with whom Justice O’Connor, Justice Kennedy, and Justice Souter join, concurring in part and concurring in the judgment.
I agree with the Court that 26 U. S. C. § 7443A allows the Chief Judge of the Tax Court to assign special trial judges to preside over proceedings like those involved here, and join Parts I, II, and III of its opinion. I disagree, however, with the Court’s decision to reach, as well as its resolution of, the Appointments Clause issue.
I — I
As an initial matter, I think the Court errs by entertaining petitioners’ constitutional challenge on the merits. Petitioners not only failed to object at trial to the assignment of their case to a special trial judge, but expressly consented to that assignment. It was only after the judge ruled against them that petitioners developed their current concern over whether his appointment violated Article II, § 2, cl. 2, of the *893Constitution. They raised this important constitutional question for the first time in their appeal to the Fifth Circuit. That court concluded that petitioners had “waived this objection” by consenting to the assignment of their case. 904 F. 2d 1011, 1015, n. 9 (1990). When we granted certiorari, we asked the parties to brief and argue the following additional question: “Does a party’s consent to have its case heard by a special tax judge constitute a waiver of any right to challenge the appointment of that judge on the basis of the Appointments Clause, Art. II, § 2, cl. 2?” 498 U. S. 1066 (1991).
Petitioners would have us answer that question “no” by adopting a general rule that “structural” constitutional rights as a class simply cannot be forfeited, and that litigants are entitled to raise them at any stage of litigation. The Court neither accepts nor rejects that proposal — and indeed, does not even mention it, though the opinion does dwell upon the structural nature of the present constitutional claim, ante, at 878-880. Nor does the Court in any other fashion answer the question we specifically asked to be briefed, choosing instead to say that, in the present case, it will “exercise our discretion” to entertain petitioners’ constitutional claim. Ante, at 879. Thus, when there occurs a similar forfeiture of an Appointments Clause objection — or of some other allegedly structural constitutional deficiency — the courts of appeals will remain without guidance as to whether the forfeiture must, or even may, be disregarded. (The Court refers to this case as “one of th[e] rare” ones in which forfeiture will be ignored, ibid, —but since all forfeitures of Appointments Clause rights, and arguably even all forfeitures of structural constitutional rights, can be considered “rare,” this is hardly useful guidance.) Having asked for this point to be briefed and argued, and having expended our time in considering it, we should provide an answer. In my view the answer is that Appointments Clause claims, and other structural constitutional claims, have no special entitlement to review. A party forfeits the right to advance on appeal a nonjuris-*894dictional claim, structural or otherwise, that he fails to raise at trial. Although I have no quarrel with the proposition that appellate courts may, in truly exceptional circumstances, exercise discretion to hear forfeited claims, I see no basis for the assertion that the structural nature of a constitutional claim in and of itself constitutes such a circumstance; nor do I see any other exceptional circumstance in the present case. Cf. Peretz v. United States, post, at 954-955 (Scalia, J., dissenting). I would therefore reject petitioners’ sweeping proposition that structural constitutional rights as a class cannot be waived or forfeited and would refuse to entertain the constitutional challenge presented here.1
“No procedural principle is more familiar to this Court than that a constitutional right may be forfeited in criminal as well as civil cases by the failure to make timely assertion of the right before a tribunal having jurisdiction to determine it.” Yakus v. United States, 321 U. S. 414, 444 (1944); see also United States v. Socony-Vacuum Oil Co., 310 U. S. 150, 238-239 (1940). Forfeiture2 is “not a mere technicality and *895is essential to the orderly administration of justice.” 9 C. Wright & A. Miller, Federal Practice and Procedure § 2472, p. 455 (1971). In the federal judicial system, the rules generally governing the forfeiture of claims are set forth in Federal Rules of Criminal Procedure 51 and 52(b) and Federal Rule of Civil Procedure 46. The Tax Court, which is not an Article III court, has adopted a rule virtually identical to the latter, Tax Court Rule 144. These rules reflect the principle that a trial on the merits, whether in a civil or criminal case, is the “main event,” and not simply a “tryout on the road” to appellate review. Cf. Wainwright v. Sykes, 433 U. S. 72, 90 (1977). The very word “review” presupposes that a litigant’s arguments have been raised and considered in the tribunal of first instance. To abandon that principle is to encourage the practice of “sandbagging”: suggesting or permitting, for strategic reasons, that the trial court pursue a certain course, and later — if the outcome is unfavorable-claiming that the course followed was reversible error.
The blanket rule that “argument[s] premised on the Constitution’s structural separation of powers [are] not a matter of personal rights and therefore [are] not waivable,” Brief for Petitioners 43-44, would erode this cardinal principle of sound judicial administration. It has no support in principle or in precedent or in policy.
As to principle: Personal rights that happen to bear upon governmental structure are no more laden with public interest (and hence inherently nonwaivable by the individual) than many other personal rights one can conceive of. First *896Amendment free-speech rights, for example, or the Sixth Amendment right to a trial that is “public,” provide benefits to the entire society more important than many structural guarantees; but if the litigant does not assert them in a timely fashion, he is foreclosed. See, e. g., Head v. New Mexico Bd. of Examiners in Optometry, 374 U. S. 424, 432-433, n. 12 (1963) (First Amendment); Levine v. United States, 362 U. S. 610, 619 (1960) (Sixth Amendment). Nor is it distinctively true of structural guarantees that litigants often may undervalue them. Many criminal defendants, for example, would prefer a trial from which the press is excluded.
It is true, of course, that a litigant’s prior agreement to a judge’s expressed intention to disregard a structural limitation upon his power cannot have any legitimating effect— i. e., cannot render that disregard lawful. Even if both litigants not only agree to, but themselves propose, such a course, the judge must tell them no. But the question before us here involves the effect of waiver not ex ante but ex post — its effect not upon right but upon remedy: Must a judgment already rendered be set aside because of an alleged structural error to which the losing party did not properly object? There is no reason in principle why that should always be so. It will sometimes be so — not, however, because the error was structural, but because, whether structural or not, it deprived the federal court of its requisite subject-matter jurisdiction. .Such an error may be raised by a party, and indeed must be noticed sua sponte by a court, at all points in the litigation, see, e. g., American Fire & Casualty Co. v. Finn, 341 U. S. 6, 17-18 (1951); Mansfield, C. & L. M. R. Co. v. Swan, 111 U. S. 379, 382 (1884); Capron v. Van Noorden, 2 Cranch 126, 127 (1804). Since such a jurisdictional defect deprives not only the initial court but also the appellate court of its power over the case or controversy, to permit the appellate court to ignore it because of waiver would be to give the waiver legitimating, as opposed to merely remedial, effect, i. e., the effect of approving, ex *897ante, unlawful action by the appellate court itself. That this, rather than any principle of perpetual remediability of structural defects, is the basis for the rule of “nonwaivability” of lack of subject-matter jurisdiction is demonstrated by the fact that a final judgment cannot be attacked collaterally— i. e., before a court that does have jurisdiction — on the ground that a subject-matter jurisdictional limitation (structural or not) was ignored. See, e. g., Insurance Corp. of Ireland v. Compagnie des Bauxites de Guinee, 456 U. S. 694, 702, n. 9 (1982).
