Concurring in Part and Dissenting in Part:
The majority today decides questions of great importance: May a State, consistent with the Constitution, impose a lifetime term limitation upon elected officials for state office, and, if so, how may the State go about enacting such a restriction? These are questions of first impression among the federal courts of appeal. With regret, I must conclude that the majority has not given them the careful consideration that they deserve.
I.
The State of California argues, as do Judge Rymer and Judge Schroeder, that the question of the constitutionality of Proposition 140 is res judicata as between the parties to this action, and that we should not reach the merits of the plaintiffs’ claims. It is, of course, well established that federal courts must give full faith and credit to a valid judgment rendered by a state court of competent jurisdiction. See 28 U.S.C. § 1738. More precisely, a federal court must accord such a judgment the same preclusive effect that it would receive in the rendering state. See Matsushita Elec. Indus. Co. v. Epstein, — U.S. -, -, 116 S.Ct. 873, 877, 134 L.Ed.2d 6 (1996). California and Judges Rymer and Schroeder argue that the parties to the present action are the same parties that challenged the constitutionality of Proposition 140 in Legislature v. Eu, 54 Cal.3d 492, 286 Cal.Rptr. 283, 816 P.2d 1309 (1991), or at least that they were in privity with the plaintiffs in that case, and hence that they should be bound by the judgment of the California Supreme Court, and their present claims should be barred. Without deciding whether Bates and the other plaintiffs are actually bound by the Eu decision, the majority concludes (as did the three-judge panel, in an alternative holding on the question) that California would apply its “public interest” exception to res judicata if this case were brought in state court, and so res judicata is no bar to the present action. See supra, majority opinion at 845-846; Jones v. Bates, 127 F.3d 839, 850-51 (9th Cir.1997). I continue also to subscribe to the panel majority’s opinion that these plaintiffs are not barred by res judicata because they were not parties to the prior state court proceeding. See id. at 848-49.
This leads me to respond to Judge Rymer’s disagreement on this point contained in her separate concurrence on this issue. Judge Rymer invokes the Rooker-Feldman doctrine, see District of Columbia Court of Appeals v. Feldman, 460 U.S. 462, 103 S.Ct. 1303, 75 L.Ed.2d 206 (1983); Rooker v. Fidelity Trust Co., 263 U.S. 413, 44 S.Ct. 149, 68 L.Ed. 362 (1923), in order to argue that the district court had no jurisdiction even to hear Bates’ challenge to Proposition 140. In doing so, she apparently argues that, when the' highest court of a state issues a ruling on the constitutionality of a state statute, the RookerFeldman doctrine prevents lower federal courts from ever conducting an independent review of the statute’s constitutionality, regardless of whether the parties before the federal court are different from the parties in the prior state court proceeding. The majority properly declines to embrace this extraordinary proposition, which constitutes a serious misreading of the Rooker-Feldman doctrine.1
*862A.
As Judge Rymer correctly explains, the Rooker-Feldman doctrine holds that federal courts have no jurisdiction to engage in appellate review of the merits of state court judgments. With some exceptions that are not relevant here, the jurisdiction of federal district courts is original, not appellate. See 28 U.S.C. § 1331. Congress has reserved the duty to hear appeals from individual state court judgments, even on issues of federal constitutional law, exclusively to the U.S. Supreme Court in our judicial system. See 28 U.S.C. § 1257; Dubinka v. Judges of the Superior Court, 23 F.3d 218, 221 (9th Cir.1994).
Because it is a doctrine concerning the integrity of individual judgments, Rooker-Feldman, like the doctrine of res judicata, is applicable only when the parties in a second action were also parties, or in privity with parties, in a previous state court proceeding. It is well established that “one is not bound by a judgment in personam in a litigation in which [one] is not designated as a party,” unless it is through a privity relationship or a certified representative. Richards v. Jefferson Cty., Ala., 517 U.S. 793, -, 116 S.Ct. 1761, 1766, 135 L.Ed.2d 76 (1996). Someone who is not a party to an action cannot, of course, request appellate review of that action. See, e.g., Fed. R.App. Proc. 3(c) (“A notice of appeal must specify the party or parties taking the appeal.”) By the same token, someone who was not a party to a state court action cannot be charged with attempting to seek “appellate review” of that action, within the meaning of RookerFeld-man, by bringing a separate action in federal court that happens to address issues that are similar or identical to the issues decided in the state court proceedings. The Rooker-Feldman doctrine prevents federal courts from reviewing state court judgments; it does not prevent federal courts from deciding issues that happen previously to have been decided in unrelated state court proceedings.
- One might well ask what the difference is between the Rooker-Feldman doctrine and traditional principles of res judicata. In many cases, the answer will be that the difference is a purely formal one. As Judge Easterbrook has recently written, “a judgment that is not entitled to full faith and credit does not acquire extra force via the Rooker-Feldman doctrine.” Kamilewicz v. Bank of Boston Corp., 100 F.3d 1348, 1350 (7th Cir.1996) (Easterbrook, J., dissenting from denial of rehearing en banc). This observation holds true both in cases where a party claims that the state proceeding that produced a challenged judgment suffered from a constitutional infirmity (the type of claim that Judge Easterbrook was analyzing in Kamilewicz) and in eases where a party claims that the state court judgment should not bar her action in federal court because it would not bar the same action in a court of the rendering state. Thus, in most cases, the difference between the Rooker-Feldman doctrine and the principle of res judicata amounts to the difference between asking a federal court to review the merits of a binding state court judgment and asking a federal court to address the same issues de novo without reference to the prior judgment. If a party comes to federal court and asks to be relieved of a state court judgment because it was wrongly decided, Rooker-Feldman directs the district court to respond that it has no jurisdiction to engage in such review. If the same party comes into federal court and simply pleads the same claims that were decided in the state court judgment, without making reference to that judgment or asking the district court to review its merits, the Full Faith and Credit Act would require the federal courts to recognize the affirmative defense of res judicata when a defendant interposes the state court judgment as a bar to the plaintiffs claims.2
*863It is difficult to articulate a general rule for identifying the circumstances under which the applicability of Rooker-Feldman and of res judicata are not essentially coextensive. This may be what inspired a leading commentator in the field of Civil Procedure to describe the Rooker-Feld-man doctrine as “somewhat peculiar.” Marcel Kahan & Linda Silberman, Matsu-shita and Beyond: The Role of State Courts in Class Actions Involving Exclusive Federal Claims, 1996 Sup.Ct. Rev. 219, 283 n.122. Perhaps the best explanation of the Rooker-Feldman doctrine is to be found by examining the highly unusual interplay between state and federal courts that obtained in the Feldman case itself.
