Per Curiam Opinion; Opinion by Judge Collins; Partial Concurrence and Partial Dissent by Judge Bea;

Concurrence by Judge R. Nelson;

Concurrence by Judge VanDyke;

Dissent by Chief Judge Murguia;

Dissent by Judge Lee

SUMMARY *

Religious Freedom Restoration Act / Free Exercise

Clause

The en banc court filed (1) an order denying a petition for rehearing en banc before the full court and amending Judge Collins’s opinion, and (2) Judge Collins’s amended opinion in a case in which the en banc court affirmed the district court’s order denying Apache Stronghold’s motion for a preliminary injunction against the federal government’s transfer of Oak Flat—federally owned land within the Tonto National Forest—to a private company, Resolution Copper.

Oak Flat is a site of great spiritual value to the Western Apache Indians and also sits atop the world’s third-largest deposit of copper ore. To take advantage of that deposit, Congress by statute—the Land Transfer Act—directed the federal government to transfer the land to Resolution Copper, which would then mine the ore.

Apache Stronghold, an organization that represents the interests of certain members of the San Carlos Apache Tribe, sued the government, seeking an injunction against the land transfer on the ground that the transfer would violate its members’ rights under the Free Exercise Clause of the First Amendment, the Religious Freedom Restoration Act (“RFRA”), and an 1852 treaty between the United States and the Apaches.

* This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.

The per curiam opinion provides an overview of the votes of the en banc court:

• A majority of the en banc court (Chief Judge

Murguia, and Judges Gould, Berzon, R. Nelson, Lee and Mendoza) concluded that (1) the Religious Land Use and Institutionalized Persons Act of 2000 (“RLUIPA”), and RFRA are interpreted uniformly; and (2) preventing access to religious exercise is an example of substantial burden. A majority of the en banc court therefore overruled the narrow definition of substantial burden under RFRA in Navajo Nation v. U.S. Forest Service, 535 F.3d 1058 (9th Cir. 2008) (en banc).

• A different majority of the en banc court (Judges

Bea, Bennett, R. Nelson, Collins, Forrest, and VanDyke) concluded that (1) RFRA subsumed, rather than overrode, the outer limits that Lyng v. Northwest Indian Cemetery Protective Ass’n, 485 U.S. 439 (1988), placed on what counts as a governmental imposition of a substantial burden on religious exercise; and (2) under Lyng, a disposition of government real property does not impose a substantial burden on religious exercise when it has “no tendency to coerce individuals into acting contrary to their religious beliefs,” does not “discriminate” against religious adherents, does not “penalize” them, and does not deny them “an equal share of the rights, benefits, and privileges enjoyed by other citizens.” Apache Stronghold’s claims under the Free Exercise Clause and RFRA failed under these Lyng-based standards and the claims based on the 1852 treaty failed for separate reasons.

In his amended opinion for the court, Judge Collins, joined by Judges Bea, Bennett, R. Nelson, Forrest, and VanDyke, held that Apache Stronghold was unlikely to succeed on the merits on any of its three claims before the court, and consequently was not entitled to preliminary injunctive relief.

• Apache Stronghold’s claim that the transfer of Oak

Flat to Resolution Copper would violate the Free Exercise Clause failed under the Supreme Court’s controlling decision in Lyng because the project challenged here is indistinguishable from that in Lyng. As in Lyng, the government’s actions with respect to “publicly owned land” would “interfere significantly with private persons’ ability to pursue spiritual fulfillment according to their religious beliefs,” but it would have no “tendency to coerce” them “into acting contrary to their religious beliefs.” Also, as in Lyng, the challenged transfer of Oak Flat for mining operations did not discriminate against Apache Stronghold’s members, did not penalize them, or deny them an “equal share of the rights, benefits, and privileges enjoyed by other citizens.”

• Apache Stronghold’s claim that the transfer of Oak

Flat to Resolution Cooper would violate RFRA failed for the same reasons because what counts as “substantially burden[ing] a person’s exercise of religion” must be understood as subsuming, rather than abrogating, the holding of Lyng. • Apache Stronghold’s claim that the 1852 Treaty of Sante Fe created an enforceable trust obligation that would be violated by the transfer of Oak Flat failed because the government’s statutory obligation to transfer Oak Flat abrogated any contrary treaty obligation.

Concurring in part and dissenting in part, Judge Bea, joined by Judge Forrest except for footnote 1 and by Judge Bennett with respect to Part II, dissented from paragraph one of the per curiam opinion’s purported overruling of Navajo Nation because a majority of the panel already affirmed the district court, under the different rationale in Judge Collins’s majority opinion, the district court’s finding that the transfer of Oak Flat will impose no substantial burden under RFRA. He concurred in full with Judge Collins’s majority opinion, and wrote separately to provide additional reasons in support of the conclusion that Apache Stronghold cannot obtain relief under RFRA.

Concurring, Judge R. Nelson stated that en banc review was warranted to correct the faulty legal test (not outcome) in Navajo Nation. He explained that since Navajo Nation was decided, it has become clear that “substantial burden” means more in RLUIPA than the narrow definition Navajo Nation gave it under RFRA, and a majority of the en banc court now rejects the narrow construction of “substantial burden” in Navajo Nation. While the dissent raises a plausible textual interpretation of “substantial burden” under RFRA, Judge R. Nelson ultimately disagrees with it. Because RFRA does not overrule the Supreme Court’s binding precedent in Lyng, Apache Stronghold has no viable RFRA claim.

Concurring, Judge VanDyke agreed with the majority that this decision is controlled by Lyng, and wrote separately to elaborate on why the alleged “burden” in this case is not cognizable under RFRA and to explain why reinterpreting RFRA to impose affirmative obligations on the government to guarantee its own property for religious use would inevitably result in religious discrimination.

Dissenting, Chief Judge Murguia, joined by Judges Gould, Berzon, and Mendoza, and by Judge Lee as to all but Part II.H, wrote that the utter destruction of Oak Flat, a site sacred to the Western Apaches since time immemorial, is a “substantial burden” on the Apaches’ sincere religious exercise under RFRA. Navajo Nation wrongly defined “substantial burden” as a narrow term of art and foreclosed relief. In light of the plain meaning of “substantial burden,” RFRA prohibits government action that “oppresses” or “restricts” “any exercise of religion, whether or not compelled by, or central to, a system of religious belief,” to a “considerable amount,” unless the government can demonstrate that imposition of the burden is in furtherance of a compelling governmental interest and the least restrictive means of furthering that compelling governmental interest. Chief Judge Murguia would hold that Apache Stronghold has shown that it is likely to succeed on the merits of its RFRA claim, and would remand for the district court to determine whether the Land Transfer Act is justified by a compelling interest pursued through the least restrictive means. Finally, Chief Judge Murguia rejected the government’s eleventh-hour argument that RFRA does not apply to the Land Transfer Act.

Dissenting, Judge Lee joined all of Chief Judge Murguia’s dissent except for Section II.H because the government waived the argument that RFRA cannot apply to the Land Transfer Act.

COUNSEL

Luke W. Goodrich (argued), Mark L. Rienzi, Diana M. Verm Thompson, Joseph C. Davis, Christopher Pagliarella, Daniel D. Benson, and Kayla A. Toney, The Becket Fund for Religious Liberty, Washington, D.C.; Michael V. Nixon, Michael V. Nixon JD, Portland, Oregon; Clifford I. Levenson, Law Office of Clifford Levenson, Phoenix, Arizona; for Plaintiffs-Appellants. Stephanie H. Barclay (argued) and Francesca Matozzo, University of Notre Dame Law School Religious Liberty Clinic, Notre Dame, Indiana; Meredith H. Kessler, Religious Liberty Clinic, Notre Dame, Indiana; Michalyn Steele, Brigham Young University Law School, Provo, Utah; for Amicus Curiae National Congress of American Indians, a Tribal Elder and other Federal Indian Law Scholars, and Organizations.

Miles E. Coleman, Nelson Mullins Riley & Scarborough LLP, Greenville, South Carolina; Thomas Hydrick, Assistant Deputy Solicitor General, South Carolina Attorney General’s Office, Columbia, South Carolina; Hunter Windham, Duffy & Young LLC, Charleston, South Carolina; Thomas C. Berg, Religious Liberty Appellate Clinic, University of St. Thomas School of Law, Minneapolis, Minnesota; W. Thomas Wheeler, Fredrikson & Byron PA, Minneapolis, Minnesota; for Amici Curiae Religious Liberty Law Scholars.

James C. Phillips, Chapman University, Dale E. Fowler School of Law, Orange, California; Gene C. Schaerr, Joshua J. Prince, Edward H. Trent, Riddhi Dasgupta, and Megan Shoell, Schaerr Jaffe LLP, Washington, D.C.; for Amici Curiae The Jewish Coalition for Religious Liberty, The International Society for Krishna Consciousness, The Sikh Coalition, and Protect the 1st.

Joshua C. McDaniel, Kelsey Baer Flores, Matthew E. Myatt, and Parker W. Knight III, Harvard Law School Religious Freedom Clinic, Cambridge, Massachusetts, for Amicus Curiae The Sikh Coalition.

James C. Phillips, Chapman University, Dale E. Fowler School of Law, Orange, California; Alexander Dushku, R. Shawn Gunnarson, Justin W. Starr, and Jarom Harrison, Kirton McConke, Salt Lake City, Utah; for Amici Curiae The Church of Jesus Christ of Latter-Day Saints, The General Conference of Seventh-Day Adventists, The Islam and Religious Freedom Action Team of the Religious Freedom Institute, and The Christian Legal Society. Jason Searle and Beth Wright, Native American Rights Fund, Boulder, Colorado; April Youpee-Roll, Munger Tolls & Olson LLP, Los Angeles, California; for Amici Curiae Tribal Nations and Tribal Organizations. David T. Raimer, Megan L. Owen, and Anika M. Smith, Jones Day, Washington, D.C., for Amicus Curiae The Mennonite Church USA and the Pacific Southwest Mennonite Conference.

Joan M. Pepin (argued), Andrew C. Mergen, Tyler M. Alexander, Attorneys; Jean E. Williams, Acting Assistant Attorney General; Todd Kim, Assistant Attorney General; United States Department of Justice, Environment and Natural Resources Division, Washington, D.C.; Katelin Shugart-Schmidt, Attorney, United States Department of Justice, Environment & Natural Resources Division, Denver, Colorado; for Defendants-Appellees. David Debold (argued), Thomas G. Hungar, and Matthew S. Rozen, Gibson Dunn & Crutcher LLP, for Amicus Curiae American Exploration & Mining Association, Women’s Mining Coalition, and Arizona Rock Products Association. William E. Trachman, Mountain States Legal Foundation, Lakewood, Colorado; Timothy Sandefur, Goldwater Institute, Phoenix, Arizona; for Amicus Curiae Towns of Superior and Hayden, Arizona, and Jamie Ramsey, the Mayor of Kearny, Arizona.

Kathryn M. Barber and Matthew A. Fitzgerald, McGuireWoods LLP, Richmond, Virginia, for Amici Curiae Pinal Partnership, Valley Partnership, PHX East Valley Partnership, The Honorable Scott J. Davis, The Honorable Myron Lizer, and Joshua Tahsuda, III. Anthony J. Ferate, Andrew W. Lester, and Courtney D. Powell, Spencer Fane LLP, Oklahoma City, Oklahoma, for Amicus Curiae Arizona Chamber of Commerce and Industry.

Christopher E. Mills, Spero Law LLC, Charleston, South Carolina, for Amici Curiae 38 Religious and Tribal Organizations.

Stephanie H. Barclay, Professor of Law, Religious Liberty Initiative Director; Meredith H. Kessler and Francesca Matozzo; Notre Dame Law School Religious Liberty Clinic, Notre Dame, Indiana; for Amici Curiae International Council of Thirteen Indigenous Grandmothers, Mica Group, and a Tribal Elder.

Joshua C. McDaniel, Parker W. Knight, III, and Kathryn F. Mahoney, Harvard Law School Religious Freedom Clinic, Harvard Law School, Cambridge, Massachusetts, for Amici Curiae the Sikh Coalition, the Christian Legal Society, and the Islam and Religious Freedom Action Team of the Religious Freedom Institute.

Gene C. Schaerr, Erika S. Jaffe, Annika B. Barkdull, and Megan Shoell, Schaerr Jaffe LLP. Washington, D.C., for Amicus Curiae Protect the First Foundation. Beth M. Wright and Jason Searle, Native American Rights Fund, Boulder, Colorado, for Amici Curiae Tohono O’Odham Nation and Tribal Organizations. Heather D. Whiteman Runs Him, Tribal Justice Clinic, Rogers College of Law, University of Arizona, Tucson, Arizona; Gerald Torres, Yale Law School, New Haven, Connecticut; for Amicus Curiae the National Native American Law Students Association Inc., Yale Native American Law Students Association, and Michigan Native American Law Students Association.

Eric N. Kniffin, Ethics & Public Policy Center, Washington, D.C., for Amici Curiae the Mennonite Church USA and 19 Additional Mennonite Organizations.

ORDER

The slip opinion filed on March 1, 2024 is amended as follows:

1) On page 33, after “(quoting Lyng, 485 U.S. at 451).”, delete the remainder of the paragraph through and including “neutral and generally applicable.” Immediately after that shortened paragraph, add the following new paragraph:

But the Court has not said, and could not have said, that the holding of Lyng rested on the view that Lyng was itself a case involving a neutral and generally applicable law. As we have set forth, Lyng rested on a holding about the scope of the term “prohibiting” under the Free Exercise Clause and never mentioned or endorsed a Smith-style rule. At most, the Court has suggested in dicta that Lyng fits a pattern of cases in which the Court had upheld laws that were “neutral and generally applicable without regard to religion” in the sense that they did not “‘penalize religious activity by denying any person an equal share of the rights, benefits, and privileges enjoyed by other citizens.’” Trinity Lutheran Church of Columbia, Inc. v. Comer, 582 U.S. 449, 460 (2017) (quoting Lyng, 485 U.S. at 449). But Trinity Lutheran did not have before it the more focused question whether, in light of the parcel-specific rigging of the statutory framework in Lyng, the underlying statute at issue in Lyng could be properly deemed to qualify as “neutral and generally applicable” under the details of Smith’s framework. As we have explained, Lyng involved a situation in which, after religious objections had been raised to the G-O road and the road’s construction had been enjoined, Congress proceeded to adopt an explicit statutory gerrymander for the precise parcel at issue. See supra at 27–28. That manifestly would not fit the Court’s current understanding of a case involving a neutral and generally applicable law. See, e.g., Church of the Lukumi, 508 U.S. at 542 (emphasizing that “categories of selection” in legislative drafting “are of paramount concern when a law has the incidental effect of burdening religious practice”). In all events, even if the law in Lyng were deemed, in hindsight, to be neutral and generally applicable within the meaning of Smith, the fact remains that the holding of Lyng did not rest on any such premise, but instead on the view that the challenged actions there lacked the sort of features that would qualify as “prohibiting” the free exercise of religion.

2) On page 43, in the sentence that begins “Consequently,” add “pre-Smith” immediately before “framework for applying”.

An amended version of the opinion, reflecting these changes, accompanies this order. The per curiam opinion, the concurrences, and the dissents are unchanged. The full court has been advised of the petition for rehearing en banc before the full court filed on April 15, 2024 (Dkt. No. 184), and no judge of the court has requested a vote on whether to rehear the matter en banc. Fed. R. App. P. 35; Ninth Circuit General Order 5.8. Accordingly, the petition for rehearing en banc before the full court is DENIED. No further petitions for rehearing will be entertained.

OPINION

PER CURIAM:

A majority of the en banc court (Chief Judge MURGUIA and Judges GOULD, BERZON, R. NELSON, LEE, and MENDOZA) concludes that (1) the Religious Land Use and Institutionalized Persons Act of 2000 (“RLUIPA”), 42 U.S.C. § 2000cc, et seq., and the Religious Freedom Restoration Act (“RFRA”), 42 U.S.C. § 2000bb, et seq., are interpreted uniformly; and (2) preventing access to religious exercise is an example of substantial burden. A majority of the en banc court therefore overrules Navajo Nation v. U.S. Forest Service to the extent that it defined a “substantial burden” under RFRA as “imposed only when individuals are forced to choose between following the tenets of their religion and receiving a governmental benefit (Sherbert) or coerced to act contrary to their religious beliefs by the threat of civil or criminal sanctions (Yoder).” 535 F.3d 1058 (9th Cir. 2008) (emphasis added).

A different majority (Judges BEA, BENNETT, R. NELSON, COLLINS, FORREST, and VANDYKE) concludes that (1) RFRA subsumes, rather than overrides, the outer limits that the Supreme Court’s decision in Lyng v. Northwest Indian Cemetery Protective Ass’n, 485 U.S. 439 (1988), places on what counts as a governmental imposition of a substantial burden on religious exercise; and (2) under Lyng, a disposition of government real property does not impose a substantial burden on religious exercise when it has “no tendency to coerce individuals into acting contrary to their religious beliefs,” does not “discriminate” against religious adherents, does not “penalize” them, and does not deny them “an equal share of the rights, benefits, and privileges enjoyed by other citizens.” Lyng, 485 U.S. at 449–50, 453. The same majority holds that Apache Stronghold’s claims under the Free Exercise Clause and RFRA fail under these Lyng-based standards and that the claims based on the 1852 Treaty fail for separate reasons.

We therefore AFFIRM the district court’s order denying the motion for a preliminary injunction. COLLINS, Circuit Judge, delivered the following opinion for the court, in which BEA, BENNETT, R. NELSON, FORREST, and VANDYKE, Circuit Judges, join:

Oak Flat, an area located on federally owned land within Tonto National Forest, is a site of great spiritual value to the Western Apache Indians, who believe that it is indispensable to their religious worship. But Oak Flat also sits atop the world’s third-largest deposit of copper ore. To take advantage of that deposit, Congress by statute directed the federal Government to transfer the land to a private company, Resolution Copper, which would then mine the ore. Apache Stronghold, an organization that represents the interests of certain members of the San Carlos Apache Tribe, sued the Government, seeking an injunction against the land transfer on the ground that the transfer would violate its members’ rights under the Free Exercise Clause of the First Amendment, the Religious Freedom Restoration Act (“RFRA”), and an 1852 treaty between the United States and the Apaches. The district court denied Apache Stronghold’s request for a preliminary injunction on the ground that Apache Stronghold had not shown a likelihood of success on the merits. See Apache Stronghold v. United States, 519 F. Supp. 3d 591, 598 (D. Ariz. 2021). We affirm.

I A Apache Stronghold is an Arizona nonprofit corporation “based in the Western Apache lands of the San Carlos Apache Tribe.” It describes itself as “connecting Apaches and other Native and non-Native allies from all over the world.” Its declared mission is “to battle continued colonization, defend Holy sites and freedom of religion, and . . . build[] a better community through neighborhood programs and civic engagement.” The San Carlos Apache Tribe of the San Carlos Reservation is a federally recognized Indian tribe located on the San Carlos Reservation, roughly 100 miles east of Phoenix.

Apache Stronghold’s members engage in traditional Western Apache religious practices. Among the locations that are central to their religion is a place called “Chí’chil Biłdagoteel,” which in English means “Emory Oak Extends on a Level.” That accounts for the site’s more common name, which is “Oak Flat.” According to Apache Stronghold’s expert witness, Western Apache religious practices at Oak Flat date back at least a millennium. The Western Apache believe that Oak Flat is a “sacred place” that serves as a “direct corridor” to “speak to [their] creator.” Specifically, they believe that Oak Flat is the site where one of the “Ga’an”—spirit messengers between the Western Apache and their Creator—“has made its imprint, its spirit.” The Western Apache believe that the Ga’an, and the Western Apaches’ interaction with the Ga’an, constitute “a crucial part” of their “personal being,” and that Oak Flat thus provides them “a unique way . . . to communicate” with their Creator.

Members of the tribe report that they “cannot have this spiritual connection with the land anywhere else on Earth.” Oak Flat is “the only area” with these unique features, making it “crucial” to Western Apache religious life. As one example, members of the tribe stated that certain Western Apache religious practices must occur at Oak Flat and cannot take place anywhere else. And even among those religious practices that need not necessarily occur at Oak Flat, some trace their origins to practices that were first begun there. One such practice is the “Sunrise Ceremony,” a rite of passage for Western Apache girls to recognize “the gift of life and the bearing of children to the female.” The Western Apache believe that “the place the ceremony takes place is the life thread forever connecting the place and the girls who have their ceremony there.” One member testified that “the most important part about” the Sunrise Ceremony “is that everything that we are able to use for the ceremony comes from Chí’chil Biłdagoteel, Oak Flat.” Accordingly, in Western Apache religious belief, harms to Oak Flat work a corresponding spiritual harm to those who performed their Sunrise Ceremonies there, damaging their “life and their connection to their rebirth.”

B In addition to being a sacred site for the Western Apache, Oak Flat is also a place of considerable economic significance. Located near the “Copper Triangle,” Oak Flat sits atop the third-largest known copper deposit in the world. Roughly 4,500 to 7,000 feet beneath Oak Flat is an ore deposit containing approximately two billion tons of “copper resource.” The U.S. Forest Service estimates that, if mined, this deposit could yield around “40 billion pounds of copper.” For that reason, there has long been considerable interest among mining companies in gaining access to the Oak Flat deposit.

Believing the copper beneath Oak Flat to be a significant asset, various members of Arizona’s congressional delegation drafted legislation to compel the Government to transfer Oak Flat and its surroundings to Resolution Copper, a private mining company. Such legislation was introduced in each Congress from 2005 through 2014. 1 Although these bills were the subject of numerous hearings and other congressional action over the years, 2 these legislative efforts 1 See, e.g., Southeast Arizona Land Exchange and Conservation Act of 2005, H.R. 2618, 109th Cong. (2005); Southeast Arizona Land Exchange and Conservation Act of 2005, S. 1122, 109th Cong. (2005); Southeast Arizona Land Exchange and Conservation Act of 2006, H.R. 6373, 109th Cong. (2006); Southeast Arizona Land Exchange and Conservation Act of 2006, S. 2466, 109th Cong. (2006); Southeast Arizona Land Exchange and Conservation Act of 2007, H.R. 3301, 110th Cong. (2007); Southeast Arizona Land Exchange and Conservation Act of 2007, S. 1862, 110th Cong. (2007); Southeast Arizona Land Exchange and Conservation Act of 2008, S. 3157, 110th Cong. (2008); Southeast Arizona Land Exchange and Conservation Act of 2009, H.R. 2509, 111th Cong. (2009); Southeast Arizona Land Exchange and Conservation Act of 2009, S. 409, 111th Cong. (2009); Southeast Arizona Land Exchange and Conservation Act of 2011, H.R. 1904, 112th Cong. (2011); Southeast Arizona Land Exchange and Conservation Act of 2013, H.R. 687, 113th Cong. (2013); Southeast Arizona Land Exchange and Conservation Act of 2013, S. 339, 113th Cong. (2013).

2 A House subcommittee held a hearing on H.R. 3301 in the 110th Congress, but no further action was taken on that bill. See H.R. 3301, Southeast Arizona Land Exchange and Conservation Act of 2007: Hearing Before the Subcomm. on Nat’l Parks, Forests, & Pub. Lands of the H. Comm. on Nat. Res., SERIAL NO. 110-52 (Nov. ^[1], 2007). In the 111th Congress, a Senate subcommittee held a hearing on S. 409 on June 17, 2009, and that bill was subsequently reported on March 2, 2010 to the Senate floor, where no further action was taken. See Public Lands and Forests Bills: Hearing Before the Subcomm. on Pub. Lands & Forests of the S. Comm. on Energy & Nat. Res., S. HRG. NO. 111-65 (June 17, 2009); S. REP. NO. 111-129 (March 2, 2010). In the 112th Congress, H.R. 1904 was considered at a June 14, 2011 House subcommittee hearing, reported out of committee on October 14, 2011, and passed by the full House on October 26, 2011. See H.R. 473, et al.: Hearing Before the Subcomm. on Nat’l Parks, Forests, & Pub. Lands of the H. Comm. on Nat. Res., SERIAL NO. 112-40 (June 14, 2011); H.R. REP. NO. 112-246 (Oct. 14, 2011); 157 CONG. REC. H7090–110 (Oct. 26, 2011). A Senate committee then held a hearing on H.R. 1904 on Feb. did not bear fruit until late 2014, when Congress passed, and the President signed, the Carl Levin and Howard P. “Buck” McKeon National Defense Authorization Act for Fiscal Year 2015 (“NDAA”). See Pub. L. No. 113-291, 128 Stat. 3292 (2014). Included as § 3003 of the NDAA was a version of the previously oft-proposed “Southeast Arizona Land Exchange and Conservation Act.” 3 Id. § 3003, 128 Stat. at 3732–41 (classified to § 539p of the unenacted title 16 of the United States Code).

Section 3003’s declared purpose is “to authorize, direct, facilitate, and expedite the exchange of land between Resolution Copper and the United States.” 16 U.S.C. § 539p(a). To that end, it directs that “if Resolution Copper offers to convey to the United States all right, title, and interest of Resolution Copper” in certain “non-Federal land,” then “the Secretary [of Agriculture] is authorized and directed to convey to Resolution Copper, all right, title, and interest of the United States in and to the Federal land.” Id. 9, 2012. See Resolution Copper: Hearing Before the S. Comm. on Energy & Nat. Res., S. HRG. NO. 112-486 (Feb. ^[9], 2012). In 2013, both House and Senate subcommittees held further hearings in the 113th Congress on the respective versions of the legislation, and the House bill was reported to the House floor on July 22, 2013. See Oversight Hearing Titled “America’s Mineral Resources: Creating Mining and Manufacturing Jobs and Securing America”: Hearing on H.R. 1063, et al., Before the Subcomm. on Energy & Mineral Res. of the H. Comm. on Nat. Res., SERIAL NO. 113-7 (March 21, 2013); Current Public Lands, Forests, and Mining Bills: Hearing Before the Subcomm. on Pub. Lands, Forests, & Mining of the S. Comm. on Energy & Nat. Res., S. HRG. NO. 113-342 (November 20, 2013); H.R. REP. NO. 113-167 (July 22, 2013). 3 Apache Stronghold derides § 3003 as a “midnight” rider attached to a “must-pass” bill, but that characterization ignores the extensive hearings and congressional consideration given to the land transfer proposal over the previous seven years. See supra note 2. § 539p(c)(1). The referenced “Federal land” consists of “approximately 2,422 acres of land located in Pinal County, Arizona,” including Oak Flat and the surrounding area. Id. § 539p(b)(2); see U.S. Forest Service, Resolution Copper Project & Land Exchange, Map of Land Exchange Parcels, (2015), https://www.resolutionmineeis.us/documents/usfs- resolution-land-exchange-parcels-2016 [https://perma.cc/JEC7-GUC4].

The land exchange is subject to certain conditions. For example, title to the land the Government would receive from Resolution Copper must be in a form that is acceptable to the Secretaries of Agriculture and the Interior, and must conform to the Department of Justice’s “title approval standards.” 16 U.S.C. § 539p(c)(2)(A), (B). The federal and non-federal land must be independently appraised, id. § 539p(c)(4), and the value of the exchanged land equalized as set forth in the statute, id. § 539p(c)(5). Other provisions of § 3003 provide direction concerning ancillary matters related to the exchange. E.g., id. § 539p(i).

In recognition of the Western Apaches’ religious beliefs, Congress incorporated an accommodation provision into § 3003. That provision directs the Secretary of Agriculture to “engage in government-to-government consultation with affected Indian tribes” to address concerns “related to the land exchange.” 16 U.S.C. § 539p(c)(3)(A). Further, the statute obligates the Secretary to work with Resolution Copper to address those concerns and to mitigate any possible “adverse effects on the affected Indian tribes.” Id. § 539p(c)(3)(B). The statute also requires Resolution Copper to keep Oak Flat accessible to the public for as long as safely possible, id. § 539p(i)(3), and Congress explicitly set aside another religiously significant area, Apache Leap, in order to “preserve [its] natural character” and “allow for traditional uses of the area.” Id. § 539p(g)(2).

Lastly, Congress expressly stated that the land exchange would generally be governed by the National Environmental Policy Act (“NEPA”), 42 U.S.C. § 4321 et seq. Thus, § 3003 requires that an environmental impact statement (“EIS”) be prepared under NEPA prior to the Secretary executing the land exchange. ^[16] U.S.C. § 539p(c)(9)(B). Congress supplemented the ordinary NEPA requirements for such statements and required that the EIS for the land transfer also “assess the effects of the mining” on “cultural and archaeological resources” in the area and “identify measures . . . to minimize potential adverse impacts on those resources.” Id. § 539p(c)(9)(C). The EIS was then to form “the basis for all decisions under Federal law related to the proposed mine,” such as “the granting of any permits, rights- of-way,” and construction approvals. Id. § 539p(c)(9)(B).

The statute commands that the land transfer take place “[n]ot later than 60 days after” the publication of the EIS. ^[16] U.S.C. § 539p(c)(10). Nowhere in § 3003 does Congress confer on the Government discretion to halt the transfer. The statute mandates that the Government secure an appraisal of the land, id. § 539p(c)(4)(A); that it prepare the EIS, id. § 539p(c)(9)(B); and that it then transfer the land, id. § 539p(c)(10). Although Resolution Copper could theoretically prevent the transfer by refusing “to convey to the United States all right, title, and interest . . . in and to the non-Federal land,” id. § 539p(c)(1), no corresponding authority exists for the Government.

Once the land transfer takes place, Resolution Copper plans to extract the ore by using “panel caving,” a technique that entails digging a “network of shafts and tunnels below the ore body.” Resolution Copper will then detonate explosives to fracture the ore, which will “move[] downward” as a result. That, in turn, will cause the ground above to begin to collapse inward. Over the next 41 years, Resolution Copper will remove progressively more ore from below Oak Flat, causing the surface geography to become increasingly distorted. The resulting subsidence will create a large surface crater, which the Forest Service estimates will span approximately 1.8 miles in diameter and involve a depression between 800 and 1,115 feet deep.

This collapse will not occur immediately upon transfer of the land. Even once Resolution Copper begins construction on the mine, it will be as much as six years before the mining facilities will be operational. And during that time, Resolution Copper is required by the terms of § 3003 to keep Oak Flat accessible to “members of the public, including Indian tribes, to the maximum extent practicable, consistent with health and safety requirements.” 16 U.S.C. § 539p(i)(3). Even so, the Government conceded at argument that “the access will end before subsidence occurs, because it wouldn’t be safe to have people accessing the land when it could subside.” Once the mine is operational, the Forest Service estimates that it will produce ore for at least 40 years before closure and reclamation activities commence to decommission the mine.

C On January 4, 2021, the Forest Service announced that the EIS for the land transfer would be published in 11 days, on January 15. That publication would trigger the 60-day window for the federal Government to transfer title to the land. ^[16] U.S.C. § 539p(c)(10). Seeking to halt the transfer, Apache Stronghold sued the federal Government and its relevant officials on January 12, requesting declaratory relief, “a permanent injunction prohibiting” the “Land Exchange Mandate,” and ancillary fees and costs. Three days later, on January 15, the Government released the EIS as planned.

