Minerva Surgical, Inc. v. Hologic, Inc.

594 U.S. 559, 210 L. Ed. 2d 689, 141 S. Ct. 2298

• Court: Supreme Court of the United States
• Decided: June 29, 2021
• Docket: 20-440
• Status: Published

Opinion

(Slip Opinion) OCTOBER TERM, 2020 1

Syllabus

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

SUPREME COURT OF THE UNITED STATES

MINERVA SURGICAL, INC. v. HOLOGIC, INC., ET AL.

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

No. 20–440. Argued April 21, 2021—Decided June 29, 2021 In the late 1990s, Csaba Truckai invented a device to treat abnormal uterine bleeding. The device, known as the NovaSure System, uses a moisture-permeable applicator head to destroy targeted cells in the uterine lining. Truckai filed a patent application and later assigned the application, along with any future continuation applications, to his company, Novacept, Inc. The PTO issued a patent for the device. No- vacept, along with its portfolio of patents and patent applications, was eventually acquired by respondent Hologic, Inc. In 2008, Truckai founded petitioner Minerva Surgical, Inc. There, he developed a sup- posedly improved device to treat abnormal uterine bleeding. Called the Minerva Endometrial Ablation System, the new device uses a moisture-impermeable applicator head to remove cells in the uterine lining. The PTO issued a patent, and the FDA approved the device for commercial sale. Meanwhile, Hologic filed a continuation application with the PTO, seeking to add claims to its patent for the NovaSure System. Hologic drafted one of its claims to encompass applicator heads generally, without regard to whether they are moisture perme- able. The PTO issued the altered patent in 2015. Hologic then sued Minerva for patent infringement. As relevant here, Minerva rejoined that Hologic’s patent was invalid because the newly added claim did not match the invention’s written description, which addresses applicator heads that are water permeable. In re- sponse, Hologic invoked the doctrine of assignor estoppel. Because Truckai had assigned the original patent application, Hologic argued, he and Minerva could not impeach the patent’s validity. The District Court agreed that assignor estoppel barred Minerva’s invalidity de- fense. The Court of Appeals for the Federal Circuit affirmed in rele- vant part. Minerva now asks this Court to abandon or narrow assignor 2 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

estoppel. Held: Assignor estoppel is well grounded in centuries-old fairness prin- ciples, and the Federal Circuit was right to uphold it. But assignor estoppel applies only when the assignor’s claim of invalidity contra- dicts explicit or implicit representations he made in assigning the pa- tent. Pp. 5–17. (a) Courts have long applied the doctrine of assignor estoppel to deal with inconsistent representations about a patent’s validity. The doc- trine got its start in late 18th-century England and crossed the Atlan- tic about a hundred years later. This Court first considered and ap- proved the doctrine in Westinghouse Elec. & Mfg. Co. v. Formica Insulation Co., 266 U. S. 342. The Court grounded the doctrine in a principle of fairness: “If one lawfully conveys to another a patented right,” Westinghouse reasoned, “fair dealing should prevent him from derogating from the title he has assigned.” Id., at 350. The Court made clear, however, that the doctrine has limits. Although the as- signor cannot assert invalidity in an infringement suit, he can argue about how to construe the patent’s claims. Id., at 350–351. The Court left for another day other questions about the doctrine’s scope, includ- ing how it would apply to the assignment of patent applications. Id., at 352–353. Pp. 5–9. (b) The Court rejects Minerva’s contention that assignor estoppel should be abandoned. Minerva’s first argument on that score—that Congress abrogated the doctrine in the Patent Act of 1952—is unper- suasive. Minerva relies on statutory language providing that “[i]nva- lidity” of the patent “shall be [a] defense[ ] in any action involving” in- fringement. 35 U. S. C. §282(b). According to Minerva, that language “instructs that invalidity must be available as a defense in every ac- tion,” thus leaving no room for assignor estoppel. Brief for Petitioner 17–18. But similar language appeared in the patent statute when the Court decided Westinghouse. Anyway, Minerva’s view is untenable be- cause it would foreclose applying in patent cases a whole host of com- mon-law preclusion doctrines—a broad result that would conflict with this Court’s precedents. See, e.g., SCA Hygiene Products Aktiebolag v. First Quality Baby Products, LLC, 580 U. S. ___, ___. And it would subvert congressional design, for Congress in 1952 “legislate[d] against a background of common-law adjudicatory principles,” includ- ing assignor estoppel. Astoria Fed. Sav. & Loan Assn. v. Solimino, 501 U. S. 104, 108. The Court also rejects Minerva’s view that two post-Westinghouse decisions have already interred assignor estoppel. In Scott Paper Co. v. Marcalus Mfg. Co., 326 U. S. 249, the Court did nothing more than decline to apply assignor estoppel in a novel and extreme circum- stance. The Court did not question—indeed, it restated—the “basic Cite as: 594 U. S. ____ (2021) 3

principle” of fairness on which the doctrine rests. Id., at 251. In Lear, Inc. v. Adkins, 395 U. S. 653, the Court considered and toppled a dif- ferent patent estoppel doctrine—licensee estoppel—but did not pur- port to decide the fate of assignor estoppel. To the contrary, the Court stated that the patent holder’s “equities” in the assignment context “were far more compelling than those presented in the typical licensing arrangement.” Id., at 664. Together, Scott Paper and Lear maintained assignor estoppel, but suggested that the doctrine needed to stay at- tached to its equitable moorings. Finally, the Court rejects Minerva’s claim that contemporary patent policy—specifically, the need to weed out bad patents—supports over- throwing assignor estoppel. Assignor estoppel reflects a demand for consistency in dealing with others. When a person sells his patent rights, he makes an (at least) implicit representation to the buyer that the patent at issue is valid. In later raising an invalidity defense, the assignor disavows that implied warranty. By saying one thing and then saying another, the assignor wants to profit doubly—by gaining both the price of assigning the patent and the continued right to use the invention it covers. That course of conduct by the assignor is unfair dealing. And the need to prevent such unfairness outweighs any loss to the public from leaving an invalidity defense to someone other than the assignor. Pp. 9–14. (c) Assignor estoppel comes with limits: it applies only when its un- derlying principle of fair dealing comes into play. That principle de- mands consistency in representations about a patent’s validity. When an assignor warrants that a patent claim is valid, his later denial of validity breaches norms of equitable dealing. But when the assignor has made neither explicit nor implicit representations in conflict with an invalidity defense, then there is no unfairness in its assertion—and so there is no ground for applying assignor estoppel. One example of non-contradiction is when an assignment occurs before an inventor can possibly make a warranty of validity as to specific patent claims. That situation arises in certain employment arrangements, when an em- ployee assigns to his employer patent rights in any future inventions he may develop during his employment. A second example is when a later legal development renders irrelevant the warranty given at the time of assignment. Third, and most relevant here, a post-assignment change in patent claims can remove the rationale for applying assignor estoppel. The last situation arises most often when an inventor as- signs a patent application, rather than an issued patent. There, the assignee may return to the PTO to enlarge the patent’s claims. As- suming that the new claims are materially broader than the old ones, the assignor did not warrant to the new claims’ validity. And if he made no such representation, then he can challenge the new claims in 4 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

litigation: Because there is no inconsistency in his positions, there is no estoppel. The Federal Circuit failed to recognize these boundaries. Minerva argued that estoppel should not apply because it was challenging a claim that was materially broader than the ones Truckai had assigned. The Federal Circuit declined to consider the alleged disparity, deeming “irrelevant” the question whether Hologic had expanded the assigned claims. But if Hologic’s new claim is materially broader than the ones Truckai assigned, then Truckai could not have warranted its validity in making the assignment. And without such a prior inconsistent rep- resentation, there is no basis for estoppel. The judgment of the Federal Circuit is therefore vacated, and the case is remanded for the Court of Appeals to address whether Hologic’s new claim is materially broader than the ones Truckai assigned. Pp. 14–17. 957 F. 3d 1256, vacated and remanded.

KAGAN, J., delivered the opinion of the Court, in which ROBERTS, C. J., and BREYER, SOTOMAYOR, and KAVANAUGH, JJ., joined. ALITO, J., filed a dissenting opinion. BARRETT, J., filed a dissenting opinion, in which THOMAS and GORSUCH, JJ., joined. Cite as: 594 U. S. ____ (2021) 1

Opinion of the Court

NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Wash- ington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.

