2530 120 supreme court reporter 530 u.s. 791 120 supreme court reporter 530 u.s. 791 ... in...

2530 120 SUPREME COURT REPORTER 530 U.S. 791

tions in early term, those who oppose itare remitted to debate the issue in itsmoral dimensions. In a cruel way, theCourt today turns its back on that balance.It in effect tells us the moral debate is notso important after all and can be conduct-ed just as well through a bullhorn from an8–foot distance as it can through a peace-ful, face-to-face exchange of a leaflet. Thelack of care with which the Court sustainsthe Colorado statute reflects a most trou-bling abdication of our responsibility toenforce the First Amendment.

S 792There runs through our First Amend-ment theory a concept of immediacy, theidea that thoughts and pleas and petitionsmust not be lost with the passage of time.In a fleeting existence we have but littletime to find truth through discourse. Nobetter illustration of the immediacy ofspeech, of the urgency of persuasion, ofthe preciousness of time, is presented thanin this case. Here the citizens who claimFirst Amendment protection seek it forspeech which, if it is to be effective, musttake place at the very time and place agrievous moral wrong, in their view, isabout to occur. The Court tears awayfrom the protesters the guarantees of theFirst Amendment when they most need it.So committed is the Court to its coursethat it denies these protesters, in the faceof what they consider to be one of life’sgravest moral crises, even the opportunityto try to offer a fellow citizen a littlepamphlet, a handheld paper seeking toreach a higher law.

I dissent.

,

530 U.S. 793, 147 L.Ed.2d 660

S 793Guy MITCHELL, et al., Petitioners,

v.

Mary L. HELMS et al.No. 98–1648.

Argued Dec. 1, 1999.

Decided June 28, 2000.Rehearing Denied Aug. 28, 2000.See 530 U.S. 1296, 121 S.Ct. 15.

Action was brought challenging con-stitutionality of state and federal school aidprograms as applied to parochial schools inJefferson Parish, Louisiana. The UnitedStates District Court for the Eastern Dis-trict of Louisiana, Frederick J.R. Heebe,J., 856 F.Supp. 1102, found EstablishmentClause violation, but, in granting motionsfor reconsideration, the District Court,Marcel Livaudais, J., 1997 WL 35283, up-held Chapter 2 program. On appeal, theCourt of Appeals for the Fifth Circuit, 151F.3d 347, reversed. Certiorari was grant-ed. The Supreme Court, Justice Thomas,held that Chapter 2 of Title I of the Ele-mentary and Secondary Education Act of1965 (ESEA), under which federal govern-ment distributes funds to state and localgovernmental agencies, which in turn lendeducational materials and equipment topublic and private schools, does not violateEstablishment Clause of the First Amend-ment as applied in Jefferson Parish; over-ruling Meek v. Pittenger, 95 S.Ct. 1753(1975), and Wolman v. Walter, 97 S.Ct.2593 (1977).

Reversed.

Justice Thomas announced the judg-ment of the Court and delivered an opin-ion, in which Chief Justice Rehnquist andJustices Scalia and Kennedy joined.

Justice O’Connor filed opinion concur-ring in the judgment, in which JusticeBreyer joined.

Justice Souter filed dissenting opinion,in which Justices Stevens and Ginsburgjoined.

2531MITCHELL v. HELMSCite as 120 S.Ct. 2530 (2000)

530 U.S. 793

1. Constitutional Law O84.5(4.1) Schools O3, 75

Chapter 2 of Title I of the Elementaryand Secondary Education Act of 1965(ESEA), under which the federal govern-ment distributes funds to state and localgovernmental agencies, which in turn lendeducational materials and equipment topublic and private schools, does not violateEstablishment Clause of the First Amend-ment as applied in Jefferson Parish, Loui-siana. (Per Justice Thomas, with threeJustices concurring and two Justices con-curring in the judgment); overruling Meekv. Pittenger, 95 S.Ct. 1753 (1975), and Wol-man v. Walter, 97 S.Ct. 2593 (1977).U.S.C.A. Const.Amend. 1; Elementaryand Secondary Education of 1965,§§ 6001–6403, as amended, 20 U.S.C.A.§§ 7301–7373.

2. Constitutional Law O84.5(4.1)If aid to schools, even ‘‘direct aid,’’ is

neutrally available and, before reaching orbenefiting any religious school, first passesthrough the hands, literally or figuratively,of numerous private citizens who are freeto direct the aid elsewhere, the govern-ment has not provided any ‘‘support ofreligion’’ for Establishment Clause pur-poses. (Per Justice Thomas, with threeJustices concurring and two Justices con-curring in the judgment). U.S.C.A. Const.Amend. 1.

3. Constitutional Law O84.5(4.1)Establishment Clause does not re-

quire the exclusion of pervasively sectarianschools from otherwise permissible aidprograms. (Per Justice Thomas, with threeJustices concurring and two Justices con-curring in the judgment.) U.S.C.A. Const.Amend. 1.

Syllabus *

Chapter 2 of the Education Consolida-tion and Improvement Act of 1981 chan-

nels federal funds via state educationalagencies (SEA’s) to local educational agen-cies (LEA’s), which in turn lend education-al materials and equipment, such as li-brary and media materials and computersoftware and hardware, to public and pri-vate elementary and secondary schools toimplement ‘‘secular, neutral, and nonideo-logical’’ programs. The enrollment of eachparticipating school determines theamount of Chapter 2 aid that it receives.In an average year, about 30% of Chapter2 funds spent in Jefferson Parish, Louisi-ana, are allocated for private schools, mostof which are Catholic or otherwise reli-giously affiliated. Respondents filed suitalleging, among other things, that Chapter2, as applied in the parish, violated theFirst Amendment’s Establishment Clause.Agreeing, the Chief Judge of the DistrictCourt held, under Lemon v. Kurtzman,403 U.S. 602, 612–613, 91 S.Ct. 2105 thatChapter 2 had the primary effect of ad-vancing religion because the materials andequipment loaned to the Catholic schoolswere direct aid and the schools were per-vasively sectarian. He relied primarily onMeek v. Pittenger, 421 U.S. 349, 95 S.Ct.1753, 44 L.Ed.2d 217, and Wolman v. Wal-ter, 433 U.S. 229, 97 S.Ct. 2593, 53 L.Ed.2d714, in which programs providing many ofthe same sorts of materials and equipmentas does Chapter 2 were struck down, eventhough programs providing for the loan ofpublic school textbooks to religious schoolswere upheld. After the judge issued anorder permanently excluding pervasivelysectarian schools in the parish from receiv-ing any Chapter 2 materials or equipment,he retired. Another judge then reversedthat order, upholding Chapter 2 under,inter alia, Zobrest v. Catalina FoothillsSchool Dist., 509 U.S. 1, 113 S.Ct. 2462,125 L.Ed.2d 1, in which a public schooldistrict was allowed to provide a sign-language interpreter to a deaf student at aCatholic high school as part of a federalprogram for the disabled. While respon-

* The syllabus constitutes no part of the opinionof the Court but has been prepared by theReporter of Decisions for the convenience of

the reader. See United States v. Detroit Tim-ber & Lumber Co., 200 U.S. 321, 337, 26 S.Ct.282, 50 L.Ed. 499.


dents’ appeal was pending, this Court de-cided Agostini v. Felton, 521 U.S. 203, 117S.Ct. 1997, 138 L.Ed.2d 391, approving aprogram under Title I of the Elementaryand Secondary Education Act of 1965 thatprovided public employees to teach reme-dial classes at religious and other privateschools. Concluding that Agostini hadneither directly overruled Meek and Wol-man nor rejected their distinction betweentextbooks and other in-kind aid, the FifthCircuit relied on those two cases to invali-date Chapter 2.

S 794Held: The judgment is reversed.

151 F.3d 347, reversed.

Justice THOMAS, joined by THECHIEF JUSTICE, Justice SCALIA, andJustice Kennedy, concluded that Chapter2, as applied in Jefferson Parish, is not alaw respecting an establishment of religionsimply because many of the private schoolsreceiving Chapter 2 aid in the parish arereligiously affiliated. Pp. 2540–2556.

(a) In modifying the Lemon test—which asked whether a statute (1) has asecular purpose, (2) has a primary effect ofadvancing or inhibiting religion, or (3) cre-ates an excessive entanglement betweengovernment and religion, see 403 U.S., at612–613, 91 S.Ct. 2105—Agostini examinedonly the first and second of those factors,see 521 U.S., at 222–223, 117 S.Ct. 1997,recasting the entanglement inquiry as sim-ply one criterion relevant to determining astatute’s effect, id., at 232–233, 117 S.Ct.1997. The Court also acknowledged thatits cases had pared somewhat the factorsthat could justify a finding of excessiveentanglement. Id., at 233–234, 117 S.Ct.1997. It then set out three primary crite-ria for determining a statute’s effect: Gov-ernment aid has the effect of advancingreligion if it (1) results in governmentalindoctrination, (2) defines its recipients byreference to religion, or (3) creates anexcessive entanglement. Ibid. In thiscase, the inquiry under Agostini’s purposeand effect test is a narrow one. Because

the District Court’s holding that Chapter 2has a secular purpose is not challenged,only Chapter 2’s effect need be considered.Further, in determining that effect, onlythe first two Agostini criteria need beconsidered, because the District Court’sholding that Chapter 2 does not create anexcessive entanglement is not challenged.P. 2540.

(b) Whether governmental aid to reli-gious schools results in religious indoctri-nation ultimately depends on whether anyindoctrination that occurs could reasonablybe attributed to governmental action. See,e.g., Agostini, 521 U.S., at 226, 117 S.Ct.1997. Moreover, the answer to the indoc-trination question will resolve the questionwhether an educational aid program ‘‘sub-sidizes’’ religion. See id., at 230–231, 117S.Ct. 1997. In distinguishing between in-doctrination that is attributable to theState and indoctrination that is not, theCourt has consistently turned to the neu-trality principle, upholding aid that is of-fered to a broad range of groups or per-sons without regard to their religion. As away of assuring neutrality, the Court hasrepeatedly considered whether any gov-ernmental aid to a religious institution re-sults from the genuinely independent andprivate choices of individual parents, e.g.,id., at 226, 117 S.Ct. 1997. Agostini’s sec-ond primary criterion—whether an aidprogram defines its recipients by referenceto religion, id., at 234, 117 S.Ct. 1997—isclosely related to the first. It looks to thesame facts as the neutrality inquiry, seeid., at 225–226, 117 S.Ct. 1997, but usesS 795those facts to answer a somewhat differ-ent question—whether the criteria for allo-cating the aid create a financial incentiveto undertake religious indoctrination, id.,at 231, 117 S.Ct. 1997. Such an incentiveis not present where the aid is allocated onthe basis of neutral, secular criteria thatneither favor nor disfavor religion, and ismade available to both religious and secu-lar beneficiaries on a nondiscriminatorybasis. Ibid. Pp. 2541–2544.


530 U.S. 796

(c) Two rules offered by respondentsto govern the determination whetherChapter 2 has the effect of advancing reli-gion are rejected. Pp. 2544–2549.

