NOTE: Where it is feasible, a syllabus (headnote) will be released, as is being done in connection with this case, at the time the opinion is issued.
The syllabus constitutes no part of the opinion of the Court but has been prepared by the Reporter of Decisions for the convenience of the reader.
See United States v. Detroit Timber & Lumber Co., 200 U. S. 321, 337.

SUPREME COURT OF THE UNITED STATES

Syllabus

T. M. v. UNIVERSITY OF MARYLAND MEDICAL SYSTEM CORPORATION et al.

certiorari to the united states court of appeals for the fourth circuit

No. 25–197. Argued April 20, 2026—Decided June 18, 2026

The Rooker-Feldman doctrine bars federal district courts from exercising jurisdiction over cases brought by state-court losers seeking review and rejection of state-court judgments rendered before district court proceedings commenced. See Exxon Mobil Corp. v. Saudi Basic Industries Corp., 544 U. S. 280, 284. This case asks whether the doctrine applies only to final judgments rendered by the highest court of a State in which a decision could be had, or whether it also bars suit when the state-court judgment at issue remains subject to further review in state appellate proceedings.

Petitioner T. M. alleges that she has a medical condition that causes changes in her mental status, including psychosis, when she ingests gluten. In March 2023, T. M. accidentally ingested gluten and was taken to the emergency room at Baltimore Washington Medical Center, where, after an administrative hearing and over her and her father’s objections, she was involuntarily committed for about three months. During her stay, T. M.’s treating psychiatrist and the medical center obtained an order authorizing the facility to forcibly inject T. M. with antipsychotic medication. These events prompted a flurry of litigation, with T. M. and her parents filing several state and federal lawsuits seeking T. M.’s release and to avoid the forced injections. T. M. and respondents negotiated a settlement agreement to facilitate T. M.’s discharge, and the state judge presiding over T. M.’s state habeas petition entered the agreement as a consent order on June 12, 2023. The consent order provided for T. M.’s immediate release subject to several conditions, including that T. M. obtain a new treating psychiatrist, continue taking her prescribed medications, and dismiss with prejudice all pending actions against respondents. Ten days after the state court entered the consent order, T. M. and her parents obtained new counsel and sued respondents in Federal District Court for the District of Maryland, seeking a declaration that the consent order violated T. M.’s federal and state due-process rights, a declaration that the order was obtained under duress, and an injunction preventing its enforcement. Meanwhile, T. M. appealed the consent order to the Appellate Court of Maryland and raised similar arguments. T. M. later successfully moved to stay the state-court appeal to prevent inconsistent rulings. The District Court dismissed the complaint sua sponte for lack of subject matter jurisdiction under the Rooker-Feldman doctrine. The Fourth Circuit affirmed, rejecting T. M.’s argument that Rooker-Feldman applies only to judgments that are final judgments from the highest court of a State in which the decision could be had.

Held: The Rooker-Feldman doctrine bars federal district court jurisdiction over cases brought by state-court losers complaining of injuries caused by state-court judgments rendered before the district court proceedings commenced and seeking district court review and rejection of those judgments, regardless of whether the state-court judgment remains subject to further review in state appellate proceedings. Pp. 5–18.

(a) Federal district courts generally lack “any power to review directly cases from state courts.” Atlantic Coast Line R. Co. v. Locomotive Engineers, 398 U. S. 281, 286. The Supreme Court is the only federal court with appellate jurisdiction to review state-court judgments, and that jurisdiction extends only to “[f]inal judgments or decrees rendered by the highest court of a State in which a decision could be had,” 28 U. S. C. §1257(a). In Rooker v. Fidelity Trust Co., 263 U. S. 413, the Court held that a Federal District Court lacked jurisdiction to declare a state-court judgment “null and void” because doing so would be “an exercise of appellate jurisdiction,” which only the Supreme Court possesses, and the “jurisdiction possessed by the District Courts is strictly original.” Id., at 414–416. In District of Columbia Court of Appeals v. Feldman, 460 U. S. 462, the Court held that a Federal District Court lacked jurisdiction to “review” a final judicial determination of the D. C. high court because such “[r]eview . . . can be obtained only in [the Supreme] Court” under §1257. Id., at 476, 482.

Rooker and Feldman rest on two closely related bases of reasoning. First, when plaintiffs “essentially invit[e] federal courts of first instance to review and reverse unfavorable state-court judgments,” they are seeking an exercise of appellate jurisdiction. Exxon, 544 U. S., at 283–284. Second, such “appellate jurisdiction to reverse or modify a state-court judgment is lodged, . . . by 28 U. S. C. §1257, exclusively in [the Supreme] Court.” 544 U. S., at 283. This Court has since refused to expand the Rooker-Feldman doctrine but has also reaffirmed its rule in the narrow ground it occupies. Exxon, 544 U. S., at 284. Pp. 5–8.

(b) T. M.’s case falls within “the narrow ground occupied by Rooker-Feldman,” Exxon, 544 U. S., at 284, because she is complaining of injuries caused by, and seeking relief from, the state-court judgment itself, arguing that the consent order violates her federal and state due process rights and was entered into under duress. The consent order was rendered 10 days before T. M. commenced this federal lawsuit, and she plainly seeks “review and rejection” of the “ ‘Consent Order’ ” by asking the District Court to declare it “unconstitutional, unenforceable, and void ab initio” and to enjoin its “enforcement.” 139 F. 4th 344, 347. P. 8.

(c) The Court rejects T. M.’s argument that Rooker-Feldman should apply only to final judgments rendered by the highest court of a State in which a decision could be had, i.e., the kinds of judgments that strictly fall within this Court’s jurisdiction under 28 U. S. C. §1257. Pp. 9–18.

(1) T. M.’s theory cannot be squared with the Court’s precedents. T. M. reads the Court’s prior Rooker-Feldman cases not to evince any concern about district courts exercising what amounts to appellate jurisdiction over state-court judgments, but instead as relying on a strict negative inference from §1257: when this Court has jurisdiction over a state-court judgment, district courts must not. But this Court’s precedents plainly adopted a more functional view of what constitutes original and appellate jurisdiction under §1331. In Rooker, nothing in the Court’s reasoning suggested the outcome would have been different had the judgment still been on appeal; the Court held that seeking to “reverse or modify the [state-court] judgment” would be an “exercise of appellate jurisdiction,” which the District Court lacks because it has “strictly original” jurisdiction. 263 U. S., at 416. Later cases confirmed that Rooker-Feldman bars suits in federal district court that “see[k] what in substance would be appellate review of [a] state judgment,” regardless of whether the judgment formally falls within this Court’s §1257 jurisdiction. Johnson v. De Grandy, 512 U. S. 997, 1005–1006. In fact, Feldman explicitly rejected the notion that district courts have jurisdiction to review state-court judgments so long as the judgments are not yet within this Court’s §1257 jurisdiction. 460 U. S., at 483–484, n. 16 (explaining that irrespective of the Court’s “jurisdiction to review a final state court judgment . . .  ‘lower federal courts possess no power whatever to sit in direct review of state court decisions’ ”). Pp. 9–13.

(2) If departure from this Court’s precedents alone were not enough to affirm the dismissal of T. M.’s suit under Rooker-Feldman, adopting T. M.’s rule would also create anomalous outcomes and undermine federalism principles. Allowing federal district courts to review state-court judgments while they are on appeal in the state-court system would undermine the “[c]ooperation and comity” on which the Nation’s federal system is built. Ruhrgas AG v. Marathon Oil Co., 526 U. S. 574, 586. T. M.’s rule would also produce arbitrarily different results depending on when a federal suit seeking review of a state-court judgment is filed, encouraging parties to file earlier in federal court while the state appellate proceedings are pending, and to duplicate their efforts even though the state process may resolve the dispute in their favor.

T. M.’s contention that abstention and preclusion doctrines can do nearly all of the work that Rooker-Feldman does is unavailing. It is unclear if such doctrines even apply in cases where a plaintiff complains of injuries that stem directly from a state-court judgment rather than attempting to relitigate the same claims. Federalism principles are thus best served by continuing to apply Rooker-Feldman to federal plaintiffs seeking review of state-court judgments, regardless of whether those judgments are final trial-court judgments or those of a State’s highest court. Pp. 13–16.

(3) T. M. contends that the Court should adopt her position because it will generally cabin Rooker-Feldman, a doctrine she claims has caused confusion and is overused as a docket-clearing mechanism in the lower courts.  T. M.’s proposed rule would fail to address the source of confusion in current Rooker-Feldman doctrine—determining when a plaintiff is seeking federal review and rejection of a state-court judgment—and would add a new source of confusion by requiring courts to determine whether a state-court judgment is a reviewable final judgment under §1257(a). In any event, the animating force behind T. M.’s arguments is the belief that Rooker and Feldman were wrongly decided, but this issue is not fairly included in the question presented. The Court today neither expands nor constrains Rooker-Feldman but leaves the doctrine as it found it. Pp. 16–18.

139 F. 4th 344, affirmed.

Sotomayor, J., delivered the opinion of the Court, in which Thomas, Alito, Kavanaugh, and Jackson, JJ., joined. Thomas, J., filed a concurring opinion. Barrett, J., filed a dissenting opinion, in which Roberts, C. J., and Kagan and Gorsuch, JJ., joined.

NOTICE: This opinion is subject to formal revision before publication in the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, pio@supremecourt.gov, of any typographical or other formal errors.