SUPREME COURT OF THE UNITED STATES _________________

No. 20–440 _________________

MINERVA SURGICAL, INC., PETITIONER v. HOLOGIC, INC., ET AL. ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FEDERAL CIRCUIT [June 29, 2021]

JUSTICE KAGAN delivered the opinion of the Court. In Westinghouse Elec. & Mfg. Co. v. Formica Insulation Co., 266 U. S. 342, 349 (1924), this Court approved the “well settled” patent-law doctrine of “assignor estoppel.” That doctrine, rooted in an idea of fair dealing, limits an inven- tor’s ability to assign a patent to another for value and later contend in litigation that the patent is invalid. The ques- tion presented here is whether to discard this century-old form of estoppel. Continuing to see value in the doctrine, we decline to do so. But in upholding assignor estoppel, we clarify that it reaches only so far as the equitable principle long understood to lie at its core. The doctrine applies when, but only when, the assignor’s claim of invalidity con- tradicts explicit or implicit representations he made in as- signing the patent. I Inventors look to the patent system to obtain valuable rights. A typical patent application, filed with the U. S. Pa- tent and Trademark Office (PTO), includes a written de- scription and drawing of the invention and one or more claims particularly setting out the invention’s scope. See 2 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

35 U. S. C. §§111–113. The application also usually con- tains an inventor’s oath—a statement attesting that the ap- plicant is “the original inventor” of the “claimed invention,” so that he is entitled to the patent sought. §115. If the PTO decides that the invention meets the “conditions for patent- ability”—mainly, that the invention is useful, novel, and non-obvious—it will issue a patent to the inventor. See §§101–103. That award gives the inventor the right to ex- clude others from making, using, or selling the invention until the patent expires (currently, 20 years after the appli- cation date). See §154. The invention sparking this lawsuit is a device to treat abnormal uterine bleeding, a medical condition affecting many millions of women. Csaba Truckai, a founder of the company Novacept, Inc., invented the device—called the NovaSure System—in the late 1990s. He soon afterward filed a patent application, and assigned his interest in the application—as well as in any future “continuation applica- tions”—to Novacept. 1 The NovaSure System, as described in Truckai’s patent application, uses an applicator head to destroy targeted cells in the uterine lining. To avoid unin- tended burning or ablation (tissue removal), the head is “moisture permeable,” meaning that it conducts fluid out of the uterine cavity during treatment. The PTO issued a pa- tent, and in 2001 the Food and Drug Administration (FDA) approved the device for commercial distribution. But nei- ther Truckai nor Novacept currently benefits from the NovaSure System patent. In 2004, Novacept sold its assets, including its portfolio of patents and patent applications, to —————— 1 A continuation application enables an inventor to add to or modify

the claims set out in his original application. See 35 U. S. C. §120; Man- ual of Patent Examining Procedure §201.07 (9th ed., June 2020). But the continuation application may not materially change the written de- scription of the invention. See Manual of Patent Examining Procedure §211.05. So the new or altered claims must align with the original de- scription. Ibid.; see 35 U. S. C. §112. Cite as: 594 U. S. ____ (2021) 3

another company (netting Truckai individually about $8 million). And in another sale, in 2007, respondent Hologic, Inc. acquired all patent rights in the NovaSure System. To- day, Hologic sells that device throughout the United States. Not through with inventing, Truckai founded in 2008 pe- titioner Minerva Surgical, Inc. There, he developed a sup- posedly improved device to treat abnormal uterine bleed- ing. Called the Minerva Endometrial Ablation System, the device (like the NovaSure System) uses an applicator head to remove cells in the uterine lining. But the new device, relying on a different way to avoid unwanted ablation, is “moisture impermeable”: It does not remove any fluid dur- ing treatment. The PTO issued a patent for the device, and in 2015 the FDA approved it for commercial sale. Meanwhile, in 2013, Hologic filed a continuation applica- tion requesting to add claims to its patent for the NovaSure System. Aware of Truckai’s activities, Hologic drafted one of those claims to encompass applicator heads generally, without regard to whether they are moisture permeable. The PTO in 2015 issued the altered patent as requested. A few months later, Hologic sued Minerva for patent in- fringement. Minerva rejoined that its device does not in- fringe. But more relevant here, it also asserted that Ho- logic’s amended patent is invalid. The essential problem, according to Minerva, is that the new, broad claim about applicator heads does not match the invention’s description, which addresses their water-permeability. See Defendant Minerva’s Opening Brief in Support of Its Motion for Partial Summary Judgment in No. 15–cv–1031 (Del.), Doc. 300, pp. 8–9, 13–15; see also Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U. S. 722, 736 (2002) (“What is claimed by the patent application must be the same” as what is described). In response, Hologic invoked the doc- trine of assignor estoppel. Because Truckai assigned the original patent application, Hologic argued, he and Minerva (essentially, his alter-ego) could not impeach the patent’s 4 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

validity. The District Court agreed that assignor estoppel barred Minerva’s invalidity defense, and also ruled that Mi- nerva had infringed Hologic’s patent. See 325 F. Supp. 3d 507, 524–525, 532 (Del. 2018). At a trial on damages, a jury awarded Hologic about $5 million. The Court of Appeals for the Federal Circuit mainly up- held the judgment, focusing on assignor estoppel. The court first “decline[d] Minerva’s invitation to ‘abandon [that] doc- trine.’ ” 957 F. 3d 1256, 1267 (2020). Citing both this Court’s precedents and equitable principles, the court af- firmed the doctrine’s “continued vitality.” Id., at 1268. An assignor, the court stated, “should not be permitted to sell something and later to assert that what was sold is worth- less, all to the detriment of the assignee.” Id., at 1265. The assignor makes an “implicit representation” that the rights “he is assigning (presumably for value) are not worthless.” Ibid. It would “work an injustice,” the court reasoned, to “allow the assignor to make that representation at the time of assignment (to his advantage) and later to repudiate it (again to his advantage).” Ibid. (quoting Diamond Scien- tific Co. v. Ambico, Inc., 848 F. 2d 1220, 1224 (CA Fed. 1988)). The court then applied assignor estoppel to bar Truckai and Minerva from raising an invalidity defense. Here, the court rejected Minerva’s argument that because “Hologic broadened the claims” after “Truckai’s assign- ment,” it would “be unfair to block Truckai (or Minerva) from challenging the breadth of those claims.” 957 F. 3d, at 1268. Relying on circuit precedent, the court deemed it “ir- relevant that, at the time of the assignment, the inventor’s patent application[ ] w[as] still pending” and that the as- signee “may have later amended the claims” without the in- ventor’s input. Ibid. (quoting Diamond Scientific, 848 F. 2d, at 1226). We granted certiorari, 592 U. S. ___ (2021), to consider the important issues raised in the Federal Circuit’s judg- ment. Assignor estoppel, we now hold, is well grounded in Cite as: 594 U. S. ____ (2021) 5

centuries-old fairness principles, and the Federal Circuit was right to uphold it. But the court failed to recognize the doctrine’s proper limits. The equitable basis of assignor es- toppel defines its scope: The doctrine applies only when an inventor says one thing (explicitly or implicitly) in assign- ing a patent and the opposite in litigating against the pa- tent’s owner. II Courts have long applied the doctrine of assignor estoppel to deal with inconsistent representations about a patent’s validity. The classic case (different in certain respects from the one here) begins with an inventor who both applies for and obtains a patent, then assigns it to a company for value. Later, the inventor/assignor joins a competitor business, where he develops a similar—and possibly infringing— product. When the assignee company sues for infringe- ment, the assignor tries to argue—contrary to the (explicit or implicit) assurance given in assigning the patent—that the invention was never patentable, so the patent was never valid. That kind of about-face is what assignor estop- pel operates to prevent—or, in legalese, estop. As one of the early American courts to use the doctrine held: The as- signor is not “at liberty to urge [invalidity] in a suit upon his own patent against a party who derives title to that pa- tent through him.” Woodward v. Boston Lasting Mach. Co., 60 F. 283, 284 (CA1 1894). Or as the Federal Circuit held in modern times: The assignor’s explicit or “implicit repre- sentation” that the patent he is assigning is “not worth- less . . . deprive[s] him of the ability to challenge later the [patent’s] validity.” Diamond Scientific, 848 F. 2d, at 1224. Assignor estoppel got its start in late 18th-century Eng- land and crossed the Atlantic about a hundred years later. In the first recorded case, Lord Kenyon found that a patent assignor “was by his own oath and deed estopped” in an in- fringement suit from “attempt[ing] to deny his having had 6 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