(i) Respondents’ chief argument—thatdirect, nonincidental aid to religiousschools is always impermissible—is incon-sistent with this Court’s more recent cases.The purpose of the direct/indirect distinc-tion is to prevent ‘‘subsidization’’ of reli-gion, and the Court’s more recent casesaddress this concern through the principleof private choice, as incorporated in thefirst Agostini criterion (i.e., whether anyindoctrination could be attributed to thegovernment). If aid to schools, even ‘‘di-rect aid,’’ is neutrally available and, beforereaching or benefiting any religious school,first passes through the hands (literally orfiguratively) of numerous private citizenswho are free to direct the aid elsewhere,the government has not provided any‘‘support of religion.’’ Witters v. Washing-ton Dept. of Servs. for Blind, 474 U.S. 481,489, 106 S.Ct. 748, 88 L.Ed.2d 846. Al-though the presence of private choice iseasier to see when aid literally passesthrough individuals’ hands, there is no rea-son why the Establishment Clause re-quires such a form. Indeed, Agostini ex-pressly rejected respondents’ absolute line.521 U.S., at 225, 117 S.Ct. 1997. To theextent respondents intend their direct/indi-rect distinction to require that any aid beliterally placed in schoolchildren’s handsrather than given directly to their schools,Meek and Wolman, the cases on whichthey rely, demonstrate the irrelevance ofsuch formalism. Further, respondents’formalistic line breaks down in the applica-tion to real-world programs. Whether aprogram is labeled ‘‘direct’’ or ‘‘indirect’’ isa rather arbitrary choice that does notfurther the constitutional analysis. SeeBoard of Ed. of Central School Dist. No. 1v. Allen, 392 U.S. 236, 243–245, 88 S.Ct.1923, 20 L.Ed.2d 1060. Although ‘‘specialEstablishment Clause dangers’’ may existwhen money is given directly to religiousschools, see, e.g., Rosenberger v. Rectorand Visitors of Univ. of Va., 515 U.S. 819,

842, 115 S.Ct. 2510, 132 L.Ed.2d 700, suchdirect payments are not at issue here. Pp.2544–2547.

(ii) Respondents’ second argument—that provision to religious schools of aidthat is divertible to religious use is alwaysimpermissible—is also inconsistent withthe Court’s more recent cases, particularlyZobrest, supra, at 18–23, 113 S.Ct. 2462,and Witters, and is also unworkable.Meek and Wolman, on which respondentsappear to rely for their divertibility rule,offer little, if any, support for their rule.The issue is not divertiSbility796 but whetherthe aid itself has an impermissible content.Where the aid would be suitable for use ina public school, it is also suitable for use inany private school. Similarly, the prohibi-tion against the government providing im-permissible content resolves the Establish-ment Clause concerns that exist if aid isactually diverted to religious uses. See,e.g., Agostini, supra, at 224–226, 117 S.Ct.1997. A concern for divertibility, as op-posed to improper content, is also mis-placed because it is boundless—envelopingall aid, no matter how trivial—and thushas only the most attenuated (if any) linkto any realistic concern for preventing anestablishment of religion. Finally, anyaid, with or without content, is ‘‘divertible’’in the sense that it allows schools to ‘‘di-vert’’ resources. Yet the Court has notaccepted the recurrent argument that allaid is forbidden because aid to one aspectof an institution frees it to spend its otherresources on religious ends. E.g., Com-mittee for Public Ed. and Religious Liber-ty v. Regan, 444 U.S. 646, 658, 100 S.Ct.840, 63 L.Ed.2d 94. Pp. 2547–2549.

(d) Additional factors cited by the dis-sent—including the concern for politicaldivisiveness that post-Aguilar v. Felton,473 U.S. 402, 105 S.Ct. 3232, 87 L.Ed.2d290, cases have disregarded, see, e.g.,Agostini, supra, at 233–234, 117 S.Ct.1997, are rejected. In particular, whethera recipient school is pervasively sectarian,a factor that has been disregarded in re-cent cases, e.g., Witters, supra, is not rele-


vant to the constitutionality of a school-aidprogram. Pp. 2550–2552.

(e) Applying the two relevant Agosti-ni criteria reveals that there is no basis forconcluding that Jefferson Parish’s Chapter2 program has the effect of advancingreligion. First, Chapter 2 does not defineits recipients by reference to religion, sinceaid is allocated on the basis of neutral,secular criteria that neither favor nor dis-favor religion, and is made available toboth religious and secular beneficiaries ona nondiscriminatory basis. 521 U.S., at231, 117 S.Ct. 1997. There is no improperincentive because, under the statute, aid isallocated based on school enrollment. Sec-ond, Chapter 2 does not result in govern-mental indoctrination of religion. It deter-mines eligibility for aid neutrally, making abroad array of schools eligible without re-gard to their religious affiliations or lackthereof. See id., at 225–226, 117 S.Ct.1997. It also allocates aid based on theprivate choices of students and their par-ents as to which schools to attend. Seeid., at 222, 117 S.Ct. 1997. Thus, it is notproblematic that Chapter 2 could fairly bedescribed as providing ‘‘direct’’ aid. Final-ly, the Chapter 2 aid provided to religiousschools does not have an impermissiblecontent. The statute explicitly requiresthat such aid be ‘‘secular, neutral, andnonideological,’’ and the record indicatesthat the Louisiana SEA and the JeffersonParish LEA have faithfully enforced thisrequirement insofar as relevant to thiscase. Although there is evidence thatequipment has been, or at least easilycould be, diverted for use in religiousclasses, that evidence is not relevant to theconstitutional analysis. S 797Scattered deminimis statutory violations of the restric-tions on content, discovered and remediedby the relevant authorities themselves be-fore this litigation began almost 15 yearsago, should not be elevated to such a levelas to convert an otherwise unobjectionableparishwide program into a law that has the

effect of advancing religion. Pp. 2552–2555.

(f) To the extent that Meek and Wol-man conflict with the foregoing analysis,they are overruled. Pp. 2555–2556.

Justice O’CONNOR, joined by JusticeBREYER, concluded that Agostini v. Fel-ton, 521 U.S. 203, 117 S.Ct. 1997, 138L.Ed.2d 391, controls the constitutional in-quiry presented here, and requires rever-sal of the Fifth Circuit’s judgment that theChapter 2 program is unconstitutional asapplied in Jefferson Parish. To the extentMeek v. Pittenger, 421 U.S. 349, 95 S.Ct.1753, 44 L.Ed.2d 217, and Wolman v. Wal-ter, 433 U.S. 229, 97 S.Ct. 2593, 53 L.Ed.2d714, are inconsistent with the Court’s judg-ment today, they should be overruled. Pp.2556–2572.

(a) The plurality announces a rule ofunprecedented breadth for the evaluationof Establishment Clause challenges to gov-ernment school-aid programs. That ruleis particularly troubling because, first, itstreatment of neutrality comes close to as-signing that factor singular importance inthe future adjudication of EstablishmentClause challenges to school-aid programs.Although neutrality is important, see, e.g.,Agostini, 521 U.S., at 228, 231–232, 117S.Ct. 1997 the Court has never held that agovernment-aid program passes constitu-tional muster solely because of the neutralcriteria it employs as a basis for distribut-ing aid. Rather, neutrality has heretoforebeen only one of several factors the Courtconsiders. See, e.g., id., at 226–228, 117S.Ct. 1997. Second, the plurality’s approv-al of actual diversion of government aid toreligious indoctrination is in tension withthis Court’s precedents. See, e.g., id., at226–227, 117 S.Ct. 1997. Actual diversionis constitutionally impermissible. E.g.,Bowen v. Kendrick, 487 U.S. 589, 621–622,624, 108 S.Ct. 2562, 101 L.Ed.2d 520. TheCourt should not treat a per-capita-aidprogram like Chapter 2 the same as thetrue private choice programs approved inWitters v. Washington Dept. of Servs. forBlind, 474 U.S. 481, 106 S.Ct. 748, 88L.Ed.2d 846, and Zobrest v. Catalina Foot-


530 U.S. 799

hills School Dist., 509 U.S. 1, 113 S.Ct.2462, 125 L.Ed.2d 1. Because Agostinirepresents the Court’s most recent at-tempt to devise a general framework forapproaching questions concerning neutralschool-aid programs, and involved an Es-tablishment Clause challenge to a school-aid program closely related to the instantprogram, the Agostini criteria should con-trol here. Pp. 2556–2560.

(b) Under Agostini, the Court askswhether the government acted with thepurpose of advancing or inhibiting religionand whether the aid has the ‘‘effect’’ ofdoing so. 521 U.S., at 222–223, 117 S.Ct.1997. The specific criteria used to deter-mine an impermissible effect have changedin recent cases, see id., at 223, 117 S.Ct.1997, which disclose three primary criteriato guide the determination: (1) whetherthe aid results in governmental indoctriSna-tion,798 (2) whether the program defines itsrecipients by reference to religion, and (3)whether the aid creates an excessive en-tanglement between government and reli-gion, id., at 234, 117 S.Ct. 1997. Finally,the same criteria can be reviewed to deter-mine whether a program constitutes en-dorsement of religion. Id., at 235, 117S.Ct. 1997. Respondents neither questionthe Chapter 2 program’s secular purposenor contend that it creates an excessiveentanglement. Accordingly, the Courtneed ask only whether Chapter 2, as ap-plied in Jefferson Parish, results in gov-ernmental indoctrination or defines its re-cipients by reference to religion. It isclear that Chapter 2 does not so define aidrecipients. Rather, it uses wholly neutraland secular criteria to allocate aid to stu-dents enrolled in religious and secularschools alike. As to the indoctrination in-quiry, the Chapter 2 program bears thesame hallmarks of the program upheld inAgostini: Aid is allocated on the basis ofneutral, secular criteria; it is supplementa-ry to, and does not supplant, nonfederalfunds; no Chapter 2 funds reach the cof-fers of religious schools; the aid is secular;evidence of actual diversion is de minimis;

and the program includes adequate safe-guards. Regardless of whether these fac-tors are constitutional requirements, theyare sufficient to find that the program atissue does not have the impermissible ef-fect of advancing religion. For the samereasons, the Chapter 2 program cannotreasonably be viewed as an endorsementof religion. Pp. 2560–2562.

(c) Respondents’ contentions thatAgostini is distinguishable and that Meekand Wolman are controlling here must berejected. Meek and Wolman created aninexplicable rift within the Court’s Estab-lishment Clause jurisprudence. Those de-cisions adhered to the prior holding inBoard of Ed. of Central School Dist. No. 1v. Allen, 392 U.S. 236, 88 S.Ct. 1923, 20L.Ed.2d 1060, that statutes authorizing thelending of textbooks to religious schoolstudents did not violate the EstablishmentClause, see, e.g., Meek, 421 U.S., at 359–362, 95 S.Ct. 1753 (plurality opinion), butinvalidated the lending of instructional ma-terials and equipment to religious schools,e.g., id., at 362–366, 95 S.Ct. 1753, on theground that any assistance in support ofthe pervasively sectarian schools’ edu-cational missions would inevitably have theimpermissible effect of advancing religion,see, e.g., id., at 365–366, 95 S.Ct. 1753.The irrationality of this distinction is pat-ent. See Wallace v. Jaffree, 472 U.S. 38,110, 105 S.Ct. 2479, 86 L.Ed.2d 29. Re-spondents’ assertion that materials andequipment, unlike textbooks, are reason-ably divertible to religious uses is rejectedbecause it does not provide a logical dis-tinction: An educator can use virtually anyinstructional tool, even a textbook, to teacha religious message. Pp. 2562–2567.

(d) The Court should follow the ruleapplied in the context of textbook lendingprograms: To establish a First Amend-ment violation, plaintiffs must prove thatthe aid actually is, or has been, used forreligious S 799purposes. See, e.g., Allen, su-pra, at 248, 88 S.Ct. 1923. Agostini andthe cases on which it relied have under-


mined the assumptions underlying Meekand Wolman. Agostini’s definitive rejec-tion of the presumption that public-schoolemployees teaching in religious schoolswould inevitably inculcate religion alsostood for—or at least strongly pointed to—the broader proposition that such pre-sumptions of religious indoctrination arenormally inappropriate when evaluatingneutral school-aid programs under the Es-tablishment Clause. Respondents’ conten-tions that Agostini should be limited to itsfacts, and that a presumption of religiousinculcation for instructional materials andequipment should be retained, must berejected. The assumption that religious-school instructors can abide by restrictionson the use of government-provided text-books, see Meek, supra, at 384, 95 S.Ct.1753, should extend to instructional mate-rials and equipment. School Dist. ofGrand Rapids v. Ball, 473 U.S. 373, 399–400, 105 S.Ct. 3248 (O’CONNOR, J., con-curring in judgment in part and dissentingin part), distinguished. Pp. 2567–2568.