SUPREME COURT OF THE UNITED STATES

_________________

No. 25–197

_________________

T. M., PETITIONER v. UNIVERSITY OF MARYLAND MEDICAL SYSTEM CORPORATION, et al.

on writ of certiorari to the united states court of appeals for the fourth circuit

[June 18, 2026]

Justice Sotomayor delivered the opinion of the Court.

Under what has become known as the Rooker-Feldman doctrine, federal district courts lack jurisdiction over “cases brought by state-court losers complaining of injuries caused by state-court judgments rendered before the district court proceedings commenced and inviting district court review and rejection of those judgments.” Exxon Mobil Corp. v. Saudi Basic Industries Corp., 544 U. S. 280, 284 (2005). This case asks whether this rule bars suit when the state-court judgment at issue is subject to further review in state appellate proceedings. A straightforward application of the logic and reasoning underlying Rooker-Feldman leads to one conclusion: It does. Because this suit falls within the narrow doctrine’s limits, the Court of Appeals for the Fourth Circuit properly affirmed its dismissal.

I

A

Petitioner T. M. alleges that she has a medical condition that “ ‘causes changes in [her] mental status upon ingesting any amount of gluten’ and can result in ‘episodes of psychosis.’ ” 139 F. 4th 344, 347 (CA4 2025) (alteration in original). In March 2023, T. M. accidentally ingested gluten and was taken to the emergency room at Baltimore Washington Medical Center. After an administrative hearing before a Maryland administrative law judge (ALJ), and over her and her father’s objection, she was involuntarily committed to the facility, where she stayed for about three months. During her stay, T. M.’s treating psychiatrist and the medical center sought, and were granted by a clinic review panel, an order authorizing the facility forcibly to inject T. M. with antipsychotic medication. That order was later affirmed by an ALJ.

These events jumpstarted a flurry of litigation, with T. M. and her parents filing several state and federal lawsuits seeking to have T. M. released from involuntary commitment and to avoid the forced injections. In state court, T. M. filed two petitions for judicial review of the ALJ decisions allowing her involuntary admission and upholding the forced-injection order. She also filed a state habeas petition and a separate civil action, with several emergency motions, seeking release from the medical center. In federal court, T. M. filed a due-process suit seeking damages and an injunction requiring her discharge. T. M.’s father, for his part, filed a state-court action against respondents seeking an order requiring them to recognize T. M.’s purported advance medical directive listing him as T. M.’s healthcare agent if she became incompetent to make decisions.

In the midst of those lawsuits, T. M. and respondents negotiated a settlement agreement to facilitate T. M.’s discharge, and the state judge presiding over the state habeas petition entered the agreement as a consent order. The consent order provided for T. M.’s immediate release subject to several conditions, including that T. M. would “(1) obtain a new treating psychiatrist and continue to take her hospital-prescribed medications; (2) regularly meet and consult with a third-party provider regarding her treatment and medication; (3) accept a referral to [an outpatient clinic] and follow their recommendations; (4) take all prescribed medications; and (5) dismiss with prejudice all of her [and her parents’] pending actions against [respondents].” App. to Pet. for Cert. 24a. The consent order also required T. M.’s parents to monitor her use of her medications and report if she stopped taking them. The state court, counsel for T. M., and respondents signed the consent order on June 12, 2023, and respondents discharged T. M.

B

Unfortunately, the consent order did not end the parties’ dispute in either state or federal court. Ten days after the state court entered the consent order, T. M. and her parents obtained new counsel and sued respondents in the District Court for the District of Maryland. This lawsuit is now before the Court. The complaint sought a declaration that the consent order violated T. M.’s federal and state due-process rights, a declaration that the order was obtained under duress, and an injunction preventing the order’s enforcement.

Meanwhile, back in state court, T. M. appealed the consent order to the Appellate Court of Maryland and raised similar arguments. About four months later, T. M. moved to stay the pending appeal in the Appellate Court of Maryland, contending that a stay was warranted to “prevent inconsistent rulings” because both the state appeal and the pending federal lawsuit were “based on the same facts and events.” Motion to Stay in No. 1:23–cv–1684, ECF Doc. 81–1, p. 1. The Appellate Court of Maryland granted the stay, which remains in effect.

In the federal case, the Maryland District Court dismissed the complaint sua sponte for lack of subject matter jurisdiction, holding that T. M.’s claims were barred by the Rooker-Feldman doctrine. That doctrine, addressed in detail below, generally holds that federal courts lack jurisdiction over lawsuits brought by parties who received adverse judgments in state courts and then turn to federal courts seeking “review and rejection of ” those judgments. Exxon, 544 U. S., at 284.

The Fourth Circuit affirmed. The panel observed that in Rooker v. Fidelity Trust Co., 263 U. S. 413 (1923), this Court held that a Federal District Court lacked jurisdiction to “ ‘declar[e]’ that [a] state court’s judgment was ‘null and void’ ” under the Federal Constitution because doing so “ ‘would be an exercise of appellate jurisdiction,’ ” which only the Supreme Court possesses. 139 F. 4th, at 348 (quoting 263 U. S., at 414–415, 416). Because T. M. was asking the District Court here to do precisely that, the Fourth Circuit reasoned that this “case is too much like Rooker to justify a different” conclusion and thus required dismissal based on Rooker-Feldman. 139 F. 4th, at 346. In so holding, the Fourth Circuit also rejected T. M.’s argument that Rooker-Feldman applies only to judgments that are “ ‘final judgment[s] from the highest court of a State in which the decision could be had,’ ” reasoning that this Court has never imposed such a requirement. 139 F. 4th, at 353–354.

T. M. filed a timely petition for a writ of certiorari seeking review of the Fourth Circuit’s conclusion that the Rooker-Feldman doctrine can be triggered by a state-court judgment that remains subject to further review in state court. The Sixth Circuit has taken the same position as the Fourth Circuit. See RLR Invs., LLC v. Pigeon Forge, 4 F. 4th 380, 389–390, 396 (2021). Other Circuits have held that Rooker-Feldman applies only if the state-court proceedings have “ended.” See, e.g., Federación de Maestros de P. R. v. Junta de Relaciones del Trabajo de P. R., 410 F. 3d 17, 24–25 (CA1 2005) (outlining three-part test to determine whether state proceedings have functionally “ended”). The Court granted certiorari to resolve the conflict among the Courts of Appeals. 607 U. S. 1079 (2025).

II

A

“Federal courts are courts of limited jurisdiction and generally can resolve only the cases that Congress grants them power to hear.” Hain Celestial Group, Inc. v. Palmquist, 607 U. S. 421, 424 (2026). As relevant here, Congress in 28 U. S. C. §1331 granted federal district courts “original jurisdiction of all civil actions” raising federal questions. District courts generally lack “any power to review directly cases from state courts.” Atlantic Coast Line R. Co. v. Locomotive Engineers, 398 U. S. 281, 286 (1970). Instead, this Court is the only federal court with appellate jurisdiction to review state-court judgments, and that jurisdiction extends only to “[f]inal judgments or decrees rendered by the highest court of a State in which a decision could be had.” §1257(a). Thus, “from the beginning we have had in this country two essentially separate legal systems” with “[e]ach system proceed[ing] independently of the other with ultimate review in this Court of the federal questions raised in either system.” Atlantic Coast Line, 398 U. S., at 286.

The Court delineated some of the bounds that separate these two systems in Rooker and Feldman. First, in Rooker, the Court held that a Federal District Court lacked jurisdiction to resolve a case that sought “to have a judgment of a circuit court in Indiana, which was affirmed by the Supreme Court of the State, declared null and void” because it violated the Constitution’s due process, contracts, and equal protection clauses. 263 U. S., at 414. The Court explained that “entertain[ing] a proceeding to reverse or modify [a] judgment for errors of that character” would be “an exercise of appellate jurisdiction.” Id., at 415–416. As a result, the suit was “plainly not within the District Court’s jurisdiction” because the “jurisdiction possessed by the District Courts is strictly original.” Ibid. The Court also laid out the path to challenging a state-court judgment: “If the constitutional questions . . . actually arose” in the suit, “it was the province and duty of the state courts to decide them” in the first instance. Id., at 415. “If the decision was wrong, that did not make the judgment void, but merely left it open to reversal or modification in an appropriate and timely appellate proceeding,” including through appellate review in this Court. Ibid.

Second, in District of Columbia Court of Appeals v. Feldman, 460 U. S. 462 (1983), the Court held that a Federal District Court lacked jurisdiction to “review” a final judicial determination of the D. C. high court because such “[r]eview . . . can be obtained only in this Court” under §1257. Id., at 476, 482. The plaintiffs there had challenged as unconstitutional a court rule requiring D. C. bar applicants to have graduated from an approved law school. They had also sought a waiver of that rule. This Court held that the Federal District Court had no jurisdiction to review the D. C. high court’s denial of the plaintiffs’ waiver petitions, but possessed jurisdiction to adjudicate the validity of the rule itself because, in the latter situation, the “district court is not reviewing a state-court judicial decision.” Id., at 486.

Rooker and Feldman are thus built on two closely related bases of reasoning. The first is that, when plaintiffs “essentially invit[e] federal courts of first instance to review and reverse unfavorable state-court judgments,” they are seeking an exercise of appellate jurisdiction. See Exxon, 544 U. S., at 283–284 (describing Rooker and Feldman). The second is that such “appellate jurisdiction to reverse or modify a state-court judgment is lodged, . . . by 28 U. S. C. §1257, exclusively in this Court.” 544 U. S., at 283. Federal district courts, by contrast, are “empowered to exercise [only] original, not appellate, jurisdiction.” Ibid.