As to precedent: Petitioners place primary reliance on some broad language in Commodity Futures Trading Comm’n v. Schor, 478 U. S. 833 (1986). We said in that case that “[w]hen these Article III limitations are at issue” (referring not to all structural limitations of the Constitution, but only to those of Article III, §§ 1 and 2) “notions of consent and waiver cannot be dispositive.” 478 U. S., at 851. But the claim before us in Schor involved a particular sort of structural defect (a tribunal’s exceeding its subject-matter jurisdiction) which, as I have just described, had traditionally been held nonwaivable on appeal in the context of Article III tribunals. To extend the same treatment to similar defects in the context of non-Article III tribunals is quite natural, whether or not it is analytically required. Cf., e. g., Clapp v. Commissioner, 875 F. 2d 1396, 1399 (CA9 1989) (“While the Tax Court is an Article I court, . . . the few cases discussing the differences between the Tax Court and an Article III court indicate that questions of Tax Court jurisdiction are to be resolved in the same manner as for an Article III court”). It is clear from our opinion in Schor that we had the analogy to Article III subject-matter jurisdiction in mind. “To the extent that this structural principle is implicated in a given case,” we said, “the parties cannot by consent cure the constitutional difficulty for the same reason that the parties by consent cannot confer on federal courts subject-matter jurisdiction beyond the limitations imposed by Article III, § 2.” *898478 U. S., at 850-851.3 I would not extend that nonwaiver rule — a traditional rule in its application to Article III courts, and understandably extended to other federal adjudicative tribunals — to structural defects that do not call into question the jurisdiction of the forum. The subject-matter jurisdiction of the forum that issued the judgment, the Tax Court, is not in question in the present case.
Petitioners only other appeal to precedent is Glidden Co. v. Zdanok, 370 U. S. 530 (1962). That case did address a nonjurisdictional structural defect that had not been raised below. As the Court explains, however, that was a structural defect that went to the validity of the very proceeding under review, ante, at 879, as opposed to one that merely affected the validity of the claim — for example, improper appointment of the Executive officer who issued the regulation central to the controversy. That was considered significant by the only opinion in the case (that of Justice Harlan, joined by Justices Brennan and Stewart) to address the waiver point. (“The alleged defect of authority here relates to basic constitutional protections designed in part for the benefit of litigants.” Id., at 536 (emphasis added).) The formulation petitioners advance, of course, is much broader than that. (“[A]n argument premised on the Constitution’s structural separation of powers is not a matter of personal rights and *899therefore is not waivable” Brief for Petitioners 43-44 (emphasis added)). “There can be no hierarchy among separation of powers principles, in which some are fundamental and nonwaivable while the vindication of others may be relegated to the vagaries of individual litigation strategies.” Id., at 45.) Even more important, Justice Harlan’s plurality opinion in Glidden does not stand for the proposition that forfeiture can never be imposed, but rather the more limited proposition, which the Court reiterates today, that forfeiture need not always be imposed.
Several recent opinions flatly contradict petitioners’ blanket assertion that structural claims cannot be waived. Surely under our jurisprudence the so-called negative Commerce Clause is structural. See Dennis v. Higgins, 498 U. S. 439, 447 (1991). And surely the supposed guarantee against waivability must apply in state courts as well as in federal courts, since according to petitioners it emanates from the structural rights themselves. Yet only last Term, in Jimmy Swaggart Ministries v. Board of Equalization of California, 493 U. S. 378, 397 (1990), we declined to consider a negative Commerce Clause challenge to a state tax because state courts had found the issue procedurally barred as a result of petitioner’s failure to raise it in his administrative proceeding for a tax refund. And in G. D. Searle & Co. v. Cohn, 455 U. S. 404, 414 (1982), we declined to reach a negative Commerce Clause claim in litigation arising in the federal courts; we remanded the case for consideration of that issue by the Court of Appeals, “if it was properly raised below. ” (Emphasis added.) The Federal Courts of Appeals (even after Schor) have routinely applied the ordinary rules of forfeiture to structural claims not raised below. See, e. g., United States v. Doremus, 888 F. 2d 630, 633, n. 3 (CA9 1989) (separation of powers claims), cert. denied, 498 U. S. 1046 (1991); Opdyke Investment Co. v. Detroit, 883 F. 2d 1265, 1276 (CA6 1989) (same); Interface Group, Inc. v. *900Massachusetts Port Authority, 816 F. 2d 9, 16 (CA1 1987) (Breyer, J.) (Supremacy and Commerce Clause claims).
Finally, as to policy: The need for the traditional forfeiture rule — in cases involving structural claims as in all others — is obvious. Without that incentive to raise legal objections as soon as they are available, the time of lower court judges and of juries would frequently be expended uselessly, and appellate consideration of difficult questions would be less informed and less complete. Besides inviting these evils, the categorical rule petitioners advance would require the development of a whole new body of jurisprudence concerning which constitutional provisions are “structural” — a question whose answer is by no means clear. Cf. Sunstein, Government Control of Information, 74 Calif. L. Rev. 889, 915 (1986) (arguing that the First Amendment is structural). Moreover, since that rigid rule would cut so much against the grain of reason and practice, it would have the side effect of distorting substantive law. The maxim volenti non fit injuria has strong appeal in human affairs, and, by requiring it to be absolutely and systematically disregarded in cases involving structural protections of the Constitution, we would incline ourselves towards finding that no such structural protection exists.