(e) Respondents’ contention that theactual administration of Chapter 2 in Jef-ferson Parish violated the EstablishmentClause is rejected. The limited evidenceamassed by respondents during 4 years ofdiscovery (which began approximately 15years ago) is at best de minimis andtherefore insufficient to affect the constitu-tional inquiry. Their assertion that thegovernment must have a failsafe mecha-nism capable of detecting any instance ofdiversion was rejected in Agostini, supra,at 234, 117 S.Ct. 1997. Because the pre-sumption adopted in Meek and Wolmanrespecting the use of instructional materi-als and equipment by religious-schoolteachers should be abandoned, there is noconstitutional need for pervasive monitor-ing under the Chapter 2 program. More-over, a review of the specific safeguardsemployed under Chapter 2 at the federal,state, and local levels demonstrates thatthey are constitutionally sufficient. Re-spondents’ evidence does not demonstrateany actual diversion, but, at most, proves

the possibility of diversion in two isolatedinstances. The evidence of violations ofChapter 2’s supplantation and secular-con-tent restrictions is equally insignificantand, therefore, should be treated the same.This Court has never declared an entireaid program unconstitutional on Establish-ment Clause grounds solely because of vio-lations on the minuscule scale of those atissue here. The presence of so few exam-ples tends to show not that the ‘‘no-diver-sion’’ rules have failed, but that they haveworked. Pp. 2568–2572.

THOMAS, J., announced thejudgment of the Court and delivered anopinion, in which REHNQUIST, C.J., andSCALIA and KENNEDY, JJ., joined.O’CONNOR, J., filed an opinionconcurring in the judgment, in whichS 800BREYER, J., joined, post, p. 2556.SOUTER, J., filed a dissenting opinion, inwhich STEVENS and GINSBURG, JJ.,joined, post, p. 2572.

Michael W. McConnell, for petitioners.

Barbara D. Underwood, for UnitedStates in support of petitioners.

Lee Boothby, for Mary L. Helms, et al.

For U.S. Supreme Court briefs, see:1999 WL 639126 (Pet.Brief)1999 WL 792124 (Resp.Brief)1999 WL 1016734 (Reply.Brief)1999 WL 1016735 (Reply.Brief)

S 801Justice THOMAS announced thejudgment of the Court and delivered anopinion, in which THE CHIEF JUSTICE,Justice SCALIA, and Justice KENNEDYjoin.

As part of a longstanding school-aid pro-gram known as Chapter 2, the FederalGovernment distributes funds to state andlocal governmental agencies, which in turnlend educational materials and equipmentto public and private schools, with theenrollment of each participating school de-termining the amount of aid that it re-ceives. The question is whether Chapter


530 U.S. 803

2, as applied in Jefferson Parish, Louisi-ana, is a law respecting an establishmentof religion, because many of the privateschools receiving Chapter 2 aid in thatparish are religiously affiliated. We holdthat Chapter 2 is not such a law.

I

AChapter 2 of the Education Consolida-

tion and Improvement Act of 1981, Pub.L.97–35, 95 Stat. 469, as amended, 20 U.S.C.§§ 7301–7373,1 has its origins in the Ele-mentary and Secondary Education Act of1965 (ESEA), Pub.L. 89–10, 79 Stat. 55,and is a close cousin of the provision of theESEA S 802that we recently considered inAgostini v. Felton, 521 U.S. 203, 117 S.Ct.1997, 138 L.Ed.2d 391 (1997). Like theprovision at issue in Agostini, Chapter 2channels federal funds to local educationalagencies (LEA’s), which are usually publicschool districts, via state educational agen-cies (SEA’s), to implement programs toassist children in elementary and second-ary schools. Among other things, Chapter2 provides aid

‘‘for the acquisition and use of instruc-tional and educational materials, includ-ing library services and materials (in-cluding media materials), assessments,reference materials, computer softwareand hardware for instructional use, andother curricular materials.’’ 20 U.S.C.§ 7351(b)(2).

LEA’s and SEA’s must offer assistanceto both public and private schools (al-though any private school must be non-profit). §§ 7312(a), 7372(a)(1). Participat-ing private schools receive Chapter 2 aidbased on the number of children enrolledin each school, see § 7372(a)(1), and alloca-tions of Chapter 2 funds for those schoolsmust generally be ‘‘equal (consistent withthe number of children to be served) to

expenditures for programs TTT for childrenenrolled in the public schools of the[LEA],’’ § 7372(b). LEA’s must in allcases ‘‘assure equitable participation’’ ofthe children of private schools ‘‘in the pur-poses and benefits’’ of Chapter 2.§ 7372(a)(1); see § 7372(b). Further,Chapter 2 funds may only ‘‘supplementand, to the extent practical, increase thelevel of funds that would TTT be madeavailable from non-Federal sources.’’§ 7371(b). LEA’s and SEA’s may not op-erate their programs ‘‘so as to supplantfunds from non-Federal sources.’’ Ibid.

Several restrictions apply to aid to pri-vate schools. Most significantly, the ‘‘ser-vices, materials, and equipment’’ providedto private schools must be ‘‘secular, neu-tral, and nonideological.’’ § 7372(a)(1).In addition, private schools may not ac-quire control of Chapter 2 funds or title toChapter 2 S 803materials, equipment, orproperty. § 7372(c)(1). A private schoolreceives the materials and equipment list-ed in § 7351(b)(2) by submitting to theLEA an application detailing which itemsthe school seeks and how it will use them;the LEA, if it approves the application,purchases those items from the school’sallocation of funds, and then lends them tothat school.

In Jefferson Parish (the Louisiana gov-ernmental unit at issue in this case), as inLouisiana as a whole, private schools haveprimarily used their allocations for nonre-curring expenses, usually materials andequipment. In the 1986–1987 fiscal year,for example, 44% of the money budgetedfor private schools in Jefferson Parish wasspent by LEA’s for acquiring library andmedia materials, and 48% for instructionalequipment. Among the materials andequipment provided have been librarybooks, computers, and computer software,and also slide and movie projectors, over-head projectors, television sets, tape re-

1. Chapter 2 is now technically Subchapter VIof Chapter 70 of 20 U.S.C., where it wascodified by the Improving America’s SchoolsAct of 1994, Pub.L. 103–382, 108 Stat. 3707.

For convenience, we will use the term ‘‘Chap-ter 2,’’ as the lower courts did. Prior to 1994,Chapter 2 was codified at 20 U.S.C. §§ 2911–2976 (1988 ed.).


corders, VCR’s, projection screens, labora-tory equipment, maps, globes, filmstrips,slides, and cassette recordings.2

It appears that, in an average year,about 30% of Chapter 2 funds spent inJefferson Parish are allocated for privateschools. For the 1985–1986 fiscal year, 41private schools participated in Chapter 2.For the following year, 46 participated,and the participation level has remainedrelatively constant since then. See App.132a. Of these 46, 34 were Roman Catho-lic; 7 were otherwise religiously affiliated;and 5 were not religiously affiliated.

BRespondents filed suit in December

1985, alleging, among other things, thatChapter 2, as applied in Jefferson Parish,S 804violated the Establishment Clause ofthe First Amendment of the Federal Con-stitution. The case’s tortuous history overthe next 15 years indicates well the degreeto which our Establishment Clause juris-prudence has shifted in recent times, whilenevertheless retaining anomalies withwhich the lower courts have had to strug-gle.

In 1990, after extended discovery, ChiefJudge Heebe of the District Court for theEastern District of Louisiana grantedsummary judgment in favor of respon-dents. Helms v. Cody, Civ. A. No. 85–5533, 1990 WL 36124 (Mar. 27), App. toPet. for Cert. 137a. He held that Chapter2 violated the Establishment Clause be-cause, under the second part of our three-part test in Lemon v. Kurtzman, 403 U.S.602, 612–613, 91 S.Ct. 2105, 29 L.Ed.2d 745(1971), the program had the primary effectof advancing religion. Chapter 2 had sucheffect, in his view, because the materialsand equipment loaned to the Catholicschools were direct aid to those schoolsand because the Catholic schools were, heconcluded after detailed inquiry into their

doctrine and curriculum, ‘‘pervasively sec-tarian.’’ App. to Pet. for Cert. 151a.Chief Judge Heebe relied primarily onMeek v. Pittenger, 421 U.S. 349, 95 S.Ct.1753, 44 L.Ed.2d 217 (1975), and Wolmanv. Walter, 433 U.S. 229, 97 S.Ct. 2593, 53L.Ed.2d 714 (1977), in which we held un-constitutional programs that providedmany of the same sorts of materials andequipment as does Chapter 2. In 1994,after having resolved the numerous otherissues in the case, he issued an orderpermanently excluding pervasively sectari-an schools in Jefferson Parish from receiv-ing any Chapter 2 materials or equipment.

Two years later, Chief Judge Heebehaving retired, Judge Livaudais receivedthe case. Ruling in early 1997 on post-judgment motions, he reversed the deci-sion of former Chief Judge Heebe andupheld Chapter 2, pointing to several sig-nificant changes in the legal landscapeover the previous seven years. Helms v.Cody, 1997 WL 35283 (Jan. 28), App. toPet. for Cert. 79a. In particular, JudgeLivaudais cited our 1993 decision in Zo-brest v. Catalina Foothills School Dist.,509 U.S. 1, 113 S.Ct. 2462, 125 L.Ed.2d 1,in which we held that a State could, as partS 805of a federal program for the disabled,provide a sign-language interpreter to adeaf student at a Catholic high school.

Judge Livaudais also relied heavily on a1995 decision of the Court of Appeals forthe Ninth Circuit, Walker v. San Francis-co Unified School Dist., 46 F.3d 1449, up-holding Chapter 2 on facts that he found‘‘virtually indistinguishable.’’ The NinthCircuit acknowledged in Walker, as JudgeHeebe had in his 1990 summary judgmentruling, that Meek and Wolman appearedto erect a constitutional distinction be-tween providing textbooks (permissible)and providing any other in-kind aid (im-permissible). 46 F.3d, at 1464–1465; see

2. Congress in 1988 amended the section gov-erning the sorts of materials and equipmentavailable under Chapter 2. Compare 20U.S.C. § 3832(1)(B) (1982 ed.) with

§ 7351(b)(2) (1994 ed.). The record in thiscase closed in 1989, and the effect of theamendment is not at issue.


530 U.S. 807

Board of Ed. of Central School Dist. No. 1v. Allen, 392 U.S. 236, 88 S.Ct. 1923, 20L.Ed.2d 1060 (1968) (upholding textbookprogram). The Court of Appeals viewedthis distinction, however, as ‘‘thin’’ and‘‘unmoored from any Establishment Clauseprinciples,’’ and, more importantly, as‘‘rendered untenable’’ by subsequent cases,particularly Zobrest. 46 F.3d, at 1465–1466. These cases, in the Ninth Circuit’sview, revived the principle of Allen and ofEverson v. Board of Ed. of Ewing,3 that‘‘state benefits provided to all citizenswithout regard to religion are constitution-al.’’ 46 F.3d, at 1465. The Ninth Circuitalso relied, id., at 1467, on our observationin Board of Ed. of Kiryas Joel VillageSchool Dist. v. Grumet, 512 U.S. 687, 114S.Ct. 2481, 129 L.Ed.2d 546 (1994), that‘‘we have frequently relied explicitly on thegeneral availability of any benefit providedreligious groups or individuals in turningaside Establishment Clause challenges,’’id., at 704, 114 S.Ct. 2481. The NinthCircuit purported to distinguish Meek andWolman based on the percentage ofschools receiving aid that were parochial (alarge percentage in those cases and a mod-erate percentage in Walker), 46 F.3d, at1468, but that court undermined this dis-tinction when it observed that Meek alsoupheld ‘‘the massive provision oftextSbooks806 to parochial schools.’’ 46F.3d, at 1468, n. 16. Thus, although theNinth Circuit did not explicitly hold thatMeek and Wolman were no longer goodlaw, its reasoning seemed to require thatconclusion.