Feldman arose out of the refusal by the District of Columbia Court of Appeals to admit Feldman to the practice of law in that jurisdiction. The Court of Appeals in the District had a blanket rule against admitting individuals, like Feldman, who had not ah-tended an accredited law school, and it refused to make an exception in Feldman’s case. See Feldman, 460 U.S. at 466-70, 103 S.Ct. at 1305-08. The “judgment” that Feld-man challenged in federal district court was the refusal by the state court3 to grant him an exception and admit him to the D.C. bar. On review, the Supreme Court held that the denial of Feldman’s petition did in fact constitute a “judgment” within the meaning of 28 U.S.C. § 1257, and hence that review of that denial could be had, if at all, only in the U.S. Supreme Court. See id. at 479-82, 486-88, 103 S.Ct. at 1313-14, 1316-17.
What is unusual about Feldman is that the claim of right in that case — the denial to a particular applicant of admission to a state bar — involved a matter of policy that was under the exclusive control of the court. A claimant has occasion to challenge his denial to the bar of a state only after denial of admission by the state court itself. Since the Court found in Feldman that a denial of admission to the bar constitutes a “judicial proceeding,” id. at 479-82, 103 S.Ct. at 1313-14 such denials form a unique class of cases in which every claimant will always already be saddled with an unfavorable state court “judgment” before ever being able to claim that he has been improperly denied admission.
In explaining that federal district courts have no jurisdiction to review decisions in this class of cases, the Court carefully distinguished between “general challenges to state bar admission rules and claims that a state court has unlawfully denied a particular applicant admission.” Id. at 485, 103 S.Ct. at 1316. It is only in the latter class of cases, the Court explained, that an erroneous decision by the state court is, as it were, an element of the petitioner’s claim of right. In contrast, a claimant can bring a general challenge to the rules promulgated by a state without first obtaining a “judgment” under those rules. Therefore, 28 U.S.C. § 1257 poses no barrier to a district court exercising jurisdiction over such a challenge.
Challenges to the constitutionality of state bar rules ... do not necessarily require a United States district court to review a final state-court judgment in a judicial proceeding. Instead, the district court may simply be asked to assess the validity of a rule promulgated in a judicial proceeding. If this is the- case, the district court is not reviewing a state-court judicial decision. In this regard, 28 U.S.C. § 1257 does not act as a bar to the district court’s consideration of the case and because the proceedings giving rise to the rule are nonjudicial the policies prohibiting United States district court review of final state-court judgments are not implicated. United States *864district courts, therefore, have subject-matter jurisdiction over general challenges to state bar rules, promulgated by state courts in nonjudicial proceedings, which do not require review of a final state-court judgment in a particular case. They do not have jurisdiction, however, over challenges to state-court decisions in particular cases arising -out of judicial proceedings even if those challenges allege that the state court’s action was unconstitutional. Review of those decision may be had only in this Court. 28 U.S.C. § 1257.
Id. at 486, 103 S.Ct. at 1316.
Thus, the impetus behind the Court’s decision in Feldman is its identification of a class of cases in which res judicata will always constitute a defense to a party’s attempt to bring a claim in federal district court, because the nature of the claim is such that it never arises except as the result of a judicial proceeding in a state court. As the Feldman Court explains, 28 U.S.C. § 1257 strips the federal district courts of jurisdiction over this unusual class of claims, depriving parties of the opportunity to bring their fruitless, always-already-barred claims in that forum.
B.
Judge Rymer reads Feldman much more broadly. She argues that, under Rooker-Feldman,- the California Supreme Court’s decision in Eu prevents the lower federal courts from ever reviewing the constitutionality of Proposition 140. Since a decision by a federal court on the constitutionality of a state statute would “effectively ... ‘overrule’ ” the merits of a decision by a state court regarding the same statute, Judge Rymer argues that the federal courts are disabled from ever reviewing the statute once a state court has rendered a final decision on the matter. See supra, Rymer concurrence at 857. So long as “the constitutional challenge raised in the federal court was ‘inextricably intertwined’ with the state court’s decision,” id., Judge Rymer would hold that state courts can permanently foreclose review of state statutes in lower federal courts.
, Where Judge Rymer goes seriously wrong is in losing sight of the importance of the identity of parties to the application of the Rooker-Feldman doctrine. As the Third Circuit has explained, “Rooker-Feldman does not bar individual constitutional claims by persons not parties to earlier state court litigation.” Valenti v. Mitchell, 962 F.2d 288, 298 (3d Cir.1992); see also Dubinka, 23 F.3d at 221 (Under Rooker-Feldman analysis, “we must determine whether [a] constitutional claim [is] ‘inextricably intertwined’ with [a] state court’s rulings in a particular plaintiff’s case.”) (emphasis added). “[T]he interests served by Rooker-Feldman are quite similar to those served by giving a state court judgment res judicata effect in a subsequent federal proceeding.” E.B. v. Verniero, 119 F.3d 1077, 1091 (3rd Cir.1997). Thus, the identity of the parties in similar state and federal proceedings is vital to determining what effect, if any, the Rooker-Feldman doctrine will have upon the jurisdiction of a federal district court. Wfiiere A and B bring a challenge to the same state statute in the same federal court proceeding,--and A, but not B, has already brought a similar challenge against the same party in state court, then Rooker-Feldman may deprive lower federal courts of jurisdiction to hear A’s claim, but it will have no impact on their jurisdiction to hear B’s claim. This was precisely the result that obtained in Vemiero, in which the Third Circuit was called upon by two different sets of litigants to review the constitutionality of “Megan’s law,” a compulsory notification law for sex-offenders. That court found that Rooker-Feldman deprived the federal courts of jurisdiction over the claim of one litigant, who had already obtained a judgment on the same claim in state court, but had no impact on the claims of other litigants who had not previously brought suit.