Thus, the structural nature of the claim here is not sufficient reason to ignore its forfeiture — and the Court (though it discusses the virtues of structure at some length) does not pretend otherwise. There must be some additional reason, then, why the Court “exercise[s] our discretion,” ante, at 879, to disregard the forfeiture. To disregard it without sufficient reason is the exercise not of discretion but of whimsy. Yet beyond its discussion of structure, the only reason the Court gives is no reason at all: “we are faced with a constitutional challenge that is neither frivolous nor disingenuous,” ibid. That describes the situation with regard to the vast majority of forfeited claims raised on appeal. As we make clear in another case decided this very day, waiver generally *901extends not merely to “frivolous” and “disingenuous” challenges, but even to “[t]he most basic rights of criminal defendants.” Peretz, post, at 936. Here, petitioners expressly consented to the Special Trial Judge. Under 26 U. S. C. § 7443A, the Chief Judge of that court has broad discretion to assign cases to special trial judges. Any party who objects to such an assignment, if so inclined, can easily raise the constitutional issue pressed today. Under these circumstances, I see no reason why this should be included among those “rare cases in which we should exercise our discretion,” ante, at 879, to hear a forfeited claim.
II
Having struggled to reach petitioners’ Appointments Clause objection, the Court finds it invalid. I agree with that conclusion, but the reason the Court assigns is in my view both wrong and full of danger for the future of our system of separate and coequal powers.
The Appointments Clause provides:
“[T]he Congress may by Law vest the Appointment of such inferior Officers, as they think proper, in the President alone, in the Courts of Law, or in the Heads of Departments.” Art. II, §2, cl. 2.
I agree with the Court that a special trial judge is an “inferior Office[r]” within the meaning of this Clause, with the result that, absent Presidential appointment, he must be appointed by a court of law or the head, of a department. I do not agree, however, with the Court’s conclusion that the Tax Court is a “Cour[t] of Law” within the meaning of this provision. I would find the appointment valid because the Tax Court is a “Departmen[t]” and the Chief Judge is its head.
A
A careful reading of the Constitution and attention to the apparent purpose of the Appointments Clause make it clear that the Tax Court cannot be one of those “Courts of Law” *902referred to there. The Clause does not refer generally to “Bodies exercising judicial Functions,” or even to “Courts” generally, or even to “Courts of Law” generally. It refers to “the Courts of Law.” Certainly this does not mean any “Cour[t] of Law” (the Supreme Court of Rhode Island would not do). The definite article “the” obviously narrows the class of eligible “Courts of Law” to those courts of law envisioned by the Constitution. Those are Article III courts, and the Tax Court is not one of them.
The Court rejects this conclusion because the Appointments Clause does not (in the style of the Uniform Commercial Code) contain an explicit cross-reference to Article III. Ante, at 888-889. This is no doubt true, but irrelevant. It is equally true that Article I, § 8, cl. 9, which provides that Congress may “constitute Tribunals inferior to the supreme Court,” does not explicitly say “Tribunals under Article III, below.” Yet, this power “plainly relates to the 'inferior Courts’ provided for in Article III, § 1; it has never been relied on for establishment of any other tribunals.” Glidden Co. v. Zdanok, 370 U. S., at 543 (opinion of Harlan, J.); see also 3 J. Story, Commentaries on the Constitution of the United States § 1573, p. 437 (1833). Today’s Court evidently does not appreciate, as Chief Justice Marshall did, that the Constitution does not “partake of the prolixity of a legal code,” McCulloch v. Maryland, 4 Wheat. 316, 407 (1819). It does not, like our opinions, bristle with “supras,” “infras,,” and footnotes. Instead of insisting upon such legalisms, we should, I submit, follow the course mapped out in Buckley v. Valeo, 424 U. S. 1, 124 (1976) (per curiam), and examine the Appointments Clause of the Constitution in light of the “cognate provisions” of which it is a central feature: Article I, Article II, and Article III. The only “Courts of Law” referred to there are those authorized by Article III, § 1, whose judges serve during good behavior with undiminishable salary. Art. III, § 1. See Glidden Co. v. Zdanok, supra, at 543 (opinion of Harlan, J.); United States v. Mouat, 124 *903U. S. 303, 307 (1888) (“courts of justice”) (dictum). The Framers contemplated no other national judicial tribunals. “According to the plan of the convention, all judges who may be appointed by the United States are to hold their offices during good behavior . . . .” The Federalist No. 78, p. 465 (C. Rossiter ed. 1961) (A. Hamilton) (second emphasis in original).
We recognized this in Buckley, supra, and it was indeed an essential part of our reasoning. Responding to the argument that a select group of Congressmen was a “Department,” we said:
“The phrase 'Heads of Departments/ used as it is in conjunction with the phrase ‘Courts of Law,’ suggests that the Departments referred to are themselves in the Executive Branch or at least have some connection with that branch. While the Clause expressly authorizes Congress to vest the appointment of certain officers in the ‘Courts of Law/ the absence of similar language to include Congress must mean that neither Congress nor its officers were included within the language ‘Heads of Departments’ in this part of cl. 2.
“Thus, with respect to four of the six voting members of the Commission, neither the President, the head of any department, nor the Judiciary has any voice in their selection.” Id., at 127 (emphasis added).
The whole point of this passage is that “the Heads of Departments” must reasonably be understood to refer exclusively to the Executive Branch (thereby excluding officers of Congress) because “the Courts of Law” obviously refers exclusively to the Judicial Branch. We were right in Buckley, and the Court is wrong today.
Even if the Framers had no particular purpose in making the Appointments Clause refer only to Article III courts, we would still of course be bound by that disposition. In fact, however, there is every reason to believe that the Appointments Clause’s limitation to Article III tribunals was adopted *904with calculation and forethought, faithfully implementing a considered political theory for the appointment of officers.
The Framers’ experience with postrevolutionary self-government had taught them that combining the power to create offices with the power to appoint officers was a recipe for legislative corruption.4 The foremost danger was that legislators would create offices with the expectancy of occupying them themselves. This was guarded against by the Incompatibility and Ineligibility Clauses, Article I, § 6, cl. 2. See Buckley, supra, at 124. But real, if less obvious, dangers remained. Even if legislators could not appoint themselves, they would be inclined to appoint their friends and supporters. This proclivity would be unchecked because of the lack of accountability in a multimember body — as *905James Wilson pointed out in his criticism of a multimember executive:
“[A]re impartiality and fine discernment likely to predominate in a numerous [appointing] body? In proportion to their own number, will be the number of their friends, favorites and dependents. An office is to be filled. A person nearly connected, by some of the foregoing ties, with one of those who [is] to vote in filling it, is named as a candidate. . . . Every member, who gives, on his account, a vote for his friend, will expect the return of a similar favor on the first convenient opportunity. In this manner, a reciprocal intercourse of partiality, of interestedness, of favoritism, perhaps of venality, is established; and in no particular instance, is there a practicability of tracing the poison to its source. Ignorant, vicious, and prostituted characters are introduced into office; and some of those, who voted, and procured others to vote for them, are the first and loudest in expressing their astonishment, that the door of admission was ever opened to men of their infamous description. The suffering people are thus wounded and buffeted, like Homer’s Ajax, in the dark; and have not even the melancholy satisfaction of knowing by whom the blows are given.” 1 Works of James Wilson 359-360 (J. Andrews ed. 1896).