any title to convey.” Oldham v. Langmead (1789), as de- scribed in J. Davies, Collection of the Most Important Cases Respecting Patents of Invention and the Rights of Patent- ees 442 (1816); see Hayne v. Maltby, 3 T. R. 439, 441, 100 Eng. Rep. 665, 666 (K. B. 1789) (recognizing the Oldham holding). That rule took inspiration from an earlier doc- trine—estoppel by deed—applied in real property law to prevent a conveyor of land from later asserting that he had lacked good title at the time of sale. See 2 E. Coke, The First Part of the Institutes of Laws of England 352a (Har- grave & Butler eds., 19th ed. 1832) (1628). Lord Kenyon’s new patent formulation of the doctrine grew in favor throughout the 1800s as an aspect of fair dealing: When “the Defendant sold and assigned th[e] patent to the Plain- tiffs as a valid one,” it “does not lie in his mouth to say that the patent is not good.” Chambers v. Crichley, 33 Beav. 374, 376, 55 Eng. Rep. 412 (1864); see Walton v. Lavater, 8 C. B. N. S. 162, 187, 141 Eng. Rep. 1127, 1137 (C. P. 1860) (“The defendant, who has received a large sum for the sale of this patent, ought not to be allowed to raise any question as to its validity”). The earliest American decision applying the doctrine dates from 1880. See Faulks v. Kamp, 3 F. 898 (CC SDNY). Within a decade or two, the doctrine was “so well established and generally accepted that citation of author- ity is useless.” Griffith v. Shaw, 89 F. 313, 315 (CC SD Iowa 1893); see 2 W. Robinson, Law of Patents for Useful Inven- tions §787 (1890) (collecting cases). This Court first considered—and unanimously ap- proved—assignor estoppel in 1924, in Westinghouse v. For- mica. Speaking through Chief Justice Taft, the Court ini- tially invoked the doctrine’s uniform acceptance in the lower courts. The first decision applying assignor estoppel, the Court recounted, was soon “followed by a myriad.” 266 U. S., at 349. “[L]ater cases in nearly all the Circuit Courts of Appeal” were “to the same point” as the first, adding up Cite as: 594 U. S. ____ (2021) 7

to a full “forty-five years of judicial consideration and con- clusion.” Ibid. Such a “well settled” rule, in the Court’s view, should “not [be] lightly disturb[ed].” Ibid. And so it was not disturbed, lightly or otherwise. Rather, the Court added its own voice to that pre-existing “myriad.” We an- nounced that an assignor “is estopped to attack” the “valid- ity of a patented invention which he has assigned.” Ibid. “As to the rest of the world,” the Court explained, “the pa- tent may have no efficacy”; but “the assignor can not be heard to question” the assignee’s rights in what was con- veyed. Ibid. Westinghouse, like its precursor decisions, grounded as- signor estoppel in a principle of fairness. “If one lawfully conveys to another a patented right,” the Court reasoned, “fair dealing should prevent him from derogating from the title he has assigned.” Id., at 350. After all, the “grantor purports to convey the right to exclude others”; how can he later say, given that representation, that the grantee in fact possesses no such right? Ibid. The Court supported that view of equity by referring to estoppel by deed. See supra, at 6. Under that doctrine, the Court explained, “a grantor of a deed of land” cannot “impeach[ ] the effect of his solemn act” by later claiming that the grantee’s title is no good. Westinghouse, 266 U. S., at 350. “The analogy” was “clear”: There was “no reason why the principles of estoppel by deed should not apply to [the] assignment of a patent right.” Id., at 348, 350. In the latter context too, the Court held, the assignor could not fairly “attack” the validity of a right he had formerly sold. Id., at 349. After thus endorsing assignor estoppel, the Court made clear that the doctrine has limits. Although the assignor cannot assert in an infringement suit that the patent is in- valid, the Court held that he can argue about how to con- strue the patent’s claims. Here, the Court addressed the role in patent suits of prior art—the set of earlier inventions 8 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

(and other information) used to decide whether the speci- fied invention is novel and non-obvious enough to merit a patent. Id., at 350. “Of course,” the Court said, the assignor cannot use prior art in an infringement suit “to destroy the patent,” because he “is estopped to do this.” Id., at 351. But he can use prior art to support a narrow claim construc- tion—to “construe and narrow the claims of the patent, con- ceding their validity.” Id., at 350–351. “Otherwise,” the Court explained, a judge “would be denied” the “most satis- factory means” of “reaching a just conclusion” about the pa- tent’s scope—a conclusion needed to resolve the infringe- ment charge. Id., at 350–351. “The distinction” thus established, the Court thought, “may be a nice one, but seems to be workable.” Id., at 351. And, indeed, the Court applied it to decide the case at hand for the assignor, finding that he had not infringed the properly narrowed claim. 2 Finally, the Court left for another day several other ques- tions about the contours of assignor estoppel. One con- cerned privity: When was an assignor so closely affiliated with another party that the latter would also be estopped? See id., at 355. Another related to consideration: What if an assignor had received only a nominal amount of money for transferring the patent? See ibid. But the question that most interested the Court was whether estoppel should op- erate differently if the assignment was not of a granted pa- tent but of a patent application—as in fact was true in that case. The Court saw a possible distinction between the two. —————— 2 The limit set out in Westinghouse is not often invoked today, because

modern courts construe patent claims (as they construe statutes) mainly by reference to their text. See Lemley, Rethinking Assignor Estoppel, 54 Houston L. Rev. 513, 523 (2016). Only when the claims are “still ambig- uous,” after consideration of text and canons, will a court think about narrowing a claim by reference to prior art. Phillips v. AWH Corp., 415 F. 3d 1303, 1327 (CA Fed. 2005) (internal quotation marks omitted). The critical point for our purpose is that even while affirming the assignor estoppel doctrine, the Court made clear that it did not always bar assign- ors from effectively defending against infringement suits. Cite as: 594 U. S. ____ (2021) 9

In a patent application, the Court began, the inventor “swor[e] to” a particular “specification.” Id., at 352. But the exact rights at issue were at that point “inchoate”—not “cer- tainly defined.” Ibid. And afterward, the Court (presci- ently) observed, the claims might be “enlarge[d]” at “the in- stance of the assignee” beyond what the inventor had put forward. Id., at 353. That might weaken the case for estop- pel. But the Court decided not to decide the issue, given its holding that the assignor had not infringed the (narrowed) patent claim anyway. III Minerva’s main argument here, as in the Federal Circuit, is that “assignor estoppel should be eliminated”—and in- deed has been already. Brief for Petitioner 17. We reject that view. The doctrine has lasted for many years, and we continue to accept the fairness principle at its core. Mi- nerva’s back-up contention is that assignor estoppel “should be constrained.” Id., at 41. On that score, we find that the Federal Circuit has applied the doctrine too expansively. Today, we clarify the scope of assignor estoppel, including in the way Westinghouse suggested. A In its quest to abolish assignor estoppel, Minerva lodges three main arguments. The first two offer different reasons for why the doctrine is already defunct: because Congress repudiated it in the Patent Act of 1952 and because, even if not, this Court’s post-Westinghouse cases “leave no room for the doctrine to continue.” Brief for Petitioner 20. The third, by contrast, is a present-day policy claim: that assignor es- toppel “imposes” too high a “barrier to invalidity chal- lenges” and so keeps bad patents alive. Id., at 38. (The principal dissent essentially endorses the first two argu- ments, but not the third. See post, at 1, 4–6, 11 (opinion of BARRETT, J.).) 10 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

On the first point, we do not agree that the Patent Act of 1952 abrogated assignor estoppel. The statutory language Minerva relies on provides that “[i]nvalidity” of the patent “shall be [a] defense[ ] in any action involving” infringe- ment. 35 U. S. C. §282(b). According to Minerva, that lan- guage “instructs that invalidity must be available as a de- fense in every [infringement] action,” thus “leav[ing] no room for assignor estoppel.” Brief for Petitioner 17–18 (em- phasis in original). But to begin with, similar language, en- titling a defendant to plead invalidity in any infringement action, was in the patent statute when Westinghouse was decided. See Patent Act of 1897, ch. 391, §2, 29 Stat. 692 (“In any action for infringement the defendant may plead” invalidity). And anyway, Minerva’s view is untenable be- cause it would foreclose applying in patent cases a whole host of common-law preclusion doctrines—not just assignor estoppel, but equitable estoppel, collateral estoppel, res ju- dicata, and law of the case. That broad result would conflict with this Court’s precedents. See, e.g., SCA Hygiene Prod- ucts Aktiebolag v. First Quality Baby Products, LLC, 580 U. S. ___, ___ (2017) (slip op., at 16) (recognizing equitable estoppel in a patent suit). And it would subvert congres- sional design. For Congress “legislate[s] against a back- ground of common-law adjudicatory principles,” and it “ex- pect[s]” those principles to “apply except when a statutory purpose to the contrary is evident.” Astoria Fed. Sav. & Loan Assn. v. Solimino, 501 U. S. 104, 108 (1991) (internal quotation marks omitted). Assignor estoppel was by 1952 just such a background principle of patent adjudication, and Congress gave no indication of wanting to terminate it or disturb its development. Nor has Congress done so since that time. We likewise do not accept Minerva’s view that two of our post-Westinghouse decisions have already interred assignor estoppel. According to Minerva (quoting the case’s dissent), Cite as: 594 U. S. ____ (2021) 11