There can be no doubt that a federal court’s pronouncement on the constitutionality of a state statute will have a serious impact on the vitality of previous state-court judgments concerning the same issue. But a litigant cannot invoke the Rooker-Feldman doctrine to prevent a federal court from “reviewing the merits” of a prior state judicial proceeding unless (at a minimum) he and his opponent were both parties to that proceed*865ing, just as a litigant cannot raise a defense of res judicata by interposing an earlier judgment unless (at a minimum) his opponent is personally bound by that judgment. It is true that state and federal courts sometimes reach contrary conclusions as to the constitutionality of state statutes. Such clashes are unfortunate, and, as Judge Rymer rightly points out, it is sometimes appropriate for federal courts to abstain from deciding such issues in order to give states a chance.to interpret their own statutes and thereby possibly to avoid constitutional problems. See supra, Rymer concurrence at 857 n.1; Railroad Comm’n of Texas v. Pullman Co., 312 U.S. 496, 61 S.Ct. 643, 85 L.Ed. 971 (1941). But the Rooker-Feldman doctrine, like the principle of res judicata, is concerned only with preserving the integrity of individual state court judgments, binding upon particular parties, from improper review by lower federal courts. Judge Rymer’s attempt to transform this extremely narrow doctrine into an instrument for restructuring the entire relationship between state and federal courts is both unfortunate and misguided.
II.
Given the importance of the right of the citizenry of a state to decide the form that its government will take (within the constraints imposed by the Federal Constitution in Article IV section 4, the Fourteenth Amendment, and elsewhere), I should not think it a remarkable proposition that, when the citizenry is asked to bring about profound changes in that government through a popular referendum, it must have adequate notice of what those profound changes will be. I do not suggest that such a vote may be invalidated simply upon proof that the voters do not understand, or have not made the effort to understand, the meaning of the initiative measure upon which they cast their votes.4 I do suggest that such a vote must be invalidated when the state does not provide information sufficient to allow any voter to know what it is that he is voting for.
Judge Rymer makes an important point. The initiative in this case deals with an issue of critical importance to the voters of the state: who will govern, citizen legislators or career legislators. While we may question the wisdom of submitting to the initiative process matters that might be better suited to the regular legislative process that is contemplated by traditional notions of republican government, this issue is, quintessentially, one that should be decided by a citizen vote (restricted, of course, by the demands of the Constitution). It is precisely because the issue before us is of such importance that we must ask whether the voters were given adequate notice of the essentials of the measure. It is in this context that the three-judge panel majority paid close attention to the information furnished by the state.
Several aspects of this litigation have tended to obscure the thoughtful attention that the important issues presented here should receive. Timing is the enemy of us all. Unfortunately, the political ramifications of this case have brought about unseemly haste, to the detriment of the deliberative process. In addition, the State of California has fundamentally shifted its position between its appearances before the California Supreme Court and before our court, leaving it in an awkward position, to say the least. Before the Supreme Court of California, the State vigorously argued that the initiative did not impose lifetime bans and urged the Supreme Court of California to ensure the constitutionality of the initiative by construing it to impose restrictions only on consecutive terms. The opponents of the initiative asserted that it was unconstitutional because it imposed a lifetime ban.
Now, before the federal courts, the State of California asserts two propositions: First, that the drafters of the initiative intended that it impose lifetime bans; but second, that *866a certain degree of ambiguity about the matter is not disqualifying and indeed may be injected purposefully to garner broader support from the voters. It is in this context that I respond to the majority’s cursory dismissal of the important issue of the adequacy of the notice to the voters in this case, and to Judge O’Scannlain’s suggestion that .voters need not be given the information necessary to understand the implication of their votes.
A.
The majority holds that the voters of California received adequate notice that, by voting in favor of Proposition 140, they were voting in favor of a lifetime ban. The majority opinion in the decision by the three-judge panel of this court more fully explains my opposition to that position. See Bates, 127 F.3d at 856-63. However, I add the following comments specifically in response to the majority’s and Judge O’Scannlain’s analyses.
The majority opinion is incorrect in its assertion that its holding, that the voters received adequate notice of the effect of Proposition 140, is “consistent with the California Supreme Court.” Despite- Judge O’Scannlain’s implication to the contrary, the California Supreme Court did not rule on whether or not the voters received adequate notice; rather, the court answered only “the interpretive question whether Proposition 140 imposes a ‘lifetime ban’.” Eu, 286 Cal.Rptr. 283, 816 P.2d at 1314. The task of determining whether or not the voters received notice is fundamentally different from the task of interpreting the meaning of a voter initiative; in the latter context, the California Supreme Court acknowledged that it was concerned only with what the “average” voter “likely” believed. Id. at 1316. Not surprisingly, in making that assessment, the California Supreme Court explicitly admitted that “Proposition 140 is ambiguous as to its intent to impose a lifetime ban.” Id. at 1315. It made this statement notwithstanding the State of California’s contrary assertion that Proposition 140 “by its terms ... only limits the number of successive terms which incumbents in the elective offices may serve.” Respondents’ Brief in Legislature v. Eu [“Eu Brief"] at 4. Surely the adequacy of notice cannot turn on what a person of average intelligence “likely” understood regarding an admittedly ambiguous initiative; notice regarding the effect of one’s vote requires clarity, see, e.g., United States v. Bass, 404 U.S. 336, 349, 92 S.Ct. 515, 523, 30 L.Ed.2d 488 (1971); “careful and purposeful consideration,” Greene v. McElroy, 360 U.S. 474, 507, 79 S.Ct. 1400, 1419, 3 L.Ed.2d 1377 (1959); or “considered evaluation,” Hampton v. Mow Sun Wong, 426 U.S. 88, 115, 96 S.Ct. 1895, 1910, 48 L.Ed.2d 495 (1976).
