRENNAN, J., dissenting). See also Jackson, 98 Yale L. J., at 13-39. However, if our post-Hans state immunity cases are instead understood as premised on a prudential balancing of state and federal interests, these cases are easily explained: a state-court decision defining federal law tips the balance in favor of federal review. Cf. ; . Respondent approaches the policy question with the view that limitless state liability under CERCLA is the best means to achieve the statute‘s ends. However, Congress clearly did not think so: it limited state and local governmental liability under § 9607 in several respects. First, there is the involuntary-ownership exclusion of § 9601(20)(D), adopted in the 1986 SARA amendments, that is discussed in detail in Part I-B, infra.

In addition, Congress also adopted in SARA a limitation on state and local government liability (to the Federal Government) for actions taken at toxic waste sites in response to emergencies. Pub. L. 99-499, § 107(d)(2), 100 Stat. 1629; 42 U.S.C. § 9607(d)(2) (1982 ed., Supp. IV). As the House Commerce Committee observed, this legislative exemption was designed to “remov[e] a disincentive for governments to respond to emergencies covered by CERCLA.” H. R. Rep. No. 99-253, pt. 1, p. 73 (1985). Thus, Congress did not view ever expanding governmental liability as the only way to achieve CERCLA‘s ends.

Of course, even if policy reasons did counsel expansive state liability under CERCLA, our “clear statement” rule mandates that the choice is to be left to Congress—to resolve with an explicit declaration of its decision—and not to be implied by this Court.

To the extent state immunity from suit in federal court is based on a concern for comity, and not on a limitation on Article III power, Congress is just as free to “declare its will” that this presumption come to an end as are States to decide not to accord one another immunity from suit in state court. See . A similar observation explains another section of SARA which the Court, ante, at 9-10, attempts to use as support for its reading of § 9601(20)(D): § 9607(d)(2), which was enacted by Congress to encourage state and local governments to conduct emergency cleanups of waste sites by exempting them from potential liability for those cleanup activities. See 42 U.S.C. § 9607(d)(2) (1982 ed., Supp. IV); H. R. Conf. Rep. No. 99-962, pp. 203-204 (1986). About this amendment, the Court again suggests that “Congress need not exempt States from liability unless they would otherwise be liable.” Ante, at 10.

As with § 9601(20)(D), however, this limitation is best understood as a limit on state liability to the United States; it need not be read as an implicit statement that elsewhere the Eleventh Amendment has been waived for private lawsuits, in order to make it a vital part of the statute. Cf.

Employees v. Missouri Dept. of Public Health and Welfare, 411 U.S. 279, 285-287 (1973).

The Court also rejects this conclusion by saying that the inclusion of the liability-creating exception to the liability-limiting exception of § 9601(20)(D) serves to enlighten us as to Congress’ “background understanding” of the effect of CERCLA in the first place: that States would be liable under § 9607. In this instance, and throughout, see n. 1, supra, the Court does not make it clear whether it is the SARA amendments of 1986, or CERCLA itself, that renders the States liable to suit under § 9607.

Yet the difference may be a significant one. Section 9607 is a strict-liability provision. See, e. g.,

New York v. Shore Realty Corp., 759 F.2d 1032, 1042 (CA2 1985); . If CERCLA as originally enacted—without any help from SARA—rendered States liable to private suits under § 9607, then they must be subject to that section‘s strict-liability rule as well.

But under § 9601(20)(D), state and local governments are liable only if they have “caused or contributed” to a release of toxic materials. If § 9601(20)(D) is the source of the Eleventh Amendment waiver, and if, as the Court contends, its provisions are meant to address all state and local governments that own or operate toxic sites, then perhaps Congress abrogated the Eleventh Amendment only far enough to make States liable under this less stringent rule—whether they are voluntary or involuntary owners of a site.

