How an Overruling Actually Happens, Step by Step

A precedent falls only when the Supreme Court, in a later case squarely presenting the question, holds the earlier rule wrong and replaces it with a different rule. No other institution performs that act, no lower court anticipates it, and no passage of time alone accomplishes it. The overruling happens inside an ordinary appellate decision, with a named case, a record, parties, and a judgment that resolves a live dispute. An overruling is a judgment, reached through a named set of factors, that a precedent is so wrong, so unworkable, or so eroded that keeping it costs the law more than the stability stare decisis exists to protect. The Court weighs error against reliance, workability, and legitimacy, and it announces the result as law for the case at hand and for future cases governed by the same question.

The procedural path begins long before argument, when a litigant files a petition for a writ of certiorari that asks the Court to overrule a precedent by name. The petition must frame the request as a question presented, tie the request to the facts of the dispute, and explain why the existing rule controls the outcome below. In Ramos v. Louisiana, 590 U.S. 83 (2020), the request to reconsider the nonunanimous jury rule of Apodaca v. Oregon, 406 U.S. 404 (1972), reached the Court because the conviction in the case at hand depended on that rule. The sequence matters because the Court decides cases, not abstract propositions, and an overruling detached from a judgment would exceed the judicial role described in Article III of the Constitution.

After the grant of certiorari, briefing turns toward the precedent itself. The party seeking the overruling briefs error in the original reasoning, intervening developments that undercut the rule, and the costs the rule imposes in practice. The party defending the precedent briefs reliance, congressional acquiescence where a statute is involved, and the institutional cost of reversal. Amicus briefs often supply the empirical record on reliance and workability, naming affected transactions, administrative systems, or trial practices with specificity. In Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), the briefing addressed Roe v. Wade, 410 U.S. 113 (1973), and Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), directly, because the parties and the Court understood that the precedent was the central issue. Oral argument then tests whether counsel will defend the precedent on its merits or only on stability grounds, a distinction that often signals how the justices perceive the strength of the original reasoning.

Conference and drafting complete the act. The justices vote on the judgment and on the treatment of the precedent, and the assigned author must assemble a majority for both. The final opinion identifies the precedent, states the holding of error, applies the governing factors, announces the replacement rule, and disposes of the case. Publication in the United States Reports gives the decision its formal authority. From that moment, the displaced rule no longer controls, the new rule governs pending and future cases within the limits of retroactivity doctrine, and the overruled decision remains in the Reports as history, as reasoning that may still persuade on unrelated points.

What Stare Decisis Means and Where the Obligation Comes From

Stare decisis comes from the Latin phrase stare decisis et non quieta movere, commonly shortened to the instruction to stand by things decided and not disturb settled points. The phrase captures a practice older than the United States, rooted in English common law courts that treated prior judgments as evidence of the law and as constraints on later judges deciding like cases in like ways. In that tradition, following precedent served consistency, predictability, and equality of treatment, because litigants in similar positions could expect similar outcomes, and judges could not remake rules according to personal preference in each case. The American system inherited the practice through colonial courts, state courts after independence, and federal courts exercising common law reasoning in statutory and constitutional interpretation. No party to the founding debates inserted a stare decisis clause into the constitutional text, and no statute commands the Supreme Court to follow its own prior decisions.

The absence of a textual command defines the doctrine. Stare decisis operates as a principle of policy and judicial method that the Court imposes on itself, describes as foundational, and departs from only for stated reasons. In Kimble v. Marvel Entertainment, LLC, 576 U.S. 446 (2015), Justice Elena Kagan, writing for the majority, described stare decisis as a foundation stone of the rule of law. The metaphor in Kimble carried a specific meaning: decided rules support planning, investment, and ordered liberty in the same way a physical foundation supports a structure built over time, and removing the stone requires more than disagreement with the design. The Court in Kimble refused to overrule a patent precedent in part because Congress, not the Court, held the primary power to revise the underlying statute and the policy balance it embodied.

Justice Louis D. Brandeis gave the classic statement of the settlement value in Burnet v. Coronado Oil and Gas Co., 285 U.S. 393 (1932), in dissent. Brandeis argued that in many fields settlement matters more than correctness, because participants order affairs around a known rule, and a stable rule that is imperfect can serve the law better than repeated revision in search of a better rule. The Burnet dissent also drew the enduring distinction that shapes later doctrine: where correction through legislation remains available, stability deserves great weight, while errors in constitutional interpretation, which legislation cannot correct, stand in a different position.

Because the obligation comes from practice, precedent, and institutional reasoning rather than from a superior text, the Court retains authority to define its strength. The doctrine binds with different force across fields, across types of reliance, and across the age and reasoning quality of the precedent under challenge. A durable constitutional precedent shows the obligation in its strongest civic form: the citizenship rule settled in United States v. Wong Kim Ark, 169 U.S. 649 (1898), has governed birth in the United States for more than a century, and its operation is traced in the account of how birthright citizenship works. Wong Kim Ark endured because later generations treated the question as decided, built administrative practice around the holding, and litigated new disputes at the margins rather than reopening the core rule. That endurance illustrates the source of the obligation in action: each later decision to follow, distinguish, or apply the rule renewed the commitment, and the accumulation of reliance raised the cost of reversal far beyond the cost present on the day the case was decided.

The Difference Between Following, Distinguishing, Narrowing and Overruling

Following a precedent means applying its rule to a later case because the material facts and the legal question align closely enough that the earlier holding controls the later outcome. A lower court follows Supreme Court precedent as a matter of hierarchy, and the Supreme Court follows its own precedent as a matter of stare decisis unless a reason to depart emerges. In Dickerson v. United States, 530 U.S. 428 (2000), the Court followed Miranda v. Arizona, 384 U.S. 436 (1966), and refused to treat a later federal statute as displacing the constitutional warnings requirement. The act of following reaffirms the rule, extends its life, and often clarifies ambiguities at the edges without changing the core holding.

Distinguishing a precedent means holding that the earlier rule does not control because a material fact or legal element differs in a way the earlier reasoning makes relevant. The later court accepts the precedent as correct on its own facts and leaves it intact, while explaining why the dispute at hand falls outside its scope. In Moore v. Harper, 600 U.S. 1 (2023), distinctions among state legislative, executive, and judicial roles mattered because the Elections Clause analysis turned on which state actor had acted and under what constraint. A sound distinction identifies the fact that mattered to the earlier majority, shows its absence or difference, and explains why that difference changes the result under the precedent’s own logic. A distinction that rests on an immaterial detail functions as evasion, and later courts often treat strained distinctions as signals that the precedent has lost support even before any formal overruling occurs.

Narrowing a precedent means confining its rule to a smaller domain than earlier language or later applications had suggested, while stopping short of declaring the case wrongly decided. A court narrows by reading a broad statement as tied to the facts presented, by limiting a rule to the statutory provision construed in the original case, or by holding that a later doctrinal development has confined the precedent to a specific context. For example, a holding announced in sweeping language about agency authority can be narrowed in a later case to the particular delegation scheme, record, and remedy present in the original dispute, so the rule survives for its core application and loses force beyond it. The method carries risk, because repeated narrowing can leave a precedent that exists in name but governs little, a condition the Court in Casey described through the separate idea of doctrinal erosion.

Overruling differs from each operation in candor and in effect. An overruling declares that the earlier decision was wrong when decided or has become wrong as law through intervening developments, withdraws its authority as precedent, and substitutes a governing rule. In Brown v. Board of Education, 347 U.S. 483 (1954), the Court held that separate educational facilities are inherently unequal, a holding incompatible with the reasoning of Plessy v. Ferguson, 163 U.S. 537 (1896), in the field of public education. The distinction among the four operations turns on what the later court says about the earlier rule: following extends it, distinguishing places the new case outside it, narrowing contracts it, and overruling removes it. Litigants, judges, and administrators depend on that clarity, because each operation carries different consequences for pending cases, for reliance interests built under the old rule, and for the legitimacy of a court that claims to decide according to law rather than preference.

The Elements of an Overruling Decision

An overruling decision must first identify the precedent with precision. The opinion names the case, states the proposition for which the case stands, and separates the holding from dicta, concurrences, and later glosses that may have expanded or contracted the original rule. In Janus v. American Federation of State, County, and Municipal Employees, Council 31, 585 U.S. 878 (2018), the majority identified Abood v. Detroit Board of Education, 431 U.S. 209 (1977), and the specific proposition that public sector agency fees could be required of nonmembers. That identification framed every later step, because the factors on error, workability, and reliance applied to the agency fee proposition, not to every statement in the Abood opinion or to union law as a whole.

The opinion must next hold that the precedent was in error, and it must give reasons that engage the original reasoning rather than merely announcing disagreement among the justices who happen to sit when the challenge arrives. Error can rest on flawed textual analysis, mistaken history, misunderstanding of a prior doctrine, or a factual premise disproved by the record developed in later litigation. In Citizens United v. Federal Election Commission, 558 U.S. 310 (2010), the majority held that Austin v. Michigan Chamber of Commerce, 494 U.S. 652 (1990), rested on a rationale inconsistent with First Amendment principles governing political speech. Whether later readers accept that assessment or reject it, the form matters: a reasoned holding of error connects the overruling to legal materials outside the preferences of the deciding majority and gives lower courts a basis for applying the new rule.

Can the Supreme Court overrule itself without a constitutional amendment?

Yes. The Supreme Court needs no amendment to overrule its own constitutional interpretations, because stare decisis is a judicial policy rather than a constitutional command. An amendment remains the route for outsiders to reverse the Court, as the Eleventh, Fourteenth, Sixteenth and Twenty-Sixth Amendments did in part against earlier rulings.

The opinion must then announce the replacement rule with enough definition to decide the case at hand and to guide courts facing the next case. A replacement can restore an older rule displaced by the overruled precedent, adopt a test drawn from a concurrence or dissent that had argued for a different path, or construct a new standard from text, history, and structure. In West Coast Hotel Co. v. Parrish, 300 U.S. 379 (1937), displacement of the liberty of contract line left a more deferential review of economic regulation as the governing approach. A vague replacement invites fragmentation, because lower courts fill gaps in divergent ways and the Court must grant review again to supply the missing definition.

Disposition completes the elements. The opinion applies the new rule to the parties, enters judgment, and addresses transition questions that the judgment itself raises, including whether related holdings survive and how severability or remedy doctrines affect the result. The disposition proves that the overruling was necessary to the decision rather than an advisory essay on a disfavored precedent. Where an overruling reaches beyond the facts needed to resolve the dispute, dissenting opinions often raise that breadth as a legitimacy objection, as Justice Stephen Breyer did in dissent in Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), in joint dissent with Justice Sonia Sotomayor and Justice Elena Kagan. The four elements together, identification, error, replacement, and disposition, convert a change of mind into an act of law.

Who May Overrule a Supreme Court Precedent, and Who May Not

Only the Supreme Court may overrule a Supreme Court precedent. That exclusive authority follows from hierarchy: a precedent of the highest court binds every other court in the federal system and binds state courts on questions of federal law, and a bound court lacks power to declare the binding rule void. The rule protects uniformity, because federal law would fragment if each district court, circuit court, or state court could retire a Supreme Court holding it considered mistaken. It also protects candor, because lower courts can criticize a precedent, explain its costs, and invite reconsideration while still applying the rule, leaving the formal act of removal to the institution that created the precedent and bears responsibility for the system of precedent as a whole.

The Court stated the lower court rule with unusual directness in Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U.S. 477 (1989). The majority instructed lower courts to follow a directly controlling precedent even when the precedent appeared to rest on reasoning rejected in another line of decisions, and to leave to the Supreme Court the prerogative of overruling its own decisions. In Agostini v. Felton, 521 U.S. 203 (1997), the Court reaffirmed that instruction while itself reconsidering Aguilar v. Felton, 473 U.S. 402 (1985). Agostini matters for two reasons: the majority applied the Rodriguez de Quijas discipline to the lower courts in the litigation history, and the majority then exercised the exclusive power that the discipline reserves, holding that intervening doctrine had eroded Aguilar. The pairing shows the system working as designed, restraint below and responsibility at the top.