The majority also inappropriately looks to the twenty-second amendment to the United States Constitution for support of its position. The twenty-second amendment states that “[n]o person shall be elected to the office of the President more than twice_” U.S. Const, amend. XXII (emphasis added). By contrast, Proposition 140 stated, “[n]o Senator may serve more than 2 terms,” and “[n]o member of the Assembly may serve more than two terms.” (emphasis added). As the state itself pointed out, this distinction has a specific meaning:
This verbal distinction between restrictions on the incumbent office-holder and restrictions on the individual person who at one time holds the' office may be found in the constitutions of 18 states which impose term limitations upon their Governors. In each constitution, a restriction on the incumbent while he is an incumbent uses the title of the office ... while a restriction on the individual which extends after he leaves the office uses the word “person.” ... [The state] knows of no constitution in which the title of an office is used to designate a person after he no longer holds that office.
Eu Brief at 23-24 n.ll. Thus, the wording of the twenty-second amendment actually supports the opposite position from that advocated by the majority today. At the very least, it creates a serious question as to the intended meaning of Proposition 140, despite the majority’s position that the. meaning was clear to the voter.
The majority’s reliance on the statements found in the materials distributed by those opposing Proposition 140 is also puzzling. *867While the opposition did suggest that the enactment of Proposition 140 would result in a lifetime ban, the state and those in favor .of Proposition 140 never responded to that allegation. The State in fact never stated to the voters that a lifetime ban was the intent or the effect of the measure. As the state acknowledged, “this is not particularly surprising” since it allowed the supporters of Proposition 140 to “benefit from the ambiguity.” Eu Brief at 21. Indeed, the state in oral argument before the en banc court went so far as to suggest that the voters have a right to vote for ambiguous legislation. Given the state’s explicit admission that the state benefits by purposefully refusing to clarify the meaning and effect of an initiative on the ballot, I am amazed that the majority could find that the voters received adequate notice that by voting in favor of Proposition 140, they knew that they were voting to enact a lifetime ban.
Finally, the “surrounding circumstances” did not provide notice to the voters that they were enacting a lifetime ban. The majority relies upon the fact that the voters had two initiatives to choose from on the ballot— Proposition 140, and Proposition 131. Proposition 131, among other things, imposed consecutive term limit bans. The majority concludes from this fact that voters therefore must have known that because Proposition 131 involved consecutive term limit bans, Proposition 140 involved lifetime bans. I respectfully disagree. As the state itself argued, “[i]t simply cannot be said that the voters rejected Proposition 131 solely because the term limitation provisions of that measure” imposed a consecutive rather than a lifetime ban. Eu Brief at 25 n.13. Proposition 131 “covered several, complex issues” and was -different from Proposition 140 .in significant ways concerning funding and revenues. While editorials and the opponents’ materials compared the effect of the two propositions in terms of the severity of the term limitations imposed by each, it cannot be said that such materials constitute “notice,” particularly in the face of not a single reference to a lifetime ban in the text of the amendment, the proponents’ ballot arguments, or the official statement prepared by the state.5
B.
The evidence upon which the majority relies suggests at best that some of those who voted in favor of or against Proposition 140 believed that by doing so, they were voting in favor of or against a lifetime ban.6 It necessarily follows, however, that others who voted in favor of or against the proposition had no notice that they were voting in favor of or against a lifetime ban and so had no notice that they were voting on a severe limitation to their fundamental right to vote for candidates of their choice. Reynolds, 377 U.S. at 555, 84 S.Ct. at 1378. We ask little of the state when we request that it give voters adequate notice of the effect of their vote, particularly when the state asks voters to restrict their own fundamental rights. In contrast, the state loses little when we take away the advantage it gains by fostering an ambiguity as to the effect of a voter initiative.
*868As Judge Rymer’s concurrence appropriately emphasizes, “decisions about how people are to participate in their own representative government and how they choose to represent themselves is ‘the very essence of all self-government.’ ” See supra, Rymer concurrence at 15118 (quoting Laurence H. Tribe, American Constitutional Law 398 (2d ed.1988)) (emphasis in original). Ironically, however, the majority would have the voters of California make those decisions while having to guess at the likely outcome of their votes, having to rely upon newspaper editorials and materials circulated by the opponents of the voter initiatives, rather than upon the plain language of the initiative itself and the explanatory materials produced by the state. It is difficult to determine how the right to vote — particularly the right to vote on issues affecting one’s own fundamental rights — can be meaningful when the state is allowed to force people to vote at their peril rather than add a few simple words to clarify the position for or against which they .are being asked to. cast their ballots.
III.
The majority disposes of the merits of petitioner’s constitutional challenge to lifetime term limits- in an opinion that barely spans one page of the federal reporter. The majority apparently believes that plaintiffs’ challenge presents an easy question. I find the question to be an extremely difficult one.
A.
When a court is called upon to analyze the constitutionality of restrictive qualifications imposed by a state upon candidates for office, it faces a dilemma to which the Supreme Court has not yet offered any answer: How is the court to determine whether a qualification is a “proper” one or not? It seems clear that some qualifications are eminently “proper” and should not require any extraordinary justification by a State in order to survive scrutiny. A requirement that candidates be adults, for example, or that they actually reside in the districts that they seek to represent, fits with our understanding of the most basic of requirements in a representative democracy: That candidates for office be citizens who are willing and able to represent the interests of a particular community. Any constitutional analysis that would require a State to offer a highly particularized explanation for imposing basic age and residency requirements upon candidates for elective office would be extraordinarily invasive.