Finally, in addition to relying on ourdecision in Zobrest and the Ninth Circuit’sdecision in Walker, Judge Livaudais in-voked Rosenberger v. Rector and Visitorsof Univ. of Va., 515 U.S. 819, 115 S.Ct.2510, 132 L.Ed.2d 700 (1995), in which, afew months after Walker, we held that theEstablishment Clause does not require apublic university to exclude a student-runreligious publication from assistance avail-

able to numerous other student-run publi-cations.

Following Judge Livaudais’ ruling, re-spondents appealed to the Court of Ap-peals for the Fifth Circuit. While thatappeal was pending, we decided Agostini,in which we approved a program that,under Title I of the ESEA, provided publicemployees to teach remedial classes at pri-vate schools, including religious schools.In so holding, we overruled Aguilar v.Felton, 473 U.S. 402, 105 S.Ct. 3232, 87L.Ed.2d 290 (1985), and partially overruledSchool Dist. of Grand Rapids v. Ball, 473U.S. 373, 105 S.Ct. 3248, 87 L.Ed.2d 267(1985), both of which had involved such aprogram.

The Fifth Circuit thus faced a dilemmabetween, on the one hand, the Ninth Cir-cuit’s holding and analysis in Walker andour subsequent decisions in Rosenbergerand Agostini, and, on the other hand, ourholdings in Meek and Wolman. To re-solve the dilemma, the Fifth Circuit aban-doned any effort to find coherence in ourcase law or to divine the future course ofour decisions and instead focused on ourparticular holdings. Helms v. Picard, 151F.3d 347, 371 (1998). It thought such anapproach required not only by the lack ofcoherence but also by Agostini’s admoni-tion to lower courts to abide by any appli-cable holding of this Court even thoughthat holding might seem inconsistent withour subsequent decisions, see Agostini,521 U.S., at 237, 117 S.Ct. 1997. The FifthCircuit acknowledged that Agostini, byrecognizing our rejection of the rule that‘‘all government aid that directly assiststhe S 807educational function of religiousschools is invalid,’’ id., at 225, 117 S.Ct.1997, had rejected a premise of Meek, butthat court nevertheless concluded thatAgostini had neither directly overruledMeek and Wolman nor rejected their dis-tinction between textbooks and other in-kind aid. The Fifth Circuit therefore con-

3. Everson v. Board of Ed. of Ewing, 330 U.S.1, 67 S.Ct. 504, 91 L.Ed. 711 (1947) (uphold-ing reimbursement to parents for costs of

busing their children to public or privateschool).


cluded that Meek and Wolman controlled,and thus it held Chapter 2 unconstitution-al. We granted certiorari. 527 U.S. 1002,119 S.Ct. 2336, 144 L.Ed.2d 234 (1999).

IIThe Establishment Clause of the First

Amendment dictates that ‘‘Congress shallmake no law respecting an establishmentof religion.’’ In the over 50 years sinceEverson v. Board of Ed. of Ewing, 330U.S. 1, 67 S.Ct. 504, 91 L.Ed. 711 (1947),we have consistently struggled to applythese simple words in the context of gov-ernmental aid to religious schools.4 As weadmitted in Tilton v. Richardson, 403 U.S.672, 91 S.Ct. 2091, 29 L.Ed.2d 790 (1971),‘‘candor compels the acknowledgment thatwe can only dimly perceive the boundariesof permissible government activity in thissensitive area.’’ Id., at 678, 91 S.Ct. 2091(plurality opinion); see Lemon, 403 U.S.,at 671, 91 S.Ct. 2091 (White, J., concurringin judgment).

In Agostini, however, we brought someclarity to our case law, by overruling twoanomalous precedents (one in whole, theother in part) and by consolidating some ofour previously disparate considerations un-der a revised test. Whereas in Lemon wehad considered whether a statute (1) has asecular purpose, (2) has a primary effect ofadvancing or inhibiting religion, or (3) cre-ates an excessive entanglement betweengovernment and religion, see 403 U.S., at612–613, 91 S.Ct. 2105, in Agostini wemodified Lemon for purposes of evaluatingaid to schools and examined only the firstand second factors, see 521 U.S., at 222–223, 117 S.Ct. 1997. We acknowledgedS 808that our cases discussing excessive en-tanglement had applied many of the sameconsiderations as had our cases discussingprimary effect, and we therefore recastLemon’s entanglement inquiry as simplyone criterion relevant to determining astatute’s effect. Agostini, supra, at 232–

233, 117 S.Ct. 1997. We also acknowl-edged that our cases had pared somewhatthe factors that could justify a finding ofexcessive entanglement. 521 U.S., at 233–234, 117 S.Ct. 1997. We then set outrevised criteria for determining the effectof a statute:

‘‘To summarize, New York City’s TitleI program does not run afoul of any ofthree primary criteria we currently useto evaluate whether government aid hasthe effect of advancing religion: It doesnot result in governmental indoctrina-tion; define its recipients by referenceto religion; or create an excessive en-tanglement.’’ Id., at 234, 117 S.Ct. 1997.

[1] In this case, our inquiry underAgostini’s purpose and effect test is anarrow one. Because respondents do notchallenge the District Court’s holding thatChapter 2 has a secular purpose, and be-cause the Fifth Circuit also did not ques-tion that holding, cf. 151 F.3d, at 369, n. 17,we will consider only Chapter 2’s effect.Further, in determining that effect, we willconsider only the first two Agostini crite-ria, since neither respondents nor theFifth Circuit has questioned the DistrictCourt’s holding, App. to Pet. for Cert.108a, that Chapter 2 does not create anexcessive entanglement. ConsideringChapter 2 in light of our more recent caselaw, we conclude that it neither results inreligious indoctrination by the governmentnor defines its recipients by reference toreligion. We therefore hold that Chapter2 is not a ‘‘law respecting an establishmentof religion.’’ In so holding, we acknowl-edge what both the Ninth and Fifth Cir-cuits saw was inescapable—Meek and Wol-man are anomalies in our case law. Wetherefore conclude that they are no longergood law.

4. Cases prior to Everson discussed the issueonly indirectly, see, e.g., Vidal v. Philadelphia,2 How. 127, 198–200, 11 L.Ed. 205 (1844);Quick Bear v. Leupp, 210 U.S. 50, 81, 28 S.Ct.690, 52 L.Ed. 954 (1908), or evaluated aid to

schools under other provisions of the Consti-tution, see Cochran v. Louisiana Bd. of Ed.,281 U.S. 370, 374–375, 50 S.Ct. 335, 74 L.Ed.913 (1930).


530 U.S. 810

S 809AAs we indicated in Agostini, and have

indicated elsewhere, the question whethergovernmental aid to religious schools re-sults in governmental indoctrination is ulti-mately a question whether any religiousindoctrination that occurs in those schoolscould reasonably be attributed to govern-mental action. See Agostini, supra, at226, 117 S.Ct. 1997 (presence of sign-lan-guage interpreter in Catholic school ‘‘ ‘can-not be attributed to state decisionmaking’ ’’(quoting Zobrest, 509 U.S., at 10, 113 S.Ct.2462) (emphasis added in Agostini )); 521U.S., at 230, 117 S.Ct. 1997 (question iswhether ‘‘any use of [governmental] aid toindoctrinate religion could be attributed tothe State’’); see also Rosenberger, 515U.S., at 841–842, 115 S.Ct. 2510; Witters v.Washington Dept. of Servs. for Blind, 474U.S. 481, 488–489, 106 S.Ct. 748, 88L.Ed.2d 846 (1986); Mueller v. Allen, 463U.S. 388, 397, 103 S.Ct. 3062, 77 L.Ed.2d721 (1983); cf. Corporation of PresidingBishop of Church of Jesus Christ of Lat-ter—day Saints v. Amos, 483 U.S. 327,337, 107 S.Ct. 2862, 97 L.Ed.2d 273 (1987)(‘‘For a law to have forbidden ‘effects’under Lemon, it must be fair to say thatthe government itself has advanced reli-gion through its own activities and influ-ence’’). We have also indicated that theanswer to the question of indoctrinationwill resolve the question whether a pro-gram of educational aid ‘‘subsidizes’’ reli-gion, as our religion cases use that term.See Agostini, 521 U.S., at 230–231, 117S.Ct. 1997; see also id., at 230, 117 S.Ct.1997.

In distinguishing between indoctrinationthat is attributable to the State and indoc-trination that is not, we have consistentlyturned to the principle of neutrality, up-holding aid that is offered to a broad rangeof groups or persons without regard totheir religion. If the religious, irreligious,and areligious are all alike eligible forgovernmental aid, no one would concludethat any indoctrination that any particularrecipient conducts has been done at the

behest of the government. For attributionof indoctrination is a relative question. Ifthe government is offering assistance torecipients who provide, so to speak, abroad range of indoctrination, the govern-ment itself is not thought responsible forany parSticular810 indoctrination. To putthe point differently, if the government,seeking to further some legitimate secularpurpose, offers aid on the same terms,without regard to religion, to all who ade-quately further that purpose, see Allen,392 U.S., at 245–247, 88 S.Ct. 1923 (dis-cussing dual secular and religious purposesof religious schools), then it is fair to saythat any aid going to a religious recipientonly has the effect of furthering that secu-lar purpose. The government, in craftingsuch an aid program, has had to concludethat a given level of aid is necessary tofurther that purpose among secular recipi-ents and has provided no more than thatsame level to religious recipients.

As a way of assuring neutrality, we haverepeatedly considered whether any gov-ernmental aid that goes to a religious insti-tution does so ‘‘only as a result of thegenuinely independent and private choicesof individuals.’’ Agostini, supra, at 226,117 S.Ct. 1997 (internal quotation marksomitted). We have viewed as significantwhether the ‘‘private choices of individualparents,’’ as opposed to the ‘‘unmediated’’will of government, Ball, 473 U.S., at 395,n. 13, 105 S.Ct. 3248 (internal quotationmarks omitted), determine what schoolsultimately benefit from the governmentalaid, and how much. For if numerous pri-vate choices, rather than the single choiceof a government, determine the distribu-tion of aid pursuant to neutral eligibilitycriteria, then a government cannot, or atleast cannot easily, grant special favorsthat might lead to a religious establish-ment. Private choice also helps guaranteeneutrality by mitigating the preference forpre-existing recipients that is arguably in-herent in any governmental aid program,see, e.g., Gilder, The Revitalization of Ev-erything: The Law of the Microcosm,


Harv. Bus. Rev. 49 (Mar./Apr.1988), andthat could lead to a program inadvertentlyfavoring one religion or favoring religiousprivate schools in general over nonreli-gious ones.

The principles of neutrality and privatechoice, and their relationship to each oth-er, were prominent not only in AgosStini,811

supra, at 225–226, 228, 230–232, 117 S.Ct.1997, but also in Zobrest, Witters, andMueller.5 The heart of our reasoning inZobrest, upholding governmental provisionof a sign-language interpreter to a deafstudent at his Catholic high school, was asfollows:

‘‘The service at issue in this case is partof a general government program thatdistributes benefits neutrally to anychild qualifying as ‘disabled’ under the[statute], without regard to the ‘sectari-an-nonsectarian, or public-nonpublic na-ture’ of the school the child attends. Byaccording parents freedom to select aschool of their choice, the statute en-sures that a government-paid interpret-er will be present in a sectarian schoolonly as a result of the private decision ofindividual parents. In other words, be-cause the [statute] creates no financialincentive for parents to choose a sectari-an school, an interpreter’s presencethere cannot be attributed to state deci-sionmaking.’’ 509 U.S., at 10, 113 S.Ct.2462.