State courts stand in the same position on federal questions. A state supreme court may interpret its own constitution to provide greater protection than federal law requires, and it may decide state statutory questions without Supreme Court control, but on the meaning of the Federal Constitution or a federal statute, a Supreme Court holding binds. In Martin v. Hunter’s Lessee, 14 U.S. 304 (1816), Justice Joseph Story defended Supreme Court appellate review of state court judgments on federal questions as essential to uniformity. The principle from Martin extends naturally to precedent: a state court that refused to follow a controlling federal holding would substitute a local interpretation for a national rule in a domain where the Supremacy Clause of Article VI requires the opposite result.

Congress holds broad power and still may not overrule a constitutional interpretation by ordinary legislation. In City of Boerne v. Flores, 521 U.S. 507 (1997), the Court held that Congress, acting under Section 5 of the Fourteenth Amendment, could enforce the amendment through remedial legislation but could not redefine the substantive right the Court had recognized in Employment Division v. Smith, 494 U.S. 872 (1990). Boerne drew the line between enforcement and revision: Congress can create remedies, gather facts, and regulate conduct within constitutional bounds, while the meaning of the constitutional provision, as interpreted by the Court, remains subject to change through the Court itself or through constitutional amendment under Article V. For statutory precedents the allocation differs, because Congress can amend the statute and thereby displace the Court’s reading through ordinary lawmaking, a power that strengthens stare decisis in statutory cases and has no parallel for constitutional holdings.

How a Challenge Reaches the Court: Vehicle, Standing and the Question Presented

A challenge reaches the Court through a vehicle, meaning a concrete case with standing, a final or reviewable judgment below, an adequate record, and a question presented that squarely raises the precedent for reconsideration. Standing requires injury, causation, and redressability under the doctrine summarized in Lujan v. Defenders of Wildlife, 504 U.S. 555 (1992), and a challenge to a precedent gains no exemption from those requirements because the legal issue appears important. A plaintiff who suffers no injury from the precedent’s rule cannot manufacture review by asking for a better rule in the abstract. The vehicle requirement disciplines timing as well as access: the Court waits for a dispute in which the precedent actually determines rights, duties, money, liberty, or governmental authority, so the consequences of overruling remain visible in facts rather than speculation.

Litigants who seek an overruling must ask expressly. A brief that argues for a favorable distinction while hinting at broader doubts rarely forces the question, because the Court can grant the distinction and leave the precedent untouched. An express request names the case to be overruled, states the proposition to be abandoned, and proposes the replacement rule, giving the opposing party fair notice and giving the justices a clean choice at conference on whether to grant review on that ground. In South Dakota v. Wayfair, Inc., 585 U.S. 162 (2018), the petition and briefing placed Quill Corp. v. North Dakota, 504 U.S. 298 (1992), directly before the Court, and the majority opinion in Wayfair overruled Quill and National Bellas Hess, Inc. v. Department of Revenue of Illinois, 386 U.S. 753 (1967), after confronting the request directly.

The Court sometimes orders reargument on the precedent itself when initial briefing or argument shows that the case cannot be decided responsibly without confronting a precedent the parties treated as background. Citizens United v. Federal Election Commission, 558 U.S. 310 (2010), supplies the leading modern example: after initial argument on a narrower statutory and as-applied path, the Court ordered reargument and supplemental briefing on whether Austin v. Michigan Chamber of Commerce, 494 U.S. 652 (1990), and part of McConnell v. Federal Election Commission, 540 U.S. 93 (2003), should be overruled. The reargument order in Citizens United expanded the briefing, delayed decision into the following term, and drew criticism from Justice John Paul Stevens in dissent about the breadth of the path chosen. The episode shows both the power and the cost of Court-initiated reconsideration, wider deliberation purchased with wider controversy about judicial reach.

A poor vehicle can delay an overruling for years. Defects that cause delay include mootness, an independent state ground, waiver or forfeiture of the key argument, an undeveloped factual record on reliance or workability, and a requested ruling broader than the dispute requires. In June Medical Services LLC v. Russo, 591 U.S. 299 (2020), the controlling opinions debated the force of Casey and Whole Woman’s Health v. Hellerstedt, 579 U.S. 582 (2016), without producing a majority overruling, and the underlying question returned in Dobbs two years later in a vehicle that raised it without the same complications. Delay of that kind frustrates advocates on both sides, preserves a precedent that may lack majority support on the merits, and confirms a central practical truth: doctrine changes through cases, and the quality of the case determines whether the doctrine can change at all in a given term.

The Burden a Challenger Carries: Special Justification

A challenger carries a burden beyond persuasion on the merits. In Payne v. Tennessee, 501 U.S. 808 (1991), Chief Justice William Rehnquist, writing for the majority, stated that stare decisis does not require adherence to a rule shown to be badly reasoned or unworkable, while also treating departure as a decision requiring justification beyond simple disagreement. Later opinions condensed that justification requirement into the phrase special justification. The phrase appears across majority, plurality, and dissenting opinions as shorthand for a showing that the precedent’s costs, defects, or erosion exceed the normal institutional reasons for stability. Payne itself overruled Booth v. Maryland, 482 U.S. 496 (1987), and South Carolina v. Gathers, 490 U.S. 805 (1989), on victim impact evidence in capital sentencing, on the ground that those decisions had misread precedent and had proved difficult to apply in a principled way.

The origin of special justification lies in a tension the Court has never resolved by rule alone. If mere error justified overruling, every precedent would stand only until five justices preferred a different answer, and law would track membership changes rather than legal reasoning. If error never justified overruling, the Court would entrench its worst mistakes and transfer correction to the amendment process, which Article V makes deliberately difficult. Special justification mediates between those failures by demanding a reason that a neutral observer could recognize as legal rather than purely political: demonstrable error in text or history, a rule that trial courts cannot administer consistently, intervening doctrine that has removed the precedent’s foundations, factual premises that later records have disproved, or reliance interests so weak that stability protects little beyond the precedent itself.

Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), gave the burden its most elaborate application in the act of refusing to overrule. The joint opinion of Justice Sandra Day O’Connor, Justice Anthony Kennedy, and Justice David Souter retained the core holding of Roe v. Wade, 410 U.S. 113 (1973), while revising its framework, and reasoned that overruling under fire without a justification beyond disagreement would damage the Court’s legitimacy. Casey treated reliance, workability, doctrinal stability, and factual change as inquiries that a challenger must win in combination sufficient to carry the special burden. Critics of Casey argued that the joint opinion converted legitimacy into a reason to preserve error, while defenders argued that Casey stated the burden honestly: a precedent on a divisive question gathers institutional stakes that a challenger must address rather than ignore.

Dickerson v. United States, 530 U.S. 428 (2000), shows the burden defeating a challenge supported by strong claims about original meaning and federal power. Chief Justice William Rehnquist, writing for the majority in Dickerson, held that Miranda v. Arizona, 384 U.S. 436 (1966), announced a constitutional rule Congress could not displace by statute, and the opinion relied in part on stare decisis, noting that Miranda had become embedded in police practice and had entered national culture as a safeguard familiar to officers and courts. The Dickerson majority found no special justification adequate to overrule a precedent that had proved workable and on which law enforcement had built training and procedure. Together, Payne, Casey, and Dickerson map the burden from both directions: Payne defines when the burden can be carried, Casey explains why legitimacy raises it, and Dickerson shows workable reliance defeating it even when the author of the majority had long questioned the precedent under challenge.

Constitutional Precedents and the Amendment Problem

Constitutional precedents occupy a distinct position because no ordinary legislative correction stands behind them. Justice Louis D. Brandeis made the point in Burnet v. Coronado Oil and Gas Co., 285 U.S. 393 (1932), in dissent, reasoning that the Court should hesitate before unsettling statutory interpretations Congress can revise, while correction of constitutional error belongs primarily to the Court itself because the amendment process offers no practical remedy in most instances. The Burnet framework does not treat constitutional precedents as weak in every respect. It treats the source of correction as different: a mistaken statute reading can be repaired by a majority in Congress and a presidential signature or veto override, while a mistaken constitutional holding can be displaced only by the Court or by a supermajority process that has succeeded only a small number of times in more than two centuries.

Payne v. Tennessee, 501 U.S. 808 (1991), converted the Brandeis insight into majority doctrine by stating that stare decisis stands at its weakest in constitutional cases. The Payne majority reasoned that when governing decisions are unworkable or badly reasoned, the Court has a responsibility to correct error in constitutional interpretation because no other institution can perform the correction through ordinary politics. Later majorities repeated the formulation when overruling constitutional precedents and qualified it when defending them, so the sentence functions as both license and warning. The license permits reconsideration of holdings that would otherwise become permanent through inertia. The warning reminds readers that weaker stare decisis still means some stare decisis, and a challenger must still carry the special justification burden described in Payne and Casey.

The amendment record proves the difficulty Brandeis and Payne identified. The Eleventh Amendment reversed the result in Chisholm v. Georgia, 2 U.S. 419 (1793), on suits against states by citizens of other states. The Fourteenth Amendment displaced Dred Scott v. Sandford, 60 U.S. 393 (1857), in part, by establishing birthright citizenship and federal power to protect civil rights against state denial. The Sixteenth Amendment answered Pollock v. Farmers’ Loan and Trust Co., 158 U.S. 601 (1895), by authorizing a federal income tax without apportionment among the states. The Twenty-Sixth Amendment reversed Oregon v. Mitchell, 400 U.S. 112 (1970), in part, by setting 18 years as the voting age floor for state as well as federal elections. Four episodes across the full span of constitutional history confirm that amendment functions as an extraordinary remedy, not a routine check.

That rarity shapes judicial behavior in predictable ways. Knowing that an entrenched constitutional error may govern for generations raises the cost of the original decision and increases pressure on later courts to correct serious mistakes despite reliance and legitimacy concerns. At the same time, the difficulty of amendment gives constitutional precedents a practical permanence that can deepen reliance by states, Congress, and private actors who build institutions on the assumption that the holding will endure. In Garcia v. San Antonio Metropolitan Transit Authority, 469 U.S. 528 (1985), the Court overruled National League of Cities v. Usery, 426 U.S. 833 (1976), on federal power over state employment practices, after concluding that the earlier line had proved unworkable in application. Garcia illustrates the amendment problem in action: no amendment stood ready to clarify state immunity from federal labor standards, so the choice lay between judicial correction and indefinite continuation of a rule the majority considered unsound and unmanageable.

Statutory Precedents and the Super Strong Presumption

Statutory precedents receive the strongest protection stare decisis offers. In Kimble v. Marvel Entertainment, LLC, 576 U.S. 446 (2015), Justice Elena Kagan, writing for the majority, refused to overrule Brulotte v. Thys Co., 379 U.S. 29 (1964), which barred patent royalties for use after patent expiration. The Kimble majority acknowledged sharp criticism of Brulotte on economic grounds and still retained the rule, because Congress had repeatedly considered the patent statute without displacing Brulotte, because parties had contracted in reliance on the known rule, and because any revision of patent policy belonged to the legislative branch with its superior capacity to gather facts and balance competing interests. Kimble turned the statutory presumption into a demonstration: even a precedent many specialists consider poorly reasoned can survive when the correction mechanism lies in competent hands outside the Court.

Patterson v. McLean Credit Union, 491 U.S. 164 (1989), supplied an earlier and equally influential statement. The Patterson majority refused to overrule Runyon v. McCrary, 427 U.S. 160 (1976), which read 42 U.S.C. Section 1981 to reach private discrimination in contracting, and the opinion stressed that Congress could amend the statute if the interpretation misread legislative intent. Congress later enacted the Civil Rights Act of 1991, revising parts of Patterson on other points while leaving the Runyon holding on Section 1981 in place, a sequence that vindicated the Patterson logic that legislative response remains possible and sometimes occurs. The example matters because statutory stare decisis rests on an empirical premise as well as a separation of powers premise: legislators monitor judicial interpretations, interest groups petition for change, and committees hold hearings that a court cannot replicate through briefing in a single case.