Even so, the Supreme Court has at least suggested that such an invasive analysis may be called for. On a number of occasions, the Court has emphasized the “fundamental principle of our representative democracy ... ‘that the people should choose whom they please to govern them.’ ” Powell v. McCormack, 395 U.S. 486, 547, 89 S.Ct. 1944, 1977, 23 L.Ed.2d 491 (1969). “The people are the best judges who ought to represent them,” the Court has written, and “[t]o dictate and control them, to tell them whom they shall not elect, is to abridge their natural rights.” U.S. Term Limits v. Thornton, 514 U.S. 779, 794-95, 115 S.Ct. 1842, 1850-51, 131 L.Ed.2d 881 (1995) (quoting 2 Elliot’s Debates 292-93 (statement of Robert Livingston)). Indeed, the Court has gone so far as to say that “[t]he right to vote freely for the candidate of one’s choice is of the essence of a democratic society, and any restrictions on that right strike at the heart of representative government.” Reynolds v. Sims, 377 U.S. 533, 555, 84 S.Ct. 1362, 1378, 12 L.Ed.2d 506 (1964) (emphasis added). Taken literally, these ringing phrases would suggest that every attempt by a state to set restrictive qualifications concerning who can and who cannot serve as a candidate for elective office implicates rights that are “fundamental” and requires careful scrutiny by a reviewing court.
It will not do to respond that the Court has only meant to secure to voters the right to vote for the qualified candidate of their choice, for this merely begs the question, what qualifications may a State impose? By narrowing the field of candidates who are eligible to run for office, lifetime term limits might permanently deprive an entire legislative district of the opportunity to vote for the otherwise-“qualified” candidate of its choice, even if that district had overwhelmingly opposed the imposition of term limits in the first place. Ascertaining the constitutional principles that should govern such a dilemma *869is of the highest importance. The qualifications for federal legislative office are absolutely fixed in the Constitution. See U.S. Const., Art. I § 2 cl. 2-3; Thornton, 514 U.S. at 837-38, 115 S.Ct. at 1871-72. There are no such clear constitutional guideposts. for state legislators. Thus, unless the Court means for lower federal courts to take literally its stirring pronouncements concerning “[t]he right to vote freely for the candidate of one’s choice,” Reynolds, 377 U.S. at 555, 84 S.Ct. at 1378 these repeated, powerful statements do not aid in answering the basic question that is before us: How is the court to determine whether a qualification is a “proper” one or not?
The majority rests its conclusion that states may freely impose lifetime term limits upon state elected officials primarily upon its assertion that “term limits on state officeholders is a neutral candidacy qualification,” one that “[does] not constitute a discriminatory restriction.” Majority opinion at 846-847. The majority looks to the Supreme Court’s cases involving restrictions on candidate access to voting ballots for the proposition that “the State’s important regulatory interests are generally sufficient to justify” such restrictions, provided that they are “reasonable” and “nondiscriminatory.” Burdick v. Takushi, 504 U.S. 428, 434, 112 S.Ct. 2059, 2063, 119 L.Ed.2d 245 (1992). The majority concludes that Proposition 140 does not “limit[ ] political participation by an identifiable political group whose members share a particular viewpoint, assoeiational preference, or economic status,” Majority opinion at 847 (quoting Anderson v. Celebrezze, 460 U.S. 780, 789, 103 S.Ct. 1564, 1570, 75 L.Ed.2d 547 (1983)), and hence that lifetime term limits do not provoke any searching judicial scrutiny.7
I find the majority’s analysis problematic, at best. In my view, U.S. Term Limits v. Thornton forecloses the analytical approach that the majority invokes. As the majority points out, Thornton involved a challenge to an attempt by the State of Arkansas to impose term limits on members of the U.S. Congress. That attempt raised serious concerns under the Qualifications Clauses, U.S. Const., Art. I § 2 cl. 2-3, concerns which are not implicated here, and the Court rested its holding in Thornton primarily upon its analysis of those particular concerns. During the course of its opinion, however, the Court also examined the Arkansas amendment in light of its ballot-access line of eases, finding that there was nothing in those cases that could save the challenged amendment. See Thornton, 514 U.S. at 834-35, 115 S.Ct. at 1869-70. That analysis is entirely at odds with the approach that the majority has adopted in this case.
The State of Arkansas argued in Thornton that its attempt to .impose term limits on members of Congress was in fact nothing more than a permissible exercise of its power, under the Elections Clause of the Federal Constitution, to regulate the “Times, Places and Manner” of federal elections. U.S. Const., Art. 1 § 4 el. 1; see Thornton, 514 U.S. at 832, 115 S.Ct. at 1868. In rejecting that argument, the Court looked to its ballot-access cases and explained that the types of regulatory procedures that those cases involved (many of which constituted permissible regulations under the Elections Clause) were qualitatively different from the imposition of a qualification for candidacy in the form of a lifetime term limitation. As the Court explained, “[t]he provisions at issue in [the ballot-access / Elections Clause cases] *870were ... constitutional because they regulated election procedures and did not even argu-' ably impose any substantive qualification rendering a class of potential candidates ineligible for ballot position.” Thornton, 514 U.S. at 835, 115 S.Ct. at 1870.
Once again, the imposition of qualifications upon state legislators does not implicate the interplay between the Qualifications Clause and the Elections Clause that the Court was primarily concerned with in Thornton. But the Court was quite clear in Thornton that its conclusion that the ballot-access cases could not save the Arkansas amendment relied both upon'the nature and extent of the burden imposed by lifetime limits and upon the unique limitation's embodied in the Qualifications Clause. Thus, the Court explained that its ballot-access cases “provide little support for the contention that a state-imposed ballot access restriction is constitutional when it is undertaken for the twin goals of disadvantaging a particular class of candidates and evading the dictates of the Qualifications Clause.” Id. at 835,115 S.Ct. at 1870 (emphasis added). As to the former consideration, the Court explained that the types of ballot regulations that the Court has upheld in cases like Burdick are qualitatively different from “measures that exclude- candidates from the ballot without reference to the candidates’ support in the electoral process.” Id.