Apache Stronghold asserted several different claims in support of its prayer for relief. First, it alleged that the Government provided too little advance notice of the publication of the EIS, thereby infringing Apache Stronghold’s members’ rights under the Due Process Clause and under the Petition Clause of the First Amendment. Next, Apache Stronghold alleged that the land transfer would violate its members’ rights under the 1852 Treaty of Sante Fe. As this treaty-based claim has been described by Apache Stronghold in this court, the 1852 treaty assertedly imposed fiduciary trust obligations on the Government to “protect the traditional uses of ancestral lands,” even if the Government “has formal title to the land.” The transfer would allegedly violate the treaty—and this corresponding federal trust obligation—because it would “allow total destruction” of the property and prevent the Western Apache from conducting their traditional religious practices.

Apache Stronghold also argued that the transfer would violate its members’ rights under the Free Exercise Clause of the First Amendment and under RFRA. With respect to its Free Exercise Clause claim, Apache Stronghold argued that § 3003 was not a neutral law of general applicability and was therefore subject to strict scrutiny. See Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 546 (1993). And, according to Apache Stronghold, the transfer was neither in support of a compelling governmental interest nor narrowly tailored to accomplish such an interest. As to RFRA, Apache Stronghold argued that the land exchange “chills, burdens, inhibits, and destroys” the religious exercise of its members, thus substantially burdening their exercise of religion in violation of RFRA. As with the Free Exercise Clause claim, Apache Stronghold’s RFRA claim asserted that the transfer was not narrowly tailored to accomplish a compelling governmental interest. See 42 U.S.C. § 2000bb-1(b). Lastly, Apache Stronghold alleged that the federal Government intentionally discriminated against its members on account of their religion in violation of the Free Exercise Clause.

Two days after filing suit, Apache Stronghold moved for a temporary restraining order (“TRO”) and preliminary injunction. Specifically, Apache Stronghold sought an order “preventing Defendants from publishing a Final Environmental Impact Statement . . . and from conveying the parcel(s) of land containing Oak Flat.”

On January 14, 2021, the district court denied Apache Stronghold’s motion for a TRO. After conducting an evidentiary hearing on February 3, the district court denied the preliminary injunction motion on February 12. Because the district court concluded that Apache Stronghold had not demonstrated “a likelihood of success on, or serious questions going to, the merits” of its claims, the district court did not consider the remaining preliminary injunction factors. See Apache Stronghold, 519 F. Supp. 3d at 598, 611. Apache Stronghold timely appealed.

On March 1, 2021, during the pendency of this appeal, the Government withdrew its EIS for the land transfer and mine. It explained that “additional time is necessary to fully understand concerns raised by Tribes” and to “ensure[] the agency’s compliance with federal law.” To date, the Government has provided the court no concrete estimate of when the EIS will be issued, except to pledge that it is not awaiting the decision in this case and to state that it will provide the court and Apache Stronghold at least 60 days’ notice prior to issuing the EIS.

II We have jurisdiction under 28 U.S.C. § 1292(a)(1). We review the district court’s refusal to issue a preliminary injunction for abuse of discretion. See AK Futures LLC v. Boyd Street Distro, LLC, 35 F.4th 682, 688 (9th Cir. 2022). We review the district court’s “underlying legal conclusions de novo” and its “factual findings for clear error.” Id.

To show that it is entitled to a preliminary injunction, Apache Stronghold “must establish [1] that [it] is likely to succeed on the merits, [2] that [it] is likely to suffer irreparable harm in the absence of preliminary relief, [3] that the balance of equities tips in [its] favor, and [4] that an injunction is in the public interest.” Winter v. Natural Res. Def. Council, Inc., 555 U.S. 7, 20 (2008). The first factor— likelihood of success on the merits—is “the most important,” and “when a plaintiff has failed to show the likelihood of success on the merits, we need not consider the remaining three [factors].” Garcia v. Google, Inc., 786 F.3d 733, 740 (9th Cir. 2015) (en banc) (citations and internal quotation marks omitted). In this court, Apache Stronghold only challenges the district court’s likelihood-of-success determination with respect to its claims under the Free Exercise Clause, RFRA, and the 1852 treaty. Because, as we shall explain, Apache Stronghold has no likelihood of success on any of those three claims, we have no occasion to address the remaining Winter factors.

III Apache Stronghold asserts that the transfer of Oak Flat from the Government to Resolution Copper would “violate the Free Exercise Clause.” This claim fails under the Supreme Court’s controlling decision in Lyng v. Northwest Indian Cemetery Protective Association, 485 U.S. 439 (1988).

A The dispute in Lyng arose from the Government’s long- running effort to build a road connecting the northwest California towns of Gasquet and Orleans (the “G-O road”). 485 U.S. at 442. One of the final components of that project involved the construction of “a 6-mile paved segment through the Chimney Rock section of the Six Rivers National Forest,” a section that had “historically been used for religious purposes by Yurok, Karok, and Tolowa Indians.” Id. As part of its preparation of a final environmental impact statement concerning the completion of the road through Chimney Rock, the Forest Service “commissioned a study of the American Indian cultural and religious sites in the area.” Id. That study recommended against completion of the road, because “any of the available routes ‘would cause serious and irreparable damage to the sacred areas which are an integral and necessary part of the belief systems and lifeway of Northwest California Indian peoples.’” Id. (citation omitted). The Forest Service nonetheless decided to proceed with the construction of the road. Id. at 443. “At about the same time, the Forest Service adopted a management plan allowing for the harvesting of significant amounts of timber in this area of the forest.” Id.

The Forest Service’s actions were promptly challenged in a federal lawsuit brought by “an Indian organization, individual Indians,” the State of California, and others. Lyng, 485 U.S. at 443. The district court permanently enjoined both the timber management plan and the construction of the remaining section of the road, holding that these actions would infringe the rights of tribal members under the Free Exercise Clause as well as violate other provisions of federal law. Id. at 443–44. While the case was pending on appeal in this court, Congress intervened by enacting the California Wilderness Act of 1984, Pub. L. No. 98-425, 98 Stat. 1619 (1984). See Lyng, 485 U.S. at 444. That statute designated much of the land governed by the Forest Service’s timber management plan as protected wilderness, thereby barring “commercial activities such as timber harvesting.” Id. However, the Act specifically “exempt[ed] a narrow strip of land, coinciding with the Forest Service’s proposed route for the remaining segment of the G-O road, from the wilderness designation.” Id. This was done precisely “to enable the completion of the Gasquet-Orleans Road project if the responsible authorities so decide.” Id. (quoting S. REP. NO. 98-582, at 29 (1984)). A panel of this court subsequently vacated the district court’s injunction to the extent that it had been mooted by the wilderness designations in the California Wilderness Act, but otherwise largely affirmed the district court. See Northwest Indian Cemetery Protective Ass’n v. Peterson, 795 F.2d 688, 698 (9th Cir. 1986); see also Lyng, 485 U.S. at 444–45.

The Supreme Court reversed. In addressing the Free Exercise Clause issue, which was a necessary component of the relief granted by the district court, the Court began by acknowledging that “[i]t is undisputed that the Indian [plaintiffs’] beliefs are sincere and that the Government’s proposed actions will have severe adverse effects on the practice of their religion.” Lyng, 485 U.S. at 447. As the Court explained, it was undisputed that the “projects at issue in this case could have devastating effects on traditional Indian religious practices,” and the Court therefore accepted the premise that “the G-O road will virtually destroy the Indians’ ability to practice their religion.” Id. at 451 (simplified); see also id. (acknowledging that the threat to the Indian plaintiffs’ “religious practices is extremely grave”). Despite these acknowledged severe impacts, the Court nonetheless held that the Government was not required to demonstrate a “compelling need” or otherwise to satisfy strict scrutiny. Id. at 447. That was true, the Court held, because the plaintiffs would not “be coerced by the Government’s action into violating their religious beliefs,” nor would that action “penalize religious activity by denying any person an equal share of the rights, benefits, and privileges enjoyed by other citizens.” Id. at 449.

The Court held that the case was, in that respect, comparable to Bowen v. Roy, 476 U.S. 693 (1986), in which the Court rejected a Free Exercise challenge to a federal statute “that required the States to use Social Security numbers in administering certain welfare programs.” Lyng, 485 U.S. at 448–49. The plaintiffs in Roy contended that the governmental assignment of a “numerical identifier” would seriously impede their ability to practice their religion by “rob[bing] the spirit of their daughter and prevent[ing] her from attaining greater spiritual power.” Id. at 448 (simplified) (quoting Roy, 476 U.S. at 696). Although the result would be a significant interference with the Roy plaintiffs’ religious beliefs, the Roy Court held that the challenged governmental action—the state and federal governments’ “internal” use of a Social Security number— nonetheless did not implicate the Free Exercise Clause. Id. As the Court explained, “[t]he Free Exercise Clause simply cannot be understood to require the Government to conduct its own internal affairs in ways that comport with the religious beliefs of particular citizens.” Id. (quoting Roy, 476 U.S. at 699). “The Free Exercise Clause affords an individual protection from certain forms of governmental compulsion; it does not afford an individual a right to dictate the conduct of the Government’s internal procedures.” Id. (quoting Roy, 476 U.S. at 700).

The Lyng Court acknowledged that “[i]t is true that this Court has repeatedly held that indirect coercion or penalties on the free exercise of religion, not just outright prohibitions, are subject to scrutiny under the First Amendment.” 485 U.S. at 450 (emphasis added). Such indirect coercion or penalties would include a denial of program benefits “based solely” on the claimant’s religious beliefs and practices, as well as any other denial of “an equal share of the rights, benefits, and privileges enjoyed by other citizens.” Id. at 449–50. But the Court held that the Free Exercise Clause’s protection against government conduct “prohibiting” the free exercise of religion, see U.S. CONST. amend. I, does not protect against the “incidental effects of government programs, which may make it more difficult to practice certain religions but which have no tendency to coerce individuals into acting contrary to their religious beliefs.” Id. at 450; see also id. at 451 (noting that the “crucial word in the constitutional text is ‘prohibit’”).

In light of these principles, the Court concluded, the claim in Lyng could not “meaningfully be distinguished” from that in Roy. Lyng, 485 U.S. at 449. Although the resulting effects on the religious practices of the Indian plaintiffs would “virtually destroy” their “ability to practice their religion,” those religious impacts nonetheless did not implicate the Free Exercise Clause because the governmental actions that caused them had “no tendency to coerce individuals into acting contrary to their religious beliefs.” Id. at 450–51. Nor was this a situation in which the Government had “discriminate[d]” against the plaintiffs, as might be the case if Congress had passed “a law prohibiting the Indian [plaintiffs] from visiting the Chimney Rock area.” Id. at 453. According to the Court, the Indian plaintiffs sought, not “an equal share of the rights, benefits, and privileges enjoyed by other citizens,” but rather a “religious servitude” that would “divest the Government of its right to use what is, after all, its land.” Id. at 449, 452– 53.

The project challenged here is indistinguishable from that in Lyng. Here, just as in Lyng, the Government’s actions with respect to “publicly owned land” would “interfere significantly with private persons’ ability to pursue spiritual fulfillment according to their own religious beliefs,” but it would have “no tendency to coerce” them “into acting contrary to their religious beliefs.” 485 U.S. at 449–50. And just as with the land use decisions at issue in Lyng, the challenged transfer of Oak Flat for mining operations does not “discriminate” against Apache Stronghold’s members, “penalize” them, or deny them “an equal share of the rights, benefits, and privileges enjoyed by other citizens.” Id. at 449, 453. Under Lyng, Apache Stronghold seeks, not freedom from governmental action “prohibiting the free exercise” of religion, see U.S. CONST. amend. I, but rather a “religious servitude” that would uniquely confer on tribal members “de facto beneficial ownership of [a] rather spacious tract[] of public property.” Lyng, 485 U.S. at 452– 53. Under Lyng, Apache Stronghold’s Free Exercise Clause claim must be rejected.

B Apache Stronghold’s various arguments for distinguishing Lyng are all without merit.

First, Apache Stronghold argues that Lyng is distinguishable because, in that case, the virtual destruction of the “Indians’ ability to practice their religion” was accomplished without actually destroying any “sites where specific rituals take place.” 485 U.S. at 451, 454. According to Apache Stronghold, Lyng’s holding is limited to cases involving only interference with “subjective” spiritual experiences and therefore does not apply to a case, such as this one, involving “physical destruction of a sacred site.” Although the dissent does not directly address the merits of Apache Stronghold’s Free Exercise Clause claim, see Dissent at 197, the dissent’s discussion of Lyng (undertaken in the context of analyzing RFRA) seeks to distinguish the case on the comparable ground that the project at issue there would not have precluded physical access to the relevant sacred sites, see Dissent at 220–26. These efforts to distinguish Lyng are refuted by Lyng itself.

In Lyng, the State of California argued that Roy was distinguishable on the ground that it involved only interference with the plaintiffs’ “religious tenets from a subjective point of view,” whereas Lyng involved a “proposed road [that] will ‘physically destroy the environmental conditions and the privacy without which the religious practices cannot be conducted.’” 485 U.S. at 449 (simplified) (emphasis added). The Court rejected this proffered subjective/physical distinction, expressly holding that there was no permissible basis to “say that the one form of incidental interference with an individual’s spiritual activities should be subjected to a different constitutional analysis than the other.” Id. at 449–50. This holding requires rejection of Apache Stronghold’s analogous proffered distinction between interference with subjective experiences and physical destruction of the means of conducting spiritual exercises.

The dissent contends that “Lyng did not specifically address government action that prevented religious exercise,” and that it therefore does not apply to a case, such as this one, in which the Government’s actions will physically destroy the site and thereby literally prevent its future use for religious purposes. See Dissent at 228–29 (emphasis added). This effort to distinguish Lyng also fails, because, once again, it ultimately relies on too expansive a notion of what counts as “prohibiting” the free exercise of religion. We readily agree that “prevent” can often be synonymous with “prohibit,” see Prohibit, WEBSTER’S THIRD NEW INTERNATIONAL DICTIONARY 1813 (1981 ed.) (“WEBSTER’S THIRD”) (“to prevent from doing or accomplishing something”), and in that sense it is true that “prevent[ing] the plaintiff from participating in an activity motivated by a sincerely held religious belief” qualifies as prohibiting free exercise. Yellowbear v. Lampert, 741 F.3d 48, 55 (10th Cir. 2014) (citing, inter alia, Lyng, 485 U.S. at 450); see also Graham v. Comm’r, 822 F.2d 844, 850–51 (9th Cir. 1987). But “prevent” also can have the broader sense of “frustrate,” “keep from happening,” or “hinder,” which is how the dissent uses the term here. See Prevent, WEBSTER’S THIRD, supra, at 1798. Lyng squarely rejected that broader notion of “prohibiting the free exercise” of religion:

The dissent begins by asserting that the “constitutional guarantee we interpret today . . . is directed against any form of government action that frustrates or inhibits religious practice.” The Constitution, however, says no such thing. Rather, it states: “Congress shall make no law . . . prohibiting the free exercise [of religion].” 485 U.S. at 456 (emphasis altered) (citations omitted).

Thus, contrary to what the dissent posits, it is not enough under Lyng to show that the Government’s management of its own land and internal affairs will have the practical consequence of “preventing” a religious exercise. Indeed, Lyng explicitly rejected that broader notion of “prohibiting” religious exercise, concluding that it was foreclosed by Roy:

. . . Bowen v. Roy rejected a First Amendment challenge to Government activities that the religious objectors sincerely believed would “‘rob the spirit’ of [their] daughter and prevent her from attaining greater spiritual power.” The dissent now offers to distinguish that case by saying that the Government was acting there “in a purely internal manner,” whereas land-use decisions “are likely to have substantial external effects.” Whatever the source or meaning of the dissent’s distinction, it has no basis in Roy. Robbing the spirit of a child, and preventing her from attaining greater spiritual power, is both a “substantial external effect” and one that is remarkably similar to the injury claimed by [the plaintiffs] in the case before us today. The dissent’s reading of Roy would effectively overrule that decision, without providing any compelling justification for doing so.

Lyng, 485 U.S. at 456 (emphasis added) (citations and further quotation marks omitted).

Second, Apache Stronghold argues that Lyng is distinguishable because it involved application of a neutral and generally applicable law, inasmuch as “the road in Lyng was carried out pursuant to the California Wilderness Act of 1984.” By contrast, according to Apache Stronghold, this case involves legislative action directed at “one ‘particular property,’” which is the antithesis of a “generally applicable” law. The dissent also endorses this ground for distinguishing Lyng, arguing that Lyng merely stands for the “proposition that the compelling interest test is ‘inapplicable’ to ‘across-the-board’ neutral laws.” See Dissent at 229 (citation omitted). Once again, Lyng itself refutes this ground for attempting to distinguish that decision.

As Lyng itself makes clear, the California Wilderness Act was not a neutral and generally appliable law in the sense that Apache Stronghold posits, because it contained an express exemption for the “narrow strip of land” that exactly “coincid[ed] with the Forest Service’s proposed route for the remaining segment of the G-O road.” 485 U.S. at 444. Thus, contrary to what Apache Stronghold claims, the relevant provisions of the statute at issue in Lyng likewise involved legislative action directed at “one ‘particular property.’” Indeed, it was precisely this feature of the challenged actions in Lyng that the plaintiffs there sought to invoke as a ground for distinguishing Roy: whereas Roy involved the “mechanical” application of a general program requirement for the welfare program at issue, Lyng involved “a case-by- case substantive determination as to how a particular unit of land will be managed.” 485 U.S. at 449. In rejecting this effort to distinguish Roy, the Lyng Court did not dispute that such a distinction existed as a factual matter between the two cases. Instead, the Court held that the distinction simply provided no grounds for distinguishing Roy. Id. at 449–50. That was true, the Court explained, because the central ingredient of a Free Exercise Claim—some “tendency to coerce individuals into acting contrary to their religious beliefs”—was absent in both cases. Id. at 450. 4 The dissent claims that, even if the Lyng decision did not view itself as resting on a rule about neutral and generally applicable laws, Employment Division, Department of Human Resources of Oregon v. Smith, 494 U.S. 872 (1990), and other post-Smith decisions have read it that way. See 4 The dissent nonetheless insists that the Forest Service’s plan and the special legislative carve-out in Lyng—both of which were tailored for the specific property at issue—were “generally applicable” because “there was no indication” that they were “made because of, rather than in disregard of,” the religious interest in that particular property. See Dissent at 232–33 (emphasis added). This contention fails, because it mixes up the distinct issues of whether a particular law is “neutral” and whether it is “generally applicable.” Even if the plan and legislation at issue in Lyng were “neutral” in the limited sense that it was not their “object . . . to infringe upon or restrict practices because of their religious motivation,” Church of the Lukumi, 508 U.S. at 533 (emphasis added), they were plainly not “generally applicable” as that phrase is currently understood, given that they were directed at one particular property. See, e.g., International Church of the Foursquare Gospel v. City of San Leandro, 673 F.3d 1059, 1066 (9th Cir. 2011) (“In this case, while the zoning scheme itself may be facially neutral and generally applicable, the individualized assessment that the City made to determine that the Church’s rezoning and CUP request should be denied is not.” (emphasis added)).

Dissent at 229–31. That is not correct. All that the Court has stated is that Smith and its progeny “drew support for [Smith’s] neutral and generally applicable standard from cases involving internal government affairs,” such as Lyng. Fulton v. City of Philadelphia, 593 U.S. 522, 536 (2021) (emphasis added). Thus, in Smith, the Court stated that its core holding—i.e., that strict scrutiny does not apply to neutral laws of general applicability—was supported by Lyng’s broader observation that the boundaries of the Free Exercise Clause “cannot depend on measuring the effects of a governmental action on a religious objector’s spiritual development.” 494 U.S. at 885 (quoting Lyng, 485 U.S. at 451).

But the Court has not said, and could not have said, that the holding of Lyng rested on the view that Lyng was itself a case involving a neutral and generally applicable law. As we have set forth, Lyng rested on a holding about the scope of the term “prohibiting” under the Free Exercise Clause and never mentioned or endorsed a Smith-style rule. At most, the Court has suggested in dicta that Lyng fits a pattern of cases in which the Court had upheld laws that were “neutral and generally applicable without regard to religion” in the sense that they did not “‘penalize religious activity by denying any person an equal share of the rights, benefits, and privileges enjoyed by other citizens.’” Trinity Lutheran Church of Columbia, Inc. v. Comer, 582 U.S. 449, 460 (2017) (quoting Lyng, 485 U.S. at 449). But Trinity Lutheran did not have before it the more focused question whether, in light of the parcel-specific rigging of the statutory framework in Lyng, the underlying statute at issue in Lyng could be properly deemed to qualify as “neutral and generally applicable” under the details of Smith’s framework. As we have explained, Lyng involved a situation in which, after religious objections had been raised to the G-O road and the road’s construction had been enjoined, Congress proceeded to adopt an explicit statutory gerrymander for the precise parcel at issue. See supra at 27– 28. That manifestly would not fit the Court’s current understanding of a case involving a neutral and generally applicable law. See, e.g., Church of the Lukumi, 508 U.S. at 542 (emphasizing that “categories of selection” in legislative drafting “are of paramount concern when a law has the incidental effect of burdening religious practice”). In all events, even if the law in Lyng were deemed, in hindsight, to be neutral and generally applicable within the meaning of Smith, the fact remains that the holding of Lyng did not rest on any such premise, but instead on the view that the challenged actions there lacked the sort of features that would qualify as “prohibiting” the free exercise of religion.

The dissent also points to Lyng’s observation that, because the “Constitution does not permit government to discriminate against religions that treat particular physical sites as sacred,” a “law prohibiting the Indian respondents from visiting the Chimney Rock area would raise a different set of constitutional questions.” 485 U.S. at 453 (emphasis added); see also Dissent at 225. According to the dissent, “the Land Transfer Act is exactly that kind of ‘prohibitory’ law.” See Dissent at 225. That contention is refuted by the fact that, under the statute, any post-transfer prohibitions that Resolution Copper may impose on public access to Oak Flat would be nondiscriminatory. See 16 U.S.C. § 539p(i)(3) (stating that, “[a]s a condition of conveyance,” Resolution Copper must “provide access to the surface of the Oak Flat Campground to members of the public, including Indian tribes, to the maximum extent practicable . . . until such time as the operation of the mine precludes continued public access for safety reasons”). To the extent that the dissent instead reads Lyng as endorsing the broader notion that the Free Exercise Clause would be violated by a nondiscriminatory law that will ultimately have the effect of precluding public access to a particular parcel of land, that view cannot be squared with Lyng’s explicit rejection of such a broad concept of “prohibiting.” Indeed, under the dissent’s expansive view, any transfer of Government land without a condition guaranteeing access to a sacred site on that parcel would amount to a prohibition on free exercise. Lyng, however, explicitly rejects the view that the Free Exercise Clause requires any such “religious servitude” on Government land, which would confer “de facto beneficial ownership of some rather spacious tracts of public property.” 485 U.S. at 452–53.

In sum, Lyng stands for the proposition that a disposition of government real property is not subject to strict scrutiny when it has “no tendency to coerce individuals into acting contrary to their religious beliefs,” does not “discriminate” against religious adherents, does not “penalize” them, and does not deny them “an equal share of the rights, benefits, and privileges enjoyed by other citizens.” Lyng, 485 U.S. at 449–50, 453. In such circumstances, the essential ingredient of “prohibiting” the free exercise of religion is absent, and the Free Exercise Clause is not violated. And because Lyng’s application of that rule in the context of that case cannot meaningfully be distinguished in this case, Apache Stronghold has no likelihood of success on its Free Exercise claim.

IV Apache Stronghold also contends that the sale of Oak Flat to Resolution Copper would violate its members’ rights under RFRA. Congress enacted RFRA in 1993 “in direct response” to Smith’s narrow construction of the Free Exercise Clause, see City of Boerne v. Flores, 521 U.S. 507, 512 (1997), and Congress did so precisely “in order to provide greater protection for religious exercise than is available” under the Free Exercise Clause as construed in Smith, see Holt v. Hobbs, 574 U.S. 352, 357 (2015). The question here is whether the broader protection afforded by RFRA has the practical effect of displacing, by statute, the pre-Smith decision in Lyng. The answer to that question is no.

A In order to understand what RFRA enacts, it is important to begin with the decision that RFRA sought to supersede, namely, Employment Division v. Smith.

Smith involved a denial of unemployment benefits to two Oregon workers who “were fired from their jobs with a private drug rehabilitation organization because they ingested peyote for sacramental purposes at a ceremony of the Native American Church, of which both [were] members.” 494 U.S. at 874. The claimants appealed that denial of benefits to the Oregon Court of Appeals, which held that the denial violated the Free Exercise Clause. Id. On the State’s further appeal, the Oregon Supreme Court agreed. Id. at 875. The U.S. Supreme Court granted certiorari, but it initially held only that, “if a State has prohibited through its criminal laws certain kinds of religiously motivated conduct without violating the First Amendment, it certainly follows that it may impose the lesser burden of denying unemployment compensation benefits to persons who engage in that conduct.” Employment Div., Dep’t of Human Res. of Oregon v. Smith, 485 U.S. 660, 670 (1988). The Court therefore remanded the case to the Oregon Supreme Court to address “whether [the plaintiffs’] sacramental use of peyote was in fact proscribed by Oregon’s controlled substance law.” Smith, 494 U.S. at 875. On remand, the Oregon Supreme Court answered that question in the affirmative and otherwise “reaffirmed its previous ruling” in the plaintiffs’ favor. Id. at 876. The U.S. Supreme Court again granted review. Id. Thus, although Smith had started out as an unemployment compensation case, it returned to the Supreme Court as squarely presenting the question of whether Oregon’s criminal prohibition on all use of peyote violated the Free Exercise Clause. Id. Accordingly, unlike Lyng, Smith presented no threshold question as to whether the challenged Oregon law actually “prohibit[ed]” the claimants’ religious exercise. See U.S. CONST. amend I.

A sharply divided Court held that there was no violation of the Free Exercise Clause. Justice Scalia’s majority opinion for five Justices acknowledged what it described as “the balancing test set forth in Sherbert v. Verner, 374 U.S. 398 (1963),” under which “governmental actions that substantially burden a religious practice must be justified by a compelling governmental interest.” Smith, 494 U.S. at 883. The Court noted that it had applied the Sherbert test in three cases to “invalidate[] state unemployment compensation rules that conditioned the availability of benefits upon an applicant’s willingness to work under conditions forbidden by his religion.” Id. The Court also observed that, in several other decisions, the Court “purported to apply the Sherbert test in contexts other than that,” but that it had “always found the test satisfied.” Id. Citing specifically to (among other decisions) Roy and Lyng, the Court further noted that, “[i]n recent years [the Court] ha[s] abstained from applying the Sherbert test (outside the unemployment compensation field) at all.” Id. The Court then held that, “[e]ven if we were inclined to breathe into Sherbert some life beyond the unemployment compensation field, we would not apply it to require exemptions from a generally applicable criminal law.” Id. at 884 (emphasis added). Reviewing its caselaw more broadly, the Court held that its decisions had “consistently held that the right of free exercise does not relieve an individual of the obligation to comply with a ‘valid and neutral law of general applicability on the ground that the law proscribes (or prescribes) conduct that his religion prescribes (or proscribes).’” Id. at 879 (citation omitted). Citing Lyng, the Court held that “[t]he government’s ability to enforce generally applicable prohibitions of socially harmful conduct, like its ability to carry out other aspects of public policy, ‘cannot depend on measuring the effects of a governmental action on a religious objector’s spiritual development.’” Id. at 885 (quoting Lyng, 485 U.S. at 451).

The Court’s holding that the Sherbert test does not apply to neutral and generally applicable prohibitions drew the sharp disagreement of four Justices, in a separate opinion written by Justice O’Connor. 5 According to Justice O’Connor, the Court’s caselaw has “respected both the First Amendment’s express textual mandate and the 5 Because Justice O’Connor ultimately concurred in the judgment even under the Sherbert test, her separate opinion was technically styled as a concurrence in the judgment. See Smith, 494 U.S. at 891–907. The other three Justices who joined Justice O’Connor’s criticism of the majority’s abandonment of the Sherbert test did not agree that the Oregon law survived that test, and they therefore only partially joined her concurrence and also filed a separate dissent. See id. at 907–21 (Blackmun, J., dissenting).

governmental interest in regulation of conduct by requiring the government to justify any substantial burden on religiously motivated conduct by a compelling state interest and by means narrowly tailored to achieve that interest.” Smith, 494 U.S. at 894 (O’Connor, J., concurring in the judgment). Citing the unemployment compensation case of Thomas v. Review Board of the Indiana Employment Security Division, 450 U.S. 707 (1981), Justice O’Connor elaborated on her understanding of what it meant for government to impose a substantial burden on religious exercise:

[T]he essence of a free exercise claim is relief from a burden imposed by government on religious practices or beliefs, whether the burden is imposed directly through laws that prohibit or compel specific religious practices, or indirectly through laws that, in effect, make abandonment of one’s own religion or conformity to the religious beliefs of others the price of an equal place in the civil community. As [the Court] explained in Thomas:

“Where the state conditions receipt of an important benefit upon conduct proscribed by a religious faith, or where it denies such a benefit because of conduct mandated by religious belief, thereby putting substantial pressure on an adherent to modify his behavior and to violate his beliefs, a burden upon religion exists.” 450 U.S., at 717–718.

Smith, 494 U.S. at 897 (O’Connor, J., concurring in the judgment). Thus, Justice O’Connor concluded, “[t]he Sherbert compelling interest test applies” to both “cases in which a State conditions receipt of a benefit on conduct prohibited by religious beliefs and cases in which a State affirmatively prohibits such conduct.” Id. at 898. In either type of case, Justice O’Connor concluded, it did not matter whether the law was a “neutral” or “generally applicable” one. Id. at 898–900. The Court’s precedents, she explained, reflected a “consistent application of free exercise doctrine to cases involving generally applicable regulations that burden religious conduct.” Id. at 892.

B Congress promptly sought to supersede, by statute, Smith’s holding that “neutral, generally applicable laws that incidentally burden the exercise of religion usually do not violate the Free Exercise Clause.” Holt, 574 U.S. at 356–57. As stated expressly in § 2 of RFRA, Congress’s primary purpose in enacting the Act was to “restore the compelling interest test as set forth in Sherbert v. Verner, 374 U.S. 398 (1963) and Wisconsin v. Yoder, 406 U.S. 205 (1972) and to guarantee its application in all cases where free exercise of religion is substantially burdened.” 42 U.S.C. § 2000bb(b)(1). That stated purpose was based on RFRA’s express finding that “laws ‘neutral’ toward religion may burden religious exercise as surely as laws intended to interfere with religious exercise.” Id. § 2000bb(a)(1).

Section 3(a) of RFRA establishes the general rule that “[g]overnment shall not substantially burden a person’s exercise of religion even if the burden results from a rule of general applicability.” 42 U.S.C. § 2000bb-1(a). In its current form, that prohibition extends to any “branch, department, agency, instrumentality, [or] official (or other person acting under color of law) of the United States” or of the District of Columbia, the Commonwealth of Puerto Rico, or the United States’ territories and possessions. Id. § 2000bb-2(1), (2). The sole exception to this general rule is contained in § 3(b), which states:

Government may substantially burden a person’s exercise of religion only if it demonstrates that application of the burden to the person— (1) is in furtherance of a compelling governmental interest; and (2) is the least restrictive means of furthering that compelling governmental interest.