The underlying logic assigns error correction according to institutional competence. When the Court misreads a statute, the text remains available for Congress to amend, the legislative history of inaction or revision informs later interpretation, and private reliance on contracts, licenses, tax planning, and compliance systems counsels against judicial disruption that regulated parties cannot anticipate through the political process. Justice Samuel Alito, dissenting in Kimble, argued that the majority gave excessive weight to a precedent with weak foundations, and that disagreement frames the boundary dispute inside statutory stare decisis: defenders stress correction by Congress and reliance, while critics stress that congressional inaction proves little about intent and that courts should not preserve a rule the deciding court now considers indefensible. Later cases described the statutory presumption as super strong, language that signals a challenger must show far more than error, including changed statutory context, conflict with later enactments, or a rule that has become a positive obstacle to the scheme Congress enacted.

Exceptions exist and prove the strength of the presumption by their rarity and reasoning. In Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U.S. 877 (2007), the Court overruled Dr. Miles Medical Co. v. John D. Park and Sons Co., 220 U.S. 373 (1911), on resale price maintenance under the Sherman Act, treating the antitrust statute as a common law statute in which Congress expected judicial development. The Leegin distinction confines easier overruling to statutes that delegate lawmaking authority to courts in common law form. For ordinary statutes with fixed texts and active legislative oversight, Kimble and Patterson govern: stability prevails, Congress holds the correction power, and the Court demands a justification that goes well beyond the claim that the earlier majority read the words incorrectly.

Worked Example: The Life Cycle of One Precedent From Decision to Displacement

The Lochner line offers a complete life cycle in outline, from decision through following, erosion, and displacement, with each stage showing a distinct mechanism. Lochner v. New York, 198 U.S. 45 (1905), held that a New York law limiting bakery employees to 60 hours per week violated liberty of contract protected by the Due Process Clause of the Fourteenth Amendment. Justice Rufus Peckham, writing for the majority in Lochner, treated freedom to buy and sell labor as a constitutional liberty that legislation could restrict only for legitimate health or safety purposes, and the majority found the bakery law an unjustified interference with that liberty. Justice John Marshall Harlan, in dissent, and Justice Oliver Wendell Holmes Jr., in dissent, disputed the majority premise, with Holmes arguing in Lochner that the Constitution does not enact an economic theory.

The following stage extended the rule beyond bakeries. In Adair v. United States, 208 U.S. 161 (1908), and Coppage v. Kansas, 236 U.S. 1 (1915), the Court applied liberty of contract reasoning to strike down laws protecting union membership, and in Adkins v. Children’s Hospital, 261 U.S. 525 (1923), the Court invalidated a minimum wage law for women in the District of Columbia. Each application reaffirmed the Lochner premise, broadened its domain, and generated new reliance among employers and new opposition among legislators seeking to regulate hours, wages, and working conditions. Following also exposed the line-drawing problem that would fuel erosion: courts had to decide which regulations counted as valid health measures and which counted as impermissible interference, and the distinctions in cases on mines, factories, and public health often turned on judgments about economic policy that the doctrine claimed to exclude from judicial discretion.

Erosion arrived through exceptions, distinctions, and shifting majorities rather than through a single repudiation. In Muller v. Oregon, 208 U.S. 412 (1908), the Court upheld an hours limit for women, and in Bunting v. Oregon, 243 U.S. 426 (1917), the Court upheld an hours law for mill workers, decisions that coexisted uneasily with Lochner by emphasizing factual records and health justifications. The factual brief filed by attorney Louis D. Brandeis in Muller, presenting social and medical evidence on working hours, demonstrated a method for sustaining regulation inside the existing framework while weakening its abstract premise. By the middle 1930s, the Lochner rule survived as a citation while its foundations, confidence in judicial review of economic legislation and skepticism toward legislative fact finding, had lost majority support.

Displacement came in West Coast Hotel Co. v. Parrish, 300 U.S. 379 (1937), which upheld a Washington minimum wage law for women and displaced the reasoning of Adkins and, in substance, the Lochner approach to economic due process. Chief Justice Charles Evans Hughes, writing for the majority in West Coast Hotel, treated regulation reasonable in relation to its subject and adopted in the interests of the community as consistent with due process, language incompatible with the Lochner presumption against wage and hour legislation. The sequence closes the life cycle: a rule announced in 1905, extended by following decisions, contracted by distinctions and factual exceptions, weakened by criticism and changed conditions, and finally displaced by a later holding that supplied a replacement approach. The detailed treatment of West Coast Hotel, its vote, and its consequences belongs to later analysis; the value of the sequence here lies in the mechanics, showing that overruling usually ends a long process rather than beginning one.

The Casey Settlement: Four Factors and a Legitimacy Argument

Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), supplied the vocabulary that later overruling opinions still borrow, contest, and rearrange. The controlling plurality opinion, written jointly by Justices Sandra Day O’Connor, Anthony Kennedy, and David Souter in 1992, faced a direct invitation to overrule Roe v. Wade, 410 U.S. 113 (1973), and declined the invitation while discarding a central part of the Roe framework. The plurality preserved what the opinion called the core holding of Roe, the recognition of a right to terminate a pregnancy before viability, and struck down the trimester framework that had structured Roe for 19 years. In place of the trimester framework, the plurality in Casey (1992) substituted the viability line and the undue burden standard for regulations of pre-viability abortion.

The plurality in Casey (1992) organized the stare decisis inquiry around four considerations, and later majorities have treated the list as the starting point for the analysis. The first consideration asked whether the earlier rule had proven unworkable in practice, meaning whether courts could apply the rule with consistency and predictability. The second consideration asked whether removal of the rule would cause reliance interests of a kind that counseled against change, meaning whether persons and institutions had ordered primary conduct in dependence on the rule. The third consideration asked whether related principles of law had developed in a way that left the earlier rule no more than a remnant of abandoned doctrine. The fourth consideration asked whether facts had changed, or had come to be understood so differently, that the earlier rule had lost significant application or justification.

Separate from the four considerations, the plurality in Casey (1992) advanced a legitimacy argument that has no exact parallel in earlier stare decisis opinions. The plurality reasoned that a decision overruling Roe under sustained political pressure, without a ground stronger than disagreement with the original outcome, would subvert the legitimacy of the Supreme Court itself. In the formulation of the plurality in Casey (1992), the promise of constancy, once given, bound the Court even when a majority doubted the original decision as an original matter. Chief Justice William Rehnquist, dissenting in Casey (1992), answered that the legitimacy of the Court rested on the correctness and persuasiveness of its reasoning rather than on adherence to a decision the dissent regarded as erroneous. Justice Antonin Scalia, dissenting separately in Casey (1992), rejected the plurality reliance analysis and argued that the Court damaged its authority by retaining a rule for institutional reasons.

The disposition in Casey (1992) showed how the four considerations operated when they pulled in different directions. On workability, the plurality concluded that the trimester framework had become rigid and had prevented the Commonwealth of Pennsylvania from expressing respect for potential life throughout pregnancy, a defect the plurality cured by substitution rather than by wholesale overruling. On reliance, the plurality concluded that women in the United States had organized intimate relationships and economic participation in reliance on the availability of abortion, a societal reliance the plurality treated as real even though the reliance resisted precise measurement. On changed facts, the plurality pointed to advances in medical technology that had moved viability earlier and improved the safety of later procedures, changes the plurality absorbed by redrawing the line rather than by abandoning the holding. The settlement in Casey (1992) thus endured for 30 years.

When Error Alone Is Not Enough: The Quality of Reasoning Factor

A consistent line in Supreme Court practice holds that a demonstration of error, standing alone, does not justify the overruling of a precedent, and the quality of reasoning factor gives that line its analytical form. The factor asks an overruling majority to grade the earlier opinion as an opinion, examining method, textual grounding, historical support, and structural coherence, rather than merely announcing disagreement with the result. Justice William Brennan, writing for the majority in Payne v. Tennessee, 501 U.S. 808 (1991), stated that stare decisis carried the least force in constitutional cases, yet even Payne (1991) required a special justification beyond the belief that the earlier case was wrongly decided. An opinion that misread the governing text, ignored controlling history, or produced a rule without a principled stopping point presents a stronger candidate for overruling than an opinion that chose between two defensible readings of an ambiguous provision.

Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), illustrates the rhetoric of the quality factor at full strength. Justice Samuel Alito, writing for the majority in Dobbs (2022), described Roe v. Wade, 410 U.S. 113 (1973), as egregiously wrong from the start, and the majority rested that characterization on an extended examination of text, history, and precedent. The majority in Dobbs (2022) argued that the Constitution of the United States made no reference to abortion, that the history of criminal regulation at the time of the Fourteenth Amendment cut against an unenumerated right, and that the reasoning in Roe (1973) and in Planned Parenthood v. Casey, 505 U.S. 833 (1992), lacked a firm grounding in constitutional text or structure. The majority in Dobbs (2022) also criticized the workability of the undue burden standard and the absence of reliance interests comparable to property and contract interests. Quality of reasoning, in that architecture, functioned as the gate through which the other factors became relevant.

Dickerson v. United States, 530 U.S. 428 (2000), presents the contrasting outcome, in which scepticism about reasoning coexisted with retention of the precedent. Chief Justice William Rehnquist, writing for the majority in Dickerson (2000), preserved Miranda v. Arizona, 384 U.S. 436 (1966), against a federal statute, 18 U.S.C. Section 3501, that had purported to restore voluntariness as the sole test for the admissibility of a confession. The majority in Dickerson (2000) did not offer an enthusiastic defense of the reasoning in Miranda (1966), and the majority treated Miranda warnings as a constitutional rule embedded in police practice rather than as a proposition freshly vindicated on the merits. The majority in Dickerson (2000) emphasized that Miranda (1966) had become part of national culture and routine law enforcement practice, and that the rule remained workable for officers and courts. In Dobbs (2022), a finding of grievous methodological failure opened the inquiry and dominated the conclusion. In Dickerson (2000), doubts about method yielded to workability and reliance, and the precedent survived.

Measured across cases, the quality factor operates by comparison against text, history, and structure, the three anchors the majority in Dobbs (2022) invoked by name. A precedent vulnerable on all three anchors presents a different stare decisis problem from a precedent vulnerable on one anchor alone, and overruling opinions since Casey (1992) have increasingly itemized the anchors separately. The discipline matters because the factor otherwise collapses into a restatement of the merits disagreement that stare decisis exists to restrain.

Workability: The Factor That Asks Whether Courts Can Apply the Rule

Workability asks a practical question that the merits inquiry does not ask, namely whether lower courts, agencies, and regulated parties can apply the governing rule and reach consistent outcomes on comparable facts. The plurality in Planned Parenthood v. Casey, 505 U.S. 833 (1992), named unworkability as a ground for reconsideration, and the majority in Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), returned to the ground at length in assessing the undue burden standard. Justice Samuel Alito, writing for the majority in Dobbs (2022), argued that the undue burden standard, and in particular the large fraction formulation drawn from Casey (1992), required judges to weigh incommensurable values without a determinate method. The majority in Dobbs (2022) pointed to divided outcomes in the United States Courts of Appeals on parental notification, waiting period, and hospital admitting privilege requirements as evidence that the standard produced contradictory results on similar records.

Other doctrines supply clearer instances of the same failure mode, and overruling opinions collect the instances as precedent for the method. Justice Stephen Breyer, dissenting in Montejo v. Louisiana, 556 U.S. 778 (2009), defended the rule of Michigan v. Jackson, 475 U.S. 625 (1986), as workable, while Justice Antonin Scalia, writing for the majority in Montejo (2009), overruled Michigan v. Jackson (1986) in part on the ground that the rule generated confusion about waiver and invocation after a defendant had received counsel. The majority in Montejo (2009) emphasized that police, prosecutors, and trial judges received inconsistent guidance from the Courts of Appeals about the circumstances in which a represented defendant could initiate contact. The comparison with the balancing test of Barker v. Wingo, 407 U.S. 514 (1972), which survived repeated criticism, shows that workability criticism requires documented divergence in application, not merely theoretical complexity.

How many times has the Supreme Court overruled itself?

Counts vary by method, because counters disagree about express overrulings, partial displacements, and statutory cases. The Congressional Research Service, in the Constitution Annotated, tracks well over 100 express overrulings of constitutional decisions, and tallies rise when statutory overrulings and partial displacements enter the count, so any single figure depends on the category chosen.