The Court’s analysis of its ballot-access cases in Thornton does not by itself require a finding that lifetime term limits on state legislators are unconstitutional. It does compel the conclusion, however, that such limits “disadvantag[e] a particular class of candidates.” Id. There is no difference between the operation of state and federal term limits in this regard. Both “exclude candidates from the ballot without reference to the candidates’ support in the electoral process.” Id. The clear import of the Court’s analysis in Thornton is that lifetime term limits cannot enjoy the deferential review that the ballot-access cases reserve to those procedural reforms that “impose only reasonable, nondiscriminatory restrictions upon the First and Fourteenth Amendment rights of voters.” Burdick, 504 U.S. at 434, 112 S.Ct. at 2063. Indeed, Thornton could be read for the broad proposition that the ballot-access cases are simply inapplicable to qualification regulations, of which lifetime term limits are an example. See supra, Rymer concurrence at 852. It at least stands for the narrower proposition that, even under the ballot-access cases, lifetime term limits have the effect of “disadvantaging a particular class of candidates,” Thornton, 514 U.S. at 835, 115 S.Ct. at 1870 and hence provoke searching judicial review. See Anderson, 460 U.S. at 793, 103 S.Ct. at 1572. Under either interpretation, Thornton forecloses the deferential approach that the majority has adopted in this case.
B.
Before going on to discuss the constitutionality of Proposition 140 under the stricter standard that Thornton requires us to employ, I think it appropriate to explore further the Supreme Court’s conclusion that lifetime term limits discriminate against “a particular class of candidates.” Thornton, 514 U.S. at 835, 115 S.Ct. at 1870. The majority in Thornton apparently took this to be a self-evident proposition, but, as the district court in the present case aptly observed, “the terms ‘reasonable’ and ‘nondiscriminatory’ [and their converse] are not self-defining.” Bates v. Jones, 958 F.Supp. 1446, 1462 (N.D.Cal.1997). Justice Kennedy took up this question in Thornton at somewhat greater length than did the majority, explaining that lifetime term limits had the dubious effect of burdening the rights of voters8 based on the manner in which they had exercised the franchise in past elections.
Not the least of the incongruities in the position advanced by Arkansas is the proposition, necessary to its case, that it can burden the rights of resident voters in federal elections by reason of the manner in which they earlier had exercised it. If the majority of the voters had been suc*871cessful in selecting a candidate, they would be penalized from exercising that same right in the future.
Thornton, 514 U.S. at 844, 115 S.Ct. at 1874 (Kennedy, J., concurring). Justice Kennedy made this observation during the course of a discussion about the burden Arkansas had placed upon federal voting rights, but he also cited Anderson and suggested that the restriction raised “First Amendment concerns.” Id. Indeed, the concern that Justice Kennedy expressed over a restriction on the right to vote or hold office that looks to “the manner in which [a voter or candidate] earlier had exercised” that right, id., echoes the Thornton majority’s suspicion of “measures that exclude candidates from the ballot without reference to the candidates’ support in the electoral process.” Id. at 885, 115 S.Ct. at 1870.
In its ballot-access eases, the Court has admonished federal courts to “examine in a realistic light the extent and nature of [candidate restrictions’] impact on voters.” Anderson, 460 U.S. at 786, 103 S.Ct. at 1568 (quotation omitted). On the most basic level, as the district court in the present case observed, “Proposition 140 prevents [California voters] from expressing their political preference for candidates with legislative experience.” Bates, 958 F.Supp. at 1463. “California voters who value legislative experience ... are not simply required to defer expressing their preference, they are forever prevented from expressing their preference at the ballot box.” Id. The district court found that this impact alone was enough to warrant strict scrutiny under Anderson and Burdick, a conclusion that was surely warranted in light of the Supreme Court’s conclusion that lifetime term limits “disadvantage] a particular class of candidates.” Thornton, 514 U.S. at 835, 115 S.Ct. at 1870.
Even so, this explanation, without'more, is not wholly satisfactory. The implicit rejoinder of the majority’s analysis is an attractive one, at least on its face. Term limits affect all candidates equally, without regard to the candidates’ parties, their political affiliations, or their resources. Thus, on their face, term limits do not seem to disadvantage any particular class of voters, unless we can say that there are voters who value legislative experience, not merely for the incumbency-based benefits that it can confer upon fortunate districts, but as a political and ideological reason for selecting a representative. I do not doubt that there are some voters who value legislative experience for precisely this reason. It is not self-evident, however, that such voters constitute a “particular class.” There surely are at least some voters who value candidates, on an ideological level, for reasons that would be implicated by the ballot-access restrictions that the Court has upheld in the past. I can easily imagine, for example, that there are voters (albeit few) who might have principled reasons for trusting only a candidate who refuses .to take part in any organized balloting and primary activities, instead insisting that he be elected, if at all, by popular acclaim through the write-in process. Cf. Burdick, 504 U.S. at 435-36, 112 S.Ct. at 2064-65. How, then, is a court to decide, in a close case, when a restriction has the disfavored effect of “disadvantaging a particular class of candidates?” Thornton, 514 U.S. at 835, U5 S.Ct. at 1870.
I am persuaded that the animating force behind the Court’s conclusion in Thornton that term limits have this effect is its observation that term limits “exclude candidates from the ballot without reference to the candidates’ support in the electoral process.” Id. (emphasis added). The regulations that the Court has upheld in its ballot-access cases have been aimed at “seeking to assure that elections are operated equitably and efficiently,” Burdick, 504 U.S. at 433, 112 S.Ct. at 2063 and at “protect[ing] the integrity and reliability of the- electoral process itself,” Anderson, 460 U.S. at 788 n. 9, 103 S.Ct. at 1570 n.9 even though they might, in■ theory, result in some voter’s ideologically preferred candidate being excluded. What those restrictions have in common is that they have asked questions that relate to a candidate’s level of engagement with the electoral process. Has the candidate made use of readily available methods of gaining access to the ballot? See Burdick, 504 U.S. at 435-36, 112 S.Ct. at 2064-65. Has the candidate demonstrated a threshold level of popular support? See Munro v. Socialist Workers Party, 479 U.S. 189, 192-93, 107 S.Ct. 533, 535-36, 93 *872L.Ed.2d 499 (1986). Has the candidate identified himself with a particular party for purposes of a given election? See Storer v. Brown, 415 U.S. 724, 726-28, 94 S.Ct. 1274, 1277-78, 39 L.Ed.2d 714. It is not enough, of course, for a candidate restriction to fit this general description. If the restriction burdens the rights of voters in too disparate or severe a fashion, it still may not pass muster. See, e.g., Anderson, 460 U.S. at 806, 103 S.Ct. at 1579. But this is the type of question that the ballot-access cases suggest that states ought to be asking in passing candidate restrictions. The candidate qualifications that I discussed at the beginning of this section — age and residency — would likewise merit deference under this approach. Clearly, residency requirements are directly concerned with the level of a candidate’s engagement with a particular community. Similarly, people below a certain age do not enjoy all the rights and responsibilities of citizenship, and reasonable age requirements help to ensure that a candidate- can fully, appreciate, and effectively represent, the most basic interests of his constituency.