In the years since Rooker and Feldman, this Court has refused to expand the Rooker-Feldman doctrine. In the “narrow ground occupied by Rooker-Feldman,” however, the Court has repeatedly reaffirmed its rule. Exxon, 544 U. S., at 284. In Exxon, the Court held that the “Rooker-Feldman doctrine . . . is confined to cases of the kind from which the doctrine acquired its name: cases brought by state-court losers complaining of injuries caused by state-court judgments rendered before the district court proceedings commenced and inviting district court review and rejection of those judgments.” 544 U. S., at 284. In those cases, Rooker-Feldman continues to require “dismiss[al] for want of subject-matter jurisdiction.” 544 U. S., at 284; see also Lance v. Dennis, 546 U. S. 459, 466 (2006) (per curiam) (“The doctrine applies only in ‘limited circumstances’ . . . where a party in effect seeks to take an appeal of an unfavorable state-court decision to a lower federal court”); Reed v. Goertz, 598 U. S. 230, 235 (2023) (“Th[e] doctrine prohibits federal courts from adjudicating cases brought by state-court losing parties challenging state-court judgments”).

B

Under these precedents, T. M.’s case falls within “the narrow ground occupied by Rooker-Feldman.” Exxon, 544 U. S., at 284. T. M. is complaining of injuries caused by, and is seeking relief from, the state-court judgment itself, arguing that the consent order violates her federal and state due process rights and was entered into under duress. See App. 40–45 (complaint alleging injuries stemming from the consent order). The consent order was rendered 10 days before T. M. commenced this federal lawsuit. T. M. plainly seeks “review and rejection” of the consent order by asking the District Court to “[d]eclare that the ‘Consent Order’ ” is “unconstitutional, unenforceable, and void ab initio” and to enjoin “enforcement of the ‘Consent Order.’ ” 139 F. 4th, at 347. Thus, T. M.’s case is the “ ‘paradigm situation in which Rooker-Feldman precludes a federal district court from proceeding’ ”: She does not like the result reached in state court and “repaired to federal court to undo the [state] judgment in [her] favor.” Exxon, 544 U. S., at 293.

III

Instead of disputing any of the points discussed above, T. M. asks this Court to add a new requirement to the Rooker-Feldman doctrine. In T. M.’s view, endorsed by the dissent, the Court should adopt a new rule specifying that Rooker-Feldman bars only federal suits seeking review and rejection of “[f]inal judgments” that are “rendered by the highest court of a State in which a decision could be had,” 28 U. S. C. §1257(a). See Brief for Petitioner 19; post, at 6 (opinion of Barrett, J.). Because that rule is inconsistent with the Court’s precedents, risks producing anomalous results, and undermines federalism interests, the Court declines to adopt it.

A

T. M.’s theory, echoed by the dissent, fails because it requires a reimagining, rather than an application, of the Court’s Rooker-Feldman precedents. Taking T. M.’s argument step-by-step illustrates the point.

In T. M.’s view, Rooker-Feldman has never rested on any distinction between what, in substance, qualifies as original and appellate jurisdiction. Instead, on her theory, an action commenced in federal district court “seeking the adjudication of a freestanding [federal-question] cause of action” always technically “invok[es] the district court’s original jurisdiction” under 28 U. S. C. §1331 and is akin to a collateral attack. Brief for Petitioner 32; see post, at 2–3 (opinion of Barrett, J.). Therefore, as T. M. sees things, the Court’s prior Rooker-Feldman cases should not be read to evince any concern about district courts exercising what amounts to appellate jurisdiction over state-court judgments, but instead should be read as relying on a strict negative inference, from §1257, that when this Court has jurisdiction over a state-court judgment, district courts must not. In T. M.’s view, it follows from here that, for Rooker-Feldman to bar a federal suit, it is not enough that a state-court loser seeks what is effectively appellate review of prior state-court judgment. For T. M., the judgment under attack must be a “[f]inal judgmen[t] or decre[e] rendered by the highest court of a State in which a decision could be had,” §1257(a), at the time the federal suit is initiated, because only then does this Court gain jurisdiction over a state-court judgment.

T. M.’s theory cannot be squared with this Court’s precedents, which plainly adopted a different, more functional view of original and appellate jurisdiction than what T. M. advances here. In Rooker, for example, it was true that the judgment at issue had been affirmed by the state high court by the time the federal suit was filed, but nothing in this Court’s reasoning suggested that the outcome would have been different had the judgment still been on appeal. There (as here), Rooker’s federal suit sought to “reverse or modify the [state-court] judgment,” which the Court held would be an “exercise of appellate jurisdiction,” 263 U. S., at 416, and not a collateral attack invoking original jurisdiction. There (as here), the District Court would lack such “appellate” jurisdiction because it has “strictly original” jurisdiction. Ibid. There (as here), it would be the “province and duty of the state courts to decide” the federal questions first, and “[i]f the decision was wrong,” Rooker could seek “reversal or modification in an appropriate and timely appellate proceeding,” including in this Court under §1257. Id., at 415.

Later cases delineating the bounds of Rooker-Feldman, see n. 4, supra, confirmed the understanding that Rooker-Feldman bars suits in federal district court that “see[k] what in substance would be appellate review of [a] state judgment,” regardless of whether the judgment formally falls within this Court’s §1257 jurisdiction. Johnson v. De Grandy, 512 U. S. 997, 1005–1006 (1994). As this Court explained in Exxon, the doctrine “ ‘recognizes that 28 U. S. C. §1331 is a grant of original jurisdiction, and does not authorize district courts to exercise appellate jurisdiction over state-court judgments, which Congress has reserved to this Court, see §1257(a).’ ” 544 U. S., at 292 (quoting Verizon Md. Inc. v. Public Serv. Comm’n of Md., 535 U. S. 635, 644, n. 3 (2002)). In Lance, too, this Court emphasized that “[t]he doctrine applies only in ‘limited circumstances’ where a party in effect seeks to take an appeal of an unfavorable state-court decision to a lower federal court.” 546 U. S., at 466 (citation omitted). None of these cases suggests that district courts have jurisdiction to review state-court judgments, and declare them void, so long as the judgments are not yet within this Court’s §1257 jurisdiction.

In fact, Feldman explicitly rejected that notion. There, the Court addressed a Fifth Circuit case holding that Rooker did not apply because the judgment “ ‘could not have been reviewed on a writ of certiorari from the United States Supreme Court,’ ” given the plaintiff ’s failure to raise the relevant federal issue in state court. 460 U. S., at 483, n. 16 (quoting Dasher v. Supreme Ct. of Tex., 658 F. 2d 1045, 1051 (1981)). The Court rejected that reasoning as “flawed” because “the fact that we may not have jurisdiction to review a final state-court judgment . . . does not mean that a United States district court should have jurisdiction over the claims.” 460 U. S., at 484, n. 16. After all, the Court explained, “ ‘lower federal courts possess no power whatever to sit in direct review of state court decisions.’ ” Id., at 483, n. 16; but see id., at 490 (Stevens, J., dissenting) (arguing, as T. M. and the dissent do here, that the Feldman majority improperly “fail[ed] to distinguish between . . . appellate review and collateral attack”). Feldman thus explicitly rejected T. M.’s view that Rooker rested solely on a strict negative inference from §1257.

The dissent, for its part, admits that this Court has relied on both §§1331 and 1257 in justifying Rooker-Feldman, but contends that Exxon “buried” and “discarded” the §1331 rationale. Post, at 7–8. Had the Court in Exxon intended to bury or discard the §1331 rationale, one might have expected it to say so rather than to quote affirmatively a case that directly describes the doctrine as resting on both §§1331 and 1257. See 544 U. S., at 292 (citing Verizon Md. Inc., 535 U. S., at 644, n. 3). Moreover, Exxon acknowledged, without disapproval, Rooker’s reasoning that “[f]ederal district courts . . . lacked the requisite appellate authority, for their jurisdiction was ‘strictly original.’ ” 544 U. S., at 284. It also summarized Rooker and Feldman as “essentially invit[ing] federal courts of first instance to review and reverse unfavorable state-court judgments,” 544 U. S., at 283, and it held that Exxon’s federal suit did not fall within Rooker-Feldman because Exxon did “not repai[r] to federal court to undo the Delaware judgment in its favor,” 544 U. S., at 293. All these statements are only consistent with the view that Rooker-Feldman is based on both §§1331 and 1257, and none even suggests that the state-court judgment at issue must be a “[f]inal judgmen[t] . . . rendered by the highest court of a State” under §1257(a).

Ultimately, adopting T. M. and the dissent’s rule would require the Court to abandon a central part of Rooker’s reasoning and reinterpret the doctrine to rest solely on a strict negative inference from §1257. See Brief for Petitioner 33–34 (acknowledging that the Court’s precedents have rested in part based on the distinction between original and appellate jurisdiction); post, at 3 (acknowledging that Rooker rested in part on §1331). The Court declines to do so.

B

The fact that T. M. and the dissent’s rule departs from this Court’s precedents is enough to affirm the dismissal of T. M.’s suit under Rooker-Feldman. If more is necessary, however, the Court also declines to distort the Rooker-Feldman doctrine in service of a rule that would create anomalous outcomes and undermine the federalism principles upon which the doctrine rests. The upshot of T. M. and the dissent’s rule would be that federal district courts could not exercise jurisdiction if a plaintiff seeks review of a state high-court judgment, but they could exercise jurisdiction if the plaintiff seeks the same review of a state lower-court judgment while that judgment is on appeal in the state-court system. That makes little sense.