A reader who keeps a separate tally of the counted decisions on paper beside a reading reference soon discovers that the denominator shifts with the definition, and the shift carries a doctrinal lesson. A narrow count limited to express overrulings of constitutional holdings produces one total, while a broader count that includes decisions disapproved in part, limited to their facts, or superseded by later reasoning produces a materially larger total. The Congressional Research Service, in the Constitution Annotated as revised in 2023, maintains the underlying tables by case name and year, and the tables reward inspection because borderline entries reveal the judgment calls that any headline number conceals. Workability analysis benefits from the same discipline. A court that calls a rule unworkable should identify the divided decisions, name the courts that divided, and show that the division tracks the rule rather than the underlying facts. The majority in Dobbs (2022) attempted that showing for the undue burden standard, and the dissent by Justices Stephen Breyer, Sonia Sotomayor, and Elena Kagan in Dobbs (2022) contested the showing by pointing to decades of lower court administration of Casey (1992).

The Factor Table That Settles the Test at a Glance

The table below settles which five questions a modern overruling opinion must answer before a majority discards an earlier holding.

Factor	What the Court asks	Where the factor decided the outcome
Quality of reasoning	Was the earlier decision wrong in method as well as result	Dobbs on Roe in 2022
Workability	Can lower courts apply the rule without contradictory results	Casey on the trimester framework in 1992
Doctrinal erosion	Have later decisions removed the rule’s foundations	Lawrence on Bowers in 2003
Reliance interests	Who ordered their affairs on the rule and at what cost	Dickerson on Miranda in 2000
Nature of the error	Does the error distort other areas of law	Loper Bright on Chevron in 2024

Read as a map rather than as a checklist, the table reveals an uneven distribution of decisive weight across the five factors, and the unevenness explains much of the predictive difficulty in stare decisis litigation. Quality of reasoning tends to carry the decision when the challenged precedent rests on a methodology that a later majority has abandoned across a field, as the majority in Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), treated the substantive due process reasoning of Roe v. Wade, 410 U.S. 113 (1973). Workability tends to carry the decision when the challenged rule governs high volume trial court administration, because contradictory outcomes accumulate quickly and visibly in reported decisions. Doctrinal erosion tends to carry the decision when intervening cases have already decided the contested question in all but name, leaving the overruling opinion to acknowledge a result that lower courts have anticipated. Reliance interests tend to carry the decision in the opposite direction, preserving precedents in property, contract, and criminal procedure administration, as the majority in Dickerson v. United States, 530 U.S. 428 (2000), preserved Miranda v. Arizona, 384 U.S. 436 (1966), against a congressional override attempt. Nature of the error tends to carry the decision when the challenged rule reallocates institutional authority, as the majority in Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), treated the deference rule of Chevron U.S.A. Inc. v. Natural Resources Defense Council, 467 U.S. 837 (1984), as a distortion of the judicial role under the Administrative Procedure Act of 1946.

Litigators who brief stare decisis after Dobbs (2022) and after Loper Bright (2024) allocate pages in proportion to that distribution, devoting the longest sections to reasoning quality and to the character of the alleged error, and treating workability and erosion as supporting grounds. Judges who write separately make the same allocation visible from the other side. Justice Elena Kagan, dissenting in Loper Bright (2024), devoted extended analysis to reliance and to the stability of administrative law, arguing that Congress had legislated against the background of Chevron (1984) for 40 years. Chief Justice John Roberts, writing for the majority in Loper Bright (2024), answered by emphasizing the quality and institutional character inquiries, arguing that Chevron (1984) had required courts to defer to agency interpretations even when a court judged another reading superior. A precedent survives when the preservation factors align, falls when the overruling factors align, and produces the hardest cases when the alignment splits, with reasoning quality on one side and reliance on the other.

Doctrinal Erosion: How Later Cases Hollow Out a Precedent Before It Falls

Doctrinal erosion describes the interval between the weakening of a precedent and the formal overruling of the precedent, an interval in which later decisions distinguish, narrow, and qualify the earlier rule until the rule governs a shrinking set of cases. The metaphor of death by a thousand distinctions, used by commentators on the common law method including Justice Benjamin Cardozo in The Nature of the Judicial Process (1921), captures the mechanism without overstating the drama. A court unwilling to overrule a precedent outright can confine the precedent to a stated set of facts, decline to extend the precedent to a neighboring problem, and decide the neighboring problem on a rationale incompatible with the broader logic of the earlier case. Each step preserves the formal authority of the precedent while reducing the work the precedent performs. Erosion therefore functions as both a cause and an indicator within the stare decisis analysis, since an overruling majority cites the erosion as evidence that the foundations have gone, while a dissent answers that the majority manufactured the erosion for that purpose.

The administrative law line that produced Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), offers the fullest modern outline of the process. Chevron U.S.A. Inc. v. Natural Resources Defense Council, 467 U.S. 837 (1984), directed courts to accept a reasonable agency interpretation of an ambiguous statute administered by the agency, and for two decades the Chevron (1984) framework governed judicial review across regulatory fields. Beginning with United States v. Mead Corp., 533 U.S. 218 (2001), the Supreme Court multiplied threshold questions about the circumstances in which Chevron deference applied at all, and Christensen v. Harris County, 529 U.S. 576 (2000), confined deference for interpretations lacking the force of law. Justice Antonin Scalia, dissenting in Mead (2001), warned that the threshold inquiry would produce confusion and ossification, a warning later opinions quoted as the qualifications accumulated. Many of the deference disputes reached the Supreme Court through the financial regulation court cases that tested the authority of banking, securities, and consumer finance regulators, and that litigation line supplied the records on which the erosion argument was built.

Chief Justice John Roberts, writing for the majority in Loper Bright (2024), completed the sequence by overruling Chevron (1984) expressly and holding that courts must exercise independent judgment in deciding whether an agency had acted within statutory authority, as required by the Administrative Procedure Act of 1946. The majority in Loper Bright (2024) presented the holding as the recognition of an erosion already far advanced, pointing to the frequency with which the Supreme Court had resolved major regulatory cases without applying the Chevron framework in the preceding decade. Justice Elena Kagan, dissenting in Loper Bright (2024), answered that the majority understated the continuing vitality of Chevron (1984) in the Courts of Appeals, where judges applied the framework in thousands of routine cases. The disagreement in Loper Bright (2024) exposes the central difficulty of the erosion factor, since both sides examined the same intervening decisions and drew opposite conclusions about whether the foundations remained. Overruling opinions therefore pair an erosion claim with a quality of reasoning claim, so that the hollowing narrative supports a methodological verdict reached on independent grounds.

Reliance Interests: Who Ordered Their Affairs on the Old Rule

Reliance interests supply the most concrete branch of the stare decisis inquiry, because reliance directs attention from the reasoning of judges to the conduct of persons and institutions that arranged primary behavior in dependence on a governing rule. The Supreme Court weighs reliance most heavily when the reliance is specific, documented, and costly to unwind, a pattern visible across property, contract, and commercial law. Justice Stephen Breyer, writing for the majority in Kimble v. Marvel Entertainment, LLC, 576 U.S. 446 (2015), refused to overrule Brulotte v. Thys Co., 379 U.S. 29 (1964), on patent royalties extending beyond the patent term, and the majority in Kimble (2015) emphasized that Congress remained free to revise the rule by statute and that parties had drafted licensing agreements against the background of Brulotte (1964). Justice Samuel Alito, dissenting in Kimble (2015), answered that the Brulotte rule rested on misconceived economic assumptions and that reliance on an inefficient rule deserved limited weight.

Planned Parenthood v. Casey, 505 U.S. 833 (1992), presents the contested instance that tests the boundaries of the principle, because the plurality in Casey (1992) credited a form of societal reliance that resisted transaction level documentation. The plurality, written jointly by Justices Sandra Day O’Connor, Anthony Kennedy, and David Souter in 1992, reasoned that the ability of women to participate equally in the economic and social life of the United States had been facilitated by the availability of abortion under Roe v. Wade, 410 U.S. 113 (1973). Critics of the Casey (1992) analysis, including Chief Justice William Rehnquist in dissent in Casey (1992), argued that reliance of that generality would insulate any longstanding precedent from reconsideration, since any durable rule shapes social arrangements over time. The majority in Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), later discounted the Casey reliance analysis, arguing that intangible reliance and contested empirical claims about social effects deserved little weight against a holding the majority regarded as gravely erroneous. The majority in Dobbs (2022) contrasted the reliance in Casey (1992) with the reliance that protects rules of property and contract, where identifiable parties change position in identifiable transactions.

What is the difference between overruling and distinguishing a case?

Overruling discards the earlier rule for future cases and denies the precedent further authority as governing law. Distinguishing accepts the earlier rule as valid law but holds that the facts of the pending case fall outside the rule, leaving the precedent intact, citable, and available to govern a future case with materially similar facts.

The distinction between the two operations shapes how reliance is argued and assessed in briefs filed after Dobbs (2022). Counsel defending a precedent documents reliance with declarations, transaction records, and administrative guidance that cites the precedent by name, following the model that succeeded in Dickerson v. United States, 530 U.S. 428 (2000), where Miranda warnings had structured police training and interrogation practice for 34 years. Counsel seeking an overruling answers by characterizing the documented reliance as sunk cost, transitional cost, or reliance on a rule that regulated government conduct rather than primary private ordering. Courts evaluating the competing characterizations ask who relied, when the reliance occurred, whether the reliance was reasonable in light of signals that the precedent stood in doubt, and whether a transition rule could mitigate the disruption.

Changed Facts and Changed Understanding: The Brown Route

Brown v. Board of Education, 347 U.S. 483 (1954), displaced Plessy v. Ferguson, 163 U.S. 537 (1896), by a route that later overruling opinions have treated as the model for the changed facts and changed understanding factor. Chief Justice Earl Warren, writing for a unanimous Court in Brown (1954), did not rest the judgment on a claim that the Fourteenth Amendment had changed meaning since 1868, and the opinion in Brown (1954) devoted its central analysis to the transformed role of public education in American life. The opinion in Brown (1954) observed that education had become the foundation of citizenship, compulsory under state law, and supported by taxation on a scale unknown when the Court decided Plessy (1896). By reframing the factual premise, the Court in Brown (1954) concluded that the separate but equal doctrine of Plessy (1896) had no place in public education, where separation itself generated a feeling of inferiority in Black schoolchildren that affected the capacity to learn. The opinion in Brown (1954) supported the conclusion with a footnote citing social science research, including Kenneth Clark and Mamie Clark, studies published in 1950 on the effects of segregation on child development.

The careful limits of the Brown (1954) method deserve emphasis, because later litigants invoke Brown (1954) for propositions the opinion did not establish. The Court in Brown (1954) confined the holding to public education and decided the companion cases on transportation and recreation separately, including Bolling v. Sharpe, 347 U.S. 497 (1954), for the District of Columbia. The Court in Brown (1954) also deferred remedy to further proceedings, resolved in Brown v. Board of Education, 349 U.S. 294 (1955), with the direction that desegregation proceed with all deliberate speed. Historians of the decision, including Richard Kluger in Simple Justice (1976), documented that the unanimity in Brown (1954) depended in part on the narrowness of the stated ground. A changed facts argument persuades most readily when the factual change is institutional, verifiable, and directly connected to the premise of the earlier opinion, rather than when the argument amounts to a claim that moral opinion has moved.

Lawrence v. Texas, 539 U.S. 558 (2003), extended the changed understanding route to a precedent of much shorter duration and showed the factor operating across 17 years rather than across six decades. Justice Anthony Kennedy, writing for the majority in Lawrence (2003), overruled Bowers v. Hardwick, 478 U.S. 186 (1986), which had upheld the criminal prohibition of homosexual sodomy in Georgia. The majority in Lawrence (2003) held that Bowers (1986) had misstated history, overstated the antiquity and uniformity of the prohibitions, and failed to account for a growing awareness that liberty protects adult persons in private intimate conduct. The majority in Lawrence (2003) pointed to the intervening decision in Planned Parenthood v. Casey, 505 U.S. 833 (1992), on liberty and autonomy, to the intervening decision in Romer v. Evans, 517 U.S. 620 (1996), on classifications directed at gay persons, and to the judgment of the European Court of Human Rights in Dudgeon v. United Kingdom (1981). Justice Antonin Scalia, dissenting in Lawrence (2003), answered that the majority had applied stare decisis selectively, preserving Casey (1992) while discarding Bowers (1986), decided only six years before Casey (1992). The exchange in Lawrence (2003) demonstrates that changed understanding arguments succeed when the intervening legal developments are named decisions with dates, rather than general appeals to social movement.