Id. § 2000bb-1(b). The net effect is that the government may substantially burden a person’s exercise of religion if and only if the government’s action can survive “strict scrutiny.” See Gonzales v. O Centro Espírita Beneficente União do Vegetal, 546 U.S. 418, 430 (2006).

Congress also made clear its intent that RFRA operate as a framework statute, “displacing the normal operation of other federal laws.” Bostock v. Clayton Cnty., 590 U.S. 644, 682 (2020). Specifically, § 6 of RFRA provides that the Act “applies to all Federal law, and the implementation of that law, whether statutory or otherwise, and whether adopted before or after” the date of RFRA’s enactment. 42 U.S.C. § 2000bb-3(a). Congress further provided that “[f]ederal statutory law adopted after [RFRA’s enactment] is subject to [RFRA] unless such law explicitly excludes such application by reference to [RFRA].” Id. § 2000bb-3(b).

RFRA does not define what it means to “substantially burden a person’s exercise of religion.” 42 U.S.C. § 2000bb- 1(a), (b). But “Congress legislates against the backdrop of existing law,” McQuiggin v. Perkins, 569 U.S. 383, 398 n.3 (2013), and the meaning of that phrase is clearly elucidated by considering the body of law discussed in the “separate opinions” in Smith, which “concerned the very issue addressed” by Congress in § 3 of RFRA. Williams v. Taylor (Terry Williams), 529 U.S. 362, 411 (2000). 6 As Terry Williams explained, in the unusual situation in which the “broader debate and the specific statements” of the Justices in a particular decision “concern[] precisely the issue” that Congress later addresses in a statute that borrows the Justices’ terminology, Congress should be understood to have “adopt[ed]” the relevant “meaning given a certain term in that decision.” 529 U.S. at 411–12. Thus, in construing the standards of review applicable in deciding habeas corpus petitions under the Antiterrorism and Effective Death Penalty Act (“AEDPA”), Terry Williams turned to “[t]he separate opinions” in Wright v. West, 505 U.S. 277 (1992), which concerned that “very issue.” 529 U.S. at 411. As Terry Williams recounted, the respective opinions of Justice 6 We refer to this case as “Terry Williams” because, in an extraordinary coincidence, the Supreme Court on the very same day decided another case named “Williams v. Taylor” (in which the petitioner was Michael Williams). See 529 U.S. 420 (2000); see also Shinn v. Martinez Ramirez, 596 U.S. 366, 381 (2022) (similarly referring to the other case as “Michael Williams”).

Thomas and Justice O’Connor in Wright vigorously debated whether habeas review should be deferential, with Justice O’Connor concluding that a federal court should review de novo whether the state court’s resolution of the federal issue was “correct,” and Justice Thomas concluding that a federal court should “simply” inquire as to whether the state decision was “reasonable.” Id. at 410–11. In addressing the issue of the appropriate standards of review in AEDPA’s amendments to the habeas statute, see 28 U.S.C. § 2254, “Congress specifically used the word ‘unreasonable,’” thereby confirming that it had effectively adopted Justice Thomas’s position and rejected Justice O’Connor’s. See Terry Williams, 529 U.S. at 411.

RFRA presents exactly the sort of distinctive situation in which the principles discussed in Terry Williams are applicable. Terry Williams invoked those principles with respect to AEDPA even though the Court conceded that there was “no indication in § 2254(d)(1) itself that Congress was ‘directly influenced’ by Justice Thomas’ opinion in Wright.” 529 U.S. at 411 (emphasis added). As the Court explained, “Congress need not mention a prior decision of this Court by name in a statute’s text in order to adopt either a rule or a meaning given a certain term in that decision.” Id. But where, as with RFRA, Congress does specifically “mention a prior decision of this Court by name in a statute’s text,” id., the inference is all the more inescapable that, when Congress borrows the Justices’ same phrasing, it does so against the backdrop of how those terms were understood in the relevant opinions accompanying that decision. Here, RFRA was enacted against the backdrop of the vigorous debate between Justice Scalia and Justice O’Connor in Smith; both of their opinions used variations of the phrase “substantially burden” in describing the pre-Smith framework for evaluating Free Exercise Clause claims 7; RFRA’s text states that its purpose is to supersede, by statute, the decision in “Employment Division v. Smith, 494 U.S. 872 (1990),” see 42 U.S.C. § 2000bb(a)(4); and, in superseding Smith, RFRA uses the phrase “substantially burden,” id. § 2000b-1(a), (b). The inference is overwhelming that Congress thereby “adopt[ed]” the “meaning given [that] certain term in that decision.” Terry Williams, 529 U.S. at 411. Consequently, RFRA unmistakably sought to enshrine, by statute, the basic principles reflected in the pre-Smith framework for applying the Free Exercise Clause that is described in those opinions, and that framework clearly includes Lyng.

Thus, for example, Justice O’Connor’s separate opinion in Smith confirms that the “substantial burden” rule established in the Court’s caselaw is consistent with, and does not abrogate, the Court’s decision in Lyng (which she wrote). As Justice O’Connor explained in her separate opinion in Smith, Lyng did not “signal” a “retreat from [the Court’s] consistent adherence to the compelling interest test” in evaluating governmental action prohibiting the free exercise of religion; instead, it reflected the underlying limits in the governmental conduct reached by the Free Exercise Clause. Smith, 494 U.S. at 900 (O’Connor, J., concurring in the judgment). She argued that, like Roy, Lyng involved the 7 See Smith, 494 U.S. at 883 (“Under the Sherbert test, governmental actions that substantially burden a religious practice must be justified by a compelling governmental interest.” (emphasis added)); id. at 894 (O’Connor, J., concurring in the judgment) (stating that, under the Court’s existing caselaw, the government is required “to justify any substantial burden on religiously motivated conduct by a compelling state interest and by means narrowly tailored to achieve that interest” (emphasis added)).

Government’s “conduct [of] its own internal affairs” in a way that did not implicate the Free Exercise Clause’s rule about “what the government cannot do to the individual.” Id. (emphasis added) (citation omitted). That view is consistent with Lyng, which—as we have exhaustively explained earlier—rests on the premise that the Government’s actions there, although substantially destructive of the Indians’ religious interests, did not involve “prohibiting the free exercise” of religion within the meaning of the Free Exercise Clause. See supra at 28–31.

Moreover, Justice O’Connor’s Smith concurrence contained a detailed explication of what counts as a cognizable burden under the Court’s then-existing caselaw, and it closely dovetails with Lyng. As she explained, such burdens may be “imposed directly through laws that prohibit or compel specific practices”; they may be imposed “indirectly through laws that, in effect, make abandonment of one’s own religion or conformity to the religious beliefs of others the price of an equal place in the civil community”; or they may involve benefit conditions that “put[] substantial pressure on an adherent to modify his behavior and to violate his beliefs.” Smith, 494 U.S. at 897 (O’Connor, J., concurring in the judgment) (emphasis added) (citation omitted).

Likewise, nothing in Justice Scalia’s majority opinion in Smith suggested that the Court thought that Lyng was inconsistent with the substantial burden test. Instead, in the course of arguing for a broader jettisoning of Sherbert’s compelling interest test, the Smith majority simply cited Lyng as an instance in which that strict scrutiny test had not been applied. See Smith, 494 U.S. at 883. As noted earlier, the Smith majority also argued that its broader position drew support from Lyng’s general observation that the limitations imposed by the Free Exercise Clause “cannot depend on measuring the effects of a governmental action on a religious objector’s spiritual development,” id. at 885 (quoting Lyng, 485 U.S. at 451), but that likewise reflects no criticism of Lyng’s holding about the scope of “prohibiting” under the Free Exercise Clause.

Indeed, the only debate that Justice Scalia and Justice O’Connor had concerning Lyng related to the majority’s use of this latter comment to bolster its broader rule about neutral laws of general applicability. Justice O’Connor objected that the majority took that comment out of Lyng’s specific context, which involved only the Government’s conduct of its “internal affairs” and therefore did not implicate the Free Exercise Clause’s rule about “what the government cannot do to the individual.” Smith, 494 U.S. at 900 (O’Connor, J., concurring in the judgment) (citation omitted). The Court responded that there was no basis for limiting the cited principle in the way that Justice O’Connor posited. Lyng’s observation should apply more broadly, the Court explained, because “it is hard to see any reason in principle or practicality why the government should have to tailor its health and safety laws to conform to the diversity of religious belief, but should not have to tailor its management of public lands, Lyng, supra, or its administration of welfare programs, Roy, supra.” Id. at 885 n.2. This debate about whether and how to extend an observation made in Lyng reflects no criticism of Lyng’s ultimate holding.

Accordingly, both Justice O’Connor’s concurrence and the majority opinion in Smith strongly confirm that, under the then-existing framework of Free Exercise Clause jurisprudence, the proposition that the government must justify, by strict scrutiny, any “substantial burden” on religious exercise is one that subsumes, rather than overrides, Lyng’s holding about the scope of government action that is reached by the constitutional phrase “prohibiting the free exercise thereof.” U.S. CONST. amend. I. As a decision about the scope of the term “prohibiting,” Lyng defines the outer bounds of what counts as a cognizable substantial burden imposed by the government. That is plainly how Justice O’Connor viewed Lyng in Smith, and the Smith majority did not disagree. When Congress copied the “substantial burden” phrase into RFRA, it must be understood as having similarly adopted the limits that Lyng placed on what counts as a governmental imposition of a substantial burden on religious exercise. See Terry Williams, 529 U.S. at 411–12; see also ANTONIN SCALIA & BRYAN A. GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS 322 (2012) (“If a statute uses words or phrases that have already received authoritative construction by the jurisdiction’s court of last resort, . . . they are to be understood according to that construction.”).

C The dissent’s exclusive reliance on its composite understanding of the dictionary definitions of “substantial” and “burden,” see Dissent at 201, contravenes the interpretive principles discussed in Terry Williams, as well as the crucial context supplied by Smith and Lyng. As a result, the dissent’s construction of the phrase elides the crucial ingredient that Lyng reflects, which is that the phrase “substantial burden” must ultimately be bounded by what counts as within the domain of the phrase “prohibiting the free exercise thereof.” U.S. CONST. amend. I (emphasis added).

It is no answer to say, as the dissent does, that we have applied that dictionary definition in construing the meaning of the identical term “substantial burden” as used in the Religious Land Use and Institutionalized Persons Act (“RLUIPA”). See Dissent at 208–10. The dissent overlooks the fact that RLUIPA expressly applies only to “substantial burdens” in two specific contexts—namely, “impos[ing] or implement[ing] a land use regulation,” 42 U.S.C. § 2000cc(a)(1), and restrictions on “a person residing in or confined to an institution” affiliated with a government, id. § 2000cc-1(a). See id. § 1997; see also Cutter v. Wilkinson, 544 U.S. 709, 715 (2005). Because both of these specific contexts inherently involve coercive restrictions, they do not raise a similar Lyng-type issue about the bounds of what counts as “prohibiting” religious exercise. In RLUIPA’s two specific contexts, where that crucial element is already baked in, the dictionary definitions of “substantial” and “burden” will adequately flesh out the concept of “substantial burden” against that backdrop. The same is true under RFRA, once it is recognized that RFRA preserves Lyng’s understanding of what counts as “prohibiting” the free exercise of religion. But the same is not true if, with respect to RFRA, the critical context supplied by Smith and Lyng is overlooked. That would yield a very different concept of “substantial burden” under RFRA, one that (unlike RLUIPA) is shorn of any requirement to show that the governmental action has a “tendency to coerce individuals into acting contrary to their religious beliefs,” “discriminate[s]” against religious adherents, “penalize[s]” them, or denies them “an equal share of the rights, benefits, and privileges enjoyed by other citizens.” Lyng, 485 U.S. at 449–50, 453. Nothing in RFRA indicates that Congress intended to eliminate this crucial element or to abrogate Lyng.

The dissent’s contrary conclusion that RFRA does supersede Lyng rests on the premise that Lyng was based on a Smith-style holding about neutral and generally applicable rules. See Dissent at 229–33. For the reasons that we have already explained, that premise is patently incorrect. The law at issue in Lyng was manifestly not generally applicable, and nothing in Lyng rests upon, or endorses, the broad rule later adopted in Smith. See supra at 28–29, 35–37. Indeed, the most that the Smith majority claimed was that one particular statement in Lyng should be extended in a way that would support differential treatment of neutral laws of general applicability. See Smith, 494 U.S. at 885.

The dissent is also wrong in asserting that a 2000 amendment to RFRA—enacted as part of RLUIPA— demonstrates Congress’s intent that RFRA not be tied to the constitutional understanding of what counts as “prohibiting” the free exercise of religion. See Dissent at 205–06. Prior to RLUIPA, RFRA defined the specific term “exercise of religion” to “mean[] the exercise of religion under the First Amendment to the Constitution.” See Pub. L. No. 103-141 § 5(4), 107 Stat. 1488, 1489 (1993). However, a circuit split developed as to whether, as a result, RFRA’s protections were limited to only those practices that are “central” to, or “mandated” by, a person’s faith. Compare Bryant v. Gomez, 46 F.3d 948, 949 (9th Cir. 1995) (adopting those limitations) with Mack v. O’Leary, 80 F.3d 1175, 1178–79 (7th Cir. 1996) (noting the circuit split and rejecting Bryant), vacated on other grounds, 522 U.S. 801 (1997). Congress, of course, cannot statutorily change the scope of the Free Exercise Clause as construed by the courts, but it could effectively abrogate decisions such as Bryant by decoupling RFRA’s definition of “exercise of religion” from the Free Exercise Clause and then giving it a broader meaning for purposes of RFRA. That is exactly what Congress did in RLUIPA. In § 7(a)(3) of RLUIPA, Congress rewrote the definition of “exercise of religion” in RFRA to mean “religious exercise, as defined in section 8 of the Religious Land Use and Institutionalized Persons Act of 2000 [42 U.S.C. § 2000cc- 5].” See Pub. L. No. 106-274, § 7(a)(3), 114 Stat. 803, 806 (2000). Section 8 of RLUIPA, in turn, defines “religious exercise” to mean “any exercise of religion, whether or not compelled by, or central to, a system of religious belief,” and further provides that the “use, building, or conversion of real property for the purpose of religious exercise shall be considered to be religious exercise.” See 42 U.S.C. § 2000cc-5(7)(A)–(B). But in thus decoupling the definition of what activities count as the “exercise of religion” from the Free Exercise Clause,” Congress did not alter the phrase “substantial burden,” nor did it suggest that that phrase should be understood as somehow being decoupled from any notion of what counts as “prohibiting” the free exercise of religion under pre-Smith caselaw. 8 The dissent further errs in contending that our construction of “substantial burden” here disregards the Supreme Court’s rejection of the view that “RFRA merely restored th[e] Court’s pre-Smith decisions in ossified form.” Burwell v. Hobby Lobby Stores, Inc., 573 U.S. 682, 715–16 8 To the extent that the dissent insinuates that the amended RFRA’s borrowing of RLUIPA’s definition of religious exercise has the effect of abrogating Lyng, see Dissent at 205–06, that is quite wrong. The dissent has not cited any authority—and we are aware of none—that would support the extraordinary proposition that RFRA and RLUIPA purport to grant freestanding rights to obtain otherwise unavailable access to the real property of others for religious use. Put simply, neither statute purports to grant persons a “religious servitude” over the property of others. Lyng, 485 U.S. at 452.

(2014); see also Dissent at 206. The proposition the Court rejected in Hobby Lobby was that RFRA protected only the particular collection of practices that happened to have been “specifically addressed in [the Court’s] pre-Smith decisions,” much like AEDPA requires a showing of “‘clearly established Federal law, as determined by the Supreme Court of the United States.’” Id. at 714 (quoting 28 U.S.C. § 2254(d)(1)). That “absurd” view, the Court explained, would mean that “resident noncitizen[s]” would not be protected by RFRA, given that there was no “pre-Smith case in which th[e] Court entertained a free-exercise claim brought by a resident noncitizen.” Id. at 715–16. Hobby Lobby thus does not stand for the quite different— and erroneous—proposition that RFRA is somehow exempt from the settled rule that “Congress legislates against the backdrop of existing law.” McQuiggin, 569 U.S. at 398 n.3. Indeed, even the dissent concedes that RFRA must be construed in light of “the Supreme Court’s pre-Smith Free Exercise jurisprudence.” See Dissent at 210–11; see also id. at 215 (noting that we have previously “relied on pre-Smith Free Exercise Clause cases to define substantial burden”).

* * * Accordingly, RFRA’s understanding of what counts as “substantially burden[ing] a person’s exercise of religion” must be understood as subsuming, rather than abrogating, the holding of Lyng. That holding therefore governs Apache Stronghold’s RFRA claim as well, and that claim therefore fails for the same reasons discussed earlier. See supra at 31.

V Finally, Apache Stronghold also argues that an 1852 treaty of “perpetual peace and amity” between the “Apache Nation of Indians” and the United States, see TREATY WITH THE APACHES, July 1, 1852, art. 2, 10 Stat. 979 (1853), created an enforceable trust obligation that would be violated by the transfer of Oak Flat. That trust obligation, Apache Stronghold argues, stems from Article 9 of the treaty, which provides, in relevant part, that Relying confidently upon the justice and the liberality of the [federal] government, and anxious to remove every possible cause that might disturb their peace and quiet, it is agreed by the aforesaid Apache’s [sic] that the government of the United States shall at its earliest convenience designate, settle, and adjust their territorial boundaries, and pass and execute in their territory such laws as may be deemed conducive to the prosperity and happiness of said Indians.

Id., art. 9; see also id., art. 11 (stating that “the government of the United States shall so legislate and act as to secure the permanent prosperity and happiness of said Indians”). Specifically, Apache Stronghold argues that the Government’s treaty obligation to “pass and execute . . . such laws as may be deemed conducive to the prosperity and happiness’” of the Apaches should be “construed to obligate the United States to preserve traditional Apache religious practices on their historic homeland.” Thus construed, Apache Stronghold contends, the Government’s obligations under the treaty override any power or obligation to transfer Oak Flat under § 3003. This contention fails. Even assuming arguendo that Apache Stronghold’s interpretation of the Government’s treaty obligations is correct, the Government’s statutory obligation to transfer Oak Flat under § 3003 clearly abrogates any contrary treaty obligation, not the other way around. ^[9] “Congress has the power to abrogate Indians’ treaty rights,” but Congress generally must “clearly express its intent to do so.” South Dakota v. Bourland, 508 U.S. 679, 687 (1993). To the extent that Apache Stronghold is correct in contending that the Government has a treaty-based trust obligation to retain Oak Flat for the benefit of the tribe and its members, § 3003 clearly and manifestly abrogates any such obligation. Section 3003 was passed to accomplish a single goal: to “authorize, direct, facilitate, and expedite the exchange of land between Resolution Copper and the United States.” 16 U.S.C. § 539p(a). The entirety of the statute is built around that ultimate objective. There are various preparatory requirements, like consultations and report generation, e.g., id. § 539p(c)(3), (c)(4), (c)(6)(A), (c)(9), and post-transfer rules about land disposition and 9 Although Apache Stronghold has adequately shown that its members face an imminent threatened injury in fact that is fairly traceable to the alleged treaty violation, see Susan B. Anthony List v. Driehaus, 573 U.S. 149, 157–58 (2014), the district court concluded that allowing its members to assert what it deemed to be the tribe’s treaty rights violated the “prudential requirement that a plaintiff ‘cannot rest his claim to relief on the legal rights or interests of third parties.’” Apache Stronghold, 519 F. Supp. 3d at 598 (quoting Warth v. Seldin, 422 U.S. 490, 499 (1975)). Because the parties’ dispute over this “prudential” requirement does not involve our subject matter jurisdiction, we are not required to resolve it before addressing the merits of the treaty issue. See FW/PBS, Inc. v. City of Dallas, 493 U.S. 215, 237 (1990) (finding that the relevant plaintiffs had Article III standing and then rejecting a claim on the merits after assuming arguendo that “prudential, jus tertii standing” was met); cf. Lexmark Int’l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 125–28 (2014) (clarifying that “‘prudential standing’ is a misnomer” and must be distinguished from the jurisdictional requirements of Article III (citation omitted)).

management, id. § 539p(d)(2), (e), (g), (h), but they all lead up to the transfer of Oak Flat. Indeed, § 3003 unambiguously states that, upon completion of the preparatory steps, “if Resolution Copper offers to convey to the United States all right, title, and interest of Resolution Copper in and to the non-Federal land, the Secretary is authorized and directed to convey to Resolution Copper, all right, title, and interest of the United States in and to the Federal land.” Id. § 539p(c)(1) (emphasis added). Section 3003’s clear direction that, after consultation with the tribe, the transfer shall occur simply cannot co-exist with Apache Stronghold’s claim that the treaty requires that it shall not occur. Section 3003 plainly abrogates any tribal treaty rights that would otherwise preclude the transfer. See Bourland, 508 U.S. at 687.

VI For the foregoing reasons, Apache Stronghold is unlikely to succeed on the merits of any of the three claims before this court. It consequently cannot show that it is entitled to preliminary injunctive relief, and we need not consider the remaining Winter factors. See Garcia, 786 F.3d at 740. The district court’s order denying Apache Stronghold’s motion for a preliminary injunction is therefore affirmed.

AFFIRMED.

BEA, Circuit Judge, dissenting in part and concurring in part, with whom Circuit Judge FORREST joins except for footnote one; Circuit Judge BENNETT joins with respect to Part II:

I.

I dissent from paragraph one of the per curiam opinion, which announces that the term “substantial burden” as used in RFRA and RLUIPA “are interpreted uniformly,” declares that Navajo Nation v. U.S. Forest Service, 535 F.3d 1058 (9th Cir. 2008), is overruled as a result of this interpretation of uniformity between RFRA and RLUIPA, and volunteers, in place of that 15-year precedent, a new test for when a government action imposes a “substantial burden” under RFRA that broadly asks whether the government conduct “prevent[s] access to religious exercise.” We also did not apply this test to arrive at the ultimate decision of this Court, and this test does not address any “issue [that is] germane to the eventual resolution of th[is] case.” United States v. Johnson, 256 F.3d 895, 914–16 (9th Cir. 2001) (separate opinion of Kozinski, J., Trott, T.G. Nelson, Silverman, JJ.) (emphasis added). That is because a majority of this panel has already affirmed, under the completely different rationale in Judge Collins’s majority opinion, the district court’s finding that the transfer of Oak Flat will impose no substantial burden under RFRA. ^[1] As outlined above, “substantial burden” was used in several pre-Smith and pre-RFRA cases and referenced a prior analytical approach. See supra § III.B; Jimmy Swaggart Ministries, 493 U.S. at 384–85; Hernandez, 490 U.S. at 699. Congress adopted “substantial burden” from those “prior Federal court rulings.” 42 U.S.C. § 2000bb-(a)(5). None of those cases define “substantial burden.” But Congress, in adopting RFRA, expressly incorporated the contours and limitations of the “substantial burden” framework into RFRA.

This aligns with how the Supreme Court described its own Free Exercise Clause jurisprudence. For example, the Court in Sherbert held that the government may not compel affirmation of a belief or penalize groups for holding certain views. 374 U.S. at 402. Same with Bowen: Free Exercise violation arises when “compulsion of certain activity with religious significance was involved.” 476 U.S. at 704. These holdings describe categories of claims protected by the First Amendment, but do not define “substantial burden” itself. There is again no definition of “substantial burden.” Thus, the legal context here reveals no technical definition or term of art.

Judge Bea next asserts that there is no evidence that Congress intended to expand or alter the definition of “substantial burden” in pre-RFRA cases. ^[6] See Bea Dissent at 87. But this again assumes, incorrectly, that there ever was a precise definition. True, RFRA’s use of “substantial burden” strongly supports the conclusion that Congress was satisfied with that portion of the test as set forth in prior federal court rulings. But that does not mean that the terms were defined as a term of art. Cf. Bea Dissent at 93.

Indeed, our sister circuits do not speak of “substantial burden” as a term of art. See, e.g., Mack, 839 F.3d at 286; U.S. Navy Seals 1-26, 27 F.4th at 336; New Doe Child #1, 891 F.3d at 578; Korte, 735 F.3d at 654; Hobby Lobby, 723 F.3d at 1114; Midrash, 366 F.3d at 1214; Murphy, 372 F.3d at 979. And for good reason: There is no definition by which they could do so. So while Lyng forecloses Apache Stronghold’s RFRA claim here, see Collins Maj. at 40, that is not because Lyng is part of any “old soil” that was used to define “substantial burden,” Bea Dissent at 79. Indeed, Lyng does not even use “substantial burden” or any analogous framing of the phrase. Lyng therefore cannot be read as establishing a precise definition of “substantial burden” “carried over into the soil” of RFRA. Taggart, 139 S. Ct. at 1801 (emphasis added).

Judge Bea’s approach, which purports to be one grounded in the statute’s text, also violates fundamental principles of textualism. See Bea Dissent at 79–93. His application of the soil theory disregards a textual analysis of half of RFRA’s statutory language. The words of a

6 The Supreme Court seems to reject that premise: “[T]here is no reason to believe . . . that [RFRA] was meant to be limited to situations that fall squarely within the holdings of pre-Smith cases.” Burwell, 573 U.S. at 706 n.18.

governing text are of paramount concern. We must analyze those words in their full context and not focus exclusively on particular provisions. See Textualism, BLACK’S LAW DICTIONARY (11th ed. 2019).

Here, Judge Bea stresses that RFRA directly cites Sherbert and Yoder. See Bea Dissent at 82–85. But this only addresses half of the relevant textual inquiry. Section 2000bb states that a purpose of RFRA is “(1) to restore the compelling interest test as set forth in Sherbert v. Verner, 374 U.S. 398 (1963) and Wisconsin v. Yoder, 406 U.S. 205 (1972).” The rest of § 2000bb, however, reads “and to guarantee its application in all cases where free exercise of religion is substantially burdened; and (2) to provide a claim or defense to persons whose religious exercise is substantially burdened by government.” Id. § 2000bb(1)– (2) (emphasis added).

Congress explicitly codified the test formulated in Sherbert and Yoder. But it did far more than that. It also extended RFRA’s reach to include any other substantial burdens (consistent with the Supreme Court’s application) on religious practice. Congress employs not one but two uses of “and.” Id. And Judge Bea ignores them both. We cannot ignore statutory language like that. If Judge Bea were correct, Congress would not need to have included language guaranteeing RFRA’s application in all cases in which there is a substantial burden. This is true even considering that Congress referenced Sherbert and Yoder to the exclusion of other cases, see Bea Dissent at 84–85, and that Congress declined to use phrases like “for example” to indicate that Sherbert and Yoder were mere examples of substantial burdens, id. at 85. The entire text of the subsection does not start and end with Sherbert and Yoder—it extends further to all substantial burdens. We cannot read Congress’s words out of existence. See TRW Inc. v. Andrews, 534 U.S. ^[19], 31 (2001) (quoting Duncan v. Walker, 533 U.S. 167, 174 (2001)) (“We are ‘reluctant to treat statutory terms as surplusage in any setting’ . . . .”).

Not only should we not read the statutory text out of existence, we also ought not read words into RFRA that are not there. That certain members of Congress made statements about RFRA’s scope as Congress debated its enactment does not provide any reliable evidence of RFRA’s meaning. See VanDyke Concurrence at 160–61. “The greatest defect of legislative history is its illegitimacy. We are governed by laws, not by the intentions of legislators.” Conroy v. Aniskoff, 507 U.S. 511, 519 (1993) (Scalia, J., concurring). The use of such legislative history has been properly criticized as being “neither compatible with our judicial responsibility of assuring reasoned, consistent, and effective application of the statutes of the United States . . . .” Blanchard v. Bergeron, 489 U.S. 87, 99 (1989) (Scalia, J., concurring); see also Does 1-6 v. Reddit, Inc., 51 F.4th 1137, 1146 (9th Cir. 2022) (R. Nelson, J., concurring). And that remains true even though one of the comments came from Senator Hatch who sponsored and championed RFRA. Particularly when legislative history supports our textual interpretation of a statute, we must even more vigilantly guard against encroaching on fundamental statutory principles of construction. 7 Therefore, our assessment of

7 Whether RFRA’s sponsor or a slew of law professors agree with our reading of prior federal law has no bearing here where the statutory text makes clear that RFRA did not overrule Lyng. Had these commentators instead suggested that RFRA overruled Lyng, that would have similarly been irrelevant. Relying on those subjective views undermines the longstanding understanding that, “It is emphatically the province and duty of substantial burden and of any implication of pre-RFRA cases, namely Lyng, must come from analysis grounded in the text. And because “substantial burden” is not a term of art with a specific definition, the soil theory is inapplicable.

B I ultimately agree with Judge Collins’s majority opinion, which relies on a more compelling theory in this case than the soil theory. See Medina Tovar v. Zuchowski, 982 F.3d 631, 644 (9th Cir. 2020) (en banc) (Callahan, J., dissenting) (“In the battle of competing aphorisms I think that ‘context matters’ prevails over the interpretive canon ‘bringing the old soil with it.’”). Judge Collins essentially invokes a different understanding of the Canon of Prior Construction. See Collins Maj. at 46–47 (citing Williams v. Taylor (Terry Williams), 529 U.S. 362 (2000)). This familiar canon is one of context: “If a statute uses words or phrases that have already received authoritative construction by the jurisdiction’s court of last resort, or even uniform construction by inferior courts or a responsible administrative agency, they are to be understood according to that construction.” Scalia & Garner, Reading Law: The Interpretation of Legal Texts 322.

But construction is different than definition. Compare Construction, BLACK’S LAW DICTIONARY (11th ed. 2019) (“The act or process of interpreting or explaining the meaning of a writing”) with Definition, BLACK’S LAW DICTIONARY (11th ed. 2019) (“The meaning of a term as explicitly stated in a drafted document such as a contract, a corporate bylaw, an ordinance, or a statute”). Here, the the judicial department to say what the law is.” Marbury v. Madison, 5 U.S. 137, 177 (1803).

Supreme Court has not defined “substantial burden.” Even so, the Court has construed the term. We apply that context to this case. Lyng is an authoritative construction that the substantial burden test codified in RFRA is inapplicable to certain challenges, including one in which the government manages its own land. True, the Smith majority rejected that the application of the Sherbert test strictly turned on “the government’s conduct of ‘its own internal affairs.’” 494 U.S. at 885 n.2 (citing Lyng, 485 U.S. at 439). But this was to justify Smith’s rule of general applicability, which was expressly overruled in RFRA. RFRA, however, does not address, nor overrule Lyng.

This said, I do not read RFRA as enshrining just Justice O’Connor’s view in her Smith concurrence. Cf. Collins Maj. at 50–51. Justice O’Connor’s articulation of Sherbert’s compelling interest test in her Smith concurrence was not her mere opinion, nor was it “her” test—it was the test established by decades of judicial precedent. Thus, in overruling Smith, Congress codified this preexisting framework in RFRA. And it follows that because RFRA’s stated purpose was to reject Smith, § 2000bb(a), and its effect was to codify the compelling interest test, id. § 2000bb(b)(1), RFRA therefore reinstated the legal framework’s parameters as well. See Parker Drilling Mgmt. Servs., Ltd. v. Newton, 139 S. Ct. 1881, 1890 (2019) (citing McQuiggin v. Perkins, 569 U.S. 383, 398 n.3 (2013)) (“Congress legislates against the backdrop of existing law.”). RFRA thus adopted the term “substantial burden” from the Court’s prior construction of the Sherbert framework. It is therefore not just Smith (or Justice O’Connor’s concurrence), but the entirety of the Court’s pre-RFRA jurisprudence, that provides the contours of substantial burden.