To start, allowing federal district courts to review state-court judgments while they are on appeal in the state-court system would undermine the “[c]ooperation and comity” on which our federal system is built. Ruhrgas AG v. Marathon Oil Co., 526 U. S. 574, 586 (1999). Indeed, that is the reason why, “ever since 1789, Congress has granted this Court the power to intervene in State litigation only after ‘the highest court of a State in which a decision in the suit could be had’ has rendered a ‘final judgment.’ ” Radio Station WOW, Inc. v. Johnson, 326 U. S. 120, 124 (1945); see North Dakota Bd. of Pharmacy v. Snyder’s Drug Stores, Inc., 414 U. S. 156, 159 (1973) (explaining that one of §1257’s purposes is to “limi[t] review of state court determinations of federal constitutional issues to leave at a minimum federal intrusion in state affairs”). It is difficult to see why Congress would have authorized district courts to review final state-court judgments pending appeal in state court when it did not authorize this Court to do so.

T. M. and the dissent’s rule also would produce arbitrarily different results depending on when, exactly, a federal suit seeking review of a state-court judgment is filed. In their view, a district court can review a state-court judgment so long as it has not been affirmed by the state high court by the time the federal action is filed. Yet a state high court may well affirm such a judgment while such a federal action is still pending. In that event, the plaintiff would be in the exact same situation Rooker disapproved: seeking to have declared null and void “a judgment of a [state court], which was affirmed by the Supreme Court of the State.” 263 U. S., at 414. Yet, under T. M. and the dissent’s rule, a district court would have jurisdiction to proceed simply because the plaintiff filed her federal suit while the state appellate proceedings were pending, rather than waiting for the state high court to act. Such a rule would only encourage parties to file earlier in federal court, duplicating their efforts even though the state process may well resolve the dispute in their favor.

T. M., her amici, and the dissent respond that federal courts have at their disposal various abstention and preclusion doctrines, which generally prevent federal courts from interfering with state proceedings or bar parties from relitigating issues already decided. According to T. M., those doctrines prevent federal-court intervention while state appellate proceedings are ongoing and thus do “nearly all,” but apparently not all, “of the work that Rooker-Feldman does.” Reply Brief 18; see Brief for Federal Courts Scholars as Amici Curiae 7–16 (listing nine doctrines); see also post, at 4–5 (opinion of Barrett, J.) (describing preclusion and abstention doctrines). It is unclear, however, if such doctrines even apply in cases, like this one, where a plaintiff does not attempt to relitigate the same claims in state and federal forums but rather complains of injuries that stem directly from a state-court judgment. See Letter Order in No. 1:23–cv–1684 (D Md.), ECF Doc. 27, pp. 3–5 (holding that neither abstention under Younger v. Harris, 401 U. S. 37 (1971), nor Colorado River Water Conservation Dist. v. United States, 424 U. S. 800 (1976), nor Maryland’s collateral-attack doctrine applied to this case); see also Lucky Brand Dungarees, Inc. v. Marcel Fashions Group, Inc., 590 U. S. 405, 411 (2020) (explaining that issue preclusion “precludes a party from relitigating an issue actually decided in a prior case and necessary to the judgment”). Federalism principles are thus best served by continuing to apply Rooker-Feldman to federal cases in which plaintiffs seek review of state-court judgments, regardless of whether those judgments are final trial-court judgments or those of a State’s highest court.

C

Finally, T. M. contends that the Court should adopt her position for the more practical reason that it will generally cabin Rooker-Feldman, a doctrine she claims has caused confusion and is overused as a docket-clearing mechanism in the lower courts. See Brief for Petitioner 35. As T. M. admits, however, most of the confusion in the lower courts “stems from the need to determine when a plaintiff is seeking federal ‘review and rejection’ of a state-court judgment.” Id., at 37; see Gilbank v. Wood Cty. Dept. of Human Servs., 111 F. 4th 754, 792–798 (CA7 2024) (en banc) (majority opinion of Kirsch, J.); id., at 760, 769–778 (dissenting opinion of Hamilton, J.) (describing the confusion over the “ ‘review and rejection’ ” requirement). That “review and rejection” question is not disputed in this case, so adopting T. M.’s rule does nothing to ameliorate any confusion over that distinct issue.

In fact, T. M. and the dissent’s rule would needlessly complicate the Rooker-Feldman analysis. It is not always straightforward to determine whether a given state-court judgment is a reviewable final judgment under §1257(a) because that question requires determining whether the judgment is both “subject to no further review or correction in any other state tribunal” and also “an effective determination of the litigation.” Market Street R. Co. v. Railroad Comm’n of Cal., 324 U. S. 548, 551 (1945). Nor is it necessarily clear when state proceedings have “ended.” See Federación de Maestros de P. R., 410 F. 3d, at 24–25 (describing test to determine whether proceedings have functionally “ended”). For example, in this case, the parties dispute whether the consent order is appealable under Maryland law. See Brief in Opposition 23–24; Reply Brief for Petitioner 9–10. If it is not, then the consent order may be final for §1257 purposes and the “end” of state-court proceedings. Further, even if a state judgment is technically subject to further review in state proceedings, it may nevertheless be reviewable by this Court under §1257 under a limited set of circumstances. See Cox Broadcasting Corp. v. Cohn, 420 U. S. 469, 477–486 (1975) (describing four sets of circumstances within the Court’s “pragmatic approach” to finality). Thus, T. M. and the dissent’s rule would both fail to address the source of confusion in the current Rooker-Feldman doctrine and add a new source to the mix.

Finally, the animating force behind many of T. M.’s and the dissent’s arguments appears to be the belief that Rooker and Feldman were wrongly decided and so should be cabined whenever possible, if not outright overruled. Cf. post, at 2–5 (raising doubts about Rooker-Feldman’s conceptual basis). This issue is not fairly included in the question presented and was not pressed at the certiorari stage. See this Court’s Rule 14.1(a).

IV

The Court today neither expands nor constrains Rooker-Feldman. Instead, the Court leaves the doctrine as it found it: narrowly confined to “cases brought by state-court losers complaining of injuries caused by state-court judgments rendered before the district court proceedings commenced and inviting district court review and rejection of those judgments.” Exxon, 544 U. S., at 284. Because T. M.’s suit falls within these strict limits, the judgment of the Court of Appeals for the Fourth Circuit is affirmed.

It is so ordered.

SUPREME COURT OF THE UNITED STATES

_________________

No. 25–197

_________________

T. M., PETITIONER v. UNIVERSITY OF MARYLAND MEDICAL SYSTEM CORPORATION, et al.

on writ of certiorari to the united states court of appeals for the fourth circuit

[June 18, 2026]

Justice Thomas, concurring.

When you lose in trial court, you go to an appeals court. The so-called Rooker-Feldman doctrine reflects that commonsense understanding. See Rooker v. Fidelity Trust Co., 263 U. S. 413 (1923); District of Columbia Court of Appeals v. Feldman, 460 U. S. 462 (1983). At its core, the doctrine says that parties who lose in state trial court do not get to appeal to a federal trial court. Instead, they may take their appeal to a state appeals court and, if a federal statute allows, ultimately seek review in this Court. That “basic fact” has been a part of our law for more than 200 years. Reed v. Goertz, 598 U. S. 230, 244 (2023) (Thomas, J., dissenting). The Court’s opinion thus properly rejects petitioner’s quest to reimagine, artificially cabin, or overrule Rooker-Feldman. Ante, at 5–18. I join it in full. I write separately to explain why Rooker, properly understood, is correct as an original matter.

I

A unanimous three-page opinion, Rooker was uncontroversial in its day. It started as an ordinary lawsuit. William and Dora Rooker sued Fidelity Trust Company in state court. See Rooker v. Fidelity Trust Co., 191 Ind. 141, 143, 131 N. E. 769, 770 (1921). The trial court entered judgment against the Rookers. Id., at 154–155, 131 N. E., at 773. The Rookers sought review in the State Supreme Court and lost. Id., at 163, 131 N. E., at 776. This Court denied certiorari. Rooker v. Fidelity Trust Co., 259 U. S. 580 (1922). The Rookers then sought review from this Court on a writ of error instead. Rooker v. Fidelity Trust Co., 261 U. S. 114, 116 (1923). But, the Court declined to grant relief on their claims of constitutional errors in the state-court judgment and dismissed the writ. Id., at 116–118.

Unsatisfied, the Rookers filed a bill in equity in Federal District Court. Rooker, 263 U. S., at 414. They asked the Federal District Court to declare the state trial court’s judgment “null and void” on the ground that it “was rendered and affirmed in contravention of the” Constitution. Id., at 414–415. The District Court dismissed the bill for lack of jurisdiction, and this Court affirmed that “plainly” correct decree. Id., at 415, 417. “Under the legislation of Congress,” this Court recognized, “no court of the United States other than this Court could entertain a proceeding to reverse or modify the judgment for” constitutional errors. Id., at 416. “To do so would be an exercise of appellate jurisdiction.” Ibid. “The jurisdiction possessed by the District Courts is strictly original.” Ibid. Accordingly, the Court held that district courts have no jurisdiction to “get rid of ” state-court judgments even “for alleged errors of law.” Ibid.

Rooker was thus based on two premises. See ante, at 5–7, 9–10. First, “to reverse or modify the” allegedly erroneous judgment of another court is to “exercise . . . appellate jurisdiction.” Rooker, 263 U. S., at 416. Second, “[t]he jurisdiction possessed by the District Courts is strictly original”; only “this Court” has “appellate jurisdiction” over state-court judgments. Ibid. From these two premises, it follows that district courts lack jurisdiction to reverse or modify the allegedly erroneous judgments of state courts. The conclusion follows from the premises, so the only question is whether the premises are true.