Signals Before the Fall: Dissents, Statements and Grants That Mark a Precedent

A precedent rarely falls without advance signals, and lawyers who monitor the signals read dissents, statements respecting denials of certiorari, concurrences, and grants of review as a connected forecasting record. A dissent from the denial of certiorari, or a concurrence in a judgment that questions an adjacent holding, announces that sitting justices regard a precedent as vulnerable and invites litigants to present the question in a suitable vehicle. Justice Clarence Thomas, in concurrences including the concurrence in Gamble v. United States, 587 U.S. 678 (2019), and the concurrence in Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), stated the position that substantive due process precedents rested on a flawed foundation and warranted reconsideration as a class. That stated position belongs to Justice Thomas as an individual justice rather than to a majority. Lower courts remain bound by the precedents so questioned, under the rule of Rodriguez de Quijas v. Shearson American Express, Inc., 490 U.S. 477 (1989), which instructs Courts of Appeals to follow directly controlling precedent even when the reasoning appears eroded, leaving to the Supreme Court the prerogative of overruling.

Erosion in adjacent doctrine supplies a second class of signal, distinct from the statements of individual justices and often more predictive, because adjacent decisions bind lower courts immediately. When the Supreme Court decides a neighboring question on a rationale that cannot coexist with the logic of an older precedent, sophisticated counsel treat the older precedent as functionally narrowed even though the older holding remains on the books. The Alien Tort Statute line offers a compact illustration of narrowing as a signal, since the account of how the Alien Tort Statute works traces Filartiga v. Pena-Irala, 630 F.2d 876 (2d Cir. 1980), confined step by step in Sosa v. Alvarez-Machain, 542 U.S. 692 (2004), in Kiobel v. Royal Dutch Petroleum Co., 569 U.S. 108 (2013), and in Jesner v. Arab Bank, PLC, 584 U.S. 241 (2018), until the cause of action recognized in Filartiga (1980) governed only a narrow residue of cases. No single decision in the Sosa (2004), Kiobel (2013), and Jesner (2018) sequence overruled Filartiga (1980), yet the sequence as a whole communicated the direction of the Court with a clarity comparable to an express statement. A grant limited to the question whether a named precedent should be overruled, as in the grant in Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), concerning Chevron U.S.A. Inc. v. Natural Resources Defense Council, 467 U.S. 837 (1984), converts speculation into a briefing schedule and ordinarily predicts that at least four justices regard the question as open and substantial.

Prudent counsel discount each signal class by a different factor, and the discounting method itself forms part of professional competence in Supreme Court practice. A lone dissent from a denial of certiorari, without a companion statement from other chambers, predicts little beyond the views of the signing justices, while a concurrence joined by multiple justices in a decided case predicts considerably more. When the overruling opinion finally arrives, as in Dobbs (2022) and in Loper Bright (2024), the reasoning frequently recites the signals as part of the justification, treating advance notice as a mitigating answer to the reliance objection.

The Political Branches in the Story: Amendment, Jurisdiction and Appointment

Stare decisis governs the relations among courts, but the Constitution of the United States distributes the response to judicial precedent across the political branches as well, and three mechanisms define the distribution: constitutional amendment, congressional regulation of jurisdiction, and presidential appointment with senatorial confirmation. The amendment route stands as the most direct repudiation of a constitutional precedent, because an amendment removes the textual ground on which the precedent rested and binds every court without further litigation. The Eleventh Amendment, ratified in 1795, repudiated Chisholm v. Georgia, 2 U.S. 419 (1793), on state sovereign immunity from private suit. The Fourteenth Amendment, ratified in 1868, repudiated the holding in Dred Scott v. Sandford, 60 U.S. 393 (1857), that persons of African descent held no rights of citizenship. The Sixteenth Amendment, ratified in 1913, repudiated Pollock v. Farmers Loan and Trust Co., 157 U.S. 429 (1895), on the classification of income taxation as a direct tax. The Twenty-Sixth Amendment, ratified in 1971, repudiated in part Oregon v. Mitchell, 400 U.S. 112 (1970), on the voting age for state elections.

Congressional regulation of jurisdiction offers a second mechanism, narrower in scope and confined by limits that the case law states with care. Article III grants Congress authority over the appellate jurisdiction of the Supreme Court subject to exceptions and regulations, and Ex parte McCardle, 74 U.S. 506 (1869), sustained the repeal of a statutory grant of habeas corpus appellate jurisdiction while a case stood pending under the repealed grant. Chief Justice Salmon Chase, writing for the Court in McCardle (1869), upheld the repeal as an exercise of the exceptions power, and the holding in McCardle (1869) is stated narrowly in later commentary because the Court in McCardle (1869) decided only the effect of the repeal on the pending appeal and preserved other avenues of habeas review, as the Court confirmed in Ex parte Yerger, 75 U.S. 85 (1869). Later decisions, including United States v. Klein, 80 U.S. 92 (1872), barred Congress from prescribing rules of decision in pending cases, and the boundary between Klein (1872) and McCardle (1869) continues to structure debate about jurisdiction stripping proposals.

Appointments supply the third mechanism and, in practice, the route by which precedent coalitions change most frequently, because the composition of the Supreme Court determines which stare decisis arguments command five votes. A president nominates and the Senate confirms, and each confirmed justice brings a methodology, a set of priors about contested precedents, and a vote on every future motion to reconsider. The process operates slowly, since vacancies arise irregularly and a single appointment alters the outcome only in closely divided lines of doctrine, yet the cumulative effect across successive appointments explains major doctrinal transitions more completely than any internal account of legal reasoning alone. The spending power illustrates the continuity that appointments leave undisturbed when no coalition forms against a precedent, since the account of how federal grant conditions work shows South Dakota v. Dole, 483 U.S. 203 (1987), still structuring federal bargains through grant conditions that states accept with knowledge of the terms. The three mechanisms together explain the institutional rhythm of overruling in the United States, in which the Court corrects constitutional doctrine internally and Congress corrects statutory doctrine legislatively.

Ramos and the Modern Restatement of the Factors

Ramos v. Louisiana, 590 U.S. 83 (2020), held that the Sixth Amendment requires a unanimous jury verdict for a criminal conviction in state court, and Ramos (2020) overruled Apodaca v. Oregon, 406 U.S. 404 (1972), which had permitted Oregon to convict by a vote of ten to two. Johnson v. Louisiana, 406 U.S. 356 (1972), had permitted a similar practice in Louisiana, and no single rationale commanded a majority in Apodaca (1972), because Justice Lewis F. Powell Jr. supplied a separate distinction between federal and state jury practice that the plurality in Apodaca (1972) never adopted.

Justice Neil Gorsuch, writing for the majority in Ramos (2020), treated the unanimity requirement as settled by the text and history of the Sixth Amendment and by the incorporation decisions of the Supreme Court of the United States carrying the Bill of Rights against the states in the same form that governed federal trials. The majority in Ramos (2020) then concluded that Apodaca (1972) had been wrongly decided, had rested on a rationale in Apodaca (1972) that no majority had ever joined, and had produced unworkable doctrine in state criminal procedure across the cases that followed Apodaca (1972). Louisiana and Oregon argued in Ramos (2020) that completed convictions and retrial burdens rested on the older rule in Apodaca (1972), and the majority in Ramos (2020) answered that a rule permitting nonunanimous convictions imposed a cost in legitimacy that reliance could not cure where the rule in Apodaca (1972) bore on the validity of the verdict itself.

Justice Brett Kavanaugh, concurring in Ramos (2020), gave the decision influence beyond the jury question by setting out a structured account of when the Supreme Court of the United States should overrule a constitutional precedent, and later majorities returned to the Ramos (2020) concurrence as a shared vocabulary in Dobbs (2022) and in Loper Bright (2024). In that concurrence in Ramos (2020), Justice Brett Kavanaugh described stare decisis as a doctrine of judicial policy rather than a command, language drawn from Payne v. Tennessee, 501 U.S. 808 (1991), which means in Ramos (2020) that precedent constrains judgment through reasons to be weighed rather than through an obligation removing judgment from the decision. Justice Brett Kavanaugh grouped the considerations in his Ramos (2020) concurrence under three questions, the first asking whether the earlier decision had been grievously or egregiously wrong in Ramos (2020) terms, a threshold set higher than simple disagreement, the second asking whether the earlier decision had caused significant negative jurisprudential or real world consequences, and the third asking whether overruling would disturb legitimate reliance of the kind protected in contract and property settings.

Justice Samuel Alito, dissenting in Ramos (2020), warned that discarding Apodaca (1972) would require review of many completed convictions and unsettle procedure in the states that had organized trials around divided verdicts, and Justice Samuel Alito argued in Ramos (2020) that the majority had understated reliance because retrials burden courts, witnesses, victims, and defendants alike. Read together, the opinions in Ramos (2020) showed a majority and a dissent using the same categories to reach opposing conclusions, so the factors in Ramos (2020) supplied the questions and the discipline of public answers, while the choice among those answers in Ramos (2020) remained an act of judgment by the Supreme Court of the United States.

Dobbs as a Worked Application of the Full Test

Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), upheld a Mississippi statute restricting abortion and overruled Roe v. Wade, 410 U.S. 113 (1973), and the controlling portions of Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), in an opinion by Justice Samuel Alito. The majority in Dobbs (2022) organized the stare decisis discussion around five factors applied in sequence, namely the nature of the error in the earlier decisions, the quality of the reasoning in those decisions, the workability of the rules the decisions imposed, the effect of those rules on other doctrines, and the presence or absence of concrete reliance interests.

On the nature of the error, the majority in Dobbs (2022) held that Roe (1973) had read a right into the Fourteenth Amendment without support in the text of that amendment or in the history surrounding ratification in 1868, and the majority in Dobbs (2022) described that error as egregious because Roe (1973) and Casey (1992) had removed a contested policy question from legislatures in every state. On the quality of reasoning, the majority in Dobbs (2022) argued that Roe (1973) resembled legislation rather than constitutional interpretation, pointing to the trimester framework in Roe (1973) as a regulatory scheme, and the majority in Dobbs (2022) also argued that Casey (1992) had retained a central holding while discarding much of the supporting rationale in Roe (1973).

On workability, the majority in Dobbs (2022) focused on the viability line that Casey (1992) had preserved as the dividing point for state regulation, and the majority in Dobbs (2022) argued that the undue burden standard attached to that line in Casey (1992) had proved difficult for lower courts to apply consistently across factual settings. On doctrinal effect, the majority in Dobbs (2022) argued that Roe (1973) and Casey (1992) had distorted standing, severability, and the rules governing facial challenges in other fields. On reliance, the majority in Dobbs (2022) held that the interests advanced for Roe (1973) and Casey (1992) were not the concrete reliance interests that stare decisis protects most strongly, contrasting planning in property and contract settings with the general societal claims in Dobbs (2022), and the majority in Dobbs (2022) concluded that all five factors supported returning the question to state legislatures.

The joint dissent of Justice Stephen Breyer, Justice Sonia Sotomayor, and Justice Elena Kagan in Dobbs (2022) answered the majority on the same five factors and reached the opposite assessment on each point in Dobbs (2022), arguing that Roe (1973) and Casey (1992) had balanced competing interests consistently with other liberty decisions, so disagreement in Dobbs (2022) did not amount to grievous error. The joint dissent in Dobbs (2022) defended the viability line in Casey (1992) as a workable rule that lower courts had administered for decades and treated the undue burden standard in Casey (1992) as comparable in difficulty to other standards in constitutional law applied by the Supreme Court of the United States. The joint dissent in Dobbs (2022) placed reliance at the center of the response in Dobbs (2022), arguing that generations had organized education, work, and family life against the background of Roe (1973) and Casey (1992), and Dobbs (2022) therefore displayed both the reach of the factor test and the dependence of the test on characterization by the deciding majority in Dobbs (2022).