The Court goes on to observe, however, that even if this interpretation is accepted as explaining almost all of the last sentence of § 9601(20)(D), it still does not account for Congress “stress[ing] that States would be liable ‘to the same extent . . . as any nongovernmental entity,‘” ante, at 11. The Court contends that the first part of the last sentence of § 9601(20)(D) (i. e., “such a State . . . shall be subject“) would have been enough to accomplish the end of merely making involuntary-owner States liable to actions by the United States; the addition of the phrase “as any nongovernmental entity” means that Congress must have intended something more. To this I have three responses.

First, Congress may have added the phrase in which the Court puts so much stock (“as any nongovernmental entity“) as a statutory “exclamation

point“: Congress may have reasoned that while state and local governments that are involuntary owners should be exempted from liability under CERCLA, those that actually cause subsequent discharges should be liable under the statute, with their involuntary ownership no defense or excuse whatsoever when the United States seeks recovery. In this view, Congress simply added the relevant phrase to strongly emphasize that involuntary ownership is no defense if a state or local government causes a discharge. Put another way, it is incongruous to attribute such sweeping significance—an Eleventh Amendment abrogation, something we have found present in only the most extraordinary circumstances—to this one phrase in the definitional portion of SARA/CERCLA.

Second, Congress could have used the phrase “as any nongovernmental entity” to insure that local governments that cause discharges at involuntarily acquired sites would be liable under § 9607. Congress may have merely wanted to be forceful in using its pre-emptive power to set aside any state-law immunity doctrines for such local government entities, without necessarily going so far as to execute an “unmistakably clear” abrogation of state government immunity. Cf.

Quern v. Jordan, 440 U.S. 332, 338-341 (1979). Finally, even if my reading of this phrase makes it somewhat superfluous to the statute, the redundancy created by my interpretation of this one clause is not nearly as severe as the redundancy created by the Court‘s reading of the statute, and discussed in the text, supra, at 47.

One additional observation concerning SARA may be made. At the time SARA was enacted, one Court of Appeals—the Third Circuit, in its initial decision in this case, —and one District Court—also as part of this litigation, —had ruled on the question whether CERCLA as it was then written abridged States’ Eleventh Amendment immunity. Both of these courts held that it did not; no federal court had ruled to the contrary.

The Court‘s view of SARA is that, in enacting § 9601(20)(D), Congress had an “unmistakably clear” intent to amend CERCLA so as to reverse the force of these holdings finding a lack of abrogation in CERCLA‘s original text. Yet just eight days after it adopted SARA, Congress enacted the Rehabilitation Act Amendments of 1986, Pub. L. 99-506, 100 Stat. 1807, which included a provision setting aside the force of our holding in

Atascadero State Hospital v. Scanlon, 473 U.S. 234 (1985), that Congress had failed to provide a clear statement of abrogation of the Eleventh Amendment. The words Congress chose in that Act are instructive: “A State shall not be immune under the Eleventh Amendment . . . from suit in Federal court for a violation of [portions of the Act].” 100 Stat. 1845.

While I would not go so far as to hold that Congress must use these precise words (i. e., make reference to the Eleventh Amendment) before it will be deemed to have abrogated States’ immunity, the words used by Congress to set aside Atascadero are legions more “unmistakably clear” than the tangled mess in § 9601(20)(D), which the Court concludes set aside the then-existing case law with respect to CERCLA.

Of course, I do not believe that only the “magic words” found in the Rehabilitation Act amendment will suffice to achieve abrogation. Cf. ante, at 13, n. 4. Instead, my view (based on our prior decisions in Atascadero and

Welch v. Texas Dept. of Highways and Public Transportation, 483 U.S. 468 (1987)) is that Congress’ intent to abrogate must be expressed clearly, in a plain statement in the text of the enactment—and is not to be derived by parsing together various fragments scattered about a statute, as if it were a legislative quote acrostic. See also n. 1, supra.

As a preliminary matter, I reiterate my view that, for the reasons stated by the plurality in , , should not be overruled.