See also Essex Result, in Memoir of Theophilus Parsons 381-382 (1859); The Federalist No. 76, pp. 455-457 (C. Rossiter ed. 1961) (A. Hamilton). And not only would unaccountable legislatures introduce their friends into necessary offices, they would create unnecessary offices into which to introduce their friends. As James Madison observed:
“The power of the Legislature to appoint any other than their own officers departs too far from the Theory which requires a separation of the great Departments of Government. One of the best securities against the ere-*906ation of unnecessary offices or tyrannical powers is an exclusion of the authors from all share in filling the one, or influence in the execution of the other.” Madison’s Observations on Jefferson’s Draft of a Constitution for Virginia, reprinted in 6 Papers of Thomas Jefferson 308, 311 (J. Boyd ed. 1952).
Por these good and sufficient reasons, then, the federal appointment power was removed from Congress. The Framers knew, however, that it was not enough simply to define in writing who would exercise this power or that. “After discriminating ... in theory, the several classes of power, as they may in their nature be legislative, executive, or judiciary, the next and most difficult task [was] to provide some practical security for each, against the invasion of the others.” The Federalist No. 48, p. 308 (C. Rossiter ed. 1961) (J. Madison). Invasion by the Legislature, of course, was the principal threat, since the “legislative authority . . . possesses so many means of operating on the motives of the other departments.” Id., No. 49, p. 314 (J. Madison). It can “mask, under complicated and indirect measures, the encroachments which it makes on the co-ordinate departments,” id., No. 48, p. 310 (J. Madison), and thus control the nominal actions (e. g., appointments) of the other branches. Cf. T. Jefferson, Notes on the State of Virginia 120 (W. Peden ed. 1955).
Thus, it was not enough simply to repose the power to execute the laws (or to appoint) in the President; it was also necessary to provide him with the means to resist legislative encroachment upon that power. The means selected were various, including a separate political constituency, to which he alone was responsible, and the power to veto encroaching laws, see Art. I, § 7, or even to disregard them when they are unconstitutional. See Easterbrook, Presidential Review, 40 Case W. Res. L. Rev. 905, 920-924 (1990). One of the most obvious and necessary, however, was a permanent salary. Art. II, § 1. Without this, “the separation of the *907executive from the legislative department would be merely nominal and nugatory. The legislature, with a discretionary power over the salary and emoluments of the Chief Magistrate, could render him as obsequious to their will as they might think proper to make him.” The Federalist No. 73, p. 441 (C. Rossiter ed. 1961) (A. Hamilton). See also id., No. 51, p. 321 (J. Madison); Mass. Const., Part The Second, Chapter II, § 1, Art. XIII (1780).
A power of appointment lodged in a President surrounded by such structural fortifications could be expected to be exercised independently, and not pursuant to the manipulations of Congress. The same is true, to almost the same degree, of the appointment power lodged in the heads of departments. Like the President, these individuals possess a reputational stake in the quality of the individuals they appoint; and though they are not themselves able to resist congressional encroachment, they are directly answerable to the President, who is responsible to his constituency for their appointments and has the motive and means to assure faithful actions by his direct lieutenants.
Like the President, the Judicial Branch was separated from Congress not merely by a paper assignment of functions, but by endowment with the means to resist encroachment-foremost among which, of course, are life tenure (during “good behavior”) and permanent salary. These structural accoutrements not only assure the fearless adjudication of cases and controversies, see The Federalist Nos. 78, 79 (A. Hamilton); Northern Pipeline Constr. Co. v. Marathon Pipe Line Co., 458 U. S. 50, 57-60 (1982) (opinion of Brennan, J.); they also render the Judiciary a potential repository of appointment power free of congressional (as well as Presidential) influence. In the same way that depositing appointment power in a fortified President and his lieutenants ensures an actual exclusion of the legislature from appointment, so too does reposing such power in an Article III court. The Court’s holding, that Congress may deposit *908such power in “legislative courts,” without regard to whether their personnel are either Article III judges or “Heads of Departments,” utterly destroys this carefully constructed scheme. And the Court produces this result, I remind the reader, not because of, but in spite of, the apparent meaning of the phrase “the Courts of Law.”
B
Having concluded, against all odds, that “the Courts of Law” referred to in Article II, § 2, are not the courts of law established by Article III, the Court is confronted with the difficult problem of determining what courts of law they are. It acknowledges that they must be courts which exercise “the judicial power of the United States” and concludes that the Tax Court is such a court — even though it is not an Article III court. This is quite a feat, considering that Article III begins “The judicial Power of the United States” — not “Some o/’the judicial Power of the United States,” or even “Most of the judicial Power of the United States” — “shall be vested in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish.” Despite this unequivocal text, the Court sets forth the startling proposition that “the judicial power of the United States is not limited to the judicial power defined under Article III.” Ante, at 889. It turns out, however — to our relief, I suppose it must be said — that this is really only a pun. “The judicial power,” as the Court uses it, bears no resemblance to the constitutional term of art we are all familiar with, but means only “the power to adjudicate in the manner of courts.” So used, as I shall proceed to explain, the phrase covers an infinite variety of individuals exercising executive rather than judicial power (in the constitutional sense), and has nothing to do with the separation of powers or with any other characteristic that might cause one to believe that is what was meant by “the Courts of Law.” As far as I can tell, the only thing to be said for this approach is that it makes the Tax *909Court a “Cour[t] of Law” — which is perhaps the object of the exercise.