Loper Bright and the Overruling of an Administrative Law Fixture

Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), overruled Chevron U.S.A. v. Natural Resources Defense Council, 467 U.S. 837 (1984), after forty years in which Chevron (1984) had supplied the governing framework for judicial review of agency interpretations of statutes in the federal courts of the United States. Chevron (1984) instructed courts to ask whether Congress had spoken directly to a question and, if a statute was ambiguous, whether the agency interpretation was reasonable, in which case a court deferred to the agency under the second step of Chevron (1984). Chief Justice John Roberts, writing for the majority in Loper Bright (2024), held that the Administrative Procedure Act of 1946 directed courts to decide questions of law, and the majority in Loper Bright (2024) grounded the replacement rule in the tradition of Marbury v. Madison, 5 U.S. 137 (1803), under which courts state what the law is.

The reasoning argument in Loper Bright (2024) held that Chevron (1984) rested on a fiction that statutory ambiguity amounted to an implicit delegation of interpretive authority from Congress to the administering agency, while ambiguity in Loper Bright (2024) analysis more often reflects drafting limits, a compromise text, or a question Congress never confronted in enacting the statute. The majority in Loper Bright (2024) treated Skidmore v. Swift and Company, 323 U.S. 134 (1944), as the sounder model, because Skidmore (1944) gave agency views weight according to the power to persuade, and under Loper Bright (2024) a court exercises independent judgment on statutory meaning while considering agency reasoning as evidence of a considered position.

The workability argument in Loper Bright (2024) answered defenders who presented Chevron (1984) as a stable settlement, because the majority in Loper Bright (2024) held that the threshold question in Chevron (1984) whether a statute is ambiguous had proved indeterminate, so similar texts produced different outcomes among judges applying Chevron (1984) in the United States Courts of Appeals. The majority in Loper Bright (2024) pointed to refinements such as the inquiry in United States v. Mead Corporation, 533 U.S. 218 (2001), into whether Chevron (1984) applied at all, and the majority in Loper Bright (2024) also noted that the Supreme Court of the United States had resolved many statutory cases without invoking Chevron (1984). Reliance received a distinct treatment in Loper Bright (2024), because the majority in Loper Bright (2024) stated expressly that prior holdings relying on Chevron (1984) to uphold specific agency actions remained undisturbed, since those holdings in Loper Bright (2024) stand as statutory precedents in their own right and remain subject to statutory stare decisis.

Justice Elena Kagan, dissenting in Loper Bright (2024), argued that the majority in Loper Bright (2024) discarded a settled allocation of interpretive authority that Congress had legislated against for decades, presenting congressional delegation in Loper Bright (2024) as the foundation of Chevron (1984), because Congress assigns broad programs to agencies expecting agencies to fill textual gaps. Justice Elena Kagan also argued in Loper Bright (2024) that agencies bring subject matter knowledge and political accountability that generalist courts lack, and on stare decisis Justice Elena Kagan argued in Loper Bright (2024) that Chevron (1984) had become background law against which Congress drafted statutes and lower courts decided thousands of cases.

What an Overruling Does to Cases Already Decided

An overruling changes the governing rule for disputes that remain open, and the law of retroactivity determines how far that change reaches into litigation that began under the older rule in cases decided by the Supreme Court of the United States. Harper v. Virginia Department of Taxation, 509 U.S. 86 (1993), addressed civil retroactivity and held that a new rule of federal law in Harper (1993) applies to all cases still open on direct review and to events arising before or after announcement of the rule in Harper (1993). The rule in Harper (1993) rejected selective prospectivity in civil cases, under which a court might apply a new rule to the winning party in Harper (1993) and deny the same benefit to other litigants in identical positions whose cases remained pending in federal court.

Does overruling a precedent make the earlier case disappear from the law?

No. The earlier opinion remains printed in the United States Reports and remains part of legal history. What changes is its authority. Later courts cite the decision as overruled, examine any reasoning that survived the overruling, and treat its governing rule as no longer controlling in future litigation.

Griffith v. Kentucky, 479 U.S. 314 (1987), stated the parallel rule for criminal cases on direct review, holding in Griffith (1987) that a newly declared rule for the conduct of criminal prosecutions applies to all cases pending on direct review or not yet final when the rule in Griffith (1987) was announced. Justice Byron White reasoned in Griffith (1987) that failure to apply a new rule to a case still moving through the ordinary appellate process in Griffith (1987) would treat similarly situated defendants differently on the accident of timing in Griffith (1987).

Teague v. Lane, 489 U.S. 288 (1989), built a stricter regime for collateral review, where a plurality in Teague (1989) held that new constitutional rules of criminal procedure generally do not apply to cases that became final before the new rules in Teague (1989) were announced. Justice Sandra Day O’Connor grounded the limit in Teague (1989) in respect for final judgments, the deterrent function of habeas review, and the interest of states in the integrity of completed prosecutions under the framework in Teague (1989). The framework in Teague (1989) recognized narrow exceptions for rules placing private conduct beyond the power of the state to proscribe and for watershed rules of criminal procedure, and Edwards v. Vannoy, 593 U.S. 255 (2021), later held in Edwards (2021) that the unanimity rule of Ramos (2020) did not apply on federal collateral review.

Final judgments stay closed under this structure even when the rule that produced a judgment has been discarded as wrong, because a party who lost under the older precedent and exhausted every appeal cannot reopen the judgment solely because the Supreme Court of the United States later adopted the argument that party had pressed in earlier litigation. Congress may change the underlying law for future cases, while Plaut v. Spendthrift Farm, 514 U.S. 211 (1995), held in Plaut (1995) that Congress may not command the reopening of a judgment that a court had finally entered, so the picture after an overruling in Plaut (1995) terms is layered rather than clean.

Partial Displacement: Overruling in Part, Pluralities and the Marks Problem

Overruling often arrives in partial form, because a later majority may discard one holding of an earlier decision while preserving another holding of that decision that rests on different reasoning in the earlier case before the Supreme Court of the United States. Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), supplied the classic structure, since the joint opinion in Casey (1992) preserved the central holding of Roe v. Wade, 410 U.S. 113 (1973), while discarding the trimester framework and the strict scrutiny rule that Roe (1973) had attached to regulation. Dobbs (2022) later described that partial preservation in Casey (1992) as a source of instability, which shows how a compromise displacement in Casey (1992) can postpone a confrontation rather than resolve the underlying disagreement in Casey (1992).

Partial displacement therefore requires a reader to identify the precise proposition that fell in a case such as Garcia (1985), because the name of an older case such as National League of Cities (1976) can survive in citations long after a central component of National League of Cities (1976) has ceased to govern. A plurality decision complicates that identification, because a plurality opinion announces the judgment of the Supreme Court of the United States in a case such as Bakke (1978) without stating a rationale that a majority of justices accepted in that case.

The rule for extracting a holding from a plurality decision comes from Marks v. United States, 430 U.S. 188 (1977), which stated in Marks (1977) that the holding may be viewed as the position taken by the justices who concurred in the judgment on the narrowest grounds in Marks (1977). The formulation in Marks (1977) works poorly when concurring opinions rest on incompatible premises, because no single position then commands support as a common denominator in Marks (1977) analysis, and courts in the United States Courts of Appeals have divided over the application of Marks (1977) in that situation.

Apodaca v. Oregon, 406 U.S. 404 (1972), remains the worked instance of the difficulty under Marks (1977), because four justices in Apodaca (1972) accepted nonunanimous verdicts in state and federal trials, four justices in Apodaca (1972) would have required unanimity in both systems, and Justice Lewis F. Powell Jr. alone in Apodaca (1972) distinguished the systems. Under the narrowest grounds approach associated with Marks (1977), the lone position of Justice Lewis F. Powell Jr. in Apodaca (1972) became the effective holding in Apodaca (1972), even though eight justices in Apodaca (1972) rejected the distinction on which that position in Apodaca (1972) rested. Ramos (2020) described that result in Apodaca (1972) as an anomaly, and Regents of the University of California v. Bakke, 438 U.S. 265 (1978), produced a comparable structure in Bakke (1978), where the opinion of Justice Lewis F. Powell Jr. in Bakke (1978) influenced admissions law for decades without a majority rationale in Bakke (1978). Grutter v. Bollinger, 539 U.S. 306 (2003), later adopted a version of that reasoning in Bakke (1978) in a majority opinion in Grutter (2003), which shows how a plurality position in Bakke (1978) can acquire settled authority through later endorsement in Grutter (2003), while a fractured decision such as Apodaca (1972) can instead receive reduced weight in Ramos (2020).

The Complication: A Test Flexible Enough to Justify Almost Anything

The strongest objection to the modern factor test holds that the test rationalizes a result reached on other grounds, because each factor in the test from Ramos (2020) and Dobbs (2022) depends on a characterization that the deciding majority in a case such as Dobbs (2022) controls. The joint dissent of Justice Stephen Breyer, Justice Sonia Sotomayor, and Justice Elena Kagan in Dobbs (2022) stated that argument in sharp form, contending in Dobbs (2022) that the only change between Casey (1992) and Dobbs (2022) was the composition of the Supreme Court of the United States. In the formulation of the joint dissent in Dobbs (2022), no new facts had emerged, no new understanding of the governing text had been demonstrated, and no development in other doctrine had undermined Roe (1973) or Casey (1992), so the overruling in Dobbs (2022) rested on the preferences of a new majority in Dobbs (2022).

Justice Louis Brandeis supplied an older version of the objection in a dissent in Burnet v. Coronado Oil and Gas Company, 285 U.S. 393 (1932), where Justice Louis Brandeis argued in Burnet (1932) that settlement of a question mattered more in many settings than settlement according to an ideal rule in Burnet (1932). Justice Louis Brandeis reasoned in Burnet (1932) that the public and the legal profession order conduct around decided cases, and that frequent correction by the Supreme Court of the United States after Burnet (1932) would teach regulated parties to treat every disfavored precedent after Burnet (1932) as provisional.

Stability under the account in Burnet (1932) is a good in itself, independent of the merits of the rule being stabilized in Burnet (1932), and a test that prices stability only through reliance in individual transactions after Burnet (1932) will undervalue that good in Burnet (1932).

The counter consideration carries real force, because the same flexibility that permits a contested overruling also carried Brown v. Board of Education, 347 U.S. 483 (1954), which discarded the doctrine of Plessy v. Ferguson, 163 U.S. 537 (1896), after fifty eight years in Brown (1954). Lawrence v. Texas, 539 U.S. 558 (2003), overruled Bowers v. Hardwick, 478 U.S. 186 (1986), after seventeen years in Lawrence (2003), and the majority in Lawrence (2003) argued that Bowers (1986) had misdescribed history and the liberty interest in Lawrence (2003) from the outset.

The thesis requires narrowing at that point, and the narrowed claim concerns justification rather than prediction, because the factors in Ramos (2020) and Dobbs (2022) cannot fix an outcome in advance where reasonable judges weigh error, workability, and reliance differently on the same record in Dobbs (2022). The factors in Ramos (2020) nevertheless discipline the justification a majority must offer in public, since a majority that wishes to overrule under Ramos (2020) must describe the earlier error in testable terms, must address the reliance the majority proposes to disturb under Ramos (2020), and must state a replacement rule whose scope later courts can apply after Ramos (2020). That discipline from Ramos (2020) is weaker than a binding rule and stronger than unrestricted choice, and the narrowed thesis claims no more than that middle position for stare decisis as restated in Ramos (2020) and applied in Dobbs (2022) by the Supreme Court of the United States.

How Other Courts and Other Countries Handle the Same Problem

The Practice Statement issued by the House of Lords in 1966 resolved the problem of self binding precedent in the United Kingdom, when Lord Chancellor Gardiner announced in the Practice Statement (1966) that the House of Lords would depart from a previous decision when a departure appeared right to do so, a step that ended the self binding rule the House of Lords had followed since London Street Tramways Company v. London County Council, decided in 1898. The Practice Statement (1966) freed the House of Lords from that complete self binding rule in London Street Tramways (1898), while warning in the Practice Statement (1966) that precedent supplies an indispensable foundation for deciding cases with certainty and that departure would be used sparingly where contracts and property settlements rested on an existing rule. The first full exercise of the power in the Practice Statement (1966) came in Conway v. Rimmer, decided in 1968, where the House of Lords departed in Conway (1968) from Duncan v. Cammell Laird and Company, decided in 1942, on the disclosure of government documents in litigation after Duncan (1942).