As this passage indicates, the privatechoices helped to ensure neutrality, andneutrality and private choices togethereliminated any possible attribution to thegovernment even when the interpretertranslated classes on Catholic doctrine.

Witters and Mueller employed similarreasoning. In Witters, we held that theEstablishment Clause did not bar a Statefrom including within a neutral programproviding tuition payments for vocationalrehabilitation a blind person studying at aChristian college to become a pastor, mis-sionary, or youth director. We explained:

‘‘Any aid TTT that ultimately flows toreligious institutions does so only as aresult of the genuinely independent andprivate choices of aid recipients. Wash-ington’s S 812program is made availablegenerally without regard to the sectari-an-nonsectarian, or public-nonpublic na-ture of the institution benefited and TTT

creates no financial incentive for stu-dents to undertake sectarian edu-cationTTTT [T]he fact that aid goes toindividuals means that the decision tosupport religious education is made bythe individual, not by the State.

. . . . .‘‘[I]t does not seem appropriate to viewany aid ultimately flowing to the InlandEmpire School of the Bible as resultingfrom a state action sponsoring or subsi-dizing religion.’’ 474 U.S., at 487–488,106 S.Ct. 748 (footnote, citations, andinternal quotation marks omitted).6

5. Justice O’CONNOR acknowledges that‘‘neutrality is an important reason for uphold-ing government-aid programs,’’ one that ourrecent cases have ‘‘emphasized TTT repeated-ly.’’ Post, at 2557 (opinion concurring injudgment).

6. The majority opinion also noted that only asmall portion of the overall aid under theState’s program would go to religious edu-cation, see Witters, 474 U.S., at 488, 106 S.Ct.748, but it appears that five Members of theCourt thought this point irrelevant. See id.,at 491, n. 3, 106 S.Ct. 748 (Powell, J., joinedby Burger, C. J., and REHNQUIST, J., con-curring) (citing Mueller v. Allen, 463 U.S. 388,401, 103 S.Ct. 3062, 77 L.Ed.2d 721 (1983),

to assert that validity of program ‘‘does notdepend on the fact that petitioner appears tobe the only handicapped student who hassought to use his assistance to pursue reli-gious training’’); 474 U.S., at 490, 106 S.Ct.748 (White, J., concurring) (agreeing with‘‘most of Justice POWELL’s concurring opin-ion with respect to the relevance of Mueller,’’but not specifying further); id., at 493, 106S.Ct. 748 (O’CONNOR, J., concurring in partand concurring in judgment) (agreeing withJustice Powell’s reliance on Mueller and ex-plaining that the program did not have animpermissible effect, because it was neutraland involved private choice, and thus ‘‘[n]oreasonable observer is likely to draw from thefacts before us an inference that the State


530 U.S. 814

Further, five Members of this Court, inseparate opinions, emphasized both the im-portance of neutrality and of privatechoices, and the relationship between thetwo. See id., at S 813490–491, 106 S.Ct. 748(Powell, J., joined by Burger, C.J., andREHNQUIST, J., concurring); id., at 493,106 S.Ct. 748 (O’CONNOR, J., concurringin part and concurring in judgment); seealso id., at 490, 106 S.Ct. 748 (White, J.,concurring).

The tax deduction for educational ex-penses that we upheld in Mueller was, inthese respects, the same as the tuitiongrant in Witters. We upheld it chieflybecause it ‘‘neutrally provides state assis-tance to a broad spectrum of citizens,’’ 463U.S., at 398–399, 103 S.Ct. 3062, and be-cause ‘‘numerous, private choices of indi-vidual parents of school-age children,’’ id.,at 399, 103 S.Ct. 3062, determined whichschools would benefit from the deductions.We explained that ‘‘[w]here, as here, aid toparochial schools is available only as aresult of decisions of individual parents no‘imprimatur of state approval’ can bedeemed to have been conferred on anyparticular religion, or on religion general-ly.’’ Ibid. (citation omitted); see id., at397, 103 S.Ct. 3062 (neutrality indicateslack of state imprimatur).

Agostini ’s second primary criterion fordetermining the effect of governmental aidis closely related to the first. The secondcriterion requires a court to considerwhether an aid program ‘‘define[s] its re-cipients by reference to religion.’’ 521U.S., at 234, 117 S.Ct. 1997. As we brieflyexplained in Agostini, id., at 230–231, 117S.Ct. 1997, this second criterion looks tothe same set of facts as does our focus,under the first criterion, on neutrality, seeid., at 225–226, 117 S.Ct. 1997, but thesecond criterion uses those facts to answer

a somewhat different question—whetherthe criteria for allocating the aid ‘‘creat[e]a financial incentive to undertake religiousindoctrination,’’ id., at 231, 117 S.Ct. 1997.In Agostini we set out the following rulefor answering this question:

‘‘This incentive is not present, however,where the aid is allocated on the basis ofneutral, secular criteria that neither fa-vor nor disfavor religion, and is madeavailable to both religious and secularbeneficiaries on a nondiscriminatory ba-sis. Under such circumstances, the aidis less likely to have the effect of ad-vancing religion.’’ Ibid.

S 814The cases on which Agostini relied forthis rule, and Agostini itself, make clearthe close relationship between this rule,incentives, and private choice. For to saythat a program does not create an incen-tive to choose religious schools is to saythat the private choice is truly ‘‘indepen-dent,’’ Witters, 474 U.S., at 487, 106 S.Ct.748. See Agostini, supra, at 232, 117 S.Ct.1997 (holding that Title I did not createany impermissible incentive, because itsservices were ‘‘available to all children whomeet the Act’s eligibility requirements, nomatter what their religious beliefs orwhere they go to school’’); Zobrest, 509U.S., at 10, 113 S.Ct. 2462 (discussing, insuccessive sentences, neutrality, privatechoice, and financial incentives, respective-ly); Witters, supra, at 488, 106 S.Ct. 748(similar). When such an incentive doesexist, there is a greater risk that one couldattribute to the government any indoctri-nation by the religious schools. See Zo-brest, supra, at 10, 113 S.Ct. 2462.

We hasten to add, what should be obvi-ous from the rule itself, that simply be-cause an aid program offers privateschools, and thus religious schools, a bene-

itself is endorsing a religious practice or be-lief’’). More recently, in Agostini v. Felton,521 U.S. 203, 117 S.Ct. 1997, 138 L.Ed.2d391 (1997), we held that the proportion of aidbenefiting students at religious schools pursu-ant to a neutral program involving privatechoices was irrelevant to the constitutional

inquiry. Id., at 229, 117 S.Ct. 1997 (refusing‘‘to conclude that the constitutionality of anaid program depends on the number of sec-tarian school students who happen to receivethe otherwise neutral aid’’); see also post, at2562 (O’CONNOR, J., concurring in judg-ment) (quoting this passage).


fit that they did not previously receivedoes not mean that the program, by reduc-ing the cost of securing a religious edu-cation, creates, under Agostini ’s secondcriterion, an ‘‘incentive’’ for parents tochoose such an education for their chil-dren. For any aid will have some sucheffect. See Allen, 392 U.S., at 244, 88S.Ct. 1923; Everson, 330 U.S., at 17, 67S.Ct. 504; see also Mueller, 463 U.S., at399, 103 S.Ct. 3062.

BRespondents inexplicably make no effort

to address Chapter 2 under the Agostinitest. Instead, dismissing Agostini as fac-tually distinguishable, they offer two rulesthat they contend should govern our deter-mination of whether Chapter 2 has theeffect of advancing religion. They arguefirst, and chiefly, that ‘‘direct, noninciden-tal’’ aid to the primary educational missionof religious schools is always impermissi-ble. Second, they argue that provision toreligious schools of aid that is divertible toreligious use is similarly impermisSsible.815

7

Respondents’ arguments are inconsistent

with our more recent case law, in particu-lar Agostini and Zobrest, and we thereforereject them.

1[2] Although some of our earlier cases,

particularly Ball, 473 U.S., at 393–394, 105S.Ct. 3248, did emphasize the distinctionbetween direct and indirect aid, the pur-pose of this distinction was S 816merely toprevent ‘‘subsidization’’ of religion, see id.,at 394, 105 S.Ct. 3248. As even the dis-sent all but admits, see post, at 2584 (opin-ion of SOUTER, J.), our more recent casesaddress this purpose not through the di-rect/indirect distinction but rather throughthe principle of private choice, as incorpo-rated in the first Agostini criterion (i.e.,whether any indoctrination could be attrib-uted to the government). If aid to schools,even ‘‘direct aid,’’ is neutrally availableand, before reaching or benefiting any reli-gious school, first passes through thehands (literally or figuratively) of numer-ous private citizens who are free to directthe aid elsewhere, the government has notprovided any ‘‘support of religion,’’ Wit-

7. Respondents also contend that Chapter 2aid supplants, rather than supplements, thecore educational function of parochial schoolsand therefore has the effect of furthering reli-gion. Our case law does provide some indi-cation that this distinction may be relevant todetermining whether aid results in govern-mental indoctrination, see Agostini, 521 U.S.,at 228–229, 117 S.Ct. 1997; Zobrest v. Catali-na Foothills School Dist., 509 U.S. 1, 12, 113S.Ct. 2462, 125 L.Ed.2d 1 (1993); but seeSchool Dist. of Grand Rapids v. Ball, 473 U.S.373, 396, 105 S.Ct. 3216, 87 L.Ed.2d 267(1985), but we have never delineated the dis-tinction’s contours or held that it is constitu-tionally required.

Nor, to the extent that the supplement/sup-plant line is separable from respondents’ di-rect/indirect and ‘‘no divertibility’’ arguments,do we need to resolve the distinction’s consti-tutional status today, for, as we have alreadynoted, Chapter 2 itself requires that aid mayonly be supplemental. 20 U.S.C. § 7371(b).See also post, at 2572 (O’CONNOR, J., con-curring in judgment) (declining to decidewhether supplement/supplant distinction is aconstitutional requirement); but see post, at2564 (explaining that computers are ‘‘neces-

sary’’ to ‘‘the educational process’’). We pre-sume that whether a parish has compliedwith that statutory requirement would be, atthe very least, relevant to whether a violationof any constitutional supplement/supplant re-quirement has occurred, yet we have no rea-son to believe that there has been any materi-al statutory violation. A statewide review bythe Louisiana SEA indicated that § 7371(b)receives nearly universal compliance. App.112a. More importantly, neither the DistrictCourt nor the Fifth Circuit even hinted thatJefferson Parish had violated § 7371(b), andrespondents barely mention the statute intheir brief to this Court, offering only theslimmest evidence of any possible violation,see id., at 63a. Respondents argue that anyChapter 2 aid that a school uses to complywith state requirements (such as those relat-ing to computers and libraries) necessarilyviolates whatever supplement/supplant linemay exist in the Constitution, but our decisionin Committee for Public Ed. and ReligiousLiberty v. Regan, 444 U.S. 646, 100 S.Ct. 840,63 L.Ed.2d 94 (1980), upholding reimburse-ment to parochial schools of costs relating tostate-mandated testing, rejects any such blan-ket rule.


530 U.S. 818

ters, supra, at 489, 106 S.Ct. 748. Seesupra, at 10–11. Although the presence ofprivate choice is easier to see when aidliterally passes through the hands of indi-viduals—which is why we have mentioneddirectness in the same breath with privatechoice, see, e.g., Agostini, 521 U.S., at 226,117 S.Ct. 1997; Witters, supra, at 487, 106S.Ct. 748; Mueller, supra, at 399, 103S.Ct. 3062—there is no reason why theEstablishment Clause requires such aform.