I agree with the unremarkable proposition that “Congress [has] wide discretion to assign the task of adjudication in cases arising under federal law to legislative tribunals.” Ante, at 889. Congress may also assign that task to subdivisions of traditional executive departments, as it did in 1924 when it created the Tax Court’s predecessor, the Tax Board of Appeals — or to take a more venerable example, as it did in 1791 when it created within the Treasury Department the Comptroller of the United States, who “decide[d] on appeal, without further review by the Secretary, all claims concerning the settlement of accounts.” Casper, An Essay in Separation of Powers: Some Early Versions and Practices, 30 Wm. & Mary L. Rev. 211, 238 (1989); see 1 Stat. 66. Such tribunals, like any other administrative board, exercise the executive power, not the judicial power of the United States. They are, in the words of the great Chief Justice, “incapable of receiving [the judicial power]” — unless their members serve for life during good behavior and receive permanent salary. American Ins. Co. v. Canter, 1 Pet. 511, 546 (1828) (Marshall, C. J.).
It is no doubt true that all such bodies “adjudicate,” i. e., they determine facts, apply a rule of law to those facts, and thus arrive at a decision. But there is nothing “inherently judicial” about “adjudication.” To be a federal officer and to adjudicate are necessary but not sufficient conditions for the exercise of federal judicial power, as we recognized almost a century and a half ago.
“That the auditing of the accounts of a receiver of public moneys may be, in an enlarged sense, a judicial act, must be admitted. So are all those administrative duties the performance of which involves an inquiry into the existence of facts and the application to them of rules of law. In this sense the act of the President in calling out the militia under the act of 1795, [Martin v. Mott,] *91012 Wheat. 19 [(1827)], or of a commissioner who makes a certificate for the extradition of a criminal, under a treaty, is judicial. But it is not sufficient to bring such matters under the judicial power, that they involve the exercise of judgment upon law and fact.” Murray’s Lessee v. Hoboken Land & Improvement Co., 18 How. 272, 280 (1856).
Accord, Bator, The Constitution as Architecture: Legislative and Administrative Courts Under Article III, 65 Ind. L. J. 233, 264-265 (1990). The first Patent Board, which consisted of Thomas Jefferson, Henry Knox, and Edmund Randolph in their capacities as Secretary of State, Secretary of War, and Attorney General, respectively, 1 Stat. 109, 110 (1790), adjudicated the patentability of inventions, sometimes hearing argument by petitioners. See 18 J. Pat. Off. Soc. 68-69 (July 1936). They were exercising the executive power. See Easterbrook, “Success” and the Judicial Power, 65 Ind. L. J. 277, 280 (1990). Today, the Federal Government has a corps of administrative law judges numbering more than 1,000, whose principal statutory function is the conduct of adjudication under the Administrative Procedure Act (APA), see 5 U. S. C. §§ 554, 3105. They are all executive officers. “Adjudication,” in other words, is no more an “inherently” judicial function than the promulgation of rules governing primary conduct is an “inherently” legislative one. See Standard Oil Co. of New Jersey v. United States, 221 U. S. 1 (1911); APA, 5 U. S. C. § 553 (“Rule making”).
It is true that Congress may commit the sorts of matters administrative law judges and other executive adjudicators now handle to Article III courts — just as some of the matters now in Article III courts could instead be committed to executive adjudicators. “[TJhere are matters, involving public rights, which may be presented in such form that the judicial power is capable of acting on them, and which are susceptible of judicial determination, but which Congress may or may not bring within the cognizance of the courts of the United *911States, as it may deem proper.” Murray’s Lessee, supra, at 284. See also Ex parte Bakelite Corp., 279 U. S. 438, 451 (1929). Congress could, for instance, allow direct review by the courts of appeals of denials of Social Security benefits. It could instead establish the Social Security Court — composed of judges serving 5-year terms — within the Social Security Administration. Both tribunals would perform identical functions, but only the former would exercise the judicial power.
In short, given the performance of adjudicatory functions by a federal officer, it is the identity of the officer — not something intrinsic about the mode of decisionmaking or type of decision — that tells us whether the judicial power is being exercised. “[0]ur cases demonstrate [that] a particular function, like a chameleon, will often take on the aspect of the office to which it is assigned.” Bowsher v. Synar, 478 U. S. 714, 749 (1986) (Stevens, J., concurring in judgment). Cf. INS v. Chadha, 462 U. S. 919, 953, n. 16 (1983). Where adjudicative decisionmakers do not possess life tenure and a permanent salary, they are “incapable of exercising any portion of the judicial power.” Ex parte Randolph, 20 F. Cas. 242, 254 (No. 11,558) (CC Va. 1833) (Marshall, C. J.).
The Tax Court is indistinguishable from my hypothetical Social Security Court. It reviews determinations by Executive Branch officials (the Internal Revenue Service) that this much or that much tax is owed — a classic executive function. For 18 years its predecessor, the Board of Tax Appeals, did the very same thing, see H. Dubroff, The United States Tax Court 47-175 (1979), and no one suggested that body exercised “the judicial power.” We held just the opposite:
“The Board of Tax Appeals is not a court. It is an executive or administrative board, upon the decision of which the parties are given an opportunity to base a petition for review to the courts after the administrative inquiry of the Board has been had and decided.” Old *912Colony Trust Co. v. Commissioner, 279 U. S. 716, 725 (1929) (Taft, C. J.).
Though renamed “the Tax Court of the United States” in 1942, it remained “an independent agency in the Executive Branch,” 26 U. S. C. § 1100 (1952 ed.), and continued to perform the same function. As an executive agency, it possessed many of the accoutrements the Court considers “quintessentially judicial,” ante, at 891. It administered oaths, for example, and subpoenaed and examined witnesses, § 1114; its findings were reviewed “in the same manner and to the same extent as decisions of the district courts in civil actions tried without a jury,” § 1141(a). This Court continued to treat it as an administrative agency, akin to the Federal Communications Commission (FCC) or the National Labor Relations Board (NLRB). See Dobson v. Commissioner, 320 U. S. 489, 495-501 (1943).
When the Tax Court was statutorily denominated an “Article I Court” in 1969, its judges did not magically acquire the judicial power. They still lack life tenure; their salaries may still be diminished; they are still removable by the President for “inefficiency, neglect of duty, or malfeasance in office.” 26 U. S. C. § 7443(f). (In Bowsher v. Synar, supra, at 729, we held that these latter terms are “very broad” and “could sustain removal ... for any number of actual or perceived transgressions.”) How anyone with these characteristics can exercise judicial power “independent. . . [of] the Executive Branch” is a complete mystery. It seems to me entirely obvious that the Tax Court, like the Internal Revenue Service, the FCC, and the NLRB, exercises executive power. Amar, Marbury, Section 13, and the Original Jurisdiction of the Supreme Court, 56 U. Chi. L. Rev. 443, 451, n. 43 (1989). See also Northern Pipeline, 458 U. S., at 113 (White, J., dissenting) (equating administrative agencies and Article I courts); Samuels, Kramer & Co. v. Commissioner, 930 F. 2d 975, 992-993 (CA2 1991) (collecting academic authorities for same proposition).