Canada states a framework through decided cases rather than a practice statement, and Canada (Attorney General) v. Bedford, 2013 SCC 72, supplies the formulation in Bedford (2013), under which a lower court in Canada may depart from binding precedent in Bedford (2013) where a new legal issue is raised or where a change in circumstances or evidence in Bedford (2013) fundamentally shifts the parameters of the debate. Chief Justice Beverley McLachlin applied that framework in Bedford (2013) to a challenge involving prostitution related provisions.

India organizes the same problem through bench strength, because a bench of the Supreme Court of India may not overrule a decision of a larger bench of the Supreme Court of India, and a judge of the Supreme Court of India who doubts a precedent of equal strength refers the question to the Chief Justice of India for constitution of a larger bench of the Supreme Court of India. Article 145 of the Constitution of India, adopted in 1950, provides in Article 145 (1950) that a substantial question involving interpretation of the Constitution of India is heard by a bench of at least five judges, known as a Constitution Bench under Article 145 (1950), and the Supreme Court of India restated the discipline in Central Board of Dawoodi Bohra Community v. State of Maharashtra, decided in 2005.

The three designs allocate the same pressures differently, because the Practice Statement (1966) in the United Kingdom trusts a final court to announce a general liberty in the Practice Statement (1966) and exercise that liberty with measurable restraint, Bedford (2013) in Canada channels change through new issues and changed evidentiary foundations in Bedford (2013), and the bench discipline in Central Board of Dawoodi Bohra Community (2005) in India makes authority depend on the size of the bench assembled in the Supreme Court of India. The Supreme Court of the United States, by contrast, sits in a single configuration for nearly every case and states governing considerations in opinions such as Ramos (2020), so the American solution in Ramos (2020) rests on articulated factors rather than panel size in India, a threshold record in Bedford (2013), or a freestanding statement in the Practice Statement (1966), and none of the four designs in Ramos (2020), Bedford (2013), or the Practice Statement (1966) removes discretion from judges.

A Practical Reading Method for Any Future Overruling Opinion

A disciplined reading of an overruling opinion from the Supreme Court of the United States begins with the holding, stated in one sentence that names the precedent discarded, the precedent preserved, and the rule that governs after judgment in a case such as Loper Bright (2024). You can test that sentence against the syllabus and the concluding pages of the majority opinion in Loper Bright (2024), where the holding in Loper Bright (2024) discarded a method of deference in Chevron (1984) and preserved holdings reached under Chevron (1984) in earlier cases. A holding sentence that cannot distinguish the fate of Chevron (1984) from the fate of a regulation upheld under Chevron (1984) has missed the structure of Loper Bright (2024), and precision at the first step in Loper Bright (2024) prevents the common error of treating strong criticism of a precedent in Dobbs (2022) as an overruling in Dobbs (2022).

The next step locates the factor the majority treats as decisive, because most overruling opinions assign unequal weight to error, reasoning, workability, doctrinal effect, and reliance even when an opinion in Dobbs (2022) addresses all five considerations from Dobbs (2022) in sequence. You can mark the passage where the majority in Ramos (2020) moved from describing Apodaca (1972) as wrong to describing Apodaca (1972) as grievously wrong in Ramos (2020), since that movement in Ramos (2020) identifies the threshold the majority in Ramos (2020) believed the record had crossed. Where the majority in Loper Bright (2024) devoted extended analysis to the Administrative Procedure Act of 1946, the textual grounding of the claimed error in Loper Bright (2024) functioned as the load bearing element of the overruling in Loper Bright (2024).

The dissent deserves reading on the same factor rather than as a general protest, because a dissent that contests the decisive premise in a case such as Loper Bright (2024) exposes the point of greatest fragility in the new rule announced in Loper Bright (2024). Justice Elena Kagan, dissenting in Loper Bright (2024), answered the majority in Loper Bright (2024) on delegation and institutional competence, so a paired reading of Loper Bright (2024) shows that the disagreement in Loper Bright (2024) concerned which institution Congress had authorized to resolve statutory gaps after Chevron (1984).

A careful reader then checks what an opinion expressly leaves undisturbed, because the preservation language in an opinion such as Loper Bright (2024) often determines the practical reach of Loper Bright (2024) for years afterward in litigation under Chevron (1984). Loper Bright (2024) preserved prior holdings that had upheld specific agency actions under Chevron (1984), and that preservation in Loper Bright (2024) placed a defined class of disputes outside the new litigation that Loper Bright (2024) otherwise invited in federal court. You can record each preservation sentence in brief form, note the class of cases a sentence in Dobbs (2022) covers, and keep the case line and factor notes together in a study file such as a reference file so that the limits in Dobbs (2022) travel with the holding in later review. The final step states the replacement rule with exact scope, including the standard a lower court must apply after Loper Bright (2024), the party who bears a burden under the new rule in Loper Bright (2024), and any class of cases the Supreme Court of the United States reserved in Loper Bright (2024) or in Dobbs (2022).

What Stays Settled and What the Doctrine Cannot Settle

Stare decisis endures in the jurisprudence of the Supreme Court of the United States not as a command that forbids correction and not as a courtesy that a majority may waive, but as a practice that requires a public reason whenever the Supreme Court of the United States displaces a settled rule in a case such as Ramos (2020). Brown v. Board of Education, 347 U.S. 483 (1954), and Ramos v. Louisiana, 590 U.S. 83 (2020), stand at different points in the history of that practice, yet each majority in Brown (1954) and Ramos (2020) accepted the burden of explaining why an older settlement in Plessy (1896) or Apodaca (1972) could no longer govern after Brown (1954) and Ramos (2020). Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), stated the burden from the side of preservation in Casey (1992), arguing in Casey (1992) that a decision to overrule under sustained public pressure would subvert the legitimacy of the Supreme Court of the United States, while Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), stated the burden from the side of correction in Dobbs (2022) through five named considerations in Dobbs (2022).

What stays settled after a century of argument is the catalogue of considerations that a responsible overruling must address, because Justice Brett Kavanaugh assembled that catalogue in Ramos (2020) from grievous error, harmful consequences, and reliance, and the majority in Dobbs (2022) refined the catalogue from Ramos (2020) into five inquiries that later litigants after Dobbs (2022) must anticipate in briefing. Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), added a preservation technique to the catalogue in Loper Bright (2024) by separating a discarded method in Chevron (1984) from holdings already entered under Chevron (1984), and Harper v. Virginia Department of Taxation, 509 U.S. 86 (1993), with Teague v. Lane, 489 U.S. 288 (1989), allocated the effects of a new rule in Harper (1993) and Teague (1989) between open cases, collateral challenges, and final judgments.

What the doctrine cannot settle is the ranking of the considerations in Ramos (2020) when those considerations point in opposite directions on a fully argued record in a case such as Dobbs (2022). Justice Louis Brandeis identified the irreducible choice in Burnet v. Coronado Oil and Gas Company, 285 U.S. 393 (1932), where settlement in Burnet (1932) claimed value apart from correctness, and no later restatement in Ramos (2020) by the Supreme Court of the United States has eliminated the tension that Justice Louis Brandeis named in Burnet (1932) between stability and correction.

The structural implication follows from the identity of the institution that administers the restraint, because stare decisis is a policy of stability administered by the institution the policy restrains, since the Supreme Court of the United States writes the test in Ramos (2020), characterizes the record under the test in Dobbs (2022), and judges whether the justification offered in Dobbs (2022) satisfies the test from Ramos (2020), with no external review of that judgment in Dobbs (2022). The strength of stare decisis is therefore measured in the reasons a majority is willing to publish in a case such as Ramos (2020), in the reliance a majority is willing to confront rather than discount in Dobbs (2022), and in the replacement rule a majority is willing to state with precision in Loper Bright (2024).

Frequently Asked Questions

Q: What does stare decisis mean in plain language?

Stare decisis is a Latin phrase that directs a court to stand by matters already decided, and judges in the United States use the expression for the habit of following earlier rulings in later disputes presenting the same legal question. The Supreme Court of the United States explained the practical meaning in Kimble v. Marvel Entertainment, 576 U.S. 446 (2015), where respect for a prior interpretation promoted consistency in the law. Under this habit, a litigant can predict how a tribunal will approach a familiar issue, and a judge must supply a special justification before discarding a governing answer. The phrase covers two directions of obedience, because a court follows rulings from higher courts in its system and also gives weight to rulings the same court issued in earlier terms. The doctrine never promised permanence, since the Supreme Court of the United States retained authority in Payne v. Tennessee, 501 U.S. 808 (1991), to revisit a holding that proved unsound.

Q: Why do courts follow precedent at all?

Courts follow precedent because a legal system that answers the same question differently from one case to the next loses predictability, and predictability allows people to arrange contracts, property transfers, and daily conduct with some confidence about legal consequences. Justice Louis Brandeis made the settlement argument in Burnet v. Coronado Oil and Gas Company, 285 U.S. 393 (1932), observing that settled law often matters more than an ideally correct rule. The Supreme Court of the United States added an equality rationale in later decisions, because treating like litigants alike requires a judge to explain any departure from an answer given to a comparable party. Precedent also conserves judicial effort, since an earlier opinion in a case such as Marbury v. Madison, 5 U.S. 137 (1803), settles foundational questions that no litigant should have to relitigate from the beginning. Finally, adherence supports institutional legitimacy, as Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), argued when the joint opinion linked public confidence to consistency.

Q: Who can ask the Supreme Court to overrule a precedent?

Any party with standing in a case properly before the Supreme Court of the United States can request that a precedent be discarded, provided the request bears on the judgment the party seeks in that dispute. A petitioner usually raises the argument in a petition for certiorari and again in merits briefing, and the Supreme Court of the United States may also direct the parties to brief whether a named decision should be overruled, as the order in Citizens United v. Federal Election Commission, 558 U.S. 310 (2010), illustrated during reargument in 2009. The Solicitor General of the United States, representing the federal government, can make the same request, and the position of the United States carried weight in Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024). The Supreme Court of the United States almost never reaches the issue without a live case, because Article III of the Constitution of the United States, effective in 1789, confines federal courts to cases and controversies.

Q: Does a precedent have to be old before it can be overruled?

No rule imposes a minimum age, and the Supreme Court of the United States has discarded holdings after intervals ranging from a few years to more than a century. Lawrence v. Texas, 539 U.S. 558 (2003), displaced Bowers v. Hardwick, 478 U.S. 186 (1986), after seventeen years, while Brown v. Board of Education, 347 U.S. 483 (1954), displaced the doctrine of Plessy v. Ferguson, 163 U.S. 537 (1896), after fifty eight years. Age still matters in the analysis, because an older holding in a case such as Swift v. Tyson, 41 U.S. 1 (1842), displaced by Erie Railroad Company v. Tompkins, 304 U.S. 64 (1938), had shaped generations of practice. Justices weigh age alongside error and reliance in Ramos v. Louisiana, 590 U.S. 83 (2020), without treating age alone as protection.

Q: What role does reliance play in keeping a precedent alive?

Reliance describes the concrete arrangements that people, businesses, and governments built on the faith of an existing rule, and strong reliance gives the Supreme Court of the United States a reason to retain a holding even when several justices doubt its original correctness. Kimble v. Marvel Entertainment, 576 U.S. 446 (2015), refused to discard a patent royalty rule from Brulotte v. Thys Company, 379 U.S. 29 (1964), partly because parties had drafted agreements around that rule for decades. Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), gave reliance a broader reading, connecting the preservation of Roe v. Wade, 410 U.S. 113 (1973), to social and economic organization over a generation. Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), narrowed that reading by asking whether the claimed reliance was concrete and measurable rather than general. Ramos v. Louisiana, 590 U.S. 83 (2020), treated reliance as one inquiry among several, which means reliance can preserve a doubtful holding.

Q: Why are statutory precedents harder to overrule than constitutional ones?

Statutory precedents receive stronger protection because Congress can correct a mistaken interpretation by passing an amendment, so the Supreme Court of the United States reasons that legislative silence after a ruling carries meaning and that stability in statutory meaning serves planning. Justice William Rehnquist expressed that logic for the Supreme Court of the United States in Patterson v. McLean Credit Union, 491 U.S. 164 (1989), where the majority declined to discard a construction of a civil rights statute from 1866. Constitutional holdings stand differently, because Article V of the Constitution of the United States, effective in 1789, makes amendment difficult, leaving the Supreme Court of the United States as the only realistic source of correction for an erroneous constitutional rule. Justice Louis Brandeis drew that distinction in Burnet v. Coronado Oil and Gas Company, 285 U.S. 393 (1932), and later majorities preserved the distinction in Ramos v. Louisiana, 590 U.S. 83 (2020).