What the Court suggested in - Thornton, however, was that a “measure[ ] that exclude[s] candidates from the ballot without reference to the candidates’ support in the electoral process,” Thornton, 514 U.S. at 835, 115 S.Ct. at 1870 is asking the wrong type of question, at least as far as the ballot-access cases are concerned. A restriction that imposes a “substantive qualification rendering a class of potential candidates ineligible for ballot position,” id., always carries with it the danger that particular classes of voters will be disadvantaged, perhaps in ways that will be difficult to foresee. The ballot-access cases, Thornton suggests, represent a compromise under which states are permitted to pass restrictions aimed at protecting the electoral process, but only so long as those restrictions limit themselves to asking about a candidate’s level of engagement and support in that process itself. See id. If the burdens that the restriction imposes are not too severe, then the restriction will enjoy a presumption of validity. If the restriction disadvantages candidates for reasons having nothing to do with a lack of engagement and support in the electoral process, however, then that presumption of validity disappears. At that point, the “fundamental principle ... ‘that the people should choose whom they please to govern them,’ ” Powell v. McCormack, 895 U.S. 486, 547, 89 S.Ct. 1944, 1977, 23 L.Ed.2d 491 (1969), reasserts itself and requires courts to look closely at the challenged restriction. Such is the case with lifetime term limits, which prevent candidates from running for office precisely because they are engaged with the political process and have found considerable support in it. See Thornton, 514 U.S. at 844, 115 S.Ct. at 1874 (Kennedy, J., concurring). We cannot accord such a regulation a casual presumption of validity. Rather, we must weigh it carefully against the interests that the State asserts in its defense.
C.
I recognize the uniquely compelling interest that states enjoy in experimenting with new forms of republican government under our federal system. Judge Rymer provides an account of that interest in her opinion that is powerful. See supra, Rymer concurrence at 851-852. As Justice Brandéis has famously written, “[i]t is one of the happy incidents of the federal system that a single courageous state may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country.” New State Ice Co. v. Liebmann, 285 U.S. 262, 311, 52 S.Ct. 371, 386, 76 L.Ed. 747 (1932) (Brandéis, J., dissenting). As Judge Rymer explains, this principle is applicable to political reform as well, for, “[tjhrough the structure of its government, and the character of those who exercise government authority, a State defines itself as a sovereign.” Gregory v. Ashcroft, 501 U.S. 452, 460, 111 S.Ct. 2395, 2400, 115 L.Ed.2d 410 (1991); see supra, Rymer concurrence at 852.
In this regard, a limitation on the number of terms that a candidate may serve in state government implicates federalism interests that pull in the opposite direction from the federalism interests implicated by a similar restriction on elected federal officials. As the Court explained in Thornton, it was of central concern to the Framers of our Con*873stitution that the States enjoy a relationship to the federal government that would not vary from state to state in order that they might be unified into a single, stable Republic. See Thornton, 514 U.S. at 806-07, 115 S.Ct. at 1856-57. “Permitting individual States to formulate diverse qualifications for their representatives would result in a patchwork of state qualifications, undermining the uniformity and the national character that the Framers envisioned and sought to ensure.” Id. at 822, 115 S.Ct. at 1864. Rather, it is the bedrock of stability that our uniform federal government offers that in turn makes it possible for States to experiment safely with novel approaches to intractable social, economic and political problems. The principle of federalism thus weighs heavily in favor of allowing California to experiment with a government of. “citizen legislators” like the one that Proposition 140 envisions.
It is necessary, however, to articulate some limitations on the ability of States to alter their structure of government on the strength of these federalism interests. Unfortunately, the ordinary paradigm of strict scrutiny analysis is not helpful in this regard. Once it is acknowledged that the States have, a compelling interest in experimenting with new forms of government, the concept of “narrow tailoring” has limited application, for any novel governmental structure is, by definition, narrowly tailored to the goal of. experimenting with that particular governmental structure. But the States do not have a completely free hand in choosing how to organize themselves. That is why the question before us has no easy solution.
It is clear, at the very least, that a State cannot discriminate in an invidious fashion-against its citizens, even in the name of sovereignty and federalism. See Ashcroft, 501 U.S. at 460, 111 S.Ct. at 2400. Disadvantaging candidates on the basis of their political ideology would unquestionably constitute such “invidious discrimination.” I have said that I am not convinced that favoring candidates on the basis of legislative experience constitutes a sufficiently distinct “political ideology” to warrant the invalidation of term limits as an invidious candidate qualification. I consider the question a close one, however. Several members of our court for various reasons would not reach the issue in this ease, nor would I, in light of my view that the process that produced Proposition 140 was infirm. It is a question that, ultimately, the Supreme Court must address, and its decision will be a weighty one. Where along the fulcrum the balance lies is not easy of resolution.
I take the occasion here, however, to express my further discomfort with any ruling that would permanently foreclose a court from concluding, after seeing term limits in action, that they were wrong: that a preference for legislative experience does, in fact, constitute a genuine and distinct voter preference or political ideology, and that lifetime term limits have deprived an identifiable group of voters of the opportunity to elect representatives who will represent their interests and promote their political ideals.