II

Rooker’s first premise was recognized long before Rooker, and its second has been an indubitable “fact of federal statutory law” for more than 200 years. Reed, 598 U. S., at 244 (opinion of Thomas, J.).

A

The power to revise or modify another court’s judgment or decree has been recognized as an exercise of appellate jurisdiction from before the ratification of the Constitution.

1

The Constitution vests the “judicial Power” in this Court and in the “inferior Courts” that Congress creates. Art. III, §1. The jurisdiction of federal courts to exercise that judicial power is limited to “Cases” and “Controversies.” §2, cl. 1. That jurisdiction, in turn, comes in two types: “original” and “appellate.” §2, cl. 2. If a court exercises appellate jurisdiction, it does not exercise original jurisdiction. Marbury v. Madison, 1 Cranch 137, 175 (1803); 3 J. Story, Commentaries on the Constitution of the United States §1698, p. 574 (1833) (Story).

Original jurisdiction is “[a] court’s power to hear and decide a matter before any other court can review the matter.” Black’s Law Dictionary 856 (7th ed. 1999). In a typical original case, a plaintiff sues a defendant. The plaintiff complains of an injury that the defendant’s allegedly illegal conduct caused and “seeks a remedy that runs against the defendant.” Reed, 598 U. S., at 243 (opinion of Thomas, J.). The case is “original” inasmuch as there is no earlier proceeding in which another court had already “determine[d] that defendant’s duties or liabilities (e.g., a judgment for money damages or an injunction).” Ibid.

Unlike the plaintiff in a typical original action, the appellant in a typical appeal cannot simply complain of an injury that the other party caused before a court has spoken. Instead, “the subject matter” under consideration “has already been instituted in, and acted upon, by some other court.” 3 Story §1755, at 627. As the party who lost in the other court, the appellant “complain[s] of an injury caused by” that court’s allegedly erroneous judgment. Exxon Mobil Corp. v. Saudi Basic Industries Corp., 544 U. S. 280, 291 (2005). He thus seeks to invoke the court’s jurisdiction to remedy an injury of that description and, ordinarily, the appellate court cannot directly redress the injury that caused the plaintiff to file his original suit. The remedy, instead, consists in “correcting that judicial action (e.g., reversal or vacatur of the challenged judgment),” not in awarding damages. Reed, 598 U. S., at 244 (opinion of Thomas, J.).

When a party asks one court to revise another’s judgment, he invokes what the Constitution calls “appellate Jurisdiction.” Art. III, §2, cl. 2. Appellate jurisdiction is a court’s “power to revise the judgments rendered” by other courts. 1 J. Kent, Commentaries on American Law 353 (11th ed. 1867) (Kent); accord, 3 W. Blackstone, Commentaries on the Laws of England 56 (1768) (Blackstone) (“reverse judgments in certain suits originally begun in [another] court”); ibid. (high court had appellate jurisdiction to correct “mistake[s] of the law, committed by [other] courts”); W. Rawle, A View of the Constitution 241 (2d ed. 1829) (Rawle) (“revision of the [decisions] of other courts”); 3 Story 626 (power to “revis[e] and correc[t] the proceedings in a cause already instituted”); 1 G. Curtis, Commentaries on the Jurisdiction, Practice, & Peculiar Jurisprudence of the Courts of the United States §113, p. 131 (1854) (Curtis) (same); Black’s Law Dictionary 79 (1891) (same); 1 A. Burrill, A New Law Dictionary 79 (1850) (“review the proceedings of another [court]”). For that reason, exercising that revising power over the judgments of “state tribunals” in cases involving “[federal] questions” is appellate in nature. 1 Kent 353.

The Founders accepted this straightforward account of appellate jurisdiction without reservation. The Articles of Confederation gave the Confederation Congress authority to create courts with “appellate power . . . to be exercised in revising the decisions of state tribunals” in cases of captures. Martin v. Hunter’s Lessee, 1 Wheat. 304, 345 (1816) (Story, J., for the Court); see Art. IX. In 1780, the Confederation Congress created a court under that authority whose jurisdiction allowed it to “hear new evidence without deference to state courts’ ” legal conclusions. W. Pryor, The Appellate Jurisdiction Clause, in The Heritage Guide to the Constitution 499 (3d ed. 2025) (Pryor). That power was “appellate” in nature precisely because it was the “power to revise the decisions of state courts.” Hunter’s Lessee, 1 Wheat., at 345; 3 Story 605–606.

When the Founders deliberated about whether to replace the Articles with the Constitution, all maintained the same view about the nature of appellate jurisdiction. At that time, different “technical sense[s] ha[d] been affixed to the term ‘appellate’ ” such that any technical interpretation would “not be understood in the same sense” in different jurisdictions. The Federalist No. 81, pp. 488–489 (C. Rossiter ed. 1961) (A. Hamilton). That fact showed “the impropriety of a technical interpretation.” Id., at 489; accord, ante, at 10. Instead, the Federalists thought that the term “appellate jurisdiction” in the proposed Constitution “denotes nothing more than the power of one tribunal to review the proceedings of another, either as to the law or fact, or both.” The Federalist No. 81, at 489. For their part, the Anti-Federalists seemed to agree. They worried that “appellate jurisdiction,” as used in the Constitution, would authorize civil-law “appeals,” which involved the power to “re-examine the whole merits of ” a previously initiated case, including review of facts. 2 H. Storing, The Complete Anti-Federalist 433 (1981); see also United States v. Wonson, 28 F. Cas. 745, 750 (No. 16,750) (CC Mass. 1812) (Story, J.). All appeared to agree, however, that “the revising power” of a court is “appellate” in nature. Hunter’s Lessee, 1 Wheat., at 344.

Traditionally, the “most usual modes of exercising appellate jurisdiction” were “by a writ of error, or by an appeal.” 3 Story §1756, at 627. But, courts could also exercise appellate jurisdiction through writs of habeas corpus or mandamus, provided that granting such a writ would involve revising the decision of another court on the merits. Ibid.; 1 Curtis §113, at 131; Marbury, 1 Cranch, at 175. Today, this Court usually exercises appellate jurisdiction by writs of certiorari. 28 U. S. C. §§1257(a), 1254(1). But, it also hears “appeal[s],” in some cases, §1253, applications for stays of lower court judgments, in others, §2101(f ), and even petitions for mandamus, §1651(a).

Thus, whether jurisdiction is appellate turns not on the specific mode of proceeding but on the nature of the relief sought. See Reed, 598 U. S., at 244 (opinion of Thomas, J.). Scholars and jurists have long recognized that “any . . . mode in which the judgment or proceedings of an inferior tribunal c[ould] be revised” was an exercise of appellate jurisdiction. 1 Curtis §113, at 131 (emphasis added). A court therefore acts as “a court of appeal” whenever it “correct[s] the errors of other jurisdictions,” 3 Blackstone 55, whatever the mode “by which a cause is transmitted to” it, Wonson, 28 F. Cas., at 748; accord, 3 Story §1755, at 627 (“[W]here the object is to revise a judicial proceeding, the mode is wholly immaterial”).

2

This Court has long understood that to exercise the revising power is to exercise appellate jurisdiction, regardless of the form of the exercise.

Most law students learn as much on the first day of constitutional-law class. In Marbury v. Madison, the Court considered whether it had the jurisdiction “to issue a mandamus” compelling “an officer for the delivery of ” Marbury’s judicial commission when no other court had before entertained the cause. 1 Cranch, at 175. Writing for the Court, Chief Justice Marshall held that the Court lacked such jurisdiction. The Constitution, he said, “define[d] the jurisdiction of the supreme court by declaring the cases in which it shall take original jurisdiction, and that in all others it shall take appellate jurisdiction.” Ibid. These categories, he reasoned, are mutually exclusive: “[I]n one class of cases its jurisdiction is original, and not appellate; in the other it is appellate, and not original.” Ibid. Because the Constitution did not give the Court original jurisdiction to issue mandamus, the Court could do so only if it would “be an exercise of appellate jurisdiction” in the case before it. Ibid.

The form of proceeding was not dispositive of the question. Even though mandamus was often issued in original cases, and the case was filed first in the Supreme Court, the Court agreed that “appellate jurisdiction may be exercised in a variety of forms, and that if it be the will of the legislature that a mandamus should be used for that purpose, that will must be obeyed.” Ibid. Because “a mandamus may be directed to courts,” the question became whether, in context, granting mandamus relief would be appellate in nature. Ibid.

On the facts of the case, the exercise of jurisdiction was original in nature. There was no decision of another court that this Court could revise. Chief Justice Marshall recognized that it “is the essential criterion of appellate jurisdiction, that it revises and corrects the proceedings in a cause already instituted.” Ibid. To issue mandamus “to an officer for the delivery of a paper, is in effect the same as to sustain an original action for that paper, and therefore seems not to belong to appellate, but to original jurisdiction.” Id., at 175–176; accord, ante, at 10.

From Marbury’s holding, then, a clear principle emerged: When a court “revises” or “corrects the proceedings in a cause already instituted,” it exercises appellate jurisdiction; when there is no previous “cause” or judicial proceeding to revise, it exercises “original” jurisdiction. 1 Cranch, at 175. And, from Marbury’s reasoning, another principle followed: A court can exercise appellate jurisdiction even if the action is filed outside of the most common appellate forms, such as a writ of error or a modern appeal, provided that there is another court’s decision to be revised. Ibid.