I also have some reservations about Judge Collins’s broad categorization of the Supreme Court’s opinion in Terry Williams. That theory allows us to infer the meaning of a word or phrase when “‘broader debate and the specific statements’ of the Justices in a particular decision concern ‘precisely the issue’ that Congress later addresses in a statute that borrows the Justices’ terminology.” Collins Maj. at 46 (quoting Terry Williams, 529 U.S. at 411–12). There is good reason to be cautious of an overapplication of this theory. The Supreme Court has not relied on it in the 23 years since Terry Williams—and we never have previously. Part of why Terry Williams has not been relied on more may be the Supreme Court’s own limitation: “It is not unusual for Congress to codify earlier precedent in the habeas context.” 529 U.S. at 380 n.11. That same principle has not been established in the First Amendment context to date.

Given these concerns, this theory should be used sparingly. But it is an appropriate application when considering a unique context like habeas in Terry Williams and an equally unique statute like RFRA where Congress explicitly adopted a term from multiple cases to codify that legal framework into law. See Smith, 494 U.S. at 883 (“Under the Sherbert test, governmental actions that substantially burden a religious practice must be justified by a ‘compelling governmental interest.’”). Thus, despite the lack of explicit definition, the body of case law from which “substantial burden” springs forecloses Apache Stronghold’s RFRA claim here. A contrary conclusion would wrongfully ignore the textualist roots of “substantial burden.” The ultimate question is whether RFRA overrules Lyng. As explained above, the stronger case is that Lyng remained part of the “substantial burden” analysis. 8 The Supreme Court has been clear: “‘If a precedent of this Court has direct application in a case,’ . . . a lower court ‘should follow the case which directly controls, leaving to this Court the prerogative of overruling its own decisions.’” Mallory v. Norfolk S. Ry. Co., 143 S. Ct. 2028, 2038 (2023) (citing Rodriguez de Quijas v. Shearson / Am. Express, Inc., 490 U.S. 477, 484 (1989)). “This is true even if the lower court thinks the precedent is in tension with ‘some other line of decisions.’” Id.

A commendable critique of Lyng might be that its holding lacks in originalist or textualist support. As Smith has been deeply criticized for its lack of original or textual grounding, the same may be said about Lyng, which Smith cites repeatedly. Cf. Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1888 (2021) (Alito, J., concurring) (Smith “can’t be squared with the ordinary meaning of the text of the Free Exercise Clause or with the prevalent understanding of the scope of the free-exercise right at the time of the First Amendment’s adoption.”). Justice Alito concludes that “the ordinary meaning of ‘prohibiting the free exercise of religion’ was (and still is) forbidding or hindering unrestrained religious practices or worship. That straightforward understanding is a far cry from the interpretation adopted in Smith.” Id. at 1896. Under that definition, perhaps it is time for the Supreme Court to revisit Lyng. But that is a task for a different Court on a different day.

8 It has been argued that because RFRA applies to all federal government action, 42 U.S.C. § 2000bb-3, it thus overrules Lyng. But RFRA also instructs courts to look to “prior Federal court rulings.” 42 U.S.C. § 2000bb(a)(5). Lyng is such a prior federal court ruling.

At any rate, Lyng remains the law. There, the Supreme Court held that the government action at issue was not a substantial burden because the First Amendment “simply cannot be understood to require the Government to conduct its own internal affairs in ways that comport with the religious beliefs of particular citizens.” 485 U.S. at 448. And because the land transfer here concerns the government’s management and alienation of its own land, which is no doubt part of its internal affairs, Lyng directly applies to any statutory application of “substantial burden” under RFRA as well. With no compelling evidence to support a finding that Lyng was overruled when Congress enacted RFRA, for the same reasons that Apache Stronghold’s claim fails under the First Amendment, it fails under RFRA too.

VII RFRA is a unique statute. While the dissent raises a plausible textual interpretation of “substantial burden,” I ultimately disagree. In adopting RFRA, Congress used a specific term—“substantial burden”—which should reasonably be read to reject Smith but incorporate prior Supreme Court construction of that term. While we lack a precise definition, we are given guideposts. And Lyng is one of those.

The phrase “substantial burden” does not exist in a vacuum. Rather, decades of Supreme Court precedent establish that only certain forms of substantial burdens are cognizable as that term is used to apply the Free Exercise Clause. And when the government seeks to manage its internal affairs and operate on its own land, no such cognizable burden exists under RFRA. Congress then codified this standard and its associated boundaries in RFRA. Because RFRA does not overrule the Supreme Court’s binding precedent in Lyng, Apache Stronghold has no viable RFRA claim here.

VANDYKE, Circuit Judge, concurring:

I agree with the majority that our decision in this case is controlled by Lyng v. Northwest Indian Cemetery Protective Association, 485 U.S. 439 (1988). I write separately to elaborate on why the alleged “burden” in this case is not cognizable under the Religious Freedom Restoration Act (RFRA) and to explain why reinterpreting RFRA to impose affirmative obligations on the government to guarantee its own property for religious use would inevitably result in religious discrimination. Occupying the background of the majority opinion is a reality central to the resolution of this case: there is no textual, historical, or precedential support for the notion that a government’s refusal to use its own property to enable or subsidize religious practice is a cognizable burden under either the Free Exercise Clause or RFRA. Even assuming it’s theoretically possible to reconceptualize Uncle Sam’s parsimony as a “burden” on religious exercise, such stinginess in the allocation of the government’s own property isn’t the sort of burden our religious freedom guarantees were ever meant to address. And because the government action here did not constitute a cognizable burden, any reliance on the substantiality of the impact of the government’s decision on the plaintiffs in this case is misguided.

I.

Enacted in response to one of the most criticized Supreme Court decisions in history, 1 RFRA was a laudable attempt to broadly restore religious liberty. But like any rights-endorsing statute, no matter its scope, RFRA has its limits. A cognizable RFRA claim arises only when (1) the government (2) substantially (3) burdens (4) religious exercise. 42 U.S.C. § 2000bb-1(a). Apache Stronghold claims that the government will burden the Apaches’ religious exercise—specifically, their use of Oak Flat to worship and conduct ceremonies—by transferring ownership of the government’s property to Resolution Copper.

Because it is undisputed that the Apaches’ desire to use Oak Flat to worship and conduct ceremonies qualifies as religious exercise, the only issue before our court is whether the transfer is an instance of the government burdening the Apaches’ religious exercise as that action has long been understood under RFRA and the Free Exercise Clause. After considering the logic underlying RFRA, and then reviewing the proper Free Exercise Clause and RFRA frameworks, it becomes apparent that the government does not burden religious exercise by refusing to ensure the government’s own property remains available to enable it.

1 Emp. Div., Dep’t of Hum. Res. of Or. v. Smith, 494 U.S. 872 (1990). A. A commonsense reading of RFRA does not suggest the government burdens religion by refusing to use its property to enable religious activity.

Notwithstanding the volume of ink spilt today by our en banc court across multiple opinions, it’s safe to say that we all agree on at least one thing: RFRA provides a claim for some—but not all—burdens that a person may experience in relation to his or her religious exercise. For starters, the burden must have been imposed by a particular entity— namely, the government. And related to that, when the government acts (or fails to act), not all of its actions (or inactions) that may have some incidental effect on an individual’s religious exercise are deemed to “burden” that person’s religious exercise within the meaning of our guarantees of religious freedom. 2 This is confirmed by both common sense and the ordinary meaning of the verb “burden,” as a few illustrations will show. Imagine, for example, that a Muslim believes he must complete a religious pilgrimage to Mecca during his lifetime. But he lacks the money to do so. If his sister has enough money to pay for the trip but refuses to give it to him, no one would seriously claim that the sister “burdened” her brother’s religious exercise by refusing to give him her money to enable his exercise. Sure, there is a sense in which the brother faces a burden on his religious exercise: he doesn’t have something he needs to enable it. But few if any 2 Indeed, Apache Stronghold’s able counsel acknowledged at oral argument that not every government action that might be characterized as a “burden” is cognizable under RFRA, including when the government refuses to sell its land to a private party to build a church on the property. would say his sister caused that burden by refusing to give him her money.

If our example were changed slightly so that the brother asked the government instead of his sister for the money, the result would be unchanged. Characterizing the government’s unwillingness to give its resources to our disadvantaged Muslim friend as a government-imposed burden on his religious exercise would be no less strange than in our first example.

That is the key to this case. Much has been said about the substantiality of the burden the Apaches will experience when the government’s Oak Flat property is traded and eventually destroyed. It is certainly true that the effect is substantial. But its substantiality is irrelevant in this case. Even assuming one could counterintuitively characterize the government’s unwillingness to give someone its property as a “burden,” such a burden is not the type of government-imposed burden that is cognizable under RFRA or the Free Exercise Clause. Few people today would characterize the government withholding its own property as the government imposing a burden. And there is no reason to think that such a peculiar conception of a government-imposed burden had any more purchase at the time of the nation’s founding, at the time of the Fourteenth Amendment’s ratification, or at the time of RFRA’s enactment. In short, Apache Stronghold’s RFRA claim fails because the government’s use of its own property simply does not impose on the Apaches’ religious exercise the type of “burden” that either RFRA or the Free Exercise Clause contemplate. B. Under the Free Exercise Clause, the government does not burden religious exercise by managing its own property.

The Free Exercise Clause comes into play when the government “prohibit[s]” the “free exercise” of religion, U.S. Const. amend. I, which courts have long interpreted as doing something that burdens such free exercise. Because this constitutional right “is written in terms of what the government cannot do to the individual, not in terms of what the individual can exact from the government,” the Supreme Court has recognized that government actions involving the government’s use of its own resources do not impose a First Amendment burden on a person’s religious exercise, even when such government actions may indirectly—and possibly even substantially—affect religious exercise. Lyng, 485 U.S. at 450–51 (emphasis added) (quoting Sherbert v. Verner, 374 U.S. 398, 412 (1963) (Douglas, J., concurring)). Since well before Smith, it has been commonly understood that the government does not impose a burden when it merely refuses to subsidize a religious exercise. See, e.g., Regan v. Tax’n with Representation of Wash., 461 U.S. 540, 549 (1983) (“We have held in several contexts that a legislature’s decision not to subsidize the exercise of a fundamental right does not infringe the right, and thus is not subject to strict scrutiny.”); Sherbert, 374 U.S. at 412 (Douglas, J., concurring) (“The fact that government cannot exact from me a surrender of one iota of my religious scruples does not, of course, mean that I can demand of government a sum of money, the better to exercise them.”). The understanding that a refusal to subsidize does not burden religious exercise is obviously not limited to just the government’s money. A Catholic priest can no more demand that the government provide him with communion wine than he can demand that the government provide him with money to buy that wine. An elder of the Church of Jesus Christ of Latter-day Saints can’t insist that the government give him either a bicycle or the cash to buy one. Nor can a pastor require that the government provide him a church on government land so that he can better serve his flock. As in our initial Mecca example, the government has not “burdened” anyone’s religious exercise in any of these examples by withholding its own resources.

Of course, every level of government in our nation distributes a variety of government benefits to a variety of recipients. And when the government does that, it cannot do so in a way that discriminates against or between religions. In Sherbert, for example, a state government provided unemployment benefits to workers who required Sunday off to practice their faith, but not to those whose religion required them to take Saturday off. 374 U.S. at 399–400, 406. The Supreme Court correctly concluded that the Free Exercise Clause disallows such discrimination between or against religions in the provision of government benefits. Id. at 404. The Court explained that such differential treatment of religious adherents in the allocation of government benefits imposes the type of “burden” on religious liberty that the Free Exercise Clause was meant to protect against. Id. Indeed, it “puts the same kind of burden upon the free exercise of religion as would a fine imposed against appellant for her Saturday worship.” Id. This is because “to condition the availability of benefits upon [a religious observer’s] willingness to violate a cardinal principle of her religious faith effectively penalizes the free exercise of her constitutional liberties.” Id. at 406. Thus, Sherbert and its progeny make clear that once the government chooses to provide government benefits, it cannot do so in a discriminatory fashion that effectively coerces potential recipients into abandoning their constitutional right to freely exercise their religion.

But of course, nowhere did Sherbert (or any case since) conclude that the government had to provide unemployment benefits to anyone in the first instance; it simply concluded that if the government chose to do so, it couldn’t religiously discriminate. See, e.g., Trinity Lutheran Church of Columbia, Inc. v. Comer, 582 U.S. 449, 467 (2017) (“[T]he exclusion of Trinity Lutheran from a public benefit for which it is otherwise qualified, solely because it is a church, is odious to our Constitution … and cannot stand.”). I’m not aware of any case applying Sherbert’s anti-discrimination principle that holds the government must either start providing or continue providing some government benefit— again, those cases simply stand for the reasonable proposition that if the government is doling out benefits, it must not discriminate against religion in the process of doing so.

Unsurprisingly, the Supreme Court has also made clear that the Free Exercise Clause protects against the government burdening religious exercise by directly imposing requirements on people that are at odds with their religious beliefs. The Supreme Court addressed this situation in Wisconsin v. Yoder, 406 U.S. 205 (1972). Wisconsin had attempted to make school attendance mandatory until the age of 16. Id. at 207. This compulsory- attendance law was “undeniably at odds with fundamental tenets of [Amish] religious beliefs” and presented the Amish with a classic dilemma: exercising their religious beliefs would lead to criminal sanctions, but compliance with the law would violate their beliefs. Id. at 218. Yoder and many cases since then stand for the straightforward proposition that, when the government says, “you must do X,” and your religion says, “you must not do X,” then the government’s demand has burdened your religious exercise.

Both the Yoder type of burden and Sherbert type of burden, while different, converge under a single concept: government coercion. Yoder involved the most direct form of coercion: violate your religious scruples or be punished. Sherbert’s coercion is less direct but not necessarily less coercive: violate your religious scruples or be denied an otherwise available government benefit. Both the Yoder and Sherbert types of government coercion are conceptually quite different from a theoretical third type: the government simply refusing to give someone its property so that he can use it to exercise his religion. ^[3] This third type of government action is different in kind from the first two. In no way is the government coercively inducing or requiring people to 3 It is important to distinguish between a Sherbert-type burden and this third potential type of claim. Both involve the government withholding its property, but in Sherbert the government is already giving its property to some religious adherents, while discriminatorily withholding its property from others of a different religion. Thus, in a Sherbert case, the baseline condition is, so to speak, that the government is already providing its property to some (but not all) religious adherents. In contrast, the baseline condition in a case like this one is that the government is not giving its property to anyone, and the religious claimants nonetheless insist that the government must uniquely provide them with government property to enable their religious exercise. Apache Stronghold has not tried to make a Sherbert-type religious discrimination claim in this case, presumably because the government isn’t discriminatorily “giving” its land to anyone but is instead trading the government-owned Oak Flat for other land owned by the mining company. In other words, the government is effectively selling Oak Flat to the mining company, and Apache Stronghold hasn’t claimed any discriminatory action on the part of the government in, say, rejecting an equivalent competing offer from Apache Stronghold. violate their religious beliefs. Instead, any coercion works in the opposite direction: people are demanding that the courts make the government enable or subsidize their religious beliefs by uniquely providing them with government property.

While an able lawyer can certainly characterize this third type of claim as a “burden,” it has been well understood since before Smith that the Free Exercise Clause does not cover any such government decisions, regardless of the label. This is most unmistakably demonstrated by Lyng. There, the federal government had permitted the building of a road and the harvesting of timber on publicly owned land. Lyng, 485 U.S. at 441–42. Some Native American tribes argued that this would burden their religious practice on the government’s land. Id. at 447. But as the Court explained, the project did not burden religious exercise within the meaning of the Free Exercise Clause. Id. at 452. Notwithstanding that the claimed effects from the roadbuilding project could be “severe” and “virtually destroy the … Indians’ ability to practice their religion,” those effects did not give rise to a cognizable burden. Id. at 447, 450–51.

The reason the Indian tribes lacked a Free Exercise Clause claim in Lyng was because, despite the “devastating” incidental effect that the government’s management of its own land would have on their religious exercise, id. at 451, the tribes would not “be coerced by the Government’s action into violating their religious beliefs; nor would [the] governmental action penalize religious activity by denying [them] … benefits,” id. at 449. As Lyng made clear, the “Free Exercise Clause affords an individual protection from certain forms of governmental compulsion; it does not afford an individual a right to dictate the conduct of the Government’s internal” affairs, particularly the government’s management of its own property. Id. at 448 (emphasis added) (quoting Bowen v. Roy, 476 U.S. 693, 699–700 (1986)).

Nothing since Lyng has cast into question the straightforward understanding that the Free Exercise Clause does not require the government to let you use its property— including its real property—to exercise your religion. Our court, sitting en banc fifteen years ago, reviewed these same cases and reached the same conclusion. See Navajo Nation v. U.S. Forest Serv., 535 F.3d 1058, 1068–73 (9th Cir. 2008) (en banc). 4 Regardless of how you label it, the government’s nondiscriminatory use of its own property has never been understood to impose a constitutionally cognizable burden on someone’s religious freedom—even when such governmental decisions incidentally have “devastating” and “severe adverse effects on the practice of [a] religion.” Lyng, 485 U.S. at 447, 451.

C. RFRA adopted the ordinary meaning of “burden” as that term had been uniformly understood in Free Exercise Clause cases. Echoing decades of Free Exercise precedent, RFRA prohibits the government from burdening a person’s religious exercise. 42 U.S.C. § 2000bb-1(a). As is typical in many statutes, RFRA defined some but not all terms that determine whether a person has a cognizable RFRA claim. For example, RFRA tells us that a person’s “religious exercise” includes “any exercise of religion, whether or not compelled by, or central to, a system of religious belief.” Id. 4 Our court reached the right result in Navajo Nation, although I might quibble with some of its rationale. at §§ 2000bb-2(4), 2000cc-5(7)(A). Since this is a clear departure from how religious exercise had been understood under the First Amendment, 5 it made sense for Congress to provide that definition. But tellingly, RFRA does not define what it means for the government to “burden” religious exercise. The obvious reason for that, given the context of RFRA’s enactment and its clear textual departures from the First Amendment in other regards, is that RFRA meant “burden” in the way it had been commonly understood in the Free Exercise Clause context. Indeed, the Supreme Court has acknowledged as much. See Tanzin v. Tanvir, 592 U.S. 43, 46–48 (2020) (citing Antonin Scalia & Bryan Garner, Reading Law: The Interpretation of Legal Texts 323 (2012)).

In pre-RFRA First Amendment caselaw, it was well understood that the government burdens religious exercise when it acts in a coercive manner, and that the government’s decisions about how it uses its own property are not coercive unless they discriminate (as in Sherbert). During and immediately after RFRA’s enactment, everyone understood that RFRA carried forward this ordinary understanding of what it means to burden religious exercise. Post-RFRA caselaw only further confirmed that RFRA adopted the ordinary meaning of how the government may impose a 5 Prior to being amended by the Religious Land Use and Institutionalized Persons Act of 2000, 42 U.S.C. § 2000cc, et seq. (RLUIPA), RFRA defined “exercise of religion” as “the exercise of religion under the First Amendment to the Constitution.” Under this standard, courts had required the burdened religious exercise to be “central to” or “compelled by” the religion. See, e.g., Graham v. C.I.R., 822 F.2d 844, 850–51 (9th Cir. 1987), aff’d sub nom. Hernandez v. Comm’r, 490 U.S. 680, 699 (1989); O’Lone v. Est. of Shabazz, 482 U.S. 342, 345 (1987); see also Bryant v. Gomez, 46 F.3d 948, 949 (9th Cir. 1995); Weir v. Nix, 114 F.3d 817, 820 (8th Cir. 1997).

burden—and specifically, as relevant to this case, that the government’s use of its own property burdens religious exercise only when it is allocated in a discriminatory manner. Here, there is no claim that the government has used its resources in a discriminatory manner, and the government therefore has not burdened the Apaches’ religious exercise within the meaning of RFRA.

i. The ordinary understanding of RFRA does not support the claim that the government burdens religious exercise by using its own resources in a nondiscriminatory manner. If RFRA’s plain text doesn’t make it obvious enough that RFRA did not depart from the ordinary meaning of “burden” under the Free Exercise Clause, the discussion surrounding the passage of RFRA further confirms that the government does not burden religious exercise by using its own resources in a nondiscriminatory manner.

When Congress enacted RFRA, it was well understood that a burden is imposed by the government’s use of its own resources only when the use of such resources discriminates against or between religions. Readily accessible examples of this widespread understanding are provided by congressional statements explicitly maintaining that RFRA “does not apply to government actions involving only management of internal Government affairs or the use of the Government’s own property or resources.” S. Rep. 103– 111, at 9 (1993); see also 139 Cong. Rec. 26193 (1993) (remarks of Sen. Hatch) (explaining that Lyng and Bowen are unaffected by RFRA). 6 Leading religious liberty 6 Judge R. Nelson mildly chastises me for engaging in supposed faint- hearted textualism by citing the congressional record. I agree with both scholars shared a similar understanding of RFRA’s effect, observing immediately after its enactment that, under RFRA, a “cognizable burden” does not exist when the government uses its resources in a nondiscriminatory manner that has only an indirect effect on religion. See Douglas Laycock & Oliver S. Thomas, Interpreting the Religious Freedom Restoration Act, 73 Tex. L. Rev. 209, 228–30 (1994) (footnotes omitted). 7 No burden exists him and Justice Scalia, whom he quotes, that “[e]ven if the members of each house wish to do so, they cannot assign responsibility or making law—or the details of law—to one of their number, or to one of their committees.” Antonin Scalia & Bryan Garner, Reading Law: The Interpretation of Legal Texts 386 (2012). But as should be sufficiently clear from context, I am not citing to the views of specific legislators for the purpose of conclusively determining what RFRA means. Nor am I (as charged) preferencing legislative history just because it happens to support my understanding of RFRA. Instead, I cite such statements as further evidence of my point—with which I believe Judge Nelson agrees—that at the time of RFRA’s enactment, nobody would have understood the government’s decision about what to do with its own land to be a cognizable burden under RFRA. Individual legislators are no more able to authoritatively speculate about how a law will apply in a certain case than anyone else. That goes for legal academics, too—who I also cite. “The interpretation of the laws is,” after all, “the proper and peculiar province of the courts,” not Congress or the academy or anyone else. Alexander Hamilton, Federalist No. 78. My point is only to demonstrate the unanimity of understanding about what did and did not constitute a burden on religious exercise at the time of RFRA’s passage, which matters here because RFRA’s text indicates that it should be understood by reference to the state of Free Exercise jurisprudence before Smith.

7 See also Luralene D. Tapahe, After the Religious Freedom Restoration Act: Still No Equal Protection for First American Worshippers, 24 N.M. L. Rev. 331, 345 (1994) (noting that pre-RFRA courts declined to extend First Amendment protection to “challenges to government control of non-Indian land” and later explaining that, “[s]ince RFRA mandates that strict scrutiny be used only if a burden is first found, Indian free exercise because citizens simply “may not demand that the Government join in their chosen religious practices” by providing the resources for such practices. Id. (quoting Lyng, 485 U.S. at 448). Everyone understood that, under RFRA, the government retains its right to use its resources according to its own preferences. 8 It does not have the claims will likely be resolved in the very same manner as before”); Ira C. Lupu, Of Time and the RFRA: A Lawyer’s Guide to the Religious Freedom Restoration Act, 56 Mont. L. Rev. 171, 202 (1995) (explaining that the “developing case law” on “substantial burden” under RFRA suggests that “religious exercise is burdened only by the combination of legal coercion and religious duty”); Daniel O. Conkle, The Religious Freedom Restoration Act: The Constitutional Significance of an Unconstitutional Statute, 56 Mont. L. Rev. 39, 73 & n.172 (1995) (noting that although “RFRA repudiates Smith, … it appears to leave the internal operations cases,” such as Lyng and Bowen, “unaffected”). 8 I of course agree with Judge Nelson that “[i]t is emphatically the province and duty of the judicial department to say what the law is.” Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177 (1803). But I respectfully disagree with his insistence that the uncontradicted view of a “slew of law professors” and legislators “has no bearing” on the proper interpretation of RFRA. I presume that Judge Nelson and I agree that it is the original public meaning of the text that controls our analysis, not some hidden or idiosyncratic meaning devised by judges. See Lynch v. Alworth-Stephens Co., 267 U.S. 364, 370 (1925) (“[T]he plain, obvious, and rational meaning of a statute is always to be preferred to any curious, narrow, hidden sense that nothing but the exigency of a hard case and the ingenuity and study of an acute and powerful intellect would discover.”). Part of the endeavor of surmising the original public meaning is understanding what the public would have originally understood the legislative enactment to mean, including the part of the public that was elected to Congress. If, for example, every law professor, every Congressman, and every other literate person in the United States were on record opining that a particular statute meant “X,” I would hope good originalists could count that as some useful evidence that its original public meaning was indeed “X,” not “Y.” See, e.g., Bostock v. Clayton County, 140 S. Ct. 1731, 1757 (Alito, J., concurring) (“As I will obligation to enable religious practice by donating its own property. ii. Cases interpreting RLUIPA are not inconsistent with this well-established understanding of RFRA.

Understandably seeking to distance themselves from the settled understanding that the government does not burden religious exercise through the mere use of its resources in a nondiscriminatory manner, Apache Stronghold and the dissent focus heavily on caselaw interpreting a different statute, RLUIPA, to argue that the government will burden the Apaches’ religious exercise because the Apaches won’t be able to access Oak Flat once it is physically destroyed. In doing so, they improperly divorce the RLUIPA cases from the comprehensive and individualized coercive context inherent in every single RLUIPA case, implicitly endorsing that the Apaches are effectively prisoners in this country and therefore indistinguishable from the actual prisoners who bring claims under RLUIPA. Applying that obviously controversial assumption—and making no attempt to show that this assumption was widely shared when RFRA was enacted in 1993—the dissent relies heavily on what has been show, there is not a shred of evidence that any Member of Congress interpreted the statutory text that way when Title VII was enacted. … And for good measure, the Court’s conclusion that Title VII unambiguously reaches discrimination on the basis of sexual orientation and gender identity necessarily means that the EEOC failed to see the obvious for the first 48 years after Title VII became law.”). That is all I mean by referencing legislative statements above—it is part of my proof that everyone who knew anything about RFRA when it was enacted understood it as not requiring holy handouts of the government’s own property. deemed a substantial burden on religious exercise in the prison context.

I agree with the dissent that the substantiality of a burden can be measured the same way under both RLUIPA and RFRA. But whether a burden is cognizable in the first instance has always been a context-dependent inquiry. And what constitutes a cognizable burden in the prison context— surely the most comprehensively coercive setting in America today—obviously may be very different from what constitutes a “burden” under RFRA. That is why, for example, a Jewish prisoner has a right under RLUIPA to require the government to provide him with kosher meals, whereas a Jewish man outside of prison has no right to insist that the government deliver him free kosher food. ^[9] The dissent’s need to resort to RLUIPA prison cases to justify its preferred outcome in this case is very telling. In 9 The other category of cases addressed by RLUIPA—land-use regulations, or “zoning”—is equally comprehensively coercive. Every zoning case involves the government telling someone what he can or can’t do with his own land. So when the government tells someone he can’t build a church on his own land, for example, that is just as coercive as forbidding someone from buying communion wine with his own money. As such, RLUIPA land-use cases, like cases in the prison context, usually don’t involve hard questions about whether the government’s regulation actually causes a burden on religious exercise. The coercive burden is obvious, inevitably making the litigated question whether the burden is substantial. See, e.g., Guru Nanak Sikh Soc’y of Yuba City v. County of Sutter, 456 F.3d 978, 988–92 (9th Cir. 2006) (discussing whether the regulation was “oppressive to a significantly great extent” (cleaned up)); Int’l Church of Foursquare Gospel v. City of San Leandro, 673 F.3d 1059, 1066 (9th Cir. 2011) (citing Guru Nanak, 456 F.3d at 987) (“[O]ur practice is to examine the particular burden imposed by the implementation of the relevant zoning code on the claimant’s religious exercise and determine, on the facts of each case, whether that burden is ‘substantial.’”). prisons, the “government exerts a degree of control unparalleled in civilian society.” Cutter v. Wilkinson, 544 U.S. 709, 720 (2005) (emphasis added). It controls every aspect of an inmate’s life and renders him fully dependent on the government by stripping him of his ability to provide for his own needs. Brown v. Plata, 563 U.S. 493, 510 (2011). It is certainly true that in RLUIPA cases, courts have concluded that the government must provide resources to prisoners for their religious exercise. But that’s for the same reason they require the government to provide prisoners with basic sustenance like food and clothing, id., or medical care, Estelle v. Gamble, 429 U.S. 97, 103 (1976), or protection from other inmates, Farmer v. Brennan, 511 U.S. 825, 833 (1994)—because the government has coercively “stripped them of virtually every means of” providing for themselves, id. In a very real sense, the prisoner depends on the grace of the government for all his needs and in all his activities. This degree of direct and immediate coercion is, again, “unparalleled in civilian society.” Cutter, 544 U.S. at 720 (emphasis added).

As a result, in the vast majority of RLUIPA cases there is no need to explicitly analyze whether the government’s action burdens religious exercise—it’s a given. The only question is substantiality. And that may also be true for some RFRA cases. But it is not true for all of them, and certainly not this one. This case presents the opposite situation encountered in most RLUIPA cases. The substantiality of the effect on the Apaches’ religious exercise is obvious; it is the legal cognizability of any burden that is at issue. Thus, the dissent’s extensive reliance on inapt RLUIPA cases analyzing the substantiality of an undisputed burden is badly misplaced.

Ultimately, the dissent cannot rely on RLUIPA prison cases without also showing that the Apaches are identically situated vis-à-vis the government as the prisoners in those cases. The dissent makes no attempt to do so, and more importantly makes no attempt to show that this was the common understanding when RFRA was enacted. Absent such a showing, the only justification for the dissent’s extensive reliance on inapt RLUIPA jurisprudence to defend its result in this case is an implicit recognition that it can’t find justification in RFRA and the Free Exercise Clause. As discussed, all the RFRA and Free Exercise Clause cases support the common understanding that, unless you’re the government’s prisoner (literally, not metaphorically), the government’s nondiscriminatory use of its own property is not the type of action that gives rise to a cognizable burden on religious exercise.

D. The government’s swap of Oak Flat for other property does not burden the Apaches’ religious exercise under RFRA.

This case is not meaningfully different from Lyng or Navajo Nation. In all three cases, the government wanted to do something with its own land. In all three cases, what the government planned to do would substantially affect how the tribes wanted to use the government’s land for their own religious exercise. In Lyng and Navajo Nation, courts rejected the First Amendment and RFRA claims because, notwithstanding the “devastating effects” on religious exercise resulting from the government’s planned use of its land, the Free Exercise Clause and RFRA simply do not recognize such burdens resulting from the government’s nondiscriminatory use of its own property. This case is no different, but the dissent would have this court reach the opposite result. In doing so, it would for the first time characterize something as a “burden” under RFRA that has never before been considered a cognizable burden. To do so would be an obvious rewriting of statutory law—a job for Congress, not the courts.