Q: What did Planned Parenthood v. Casey say about overruling Roe v. Wade?

The joint opinion in Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), written by Justice Sandra Day O’Connor, Justice Anthony Kennedy, and Justice David Souter, refused to discard Roe v. Wade, 410 U.S. 113 (1973), and offered four considerations for that refusal. Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), asked whether the rule in Roe v. Wade, 410 U.S. 113 (1973), had proved unworkable, whether removal would injure reliance, whether intervening doctrine had undermined the holding, and whether facts had changed enough to rob the holding of justification. The joint opinion in Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), answered each inquiry in favor of preservation and added a legitimacy argument, warning that surrender to political pressure would damage the standing of the Supreme Court of the United States.

Q: How did Dobbs v. Jackson Women’s Health Organization justify overruling Roe and Casey?

The majority opinion by Justice Samuel Alito in Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), applied five considerations drawn from earlier stare decisis decisions and concluded that each consideration favored discarding Roe v. Wade, 410 U.S. 113 (1973), and Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992). Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), characterized the original error as egregious, described the reasoning in Roe v. Wade, 410 U.S. 113 (1973), as remarkably loose, and treated the viability rule as lacking a principled basis in text or history. Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), also argued that the undue burden standard had proved unworkable in lower courts and that the abortion holdings had distorted doctrines governing standing and facial challenges. On reliance, Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), found the interests advanced too general and contested to count as the concrete reliance that stare decisis protects most heavily.

Q: What happened when Brown v. Board of Education rejected Plessy v. Ferguson?

Chief Justice Earl Warren, writing for a unanimous Supreme Court of the United States in Brown v. Board of Education, 347 U.S. 483 (1954), held that state mandated separation of children in public schools violated the Equal Protection Clause of the Fourteenth Amendment, ratified in 1868. Brown v. Board of Education, 347 U.S. 483 (1954), declared that separate educational facilities are inherently unequal, a sentence that withdrew the constitutional foundation from the separate but equal doctrine announced in Plessy v. Ferguson, 163 U.S. 537 (1896). The opinion in Brown v. Board of Education, 347 U.S. 483 (1954), avoided a formal announcement using the vocabulary of overruling and never named Plessy v. Ferguson, 163 U.S. 537 (1896), as discarded in express terms, although lawyers and later courts understood the effect without ambiguity. Implementation required a second decision, Brown v. Board of Education, 349 U.S. 294 (1955), which directed desegregation with all deliberate speed through the federal district courts.

Q: Which case ended the Lochner era approach to economic regulation?

West Coast Hotel Company v. Parrish, 300 U.S. 379 (1937), ended the Lochner era approach by upholding a Washington minimum wage law for women and by discarding Adkins v. Children’s Hospital, 261 U.S. 525 (1923), which had invalidated a comparable wage law. Chief Justice Charles Evans Hughes, writing in West Coast Hotel Company v. Parrish, 300 U.S. 379 (1937), rejected the premise that freedom of contract occupied a privileged place under the Due Process Clause of the Fourteenth Amendment, ratified in 1868. The target of the shift was the doctrine associated with Lochner v. New York, 198 U.S. 45 (1905), which had struck down a maximum hours law for bakers on liberty of contract reasoning. West Coast Hotel Company v. Parrish, 300 U.S. 379 (1937), announced that regulation reasonably related to a legitimate legislative end would survive review, and later decisions such as United States v. Carolene Products Company, 304 U.S. 144 (1938), consolidated that deferential posture toward economic legislation by state and federal governments.

Q: What factors did Ramos v. Louisiana list for overruling a precedent?

Justice Brett Kavanaugh, in a concurrence in Ramos v. Louisiana, 590 U.S. 83 (2020), organized the governing inquiries into three groups that synthesized decades of decisions by the Supreme Court of the United States on the retention of precedent. The Ramos v. Louisiana, 590 U.S. 83 (2020), concurrence asked first whether the precedent under review was grievously or egregiously wrong as an original matter, a threshold demanding more than ordinary disagreement. The Ramos v. Louisiana, 590 U.S. 83 (2020), concurrence asked second whether the precedent had produced significant negative consequences in law or in practice, including confusion among lower courts and distortion of related rules. The Ramos v. Louisiana, 590 U.S. 83 (2020), concurrence asked third whether discarding the precedent would injure reliance interests substantial enough to counsel retention despite established error. Justice Brett Kavanaugh applied that structure in Ramos v. Louisiana, 590 U.S. 83 (2020), to Apodaca v. Oregon, 406 U.S. 404 (1972),.

Q: How does a lower court treat a precedent it believes the Supreme Court would overrule?

A lower court must follow a directly controlling precedent of the Supreme Court of the United States even when the reasoning of that precedent appears weakened by later decisions, and the lower court must leave any act of discarding the precedent to the Supreme Court of the United States itself. The Supreme Court of the United States stated that instruction in Rodriguez de Quijas v. Shearson Lehman Brothers, 490 U.S. 477 (1989), where a court of appeals had anticipated the fall of an earlier arbitration holding. The Supreme Court of the United States repeated the instruction in Agostini v. Felton, 521 U.S. 203 (1997), cautioning lower courts against concluding that more recent cases had implicitly displaced an older holding that governed the dispute at hand. Obedience in that interval preserves hierarchy, because litigants receive a uniform rule until the Supreme Court of the United States changes the rule in a case such as Agostini v. Felton, 521 U.S. 203 (1997).

Q: Can Congress reverse a Supreme Court decision by passing a new law?

Congress can displace a decision of the Supreme Court of the United States that interprets a statute by amending the statute, and the Civil Rights Act of 1991 supplied a prominent example by revising rules from several employment decisions, including Patterson v. McLean Credit Union, 491 U.S. 164 (1989). Congress cannot reverse a constitutional holding by ordinary legislation, because Marbury v. Madison, 5 U.S. 137 (1803), established judicial review and City of Boerne v. Flores, 521 U.S. 507 (1997), rejected a congressional attempt to expand a constitutional right beyond the reading given by the Supreme Court of the United States. A constitutional amendment under Article V of the Constitution of the United States, effective in 1789, remains available, and the Eleventh Amendment, ratified in 1795, answered Chisholm v. Georgia, 2 U.S. 419 (1793), through that route.

Q: What is a plurality decision and why is its precedential value uncertain?

A plurality decision resolves a case for the Supreme Court of the United States when no single opinion attracts the support of five justices, so the judgment commands a majority while the explanation commands only a faction. The uncertainty begins with identification, because lawyers must determine which opinion, if any, states the governing rule that lower courts should extract from the fractured result in a case such as Apodaca v. Oregon, 406 U.S. 404 (1972). Marks v. United States, 430 U.S. 188 (1977), supplied the standard technique by directing attention to the position of the justices concurring in the judgment on the narrowest grounds. That technique fails to yield a clean answer when concurring opinions share no common reasoning, and panels of the United States Courts of Appeals have responded by treating some fractured decisions as binding only on closely similar facts.

Q: Why did the Court overrule Chevron deference in Loper Bright Enterprises v. Raimondo?

Chief Justice John Roberts, for the majority in Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), held that Chevron U.S.A. v. Natural Resources Defense Council, 467 U.S. 837 (1984), required a form of deference that the Administrative Procedure Act of 1946 did not authorize. Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), reasoned that a statute assigns courts the duty to decide questions of law and that ambiguity in a text does not transfer that duty to an agency. Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), also found the deference framework unworkable in practice, because the threshold inquiry into ambiguity had produced inconsistent results and repeated subsidiary tests since United States v. Mead Corporation, 533 U.S. 218 (2001). The majority in Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), treated Skidmore v. Swift and Company, 323 U.S. 134 (1944), as the proper measure of agency influence, under which an agency view persuades through quality of reasoning.

Q: What does it mean for a decision to be narrowed rather than overruled?

Narrowing a decision confines the reach of a holding to a smaller class of cases while leaving the named precedent standing, and the Supreme Court of the United States often narrows through distinguishing facts, limiting a rationale, or reading an earlier opinion in light of later doctrine. Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), narrowed Roe v. Wade, 410 U.S. 113 (1973), in practical effect by replacing the trimester structure with a viability standard, even though the joint opinion presented the move as preservation of a core holding. The cost of narrowing appears when litigants can no longer tell whether the surviving core of a precedent from Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), will govern a new dispute.

Q: How do justices signal that a precedent is in danger before it falls?

Justices signal danger through dissents that invite a future challenge, concurrences that narrow a holding to its facts, and majority opinions that criticize reasoning while stopping short of discarding the precedent under review. Justice Antonin Scalia used dissents in cases after Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), to argue that Roe v. Wade, 410 U.S. 113 (1973), should be discarded, and those opinions supplied later litigants with a roadmap for framing a challenge. Statements respecting a denial of certiorari perform a similar function, as the statement by Justice Clarence Thomas in Gamble v. United States, 587 U.S. 678 (2019), demonstrated by questioning broad claims made for stare decisis in a separate opinion in that litigation. A grant of certiorari on a question that asks expressly whether a named case should be overruled, as in Dobbs v. Jackson Women’s Health Organization, 597 U.S. 215 (2022), is the clearest signal of all.

Q: What is the difference between horizontal and vertical stare decisis?

Vertical stare decisis describes the obligation of a lower court to follow the decisions of a court above the lower court in the same judicial hierarchy, so a United States District Court must apply holdings of the Supreme Court of the United States and of the governing United States Court of Appeals. Rodriguez de Quijas v. Shearson Lehman Brothers, 490 U.S. 477 (1989), enforced the vertical form by reversing a lower court that had anticipated the abandonment of a governing holding. Horizontal stare decisis describes the weight a court gives to rulings of the same court, and for the Supreme Court of the United States that form operates as a policy of consistency rather than a binding command, as Payne v. Tennessee, 501 U.S. 808 (1991), explained. The distinction explains why the Supreme Court of the United States can discard Apodaca v. Oregon, 406 U.S. 404 (1972), in Ramos v. Louisiana, 590 U.S. 83 (2020), while no lower court could take that step.

Q: Why did Lawrence v. Texas overrule Bowers v. Hardwick after a short interval?

Justice Anthony Kennedy, writing for the majority in Lawrence v. Texas, 539 U.S. 558 (2003), held that Bowers v. Hardwick, 478 U.S. 186 (1986), had framed the claimed liberty interest too narrowly and had misread the history of laws governing private consensual conduct between adults. Lawrence v. Texas, 539 U.S. 558 (2003), reasoned that intervening decisions, including Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992), and Romer v. Evans, 517 U.S. 620 (1996), had weakened the doctrinal premises on which Bowers v. Hardwick, 478 U.S. 186 (1986), rested. Lawrence v. Texas, 539 U.S. 558 (2003), also considered a decision of the European Court of Human Rights in Dudgeon v. United Kingdom, decided in 1981, as evidence that the historical claims in Bowers v. Hardwick, 478 U.S. 186 (1986), overstated a consensus. Seventeen years separated the two decisions, a span the majority in Lawrence v. Texas, 539 U.S. 558 (2003), treated as sufficient when error and doctrinal erosion were both established.

Q: What happens to cases decided under a precedent after that precedent is overruled?

Cases still open on direct review receive the benefit of the new rule, while judgments that became final before the overruling generally remain in force, and collateral challenges face an additional barrier that depends on the nature of the new rule. Harper v. Virginia Department of Taxation, 509 U.S. 86 (1993), applied that allocation in civil litigation by extending a new rule to every case pending when the Supreme Court of the United States announced the rule. Griffith v. Kentucky, 479 U.S. 314 (1987), adopted the parallel position for criminal prosecutions still moving through direct appeal. Teague v. Lane, 489 U.S. 288 (1989), restricted the reach of new constitutional rules of criminal procedure on collateral review, and Edwards v. Vannoy, 593 U.S. 255 (2021), applied that restriction to deny relief based on Ramos v. Louisiana, 590 U.S. 83 (2020), to a prisoner whose conviction was final.