Chief Justice Marshall reaffirmed both principles for the Court just a few years later. See Rawle 227–228. Ex parte Bollman, 4 Cranch 75 (1807), involved two individuals—Bollman and Swartwout—who were implicated in the infamous Burr conspiracy. Both had been jailed by the circuit court for the District of Columbia pending their trial. Id., at 75–76. And, each sought a writ of habeas corpus from this Court. Id., at 75. The question, as in Marbury, was whether this Court had jurisdiction to issue the writs. This time the Court held that it had such jurisdiction.

The arguments before the Court concerned the nature of original and appellate jurisdiction. Counsel for Bollman argued that the Judiciary Act gave “this court the power to grant writs of habeas corpus,” and that “congress had authority, by the constitution,” to give the Court such power. 4 Cranch, at 84–85. But, this second point, he argued, depended on “whether this power or jurisdiction be in its nature original or appellate.” Id., at 85. He conceded that, if granting the writ would be an exercise of original jurisdiction, then Congress could not give this Court the power to issue the writ. Id., at 85–86. “This principle,” he acknowledged, had been “established by the case of Marbury v. Madison.” Id., at 86. Following Marbury, he argued that the “criterion which distinguishes appellate from original jurisdiction, is that it revises and corrects the decisions of another tribunal.” 4 Cranch, at 86. And, the “object of the habeas corpus now applied for,” he urged, “is to revise and correct the proceedings of the Court below . . . so far as respects the legality of such commitment.” Ibid.

Again writing for the Court, Chief Justice Marshall agreed with this approach. He began by explaining that federal courts, as creatures of “written law,” cannot “transcend th[e] jurisdiction” that is “defined by written law.” Id., at 93. “[F]or the meaning of the term habeas corpus,” he said, “resort may unquestionably be had to the common law; but the power to award the writ by any of the courts of the United States, must be given by written law.” Id., at 93–94. So, the question was the same as in Marbury: “whether by any statute, compatible with the constitution of the United States, the power to award the writ . . . has been given to this court.” 4 Cranch, at 94.

This time, though, Chief Justice Marshall answered yes. The Judiciary Act, he concluded, purported to give “the power to award writs of habeas corpus in order to examine into the cause of commitment.” Id., at 100. And, he “acknowledged that the writ was sought in the first instance in the Supreme Court.” Pryor 500. But, the Constitution did not stand in the way of the Court’s issuing the writ in the case before it. Though “the mandamus case” of Marbury required that the Court “would not exercise original jurisdiction,” he explained, “that which the court is now asked to exercise is clearly appellate.” 4 Cranch, at 100–101. He reasoned that the “decision that the individual shall be imprisoned must always precede the application for a writ of habeas corpus.” Id., at 101. In other words, a court had already addressed the issue of the imprisonment, so there was “a cause already instituted.” Marbury, 1 Cranch, at 175. Because “this writ must always be for the purpose of revising that decision,” he concluded that it is “therefore appellate in its nature.” 4 Cranch, at 101.

*  *  *

Thus, both Marbury and Bollman support Rooker’s first premise that to reverse or modify the allegedly erroneous judgment of another court is to exercise appellate jurisdiction. Federal courts must inquire into the nature of the jurisdiction a party seeks to invoke and whether it takes “an appellate form,” regardless of the label placed on the paper filed in the court. Ex parte Watkins, 7 Pet. 568, 573 (1833) (Story, J., for the Court); accord, ante, at 10. And, they further explain that “seek[ing] to revise the acts of ” another court is “appellate” in nature. Ex parte Watkins, 7 Pet., at 573. Whether the court hears the claim in the form of a mandamus petition as in Marbury, a habeas petition as in Bollman, or a bill in equity as in Rooker, if a party asks one court to revise the judgment of another court for errors of law, it invokes appellate jurisdiction. The question is then whether the court has revising power to “declar[e]” a state-court judgment “a nullity [on the merits].” Voorhees v. Jackson, 10 Pet. 449, 474 (1836).

B

Congress has long given only this Court, and not inferior courts, appellate jurisdiction to revise state-court judgments.

The Constitution defines the limits of this Court’s jurisdiction. Its “original Jurisdiction” is limited to “Cases affecting Ambassadors, other public Ministers and Consuls, and those in which a State shall be Party.” Art. III, §2, cl. 2. In “all” other cases, “in Law and Equity,” this Court “shall have appellate Jurisdiction,” subject to any “Exceptions, and under such Regulations as the Congress shall make.” §2, cls. 1–2. Only this Court “possesses jurisdiction derived immediately from the constitution.” United States v. Hudson, 7 Cranch 32, 33 (1812).

As creatures of statute, inferior federal courts “possess no jurisdiction but what is given them by the power that creates them.” Ibid. They have “none” beyond what Congress “authorize[s].” Ibid. Whether inferior federal courts have revising power thus depends on whether a statute says that they do. C. Williams, Jurisdiction and Practice of Federal Courts 433 (1917) (Williams). Today, statutes say that courts of appeals have revising power over district-court judgments, 28 U. S. C. §1291, and that district courts have revising power over bankruptcy-court judgments, §158(a)(1).

Since the founding, there has been no doubt that inferior “federal” courts have the power “to review the proceedings of ” “State courts” only if Congress, in its “discretion,” has given them “appellate jurisdiction” to do so. The Federalist Nos. 81, 82, at 488–490, 493–495 (A. Hamilton); see also Hudson, 7 Cranch, at 33. Congress “is not limited by the constitution to any particular mode, or time of exercising it,” but there must be a statutory grant of revising power over state-court judgments before inferior courts can revise them. Hunter’s Lessee, 1 Wheat., at 349; accord, 3 Story §1755, at 627 (“any form, which the legislature may choose to prescribe”).

Congress has never purported to give inferior federal courts general revising power over state-court judgments for errors of federal law. The First Congress declared that, in federal-question cases, “a final judgment or decree” of “the highest court of law or equity of a State” “may be re-examined and reversed or affirmed in” this Court. Judiciary Act of 1789, §25, 1 Stat. 85–86. Congress called that revising power “appellate jurisdiction” and identified the mode as a writ of error. §§13, 25, id., at 80–81, 85–86. It gave no other federal court revising power over “courts of the several states.” §13, id., at 81. And, it gave no other federal court civil federal-question jurisdiction, either. Cf. §§9, 11, id., at 75–79. Thus, only state courts could exercise original jurisdiction in federal-question cases, and only this federal Court could revise their judgments.

It took Congress 86 years to give federal district courts federal-question jurisdiction, but it limited that jurisdiction to original actions. See Schweiker v. Chilicky, 487 U. S. 412, 420–421 (1988). In 1875, Congress granted federal courts “original” jurisdiction, “concurrent with the courts of the several States,” over federal-question cases. Act of Mar. 3, 1875, ch. 137, §1, 18 Stat. 470. But, that grant of original jurisdiction was not a grant of revising power. No inferior federal court was “empowered . . . to enjoin the judgment of a State court.” D. Chamberlain, The State Judiciary, in Constitutional History of the United States 254 (1889); id., at 253–259. State-court “judgments and decrees” were “subject in all cases . . . arising under the Constitution,” or federal law, “to the revising or appellate jurisdiction of the Supreme Court . . . , but in all other cases controlled only by the constitutions and laws of the respective States.” Id., at 259.

Nothing relevant has changed since 1875. See Atlantic Coast Line R. Co. v. Locomotive Engineers, 398 U. S. 281, 286 (1970). Thus was the state of the law before Rooker. 263 U. S., at 416; Act of Mar. 3, 1911, §§24–27, 36 Stat. 1091–1094; Act of Sept. 6, 1916, ch. 448, §2, 39 Stat. 726; Williams 48–49 (“The appellate jurisdiction [of district courts] is very slender”). And, it remains true today that district courts have only “original jurisdiction” in federal-question cases, 28 U. S. C. §1331, and that only this Court has appellate jurisdiction over state courts in such cases, §1257(a). See ante, at 5. Congress specified no mode for district courts to exercise appellate jurisdiction in federal-question cases, §1331; the only federal court authorized to do so over state courts is this one—“by writ of certiorari,” §1257(a).

A district court therefore cannot “act as an appellate Court” to revise state-court judgments. Smith v. McIver, 9 Wheat. 532, 535 (1824) (Marshall, C. J., for the Court). To do so by “declar[ing]” a state-court judgment “void” for errors of federal law “usurps the jurisdiction of an appellate court.” Voorhees, 10 Pet., at 474.

III

As “creatures of a distinct government,” “state courts are not, in any sense of the word, inferior” to inferior federal courts. Bollman, 4 Cranch, at 97. From the founding, state courts have had original jurisdiction in federal-question cases, and their judgments in such cases could be revised only in state appellate courts or in this Court. Ante, at 13–16. Congress did not give federal district courts revising power over state-court judgments by giving them only original jurisdiction in federal-question cases. Grants of “original jurisdiction” do “not authorize district courts to exercise appellate jurisdiction.” Verizon Md. Inc. v. Public Serv. Comm’n of Md., 535 U. S. 635, 644, n. 3 (2002). Rooker thus “correctly applied the simple legal proposition that only this Court may exercise appellate jurisdiction over state-court judgments.” Lance v. Dennis, 546 U. S. 459, 467 (2006) (Stevens, J., dissenting). The Court reaffirms that longstanding proposition. Ante, at 5–8. I therefore join its opinion in full.