Scott Paper Co. v. Marcalus Mfg. Co. “eliminated any justi- fication for assignor estoppel and ‘repudiated’ the doctrine.” Reply Brief 13 (quoting 326 U. S. 249, 264 (1945) (Frank- furter, J., dissenting)). And if that were not enough, Mi- nerva continues, our decision in Lear, Inc. v. Adkins, 395 U. S. 653 (1969), also “eviscerated any basis for assignor es- toppel.” Reply Brief 13. But we think the words “elimi- nated,” “repudiated,” and “eviscerated” are far off. Scott Pa- per and Lear in fact retained assignor estoppel; all they did was police the doctrine’s boundaries (just as Westinghouse did and we do today). Whatever a worked-up dissent charged, Scott Paper did nothing more than decline to apply assignor estoppel in a novel and extreme circumstance. The petitioner in Scott Paper made the same ask Minerva does here: to abolish the Westinghouse rule. The Court expressly declined that re- quest. See 326 U. S., at 254. And it restated the “basic principle” animating assignor estoppel, describing it as “one of good faith, that one who has sold his invention may not, to the detriment of the purchaser, deny the existence of that which he has sold.” Id., at 251. The Court, to be sure, de- clined to apply the doctrine in the case before it. There, estoppel would have prevented the assignor from making a device on which the patent had expired—a device, in other words, that had already entered the public domain. The Court could not find any precedent for applying estoppel in that situation. See id., at 254. And the Court thought that doing so would carry the doctrine too far, reasoning that the public’s interest in using an already-public invention out- weighs the “interest in private good faith.” Id., at 256–257. But the Court did not question—again, it reaffirmed—the principle of fairness on which assignor estoppel rests in more common cases, where the assignee is not claiming to control a device unequivocally part of the public domain. See id., at 251. In those cases, the doctrine remained intact. Lear gives Minerva still less to work with. In that case, 12 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

the Court considered and toppled a different patent estop- pel doctrine. Called licensee estoppel, it barred (as its name suggests) a patent licensee from contesting the validity of the patent on a device he was paying to use. Minerva’s basic claim is that as goes one patent estoppel rule, so goes another. Brief for Petitioner 23–25. But Lear did not pur- port to decide the fate of the separate assignor estoppel doc- trine. To the contrary, the Court stated that the patent holder’s “equities” in the assignment context “were far more compelling than those presented in the typical licensing ar- rangement.” 395 U. S., at 664. And so they are. As ex- plained earlier, assignor estoppel rests on the idea that the assignor has made an explicit or implicit representation about the patent’s validity, and received some kind of pay- ment in return. See supra, at 5–6. No rationale of that kind supports licensee estoppel. The licensee is a buyer of patent rights, not a seller. He has given no assurances of the pa- tent’s worth. All he has done is purchase the right to use a patented device—which, if the patent is invalid, he need not have done. Lear’s refusal to bar a licensee’s claim of inva- lidity showed that the Court was alert to “the important public interest in permitting full and free competition in the use of ideas”—and so would not always apply patent estop- pel doctrines. 395 U. S., at 670. But that does not mean, at the other extreme, always rejecting those doctrines. Lear counseled careful attention to the equities at stake in dis- crete patent contexts—and expressly distinguished as- signor from licensee estoppel. In sum, Scott Paper and Lear left Westinghouse right about where they found it—as a bounded doctrine designed to prevent an inventor from first selling a patent and then contending that the thing sold is worthless. Westinghouse saw that about-face as unfair; Scott Paper and Lear never questioned that view. At the same time, Westinghouse re- alized that assignor estoppel has limits: Even in approving the doctrine, the Court made clear that not every assignor Cite as: 594 U. S. ____ (2021) 13

defense in every case would fall within its scope. See supra, at 7–8. Scott Paper and Lear adopted a similar stance. They maintained assignor estoppel, but suggested (if in dif- ferent ways) that the doctrine needed to stay attached to its equitable moorings. The three decisions together thus show not the doctrinal “eviscerat[ion]” Minerva claims, Reply Brief 13, but only the kind of doctrinal evolution typical of common-law rules. Finally, we do not think, as Minerva claims, that contem- porary patent policy—specifically, the need to weed out bad patents—supports overthrowing assignor estoppel. In re- jecting that argument, we need not rely on stare decisis: “[C]orrect judgments have no need for that principle to prop them up.” Kimble v. Marvel Entertainment, LLC, 576 U. S. 446, 455 (2015). And we continue to think the core of as- signor estoppel justified on the fairness grounds that courts applying the doctrine have always given. Assignor estop- pel, like many estoppel rules, reflects a demand for con- sistency in dealing with others. See H. Herman, The Law of Estoppel §3 (1871) (“An estoppel is an obstruction or bar to one’s alleging or denying a fact contrary to his own pre- vious action, allegation or denial”). When a person sells his patent rights, he makes an (at least) implicit representation to the buyer that the patent at issue is valid—that it will actually give the buyer his sought-for monopoly. 3 In later —————— 3 Recognizing this implicit representation is particularly appropriate

given the patent law’s demand for honesty from patent applicants. In applying for a patent, the inventor must ordinarily submit an oath—a statement attesting that he is “the original inventor” of the “claimed in- vention.” §115(b)(2); see supra, at 2. And the inventor must comply with “a duty of candor and good faith” in the patent process, including “a duty to disclose” to the PTO all information he knows “to be material to pa- tentability.” 37 CFR §1.56(a) (2020); see §1.63(c). An inventor present- ing an application to the PTO thus states his good-faith belief that his claims are patentable—that they will result in a valid patent. When the inventor then assigns those claims to another, he effectively incorporates that assurance. 14 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

raising an invalidity defense, the assignor disavows that implied warranty. And he does so in service of regaining access to the invention he has just sold. As the Federal Cir- cuit put the point, the assignor wants to make a “represen- tation at the time of assignment (to his advantage) and later to repudiate it (again to his advantage).” Diamond Scientific, 848 F. 2d, at 1224; see supra, at 4. By saying one thing and then saying another, the assignor wants to profit doubly—by gaining both the price of assigning the patent and the continued right to use the invention it covers. That course of conduct by the assignor strikes us, as it has struck courts for many a year, as unfair dealing—enough to out- weigh any loss to the public from leaving an invalidity de- fense to someone other than the assignor. 4 B Still, our endorsement of assignor estoppel comes with limits—true to the doctrine’s reason for being. Just as we guarded the doctrine’s boundaries in the past, see supra, at 7–8, 11–13, so too we do so today. Assignor estoppel should apply only when its underlying principle of fair dealing comes into play. That principle, as explained above, de- mands consistency in representations about a patent’s va- lidity: What creates the unfairness is contradiction. When an assignor warrants that a patent is valid, his later denial of validity breaches norms of equitable dealing. And the original warranty need not be express; as we have ex- plained, the assignment of specific patent claims carries with it an implied assurance. See supra, at 13. But when