*913In seeking to establish that “judicial power” in some constitutionally significant sense — in a sense different from the adjudicative exercise of executive power — can be exercised by someone other than an Article III judge, the Court relies heavily upon the existence of territorial courts. Ante, at 889-891. Those courts have nothing to do with the issue before us.5 I agree that they do not exercise the national executive power — but neither do they exercise any national judicial power. They are neither Article III courts nor Article I courts, but Article IV courts — just as territorial governors are not Article I executives but Article IV executives.
“These Courts, then, are not constitutional Courts, in which the judicial power conferred by the Constitution on the general government, can be deposited. They are incapable of receiving it. They are legislative Courts, created in virtue of the general right of sovereignty which exists in the government, or in virtue of that clause which enables Congress to make all needful rules and regulations, respecting the territory belonging to the United States. ... In legislating for them, Congress exercises the combined powers of the general, and of a state government.” American Ins. Co. v. Canter, 1 Pet., at 546 (Marshall, C. J.) (emphasis added).
*914Or as the Court later described it:
“[Territories] are not organized under the Constitution, nor subject to its complex distribution of the powers of government, as the organic law; but are the creations, exclusively, of the legislative department, and subject to its supervision and control.” Benner v. Porter, 9 How. 235, 242 (1850).
Thus, Congress may endow territorial governments with a plural executive; it may allow the executive to legislate; it may dispense with the legislature or judiciary altogether. It should be obvious that the powers exercised by territorial courts tell us nothing about the nature of an entity, like the Tax Court, which administers the general laws of the Nation. See Northern Pipeline, supra, at 75-76 (opinion of Brennan, J.).
The Court claims that there is a “longstanding practice” of permitting Article I courts to appoint inferior officers. Ante, at 890. I am unaware of such a practice. Perhaps the Court means to refer not to Article I courts but to the territorial courts just discussed, in which case the practice would be irrelevant. As I shall discuss below, an Article I court (such as the Tax Court) that is not within any other department would be able to have its inferior officers appointed by its chief judge — not under the “Courts of Law” provision of Article II, § 2, but under the “Heads of Departments” provision; perhaps it is that sort of practice the Court has in mind. It is certain, in any case, that no decision of ours has ever approved the appointment of an inferior officer by an Article I court. Ex parte Hennen, 13 Pet. 230 (1839), which the Court cites, involved appointment by an Article III tribunal.
I — I I — I
Since the Tax Court is not a court of law, unless the Chief Judge is the head of a department, the appointment of the Special Trial Judge was void. Unlike the Court, I think he is.
*915I have already explained that the Tax Court, like its predecessors, exercises the executive power of the United States. This does not, of course, suffice to make it a “Departmen[t]” for purposes of the Appointments Clause. If, for instance, the Tax Court were a subdivision of the Department of the Treasury — as the Board of Tax Appeals used to be — it would not qualify. In fact, however, the Tax Court is a freestanding, self-contained entity in the Executive Branch, whose Chief Judge is removable by the President (and, save impeachment, no one else). Nevertheless, the Court holds that the Chief Judge is not the head of a department.
It is not at all clear what the Court’s reason for this conclusion is. I had originally thought that the Court was adopting petitioners’ theory — wrong, but at least coherent — that “Heads of Departments” means Cabinet officers. This is suggested by the Court’s reliance upon the dictum in Burnap v. United States, 252 U. S. 512, 515 (1920), to the effect that the head of a department must be “‘the Secretary in charge of a great division of the executive branch of the Government, like the State, Treasury, and War, who is a member of the Cabinet,”’ ante, at 886 (emphasis added); by the Court’s observation that “[t]he Cabinet-level departments are limited in number and easily identified,” ibid.; and by its reliance upon the fact that in the Twenty-fifth Amendment “the term ‘department’ refers to Cabinet-level entities,” ante, at 887. Elsewhere, however, the Court seemingly disclaims Cabinet status as the criterion, see ante, at 887, n. 4, and says that the term “Departmen[t]” means “executive divisions like the Cabinet-level departments,” ante, at 886 (emphasis added). Unfortunately, it never specifies what characteristic it is that causes an agency to be “like a Cabinet-level department,” or even provides any intelligible clues as to what it might have in mind. It quotes a congressional Committee Report that seemingly equates Cabinet status with inclusion within the statutory defintion of “ ‘department’ ” in 5 U. S. C. § 101, ante, at 887 (quoting H. R. Rep. *916No. 203, 89th Cong., 1st Sess., 3 (1965)), but this hint is canceled by a footnote making it clear that “Cabinet-like” character, whatever it is, is not “strictly limit[ed]” by that provision, ante, at 887, n. 4. Its approving quotation of Burnap’s reference to “a great division of the executive branch” might invite the guess that numerosity is the key — but the Department of Education has fewer than 5,000 employees, and the FCC, which the Court also appears willing to consider such a “‘great division,’” ante, at 886, fewer than 1,800. See Employment and Trends as of March 1991, Office of Personnel Management, Table 2. The Court reserves the right to consider as “Cabinet-like” and hence as “Departments” those agencies which, above all others, are at the farthest remove from Cabinet status, and whose heads are specifically designed not to have the quality that the Court earlier thinks important, of being “subject to the exercise of political oversight and shar[ing] the President’s accountability to the people,” ante, at 886 — namely, independent regulatory agencies such as the Federal Trade Commission and the Securities and Exchange Commission, ante, at 887, n. 4. Indeed, lest any conceivable improbability be excluded, the Court even reserves the right to consider as a “Departmen[t]” an entity that is not headed by an officer of the United States — the Federal Reserve Bank of St. Louis, whose president is appointed in none of the manners constitutionally permitted for federal officers, but rather by a Board of Directors, two-thirds of whom are elected by regional banks, see 12 U. S. C. §§302, 304, and 341. It is as impossible to respond to this random argumentation as it is to derive a comprehensible theory of the appointments power from it. I shall address, therefore, what was petitioners’ point, what I originally took to be the point of the Court’s opinion, and what is the only trace of a flesh-and-blood point that subsists: the proposition that “Department” means “Cabinet-level agency.”