SUPREME COURT OF THE UNITED STATES

_________________

No. 25–197

_________________

T. M., PETITIONER v. UNIVERSITY OF MARYLAND MEDICAL SYSTEM CORPORATION, et al.

on writ of certiorari to the united states court of appeals for the fourth circuit

[June 18, 2026]

Justice Barrett, with whom The Chief Justice, Justice Kagan, and Justice Gorsuch join, dissenting.

Twenty years ago, this Court held that the Rooker-Feldman doctrine is “confined” to the procedural circumstances of the two cases from which the doctrine draws its name. Exxon Mobil Corp. v. Saudi Basic Industries Corp., 544 U. S. 280, 284, 291 (2005); see Rooker v. Fidelity Trust Co., 263 U. S. 413 (1923); District of Columbia Court of Appeals v. Feldman, 460 U. S. 462 (1983). In “both cases,” we emphasized, the federal action was brought “after the state proceedings ended.” Exxon, 544 U. S., at 291. Seven Courts of Appeals took us at our word, refusing to apply Rooker-Feldman when the underlying state action remained pending. They were right to hold the line. Because the Court has chosen to relax it, I respectfully dissent.

I

Until today, this Court has applied the Rooker-Feldman doctrine “only twice.” Exxon, 544 U. S., at 283. That is no accident: Because Rooker-Feldman stands on shaky ground, we have consciously kept its footprint small. Its rationale has gotten no firmer, so we should make the doctrine no larger.

Rooker-Feldman is a “maddeningly elusive” doctrine. B. Friedman & J. Gaylord, Rooker-Feldman, from the Ground Up, 74 Notre Dame L. Rev. 1129, 1133 (1999). We have primarily justified it as a negative inference from 28 U. S. C. §1257: Because Congress gave this Court appellate jurisdiction over certain state-court judgments, district courts must lack original jurisdiction to entertain a collateral attack against those judgments. See Rooker, 263 U. S., at 415–416; Feldman, 460 U. S., at 476. This Court’s exercise of appellate jurisdiction, the argument runs, is “functional[ly]” equivalent to a district court’s exercise of original jurisdiction over a collateral attack. Ante, at 10. The problem? A collateral attack is not functionally equivalent to an appeal.

Consider just a few of the differences. An appeal “removes the record into the supervising tribunal.” Cohens v. Virginia, 6 Wheat. 264, 410 (1821). A collateral attack does not. See Fed. Rules Civ. Proc. 3, 7. On appeal, new evidence is generally unwelcome. See Fed. Rule App. Proc. 10; United States v. Coe, 155 U. S. 76, 83–84 (1894). In a collateral attack, new evidence is fair game. See, e.g., Fed. Rule Civ. Proc. 26. Most obviously, preclusion does not apply on appeal. This Court is not barred from considering an issue just because the state high court already decided it—reconsidering the state court’s judgment is the whole point. Not so with a collateral attack; preclusion restrains district courts from simply redoing the state court’s work. See, e.g., Parsons Steel, Inc. v. First Alabama Bank, 474 U. S. 518, 523 (1986). Finally, while this Court can vacate or reverse the judgment of the state high court and remand for further proceedings, a district court entertaining a collateral attack cannot. See 7 J. Lucas, Moore’s Federal Practice ¶60.36, pp. 60–365 to 60–371 (2d ed. Cum. Supp. 1996–1997) (explaining that collateral attacks “d[o] not operate to vacate or annul the [original] judgment”). Because of the many differences between appellate jurisdiction and original jurisdiction over a collateral attack, it is doubtful that by giving this Court the former, Congress intended to deprive district courts of the latter.

Occasionally, we have also cited 28 U. S. C. §1331 as support for the Rooker-Feldman doctrine, observing that it gives district courts “original,” not appellate, jurisdiction. Rooker, 263 U. S., at 416. This observation adds nothing—as just explained, a collateral attack is not the functional equivalent of an appeal. Besides, §1331 cannot explain why Rooker-Feldman has always been limited to collateral attacks on state-court judgments. If entertaining a collateral attack goes beyond “original jurisdiction,” then it should not matter whether the underlying judgment is from a state court or a federal one. Yet it is well settled that district courts can entertain collateral attacks on federal judgments—for example, an independent action for relief from a judgment to prevent a “grave miscarriage of justice.” United States v. Beggerly, 524 U. S. 38, 45–47 (1998); see Fed. Rule Civ. Proc. 60(d) (Rule 60 does not limit a court’s power to “entertain an independent action to relieve a party from a judgment, order, or proceeding”). It is little wonder that (at least until today) we had backed away from the §1331 rationale.

Beyond the difficulties of finding a statutory basis for Rooker-Feldman, parts of the United States Code undercut it. The doctrine is in tension, to say the least, with statutes granting district courts jurisdiction in categorical terms. Section 1331 gives district courts “original jurisdiction” over “all civil actions” presenting a federal question—not “all civil actions” presenting a federal question “minus those that attack a state-court judgment.” The other jurisdictional grants are similar. See, e.g., 28 U. S. C. §1332 (“The district courts shall have original jurisdiction of all civil actions” meeting certain diversity and amount-in-controversy requirements). None has a carveout for collateral attacks.

More to the point, in trying to read between the lines of §1257 and §1331, Rooker-Feldman misses what is further down the page: Congress has told us how to treat state-court judgments. The Full Faith and Credit Act, 28 U. S. C. §1738, provides that the “judicial proceedings of any court” of any State “shall have the same full faith and credit in every court within the United States . . . as they have by law or usage in the courts of such State . . . from which they are taken.” In other words, federal courts “must give the same preclusive effect to a state-court judgment as another court of that State would give.” Parsons Steel, 474 U. S., at 523.

Conceptual problems aside, it is not clear what Rooker-Feldman brings to the table. Preclusion already bars parties from relitigating issues and claims that were decided in state court. See, e.g., ibid. And when state litigation is ongoing, various abstention doctrines prevent federal courts from getting involved. Younger abstention, for example, precludes federal courts from interfering with state criminal prosecutions and certain state civil proceedings. See Younger v. Harris, 401 U. S. 37 (1971); Sprint Communications, Inc. v. Jacobs, 571 U. S. 69, 72–73 (2013); see also Colorado River Water Conservation Dist. v. United States, 424 U. S. 800, 818–821 (1976) (Colorado River abstention); Railroad Comm’n of Tex. v. Pullman Co., 312 U. S. 496, 500–501 (1941) (Pullman abstention); Burford v. Sun Oil Co., 319 U. S. 315, 332–334 (1943) (Burford abstention). Finally, the Anti-Injunction Act generally bars federal courts from enjoining state proceedings. 28 U. S. C. §2283.

None of this is to say that Rooker-Feldman is inconsequential. Crucially, it is jurisdictional—so courts must analyze it, even if no one raises it. And because Rooker-Feldman lacks both a clear role and a clear rationale, it is hard to contain. After Feldman, the doctrine became “perhaps the primary docket-clearing workhorse for the federal courts.” S. Bandes, The Rooker-Feldman Doctrine: Evaluating Its Jurisdictional Status, 74 Notre Dame L. Rev. 1175 (1999). “In the process,” Rooker-Feldman “mushroomed well beyond the §1257 explanation that gave it birth, as the federal courts found one claim after another closely intertwined with claims raised, resolved, sometimes still pending, in the state courts.” VanderKodde v. Mary Jane M. Elliott, P. C., 951 F. 3d 397, 406 (CA6 2020) (Sutton, J., concurring).

II

A

Recognizing that things had gone awry, this Court intervened. In Exxon, we neutralized Rooker-Feldman by confining it to the procedural circumstances of the Rooker and Feldman cases. 544 U. S., at 284, 291. Our treatment of the doctrine was so devastating that one scholar penned its obituary. See S. Bray, Rooker Feldman (1923–2006), 9 Green Bag 2d 317 (2006).

Faithfully read, Exxon requires us to reject the application of Rooker-Feldman to T. M.’s case. In Exxon, we lamented that lower courts had construed the doctrine “to extend far beyond the contours of the Rooker and Feldman cases.” 544 U. S., at 283. We acknowledged its tension with both jurisdictional statutes like §1331 and the Full Faith and Credit Act, observing that the doctrine risks “overriding Congress’ conferral of federal-court jurisdiction concurrent with jurisdiction exercised by state courts” and “superseding the ordinary application of preclusion law pursuant to 28 U. S. C. §1738.” Ibid.; see id., at 284. And we drew a line: Rooker-Feldman “is confined to cases of the kind from which the doctrine acquired its name.” 544 U. S., at 284. We then identified the key features of those cases. Id., at 291. Among them: In “both” Rooker and Feldman, the losing party sued in federal court “after the state proceedings ended.” 544 U. S., at 291 (emphasis added).

This limitation flows directly from Exxon’s logic. Exxon confines Rooker-Feldman to cases like Rooker and Feldman for a reason: Those two cases “exhibit the limited circumstances in which this Court’s appellate jurisdiction over state-court judgments” under §1257 “precludes a United States district court from exercising subject-matter jurisdiction in an action it would otherwise be empowered to adjudicate under a congressional grant of authority, e.g., §1330 (suits against foreign states), §1331 (federal question), and §1332 (diversity).” 544 U. S., at 291. This Court’s appellate jurisdiction is limited to “final judgments or decrees rendered by the highest court of a State in which a decision could be had.” §1257(a). If there is no final judgment from a state high court, §1257 does not confer jurisdiction in this Court; thus, there is nothing to “preclud[e]” district courts from exercising the jurisdiction they would “otherwise” have. Id., at 291.