II.

Reconceiving the government’s nondiscriminatory use of its own property as a cognizable burden under RFRA would not only require a judicial rewrite of the statute; it would turn the statute on its head, requiring instead of reducing religious discrimination. Because the government’s resources are not infinite, the expansion of RFRA advocated by Apache Stronghold and the dissent would inevitably require the government to discriminate between competing religious claimants. While no doubt some such claims—including those made by Apache Stronghold in this case—would be sympathetic, there is no way to resolve this case in the Apaches’ favor without endorsing a rule that would one day soon force the government to pick religious winners and losers. So even if this court did require the government to effectively hand over Oak Flat as a religious offering to the Apaches, only some religions would benefit from the precedent created by such a decision. ^[10] 10 In Part I of this opinion, I have endeavored to explain why I think the dissent’s proposed interpretation of RFRA is wrong as a legal matter. And now, in Part II, I explain why that view is also wrongheaded. Judge Nelson misunderstands this approach, confusing the reasons I agree with the majority’s interpretation of RFRA (Part I) with the warnings I make about religious discrimination that would inevitably result if the dissent’s rewrite of RFRA was adopted (Part II). But to be clear, I agree with Judge Nelson that “[t]he dissenters are not wrong … because under their view ‘only some religions would benefit from the precedent created by such a decision.’” The reason the dissenters are wrong is because they Eventually, lines limiting the court-enforced distribution of the government’s largesse would need to be drawn. And because, as explained above, the dissent’s novel approach has no basis in the text or original understanding of RFRA, any judicially created distinctions limiting the extent of the resulting religious entitlement would similarly lack any statutory justification. Worse, such distinctions would necessarily discriminate between religions, offering government property to some and not others and turning RFRA into a tragic parody of itself. One need look no further than the dissent itself to see early indications of the kind of discriminatory distinctions that might flow from this atextual understanding of RFRA.

A. The dissent would establish a discriminatory preference in favor or older religions and against newer ones.

Not far into the dissent, the reader encounters the first such distinction: religious practices with a lengthy historical pedigree apparently deserve more protection than newly established ones. Parroting Apache Stronghold’s repeated emphasis that the Apaches have worshipped at Oak Flat “since time immemorial,” the dissent heavily implies the Apaches should be treated preferentially because their religious exercise is a long-established practice. 11 advance a view of RFRA that has no basis in its original public meaning. My point here is that in addition to being the legally wrong interpretation, the dissenters’ judicial revision of RFRA would also undermine the equal protection of religion that RFRA was enacted to protect. 11 The dissent is not alone in emphasizing the ancient nature of the Apaches’ religious practice. Both the panel and motion-stage dissents The trouble with emphasizing the lengthy history of the Apaches’ religious practice at Oak Flat is that it is entirely irrelevant to our analysis under RFRA and the Free Exercise Clause. Our religious liberty protections “apply to all citizens alike,” Lyng, 485 U.S. at 452, and with equal force to a religion founded yesterday as to one with roots deep in prehistory. How long a person has practiced a religion, or how old that religion is, should be “immaterial to our determination that … free exercise rights have been burdened; the salient inquiry under” both RFRA and the Free Exercise Clause “is the burden involved.” Hobbie v. Unemployment Appeals Comm’n of Fla., 480 U.S. 136, 144 (1987). It is bad enough that Apache Stronghold’s counsel made this discriminatory argument. Our court has thankfully refused to make things worse by imbuing it with the force of law. ^[12] did so also. See, e.g., Apache Stronghold v. United States, 38 F.4th 742, 774 (9th Cir. 2022) (Berzon, J., dissenting). 12 It’s not hard to see how invidious this argument is when you consider a sincere religious observer whose newer religion requires the ceremonial use of Oak Flat, just like the Apaches. The government’s action of trading Oak Flat for other land would have exactly the same effect on both the observer of a newer religion and an Apache: neither would be able to use Oak Flat for religious ceremonies. But accepting the dissent’s implicit premise that the “time-immemorial” nature of the Apaches’ religious practice at Oak Flat is legally significant could lead to a different result in each of the two cases: the transfer of Oak Flat would burden the Apaches’ religious exercise, but the same transfer might not burden a similarly situated practitioner of the newer religion simply because the person (or, more precisely, the person’s predecessors) had not used the land before or for long enough. And what about a religion of intermediate age—say, a hundred years or so? How long is “long enough” to warrant protection under RFRA? By introducing the age of a religion and the length of religious practice as variables relevant to the analysis, the dissent offers an arbitrary and Of course, the suggestion that long-established religious practices should receive favorable treatment under RFRA is made only lightly. The dissent stops short of a full-throated defense of such a rule. Instead, it contents itself to repeatedly emphasize the longstanding nature of the Apaches’ religious practice and leaves the legal significance of that fact to implication. Making the argument explicitly would lay its blatantly discriminatory character bare, but subtle though it may be, the dissent unmistakably lays the groundwork for a discriminatory limiting principle that (need it be said?) could never be supported under either the Free Exercise Clause or RFRA.

B. The dissent’s interpretation of RFRA also discriminates by providing more protection against burdens accompanied by significant physical or environmental impacts.

Both the dissent and Apache Stronghold also take care to emphasize the extent of the physical destruction associated with the transfer of Oak Flat. The import of such argument is clear: as with age, the dissent and the Apaches would also establish a discriminatory preference in favor of protecting burdens on religious exercise with a significant physical or environmental component when compared to burdens associated with less physical manifestations. But doing so would be double error, both because such a rule wrongly implies that a practitioner’s religious harm under RFRA claim is somehow predicated on the physical attributes of the intrusion, and because it invites courts to measure the comparative significance of religious harms in physical terms, a behavior strictly prohibited in our jurisprudence. discriminatory distinction between observers of newer religions and long-established ones—a distinction that has no basis in RFRA. Ultimately, this distinction too is contrary to both the text of RFRA and the background precedent that informed its understanding, and if adopted, it would likewise perpetuate religious discrimination. i. Attempting to distinguish Lyng and Navajo

Nation by focusing on the extent of the physical impact reads a discriminatory preference for land-based religious practices into RFRA.

The biggest hurdle faced by the dissent and the Apaches is that this case is strikingly similar to both the Supreme Court’s decision in Lyng and our court’s en banc decision in Navajo Nation. To get around these cases, which doom its claims, Apache Stronghold attempts to distinguish them by emphasizing the physical differences between the government’s actions in those cases and this one. Navajo Nation and Lyng are different, they contend, because “neither … involved physical destruction of a sacred site.” The dissent employs similar logic, distinguishing Lyng on the basis that the transfer will result in the “utter destruction” of Oak Flat, which “will prevent the Western Apaches from visiting Oak Flat for eternity.” Not only does this argument fail to provide a suitable basis to distinguish Lyng and Navajo Nation, but it also introduces another arbitrary and discriminatory limitation on the scope of RFRA’s protection.

In Navajo Nation, the government allowed a mountain sacred to multiple Indian tribes to be showered daily with 1.5 million gallons of poopy water that, according to those tribes, would desecrate the mountain, render it impure, and destroy their ability to perform certain religious ceremonies. 535 F.3d at 1062–63; id. at 1081 (Fletcher, J., dissenting). So both Navajo Nation and this case present precisely the same impact on religious exercise from government land-use decisions: elimination of the ability to perform religious ceremonies. The dissent here, however, distinguishes Navajo Nation by asserting that “nothing ‘with religious significance … would be physically affected’” by the government’s decision to spray recycled wastewater containing human waste onto a sacred mountain (emphasis added). But that downplays the spiritual significance of the government’s action in Navajo Nation and ignores the court’s later reasoning in the same opinion that “[e]ven were we to assume … that the government action in this case w[ould] ‘virtually destroy the … Indians’ ability to practice their religion,’” the result would not have changed. Navajo Nation, 535 F.3d at 1072 (quoting Lyng, 485 U.S. at 451).

The dissent similarly distinguishes and downplays the government’s land-use decisions in Lyng—notwithstanding their “severe” and “devastating effects on traditional Indian religious practices”—by highlighting the limited physical effects of the government’s actions in Lyng. In the face of Lyng and Navajo Nation, it nevertheless continues to rely on the extent of the physical impact that will result from the government’s decision to transfer Oak Flat.

There is little doubt that the government’s decision to transfer Oak Flat will have consequences for the physical environment in and around that area, but as much as some may wish otherwise, this is not an environmental case. This is a case about religious injury, and the measure of that injury is the harm to religious exercise. That harm is precisely the same here as it was in Lyng and Navajo Nation: the complete inability of Native Americans to conduct certain religious ceremonies because of government decisions about how it uses government land.

The desire to distinguish Lyng and Navajo Nation by emphasizing the physical impact of the challenged government decision is certainly understandable from an environmentalist’s perspective, but doing so would result in an unfortunate perversion of RFRA. The view advocated by Apache Stronghold and endorsed by the dissent threatens to turn RFRA into a statute that arbitrarily gives greater protection to burdens on religious exercise that are more physical in nature, while downplaying equally significant burdens on other forms of religious exercise simply because they don’t similarly affect the physical environment. Such an approach privileges forms of religious exercise that preserve the physical environment at the expense of other religious exercise that might arguably lack similar positive environmental externalities. Again, it is understandable why this might be an attractive rewrite of RFRA for some modern judges—one could say that environmentalism is the favored religion du jour 13—it just has no basis whatsoever in RFRA’s text or original meaning.

13 See Joel Garreau, Environmentalism as Religion, The New Atlantis, Summer 2010, at 61 (“For some individuals and societies, the role of religion seems increasingly to be filled by environmentalism.”); Freeman Dyson, The Question of Global Warming, The New York Review of Books (June 12, 2008), https://www.nybooks.com/articles/2008/06/12/the-question-of-global- warming/ (“There is a worldwide secular religion which we may call environmentalism …. Environmentalism has replaced socialism as the leading secular religion.”); Robert H. Nelson, Environmental Religion: A Theological Critique, 55 Case W. Res. L. Rev. 51, 51 (2004) (“Environmentalism is a type of modern religion.… Indeed, many leading environmentalists have characterized their own efforts in religious terms.”); Andrew Sullivan, Green Faith, The Atlantic (March 28, 2007), https://www.theatlantic.com/daily- dish/archive/2007/03/green-faith/229789/; Andrew P. Morriss &

ii. A rule that distinguishes religious harms by their physical measurability finds no support in either the text of RFRA or the body of caselaw supporting it.

The physical impact of the government’s actions has no basis in the text of RFRA, and it is just as foreign to the pre-Smith understanding of the Free Exercise Clause that informed RFRA. But it is not simply the case that the dissent’s approach finds no support in RFRA’s text or caselaw; it has already been affirmatively rejected. Focusing on the physical destruction of Oak Flat resurrects an argument that the Supreme Court rejected outright in Lyng.

In Lyng, the government sought to build a road that would result in the physical destruction of wilderness conditions necessary for the plaintiffs’ religious exercise, including “privacy, silence, and an undisturbed natural setting.” 485 U.S. at 442. The Court recognized that “too much disturbance of the area’s natural state would clearly render any meaningful continuation of traditional practices impossible,” meaning the “projects at issue … could have devastating effects on traditional Indian religious practices.” Id. at 451. The Court nevertheless explained that the incidental religious effect of such government action on native tribal religious activity—“devastating” though it might be—could not “meaningfully be distinguished from the use of a Social Security number” in Bowen v. Roy, in which a religious practitioner sincerely believed that merely issuing a Social Security number (which had the slightest of physical components) to a child would rob the child of her spirit. Id. at 449, 456. “In both cases, the challenged Benjamin D. Cramer, Disestablishing Environmentalism, 39 Env’t L. 309, 323–42 (2009).

Government action would interfere significantly with private persons’ ability to pursue spiritual fulfillment according to their own religious beliefs.” Id. at 449. Thus, notwithstanding the significantly different physical effects of the government action in each case, the religious harms suffered were indistinguishable for purposes of determining whether a burden existed. Id. at 449–50. The presence or absence of the burden on religious exercise turns not on the degree of any physical impact from the government’s activity, as urged by Apache Stronghold and the dissent, but on the asserted harm to religious exercise, as explained in Lyng and Bowen.

iii. Analyzing burdens on religious exercise with reference to their associated physical impacts is inherently discriminatory.

Text and caselaw aside, it is also inequitable to let the physical consequences of a government action determine whether religious exercise has been burdened because religions differ in what might burden their exercise. Some religions place more emphasis on the material world, while others are more spiritually directed. Some center their devotion on historic rites held in set-apart, holy places, while others are not as ceremonially or geographically constrained. And of course, many faiths incorporate degrees of some or all of these defining characteristics into their religious practice. The dissent’s misguided emphasis on the environmental consequences of the government’s action preferences some of these religious aspects over others, and if it were afforded legal significance, it would ensure that RFRA would be applied discriminatorily going forward. Religions that experience a substantial burden to their exercise due to government action that also has a substantial physical manifestation would be treated favorably. Inversely, religions affected by government actions with less physical impact would be sent to the back of the bus. But our religious liberty protections were designed to extend to all religions, not just to those that may suffer a tangibly “objective” and “measurable” burden (whatever that might mean) evaluated in physical terms. A test that relies on the physical effects of government action could significantly reduce protection for religions that do not rely on tangible relics, material artifacts, or other paraphernalia. Such a test would threaten to overtly discriminate against and overwhelmingly under-protect religions less tied to the material world.

C. The dissent encourages discrimination by creating a baseless distinction between the government’s real property and its other property.

The dissent relatedly appears to infer that there’s something legally special about the religious use of government-owned real property that makes it materially distinguishable from other forms of government resources. But again, this distinction bears no connection to anything in RFRA itself, and it too would invite future discrimination between religious groups.

As a legal matter, limiting the dissent’s preferred rule that the government must give out its resources for religious exercise to religions that use particular real property in the government’s control is clearly disconnected from RFRA’s text. The practice of essentially every religion is resource constrained, and nothing in the statutory text supports distinguishing between the types of resources that religious observers need to conduct their religious exercises. Some need land, some need vehicles, some need cash (or Venmo). Regardless of what they need in a particular instance to exercise their religion, one commonality among religious observers is that they are often limited in what religious activities they can engage in based on the resources they have available to them. And if the government owns the resources they need, they face the exact same problem— regardless of whether it’s land or legal tender, the government’s refusal to contribute its stuff is hindering their religious exercise.

Grafting onto RFRA a special rule favoring religions that happen to require land would clearly discriminate against other religions. What makes real property special, particularly under RFRA? Is needing specific real property to conduct a ceremony different under RFRA from needing a bike to proselytize? Or needing a sweat lodge made from certain trees under government control? There is no logical or textual basis in RFRA for the dissent’s suggestion that land is somehow special. While certain tracts of government-owned land are religiously special for many Native Americans, other government property may be (or become) religiously special for other religions. Under the dissent’s approach, the latter would be treated worse than the former without any textual basis for the difference in treatment.

The dissent tries to limit the discriminatory impact of the rule it offers by limiting it to circumstances where the government has unique control over access to religious resources. But that’s no limitation at all. The government has unique control over all its resources. Every dollar bill in circulation was at one point owned and “uniquely controlled” by the government—after all, the government alone prints legal tender. So if a religious observer sincerely believes he needs a government resource to exercise his religion, including cash, the dissent’s “unique control” principle offers no practical limitation on what resources the government may need to give the religious observer. Arbitrarily carving out government favors for a religion that requires specific real property would invite discrimination against religions with different property needs. 14 14 So to recap: I not only think it would badly misinterpret RFRA to revise it the way the dissent does (Part I above), but I also think it would be a bad idea that would necessarily force the government to discriminatorily pick religious winners and losers in the distribution of its largesse (this Part II). Judge Nelson does not dispute my prediction that it would result in discrimination, but instead disputes my premise that such discrimination would be odious to the promise of religious liberty contained in both RFRA and the Constitution’s religion clauses. That surprises me. Since long before Smith was decided, it has been a bedrock principle of American religious liberty law that the government “cannot prefer one religion over another.” Larson v. Valente, 456 U.S. 228, 246 (1982) (quoting Everson v. Bd. of Educ., 330 U.S. ^[1], 15 (1947)). With that time-honored principle in mind, I’m not sure what Judge Nelson is suggesting in his three hypotheticals. I would think it is beyond dispute that the government cannot discriminate by allowing a devout Muslim prisoner to grow a beard for religious reasons while disallowing the same or a similar religious exception for devout Jewish or Native American prisoners. See, e.g., Warsoldier v. Woodford, 418 F.3d 989 (9th Cir. 2005); Sprouse v. Ryan, 346 F. Supp. 3d 1347 (D. Ariz. 2017). Is Judge Nelson seriously contending we could require a religious zoning exemption for a Catholic cathedral to build a 100-foot steeple, yet deny a mosque across the street the same exemption to build a 100-foot minaret? And does anyone seriously believe that a school-choice program that gave voucher money to Catholic schools but not Lutheran schools would pass constitutional muster? It has taken too long for the Supreme Court to recognize that discrimination against religion vis-à-vis supposedly “secular” counterparts is constitutionally problematic. See, e.g., Locke v. Davey, 540 U.S. 712 (2004). But there has always been widespread acceptance that discrimination between religions is repugnant to the Constitution.

D. The dissent further encourages discrimination by reading a reparations theory into RFRA. Ultimately, none of the distinctions either explicitly or implicitly relied on by the dissent to rationalize its rewrite of RFRA have any basis in its text or original meaning. So what might better explain the result the dissent would prefer this court to reach? It appears that, buttressed by the argument of academics who appeared as amici in this case, what the dissent is really advocating for is what might best be called a reparations version of RFRA. See Stephanie H. Barclay & Michalyn Steele, Rethinking Protections for Indigenous Sacred Sites, 134 Harv. L. Rev. 1294 (2021).

Under this “reconceptualized” and “alternative” theory of RFRA, Native Americans have a special historical and religious need for government-owned land because that land once belonged to them. As the academics explain, because the ancestors of Native Americans were mistreated and their land was taken, RFRA (and other laws) should be re-read to give current tribal members “unique” access to federal land. Id. at 1297–1303. Whatever the merits of these academic arguments, this court rightly declined to rewrite RFRA in service to them. If Native Americans are going to get unique protection of their religious exercise, they need to obtain it from Congress, not ask the courts to pretend they already got it from Congress. i. Amici’s reparations theory of RFRA has no basis in RFRA.

For starters, the academic argument motivating the dissent’s approach has no basis in the text or original meaning of RFRA, nor does it pretend to. The scholars pushing their theory openly acknowledge that courts have historically interpreted RFRA and the Free Exercise Clause to the contrary, id. at 1297, and that their approach requires courts to “recontextualize the way in which the law … view[s] coercion”—and thus what constitutes a burden—under RFRA, id. at 1302. Boiled down, theirs is a reparations theory of religious liberty for Native Americans, and Native Americans alone. Obviously, the reader will search RFRA in vain for any intergenerational theory of reparations, for Native Americans or otherwise. There is simply nothing in the text to that effect, and unsurprisingly, nobody at the time of RFRA’s enactment thought it was providing some type of reparations benefit.

To overcome RFRA’s obvious textual silence, these scholars try to draw an analogy from religious accommodations in inherently coercive contexts—namely, prisons. If this sounds familiar, that’s because it’s the same analogy suggested by the dissent, which asserts that the transfer of Oak Flat “prevents the Apaches from practicing their religious beliefs … just as would an outright ban or religious worship … in prison.” They correctly observe that the reason religious inmates are entitled to receive government property in prison to practice their religions under RLUIPA is because of the inherently coercive environment of prison. Id. at 1333. Just as prisons are under exclusive government control, the argument goes, many sites sacred to Native Americans are under exclusive government control, and therefore the government should more proactively give its property to indigenous persons to offset the coercion suffered by their ancestors when the government took their land in the first place. Id. at 1339–43.

It’s an interesting academic theory, and not one entirely devoid of moral force. But as already noted, nothing shows that Congress was attempting to do anything reparations- related when it passed RFRA. Even assuming the coercive removal of Native Americans from their lands can be analogized in some way to the coercion experienced by prison inmates, direct and immediate coercion is entirely different from ancestral coercion. The religious liberty of an inmate is directly and immediately implicated by the extreme version of coercion the government has imposed on that inmate. In contrast, the “reconceptualized” version of coercion relied on by the scholars’ attempted rewrite of RFRA is the governmental coercion of the ancestors of present-day Native Americans. This reparations-based theory is not entirely different from saying the Fourth Amendment should be applied specially to modern-day African Americans because of the lingering effects of slavery. Again, regardless of whether the theory has any merit, the idea that RFRA meant this when it was enacted in 1993 is entirely unfounded. RFRA was enacted to protect religious freedoms from current and future interference, not to turn back the clock and hunt for past burdens for which future religious devotees might be remunerated.

ii. To avoid discrimination, a reparations theory of RFRA would entitle a wide variety of religions to government handouts.

But that isn’t the only problem with a reparations theory of RFRA. Even assuming that religious reparations for ancestral coercion were somehow legitimate, what is the limiting principle? Should every religious person who can plausibly claim ancestral discrimination be entitled to religious reparations? RFRA is supposed to be generally applicable to protect all religions, so surely if reparations for government-sanctioned ancestral coercion of Native Americans are available under RFRA, they should also be available to others. Native Americans are not the only recipients of past government-imposed or government- allowed mistreatment arguably affecting their modern-day religious exercise. Indeed, if the dissent’s reparations theory of RFRA were ever adopted, one could expect swaths of religious claimants to line up for government benefits, each carrying the historical pedigree of discrimination against their respective religious tradition in tow.

Baptists in colonial Virginia were horsewhipped and their ministers were imprisoned when the Church of England enjoyed a monopoly there. 15 Catholics were deprived of their political and civil rights at various times in all thirteen colonies, 16 antebellum mobs burned down their churches and occasionally massacred them, 17 and efforts to ratify a constitutional amendment designed to clamp down on their parochial schools—the “Blaine Amendment of 1870”— gained widespread traction after the Civil War. 18 Mormons 15 Michael W. McConnell, The Origins and Historical Understanding of Free Exercise of Religion, 103 Harv. L. Rev. 1409, 1421–23 (1990). 16 Forrest McDonald, Novus Ordo Seclorum: The Intellectual Origins of the Constitution 42 (1985).

17 E.g., Sydney E. Ahlstrom, A Religious History of the American People 561 (2d ed. 2004) (describing anti-Catholic riots in Boston), 563 (describing riots in Philadelphia and New York), 1090 (In the United States, “Catholics were subjected to disabilities, intolerance, and violence from the earliest times.”); Sean Wilentz, The Rise of American Democracy: Jefferson to Lincoln 451 (2005); Kurt T. Lash, The Second Adoption of the Establishment Clause: The Rise of the Nonestablishment Principle, 27 Ariz. St. L.J. 1085, 1118–20 (1995) (describing a massacre of Catholics in Kentucky).

18 Espinoza v. Montana Dep’t of Revenue, 140 S. Ct. 2246, 2259 (2020) (quoting Mitchell v. Helms, 530 U.S. 793, 828–29 (2000) (“The Blaine Amendment was ‘born of bigotry’ and ‘arose at a time of pervasive hostility to the Catholic Church and to Catholics in general’; many of its state counterparts have similarly shameful pedigree.”)); see Richard White, The Republic for Which It Stands: The United States During were violently expelled from Missouri in 1838,19 denied the right to vote in Idaho in the 1880s, 20 and had their settlements in Utah undercut by the federal government in favor of Native Americans. 21 The first Jews to arrive in the colonies were nearly expelled because of their religion,22 Ulysses S. Grant’s notorious “General Orders No. 11” expelled Jews from defeated Confederate territories, 23 and “anti-Semitism began to grow virulent as soon as the Jewish immigration rate started to rise during the 1880s.” 24 And of course, one could surely argue that some African Americans today continue to experience the lingering effects of slavery and segregation as resource constraints on the uninhibited exercise of their religion. 25 Black churches were Reconstruction and the Gilded Age, 1865–1896, at 317–21, in 7 Oxford Hist. of the United States (David M. Kennedy ed. 2017). See generally John C. Jeffries & James E. Ryan, A Political History of the Establishment Clause, 100 Mich. L. Rev. 279, 301–05 (2001). 19 See, e.g., Marie H. Nelson, Anti-Mormon Mob Violence and the Rhetoric of Law and Order in Early Mormon History, 21 Legal Stud. F. 353, 358–73 (1997).

20 Davis v. Beason, 133 U.S. 333, 345–48 (1890), overruled by Romer v. Evans, 517 U.S. 620, 634 (1996).

21 See Uintah Ute Indians of Utah v. United States, 28 Fed. Cl. 768, 772– 73 (1993).

22 Eli Faber, America’s Earliest Jewish Settlers, 1654–1820, at 25, in The Columbia Hist. of Jews and Judaism in Am. (Marc Lee Raphael ed. 2008).

23 See, e.g., Eric Muller, All the Themes but One, 66 U. Chi. L. Rev. 1395, 1420–24 (1999).

24 Ahlstrom, supra, at 973–74, 1090. 25 See, e.g., In re African-American Slave Descendants Litig., 471 F.3d 754, 759–60 (7th Cir. 2006); Cato v. United States, 70 F.3d 1103, 1105– 06, 1109–11 (9th Cir. 1995); see also Margaret Russell, Cleansing sporadically suppressed by Southern states before the Civil War, 26 Bull Connor arrested congregants by the busload as they left the safety of the sanctuary to march for equal rights in the streets, 27 and some of the church buildings they left behind were bombed in their absence. 28 History is replete with examples of the mistreatment of groups of people by other groups, and this nation’s history is unfortunately not exempt. Given this reality, it’s unclear why the reparations theory of RFRA offered by the dissent would stop with Native Americans and not extend to Baptists, Catholics, Mormons, Jews, and descendants of slaves, to name but a few possible groups.

Regardless of the philosophical arguments for and against reparations, RFRA was not designed to create reparations for any aggrieved religious group. There is zero legal or textual basis for reading such a program into RFRA. If reparations are ever to come from any source, it must be from Congress, not the courts. And until Congress enacts religious reparations for Native Americans, courts should Moments and Retrospective Justice, 101 Mich. L. Rev. 1225, 1240 (2003).

26 Steven Hahn, A Nation Under Our Feet: Black Political Struggles in the Rural South from Slavery to the Great Migration 45 (2003). 27 Taylor Branch, Pillar of Fire: America During the King Years 1963– 65 77 (1998).

28 Id. at 137–38; see also Church Fires in the Southeast: Hearing Before the H. Comm. on the Judiciary, 104th Cong. 9–13 (1996) (statement of Donald L. Payne, Representative in Congress from the State of New Jersey, summarizing church burning incidents under criminal investigation in 1995–1996 in the Southeast states). See generally S. Willoughby Anderson, The Past on Trial: Birmingham, the Bombing, and Restorative Justice, 96 Calif. L. Rev. 471 (2008). studiously avoid inventing such remedies under the auspices of RFRA, a statute designed to protect religious liberty for all. RFRA does not play favorites, and neither should we. For these reasons, I wholeheartedly agree with the majority’s refusal to rewrite RFRA to include an affirmative mandate to discriminate.

MURGUIA, Chief Judge, dissenting, with whom GOULD, BERZON, and MENDOZA, Circuit Judges, join, and LEE, Circuit Judge, joins as to all but Part II.H:

We are asked to decide whether the utter destruction of Chí’chil Biłdagoteel, a site sacred to the Western Apaches since time immemorial, is a “substantial burden” on the Apaches’ sincere religious exercise under the Religious Freedom Restoration Act (“RFRA”), 42 U.S.C. §§ 2000bb to bb-4. Under any ordinary understanding of the English language, the answer must be yes. This conclusion comports with the First Amendment’s protection against government conduct prohibiting the free exercise of religion, because the destruction of the Apaches’ sacred site will prevent worshipers from ever again exercising their religion. See U.S. Const. amend. I.

Our decision in Navajo Nation v. United States Forest Service, 535 F.3d 1058 (9th Cir. 2008) (en banc), wrongly defined “substantial burden” as a narrow term of art and foreclosed any relief. Although a majority of this en banc court rejects Navajo Nation’s reasoning, see Nelson Op. at 130; Collins Op. at 51–52 (no mention of Navajo Nation while recognizing that in certain instances “substantial burden” under RFRA can be read by its plain meaning), a different majority concludes that the Apaches’ RFRA claim fails under Lyng v. Northwest Indian Cemetery Protective Association, 485 U.S. 439 (1988). Relying on Lyng, Judge Collins’ majority opinion (“the majority”) holds that the destruction of a sacred site cannot be described as a substantial burden no matter how devastating the impact on religious exercise, erroneously concluding that preventing a religious practice is neither prohibitory nor coercive. In so doing, the majority misreads RFRA, Supreme Court precedent, and our own case law. And rather than using the rare opportunity of sitting en banc to provide clarity, the majority leaves litigants in the dark as to what “substantial burden” means. I respectfully dissent.

I. Background

In a rider to a must-pass defense spending bill, Congress directed the Secretary of Agriculture to transfer 2,422 acres of federal land to Resolution Copper Mining, a foreign- owned limited liability company, to build an underground copper mine. The copper ore is located beneath Chí’chil Biłdagoteel, also known as Oak Flat, a sacred place where Western Apache people have worshiped and conducted ceremonies since time immemorial. ^[1] Once the land transfer occurs, Resolution Copper will mine the ore through a panel caving process, causing the land to subside and eventually creating a crater nearly two miles wide and a thousand feet deep. It is undisputed that this subsidence will destroy the Apaches’ historical place of worship, preventing them from ever again engaging in religious exercise at their sacred site.

1 Western Apache generally refers to the Apaches living in modern day Arizona, including ancestors of the White Mountain, San Carlos, Cibecue, and Tonto Apache.

The land transfer, however, is subject to RFRA. Congress enacted RFRA to protect the right to engage in religious practice without substantial government interference, which “the framers of the Constitution” understood “as an unalienable right.” 42 U.S.C. § 2000bb(a)(1). Thus, under RFRA, the federal government must provide a “compelling” justification pursued by the least restrictive means for any action that “substantially burden[s]” sincere religious exercise. Id. § 2000bb-1(b). Apache Stronghold, an Arizona nonprofit organization founded by a former Chairman of the San Carlos Apache Tribe to preserve Indigenous sacred sites, sued to enjoin the land transfer, arguing that, among other things, it violates RFRA. The district court, relying on our decision in Navajo Nation, declined to preliminarily enjoin the transfer, concluding that the destruction of Oak Flat did not amount to a substantial burden on the Apaches’ religious exercise. The district court therefore did not determine whether the government had provided sufficient justification for the land transfer.

Because the land transfer will prevent Apache worshippers from engaging in sincere religious exercise at their sacred site, I would hold that Apache Stronghold is likely to succeed in establishing that the government has imposed a “substantial burden” on the Apaches’ religious exercise. Such a holding stems from the Supreme Court’s jurisprudence before and after the enactment of RFRA, as well as our own case law, which have long recognized that preventing people from engaging in religious exercise impermissibly burdens that exercise. And such a decision reflects the government’s unique control of access to Oak Flat, a degree of control that is rare outside the prison and land-use context. I would therefore reverse the district court’s order concluding that there is no substantial burden, vacate the rest of the order, and remand to the district court to determine whether the government can demonstrate that the substantial burden posed by the land transfer is justified under subsection 2000bb-1(b).