There is no basis in text or precedent for this position. The term “Cabinet” does not appear in the Constitution, the *917Founders having rejected proposals to create a Cabinet-like entity. See H. Learned, The President’s Cabinet 74-94 (1912); E. Corwin, The President 97, 238-240 (5th rev. ed. 1984). The existence of a Cabinet, its membership, and its prerogatives (except to the extent the Twenty-fifth Amendment speaks to them), are entirely matters of Presidential discretion. Nor does any of our cases hold that “the Heads of Departments” are Cabinet members. In United States v. Germaine, 99 U. S. 508 (1879), we merely held that the Commissioner of Pensions, an official within the Interior Department, was not the head of a department. And, in Bitmap, supra, we held that the Bureau of Public Buildings and Grounds, a bureau within the War Department, was not a department.
The Court’s reliance on the Twenty-fifth Amendment is misplaced. I accept that the phrase “the principal officers of the executive departments” is limited to members of the Cabinet. It is the structural composition of the phrase, however, and not the single word “departments” which gives it that narrow meaning — “the principal officers” of the “executive departments” in gross, rather than (as in the Opinions Clause) “the principal Officer in each of the executive Departments,” or (in the Appointments Clause) simply “the Heads” (not “principal Heads”) “of Departments.”
The only history on the point also militates against the Court’s conclusion. The 1792 Congress passed an “Act to establish the Post-Office and Post Roads within the United States,” creating a Postmaster General and empowering him to appoint “an assistant, and deputy postmasters, at all places where such may be found necessary.” § 3, 1 Stat. 234. President Washington did not bring the Postmaster into his Cabinet. See Learned, supra, at 233-249. It seems likely that the Assistant Postmaster General and Deputy Postmasters were inferior officers — which means either that “the Heads of Departments” include principal officers other than the Cabinet, or that the Second Congress and President *918Washington did not understand the Appointments Clause. In any case, it is silly to think that the Second Congress (or any later Congress) was supposed to guess whether the President would bring the new agency into his Cabinet in order to determine how the appointment of its inferior officers could be made.
Modern practice as well as original practice refutes the distinction between Cabinet and non-Cabinet agencies. Congress has empowered non-Cabinet agencies to appoint inferior officers for quite some time. See, e. g., 47 U. S. C. § 155(f) (FCC — managing director); 15 U. S. C. § 78d(b) (Securities and Exchange Commission — “such officers ... as may be necessary”); 15 U. S. C. § 42 (Federal Trade Commission-secretary); 7 U. S. C. § 4a(c) (Commodity Futures Trading Commission — general counsel). In fact, I know of very few inferior officers in the independent agencies who are appointed by the President, and of none who is appointed by the head of a Cabinet department. The Court’s interpretation of “Heads of Departments” casts into doubt the validity of many appointments and a number of explicit statutory authorizations to appoint.
A number of factors support the proposition that “Heads of Departments” includes the heads of all agencies immediately below the President in the organizational structure of the Executive Branch. It is quite likely that the “Departments” referred to in the Opinions Clause (“The President. . . may require the Opinion, in writing, of the principal Officer in each of the executive Departments,” Art. II, § 2) are the same as the “Departments” in the Appointments Clause. See Germaine, supra, at 511. In the former context, it seems to me, the word must reasonably be thought to include all independent establishments.. The purpose of the Opinions Clause, presumably, was to assure the President’s ability to get a written opinion on all important matters. But if the “Departments” it referred to were only Cabinet departments, *919it would not assure the current President the ability to receive a written opinion concerning the operations of the Central Intelligence Agency, an agency that is not within any other department, and whose Director is not a member of the Cabinet.
This evident meaning — that the term “Departments” means all independent executive establishments — is also the only construction that makes sense of Article II, §2’s sharp distinction between principal officers and inferior officers. The latter, as we have seen, can by statute be made appoint-able by “the President alone, . . . the Courts of Law, or . . . the Heads of Departments.” Officers that are not “inferior Officers,” however, must be appointed (unless the Constitution itself specifies otherwise, as it does, for example, with respect to officers of Congress) by the President, “by and with the Advice and Consent of the Senate.” The obvious purpose of this scheme is to make sure that all the business of the Executive will be conducted under the supervision of officers appointed by the President with Senate approval; only officers “inferior,” i. e., subordinate, to those can be appointed in some other fashion. If the Appointments Clause is read as I read it, all inferior officers can be made ap-pointable by their ultimate (sub-Presidential) superiors; as petitioners would read it, only those inferior officers whose ultimate superiors happen to be Cabinet members can be. All the other inferior officers, if they are to be appointed by an Executive official at all, must be appointed by the President himself or (assuming cross-department appointments are permissible) by a Cabinet officer who has no authority over the appointees. This seems to me a most implausible disposition, particularly since the makeup of the Cabinet is not specified in the Constitution, or indeed the concept even mentioned. It makes no sense to create a system in which the inferior officers of the Environmental Protection Agency, *920for example — which may include, inter alios, bureau chiefs, the general counsel, and administrative law judges — must be appointed by the President, the courts of law, or the “Secretary of Something Else.”
In short, there is no reason, in text, judicial decision, history, or policy, to limit the phrase “the Heads of Departments” in the Appointments Clause to those officials who are members of the President’s Cabinet. I would give the term its ordinary meaning, something which Congress has apparently been doing for decades without complaint. As an American dictionary roughly contemporaneous with adoption of the Appointments Clause provided, and as remains the case, a department is “[a] separate allotment or part of business; a distinct province, in which a class of duties are allotted to a particular person . . . .” IN. Webster, American Dictionary 58 (1828). I readily acknowledge that applying this word to an entity such as the Tax Court would have seemed strange to the Founders, as it continues to seem strange to modern ears. But that is only because the Founders did not envision that an independent establishment of such small size and specialized function would be created. They chose the word “Departmen[tj,” however, not to connote size or function (much less Cabinet status), but separate organization — a connotation that still endures even in colloquial usage today (“that is not my department”). The Constitution is clear, I think, about the chain of appointment and supervision that it envisions: Principal officers could be permitted by law to appoint their subordinates. That should subsist, however much the nature of federal business or of federal organizational structure may alter.