—————— 4 Even beyond promoting fairness, assignor estoppel furthers some pa-

tent policy goals. Assignors are especially likely infringers because of their knowledge of the relevant technology. By preventing them from raising an invalidity defense in an infringement suit, the doctrine gives assignees confidence in the value of what they have purchased. That raises the price of patent assignments, and in turn may encourage inven- tion. Cite as: 594 U. S. ____ (2021) 15

the assignor has made neither explicit nor implicit repre- sentations in conflict with an invalidity defense, then there is no unfairness in its assertion. And so there is no ground for applying assignor estoppel. One example of non-contradiction is when the assign- ment occurs before an inventor can possibly make a war- ranty of validity as to specific patent claims. Consider a common employment arrangement. An employee assigns to his employer patent rights in any future inventions he develops during his employment; the employer then decides which, if any, of those inventions to patent. In that sce- nario, the assignment contains no representation that a pa- tent is valid. How could it? The invention itself has not come into being. See Lemley, Rethinking Assignor Estop- pel, 54 Houston L. Rev. 513, 525–527 (2016). And so the employee’s transfer of rights cannot estop him from alleging a patent’s invalidity in later litigation. A second example is when a later legal development ren- ders irrelevant the warranty given at the time of assign- ment. Suppose an inventor conveys a patent for value, with the warranty of validity that act implies. But the governing law then changes, so that previously valid patents become invalid. The inventor may claim that the patent is invalid in light of that change in law without contradicting his ear- lier representation. What was valid before is invalid today, and no principle of consistency prevents the assignor from saying so. Most relevant here, another post-assignment develop- ment—a change in patent claims—can remove the ra- tionale for applying assignor estoppel. Westinghouse itself anticipated this point, which arises most often when an in- ventor assigns a patent application, rather than an issued patent. As Westinghouse noted, “the scope of the right con- veyed in such an assignment” is “inchoate”—“less certainly defined than that of a granted patent.” 266 U. S., at 352– 353; see supra, at 9. That is because the assignee, once he 16 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

is the owner of the application, may return to the PTO to “enlarge[ ]” the patent’s claims. 266 U. S., at 353; see 35 U. S. C. §120; 37 CFR §1.53(b). And the new claims result- ing from that process may go beyond what “the assignor in- tended” to claim as patentable. 266 U. S., at 353. Westing- house did not need to resolve the effects of such a change, but its liberally dropped hints—and the equitable basis for assignor estoppel—point all in one direction. Assuming that the new claims are materially broader than the old claims, the assignor did not warrant to the new claims’ va- lidity. And if he made no such representation, then he can challenge the new claims in litigation: Because there is no inconsistency in his positions, there is no estoppel. The lim- its of the assignor’s estoppel go only so far as, and not be- yond, what he represented in assigning the patent applica- tion. The Federal Circuit, in both its opinion below and prior decisions, has failed to recognize those boundaries. Mi- nerva (recall, Truckai’s alter-ego) argued to the court that estoppel should not apply because it was challenging a claim that was materially broader than the ones Truckai had assigned. But the court declined to consider that al- leged disparity. Citing circuit precedent, the court held it “irrelevant” whether Hologic had expanded the assigned claims: Even if so, Minerva could not contest the new claim’s validity. 957 F. 3d, at 1268 (quoting Diamond Sci- entific, 848 F. 2d, at 1226); see supra, at 4. For the reasons given above, that conclusion is wrong. If Hologic’s new claim is materially broader than the ones Truckai assigned, then Truckai could not have warranted its validity in mak- ing the assignment. And without such a prior inconsistent representation, there is no basis for estoppel. We remand this case to the Federal Circuit to now ad- dress what it thought irrelevant: whether Hologic’s new claim is materially broader than the ones Truckai assigned. The parties vigorously disagree about that issue. In Cite as: 594 U. S. ____ (2021) 17

Truckai’s view, the new claim expanded on the old by cov- ering non-moisture-permeable applicator heads. In Ho- logic’s view, the claim matched a prior one that Truckai had assigned. Resolution of that issue in light of all relevant evidence will determine whether Truckai’s representations in making the assignment conflict with his later invalidity defense—and so will determine whether assignor estoppel applies. IV This Court recognized assignor estoppel a century ago, and we reaffirm that judgment today. But as the Court rec- ognized from the beginning, the doctrine is not limitless. Its boundaries reflect its equitable basis: to prevent an as- signor from warranting one thing and later alleging an- other. Assignor estoppel applies when an invalidity defense in an infringement suit conflicts with an explicit or implicit representation made in assigning patent rights. But absent that kind of inconsistency, an invalidity defense raises no concern of fair dealing—so assignor estoppel has no place. For these reasons, we vacate the judgment of the Federal Circuit and remand the case for further proceedings con- sistent with this opinion. It is so ordered. Cite as: 594 U. S. ____ (2021) 1

ALITO, J., dissenting

MINERVA SURGICAL, INC., PETITIONER v. HOLOGIC, INC., ET AL. ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FEDERAL CIRCUIT [June 29, 2021]

JUSTICE ALITO, dissenting. We granted review in this case to decide whether the doc- trine of assignor estoppel bars petitioner from challenging the validity of a patent indirectly assigned to respondents, and I do not see how we can answer that question without deciding whether Westinghouse Elec. & Mfg. Co. v. Formica Insulation Co., 266 U. S. 342 (1924), which recognized as- signor estoppel, should be overruled. Both the majority and the principal dissent go to great lengths to avoid that ques- tion, but in my judgment, their efforts are unsuccessful. The majority says it has no need to invoke precedent, see ante, at 13, but without that support, the majority’s holding cannot stand. Not one word in the patent statutes supports assignor estoppel, and the majority does not claim other- wise. “[T]his Court [doesn’t] usually read into statutes words that aren’t there,” Romag Fasteners, Inc. v. Fossil Group, Inc., 590 U. S. ___, ___ (2020) (slip op., at 3), but that is just what the majority has done in this case. With so little support for its reasoning, it is more than a little surprising that the majority forswears reliance on precedent. See ante, at 13. Not too long ago, in Kimble v. Marvel Entertainment, LLC, 576 U. S. 446 (2015), another case involving a judicially created rule of patent law, the Court applied a “superpowered form of stare decisis.” Id., 2 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

at 458. Yet the majority refuses to wield the nearly impreg- nable Kimble shield. Instead, it adopts a text-blind method of statutory interpretation with which I cannot possibly agree. The Court’s evasion of stare decisis is fully matched by the principal dissent. That opinion sees no need to address stare decisis because, in its view, Westinghouse has not been a precedent for the past 69 years. According to the principal dissent, Westinghouse was abrogated by the Patent Act of 1952. It reasons as follows: Westinghouse interpreted the Patent Act of 1870, post, at 1 (opinion of BARRETT, J.); be- cause that Act was superseded by the Patent Act of 1952, we must decide whether the new Act “ratified” Westing- house, post, at 1; and in order to show that Westinghouse was ratified, the defenders of assignor estoppel must per- suade us that “(1) the interpretation [adopted in Westing- house was] so well settled that we can ‘presume Congress knew of and endorsed it’ at the time of the reenactment, and (2) the statute [was] reenacted without material change,” post, at 2–3 (quoting Jama v. Immigration and Customs Enforcement, 543 U. S. 335, 349 (2005)). This reasoning is unprecedented and troubling. To start, it is quite misleading to suggest that Westinghouse was based on an interpretation of the 1870 Patent Act. Neither of the two statutory provisions Westinghouse mentioned said anything that supported the Court’s decision, and the Court did not claim otherwise. 1 Instead, the decision rested on different grounds. It relied on the principle of “fair deal- ing,” an analogy to the doctrine of estoppel by deed (a fea- ture of the law of real property), and perhaps most im- portantly, a body of lower court case law. 266 U. S., at 348– 352. If Westinghouse had been based on an interpretation —————— 1 One provision permitted the assignment of patent rights, see West-

inghouse, 266 U. S., at 348–349, and the other, which authorized the granting of a patent to an assignee, was discussed in relation to the scope of the doctrine of assignor estoppel, not its existence, see id., at 352. Cite as: 594 U. S. ____ (2021) 3

of language in the 1870 Act and if the 1952 Act had changed that language, there might be a basis for finding abroga- tion. 2 But that is not the situation here. After starting with this flawed premise, the principal dis- sent adopts an ill-suited standard for determining whether one of our precedents has been abrogated by Congress. It applies a rule of interpretation that is patently designed for a different purpose, i.e., determining whether congressional reenactment of a statute should be understood as a ratifi- cation of a preexisting body of lower court case law. 3 That was the issue in Jama, 543 U. S., at 349–352, the case the principal dissent quotes, and that is why the rule asks whether the interpretation in question is “well settled.” That question makes sense as applied to a body of lower court cases, but what does that mean with respect to one of our precedents? I would think that endorsement by a ma- jority of this Court is consensus enough. Suggesting that a rule announced in a decision of this Court can cease to be a precedent if it is not “well settled” is very strange. Equally strange is the question whether we can presume that Congress “knew of ” a decision of this Court interpret- ing a statutory provision that it reenacts. Perhaps it is hu- bris, but we have often presumed that Congress is aware of our decisions. See, e.g., Ryan v. Valencia Gonzales, 568