Indeed, Agostini expressly rejected theabsolute line that respondents would haveus draw. We there explained that ‘‘wehave departed from the rule relied on inBall that all government aid that directlyassists the educational function of religiousschools is invalid.’’ 521 U.S., at 225, 117S.Ct. 1997. Agostini relied primarily onWitters for this conclusion and made clearthat private choice and neutrality wouldresolve the concerns formerly addressedby the rule in Ball. It was undeniable inWitters that the aid (tuition) would ulti-mately go to the Inland Empire School ofthe Bible and would support religious edu-cation. We viewed this arrangement, how-ever, as no different from a governmentissuing a paycheck to one of its employeesknowing that the employee would directthe funds to a religious institution. Botharrangements would be valid, for the samereason: ‘‘[A]ny money that ultimately wentto religious institutions did so ‘only as aresult of the genuinely independent andprivate choices of’ individuals.’’S 817Agostini, supra, at 226, 117 S.Ct. 1997(quoting Witters, 474 U.S., at 487, 106S.Ct. 748). In addition, the program inWitters was neutral. 521 U.S., at 225, 117S.Ct. 1997 (quoting Witters, supra, at 487,106 S.Ct. 748).

As Agostini explained, the same reason-ing was at work in Zobrest, where weallowed the government-funded interpret-er to provide assistance at a Catholicschool, ‘‘even though she would be amouthpiece for religious instruction,’’ be-cause the interpreter was provided accord-

ing to neutral eligibility criteria and pri-vate choice. 521 U.S., at 226, 117 S.Ct.1997. Therefore, the religious messagesinterpreted by the interpreter could not beattributed to the government, see ibid.(We saw no difference in Zobrest betweenthe government hiring the interpreter di-rectly and the government providing fundsto the parents who then would hire theinterpreter. 509 U.S., at 13, n. 11, 113S.Ct. 2462.) We rejected the dissent’s ob-jection that we had never before allowed‘‘a public employee to participate directlyin religious indoctrination.’’ See id., at 18,113 S.Ct. 2462 (opinion of Blackmun, J.).Finally, in Agostini itself, we used thereasoning of Witters and Zobrest to con-clude that remedial classes provided underTitle I of the ESEA by public employeesdid not impermissibly finance religious in-doctrination. 521 U.S., at 228, 117 S.Ct.1997; see id., at 230–232, 117 S.Ct. 1997.We found it insignificant that students didnot have to directly apply for Title I ser-vices, that Title I instruction was providedto students in groups rather than individu-ally, and that instruction was provided inthe facilities of the private schools. Id., at226–229, 117 S.Ct. 1997.

To the extent that respondents intendtheir direct/indirect distinction to requirethat any aid be literally placed in thehands of schoolchildren rather than givendirectly to the school for teaching thosesame children, the very cases on whichrespondents most rely, Meek and Wolman,demonstrate the irrelevance of such for-malism. In Meek, we justified our rejec-tion of a program that loaned instructionalmaterials and equipment by, among otherthings, pointing out that the aid wasloaned to the schools, and thus was ‘‘directS 818aid.’’ 421 U.S., at 362–363, 95 S.Ct.1753. The materials-and-equipment pro-gram in Wolman was essentially identical,except that the State, in an effort to com-ply with Meek, see Wolman, 433 U.S., at233, 250, 97 S.Ct. 2593, loaned the aid tothe students. (The revised program oper-ated much like the one we upheld in Allen.


Compare Wolman, supra, at 248, 97 S.Ct.2593, with Allen, 392 U.S., at 243–245, 88S.Ct. 1923.) Yet we dismissed as ‘‘techni-cal’’ the difference between the two pro-grams: ‘‘[I]t would exalt form over sub-stance if this distinction were found tojustify a result different from that inMeek.’’ 433 U.S., at 250, 97 S.Ct. 2593.Wolman thus, although purporting to reaf-firm Meek, actually undermined that deci-sion, as is evident from the similarity be-tween the reasoning of Wolman and thatof the Meek dissent. Compare Wolman,supra, at 250, 97 S.Ct. 2593 (The ‘‘technicalchange in legal bailee’’ was irrelevant),with Meek, supra, at 391, 95 S.Ct. 1753(REHNQUIST, J., concurring in judgmentin part and dissenting in part) (‘‘Nor canthe fact that the school is the bailee beregarded as constitutionally determina-tive’’). That Meek and Wolman reachedthe same result, on programs that wereindistinguishable but for the direct/indirectdistinction, shows that that distinctionplayed no part in Meek.

Further, respondents’ formalistic linebreaks down in the application to real-world programs. In Allen, for example,although we did recognize that studentsthemselves received and owned the text-

books, we also noted that the books pro-vided were those that the private schoolsrequired for courses, that the schools couldcollect students’ requests for books andsubmit them to the board of education,that the schools could store the textbooks,and that the textbooks were essential tothe schools’ teaching of secular subjects.See 392 U.S., at 243–245, 88 S.Ct. 1923.Whether one chooses to label this program‘‘direct’’ or ‘‘indirect’’ is a rather arbitrarychoice, one that does not further the con-stitutional analysis.

Of course, we have seen ‘‘special Estab-lishment Clause dangers,’’ Rosenberger,515 U.S., at 842, 115 S.Ct. 2510, whenmoney is S 819given to religious schools orentities directly rather than, as in Wittersand Mueller, indirectly. See 515 U.S., at842, 115 S.Ct. 2510 (collecting cases); id.,at 846–847, 115 S.Ct. 2510 (O’CONNOR,J., concurring); see also Bowen v. Ken-drick, 487 U.S. 589, 608–609, 108 S.Ct.2562, 101 L.Ed.2d 520 (1988); compareCommittee for Public Ed. and ReligiousLiberty v. Regan, 444 U.S. 646, 100 S.Ct.840, 63 L.Ed.2d 94 (1980), with Levitt v.Committee for Public Ed. & ReligiousLiberty, 413 U.S. 472, 93 S.Ct. 2814, 37L.Ed.2d 736 (1973).8 But S 820direct pay-

8. The reason for such concern is not that theform per se is bad, but that such a formcreates special risks that governmental aidwill have the effect of advancing religion (or,even more, a purpose of doing so). An indi-rect form of payment reduces these risks.See Mueller, 463 U.S., at 399, 103 S.Ct. 3062(neutral tax deduction, because of its indirectform, allowed economic benefit to religiousschools only as result of private choice andthus did not suggest state sanction of schools’religious messages). It is arguable, however,at least after Witters v. Washington Dept. ofServs. for Blind, 474 U.S. 481, 106 S.Ct. 748,88 L.Ed.2d 846 (1986), that the principles ofneutrality and private choice would be ade-quate to address those special risks, for it ishard to see the basis for deciding Wittersdifferently simply if the State had sent thetuition check directly to whichever schoolWitters chose to attend. See Rosenberger v.Rector and Visitors of Univ. of Va., 515 U.S.819, 848, 115 S.Ct. 2510, 132 L.Ed.2d 700(1995) (O’CONNOR, J., concurring) (explain-ing Witters as reconciling principle of neutral-

ity with principle against public funding ofreligious messages by relying on principle ofprivate choice). Similarly, we doubt it wouldbe unconstitutional if, to modify Witters’ hypo-thetical, see 474 U.S., at 486–487, 106 S.Ct.748; supra, at 2544–2545, a government em-ployer directly sent a portion of an employee’spaycheck to a religious institution designatedby that employee pursuant to a neutral chari-table program. We approved a similar ar-rangement in Quick Bear, 210 U.S., at 77–82,28 S.Ct. 690, and the Federal Governmentappears to have long had such a program, see1999 Catalog of Caring: Combined FederalCampaign of the National Capital Area 44, 45,59, 74–75 (listing numerous religious organi-zations, many of which engage in religiouseducation or in proselytizing, to which federalemployees may contribute via payroll deduc-tions); see generally Cornelius v. NAACP Le-gal Defense & Ed. Fund, Inc., 473 U.S. 788,105 S.Ct. 3439, 87 L.Ed.2d 567 (1985) (dis-cussing Combined Federal Campaign). Fi-nally, at least some of our prior cases striking


530 U.S. 821

ments of money are not at issue in thiscase, and we refuse to allow a ‘‘special’’case to create a rule for all cases.

2

Respondents also contend that the Es-tablishment Clause requires that aid toreligious schools not be impermissibly reli-gious in nature or be divertible to religioususe. We agree with the first part of thisargument but not the second. Respon-dents’ ‘‘no divertibility’’ rule is inconsistentwith our more recent case law and is un-workable. So long as the governmentalaid is not itself ‘‘unsuitable for use in thepublic schools because of religious con-tent,’’ Allen, supra, at 245, 88 S.Ct. 1923,and eligibility for aid is determined in aconstitutionally permissible manner, anyuse of that aid to indoctrinate cannot beattributed to the government and is thusnot of constitutional concern. And, ofcourse, the use to which the aid is put doesnot affect the criteria governing the aid’sallocation and thus does not create anyimpermissible incentive under Agostini ’ssecond criterion.

Our recent precedents, particularly Zo-brest, require us to reject respondents’argument. For Zobrest gave no consider-ation to divertibility or even to actual di-version. Had such things mattered to theCourt in Zobrest, we would have foundthe case to be quite easy—for strikingdown rather than, as we did, upholdingthe program—which is just how the dis-sent saw the case. See, e.g., 509 U.S., at18, 113 S.Ct. 2462 (Blackmun, J., dissent-ing) (‘‘Until now, the Court never has au-thorized a public employee to participatedirectly in religious indoctrination’’); id.,at 22, 113 S.Ct. 2462 (‘‘[G]overnmentcrosses the boundary when it furnishesthe medium for communication of a reli-

gious messageTTTT [A] state-employedsign-language interpreter would serve asthe conduit for James’ religious education,thereby assisting Salpointe [High School]in its mission of religious indoctrination’’);id., at 23, 113 S.Ct. 2462 (interpreterS 821‘‘is likely to place the imprimatur ofgovernmental approval upon the favoredreligion’’); see generally id., at 18–23, 113S.Ct. 2462. Quite clearly, then, we didnot, as respondents do, think that the useof governmental aid to further religiousindoctrination was synonymous with reli-gious indoctrination by the government orthat such use of aid created any improperincentives.

Similarly, had we, in Witters, been con-cerned with divertibility or diversion, wewould have unhesitatingly, perhaps sum-marily, struck down the tuition-reimburse-ment program, because it was certain thatWitters sought to participate in it to ac-quire an education in a religious careerfrom a sectarian institution. Diversionwas guaranteed. Mueller took the sameview as Zobrest and Witters, for we did notin Mueller require the State to show thatthe tax deductions were only for the costsof education in secular subjects. We de-clined to impose any such segregation re-quirement for either the tuition-expensedeductions or the deductions for itemsstrikingly similar to those at issue in Meekand Wolman, and here. See Mueller, 463U.S., at 391, n. 2, 103 S.Ct. 3062; see alsoid., at 414, 103 S.Ct. 3062 (Marshall, J.,dissenting) (‘‘The instructional materialswhich are subsidized by the Minnesota taxdeduction plainly may be used to inculcatereligious values and belief’’).