I must confess that in the case of the Tax Court, as with some other independent establishments (notably, the so-called “independent regulatory agencies” such as the FCC and the Federal Trade Commission) permitting appointment of inferior officers by the agency head may not ensure the *921high degree of insulation from congressional control that was the purpose of the appointments scheme elaborated in the Constitution. That is a consequence of our decision in Humphrey’s Executor v. United States, 295 U. S. 602 (1935), which approved congressional restriction upon arbitrary dismissal of the heads of such agencies by the President, a scheme avowedly designed to made such agencies less accountable to him, and hence he less responsible for them. Depending upon how broadly one reads the President’s power to dismiss “for cause,” it may be that he has no control over the appointment of inferior officers in such agencies; and if those agencies are publicly regarded as beyond his control — a “headless Fourth Branch” — he may have less incentive to care about such appointments. It could be argued, then, that much of the raison d’etre for permitting appointive power to be lodged in “Heads of Departments,” see supra, at 903-908, does not exist with respect to the heads of these agencies, because they, in fact, will not be shored up by the President and are thus not resistant to congressional pressures. That is a reasonable position — though I tend to the view that adjusting the remainder of the Constitution to compensate for Humphrey’s Executor is a fruitless endeavor. But in any event it is not a reasonable position that supports the Court’s decision today — both because a “Cour[t] of Law” artificially defined as the Court defines it is even less resistent to those pressures, and because the distinction between those agencies that are subject to full Presidential control and those that are not is entirely unrelated to the distinction between Cabinet agencies and non-Cabinet agencies, and to all the other distinctions that the Court successively embraces. (The Central Intelligence Agency and the Environmental Protection Agency, for example, though not Cabinet agencies or components of Cabinet agencies, are not “independent” agencies in the sense of independence from Presidential control.) *922In sum, whatever may be the distorting effects of later innovations that this Court has approved, considering the Chief Judge of the Tax Court to be the head of a department seems to me the only reasonable construction of Article II, §2.
=1= * *
For the above reasons, I concur in the judgment that the decision below must be affirmed.
I have no quarrel with the Court’s decision to entertain petitioners’ statutory claim on the merits, as that claim was resolved on the merits by the Court of Appeals. See Virginia Bankshares, Inc. v. Sandberg, post, at 1099, n. 8 (1991).
The Court uses the term “waive” instead of “forfeit,” see ante, at 878-880. The two are really not the same, although our cases have so often used them interchangeably that it may be too late to introduce precision. Waiver, the “intentional relinquishment or abandonment of a known right or privilege,” Johnson v. Zerbst, 304 U. S. 458, 464 (1938), is merely one means by which a forfeiture may occur. Some rights may be forfeited by means short of waiver, see, e. g., Levine v. United States, 362 U. S. 610, 619 (1960) (right to public trial); United States v. Bascaro, 742 F. 2d 1335, 1365 (CA11 1984) (right against double jeopardy), cert. denied sub nom. Hobson v. United States, 472 U. S. 1017 (1985); United States v. Whitten, 706 F. 2d 1000, 1018, n. 7 (CA9 1983) (right to confront adverse witnesses), cert. denied, 465 U. S. 1100 (1984), but others may not, see, e. g., Johnson, supra (right to counsel); Patton v. United States, 281 U. S. 276, 312 (1930) (right to trial by jury). A right that cannot be waived cannot be *895forfeited by other means (at least in the same proceeding), but the converse is not true.
In this case, petitioners expressly consented to the Special Trial Judge’s role. As far as my analysis is concerned, however, it would not matter if an even more inadvertent forfeiture were involved — that is, if petitioners had not even consented but had merely failed to object in timely fashion. I shall not try to retain the distinction between waiver and forfeiture throughout this opinion, since many of the sources I shall be using disregard it.
Ironically enough, the categorical “no-waiver” rule that petitioners propose would destroy the very parallelism between administrative and judicial tribunals that Schor sought to achieve. For we have held that, in the administrative context, the use of unauthorized personnel to conduct a hearing (a hearing examiner not properly appointed pursuant to the Administrative Procedure Act) would not justify judicial reversal of the agency decision where no objection was lodged before the agency itself: “[W]e hold that the defect in the examiner’s appointment was an irregularity which would invalidate a resulting order if the Commission had overruled an appropriate objection made during the hearings. But it is not one which deprives the Commission of power or jurisdiction, so that even in the absence of timely objection its order should be set aside as a nullity.” United States v. L. A. Tucker Truck Lines, Inc., 344 U. S. 33, 38 (1952).
The Court apparently thinks that the Appointments Clause was designed to check executive despotism. Ante, at 883-884. This is not what we said in Buckley v. Valeo, 424 U. S. 1, 129 (1976), and it is quite simply contrary to historical fact. The quotations on which the Court relies describe abuses by the unelected royal governors and the Crown, who possessed the power to create and fill offices. The drafters of several early State Constitutions reacted to these abuses by lodging the appointment power in the legislature. See, e. g., Va. Const. (1776) (legislature appoints judges); cf. Articles of Confederation, Art. IX (Congress appoints courts and officers of land forces). Americans soon learned, however, that “in a representative republic where the executive magistracy is carefully limited ... it is against the enterprising ambition of the [legislative] department that the people ought to indulge all their jealousy and exhaust all their precautions.” The Federalist No. 48, p. 309 (C. Rossiter ed. 1961) (J. Madison). Soon after the revolution, “[t]he appointing authority which in most constitutions had been granted to the assemblies had become the principal source of division and faction in the states.” G. Wood, The Creation of The American Republic, 1776-1787, p. 407 (1969). By 1780, States were reacting to these abuses by reposing appointment authority in the executive. See Mass. Const., Part The Second, Chapter II, § 1, Art. IX (1780); N. H. Const. (1784) (officers appointed by president and a council). On legislative despotism, see generally Wood, supra, at 403-409. The Framers followed the lead of these later Constitutions. The Appointments Clause is, intentionally and self-evidently, a limitation on Congress.
Sadly, the Court also relies on dicta in Williams v. United States, 289 U. S. 553 (1933), an opinion whose understanding of the principles of separation of powers ought not inspire confidence, much less prompt emulation. It includes, for example, the notion that all disputes over which Article III provides jurisdiction can only be committed to Article III courts, id., at 580-581; see also D. Currie, Federal Courts 145-146 (1982) — which would make the Tax Court unconstitutional. Williams has been declared an “intellectual disaster” by commentators. P. Bator, D. Meltzer, P. Miskin, & D. Shapiro, Hart & Wechsler’s The Federal Courts and The Federal System 468 (3d ed. 1988); Bator, The Constitution as Architecture: Legislative And Administrative Courts Under Article III, 65 Ind. L. J. 233, 242-243, n. 30 (1990) (“I could devote a whole lecture to the ways in which [the reasoning of Williams] is erroneous”).