A. Oak Flat and the Land Transfer

The Western Apache believe that their ancestral landscape is imbued with diyah, or spiritual power. This is especially true for Chí’chil Biłdagoteel, which means “Emory Oak Extends on a Level” or “Flat with Acorn Trees” or more simply “Oak Flat,” a 6.7-square-mile sacred site located primarily in the Tonto National Forest. Oak Flat is situated between Ga’an Bikoh (Devil’s Canyon), a canyon east of Oak Flat, and Dibecho Nadil (Apache Leap), the edge of a plateau west of Oak Flat.

Oak Flat, Devil’s Canyon, and Apache Leap comprise a hallowed area where the Apaches believe that the Ga’an— the “guardians” and “messengers” between Usen, the Creator, and people in the physical world—dwell. Usen created the Ga’an as “the buffer between heaven and earth” and created specific “blessed places” for the Ga’an to reside. The Ga’an are “the very foundation of [Apache] religion,” and they protect and guide the Apache people. The Apaches describe the Ga’an as their “creators, [their] saints, [their] saviors, [and their] holy spirits.”

Through Usen and the Ga’an, the Apaches believe that everything has life, including air, water, plants, animals, and Nahagosan—Mother Earth herself. The Apaches strive to remain “intertwined with the earth, with the mother” so they can “communicate with what [is] spiritual, from the wind to the trees to the earth to what [is] underneath.” Because of the importance of remaining connected to the land, the Apaches view Oak Flat as a “direct corridor” to their Creator’s spirit and as the place where the Ga’an “live and breathe.” Oak Flat is thus “uniquely endowed with holiness and medicine,” and neither “the powers resident there, nor [the Apaches’] religious activities . . . can be ‘relocated.’” The Ga’an come “to ceremonies to impart well-being to” the Apaches “to heal, and to help the people stay on the correct path.” Oak Flat thus serves as a sacred ceremonial ground, and these ceremonies cannot take place “anywhere else.” For instance, young Apache women have a coming- of-age ceremony, known as a “Sunrise Ceremony,” in which each young woman will “connect her soul and her spirit to the mountain, to Oak Flat.” Similarly, “young boys that are coming into manhood” have a sweat lodge ceremony at Oak Flat. There, the Apaches also conduct a Holy Grounds Ceremony, which is a “blessing and a healing ceremony . . . for people who are sick, have ailments[,] or seek guidance.” The Apaches gather “sacred medicine plants, animals, and minerals essential to [these] ceremonies” from Oak Flat, and they use “the sacred spring waters that flow[] from the earth with healing powers” that are not present elsewhere. “Because the land embodies the spirit of the Creator,” if the land is desecrated, then the “spirit is no longer there. And so without that spirit of Chí’chil Biłdagoteel, [Oak Flat] is like a dead carcass.” Apache Stronghold v. United States, 519 F. Supp. 3d 591, 604 (D. Ariz. 2021).

The Apaches have held Oak Flat sacred since long before the United States government and its people ventured west of the Rio Grande. The Apaches, however, were dispossessed from their ancestral land during the nineteenth century, when miners and settlers moved west and clashed repeatedly with the local Apaches. To make peace, various Apache leaders signed the Treaty of Santa Fe in 1852, wherein the United States government promised the Apaches that it would “designate, settle, and adjust their territorial boundaries” and “pass and execute” laws “conducive to the prosperity and happiness of” their people. Despite the treaty, conflict continued as more settlers, miners, and United States soldiers entered the Apaches’ ancestral land, resulting in several massacres of the Apaches by soldiers and civilians. By the late 1870s, the United States government forcibly removed the Apaches from their ancestral homelands and onto reservations, so that today, the Apaches no longer live on lands encompassing their sacred places. Nonetheless, the Apaches “remain connected to their spirituality” and “the earth,” and they continue to come to Oak Flat to worship, conduct ceremonies, sing and pray, and gather sacred plants. Apache Stronghold, 519 F. Supp. 3d at 603–04.

In the twentieth century, the United States government took steps to protect Oak Flat from mining activity. In 1955, President Eisenhower reserved 760 acres of Oak Flat for “public purposes” to protect it from mineral exploration or other mining-related activities. 20 Fed. Reg. 7319, 7336–37 (Oct. ^[1], 1955). President Nixon renewed that protection in 1971. 36 Fed. Reg. 18,997, 19,029 (Sept. 25, 1971). That approach changed in 1995, after miners discovered a large copper deposit 7,000 feet beneath Oak Flat. The following decades saw several congressional attempts to transfer Oak Flat to Resolution Copper. Those efforts reached fruition in 2014, when Congress passed the National Defense Authorization Act for Fiscal Year 2015, Pub. L. No. 113-291 (2014) (“NDAA”). The NDAA included a rider that stripped Oak Flat’s mining protections and “authorized and directed” the Secretary of Agriculture to convey 2,422 acres of federal land, including Oak Flat, to Resolution Copper in exchange for 5,344 acres of Arizona land currently owned by the company. See id. § 3003, 128 Stat. 3292 (codified at 16 U.S.C. § 539p) (the “Land Transfer Act”). 2 Congress’s stated purpose for authorizing the exchange is to “carry out mineral exploration activities under” Oak Flat. ^[16] U.S.C. § 539p(c)(6)(A)(i).

Under the Land Transfer Act, the Secretary of Agriculture must prepare an environmental impact statement (“EIS”) before the land transfer may take place. See id. § 539p(c)(9)(B). 3 This EIS will “be used as the basis for all” federal government decisions “significantly affecting the quality of the human environment,” including permitting necessary for any development of the transferred land. Id. The EIS must “assess the effects of the mining and related activities on the Federal land conveyed to Resolution Copper under [the Land Transfer Act] on the cultural and archeological resources that may be located on [that] land” and “identify measures that may be taken, to the extent practicable, to minimize potential adverse impacts on those resources.” Id. § 539p(c)(9)(C). Within sixty days of the Final EIS’s publication, and regardless of its contents, “the Secretary shall convey” the land to Resolution Copper. Id. § 539p(c)(10).

In January 2021, the Forest Service, a division of the Department of Agriculture, issued an EIS, which has since 2 The 2,422-acre tract is known as the “Oak Flat Federal Parcel,” and includes the 760-acre section of land originally protected by President Eisenhower in 1955 (known as the “Oak Flat Withdrawal Area”) as well as additional National Forest Service lands near Oak Flat. The copper deposit sits primarily beneath the Oak Flat Withdrawal Area. ^[3] The Land Transfer Act is subject to several other conditions not at issue here. See, e.g., 16 U.S.C. § 539p(c)(2)(A), (B). been withdrawn. In that EIS, the Forest Service concluded that the land transfer would remove Oak Flat from the Forest Service’s jurisdiction, making the Forest Service unable to “regulate” the mining activity under applicable environmental laws. The Forest Service found that the mine would be “one of the largest” and “deepest” “copper mines in the United States,” with an estimated 1,970 billion metric tons of copper situated 4,500 to 7,000 feet beneath Oak Flat. Resolution Copper will use an underground mining technique known as panel caving that carves a network of tunnels below the ore. As the ore is removed, the land above the ore “moves downward or ‘subsides.’” This “subsidence zone” or crater will reach between 800 and 1,115 feet deep and nearly two miles wide. The crater would start to appear within six years of active mining. The crater and related mining activity will have a lasting impact on the land of approximately eleven square miles. The Forest Service “assessed alternative mining techniques in an effort to prevent subsidence, but alternative methods were considered unreasonable.”

As a result of the crater, the Forest Service determined that “access to Oak Flat and the subsidence zone will be curtailed once it is no longer safe for visitors.” The Forest Service therefore concluded that the mine would cause “immediate, permanent, and large in scale” destruction of “archaeological sites, tribal sacred sites, cultural landscapes, and plant and mineral resources.” 4 Oak Flat would “be permanently affected,” and tribal members would 4 Removing the ore will also create roughly one-and-a-half billion tons of waste that will need to be stored “in perpetuity” at a site close to Oak Flat. The Forest Service determined that development of the storage facility will “permanently bury or otherwise destroy many prehistoric and historic cultural artifacts, potentially including human burials.” irreversibly lose access to the area for “religious purposes,” thus resulting in “an indescribable hardship to [Indigenous] peoples.” “[T]he impacts of the Resolution Copper [mine] . . . are substantial and irreversible due to the changes that would occur at Oak Flat.” The Forest Service also found that there are no mitigation measures that could “replace or replicate the historic properties that would be destroyed by project construction. . . . Archaeological sites cannot be reconstructed once disturbed, nor can they be fully mitigated.”

In March 2021, the Department of Agriculture ordered the Forest Service to rescind the EIS. The Department explained that the government needed “additional time” to “fully understand concerns raised by Tribes and the public” and to “ensure the agency’s compliance with federal law.” While counsel for the government informed the en banc panel at oral argument in March 2023 that the environmental analysis would be completed and the EIS republished by the summer, the Forest Service has not yet issued a revised Final EIS.

B. Procedural History

Apache Stronghold filed this action several days before the government issued the now-withdrawn EIS. 5 As 5 Besides this case, there are two other pending cases seeking to prevent the land transfer. In January 2021, the San Carlos Apache Tribe sued the Forest Service to stop the land transfer under RFRA, the Free Exercise Clause, and the 1852 Treaty of Santa Fe, and moved to vacate the now withdrawn EIS as deficient under the Administrative Procedure Act (“APA”), the National Environmental Policy Act (“NEPA”), the Land Transfer Act, and the National Historic Preservation Act. See San Carlos Apache Tribe v. U.S. Forest Serv., No. 21-cv-0068 (D. Ariz.). Also in January 2021, a coalition of environmental and tribal groups sued the Forest Service to enjoin the land transfer and vacate the EIS as deficient relevant on appeal, Apache Stronghold alleges that the Land Transfer Act violates RFRA, the First Amendment’s Free Exercise Clause, and trust duties created by the 1852 Treaty of Santa Fe. Two days after filing its complaint, Apache Stronghold filed a motion for a temporary restraining order and for a preliminary injunction to prevent the government from transferring the land to Resolution Copper. The district court denied the temporary restraining order, reasoning that Apache Stronghold could not show immediate and irreparable injury. Apache Stronghold, 519 F. Supp. 3d at 597.

The district court then held a hearing and took evidence before denying Apache Stronghold’s motion for a preliminary injunction. Id. at 611. The district court found that Apache Stronghold was unlikely to succeed on the merits of its RFRA, Free Exercise Clause, and breach of trust claims. See id. at 598–609. As to the RFRA claim, the district court concluded that although the “Government’s mining plans on Oak [Flat] will have a devastating effect on under the APA, NEPA, the Land Transfer Act, the Forest Service Organic Act, the Federal Land Policy and Management Act, and other statutory grounds. See Ariz. Mining Reform Coal. v. U.S. Forest Serv., No. 2:21-cv-0122-DLR (D. Ariz.). Resolution Copper intervened in both cases, and the Defendants moved to consolidate all three cases. The district court in this case denied that motion, concluding that “there is minimal overlap in controlling questions of law between the pending cases” given the different legal theories advanced by the three plaintiffs.

The parties agreed to stay both cases after the Forest Service withdrew its original EIS. See San Carlos Apache Tribe, No. 21-cv-0068 (D. Ariz. Mar. 15, 2021); Ariz. Mining Reform Coal., No. 21-cv-0122 (D. Ariz. Mar. 15, 2021). Those cases remain stayed, and the parties have filed regular joint status reports. The government has stated that it will give the defendants sixty days’ notice prior to filing an updated Final EIS. As of now, that notice has not been given. the Apache people’s religious practices,” there was no “substantial burden” under this circuit’s limited definition of that term. Id. at 605–08 (citing Navajo Nation, 535 F.3d at 1063–72). The district court therefore did not determine whether the government could establish a compelling interest to justify its actions, nor did the district court analyze the other preliminary injunction factors under Winter v. Natural Resources Defense Council, Inc., 555 U.S. 7, 20 (2008). See Apache Stronghold, 519 F. Supp. 3d at 611. Apache Stronghold appealed, and moved for an injunction pending appeal.

After the district court denied Apache Stronghold’s preliminary injunction motion, the Forest Service withdrew the Final EIS. The three-judge motions panel that considered Apache Stronghold’s motion for an injunction pending appeal therefore concluded that Apache Stronghold had failed to show that it needed immediate relief to “avoid irreparable harm,” because the Forest Service expected to take “months” to complete its revised environmental review and the land transfer would not occur until then. Apache Stronghold v. United States, No. 21-15295, 2021 U.S. App. LEXIS 6562, at *2 (9th Cir. March 5, 2021) (“Injunction Order”). Accordingly, the divided motions panel denied Apache Stronghold’s motion. Id. In dissent, Judge Bumatay stated that he would have granted the motion and held that the land transfer violated RFRA because “the complete destruction of the land . . . . is an obvious substantial burden on [the Apaches’] religious exercise, and one that the Government has not attempted to justify.” Id. at *5 (Bumatay, J., dissenting).

On the merits, a divided three-judge panel affirmed the district court’s order. Apache Stronghold v. United States, 38 F.4th 742 (9th Cir. 2022). We granted rehearing en banc. Apache Stronghold v. United States, 56 F.4th 636 (9th Cir. 2022). 6 II. Discussion In Winter, the Supreme Court emphasized that injunctive relief, whether temporary or permanent, is an “extraordinary remedy never awarded as of right.” 555 U.S. at 24. A party seeking a preliminary injunction must show that: (1) it is “likely to succeed on the merits”; (2) it is “likely to suffer irreparable harm in the absence of preliminary relief”; (3) “the balance of equities tips in [its] favor”; and (4) “an injunction is in the public interest.” Id. at 20. “Where, as here, the government opposes a preliminary injunction, the third and fourth factors merge into one inquiry.” Porretti v. Dzurenda, 11 F.4th 1037, 1047 (9th Cir. 2021).

The district court concluded that Apache Stronghold could not establish a likelihood of success on any of its three claims, so it denied the motion for a preliminary injunction. See Apache Stronghold, 519 F. Supp. 3d at 598–609. Because I conclude that Navajo Nation’s reasoning is incorrect and because I would hold that preventing a person from engaging in sincere religious exercise is a substantial burden under RFRA, I would reverse and remand. I would therefore consider neither the other two claims nor the remaining Winter factors. Finally, I conclude that RFRA applies to the Land Transfer Act. Because a majority of judges have voted to affirm, I respectfully dissent.

6 After oral argument, Resolution Copper intervened in this case before the district court, as well as before this court, for the limited purpose of participating in potential future litigation before the Supreme Court. A. RFRA and the Religious Land Use and Institutionalized Persons Act In RFRA, Congress crafted a statutory right to the free exercise of religion broader than the corresponding constitutional right delineated by the Supreme Court in Employment Division, Department of Human Resources of Oregon v. Smith, 494 U.S. 872 (1990). In Smith, the Supreme Court held that the First Amendment tolerates neutral, generally applicable laws even when those laws burden or prohibit religious acts. Id. at 885–90. The Supreme Court explained that so long as the government’s burden on religious exercise, even if substantial, was not the “object of” a law, “the First Amendment has not been offended” and the government need not demonstrate a narrowly tailored, compelling governmental interest to justify it. Id. at 878–79; see also id. at 886 n.3 (“[G]enerally applicable, religion-neutral laws that have the effect of burdening a particular religious practice need not be justified by a compelling governmental interest.”).

In response, in 1993, Congress enacted RFRA. Congress disagreed with the Supreme Court’s decision in Smith to “virtually eliminate[] the requirement that the government justify burdens on religious exercise imposed by laws neutral toward religion.” 42 U.S.C. § 2000bb(a)(4). Instead, Congress found that “the framers of the Constitution[] recogniz[ed the] free exercise of religion as an unalienable right,” and that governments, therefore, “should not substantially burden religious exercise without compelling justification.” Id. § 2000bb(a)(1), (3). Congress further determined that “the compelling interest test”—i.e., strict scrutiny—“is a workable test for striking sensible balances between religious liberty and competing prior governmental interests.” Id. § 2000bb(a)(5); see Gonzales v. O Centro Espírita Beneficente Uniaõ do Vegetal, 546 U.S. 418, 430 (2006). Congress then stated that RFRA’s two “purposes” were (1) “to restore the compelling interest test as set forth in Sherbert v. Verner, 374 U.S. 398 (1963) and Wisconsin v. Yoder, 406 U.S. 205 (1972)[,] and to guarantee its application in all cases where free exercise of religion is substantially burdened,” and (2) “to provide a claim or defense to persons whose religious exercise is substantially burdened by government.” 42 U.S.C. § 2000bb(b). RFRA therefore goes “far beyond what . . . is constitutionally required” under the Free Exercise Clause, and thus “provide[s] very broad protection for religious liberty.” Burwell v. Hobby Lobby Stores, Inc., 573 U.S. 682, 706 (2014); see Ramirez v. Collier, 595 U.S. 411, 424 (2022).

Four years later, however, the Supreme Court struck down the portion of RFRA regulating state and local governments, concluding that Congress had exceeded its power under § 5 of the Fourteenth Amendment to regulate states. City of Boerne v. Flores, 521 U.S. 507, 511, 536 (1997). To repair RFRA’s constitutional defect, Congress enacted the Religious Land Use and Institutionalized Persons Act of 2000 (“RLUIPA”), 114 Stat. 803, 42 U.S.C. §§ 2000cc to cc-5, “which applies to the States and their subdivisions and invokes congressional authority under the Spending and Commerce Clauses.” Holt v. Hobbs, 574 U.S. 352, 357 (2015). Recognizing their history and overlapping purposes, the Supreme Court has characterized RLUIPA and RFRA as “sister statute[s]” that “impose[] the same general test,” distinguished only in that they apply to different “categor[ies] of governmental actions.” Hobby Lobby, 573 U.S. at 695, 730. In contrast to RFRA’s more general application to all federal government action, including federal prisons and federal land-use regulations by the District of Columbia or U.S. territories, see 42 U.S.C. §§ 2000bb-1, 2000bb-3, RLUIPA governs only state land-use regulations, see id. § 2000cc, and religious exercise by institutionalized persons, typically in the state prison context, see id. § 2000cc-1. RLUIPA otherwise generally “mirrors RFRA.” Holt, 574 U.S. at 357–58; compare 42 U.S.C. § 2000cc-1(a) (providing that a “substantial burden” in the state prison context must be justified by a compelling governmental interest pursued through the least restrictive means); with id. § 2000bb-1(b) (same test for federal government action).

B. Defining “Substantial Burden” i. Plain Meaning

With that background in mind, I turn to Apache Stronghold’s claim that the government will violate RFRA by transferring Oak Flat to Resolution Copper, which will result in the destruction of the Apaches’ place of worship. Under RFRA, the federal government may not “substantially burden a person’s exercise of religion . . . except as provided in subsection (b).” 42 U.S.C. § 2000bb-1(a). Subsection (b) provides that the “Government may substantially burden a person’s exercise of religion only if it demonstrates that application of the burden to the person—(1) is in furtherance of a compelling governmental interest; and (2) is the least restrictive means of furthering that compelling governmental interest.” Id. § 2000bb-1(b). Thus, to proceed with its RFRA claim, Apache Stronghold must show that (i) its sincere religious exercise is (ii) subject to a substantial burden imposed by the government. If Apache Stronghold makes that showing, the government must then justify that burden by demonstrating that (iii) it has a compelling interest that (iv) it is pursuing through the least restrictive means.

As to the Apaches’ religious exercise, the district court found, and the government does not dispute, that the Apaches have a sincere religious belief in worshipping and conducting ceremonies at Oak Flat. See Apache Stronghold, 519 F. Supp. 3d at 603; see also 42 U.S.C. §§ 2000bb-2(4), 2000cc-5(7)(A) (defining the “exercise of religion” to include “any exercise of religion, whether or not compelled by, or central to, a system of religious belief”). 7 Because the government concedes that “it is undisputed that RFRA applies to federal land-management statutes and their implementation,” on appeal, we must determine whether the transfer and resulting destruction of Oak Flat constitutes a substantial burden on the Apaches’ religious exercise.

To define “substantial burden,” I begin with RFRA’s text. Tanzin v. Tanvir, 592 U.S. 43, 46 (2020); Williams v. Taylor, 529 U.S. 420, 431 (2000). Because RFRA does not define “substantial burden,” I “turn to the phrase’s plain meaning at the time of enactment.” Tanzin, 592 U.S. at 48; see also FCC v. AT & T Inc., 562 U.S. 397, 403 (2011). Indeed, when grappling with RFRA’s undefined terms, the Supreme Court has done just that. Tanzin, 592 U.S. at 45– 49 (looking to RFRA’s plain meaning, using dictionaries, to conclude that “appropriate relief” encompasses claims for money damages against government officials in their individual capacities).

7 RFRA appropriately does not permit courts to judge the significance or “centrality” of a particular belief or practice, given that courts are not the proper arbiters of religious doctrine. See 42 U.S.C. §§ 2000bb-2(4), 2000cc-5(7)(A). Courts can only inquire into the sincerity of the professed religiosity. See Hobby Lobby, 573 U.S. at 696, 717 n.28; cf. Cutter v. Wilkinson, 544 U.S. 709, 725 n.13 (2005).

At the time of RFRA’s passage, a “burden” was defined as “[s]omething oppressive” or “anything that imposes either a restrictive or onerous load” on an activity. Burden, Black’s Law Dictionary (6th ed. 1990); Webster’s Third New International Dictionary 298 (1986) (defining burden as “something that weighs down [or] oppresses”). A burden is “substantial” if it is “[o]f ample or considerable amount, quantity, or dimensions.” Substantial, Oxford English Dictionary 66–67 (2d ed. 1989). And “substantial” does not mean complete or total. Substantial, Black’s Law Dictionary (6th ed. 1990) (defining “substantial” as something “considerable”; not “nominal”). In light of the plain meaning of substantial burden, therefore, RFRA prohibits government action that “oppresses” or “restricts” “any exercise of religion, whether or not compelled by, or central to, a system of religious belief,” to a “considerable amount,” unless the government can demonstrate that imposition of the burden is in furtherance of a compelling governmental interest and the least restrictive means of furthering that compelling governmental interest. Accord Injunction Order, 2021 U.S. App. LEXIS 6562, at *8–9 (Bumatay, J., dissenting). ii. Navajo Nation’s Flawed Reasoning

Our decision in Navajo Nation, relied upon by the district court, rejected a plain meaning reading of “substantial burden.” There, Native American tribes and their members sought to enjoin the use of artificial snow, made from recycled wastewater, on a public mountain sacred to their religion. Navajo Nation, 535 F.3d at 1062–63. This court concluded that using artificial snow was not a substantial burden under RFRA, because “the sole effect of the artificial snow is on the Plaintiffs’ subjective spiritual experience.” Id. at 1063, 1070 (emphasis added). Aside from holding that subjective interference with religious exercise is not a substantial burden under RFRA, Navajo Nation also concluded that because Congress “incorporated” Sherbert and Yoder into RFRA, the only two categories of burden that could constitute a “substantial burden” are the specific types of burdens at issue in those cases. 535 F.3d at 1069–70; see also id. at 1063. Navajo Nation therefore held:

Under RFRA, a “substantial burden” is imposed only when individuals are forced to choose between following the tenets of their religion and receiving a governmental benefit (Sherbert) or coerced to act contrary to their religious beliefs by the threat of civil or criminal sanctions (Yoder). Any burden imposed on the exercise of religion short of that described by Sherbert and Yoder is not a “substantial burden” within the meaning of RFRA, and does not require the application of the compelling interest test set forth in those two cases.

Id. at 1069–70. This is erroneous for six reasons. First, Navajo Nation made too much of the fact that RFRA explicitly mentions Sherbert and Yoder by name in explaining the statute’s purpose. See 535 F.3d at 1074–75. Reading “substantial burden” by its plain language is fully consistent with RFRA’s statements of purpose. Congress explained that RFRA’s two “purposes” are (1) “to restore the compelling interest test as set forth in Sherbert and Yoder[,] and to guarantee its application in all cases where free exercise of religion is substantially burdened,” and (2) “to provide a claim or defense to persons whose religious exercise is substantially burdened by government.” 42 U.S.C. § 2000bb(b) (emphasis added) (citations omitted). Section 2000bb(b) thus links Sherbert and Yoder to the “compelling interest test,” not to the “substantial burden” inquiry. See 42 U.S.C. § 2000bb(b) (not mentioning Sherbert or Yoder in RFRA’s second purpose). Consonant with the statute’s purposes, the Supreme Court has recognized that “RFRA expressly adopted the compelling interest test ‘as set forth in Sherbert and Yoder.’” Gonzales, 546 U.S. at 431 (quoting 42 U.S.C. § 2000bb(b)(1) (emphasis added) (citations omitted)). “In each of those cases, [the] Court looked beyond broadly formulated interests justifying the general applicability of government mandates and scrutinized the asserted harm of granting specific exemptions to particular religious claimants.” Id.

In other words, when enacting RFRA, Congress was focused on governments’ justifications for burdens on religious exercise created by generally applicable laws—the requirement present in Sherbert and Yoder that Smith eliminated—not the definition of substantial burden. Justice O’Connor, concurring only in the judgment in Smith, made this point when she critiqued the Smith majority for dropping the “Sherbert compelling interest test” and argued that “[r]ecent cases have instead affirmed that [compelling interest] test as a fundamental part of our First Amendment doctrine. The cases cited by the [majority] signal no retreat from our consistent adherence to the compelling interest test.” Smith, 494 U.S. at 898, 900 (O’Connor, J., concurring in the judgment) (emphasis added) (cleaned up). Justice O’Connor notably did not describe the test as the “Sherbert substantial burden test,” because her disagreement with the Smith majority was not with the meaning of substantial burden but with the level of scrutiny. And the Smith majority never defined substantial burden because it concluded the Sherbert test was entirely “inapplicable” in cases challenging neutral, generally applicable laws. See id. at 884–85.

Second, neither Sherbert nor Yoder contains the term “substantial burden.” It would therefore be surprising for Congress to invoke an interpretation of a purported term of art by referencing two cases, neither of which uses the term. See Sherbert, 374 U.S. at 406 (“substantial infringement”); Yoder, 406 U.S. at 220 (“unduly burdens”). Navajo Nation’s argument that “substantial burden” is a term of art from the Supreme Court’s pre-RFRA First Amendment jurisprudence makes little sense given that neither case includes that term. 535 F.3d at 1074. Indeed, the Supreme Court did not commonly or consistently use the term “substantial burden.”

In Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, for example, decided just months before Congress enacted RFRA, the Court explained that “[a] law burdening religious practice that is not neutral or not of general application must undergo the most rigorous of scrutiny,” without using the term “substantial burden.” 508 U.S. 520, 546 (1993). If “substantial burden” truly was a term of art, then one would expect consistent usage. See Yellen v. Confederated Tribes of Chehalis Rsrv., 141 S. Ct. 2434, 2445 (2021) (“Ordinarily . . . this Court reads statutory language as a term of art only when the language was used in that way at the time of the statute’s adoption.”).

In looking to the term’s plain meaning, I do not ignore the significance of RFRA mentioning Sherbert and Yoder by name. But rather than implausibly reading “substantial burden” as a term of art shackled to Sherbert and Yoder, I rely on those cases—along with other “Federal court rulings,” 42 U.S.C. § 2000bb(a)(5)—to properly situate “substantial burden” within RFRA. See infra § II(D). And it would unreasonably contort the English language to read “substantial burden” to exclude the utter destruction of sacred sites. “Because common sense rebels” at the majority’s interpretation of RFRA, “we should not adopt that interpretation unless the statutory language compels us to conclude that Congress intended such a startling result.” United States v. Ibarra-Galindo, 206 F.3d 1337, 1341 (9th Cir. 2000) (Canby, J., dissenting).

Third, Navajo Nation (and the majority here) proceeds as if RFRA’s coverage is identical to that of the Free Exercise Clause, frozen in time at the moment of the statute’s enactment. But Congress amended RFRA in 2000 and repealed RFRA’s previous definition of the “exercise of religion” as “the exercise of religion under the First Amendment to the Constitution.” Pub. L. No. 103-141, § 5 (1993). As the Supreme Court explained: “[t]hat amendment deleted the prior reference to the First Amendment,” and it is unclear “why Congress did this if it wanted to tie RFRA coverage tightly to the specific holdings of our pre-Smith free-exercise cases.” Hobby Lobby, 573 U.S. at 714. Congress also broadened the definition of “religious exercise” in two ways: it eliminated any requirement that a religious exercise be “compelled by, or central to, a system of religious belief,” 42 U.S.C. § 2000cc- 5(7)(A), and it specified that “religious exercise” includes “[t]he use, building, or conversion of real property for the purpose of religious exercise,” 42 U.S.C. § 2000cc-5(7)(B). The term “substantial burden” must therefore be construed in light of Congress’s express direction that RFRA applies to the use of property for religious purposes. See U.S. Nat’l Bank of Or. v. Indep. Ins. Agents of Am., Inc., 508 U.S. 439, 455 (1993) (explaining that statutory construction “is a holistic endeavor,” so “in expounding a statute, we must not be guided by a single sentence or member of a sentence, but look to the provisions of the whole law” (quotation marks omitted) (cleaned up)). That Congress amended RFRA to expressly include religious use of property reinforces my conclusion that the denial of religious exercise at a sacred site is a substantial burden on religious exercise, contrary to the holding of Navajo Nation.

Fourth, considering this amendment to RFRA, and after Navajo Nation, the Supreme Court has rejected the notion that RFRA “merely restored [its] pre-Smith decisions in ossified form.” Hobby Lobby, 573 U.S. at 715–16. Instead, the Court explained that “the amendment of RFRA through RLUIPA surely dispels any doubt” that Congress did not intend “to tie RFRA coverage tightly to the specific holdings of our pre-Smith free-exercise cases.” Id. at 714; see also id. at 706 n.18 (explaining that there is “no reason to believe” that RFRA “was meant to be limited to situations that fall squarely within the holdings of pre-Smith cases”). I therefore rely on pre-Smith cases for guidance only.

Fifth, and relatedly, as discussed in the next section, Navajo Nation’s choice to confine “substantial burden” to a term of art cannot stand in the face of the Supreme Court’s directive that RFRA and RLUIPA impose “the same standard.” Holt, 574 U.S. at 356–58 (quoting Gonzales, 546 U.S. at 436); see also Food Mktg. Inst. v. Argus Leader Media, 139 S. Ct. 2356, 2365 (2019) (noting that courts do not “ordinarily imbue statutory terms with a specialized . . . meaning when Congress has not itself invoked” one). Finally, instead of just answering the question before it, Navajo Nation’s decision to define substantial burden as a narrow term of art swept too broadly. Cf. City of Ontario v. Quon, 560 U.S. 746, 760 (2010) (“A broad holding . . . might have implications for future cases that cannot be predicted.”). This case asks whether the utter destruction of a sacred site is a substantial burden. That is a fundamentally different question than the one Navajo Nation considered, because there, plaintiffs still had “virtually unlimited access to the mountain” to “continue to pray, conduct their religious ceremonies, and collect plants for religious use.” Navajo Nation, 535 F.3d at 1063 (emphasis added); see id. (noting that nothing “with religious significance, or religious ceremonies . . . would be physically affected”). Because the Navajo Nation majority went to great lengths to emphasize that “no places of worship [were] made inaccessible,” id., Navajo Nation should not have adopted a rule that extends to cases where places of worship will be obliterated. And by adopting such a broad holding, it erred.