Justice O’CONNOR acknowledges thatthe Court in Zobrest and Witters approvedprograms that involved actual diversion.See post, at 2558 (opinion concurring in

down direct payments involved serious con-cerns about whether the payments were trulyneutral. See, e.g., Committee for Public Ed. &Religious Liberty v. Nyquist, 413 U.S. 756,762–764, 768, 774–780, 93 S.Ct. 2955, 37L.Ed.2d 948 (1973) (striking down, by 8–to–1

vote, program providing direct grants formaintenance and repair of school facilities,where payments were allocated per-pupil butwere only available to private, nonprofitschools in low-income areas, ‘‘ ‘all or prac-


judgment). The dissent likewise does notdeny that Witters involved actual diver-sion. See post, at 2587–2588, n. 16. Thedissent does claim that the aid in Zobrest‘‘was not considered divertible,’’ post, at2587, n. 16, but the dissent in Zobrest,which the author of today’s dissent joined,understood the case otherwise. See su-pra, at 22, 113 S.Ct. 2462. As that dissentmade clear, diversion is the use of govern-ment aid to further a religious message.See Zobrest, supra, at 21–22, 113 S.Ct.2462 (Blackmun, J., dissenting); see alsopost, at 2559, 2567 (O’CONNOR, J., con-curring in judgment). By that definition,the S 822government-provided interpreter inZobrest was not only divertible, but actual-ly diverted.

Respondents appear to rely on Meekand Wolman to establish their rule against‘‘divertible’’ aid. But those cases offer lit-tle, if any, support for respondents. Meekmentioned divertibility only briefly in aconcluding footnote, see 421 U.S., at 366, n.16, 95 S.Ct. 1753, and that mention was, atmost, peripheral to the Court’s reasoningin striking down the lending of instruction-al materials and equipment. The aid pro-gram in Wolman explicitly barred diverti-ble aid, 433 U.S., at 248–249, 97 S.Ct. 2593,so a concern for divertibility could nothave been part of our reason for findingthat program invalid.

The issue is not divertibility of aid butrather whether the aid itself has an imper-missible content. Where the aid would besuitable for use in a public school, it is alsosuitable for use in any private school.Similarly, the prohibition against the gov-ernment providing impermissible contentresolves the Establishment Clause con-cerns that exist if aid is actually divertedto religious uses.9 In Agostini, we ex-plained Zobrest by making just this dis-

tinction between the content of aid and theuse of that aid: ‘‘Because the only govern-ment aid in Zobrest was the interpreter,who was herself not inculcating any reli-gious messages, no government indoctrina-tion took place.’’ 521 U.S., at 224, 117S.Ct. 1997 (second emphasis added).Agostini also acknowledged that what thedissenters in Zobrest had charged was es-sentially true: Zobrest did effect a ‘‘shiftTTT in our Establishment Clause law.’’521 U.S., at 225, 117 S.Ct. 1997. Theinterpreter herself, assuming that sheS 823fulfilled her assigned duties, see id., at224–225, 117 S.Ct. 1997, had ‘‘no inherentreligious significance,’’ Allen, 392 U.S., at244, 88 S.Ct. 1923 (discussing bus rides inEverson ), and so it did not matter (giventhe neutrality and private choice involvedin the program) that she ‘‘would be amouthpiece for religious instruction,’’Agostini, supra, at 226, 117 S.Ct. 1997(discussing Zobrest ). And just as a gov-ernment interpreter does not herself incul-cate a religious message—even when sheis conveying one—so also a governmentcomputer or overhead projector does notitself inculcate a religious message, evenwhen it is conveying one.

In Agostini itself, we approved the pro-vision of public employees to teach secularremedial classes in private schools partlybecause we concluded that there was noreason to suspect that indoctrinating con-tent would be part of such governmentalaid. See 521 U.S., at 223–225, 226–227,234–235, 117 S.Ct. 1997. Relying on Zo-brest, we refused to presume that the pub-lic teachers would ‘‘ ‘inject religious con-tent’ ’’ into their classes, 521 U.S., at 225,117 S.Ct. 1997, especially given certainsafeguards that existed; we also saw no

tically all’ ’’ of which were Catholic). Id., at768, 93 S.Ct. 2955.

9. The dissent would find an establishment ofreligion if a government-provided projectorwere used in a religious school to show aprivately purchased religious film, eventhough a public school that possessed thesame kind of projector would likely be consti-

tutionally barred from refusing to allow astudent bible club to use that projector in aclassroom to show the very same film, wherethe classrooms and projectors were generallyavailable to student groups. See Lamb’sChapel v. Center Moriches Union Free SchoolDist., 508 U.S. 384, 113 S.Ct. 2141, 124L.Ed.2d 352 (1993).


530 U.S. 825

evidence that they had done so, id., at 226–227, 117 S.Ct. 1997.

In Allen we similarly focused on content,emphasizing that the textbooks were pre-approved by public school authorities andwere not ‘‘unsuitable for use in the publicschools because of religious content.’’ 392U.S., at 245, 88 S.Ct. 1923. See Lemon,403 U.S., at 617, 91 S.Ct. 2105 (‘‘We notethat the dissenters in Allen seemed chieflyconcerned with the pragmatic difficultiesinvolved in ensuring the truly secular con-tent of the textbooks’’ (emphasis added)).Although it might appear that a book,because it has a pre-existing content, is notdivertible, and thus that lack of divertibili-ty motivated our holding in Allen, it ishard to imagine any book that could not, ineven moderately skilled hands, serve toillustrate a religious message.10 Post, at2565–2566 (O’CONNOR, J., concurring injudgment) S 824(agreeing with this point).Indeed, the plaintiffs in Walker essentiallyconceded as much. 46 F.3d, at 1469, n. 17.A teacher could, for example, easily useShakespeare’s King Lear, even though setin pagan times, to illustrate the FourthCommandment. See Exodus 20:12 (‘‘Hon-or your father and your mother’’). Thus,it is a nonsequitur for the dissent to con-tend that the textbooks in Allen were ‘‘notreadily divertible to religious teaching pur-poses’’ because they ‘‘had a known andfixed secular content.’’ Post, at 2586.

A concern for divertibility, as opposed toimproper content, is misplaced not onlybecause it fails to explain why the sort ofaid that we have allowed is permissible,but also because it is boundless—envelop-ing all aid, no matter how trivial—and thushas only the most attenuated (if any) link

to any realistic concern for preventing an‘‘establishment of religion.’’ Presumably,for example, government-provided lec-terns, chalk, crayons, pens, paper, and pa-intbrushes would have to be excluded fromreligious schools under respondents’ pro-posed rule. But we fail to see how indoc-trination by means of (i.e., diversion of)such aid could be attributed to the govern-ment. In fact, the risk of improper attri-bution is less when the aid lacks content,for there is no risk (as there is with books)of the government inadvertently providingimproper content. See Allen, supra, at255–262, 88 S.Ct. 1923 (Douglas, J., dis-senting). Finally, any aid, with or withoutcontent, is ‘‘divertible’’ in the sense that itallows schools to ‘‘divert’’ resources. Yetwe have ‘‘ ‘not accepted the recurrent ar-gument that all aid is forbidden becauseaid to one aspect of an institution frees itto spend its other resources on religiousends.’ ’’ Regan, 444 U.S., at 658, 100 S.Ct.840 (quoting Hunt v. McNair, 413 U.S.734, 743, 93 S.Ct. 2868, 37 L.Ed.2d 923(1973)).

S 825It is perhaps conceivable that courtscould take upon themselves the task ofdistinguishing among the myriad kinds ofpossible aid based on the ease of divertingeach kind. But it escapes us how a courtmight coherently draw any such line. Itnot only is far more workable, but also isactually related to real concerns about pre-venting advancement of religion by gov-ernment, simply to require, as did Zobrest,Agostini, and Allen, that a program of aidto schools not provide improper contentand that it determine eligibility and allo-cate the aid on a permissible basis.11

10. Although we did, elsewhere in Board of Ed.of Central School Dist. No. 1 v. Allen, 392 U.S.236, 88 S.Ct. 1923, 20 L.Ed.2d 1060 (1968),observe, in response to a party’s argument,that there was no evidence that the schoolswere using secular textbooks to somehow fur-ther religious instruction, see id., at 248, 88S.Ct. 1923, we had no occasion to say whatthe consequence would be were such use oc-curring and, more importantly, we think thatthis brief concluding comment cannot be

read, especially after Zobrest (not to mentionWitters, Mueller, and Agostini ) as essential tothe reasoning of Allen.

11. Justice O’CONNOR agrees that the Consti-tution does not bar divertible aid. See post,at 2567 (opinion concurring in judgment).She also finds actual diversion unproblematicif ‘‘true private-choice’’ directs the aid. Seepost, at 2559. And even when there is notsuch private choice, she thinks that some


C

The dissent serves up a smorgasbord of11 factors that, depending on the facts ofeach case ‘‘in all its particularity,’’ post, at2577, could be relevant to the constitution-ality of a school aid program. And those11 are a bare minimum. We are reas-sured that there are likely more.12 Seepost, at 2581–2582, 2583. Presumably theywill be revealed in future cases, as needed,but at least one additional factor is evidentfrom the dissent itself: The dissent resur-rects the concern for political divisivenessthat once occupied the Court but that post-Aguilar cases have rightly disregarded.Compare post, at 2572, 2575, 2591, 2595, n.27, with Agostini, 521 U.S., at 233–234, 117S.Ct. 1997; Bowen, 487 U.S., at 617, n. 14,108 S.Ct. 2562; Amos, 483 U.S., at 339–340, n. 17, 107 S.Ct. 2862. As JusticeO’CONNOR explained in dissent in Agui-lar: ‘‘It is curious indeed to base ourinterpretation of the Constitution on spec-ulation as to the likelihood of a phenome-non which the parties may create merelyby S 826prosecuting a lawsuit.’’ 473 U.S., at429, 105 S.Ct. 3232. While the dissentdelights in the perverse chaos that allthese factors produce, post, at 2589–2590;see also post, at 2573, 2581–2582, the Con-stitution becomes unnecessarily clouded,and legislators, litigants, and lower courtsgroan, as the history of this case amplydemonstrates. See Part I–B, supra.

One of the dissent’s factors deservesspecial mention: whether a school thatreceives aid (or whose students receiveaid) is pervasively sectarian. The dissentis correct that there was a period whenthis factor mattered, particularly if thepervasively sectarian school was a primaryor secondary school. Post, at 2581–2583,2586–2587, 2589, 2591–2594. But that pe-

riod is one that the Court should regret,and it is thankfully long past.

There are numerous reasons to formallydispense with this factor. First, its rele-vance in our precedents is in sharp decline.Although our case law has consistentlymentioned it even in recent years, we havenot struck down an aid program in relianceon this factor since 1985, in Aguilar andBall. Agostini of course overruled Aguilarin full and Ball in part, and today JusticeO’CONNOR distances herself from thepart of Ball with which she previouslyagreed, by rejecting the distinction be-tween public and private employees thatwas so prominent in Agostini. Comparepost, at 2567–2568, 2570 (opinion concur-ring in judgment), with Agostini, supra, at223–225, 234–235, 117 S.Ct. 1997. In Wit-ters, a year after Aguilar and Ball, we didnot ask whether the Inland Empire Schoolof the Bible was pervasively sectarian. InBowen, a 1988 decision, we refused to findfacially invalid an aid program (althoughone not involving schools) whose recipientshad, the District Court found, includedpervasively sectarian institutions. See 487U.S., at 636, 647, 648, 108 S.Ct. 2562(Blackmun, J., dissenting). Although weleft it open on remand for the DistrictCourt to reaffirm its prior finding, we tookpains to emphasize the narrowness of the‘‘pervasively sectarian’’ category, see id.,at 620–621, 108 S.Ct. 2562 (opinion of theCourt), and two S 827Members of the majori-ty questioned whether this category was‘‘well-founded,’’ id., at 624, 108 S.Ct. 2562(KENNEDY, J., joined by SCALIA, J.,concurring). Then, in Zobrest and Agosti-ni, we upheld aid programs to childrenwho attended schools that were not onlypervasively sectarian but also were pri-mary and secondary. Zobrest, in turningaway a challenge based on the pervasivelysectarian nature of Salpointe Catholic

amount of actual diversion is tolerable andthat safeguards for preventing and detectingactual diversion may be minimal, as we ex-plain further, infra, at 2553–2555.