—————— 2 By the same token, if Congress had again used that particular lan-

guage without change, there might be a basis for finding ratification. See A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 322 (2012) (Reading Law) (prior-construction canon applies when a “word or phrase has been authoritatively interpreted [and] a later ver- sion of that act perpetuat[es] the wording”). 3 This is not to say that Congress cannot pass statutes meant to incor-

porate this Court’s interpretations of specific statutory language. Com- pare post, at 3, n. 3. It obviously can. What is unusual is not the idea that Congress can ratify this Court’s decisions but instead the applica- tion of a test designed to assess the ratification of lower court decisions to assess the abrogation of a decision of this Court. 4 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

U. S. 57, 66 (2013) (“We normally assume that, when Con- gress enacts statutes, it is aware of relevant judicial prece- dent” (internal quotation marks omitted)). When we reach the final part of the principal dissent’s test for abrogating one of our precedents—whether the stat- ute in question was “reenacted without material change”— we encounter further problems. The adjective “material” can mean “[h]aving some logical connection with the conse- quential facts.” Black’s Law Dictionary 1170 (11th ed. 2019). Thus, the principal dissent can be read to say that a decision of this Court interpreting a statutory provision is abrogated whenever a change in the language of the statute provides some degree of support for a different interpreta- tion. The principal dissent cites no case in which we have set the abrogation standard so low. And this new “material change” standard for abrogating one of our precedents is a marked departure from the standard advocated in one of the principal dissent’s chosen treatises. See Reading Law 331 (“Legislative revision of law clearly established by judi- cial opinion ought to be by express language or by unavoid- ably implied contradiction”). 4 The principal dissent is forced to adopt this new, low standard because it could not otherwise muster any sort of argument for abrogation. The principal dissent cannot identify anything in the 1952 Act that does away with the judge-made doctrine of assignor estoppel “by express lan- guage or by unavoidably implied contradiction.” Ibid. It cites 35 U. S. C. §282(b), which states that the invalidity of

—————— 4 The principal dissent responds by noting that this passage in Reading

Law sets out the “ ‘express language’ ” or “ ‘unavoidably implied contra- diction’ ” standard as part of “an entirely different canon.” Post, at 7, n. 4. That is precisely the point. The passage discusses the standard that should be used when an “authoritative judicial holding” is “cast in doubt and subjected to challenge” by changes to a statutory scheme. Reading Law 331. The rule applied by the principal dissent, on the other hand, is not meant to assess congressional abrogation of our precedents. Cite as: 594 U. S. ____ (2021) 5

a patent is a defense “in any action involving the validity or infringement of a patent.” See post, at 1. But as the major- ity notes, the patent laws contained similar language when Westinghouse was decided. Ante, at 10. The only modification made by the 1952 Act that the prin- cipal dissent claims has any logical connection with as- signor estoppel is the addition of language saying that pa- tents generally have the attributes of personal property. See post, at 7. That change has a bearing on whether West- inghouse should be overruled because it undermined West- inghouse’s analogy to estoppel by deed, but it is not enough to show abrogation because Westinghouse did not rely solely on that analogy. In sum, I do not think we can decide the question that the petition in this case presents unless we decide whether Westinghouse should be overruled. 5 Because the majority and the principal dissent refuse to decide whether Westing- house should be overruled, I would dismiss the writ as im- providently granted. I therefore respectfully dissent.

—————— 5 Under similar circumstances in Kimble v. Marvel Entertainment,

LLC, 576 U. S. 446 (2015), every Member of this Court assessed a judge- made patent-law doctrine through the lens of stare decisis, see id., at 455–465; id., at 470–472 (ALITO, J., dissenting), even though “Congress ha[d] repeatedly amended . . . the specific provision . . . on which [our earlier decision nominally] rested,” id., at 456 (majority opinion). The principal dissent does not even cite Kimble, let alone make any effort to reconcile its novel approach with that in our most analogous precedent. Cite as: 594 U. S. ____ (2021) 1

BARRETT, J., dissenting

MINERVA SURGICAL, INC., PETITIONER v. HOLOGIC, INC., ET AL. ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FEDERAL CIRCUIT [June 29, 2021]

JUSTICE BARRETT, with whom JUSTICE THOMAS and JUSTICE GORSUCH join, dissenting. The Patent Act of 1952 sets forth a comprehensive scheme for the creation and protection of patent rights. But it nowhere mentions the equitable doctrine of assignor es- toppel, which precludes inventors who file patent applica- tions from later saying that the patent is invalid. To the contrary, where the Act does address invalidity defenses, it states that invalidity “shall” be a defense “in any action in- volving the validity or infringement of a patent.” 35 U. S. C. §282(b). The text includes no exception for actions in which the inventor is the defendant. So why the doctrine of assignor estoppel? Because in Westinghouse Elec. & Mfg. Co. v. Formica Insulation Co., 266 U. S. 342 (1924), we interpreted a predecessor statute, the Patent Act of 1870, to incorporate the doctrine. The question before us is whether the doctrine carried over into the Patent Act of 1952. That could have happened in one of two ways: (1) if Congress ratified Westinghouse when it reenacted the assignment provision in 1952, or (2) if as- signor estoppel was part of the well-settled common-law backdrop against which Congress legislated in 1952. The Court opts for the second theory, but in my view, neither works. 2 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

I I will take the possibility of congressional ratification first because it follows more naturally from Westinghouse. In that case, the Court did not present the doctrine of as- signor estoppel as a well-established background principle against which Congress legislated when it enacted the Pa- tent Act of 1870. 1 Nor could it have: The first American case to apply the doctrine was not decided until 1880. See Faulks v. Kamp, 3 F. 898 (CC SDNY). Instead, the West- inghouse Court identified assignor estoppel as a rule that made sense in light of the Act’s assignment provision. After examining the text of the provision, the Court explained: “[T]here seems to be no reason why the principles of estop- pel by deed should not apply to assignment of a patent right in accordance with the statute,” because “[i]t was mani- festly intended by Congress to surround the conveyance of patent property with safeguards resembling those usually attaching to that of land.” 266 U. S., at 348–349. Some lower courts had applied the doctrine to the assignment and conveyance of patents, and, giving that trend its blessing, the Court described assignor estoppel as a sensible gloss on the assignment provision of the 1870 Act. Id., at 349–350. 2 The question here is whether Congress embraced this gloss when it reenacted that provision in 1952. Congress ratifies a judicial interpretation in a reenacted statute only if two requirements are satisfied: (1) the interpretation must be so well settled that we can “presume Congress —————— 1 I do not understand the Court to have a contrary view—its position is

that assignor estoppel had become a well-established background princi- ple of patent adjudication by the time Congress enacted the Patent Act of 1952. Ante, at 10. I address this possibility in Part II, infra. 2 JUSTICE ALITO maintains that Westinghouse did not interpret the

statute because it made no effort to parse the text. Ante, at 2–3 (dissent- ing opinion). But whatever the decision’s merits, it plainly grounded as- signor estoppel in the statute’s assignment provision. See Westinghouse, 266 U. S., at 348–349. Cite as: 594 U. S. ____ (2021) 3

knew of and endorsed it” at the time of the reenactment, and (2) the statute must be reenacted without material change. Jama v. Immigration and Customs Enforcement, 543 U. S. 335, 349, 351 (2005); Forest Grove School Dist. v. T. A., 557 U. S. 230, 239–240 (2009); see also Lamar, Archer & Cofrin, LLP v. Appling, 584 U. S. ___, ___ (2018) (slip op., at 11) (noting that Congress incorporates a prior judicial in- terpretation of a statute when it uses “the materially same language” and the interpretation is “longstanding”); W. Eskridge, Interpreting Law: A Primer on How to Read Stat- utes and the Constitution 421–422 (2016) (“When Congress reenacts a statute, it incorporates settled interpretations of the reenacted statute. The rule is inapplicable when there is no settled standard Congress could have known or the re- enactment makes a material change in the text” (footnote omitted)). So here, respondents must persuade us that (1) as of 1952, Westinghouse’s construction of the assignment provision in the Patent Act of 1870 was well settled, and (2) the assignment provision in the Patent Act of 1952 is mate- rially identical to the 1870 version. Jama, 543 U. S., at 349. They cannot clear either hurdle. 3 —————— 3 JUSTICE ALITO suggests that the reenactment canon has no applica-

tion when this Court, as opposed to lower courts, has interpreted the prior version of a statute. Ante, at 3–4. He is mistaken. As a leading treatise explains, the canon has its most obvious application when a “court of last resort” interprets a statute, though it “applies as well to uniform holdings of lower courts and even to well-established agency in- terpretations.” A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 323–324 (2012) (footnote omitted); see also C. Nelson, Statutory Interpretation 479 (2011) (noting that the canon applies when “[t]he Supreme Court (or a critical mass of lower courts, or an agency that Congress has put in charge of administering the statute) adopt[s] a prominent interpretation of one of the statute’s provisions”); Forest Grove School Dist. v. T. A., 557 U. S. 230, 239 (2009) (applying the canon to a statute that we had previously interpreted and that Congress had reen- acted without material change); Shapiro v. United States, 335 U. S. 1, 16, 20 (1948) (same). To be sure, there is a practical difference between 4 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