Here, the state proceedings have not “ended.” T. M.’s original action is still pending in a Maryland appellate court. See ante, at 3. So this case falls outside the narrow confines in which Rooker-Feldman continues to operate, and the District Court had jurisdiction. See Exxon, 544 U. S., at 291.

B

Rather than hew to this straightforward analysis, the Court reshapes Rooker-Feldman. It does so by both retreating to the old (the §1331 rationale that Exxon discarded) and inventing the new (a grab bag of policy reasons for disregarding §1257’s finality requirement). This reconception of Rooker-Feldman leaves the doctrine worse off.

The Court’s first mistake is failing to believe that Exxon means what it says. The Court acknowledges that Exxon limits Rooker-Feldman to “ ‘cases of the kind from which the doctrine acquired its name.’ ” Ante, at 7 (quoting Exxon, 544 U. S., at 284). But it ignores that Exxon identifies the defining features of those cases and specifically notes that “both” had been filed “after the state proceedings ended.” 544 U. S., at 291; see Skinner v. Switzer, 562 U. S. 521, 531 (2011) (reiterating that Rooker and Feldman were filed “after the state proceedings ended”). Almost every circuit has recognized the significance of this language. See n. 1, supra. This Court dismisses it in a footnote. Ante, at 8, n. 5.

To justify giving Rooker-Feldman a broader sweep, the Court resurrects a rationale that Exxon buried: §1331. Before Exxon, we had cited §1331 as a secondary rationale for the doctrine, see ante, at 3; according to the Court, Exxon continues the thread, see ante, at 11. That is a very ambitious reading of Exxon. Rather than describing Rooker-Feldman as a negative inference from §1331, Exxon portrays it as a negative inference from §1257 that overrides §1331. As Exxon summarizes: Rooker-Feldman describes when this Court’s appellate jurisdiction “precludes” district courts from exercising the jurisdiction they would “otherwise” have. 544 U. S., at 291. The Court protests that if Exxon had intended to discard the §1331 rationale, it would have said so. See ante, at 12. But the Court does not even mention this language—Exxon’s only explanation of the doctrine’s rationale—much less explain why it does not suffice. And our more recent Rooker-Feldman cases confirm that Exxon tossed the §1331 justification. These cases present Rooker-Feldman as a “strict negative inference from §1257,” ante, at 13, and do not mention §1331 at all. See Skinner, 562 U. S., at 531–532 (explaining Rooker-Feldman solely as an inference from §1257); Lance v. Dennis, 546 U. S. 459, 463 (2006) (per curiam) (same).

Yet even as it resuscitates the §1331 rationale, the Court acknowledges that §1257 remains in the mix. Ante, at 6 (describing Rooker-Feldman as “built” on both §1331 and §1257). And by recognizing that the doctrine applies only to “state-court judgments,” ante, at 1 (emphasis added), the Court tacitly admits that §1257—not §1331—is what really determines when Rooker-Feldman kicks in. Still, the Court’s embrace of §1257 is only partial: It is unwilling to accept §1257’s requirement of a “[f]inal judgment” from the “highest court of a State.” The Court’s main reason for disregarding this limit appears to be Exxon, but as explained, Exxon compels the opposite conclusion.

Perhaps recognizing that its new approach to §1257 requires some other rationale, the Court spends several pages running through various policy considerations. Applying §1257’s finality limit, the Court says, would generate “anomalous outcomes,” disrespect “federalism,” and sow “confusion.” Ante, at 13, 17. Its theory is an innovation, but not an improvement.

Take “anomalous outcomes.” The Court reasons that it is “difficult to see why Congress would have authorized district courts to review final state-court judgments pending appeal in state court when it did not authorize this Court to do so.” Ante, at 14. But the Court’s approach also yields “anomalous outcomes.” To give one example: As the Court concedes, the doctrine applies only to “ ‘state-court judgments’ ”; it does not bar district courts from reviewing state-court interlocutory orders. Ante, at 1 (quoting Exxon, 544 U. S., at 284). Yet it is similarly “difficult to see why Congress would have authorized district courts” to review interlocutory orders subject to further review in state court “when it did not authorize this Court to do so.” Ante, at 14. The Court’s approach cannot answer this question, but mine can: Rooker-Feldman allows for this result because it is rooted in §1257, not in a vague notion of which outcomes are “anomalous.”

The Court also insists that “[f]ederalism principles are . . . best served” by applying Rooker-Feldman to dismiss T. M.’s challenge. Ante, at 15. But the word “federalism” does not appear in either Rooker or Feldman. And if it were true that general concerns about federalism animate the doctrine, Exxon would have come out the opposite way. The lower court in Exxon applied Rooker-Feldman in part because the federal suit was “an ‘insurance policy’ against an adverse result in state court”—a clear affront to federalism. 544 U. S., at 294, n. 9 (quoting 364 F. 3d 102, 105–106 (CA3 2004)). We dismissed that line of reasoning in a footnote, explaining that there is “nothing necessarily inappropriate” about “a protective action.” 544 U. S., at 294, n. 9. If federalism drives Rooker-Feldman, the Exxon Court missed the memo.

Even if federalism were a solid rationale, it is doubtful that “[f]ederalism principles are . . . best served by” dismissing T. M.’s challenge. Ante, at 15. Like other States, Maryland has adopted preclusion rules to determine when parties may relitigate issues and claims decided in a prior case. The District Court determined that T. M.’s lawsuit is not “precluded” under Maryland law—which means that T. M. could have filed a suit just like this one in a Maryland state court. Letter Order in No. 1:23–cv–1684 (D Md.), ECF Doc. 27, pp. 2–3. Yet the Court maintains that bouncing her suit from federal court best serves federalism. Why? Wouldn’t Maryland’s interests be best served by affording its judgments the effect they would have under state law?

In any event, Congress has already made the federalism calculus, and this Court is not free to override it. As Exxon recognizes, the Full Faith and Credit Act—which the Court does not mention—requires federal courts to give state judgments the effect that they would have in state courts. 544 U. S., at 293; see Lance, 546 U. S., at 466 (“Congress has directed federal courts to look principally to state law in deciding what effect to give state-court judgments”). So even if the Court could otherwise ignore state law in the name of federalism, Congress has foreclosed that path.

Grounding Rooker-Feldman in federalism is also at odds with our abstention doctrines, which already account for friction between state and federal courts but do not apply here. We have purposefully kept these doctrines narrow, stressing that district courts have a “virtually unflagging obligation . . . to exercise the jurisdiction given them.” Colorado River, 424 U. S., at 817. “[O]nly exceptional circumstances justify a federal court’s refusal to decide a case in deference to the States.” New Orleans Public Service, Inc. v. Council of City of New Orleans, 491 U. S. 350, 368 (1989). No one argues that T. M.’s challenge fits within those “exceptional circumstances.” Nonetheless, the Court repackages the same principle—“deference to the States”—as a reason why the District Court lacked jurisdiction in the first place. See ante, at 14. It is ironic that in the Court’s view, federalism concerns that are too weak to trigger abstention doctrines are strong enough to deprive district courts of jurisdiction altogether. Cf. Exxon, 544 U. S., at 284 (outside the circumstances of Rooker and Feldman, Rooker-Feldman doctrine “does not . . . augment the circumscribed doctrines that allow federal courts to stay or dismiss proceedings in deference to state-court actions”).

The Court’s reliance on federalism also sits uneasily with the fact that T. M.’s suit arises under 42 U. S. C. §1983. Section 1983 was the “product of a vast transformation from” traditional “concepts of federalism.” Mitchum v. Foster, 407 U. S. 225, 242 (1972). Its “very purpose” was “to interpose the federal courts between the States and the people, as guardians of the people’s federal rights—to protect the people from unconstitutional action under color of state law, ‘whether that action be executive, legislative, or judicial.’ ” Ibid. (quoting Ex parte Virginia, 100 U. S. 339, 346 (1880)). Given our longstanding interpretation of §1983, it is odd to cite federalism as grounds for dismissing a §1983 suit. Cf. J. Beerman, Comments on Rooker-Feldman or Let State Law Be Our Guide, 74 Notre Dame L. Rev. 1209, 1230–1231 (1999). Notably, we have construed the Anti-Injunction Act—which normally bars federal courts from enjoining state-court proceedings—to contain an exception for §1983 suits like T. M.’s. See Mitchum, 407 U. S., at 242–243.

Finally, the Court insists that drawing the line at finality would “needlessly complicate the Rooker-Feldman analysis” because “[i]t is not always straightforward to determine whether a given state-court judgment is a reviewable final judgment under §1257(a).” Ante, at 17. But the line must be drawn somewhere, and there is no reason to think that drawing it at “judgment,” as the Court does, is any more manageable. We apply the finality test with little trouble; surely lower courts can do the same.

*  *  *

The upshot of today’s decision is that the Court has muddied waters that were hardly clear to begin with. That is unfortunate, because there was a better path available: treating Rooker-Feldman as “the §1257 Rule.” VanderKodde, 951 F. 3d, at 409 (Sutton, J., concurring). Doing so would have been both clearer and more faithful to Exxon.

Still, the news is not all bad. Although the Court expands Rooker-Feldman beyond Exxon’s line, it repeatedly emphasizes that the doctrine is “narrow.” See ante, at 1, 7, 8, 18. Courts should not lose sight of that message. In the end, Rooker-Feldman has been given an inch—it should not be allowed to take a mile.