“When “there is a serious factual controversy over the feasibility of recognizing [a] claimed right without at the same time making it impossible for the State to engage in an undoubtedly legitimate exercise of power,” Washington v. Glucksberg, — U.S. -, -, 117 S.Ct. 2258, 2292, 138 L.Ed.2d 772 (1997) (Souter, J., concurring), it may sometimes be appropriate for federal courts to deny a claim of right while remaining open to'reaching a different conclusion if a demonstrable infringement on individual liberties can later be shown. This principle holds especially true when the organization of our federal system is at issue, since one of the very purposes of that system is to foster investigation and experimentation by state legislatures. See id. at -, 117 S.Ct. at 2293 (Souter, J., concurring); Glucksberg, — U.S. at -, 117 S.Ct. at 2303 (O’Connor, J., concurring). At a minimum, I suggest that the majority should recognize that what it does today is to validate an experiment that may not prove out. If experience demonstrates that the predicate that underlies the majority’s decision is incorrect, and that lifetime term limits consistently disadvantage a particular class of voters on the basis of an articulable political ideology, then our court should remain open to revisiting its holding.
*874* * *
I voice my strong dissent from the majority’s view that the voters of California received adequate notice from the State regarding the true import of their vote on Proposition 140. From reading their opinion, one might conclude also that the majority did not appreciate either the gravity or the closeness of the issues upon which they were called to deliberate.
. Throughout the course of this litigation, California has vigorously advanced the position that "state courts are not bound' by lower federal appellate court decisions on federal questions.” Petition of the State of California for a Writ of Certiorari, filed Sept. 11, 1997, at 17. Among many other cases, the State has cited Rooker in support of this position. See id. at 869. It is not clear to me whether Judge Rymer, in arguing that states can foreclose review of federal questions by federal courts, means to incorporate this position into her argument. Since she does not do so explicitly, I will not respond to that position at length. I do note, however, that the practical effect of the argument that Judge Rymer advances in her concurrence is that lower federal courts would be bound by state court decisions on federal questions. If we assume, for the sake of argument, that California is correct in its assertion that "state courts are not bound by lower federal appellate court decisions on federal *862questions,” id. at 867, then surely the State could not be heard to argue that the converse would not also be true. What is good for the goose is good for the gander.
. The difference between these two situations, while largely formal, is significant. If a district court were to engage in "appellate review” of a state court judgment, the result would be that the state court judgment would be vacated. If a district court were to disregard a state court judgment for res judicata purposes, however, the judgment would remain in existence at the close of the federal proceedings. This distinction *863maps onto the difference in the control that parties exercise over each doctrine: res judicata is a waivable defense, while infirmities to a court’s subject-matter jurisdiction (including Rooker-Feldman review) are not. Compare Clements v. Airport Auth. of Washoe Cty., 69 F.3d 321, 328 (9th Cir.1995) (res judicata), with Owen Equip. & Erection Co. v. Kroger, 437 U.S. 365, 377 n. 21, 98 S.Ct. 2396, 2404 n. 21, 57 L.Ed.2d 274 (1978) (subject-matter jurisdiction).
. For purposes, of appellate review under 28 U.S.C. § 1257, the local courts of the District of Columbia are treated as state courts. See District of Columbia Court Reform and Criminal Procedure Act of 1970, Pub.L. No. 91-358, § 172(a)(1), 84 Stat. 590; see also Feldman, 460 U.S. at 463-64, 103 S.Ct. at 1305.
. Judge O’Scannlain purposefully misinterprets my argument. He asserts that I impose a test that would require federal courts to examine whether voters “were capable of understanding what they voted for.” Supra, O’Scannlain concurrence at 853. He facetiously notes that he finds no "ignorant voter clause” in the Constitution. Of course, I do not either. Adequate notice is a requirement to assure that the voter has information from which the voter can make an informed judgment. Even the most intelligent and diligent voter must be informed of the intent and effects of that upon which he casts his ballot.
. The majority's reliance on Taxpayers to Limit Campaign Spending, 274 Cal.Rptr. 787, 799 P.2d at 1220, is misplaced. In that case, the California Supreme Court considered what to do when two voter initiatives, both dealing with campaign finance, received more than 50 percent affirmative votes. The court held that no part of the ballot initiative receiving the fewer affirmative votes could take effect. 274 Cal.Rptr. 787, 799 P.2d at 1221. While that ruling apparently was interpreted by the press as potentially affecting what would happen if both Proposition 131 and Proposition 140 were passed, that does not constitute notice that Proposition 140 involved a lifetime ban.
. I also note that some of the evidence relied upon by the en banc majority was not relied upon by the California Supreme Court, despite the majority’s statement that its decision is "consistent” with that of the California Supreme Court. The latter looked only to the Legislative Analyst's analysis and the opponents’ materials in interpreting the meaning of Proposition 140 and concluding that it imposed a lifetime ban. See Eu, 286 Cal.Rptr. 283, 816 P.2d at 1315. The California Supreme Court specifically rejected the notion that the wording of the amendment made it clear what type of ban was contemplated by Proposition 140. See id. Nor did the Eu court look to the effect of having Proposition 131 on the same ballot, or to the attention Proposition 140 received from the media.
. Judge O’Scannlain vigorously urges the position that the Supreme Court's summary dismissal in Moore v. McCartney is such "compelling authority” for the issues before us that there is not even a federal question for us to decide and we should dismiss for want of jurisdiction. See supra, O'Scannlain concurrence at 15096-102. Judge O’Scannlain’s own presentation of his argument provides a sufficient basis for disregarding it, however. Moore was a case about consecutive term limits, not a lifetime ban. The petitioner in that case — the Gqver-nor — made no claims under the First Amendment; indeed, there were no voters joined as parties in the case at all, see Maloney v. McCartney, 159 W.Va. 513, 223 S.E.2d 607, 610-11 (1976), raising a serious question as to whether the case presented a federal voting rights issue over which the Court could exercise jurisdiction. The case was a summary dismissal and has been followed by twenty years of First Amendment opinions in the related field of candidate restrictions. Judge O’Scannlain's only authority for the proposition that Moore should even inform the present inquiry is Justice Thomas’ dissenting opinion in Thornton.
. As the Court has explained, "the rights of voters and the rights of candidates do not lend themselves to neat separation; laws that affect candidates always have at least some theoretical, correlative effect on voters.” Bullock v. Carter, 405 U.S. 134, 143, 92 S.Ct. 849, 855, 31 L.Ed.2d 92 (1972).