Accordingly, I would revise Navajo Nation’s definition of “substantial burden” to the extent that it defined that phrase as a term of art limited to the kinds of burdens at issue in Sherbert and Yoder. Rather, as discussed infra § II(D), the kinds of burdens challenged in Sherbert and Yoder are examples sufficiently demonstrating a substantial burden, not those necessary to do so. 8 C. RFRA and RLUIPA Are Interpreted Uniformly RLUIPA, RFRA’s sister statute, supports my conclusion to define substantial burden by its plain meaning. RLUIPA’s 8 As reflected in the first paragraph of the per curiam opinion, a majority of this court has overruled Navajo Nation’s narrow test for a “substantial burden” under RFRA. I echo Judge Nelson’s clear refutation of any suggestion to the contrary. See Nelson Op. at 135–38. “substantial burden” test largely mirrors RFRA’s test, and like RFRA, it does not define “substantial burden.” See 42 U.S.C. §§ 2000cc, 2000cc-1, 2000cc-5(4)(A). So, as we did in San Jose Christian College v. City of Morgan Hill, I look to RLUIPA’s plain meaning to interpret “a ‘substantial burden’ on ‘religious exercise’” in the land-use context as “a significantly great restriction or onus upon such exercise.” 360 F.3d 1024, 1034 (9th Cir. 2004); id. (“When a statute does not define a term, a court should construe that term in accordance with its ordinary, contemporary, common meaning.” (quotation marks omitted)). Since then, we have relied on this plain meaning definition of substantial burden in other RLUIPA cases. See, e.g., Guru Nanak Sikh Soc. of Yuba City v. County of Sutter, 456 F.3d 978, 988–89 (9th Cir. 2006); Int’l Church of Foursquare Gospel v. City of San Leandro, 673 F.3d 1059, 1067 (9th Cir. 2011). 9 That “substantial burden” has the same meaning under both RFRA and RLUIPA is a logical application of statutory construction for several reasons. First, it is significant that these two Title 42 statutes use the same “substantial burden” and “compelling interest” language. See United States v. Nishiie, 996 F.3d 1013, 1026 (9th Cir. 2021) (“When Congress uses the same language in two statutes having similar purposes,” this Court starts with the “presum[ption] 9 Dictionaries contemporaneous with the enactments of RFRA and RLUIPA define “substantial” synonymously as either a “considerable” or a “significant” amount. To the extent there is any semantic difference, I conclude that the meaning of “substantial” is the same under both statutes, particularly given that RLUIPA was meant to restore part of RFRA’s original reach. See Holt, 574 U.S. at 357–58 (RLUIPA “mirrors RFRA”); Gonzales, 546 U.S. at 436 (RLUIPA allows incarcerated people “to seek religious accommodations pursuant to the same standard as set forth in RFRA.”).

that Congress intended that text to have the same meaning in both statutes.” (quotation marks omitted) (cleaned up)); Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 172–73 (2012) (presumption of consistent usage). The term “religious exercise” also has an identical definition in the two statutes. See 42 U.S.C. §§ 2000bb-2(4), 2000cc-5(7)(A). The two sister statutes differ only in what categories of government action they control: RFRA applies to all federal action, including federal prisons and land-use restrictions, whereas RLUIPA governs state government land-use regulations and state prisons. Diverging definitions for identical terms in the two statutes would allow federal prisons to burden religious rights more heavily than state prisons, or vice versa, which is implausible given the statutes’ history and purpose. See Gonzales, 546 U.S. at 436; Holt, 574 U.S. at 356–58 (explaining that the two statutes impose “the same standard”); Cutter, 544 U.S. at 716–17 (“To secure redress for [incarcerated persons] who encountered undue barriers to their religious observances, Congress carried over from RFRA [to RLUIPA] the ‘compelling governmental interest’/‘least restrictive means’ standard.”); see also Austin v. U.S. Navy Seals 1–26, 142 S. Ct. 1301, 1307 (2022) (Alito, J., dissenting) (explaining that RLUIPA “essentially requires prisons to comply with the RFRA standard”).

Second, the Supreme Court has cross-referenced the two statutes for support. See, e.g., Holt, 574 U.S. at 356–57 (a RLUIPA case invoking RFRA cases); Hobby Lobby, 573 U.S. at 695, 729 n.37 (a RFRA case invoking RLUIPA cases).

Third, at least seven other circuits agree with my conclusion that the two statutes’ “substantial burden” standards are one and the same. See, e.g., Mack v. Warden Loretto FCI, 839 F.3d 286, 304 n.103 (3d Cir. 2016) (“[T]he two statutes are analogous for purposes of the substantial burden test.”); Madison v. Riter, 355 F.3d 310, 315 (4th Cir. 2003) (RLUIPA “reinstate[d] RFRA’s protection against government burdens” and “mirror[s]” its provisions); A.A. ex rel. Betenbaugh v. Needville Indep. Sch. Dist., 611 F.3d 248, 264 n.64 (5th Cir. 2010) (“same ‘substantial burden’ question”); Korte v. Sebelius, 735 F.3d 654, 682–83 (7th Cir. 2013) (“same understanding”); Patel v. U.S. Bureau of Prisons, 515 F.3d 807, 813 (8th Cir. 2008) (“same definition”); Hobby Lobby Stores, Inc. v. Sebelius, 723 F.3d 1114, 1138 n.13 (10th Cir. 2013) (“interpreted uniformly”), aff’d sub nom. Hobby Lobby, 573 U.S. 682; Eternal Word Television Network, Inc. v. Sec’y of U.S. Dep’t of Health & Human Servs., 818 F.3d 1122, 1144 n.23 (11th Cir. 2016) (“same substantial burden analysis”); see also Sabir v. Williams, 52 F.4th 51, 60 & n.5 (2d Cir. 2022) (applying RLUIPA’s substantial burden precedent to a RFRA claim); EEOC v. R.G. & G.R. Harris Funeral Homes, Inc., 884 F.3d 560, 587 (6th Cir. 2018) (relying on Holt, a RLUIPA case, to define substantial burden in a RFRA case), aff’d sub nom. Bostock v. Clayton County, 140 S. Ct. 1731 (2020). The great weight of authority thus buttresses my conclusion that RFRA and RLUIPA employ the same substantial burden test defined by its plain meaning. D. Preventing a Person from Engaging in Religious Exercise Is an Example of a Substantial Burden I next consider which government actions amount to a substantial burden on religious exercise. Keeping in mind that RFRA did not “merely restore[ the Supreme] Court’s pre-Smith decisions in ossified form,” Hobby Lobby, 573 U.S. at 715, the Supreme Court’s pre-Smith Free Exercise jurisprudence, as well as our own case law, provide at least three clear examples of a substantial burden on religious exercise: where the government (1) forces a religious adherent to choose between sincere religious exercise and receiving government benefits; (2) threatens a religious adherent with civil or criminal sanctions for engaging in sincere religious exercise; or (3) prevents a person from engaging in sincere religious exercise. i. Pre-Smith Free Exercise Jurisprudence I begin with Sherbert and Yoder, the two pre-Smith cases that RFRA mentions by name. See 42 U.S.C. § 2000bb(b)(1). In Sherbert, a state employer fired a Seventh-day Adventist because she refused to work on Saturdays, her faith’s day of rest. 374 U.S. at 399. The state denied the plaintiff’s claim for unemployment compensation benefits, finding that she had failed to accept work without good cause. Id. at 399–401. The Supreme Court held that the state’s denial of unemployment compensation to the plaintiff because she was exercising her faith imposed a “substantial infringement” under the Free Exercise Clause. Id. at 403–04, 406. Such a condition unconstitutionally forced the plaintiff “to choose between following the precepts of her religion and forfeiting benefits, on the one hand, and abandoning one of the precepts of her religion in order to accept work, on the other hand.” Id. at 404. Having determined that there was a “substantial infringement” on religious exercise, the Court then “consider[ed] whether some compelling state interest enforced in the eligibility provisions of the [state] statute justifie[d] the substantial infringement of [her] First Amendment right,” and held that the state’s concern about protecting against “fraudulent [unemployment] claims” was insufficiently compelling. Id. at 406–09.

In Yoder, a state prosecuted members of the Amish faith for violating a state law that required children to attend school until the age of sixteen. 406 U.S. at 207–08. The defendants sincerely believed that their children’s attendance in high school was “contrary to the Amish religion and way of life.” Id. at 209. The Supreme Court reversed the convictions, holding that the application of the compulsory school-attendance law to the defendants “unduly burden[ed]” their exercise of religion in violation of the Free Exercise Clause. Id. at 207, 220. According to the Court, the state law “affirmatively compel[led the defendants], under threat of criminal sanction, to perform acts undeniably at odds with fundamental tenets of their religious beliefs.” Id. at 218. As to the state’s interest underlying its truancy law, the Court explained that a general interest in compulsory education was insufficiently compelling. Id. at 221.

But pre-RFRA precedents did not limit the kinds of burdens protected under the Free Exercise Clause to the types of burdens challenged in Sherbert (the choice between sincere religious exercise and receiving government benefits) and in Yoder (the threat of civil or criminal sanctions). Beyond these two cases, the Supreme Court’s pre-Smith jurisprudence recognizes at least one other category of government action that violates the Free Exercise Clause: preventing a religious adherent from engaging in religious exercise. In Cruz v. Beto, for example, a prison denied a Buddhist access to the prison chapel and prohibited him from corresponding with his religious advisor. 405 U.S. 319, 322 (1972) (per curiam). The Court reversed the dismissal of the complaint and held that, taking the allegations as true, the prison had violated the Free Exercise Clause. Id.

And in O’Lone v. Estate of Shabazz, prison officials “prevented Muslims . . . from attending Jumu’ah,” an Islamic congregational service held on Friday afternoons. 482 U.S. 342, 347 (1987). The plaintiffs sued, “alleging that the prison policies unconstitutionally denied them their Free Exercise rights under the First Amendment.” Id. The Supreme Court recognized that preventing Muslims from engaging in religious exercise gave rise to a cognizable Free Exercise Clause claim. But, at the time, before RFRA and RLUIPA, prison officials were only required to show that a policy that burdened religious exercise was “reasonable.” Id. at 350. So the Court concluded that preventing Muslims from attending religious services was “justified by concerns of institutional order and security.” Id.; see id. at 351–52 (concluding that, although there were “no alternative means of attending Jumu’ah,” the prison policy of preventing religious exercise was reasonable because “alternative means of exercising the [First Amendment] right” remained open as the plaintiffs were “not deprived of all forms of religious exercise” such as daily prayer).

In dissent, Justice Brennan agreed that preventing an adherent from engaging in religious practices was sufficient to demonstrate a Free Exercise claim, but disagreed with the majority’s reasonableness standard:

The prison in this case has completely prevented respondent inmates from attending the central religious service of their Muslim faith. I would therefore hold prison officials to the standard articulated in Abdul Wali, [which requires the government to demonstrate a compelling interest] and would find their proffered justifications wanting.

The State has neither demonstrated that the restriction is necessary to further an important objective nor proved that less extreme measures may not serve its purpose. Id. at 359 (Brennan, J., dissenting). RFRA and RLUIPA later essentially codified Justice Brennan’s dissent, eliminating the reasonableness test for evaluating prison policies and instead requiring federal and state prison policies that substantially burden religious exercise to be justified by a compelling interest furthered by the least restrictive means. See 42 U.S.C. § 2000cc-1(a); id. § 2000bb-1(b). 10 RFRA also instructs that courts look to “prior Federal court rulings.” 42 U.S.C. § 2000bb(a)(5). Like the Supreme Court, our own cases prior to Smith recognized that preventing a person from engaging in religious exercise implicates the Free Exercise Clause. For instance, in Graham v. Commissioner of Internal Revenue, we required a religious adherent, there a taxpayer, to show that the 10 Other pre-Smith examples falling outside the Sherbert/Yoder framework are Free Exercise Clause challenges to government autopsies. See Tanzin, 592 U.S. at 51 (noting that autopsies are among the cases in which RFRA grants effective relief) (citing Yang v. Sturner, 728 F. Supp. 845 (D.R.I. 1990) (autopsy of son that violated Hmong beliefs), opinion withdrawn in light of Smith, 750 F. Supp. 558 (D.R.I. 1990)); see also City of Boerne, 521 U.S. at 547 (O’Connor, J., concurring in part) (discussing Yang as an example of why Smith was wrongly decided in the context of RFRA); Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1893 & n.26 (2021) (Alito, J., concurring in judgment) (discussing the import of Yang in the lead up to Congress enacting RFRA and stating that “Smith’s impact was quickly felt, and Congress was inundated with reports of the decision’s consequences” (citing 139 Cong. Rec. 9681 (1993))).

government action “burdens the adherent’s practice of his or her religion by pressuring him or her to commit an act forbidden by the religion or by preventing him or her from engaging in conduct or having a religious experience.” 822 F.2d 844, 850–51 (9th Cir. 1987) (emphasis added), aff’d sub nom. Hernandez v. Comm’r, 490 U.S. 680 (1989).

The same is true in other cases. See, e.g., McElyea v. Babbitt, 833 F.2d 196, 197–99 (9th Cir. 1987) (citing O’Lone and recognizing a Free Exercise Clause claim where a prison had no weekly Jewish services and the plaintiff alleged that prison officials “prevented him from practicing his religion”); Allen v. Toombs, 827 F.2d 563, 567 (9th Cir. 1987) (assuming that denial of access to a sweat lodge was a viable Free Exercise Clause claim, but upholding the prison policy under the O’Lone, pre-RFRA, reasonableness test); cf. Freeman v. Arpaio, 125 F.3d 732, 736 (9th Cir. 1997) (holding, in a Free Exercise Clause case decided post-City of Boerne and pre-RLUIPA, that “[i]n order to establish a free exercise violation, [a plaintiff] must show the defendants burdened the practice of his religion, by preventing him from engaging in [religious exercise], without [proper] justification” (footnote omitted)). ii. This Circuit’s Precedents Recognize Preventing Religious Exercise Is a Substantial Burden Given this legal backdrop, it is unsurprising that in our first RFRA case in 1995, we relied on pre-Smith Free Exercise Clause cases to define substantial burden to include preventing a person from engaging in religious exercise. In Bryant v. Gomez, we held that to show a “substantial burden” under RFRA, the religious adherent has the obligation to prove that a governmental action burdens the adherent’s practice of his or her religion by preventing him or her from engaging in conduct or having a religious experience . . . . This interference must be more than an inconvenience.

46 F.3d 948, 949 (9th Cir. 1995) (per curiam) (cleaned up) (quoting Graham, 822 F.2d at 850–51). 11 The majority makes no effort to explain why we should not adhere to Bryant’s formulation of substantial burden. Nor does it distinguish our subsequent pre-Navajo Nation RFRA cases in which we consistently invoked the concept of preventing a person from engaging in religious conduct as a substantial burden in various contexts, including ones outside of the two RLUIPA contexts. For example, in a case considering a university’s mandatory student registration fee that, in part, covered abortion services, we “look[ed] to our 11 In Bryant, we rejected the plaintiff’s RFRA claim because “full Pentecostal services” were not “mandated by his faith.” 46 F.3d at 949 (stating that religious exercise must be one that “the faith mandates” or “a tenet or belief that is central to religious doctrine”). However, as discussed supra § II(B)(ii), in 2000, Congress expanded the statutory protection for religious exercise by amending RFRA and RLUIPA’s definition of “exercise of religion” to include “any exercise of religion, whether or not compelled by, or central to, a system of religious belief.” 42 U.S.C. §§ 2000bb-2(4), 2000cc-5(7)(A). So to the extent that Bryant and other cases discussed below applied a narrower definition of “religious exercise” that required it to be central to or mandated by a person’s faith, Congress has abrogated them. Similarly, RFRA and RLUIPA’s definition of “exercise of religion” is broader than O’Lone and Freeman’s definition under the Free Exercise Clause. Otherwise, Bryant’s discussion of substantial burden remains good law. decisions prior to Smith,” including a Free Exercise Clause challenge by a taxpayer, to define substantial burden to include “preventing [a person] from engaging in conduct or having a religious experience.” Goehring v. Brophy, 94 F.3d 1294, 1299 (9th Cir. 1996) (quoting Graham, 822 F.2d. at 850–51, and discussing Bryant); see also Worldwide Church of God v. Phila. Church of God, Inc., 227 F.3d 1110, 1121 (9th Cir. 2000) (citing Bryant’s substantial burden standard in a copyright case and concluding that the unauthorized use of intellectual property of religious texts was not a substantial burden under RFRA); Stefanow v. McFadden, 103 F.3d 1466, 1471 (9th Cir. 1996) (citing Bryant’s standard and finding no substantial burden because an incarcerated person was not “prevented” from “engaging in any [religious] practices” when the prison confiscated a religious text not central to his practice). 12 Similarly, before and since Navajo Nation, we have routinely recognized that preventing religious exercise qualifies as a substantial burden under RLUIPA, which applies the “same standard” as RFRA, Holt, 574 U.S. at 356– 57. See Johnson v. Baker, 23 F.4th 1209, 1215–16 (9th Cir. 2022) (recognizing that prohibiting plaintiff from possessing scented prayer oil in his cell substantially burdened his religious exercise); Foursquare Gospel, 673 F.3d at 1061, 1066–70 (recognizing that preventing the plaintiff from building a place of worship could constitute a substantial burden); Greene v. Solano Cnty. Jail, 513 F.3d 982, 988 (9th 12 The Seventh, Eighth, and Tenth Circuits have followed Bryant’s interpretation of a substantial burden under RFRA. See Mack v. O’Leary, 80 F.3d 1175, 1179 (7th Cir. 1996) (expressly drawing on Bryant); Weir v. Nix, 114 F.3d 817, 820 (8th Cir. 1997); Werner v. McCotter, 49 F.3d 1476, 1480 (10th Cir. 1995) (citing Bryant). Cir. 2008) (“We have little difficulty in concluding that an outright ban on a particular religious exercise”—i.e., a “policy of prohibiting [a person] from attending group religious worship services”—“is a substantial burden on that religious exercise.”); Guru Nanak Sikh Soc’y of Yuba City, 456 F.3d at 981–82 (holding that a county “imposed a substantial burden” on a Sikh organization’s “religious exercise” by denying applications from the group for a conditional use permit to build a temple); cf. United States v. Antoine, 318 F.3d 919, 923–24 (9th Cir. 2003) (assuming that “raz[ing]” a “house of worship” to build a freeway would be a substantial burden). 13 13 Several other circuits also recognize that denying access to or preventing religious exercise qualifies as a substantial burden under RLUIPA. See Yellowbear v. Lampert, 741 F.3d 48, 55 (10th Cir. 2014) (Gorsuch, J.); Haight v. Thompson, 763 F.3d 554, 565 (6th Cir. 2014); Lovelace v. Lee, 472 F.3d 174, 187–88 (4th Cir. 2006); Murphy v. Mo. Dep’t of Corrs., 372 F.3d 979, 988 (8th Cir. 2004); cf. C.L. for Urb. Believers v. City of Chicago, 342 F.3d 752, 761 (7th Cir. 2003). Notably, the Tenth Circuit referenced this circuit’s definition of a substantial burden when defining it to include preventing religious exercise. See Werner, 49 F.3d at 1480 (citing Bryant); Abdulhaseeb v. Calbone, 600 F.3d 1301, 1313 (10th Cir. 2010) (citing Werner).

And in a recent RLUIPA case, the Supreme Court stayed the execution of an incarcerated person who requested that “his long-time pastor be allowed to pray with him and lay hands on him while he is being executed.” Ramirez, 595 U.S. at 416; see id. at 426, 433 (holding that the state’s refusal to permit audible prayer or religious touch, denying him access to his religious rites, “substantially burdens his exercise of religion,” because “he will be unable to engage in protected religious exercise in the final moments of his life”). E. The Land Transfer Act Substantially Burdens the Exercise of Religion The foregoing firmly establishes that where the government prevents a person from engaging in religious exercise, the government has substantially burdened the exercise of religion. The plain meaning of RFRA clearly reaches such instances. The Free Exercise Clause cases prior to Smith so recognized. O’Lone, 482 U.S. at 347–52; Graham, 822 F.2d at 850–51. We held as much in our first RFRA case. See Bryant, 46 F.3d at 949. And, as Judge Bumatay pointed out in his dissent from the order declining to enjoin the land transfer pending appeal, this understanding is consistent with RLUIPA. See Injunction Order, 2021 U.S. App. LEXIS 6562, at *9 (Bumatay, J., dissenting) (“[A]s then-Judge Gorsuch wrote [in a RLUIPA case], a substantial burden exists when the government ‘prevents the plaintiff from participating in an activity motivated by a sincerely held religious belief.’” (quoting Yellowbear v. Lampert, 741 F.3d 48, 55 (10th Cir. 2014))).

I now turn to whether Apache Stronghold is likely to succeed in showing that the transfer and eventual destruction of Oak Flat constitutes a substantial burden on the Western Apaches’ religious exercise. The district court heard extensive testimony about the impact of the land transfer and mine. The district court found:

Because the land embodies the spirit of the Creator, “without any of that, specifically those plants, because they have that same spirit, that same spirit at Oak Flat, that spirit is no longer there. And so without that spirit of Chí’chil Biłdagoteel, it is like a dead carcass.” If the mining activity continues, Naelyn Pike testified, “then we are dead inside. We can’t call ourselves Apaches.” Quite literally, in the eyes of many Western Apache people, Resolution Copper’s planned mining activity on the land will close off a portal to the Creator forever and will completely devastate the Western Apaches’ spiritual lifeblood. . . . [T]he land in this case will be all but destroyed to install a large underground mine, and Oak Flat will no longer be accessible as a place of worship. Apache Stronghold, 519 F. Supp. 3d at 604, 606 (citations omitted).

As discussed supra § I(A), the Forest Service, in its now-withdrawn EIS, similarly documented the extensive, irreversible, and devastating impact of the mine’s construction, and how the mining activity would prevent Apache worshipers from engaging in religious exercise at their religious sites. The crater will start to appear within six years of active mining, and the Forest Service concluded that the mining activity will cause “immediate” and “permanent” destruction of “archaeological sites, tribal sacred sites, cultural landscapes, and plant and mineral resources.” In addition, once the government publishes its Final EIS, regardless of its contents, “the Secretary shall convey” the land to Resolution Copper within sixty days. ^[16] U.S.C. § 539p(c)(10) (emphasis added). So once the land transfer occurs, Oak Flat will be private property no longer subject to RFRA and other federal protections.

In other words, the land transfer will result in a crater that will subsume Oak Flat. The impact of the mining activity on sacred sites will be immediate and irreversible. All that will be left is a massive hole and rubble, making the site unsuitable for religious exercise. Religious worship will be impossible, and the Apaches will be prevented from ever again worshipping at Oak Flat. As I have concluded, where the government prevents a religious adherent from engaging in religious exercise, the government has restricted the exercise of religion to a considerable amount. I would therefore hold that Apache Stronghold is likely to succeed in establishing that transferring Oak Flat to Resolution Copper will amount to a substantial burden under RFRA. See 42 U.S.C. § 2000bb-1(a). Because the district court did not determine whether the government could justify that burden by demonstrating a compelling interest pursued through the least restrictive means, I would remand for the district court to make that determination in the first instance. See id. § 2000bb-1(b).

F. Lyng Is Consistent with My Analysis i. Lyng and Prohibitions on Free Exercise

The majority concludes that the destruction of a sacred site cannot be a substantial burden but cites no authority squarely supporting that proposition. Indeed, the majority fails to cite even one case foreclosing a RFRA claim where the government completely prevents a person from engaging in religious exercise. Confusingly, the majority agrees with me that then-Judge Gorsuch correctly held in Yellowbear “that ‘prevent[ing] the plaintiff from participating in an activity motivated by a sincerely held religious belief’ qualifies as prohibiting free exercise.” Collins Op. at 33 (quoting Yellowbear, 741 F.3d at 55). And the majority concedes that it is undisputed that the Land Transfer Act will categorically prevent the Apaches from participating in any worship at Oak Flat because their religious site will be obliterated. See Collins Op. at 23. If the majority agrees with Yellowbear’s formulation—which mirrors the one I have laid out above in § II(D) (explaining that preventing religious exercise is an example of a substantial burden)— and agrees that the Apaches will be prevented from worshiping at Oak Flat, Apache Stronghold’s claim cannot fail. See Injunction Order, 2021 U.S. App. LEXIS 6562, at *9–10 (Bumatay, J., dissenting) (relying on Yellowbear to conclude that the destruction of Oak Flat is a substantial burden). And yet, the majority says that it does.

Rather than acknowledge this inconsistency, the majority relies entirely on a pre-RFRA Free Exercise Clause case: Lyng v. Northwest Indian Cemetery Protective Association, 485 U.S. 439 (1988). But Lyng cannot bear the weight the majority places on it.

The Supreme Court in Lyng did not analyze whether there was a substantial burden under the Free Exercise Clause. The case is therefore not inconsistent with my RFRA analysis and cannot foreclose Apache Stronghold’s statutory claim, which rests on the “substantial burden” concept.

In its retelling of Lyng, the majority omits crucial facts. The Lyng plaintiffs challenged the federal government’s proposal to permit timber harvesting and build a road through part of a national forest that “ha[d] traditionally been used for religious purposes by members of three American Indian tribes.” 485 U.S. at 441–42. The proposed road “avoided archeological sites and was removed as far as possible from the sites used by [tribes] for specific spiritual activities.” Id. at 443. Unlike here—a fact that the majority entirely disregards—“[n]o sites where specific rituals t[ook] place were to be disturbed.” Id. at 454. The Lyng plaintiffs continued to have full access to their sacred sites to engage in religious exercise, and there were “one-half mile protective zones around all the religious sites,” insulating them from any logging activity. See id. at 441–43. However, because the road and logging activity would generally disturb the “privacy,” “silence,” “spiritual development,” and the subjective enjoyment of those sacred sites, the plaintiffs brought a Free Exercise Clause challenge. Id. at 442, 444, 454 (citing the record to note that “successful use of the area is dependent upon and facilitated by certain qualities of the physical environment, the most important of which are privacy, silence, and an undisturbed natural setting” (cleaned up)); see id. at 462 (Brennan, J., dissenting) (quoting the record to highlight that “silence, the aesthetic perspective, and the physical attributes, are an extension of the sacredness of [each] particular site”).

Assuming that the noise and general disturbance from logging would “have severe adverse effects” on the individuals’ subjective religious experience, the Supreme Court held that the government’s actions did not trigger the compelling interest test under the Free Exercise Clause. Id. at 447, 450–51. Relying on Bowen v. Roy, 476 U.S. 693 (1986), the Court concluded that the Lyng plaintiffs’ subjective spiritual harm from the loss of silence and privacy was “incidental” to the government’s “internal” affairs. Lyng, 485 U.S. at 448, 451. In Roy, the Supreme Court had rejected a religious objection to the use of Social Security numbers as a numerical identifier that, according to the plaintiffs’ religious beliefs, would “‘rob the spirit’ of [their] daughter and prevent her from attaining greater spiritual power.” 476 U.S. at 696. The Roy Court held that the “Free Exercise Clause simply cannot be understood to require the Government to conduct its own internal affairs in ways that comport with the religious beliefs of particular citizens.” Id. at 699.

Applying Roy, the Lyng Court explained that the plaintiffs’ allegations of spiritual harm “cannot meaningfully be distinguished from the use of a Social Security number in Roy”:

Similarly, in this case, it is said that disruption of the natural environment caused by the . . . road will diminish the sacredness of the area in question and create distractions that will interfere with “training and ongoing religious experience of individuals using [sites within] the area for personal medicine and growth . . . and as integrated parts of a system of religious belief and practice which correlates ascending degrees of personal power with a geographic hierarchy of power.”

485 U.S. at 448–49 (quoting the record). The Court construed the harm in both cases as “subjective” and so refused to decide whether the spiritual harm in Roy was “significantly greater” than the Lyng plaintiffs’ harm. Id. at 449. 14 14 In rejecting the plaintiffs’ challenge, the Supreme Court did not minimize the impact that the road building and logging activity would have on the plaintiffs’ “personal spiritual development.” Lyng, 485 U.S. at 451. The Court, however, did not wish to weigh the magnitude of the subjective spiritual harm. Id. at 449, 451. So it explained that the noise and invasion of privacy caused by roadbuilding and logging had only an “incidental” constitutional effect under the Free Exercise Clause because the government was not “outright prohibit[ing]” religious exercise, Lyng emphasized that the “crucial word in the constitutional text [of the Free Exercise Clause] is ‘prohibit’: ‘For the Free Exercise Clause is written in terms of what the government cannot do to the individual, not in terms of what the individual can exact from the government.’” Id. at 451 (emphasis added) (quoting Sherbert, 374 U.S. at 412 (Douglas, J., concurring)). The Court therefore concluded its analysis by reiterating that “[t]he Constitution does not permit [the] government to discriminate against religions that treat particular physical sites as sacred, and a law prohibiting the Indian respondents from visiting the [sacred] area would raise a different set of constitutional questions.” Id. at 453 (emphasis added).

The majority argues that, as in Lyng, the land transfer here is not “a situation in which the Government ha[s] ‘discriminate[d]’ against the plaintiffs, as might be the case if Congress had passed ‘a law prohibiting the Indian [plaintiffs] from visiting the [sacred] area.’” Collins Op. at 31 (quoting Lyng, 485 U.S. at 453). The majority is mistaken on two fronts. First, the Land Transfer Act is exactly that kind of “prohibitory” law. It is undisputed and indisputable that once implemented, the Act will prevent the Western Apaches from visiting Oak Flat for eternity. The majority concedes this point, but then goes on to argue that where government action only “frustrates or inhibits” religious exercise, the government does not violate RFRA. “indirect[ly] coerc[ing]” an individual to act contrary to their religious belief, or “penal[izing]” religious practice. Id. at 450–51 (citing U.S. Const. amend. I; Sherbert, 374 U.S. at 404).

This discussion also highlights that Free Exercise Clause claims are not limited to the circumstances presented in Sherbert and Yoder but include the broader concept of “prohibitions.” Id. at 450; U.S. Const. amend. I.

But Apache Stronghold does not argue that the destruction of Oak Flat merely “frustrates” their ability to worship there; they argue—and the district court found—that worship there will be “impossible,” and their spiritual practice will be eviscerated. See Apache Stronghold, 519 F. Supp. 3d at 604 (“Quite literally, in the eyes of many Western Apache people, Resolution Copper’s planned mining activity on the land will close off a portal to the Creator forever and will completely devastate the Western Apaches’ spiritual lifeblood.”); id. at 606 (“[T]he land in this case will be all but destroyed to install a large underground mine, and Oak Flat will no longer be accessible as a place of worship.”). So, contrary to the majority, this case does not ask us to determine at what point “frustrating” religious exercise qualifies as a substantial burden; 15 instead, we are confronted only with the utter erasure of a religious practice. In other words, the burden here is categorical and thus undisputedly “synonymous with ‘prohibit.’” Collins Op. at 33.