A Westinghouse’s construction of the assignment provision in the Patent Act of 1870 was far from well settled in 1952. Indeed, it is difficult to describe Westinghouse itself as much more than a “mild endorsement of assignor estoppel.” Brief for Petitioner 20. While accepting the doctrine, the Court simultaneously declined to apply it. Westinghouse, 266 U. S., at 355 (holding that an assignor could use prior art to narrow the patent claims); cf. Brief for United States as Amicus Curiae 10 (“This Court has never actually ap- plied assignor estoppel in a case before it”). Still, if West- inghouse had been the last word on assignor estoppel, one might argue that it set forth a statement of a stable princi- ple. But Westinghouse was not the last word on assignor es- toppel. The next time we considered the doctrine, we back- pedaled. See Scott Paper Co. v. Marcalus Mfg. Co., 326 U. S. 249 (1945). Deeming Westinghouse’s analysis a “logi- cal embarrassment,” we carved out an exception plainly in- consistent with the general rule of assignor estoppel: that it does not apply to an assignor who contests a patent’s valid- ity by invoking an expired patent. 326 U. S., at 253, 256– 257. We also cast doubt on the continuing validity of the doctrine, expressly declining to address the extent to which assignor estoppel “may be deemed to have survived the [Westinghouse] decision or to be restricted by it.” Id., at 254.

—————— the canon’s application to this Court and lower courts: It takes a “uniform and sufficiently numerous” body of lower court decisions to satisfy the presumption of congressional notice, whereas a single decision of this Court can be enough. Scalia, Reading Law, at 325; see, e.g., Forest Grove, 557 U. S., at 239; Manhattan Properties, Inc. v. Irving Trust Co., 291 U. S. 320, 336 (1934); cf. BP p.l.c. v. Mayor and City Council of Baltimore, 593 U. S. ___, ___ (2021) (slip op., at 11). But if a decision of this Court is oblique or thrown into doubt by subsequent precedent—both of which are true here—we have no basis for presuming that Congress was on notice of and endorsed our position. See Forest Grove, 557 U. S., at 239. Cite as: 594 U. S. ____ (2021) 5

The dissent, for its part, claimed that the Court had “repu- diated judicially” assignor estoppel altogether. Id., at 264 (opinion of Frankfurter, J.); see also Edward Katzinger Co. v. Chicago Metallic Mfg. Co., 329 U. S. 394, 400 (1947) (de- scribing Scott Paper as holding that an assignor could “chal- lenge the validity” of a patent and “defeat an action for in- fringement”). So when Congress reenacted the Patent Act in 1952, as- signor estoppel was far from well settled—if anything, it was on life support. Respondents could muster only three cases in the seven years after Scott Paper (before Congress reenacted the Patent Act) even loosely suggesting support for the doctrine. See Brief for Respondents 6. Indeed, ra- ther than embracing assignor estoppel after Scott Paper, courts questioned the doctrine’s validity. See, e.g., Douglass v. United States Appliance Corp., 177 F. 2d 98, 101 (CA9 1949). Scholars did the same; some, like the dis- sent in Scott Paper, concluded that we had “wipe[d] out es- toppel by assignment.” Lechner, Estoppel Against Patent Assignors—The Scott Paper Company Case, 28 J. Pat. Off. Soc. 325, 330 (1946). Others concluded that the law was a mess. As one scholar put it: Westinghouse’s rule “has be- come so unsettled during the past forty years of judicial con- sideration that, today, some courts apparently consider the rule to be no longer valid, others find no weakening of the rule, while still other courts apply the rule only after con- siderable speculation as to its continued validity.” Cooper, Estoppel To Challenge Patent Validity: The Case of Private Good Faith vs. Public Policy, 18 W. Res. L. Rev. 1122, 1123 (1967) (footnotes omitted; emphasis added). Post-1952 judicial decisions addressing assignor estoppel supply yet more evidence that the status of the doctrine was (at best) uncertain when Congress reenacted the Patent Act. In Lear, Inc. v. Adkins, 395 U. S. 653 (1969), for exam- ple, we disavowed the closely related doctrine of licensee es- toppel, noting along the way that the exception articulated 6 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

in Scott Paper had “undermined the very basis of the ‘gen- eral rule’ ” of assignor estoppel. 395 U. S., at 666. And fol- lowing Lear, several lower courts concluded that assignor estoppel was not just unsettled but, like licensee estoppel, dead. See, e.g., Coastal Dynamics Corp. v. Symbolic Dis- plays, Inc., 469 F. 2d 79 (CA9 1972). Tellingly, respondents could not come up with even one case applying assignor estoppel in the nearly 20-year period from Lear until the Federal Circuit resurrected the doctrine in 1988. See Diamond Scientific Co. v. Ambico, Inc., 848 F. 2d 1220, 1224–1225. And when the Federal Circuit res- urrected the doctrine, even it acknowledged that we had left the vitality of assignor estoppel unsettled: “Although the Supreme Court has examined th[e] doctrine . . . its opinions have hardly been definite or definitive.” Id., at 1222. Given all this, it is hard to see how we can “presume Con- gress knew of and endorsed” the doctrine when it adopted a new version of the Patent Act in 1952. Jama, 543 U. S., at 349. That is so even if a technical parsing of Scott Paper finds that it left assignor estoppel not quite dead, but ever- so-faintly breathing. In this circumstance, it would be strange to conclude that the vitality of assignor estoppel was so “unquestioned,” 543 U. S., at 349, that Congress would have “regard[ed] the point as settled law,” A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 325 (2012). On the contrary, the law was anything but clear. Today’s opinion confirms as much. Although the Court endorses assignor estoppel, it does not apply the doctrine as Westinghouse described it. Westinghouse stated the “rule” of assignor estoppel this way: “[A]n assignor of a patent right is estopped to attack the utility, novelty or validity of a patented invention which he has assigned or granted as against any one claiming the right under his assignment or grant.” 266 U. S., at 349. The Court describes assignor estoppel much more narrowly: “The doctrine applies only Cite as: 594 U. S. ____ (2021) 7

when an inventor says one thing (explicitly or implicitly) in assigning a patent and the opposite in litigating against the patent’s owner.” Ante, at 5. This version of assignor estoppel does not appear in Westinghouse—nor, to my knowledge, in any other judicial decision. The new version might be preferable to the old, but if Congress truly had ratified Westinghouse, it would have endorsed the Westinghouse version. B The reenactment canon has a second requirement: The reenacted statute must be materially identical to the one previously interpreted. 4 This poses another stumbling block for assignor estoppel, because the assignment provi- sion of the 1952 Act contains a significant sentence that the 1870 Act did not: “Subject to the provisions of this title, pa- tents shall have the attributes of personal property.” 5 Pa- tent Act of 1952, §261, 66 Stat. 810 (emphasis added).

—————— 4 JUSTICE ALITO argues that the “material change” standard requires

“express language” or “unavoidably implied contradiction.” Ante, at 4 (internal quotation marks omitted). He draws this heightened standard from an entirely different canon—the presumption that a later enacted statute does not impliedly repeal a former one. See ibid.; Scalia, Reading Law, at 327–333 (describing the “Presumption Against Implied Repeal”). As for the canon applicable here—the reenactment canon—our prece- dents do not support JUSTICE ALITO’s proposed standard. See, e.g., Lo- rillard v. Pons, 434 U. S. 575, 580 (1978) (“Congress is presumed to be aware of [a] judicial interpretation of a statute and to adopt that inter- pretation when it re-enacts a statute without change”); Forest Grove, 557 U. S., at 239–240 (same); see also Lamar, Archer & Cofrin, LLP v. Ap- pling, 584 U. S. ___, ___ (2018) (slip op., at 11) (Congress is presumed to be aware of a prior judicial interpretation of a statute when it uses “the materially same language” in a subsequent statute). The implied repeal canon addresses a different situation: when a court interprets two differ- ent and arguably conflicting statutory provisions. Scalia, Reading Law, at 331. 5 The Patent Act of 1870 provides, in relevant part:

“[E]very patent or any interest therein shall be assignable in law, by 8 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

Given the reasoning of Westinghouse, this is a material change. Westinghouse grounded its approval of assignor estoppel in the idea that patents are like real property. After observ- ing that the assignment provision required an assignment to be made in writing and recorded in the Patent Office within three months, the Court stated that “there seems to be no reason why the principles of estoppel by deed should not apply to assignment of a patent right in accordance with the statute.” 266 U. S., at 348–349. After all, the purpose of the writing-and-recording requirement is to “furnish written and recorded evidence of title and to protect the purchaser of the title as recorded for value without notice.” Id., at 349. Thus, the Court concluded, “[i]t was manifestly intended by Congress to surround the conveyance of patent property with safeguards” that resemble “those usually at- taching to that of land.” Ibid. But this analogy was inapt from the start. Even Westing- house admitted that deeds and patents differed in an im- portant respect: “A tract of land is easily determined by sur- vey. Not so the scope of a patent right for an invention.” Id., at 350. Moreover, unlike the grantor of a deed, who guarantees the quality of title, an assignor of a patent can- not warrant a patent’s validity. The validity of a patent in- volves a factual and legal inquiry “predicated on factors as to which reasonable men can differ widely.” Lear, 395 U. S., at 670; see Stanford, Diamond Scientific Co. v. Ambico, Inc.: Enforcing Patent Assignor Estoppel, 26 Houston L. Rev.