15 See, e.g., Braunfeld v. Brown, 366 U.S. 599, 605 (1961) (plurality opinion) (no infringement where a law merely “operates so as to make the practice of [the individual’s] religious beliefs more expensive”); Lakewood, Ohio Congregation of Jehovah’s Witnesses, Inc. v. City of Lakewood, 699 F.2d 303, 306 (6th Cir. 1983) (similar); Goehring, 94 F.3d at 1299; Worldwide Church of God, 227 F.3d at 1121; United States v. Friday, 525 F.3d 938, 947 (10th Cir. 2008) (“We are skeptical that the bare requirement of obtaining a permit can be regarded as a ‘substantial burden’ under RFRA.”); see also Adkins v. Kaspar, 393 F.3d 559, 570 (5th Cir. 2004) (no infringement where government action “merely prevents the adherent from either enjoying some benefit that is not otherwise generally available or acting in a way that is not otherwise generally allowed”); Abdulhaseeb, 600 F.3d at 1316 (“[W]e do not intend to imply that every infringement on a religious exercise will constitute a substantial burden.”).

Second, that the Land Transfer Act does not specially “discriminate” against the Western Apaches by name—i.e., that the Act is neutral and generally applicable to all who would visit Oak Flat—is irrelevant because, when enacting RFRA, Congress eliminated Smith’s neutrality test. See 42 U.S.C. § 2000bb(a)(2) (“Congress finds that . . . laws ‘neutral’ toward religion may burden religious exercise as surely as laws intended to interfere with religious exercise.”). All that matters under RFRA, as opposed to the Free Exercise Clause, is whether the government has “substantially burden[ed]” sincere religious exercise. Id. § 2000bb-1(a). The majority thus misunderstands Congress’s purpose in enshrining a broad right to religious liberty by eliminating Smith’s neutrality requirement.

The majority argues that such a reading of RFRA is too “broad.” But a clear-cut conclusion that making religious exercise impossible is a “substantial burden” can hardly be called broad, especially when it adheres closely to both RFRA’s text and the Supreme Court’s precedent. The majority also contends that claims like Apache Stronghold’s would subject the government to “religious servitude.” Yet the majority proceeds as if, once a religious adherent has satisfied the substantial burden test, the outcome is a foregone conclusion. However, Congress explicitly identified the compelling interest test as “a workable test for striking sensible balances between religious liberty and competing prior governmental interests.” 42 U.S.C. § 2000bb(a)(5).

At this stage, Apache Stronghold has only proven that there is a substantial burden. On remand, the government could demonstrate that transferring Oak Flat is justified by a compelling interest pursued through the least restrictive means. ^[16] See Thomas v. Rev. Bd. of Ind. Emp. Sec. Div., 450 U.S. 707, 718 (1981) (“The mere fact that the petitioner’s religious practice is burdened by a governmental program does not mean that an exemption accommodating his practice must be granted. The state may justify an inroad on religious liberty by showing that it is the least restrictive means of achieving some compelling state interest.”); see also Gonzales, 546 U.S. at 430, 436 (rejecting the government’s “slippery slope” argument under RFRA, and noting that Sherbert did so under the Free Exercise Clause); cf. Cutter, 544 U.S. at 722 (stating that the Supreme Court had “no cause to believe” that the compelling interest test “would not be applied in an appropriately balanced way”). So although Lyng did not specifically address government action that prevented religious exercise, contrary to the 16 The compelling interest test has not proven fatal to the government. See Douglas Laycock & Thomas C. Berg, Protecting Free Exercise Under Smith and After Smith, Cato Sup. Ct. Rev. at 44–45 & n.66 (2020–21) (noting that “the compelling-interest standard has not come close to producing the ‘anarchy’ of which Smith warned” and finding that “free-exercise claims, including RFRA claims, were the least likely to invalidate the government action” (citing Adam Winkler, Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts, 59 Vand. L. Rev. 793, 857–58, 861 (2006))).

And if the majority were correct that my reading of RFRA would subject the government to “religious servitude,” then we would necessarily have seen that concern play out in circuits that have long employed a broader reading of “substantial burden.” Neither the government nor the majority provide evidence that other circuits are inundated with such claims, and I have found no evidence hinting at that possibility. Cf. Yellowbear, 741 F.3d at 62 (Gorsuch, J.) (rejecting slippery slope argument). In addition, before Smith, the government was not yoked to religious deference—as the majority and the government fears it would be—even though the Supreme Court had read the Free Exercise Clause to cover claims about preventing religious exercise. majority’s assertions, Lyng’s discussion of “discrimination” by “prohibiting” access to a sacred site confirms that the Land Transfer Act creates a substantial burden. ii. Lyng’s Post-RFRA Limits Moreover, to the degree Lyng’s Free Exercise ruling is in any tension with my understanding of RFRA, those aspects of Lyng were not carried forward into RFRA. Smith makes that much evident, as it treats Lyng as declining to apply the compelling interest test to a neutral law of general applicability, and RFRA displaced that standard for governmental decisions governed by RFRA.

Smith held that Lyng “declined to apply Sherbert analysis to the Government’s logging and road construction activities on lands used for religious purposes by several Native American Tribes, even though it was undisputed that the activities ‘could have devastating effects on traditional Indian religious practices.’” Smith, 494 U.S. at 883 (quoting Lyng, 485 U.S. at 451). Per Smith, Lyng stood for the proposition that the compelling interest test is “inapplicable” to “across-the-board” neutral laws. Smith, 494 U.S. at 884– 85. In declining to apply the compelling interest test, Smith relied on Lyng for the point that “[t]he government’s ability to enforce generally applicable prohibitions of socially harmful conduct, like its ability to carry out other aspects of public policy, ‘cannot depend on measuring the effects of a governmental action on a religious objector’s spiritual development.’” Smith, 494 U.S. at 885 (quoting Lyng, 485 U.S. at 451). Smith then concluded that “generally applicable, religion-neutral laws that have the effect of burdening a particular religious practice need not be justified by a compelling governmental interest.” Id. at 886 n.3. In so holding, Smith emphatically rejected Justice O’Connor’s concurrence suggesting that Lyng created an exception for Free Exercise challenges to the government’s conduct of its internal affairs. 494 U.S. at 885 n.2. 17 The Smith majority first acknowledged that “Justice O’Connor seeks to distinguish Lyng and Roy on the ground that those cases involved the government’s conduct of ‘its own internal affairs.’” Id. (citations omitted). Smith then considered Justice O’Connor’s position that challenges to the government’s conduct of its internal affairs are “different because, as Justice Douglas said in Sherbert, ‘the Free Exercise Clause is written in terms of what the government cannot do to the individual, not in terms of what the individual can exact from the government.’” Id. (internal quotation marks and citation omitted). “But,” said the Smith majority in refuting the internal affairs proposition, “that quote obviously envisioned that what ‘the government cannot do to the individual’ includes not just the prohibition of an individual’s freedom of action through criminal laws but also the running of its programs . . . in such fashion as to harm the individual’s religious interests.” Id. “Moreover,” Smith continued, “it is hard to see any reason in principle or practicality why the government should have to tailor its health and safety laws to conform to the diversity of religious belief, but should not have to tailor its management of public lands, Lyng, supra.” Id. (emphasis added). 18 17 Judge Nelson’s concurring opinion so recognizes. 18 As the Smith majority alluded to, it is hard to see how an exception permitting the government to substantially burden religious exercise when “manag[ing] its internal affairs,” Nelson Op. at 148, would not encompass most government action and indeed swallow RFRA whole. Smith treated Lyng as reflecting not any special exception for challenges to the government’s internal affairs, but as concerning the type of neutral and generally applicable laws not subject to the compelling interest test under Smith. Id. at 884–85 (citing Lyng, 485 U.S. at 451). Smith’s understanding of Lyng remains controlling. See Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1878 (2021) (“Smith . . . drew support for the neutral and generally applicable standard from cases involving internal government affairs.” (citing Lyng, 485 U.S. at 439)).

Accordingly, Lyng was not about measuring the extent of burdens sufficient to trigger the compelling interest test. Nor was Lyng, as the majority and concurring opinions posit, a case concerning the borders of the Free Exercise Clause or a special carve-out category of government actions that were not covered by Smith. Instead, Lyng reflected the principle, further developed in Smith and rejected in RFRA, that the compelling interest test was categorically inapplicable to neutral and generally applicable laws. See Smith, 494 U.S. at 884–85; Fulton, 141 S. Ct. at 1878.

Smith’s controlling interpretation of Lyng thus makes clear that (1) Lyng turned on the categorical inapplicability of the compelling interest test to the Free Exercise challenge in that case; and (2) the reason the compelling interest test was inapplicable in Lyng was that “the test [is] inapplicable to such challenges” to generally applicable laws. Smith, 494 U.S. at 885. RFRA’s rejection of Smith’s rule—that the compelling interest test is inapplicable to neutral and generally applicable laws—means that Lyng likewise does not control in RFRA cases.

The majority’s flawed response to this point is that Lyng did not involve a neutral or generally applicable law. Collins Op. at 35–36. But that proposition is wrong. Indeed, elsewhere in its opinion, the majority asserts, accurately, that Lyng did not involve “a situation in which the Government had ‘discriminate[d]’ against the plaintiffs, as might be the case if Congress had passed ‘a law prohibiting the Indian [plaintiffs] from visiting the [sacred] area.’” Collins Op. at 31 (quoting Lyng, 485 U.S. at 453). A law that “does not ‘discriminate’ against religious adherents,” like the policy in Lyng, is a neutral one for purposes of Free Exercise doctrine. See Church of the Lukumi Babalu Aye, 508 U.S. at 533 (explaining that a “law is not neutral” if “the object of a law is to infringe upon or restrict practices because of their religious motivation” (citing Smith, 494 U.S. at 878–89)). The plan to build the road at issue in Lyng was indisputably neutral in this sense, as it would affect equally all who preferred leaving the wilderness untouched— environmentalists, for example, or ranchers.

Nor is the majority correct that the policy challenged in Lyng was not generally applicable. In Lyng, the Forest Service proposed building a road connecting two towns and permitting timber harvesting in the same area; the road would be open to all, and there was no suggestion that the purpose of the Forest Service’s plan was to discriminate against Native American tribes. Indeed, the Forest Service took steps to mitigate the impact on tribes by “select[ing] a route that avoided archeological sites and was removed as far as possible from the sites used by [tribes] for specific spiritual activities.” Lyng, 485 U.S. at 443. While the litigation in Lyng was pending in the court of appeals, Congress enacted the California Wilderness Act, which designated portions of the forest as a protected wilderness area but excluded the proposed route. Id. at 444. While the choice of the route in the Act was made with knowledge of the tribes’ religious interest in it, there was no indication that it was made because of, rather than in disregard of, that interest, and the impact of the choice remained generally applicable and neutral. ^[19] In short, the plan to construct a road and harvest timber in Lyng was generally applicable and “‘neutral’ toward religion” in the sense that its purpose was not to “interfere with religious exercise.” 42 U.S.C. § 2000bb(a)(2). Therefore Lyng, a Free Exercise Clause case that rejected the compelling interest test for neutral laws of general applicability, does not answer the question of whether, under RFRA, preventing a person from engaging in religious exercise by denying them access to a sacred site is a substantial burden.

iii. Terry Williams Is Inapplicable Here There is another, related problem with the majority’s treatment of Lyng. Relying on Williams v. Taylor, 529 U.S. 362, 411 (2000) (“Terry Williams”), the majority erroneously proceeds as if Congress must be understood to have adopted the term “substantial burden” as interpreted in Justice O’Connor’s concurrence in Smith, and so excepted cases similar to Lyng from that concept.

Terry Williams explained that “Congress need not mention a prior decision of this Court by name in a statute’s text in order to adopt either a rule or a meaning given a 19 Moreover, even if the majority were correct as to the impact of the California Wilderness Act, that would be beside the point. Lyng involved a challenge to the Forest Service’s plan to construct the road and harvest timber, not to the California Wilderness Act. See Lyng, 485 U.S. at 448; Collins Op. at 28 (acknowledging that the California Wilderness Act was not enacted until the litigation in Lyng “was pending on appeal in this court”).

certain term in that decision.” 529 U.S. at 411. Where “[t]he separate opinions” in a prior Supreme Court case “concerned the very issue addressed” in a subsequently enacted statute, the prior case can “confirm what [the statutory] language already makes clear.” Id. at 411–12. But the majority opinion’s premises for applying Terry Williams here are flawed.

First, the majority here is wrong that Smith “concerned the very issue” of what constitutes a cognizable substantial burden. The majority opinion asserts that “in superseding Smith, RFRA uses the phrase ‘substantially burden,’ id. § 2000b-1(a), (b),” so “[t]he inference is overwhelming that Congress thereby ‘adopt[ed]’ the ‘meaning given [that] certain term in that decision.’” Collins Op. at 48 (quoting Terry Williams, 529 U.S. at 411). From that premise, the majority concludes that “[w]hen Congress copied the ‘substantial burden’ phrase into RFRA, it must be understood as having similarly adopted the limits that Lyng places on what counts as a governmental imposition of a substantial burden on religious exercise.”

But as Judge Nelson’s concurring opinion appears to acknowledge, neither Lyng nor the Smith majority interpreted the term “substantial burden.” Nelson Op. at 140. Lyng simply refused to apply the compelling interest test. See 485 U.S. at 450–51 (explaining that Sherbert and Yoder “cannot imply that incidental effects of government programs,” without outright prohibition, coercion, or penalty, “require government to bring forward a compelling justification”); see also Smith, 494 U.S. at 883. Thus, Judge Nelson writes that Lyng is not part of any “old soil” that was used to define “substantial burden,” Bea Dissent at 79. Indeed, Lyng does not even use “substantial burden” or any analogous framing of the phrase. Lyng therefore cannot be read as establishing a precise definition of “substantial burden” “carried over into the soil” of RFRA.

Nelson Op. at 141 (citation omitted).

Likewise, Smith was about categorically excepting neutral and generally applicable laws from the compelling interest test, rather than about defining the term “substantial burden.” See 494 U.S. at 884–85; see also supra § II(F)(ii) (discussing Justice O’Connor’s Smith concurrence and explaining that the Smith majority did not apply the compelling interest test). Although Justice O’Connor’s concurring opinion took the position that the denial of unemployment benefits based on religious drug use constituted a substantial burden, she did not rely on Lyng in her discussion of that term. See Smith, 494 U.S. at 897–98 (O’Connor, J., concurring in the judgment). Moreover, the Smith majority never reached the question of what types of burdens would be required to satisfy the first step of the Sherbert test. Instead, it concluded that the test was entirely “inapplicable” in cases challenging neutral, generally applicable laws. See Smith, 494 U.S. at 884–85. So there was no “vigorous debate” in Smith on the meaning of the term substantial burden, contrary to the majority’s representation.

Furthermore, Terry Williams involved a situation in which Congress did “not mention a prior decision of this Court by name in a statute’s text.” 529 U.S. at 411. That is not the circumstance here. Instead, RFRA explicitly identified which portion of Smith Congress sought to address. Congress declared that “in Employment Division v. Smith, the Supreme Court virtually eliminated the requirement that the government justify burdens on religious exercise imposed by laws neutral toward religion.” 42 U.S.C. § 2000bb(a)(4) (citation omitted). Congress’s view, by contrast, was that “laws ‘neutral’ toward religion may burden religious exercise as surely as laws intended to interfere with religious exercise.” Id. § 2000bb(a)(2). Consequently, although the majority opinion points to RFRA’s citation to Smith as reinforcing its holding, the appropriate conclusion is the opposite: Congress was specific about the aspect of Smith that it intended to address—the rule that neutral and generally applicable laws are not subject to the compelling interest test. Congress could not have, by expressly citing Smith in the course of negating its exception for neutral and generally applicable laws, intended to incorporate the “meaning given a certain term,” Terry Williams, 529 U.S. at 411, when that term simply was not at issue in Smith.

The upshot is that RFRA’s text does not support the majority’s conclusion that Congress intended a special exception for certain types of government actions. Rather, RFRA is explicit that:

• Religious exercise includes the use of real property for the purpose of religious exercise. 42 U.S.C. § 2000bb-2(4); Id. § 2000cc-5(7)(B). • Under RFRA, the “[g]overnment shall not substantially burden a person’s exercise of religion” except when the compelling interest test is satisfied. Id. § 2000bb-1(a), (b). No other exceptions are provided.

• Government “includes a branch, department, agency, instrumentality, and official (or other person acting under color of law) of the United States, or of a covered entity.” Id. § 2000bb-2(1). • RFRA “applies to all Federal law, and the implementation of that law, whether statutory or otherwise.” Id. § 2000bb-3(a) (emphasis added) • “Nothing in” RFRA “shall be construed to authorize any government to burden any religious belief.” Id. § 2000bb-3(c). Here, Congress used the term “burden” rather than “substantial burden.” • “[T]he compelling interest test as set forth in prior Federal court rulings is a workable test for striking sensible balances between religious liberty and competing prior governmental interests.” Id. § 2000bb(a)(5).

Given these congressional directives, unlike in Terry Williams, this is not a case in which reference to Smith can “confirm what” RFRA’s statutory “language already makes clear.” Terry Williams, 529 U.S. at 411–12. Rather, for the reasons I have surveyed, what RFRA’s language makes clear is that there is a “substantial burden” when individuals are prevented from practicing their religion by governmental action; if Lyng indicates otherwise (which I do not believe), that implication of Lyng does not survive RFRA. G. This En Banc Panel Fails to Clarify Our Law “As an en banc court, we have a responsibility to bring clarity to our law.” Garfias-Rodriguez v. Holder, 702 F.3d 504, 532 (9th Cir. 2012) (en banc) (Kozinski, C.J., concurring in part). Notably, although the divided three-judge panel rejected Apache Stronghold’s RFRA claim largely under Navajo Nation, the majority makes no mention of that case. Instead, litigants are forced to piece together from a composite of opinions that a majority of judges on this en banc panel rejects Navajo Nation’s reasoning.

Furthermore, the majority opinion creates confusion as to how to define “substantial burden.” Although RFRA’s text simply provides that the federal government may not “substantially burden a person’s exercise of religion,” 42 U.S.C. § 2000bb-1(a), the majority skips the test entirely and asks only whether litigants bring a “cognizable” claim. As I have discussed, see supra § II(E), preventing religious adherents from worshipping at a sacred site is inherently prohibitory. For the majority, only once a litigant has shown that the government action is cognizably “prohibitory” can a court ask whether there is a “substantial burden.” At that point, the majority finds it “adequate[]” to apply a dictionary definition of “substantial burden” in the context of zoning and confinement under both RFRA and RLUIPA, but not in other RFRA contexts. Collins Op. at 52. But this answer is not helpful. Under the majority’s approach, dictionaries can supply the meaning of substantial burden in RFRA cases about zoning and confinement, but dictionaries appear to be irrelevant when a person challenges a different type of government action—as Apache Stronghold does here. Either the meaning of “substantial burden” is the same under RFRA and RLUIPA, or the definition under RFRA is case- dependent. It cannot be both.

And the majority provides no authority for this sort of distinction. Nor could it. If the meaning of “substantial burden” turned on the type of case, several Supreme Court Free Exercise Clause cases would have lacked any discussion of substantial burden or compelling interest. See, e.g., Hernandez, 490 U.S. at 684–85, 699 (discussing substantial burden and concluding the government had a compelling justification in a Free Exercise Clause challenge to the Internal Revenue Service’s refusal to recognize payments made by Scientologists to churches as tax- deductible charitable contributions).

The majority’s shapeshifting definition of substantial burden also finds no support in RFRA’s and RLUIPA’s text. RLUIPA’s land-use provision states that “[n]o government shall impose or implement a land use regulation in a manner that imposes a substantial burden on the religious exercise of a person.” 42 U.S.C. § 2000cc(a)(1) (emphasis added). And the institutionalized persons provision likewise states that “[n]o government shall impose a substantial burden on the religious exercise of a person residing in or confined to an institution.” Id. § 2000cc-1(a) (emphasis added). The majority argues that RLUIPA incorporates or “bake[s] in” the Free Exercise Clause’s “prohibition” requirement. But RLUIPA’s text does not use the word “prohibit,” so it is hard to see how RLUIPA incorporates the Free Exercise Clause in a way that RFRA does not. Compare id., with § 2000bb- 1(a) (“Government shall not substantially burden a person’s exercise of religion.”).

Nor does the majority meaningfully distinguish the coercion inherent in land-use cases from the coercion here. For instance, the majority contends that in the land-use context, the Free Exercise Clause’s “prohibition” requirement is inherent. Collins Op. at 52. But if a city precludes the building of a church on a parcel zoned for single-family dwellings, the city is not conditioning a benefit on forgoing religious exercise nor is it penalizing religious exercise. So how is the city’s zoning law “inherently . . . coercive” in a way that the Land Transfer Act and the destruction of Oak Flat is not? The majority offers little guidance to litigants wondering what governmental actions are sufficiently “coercive” to allow for a substantial burden analysis.

Indeed, contrary to what the majority says, Apache Stronghold’s RFRA claim “inherently involve[s] coercive restrictions.” Collins Op. at 52. As Judge Berzon noted in her panel dissent, Native American sacred sites—like the contexts of land-use and confinement—are unique in that “the government controls access to religious locations and resources.” Apache Stronghold, 38 F.4th at 776 (Berzon, J., dissenting) (citing Stephanie Hall Barclay and Michalyn Steele, Rethinking Protections for Indigenous Sacred Sites, 134 Harv. L. Rev. 1294, 1301 (2021)). In each of these contexts the government has control over religious sites and resources, and religious adherents must “practice their religion in contexts in which voluntary choice is not the baseline.” Id. As with the Western Apaches here, Native American religions are typically land-based, so many traditional Native American religious sites are located exclusively on federal land. Therefore, unlike most non- incarcerated Americans, Native Americans are “at the mercy of government permission to access sacred sites.” Id. (quoting Barclay & Steele, supra, at 1301); see also Douglas Laycock & Thomas C. Berg, Protecting Free Exercise Under Smith and After Smith, Cato Sup. Ct. Rev. at 33, 58 (2020–21) (arguing that the government “took control over the tribes’ ability to practice their traditions fully—in somewhat the same way that prisons control [incarcerated persons’] ability to practice their faith”). The Land Transfer Act thus prevents the Apaches from practicing their religion at Oak Flat, substantially burdening their religious exercise, just as would an outright ban of religious worship, meetings, or diet in prison, or a zoning law precluding a religious group from building a mosque, church, or synagogue. In other words, the government’s control over access to Oak Flat is coercive, and few other religious adherents are situated similarly to the Apache such that they need the government’s permission to worship.

H. RFRA Applies to the Land Transfer Act

For the first time in its Brief in Opposition to Rehearing En Banc, the government urges this court to affirm on the alternative ground that, under the legislative anti-entrenchment principle, RFRA cannot apply to the Land Transfer Act. Because the government did not raise that argument before the district court, and did not develop it on appeal, I would normally consider such eleventh-hour arguments waived. See Partenweederei, MS Belgrano v. Weigel, 313 F.2d 423, 425 (9th Cir. 1962). However, the issue is purely legal, and the government could and likely would raise the argument to the district court on remand. See Janes v. Wal-Mart Stores Inc., 279 F.3d 883, 888 n.4 (9th Cir. 2002). So for the sake of judicial efficiency, I address it now.

RFRA applies to “all Federal” statutes enacted after RFRA’s adoption “unless such [later-enacted] law explicitly excludes such application by reference.” 42 U.S.C. § 2000bb-3(b). The government argues that § 2000bb-3(b) holds no force whatsoever and instead maintains the Land Transfer Act supersedes RFRA because “one legislature cannot abridge the powers of a succeeding legislature.” Fletcher v. Peck, 10 U.S. (6 Cranch) 87, 135 (1810) (Marshall, C.J.). Generally, under the legislative anti-entrenchment doctrine, a prior Congressional enactment “may be repealed, amended, or disregarded by the legislature which enacted it, and is not binding upon any subsequent legislature.” United States v. Winstar Corp., 518 U.S. 839, 873 (1996) (cleaned up).

The Supreme Court has held, however, that “RFRA operates as a kind of super statute” because it applies to all federal statutes and thus “displac[es] the normal operation of other federal laws.” Bostock, 140 S. Ct. at 1754. In two RFRA cases, the Supreme Court accordingly determined that RFRA was controlling even though it conflicted with later-enacted federal law. See Little Sisters of the Poor v. Pennsylvania, 140 S. Ct. 2367, 2383 (2020) (applying RFRA to the Affordable Care Act (“ACA”), a later-enacted statute, because the “ACA does not explicitly exempt RFRA”); Hobby Lobby, 573 U.S. at 719 n.30 (rejecting an implied repeal argument for the same reason). And as the Seventh and Eleventh Circuits have recognized, RFRA is consistent with the anti-entrenchment principle because “the statute does not apply to a subsequently enacted law if it ‘explicitly excludes such application by reference to’” RFRA. Korte, 735 F.3d at 672–73 (cleaned up) (quoting 42 U.S.C. § 2000bb-3(b)); accord Cheffer v. Reno, 55 F.3d 1517, 1522 n.10 (11th Cir. 1995). In other words, because a majority of Congress can preclude the application of RFRA to any subsequently-enacted statute, Congress “remains free to repeal the earlier statute, to exempt the current statute from the earlier statute, to modify the earlier statute, or to apply the earlier statute but as modified.” Dorsey v. United States, 567 U.S. 260, 274 (2012). 20 RFRA does not therefore limit 20 Neither Judge Bea’s concurrence nor the government explain why we should depart from Korte and Cheffer and create a circuit split. See the authority of future Congresses and so does not violate the anti-entrenchment principle. See Little Sisters of the Poor, 140 S. Ct. at 2383 (RFRA “permits Congress to exclude statutes from RFRA’s protections.” (citing 42 U.S.C. § 2000bb-3(b))).

I note that RFRA’s express exemption provision is no different from the one contained in the Administrative Procedure Act (“APA”), which the Supreme Court considered in Marcello v. Bonds, 349 U.S. 302, 310 (1955). The question in Marcello was whether the Immigration and Nationality Act (“INA”) satisfied the APA’s requirement that any exemptions from its procedures be “express[],” such that the APA was inapplicable to deportation proceedings. 349 U.S. at 305–10. The INA section at issue provided that “[t]he procedure (herein prescribed) shall be the sole and exclusive procedure for determining the deportability of an alien under this section.” Marcello, 349 U.S. at 309 (emphasis added) (quotation marks omitted). The Supreme Court explained that this textual provision was a “clear and categorical direction” that the INA “was meant to exclude the application of the” APA. Id.

In other words, the Supreme Court held that the INA did not need to explicitly mention the APA or use a “magical password[]” to supersede the APA’s express repeal provision. Id. at 309–10. The INA’s express inclusion of a “notwithstanding” clause—i.e., “notwithstanding the provisions of any other law”—was sufficient. Id. Consistent with Marcello, we have recognized the inclusion of a “notwithstanding” clause as “a method—akin to an express Kelton Arms Condo. Owners Ass’n, Inc. v. Homestead Ins. Co., 346 F.3d 1190, 1192 (9th Cir. 2003) (“[W]e decline to create a circuit split unless there is a compelling reason to do so.”). reference to the superseded statute—by which Congress can demonstrate that it intended to partially repeal an [earlier] Act.” United States v. Novak, 476 F.3d 1041, 1052 (9th Cir. 2007) (en banc) (cleaned up).

In short, for a statute to exempt itself from RFRA, a simple majority of Congress need only exempt that later-enacted statute from RFRA under 42 U.S.C. § 2000bb-3(b), either by referencing RFRA specifically or by including some variation of a “notwithstanding any other law” provision under Marcello. See Lujan-Armendariz v. I.N.S., 222 F.3d 728, 747 (9th Cir. 2000), overruled on other grounds by Nunez-Reyes v. Holder, 646 F.3d 684 (9th Cir. 2011) (en banc). Such a requirement does not require a “magical password” to supersede RFRA, nor does it violate the legislative anti-entrenchment principle. Marcello, 349 U.S. at 309–10; see Korte, 735 F.3d at 672–73.

Here, the Land Transfer Act cannot escape RFRA’s reach. It neither explicitly exempts itself from RFRA, nor does it contain a “notwithstanding any other law” provision of any kind. See 16 U.S.C. § 539p. At the same time, had Congress wanted to exempt the Land Transfer Act from RFRA, it knew how to do so. The Land Transfer Act includes a specific exemption from another statute—the Federal Land Policy and Management Act of 1976— reinforcing that Congress could have, but did not, enact a similar exemption from RFRA. See 16 U.S.C. § 539p(c)(5)(B)(ii) (“The Secretary may accept a payment in excess of 25 percent of the total value of the land or interests conveyed, notwithstanding section 206(b) of the Federal Land Policy and Management Act of 1976 (43 U.S.C. 1716(b)).” (emphasis added)). If Congress meant to exempt the Land Transfer Act from RFRA, Congress could and would have done so explicitly. Accordingly, RFRA applies to the Land Transfer Act.

III. Conclusion

The majority tragically errs in rejecting Apache Stronghold’s RFRA claim solely under Lyng. Lyng does not answer the question here, where we are faced with government action that will result in a massive hole obliterating Oak Flat and categorically preventing the Western Apaches from ever again communing with Usen and the Ga’an, the very foundation of the Apache religion. The effect will be immediate and irreversible. Under RFRA, preventing religious adherents from engaging in sincere religious exercise undeniably constitutes a “substantial[] burden.” 42 U.S.C. § 2000bb-1(a). RFRA’s plain text encompasses such claims, and the Supreme Court’s and our jurisprudence have long so recognized.

I would therefore hold that, at this stage, Apache Stronghold has shown that it is likely to succeed on the merits of its RFRA claim, and I would remand for the district court to determine whether the Land Transfer Act is justified by a compelling interest pursued through the least restrictive means. 42 U.S.C. § 2000bb-1(b). Because the majority holds the opposite, I respectfully dissent. LEE, Circuit Judge, dissenting:

Chief Judge Murguia’s excellent dissent lays out why Navajo Nation v. United States Forest Service, 535 F.3d 1058 (9th Cir. 2008) (en banc), incorrectly defined “substantial burden” as a narrow term of art. Simply put, the complete obliteration of the land—which the Western Apache consider sacred and where they have worshipped and conducted ceremonies for at least a millennium— obviously imposes a substantial burden on the Apache’s religious exercise.

I join Chief Judge Murguia’s dissent except for Section II.H. I do not believe we should address the merits of the government’s last-minute argument that the Religious Freedom Restoration Act cannot apply to the Land Transfer Act. The government did not bother raising this difficult question before the district court or on appeal. Rather, the government advanced this argument for the first time in its brief opposing rehearing en banc, and now asks the en banc panel to rule in its favor on this newly developed argument. The government infrequently shows any grace when people miss deadlines or do not follow its rules. Cf. Niz-Chavez v. Garland, 141 S. Ct. 1474, 1486 (2021) (“If men must turn square corners when they deal with the government, it cannot be too much to expect the government to turn square corners when it deals with them.”). I would not show any leniency to the government and would consider this argument waived.