—————— an instrument in writing; and the patentee or his assigns or legal repre- sentatives may, in like manner, grant and convey an exclusive right un- der his patent to the whole or any specified part of the United States; and said assignment, grant, or conveyance shall be void as against any subsequent purchaser or mortgagee for a valuable consideration, without notice, unless it is recorded in the patent office within three months from the date thereof.” §36, 16 Stat. 203; see Westinghouse, 266 U. S., at 348 (citing to the revised version of the 1870 Act, “§4898”). Cite as: 594 U. S. ____ (2021) 9

761, 766 (1989) (“Since validity is never contractually trans- ferred in an assignment, there is no theoretical basis for as- signor estoppel by deed” (footnote omitted)). In any event, however persuasive we found the patent- deed analogy in Westinghouse, the Patent Act of 1952 un- raveled it. When Congress reenacted the assignment pro- vision, it specified that patents have “the attributes of per- sonal property.” 66 Stat. 810 (emphasis added). Because this language is inconsistent with the premise on which Westinghouse rested, it undercuts the argument that Con- gress ratified Westinghouse in the new assignment provi- sion. Cf. Holder v. Martinez Gutierrez, 566 U. S. 583, 593 (2012) (concluding that a reenacted provision did not ratify a prior judicial construction where the provision lacked the word on which the earlier construction was based). Respondents try to brush off the change as insignificant, asserting that this “sentence has no bearing on assignor es- toppel anyway, because estoppel by deed, upon which as- signor estoppel is based, can apply to real or personal prop- erty.” Brief for Respondents 20 (internal quotation marks omitted). But whatever the scope of estoppel by deed—re- spondents do not elaborate—Westinghouse expressly rested on “[t]he analogy between estoppel in conveyances of land and estoppel in assignments of a patent right.” 266 U. S., at 350 (emphasis added). In the one case, a “grantor pur- ports to convey the right to exclude others . . . from a de- fined tract of land, and in the other, from a described and limited field of the useful arts.” Ibid. By making clear that patents have “the attributes of personal property,” Con- gress directly undermined an interpretation that treated patents like deeds conveying land. II If Congress did not ratify our gloss on the assignment provision in the 1952 Act, how else might assignor estoppel be part of the statute? The Court comes at the interpretive 10 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

problem from a different angle: It holds that by 1952, as- signor estoppel had become “a background principle of pa- tent adjudication” against which Congress legislated. Ante, at 10. On this theory, Westinghouse is important not for its interpretation of the assignment provision but for the equi- table principle that it endorsed. In my view, this theory also fails because Westinghouse proved to be a false start for the doctrine. We have said that “where a common-law principle is well established, . . . courts may take it as given that Congress has legislated with an expectation that the principle will apply” absent statutory cues to the contrary. Astoria Fed. Sav. & Loan Assn. v. Solimino, 501 U. S. 104, 108 (1991); see also Eskridge, Interpreting Law, at 348 (“[C]ourts will assume that legislatures act against the background of the common law”). So, for example, a federal statute of limita- tions ordinarily is subject to equitable tolling even when the text is silent because “Congress must be presumed to draft limitations periods in light of this background principle.” Young v. United States, 535 U. S. 43, 49–50 (2002); see also Nelson, Statutory Interpretation, at 629 (“[C]ourts fre- quently understand federal statutes to come with some un- stated qualifications or embellishments suggested by prin- ciples of general jurisprudence”). The Court says that assignor estoppel works this way too. Ante, at 10. True, the Patent Act provides that “[i]nvalidity of the patent shall be a defense[ ] in any action involving” infringement. Ibid. (quoting 35 U. S. C. §282(b); internal quotation marks and some alterations omitted). Yet, the Court reasons, this apparently absolute language does not “foreclose applying in patent cases a whole host of common- law preclusion doctrines,” including “equitable estoppel, collateral estoppel, res judicata, and law of the case.” Ante, at 10. Assignor estoppel, the Court says, falls in that same category. Ibid. According to the Court, Westinghouse con- firmed a trend that had already begun in the lower courts Cite as: 594 U. S. ____ (2021) 11

and has continued unabated since, giving assignor estoppel a place in the pantheon of well-established common-law principles. I disagree. The common-law pedigree of assignor estop- pel differs markedly from that of the preclusion doctrines with which the Court groups it. Some of those doctrines have been around for nearly a thousand years, see, e.g., Bursak, Note, Preclusions, 91 N. Y. U. L. Rev. 1651, 1663 (2016) (“Res judicata had migrated to England no later than the early 1100s”); Millar, The Historical Relation of Estop- pel by Record to Res Judicata, 35 Ill. L. Rev. 41, 44–45 (1940) (res judicata and collateral estoppel go back to at least the 1200s), and were far more settled in early Ameri- can courts, see, e.g., Washington, Alexandria, & Georgetown Steam-Packet Co. v. Sickles, 24 How. 333, 341 (1861) (“The authority of the res judicata . . . is derived by us from the Roman law and the Canonists”); Hopkins v. Lee, 6 Wheat. 109, 114 (1821) (stating that collateral estoppel has “found its way into every system of jurisprudence”). They are firmly rooted in our jurisprudence now. Assignor estoppel, by contrast, has far from this kind of “impeccable historic pedigree.” Kirtsaeng v. John Wiley & Sons, Inc., 568 U. S. 519, 538 (2013). It is more recent and far shakier. It was introduced into patent law in the late 19th century—about a hundred years after Congress en- acted the first patent laws and a decade after Congress passed the 1870 Act. See Faulks v. Kamp, 3 F. 898 (CC SDNY 1880); ante, at 6 (opinion of the Court). And after its introduction, it lacked staying power. Westinghouse proved to be the “high-water mark of the doctrine in this Court.” Brief for Petitioner 20. As I have already explained, lower courts, commentators, the Scott Paper dissent, and even the Court itself in Lear regarded Scott Paper as having gutted the doctrine of assignor estoppel. See supra, at 4–6. It is therefore difficult for me to see how, in 1952, assignor 12 MINERVA SURGICAL, INC. v. HOLOGIC, INC.

estoppel constituted a “long-established and familiar prin- cipl[e]” like res judicata. Isbrandtsen Co. v. Johnson, 343 U. S. 779, 783 (1952). At most, the doctrine of assignor es- toppel was in a confused state by 1952—a far cry from the high bar that we have required to incorporate “well-settled” common-law background principles into a statute. See, e.g., Kirtsaeng, 568 U. S., at 538 (noting the first-sale doctrine’s “impeccable historic pedigree”); Lozano v. Montoya Alvarez, 572 U. S. 1, 10 (2014) (emphasizing that equitable tolling is a “long-established feature of American jurisprudence”). Indeed, whatever one might have said when the Court de- cided Westinghouse in 1924, Scott Paper cut sharply in the other direction. And such “contradictory signals are not typically the stuff of which background rules of common law are made.” B&B Hardware, Inc. v. Hargis Industries, Inc., 575 U. S. 138, 163 (2015) (THOMAS, J., dissenting). III Respondents insist that assignor estoppel promotes fair dealing, while petitioner protests that the supposedly equi- table doctrine is actually inequitable in practice. If we had authority to develop federal common law on the subject, we could take sides in that debate. But no one contends that we do. This case turns on whether the Patent Act of 1952 incorporates the doctrine, and because it does not, I respect- fully dissent.