12. It is thus surprising for the dissent to ac-cuse us of following a rule of ‘‘breathtakingTTT manipulability.’’ Post, at 2591, n. 19.


530 U.S. 828

High School, emphasized the presence ofprivate choice and the absence of govern-ment-provided sectarian content. 509U.S., at 13, 113 S.Ct. 2462. Agostini, inexplaining why the aid program was con-stitutional, did not bother to mention thatpervasively sectarian schools were at is-sue,13 see 521 U.S., at 226–235, 117 S.Ct.1997, a fact that was not lost on the dis-sent, see id., at 249, 117 S.Ct. 1997 (opin-ion of SOUTER, J.). In disregarding thenature of the school, Zobrest and Agostiniwere merely returning to the approach ofEverson and Allen, in which the Courtupheld aid programs to students at perva-sively sectarian schools. See post, at 2576,2581–2582 (SOUTER, J., dissenting) (not-ing this fact regarding Everson); Allen,392 U.S., at 251–252, 88 S.Ct. 1923 (Black,J., dissenting); id., at 262–264, 269–270, 88S.Ct. 1923, n. (Douglas, J., dissenting).

Second, the religious nature of a recipi-ent should not matter to the constitutionalanalysis, so long as the recipient adequate-ly furthers the government’s secular pur-pose. See supra, at 2541. If a programoffers permissible aid to the religious (in-cluding the pervasively sectarian), the are-ligious, and the irreligious, it is a mysterywhich view of religion the government hasestablished, and thus a mystery what theconstitutional violation would be. The per-vasively sectarian recipient has not re-ceived any special favor, and it is mostbizarre that the Court would, as the dis-sent seemingly does, reserve special hostil-ity for those who take their religion seri-ously, who think that their religion shouldaffect the whole S 828of their lives, or whomake the mistake of being effective intransmitting their views to children.

Third, the inquiry into the recipient’sreligious views required by a focus onwhether a school is pervasively sectarian isnot only unnecessary but also offensive.It is well established, in numerous other

contexts, that courts should refrain fromtrolling through a person’s or institution’sreligious beliefs. See Employment Div.,Dept. of Human Resources of Ore. v.Smith, 494 U.S. 872, 887, 110 S.Ct. 1595,108 L.Ed.2d 876 (1990) (collecting cases).Yet that is just what this factor requires,as was evident before the District Court.Although the dissent welcomes such prob-ing, see post, at 2592–2593, we find itprofoundly troubling. In addition, and re-lated, the application of the ‘‘pervasivelysectarian’’ factor collides with our decisionsthat have prohibited governments fromdiscriminating in the distribution of publicbenefits based upon religious status orsincerity. See Rosenberger v. Rector andVisitors of Univ. of Va., 515 U.S. 819, 115S.Ct. 2510, 132 L.Ed.2d 700 (1995);Lamb’s Chapel v. Center Moriches UnionFree School Dist., 508 U.S. 384, 113 S.Ct.2141, 124 L.Ed.2d 352 (1993); Widmar v.Vincent, 454 U.S. 263, 102 S.Ct. 269, 70L.Ed.2d 440 (1981).

Finally, hostility to aid to pervasivelysectarian schools has a shameful pedigreethat we do not hesitate to disavow. Cf.Chicago v. Morales, 527 U.S. 41, 53–54, n.20, 119 S.Ct. 1849, 144 L.Ed.2d 67 (1999)(plurality opinion). Although the dissentprofesses concern for ‘‘the implied exclu-sion of the less favored,’’ post, at 2572, theexclusion of pervasively sectarian schoolsfrom government-aid programs is justthat, particularly given the history of suchexclusion. Opposition to aid to ‘‘sectarian’’schools acquired prominence in the 1870’swith Congress’ consideration (and nearpassage) of the Blaine Amendment, whichwould have amended the Constitution tobar any aid to sectarian institutions. Con-sideration of the amendment arose at atime of pervasive hostility to the CatholicChurch and to Catholics in general, and itwas an open secret that ‘‘sectarian’’ wascode for ‘‘Catholic.’’ See generally Green,

13. Nor does Justice O’CONNOR do so todayin her analysis of Jefferson Parish’s Chapter 2

program.


The Blaine Amendment Reconsidered, 36Am. J. Legal Hist. 38 (1992).S 829Notwithstanding its history, of course,‘‘sectarian’’ could, on its face, describe theschool of any religious sect, but the Courteliminated this possibility of confusionwhen, in Hunt v. McNair, 413 U.S., at 743,93 S.Ct. 2868, it coined the term ‘‘perva-sively sectarian’’—a term which, at thattime, could be applied almost exclusively toCatholic parochial schools and which eventoday’s dissent exemplifies chiefly by ref-erence to such schools. See post, at 2582,2592–2593 (opinion of SOUTER, J.).

[3] In short, nothing in the Establish-ment Clause requires the exclusion of per-vasively sectarian schools from otherwisepermissible aid programs, and other doc-trines of this Court bar it. This doctrine,born of bigotry, should be buried now.

III

Applying the two relevant Agostini cri-teria, we see no basis for concluding thatJefferson Parish’s Chapter 2 program ‘‘hasthe effect of advancing religion.’’ Agosti-ni, supra, at 234, 117 S.Ct. 1997. Chapter2 does not result in governmental indoctri-nation, because it determines eligibility foraid neutrally, allocates that aid based onthe private choices of the parents ofschoolchildren, and does not provide aidthat has an impermissible content. Nordoes Chapter 2 define its recipients byreference to religion.

Taking the second criterion first, it isclear that Chapter 2 aid ‘‘is allocated onthe basis of neutral, secular criteria thatneither favor nor disfavor religion, and ismade available to both religious and secu-lar beneficiaries on a nondiscriminatorybasis.’’ Agostini, 521 U.S., at 231, 117S.Ct. 1997. Aid is allocated based on en-rollment: ‘‘Private schools receive Chapter2 materials and equipment based on theper capita number of students at eachschool,’’ Walker, 46 F.3d, at 1464, and allo-cations to private schools must ‘‘be equal

(consistent with the number of children tobe served) to expenditures for programsunder this subchapter for children enrolledin the public schools of the [LEA],’’ 20U.S.C. § 7372(b). LEA’s must provideChapter 2 materials and equipment for thebenefit S 830of children in private schools‘‘[t]o the extent consistent with the numberof children in the school district of [anLEA] TTT who are enrolled in private non-profit elementary and secondary schools.’’§ 7372(a)(1). See App. to Pet. for Cert.87a (District Court, recounting testimonyof head of Louisiana’s Chapter 2 programthat LEA’s are told that ‘‘ ‘for every dollaryou spend for the public school student,you spend the same dollar for the non-public school student’ ’’); §§ 7372(a)(1) and(b) (children in private schools must re-ceive ‘‘equitable participation’’). The allo-cation criteria therefore create no improp-er incentive. Chapter 2 does, by statute,deviate from a pure per capita basis forallocating aid to LEA’s, increasing the per-pupil allocation based on the number ofchildren within an LEA who are from poorfamilies, reside in poor areas, or reside inrural areas. §§ 7312(a)-(b). But respon-dents have not contended, nor do we haveany reason to think, that this deviation inthe allocation to the LEA’s leads to devia-tion in the allocation among schools withineach LEA, see §§ 7372(a)-(b), and, even ifit did, we would not presume that such adeviation created any incentive one way orthe other with regard to religion.

Chapter 2 also satisfies the first Agosti-ni criterion. The program makes a broadarray of schools eligible for aid withoutregard to their religious affiliations or lackthereof. § 7372; see § 7353(a)(3). Wetherefore have no difficulty concluding thatChapter 2 is neutral with regard to reli-gion. See Agostini, supra, at 225–226, 117S.Ct. 1997. Chapter 2 aid also, like the aidin Agostini, Zobrest, and Witters, reachesparticipating schools only ‘‘as a conse-quence of private decisionmaking.’’ Agos-tini, supra, at 222, 117 S.Ct. 1997. Privatedecisionmaking controls because of the per


530 U.S. 832

capita allocation scheme, and those deci-sions are independent because of the pro-gram’s neutrality. See 521 U.S., at 226,117 S.Ct. 1997. It is the students andtheir parents—not the government—who,through their choice of school, determinewho receives Chapter 2 funds. The aidfollows the child.

S 831Because Chapter 2 aid is providedpursuant to private choices, it is not pro-blematic that one could fairly describeChapter 2 as providing ‘‘direct’’ aid. Thematerials and equipment provided underChapter 2 are presumably used from timeto time by entire classes rather than byindividual students (although individualstudents are likely the chief consumers oflibrary books and, perhaps, of computersand computer software), and studentsthemselves do not need to apply for Chap-ter 2 aid in order for their schools toreceive it, but, as we explained in Agostini,these traits are not constitutionally signifi-cant or meaningful. See id., at 228–229,117 S.Ct. 1997. Nor, for reasons we havealready explained, is it of constitutionalsignificance that the schools themselves,rather than the students, are the bailees ofthe Chapter 2 aid. The ultimate beneficia-ries of Chapter 2 aid are the students whoattend the schools that receive that aid,and this is so regardless of whether indi-vidual students lug computers to schooleach day or, as Jefferson Parish has moresensibly provided, the schools receive thecomputers. Like the Ninth Circuit, andunlike the dissent, post, at 2583, we ‘‘seelittle difference in loaning science kits tostudents who then bring the kits to schoolas opposed to loaning science kits to theschool directly.’’ Walker, supra, at 1468,

n. 16; see Allen, 392 U.S., at 244, n. 6, 88S.Ct. 1923.

Finally, Chapter 2 satisfies the firstAgostini criterion because it does not pro-vide to religious schools aid that has animpermissible content. The statute explic-itly bars anything of the sort, providingthat all Chapter 2 aid for the benefit ofchildren in private schools shall be ‘‘secu-lar, neutral, and nonideological,’’§ 7372(a)(1), and the record indicates thatthe Louisiana SEA and the Jefferson Par-ish LEA have faithfully enforced this re-quirement insofar as relevant to this case.The chief aid at issue is computers, com-puter software, and library books. Thecomputers presumably have no pre-exist-ing content, or at least none that would beimpermissible for use in public schools.Respondents do not contend S 832otherwise.Respondents also offer no evidence thatreligious schools have received softwarefrom the government that has an imper-missible content.

There is evidence that equipment hasbeen, or at least easily could be, divertedfor use in religious classes. See, e.g., App.108a, 118a, 205a–207a. Justice O’CON-NOR, however, finds the safeguardsagainst diversion adequate to prevent anddetect actual diversion. Post, at 2569,2572 (opinion concurring in judgment).The safeguards on which she relies reduceto three: (1) signed assurances that Chap-ter 2 aid will be used only for secular,neutral, and nonideological purposes, (2)monitoring visits, and (3) the requirementthat equipment be labeled as belonging toChapter 2.14 As to the first, JusticeO’CONNOR rightly places little reliance

14. Many of the other safeguards on whichJustice O’CONNOR relies are safeguardsagainst improper content, not against diver-sion. See post, at 2569–2570, 2570 (opinionconcurring in judgment). Content is a differ-ent matter from diversion and is much easierto police than is the mutable use of materialsand equipment (which is one reason that wefind the safeguards against improper content

adequate, infra, at 2554–2555). Similarly, thestatutory provisions against supplanting non-federal funds and against paying federal fundsfor religious worship or instruction, on whichJustice O’CONNOR also relies, post, at 2569,are of little, if any, relevance to diversion—theformer because diversion need not supplant,and the latter because religious schools re-ceive no funds, 20 U.S.C. § 7372(c)(1).


on it. Post, at 2569–2570. As to the sec-ond, monitoring by SEA and LEA officialsis highly unlikely to prevent or catch diver-sion.15 As to the third, compliance withthe labeling reSquirement833 is haphazard,see App. 113a, and, even if the require-ment were followed, we fail to see how alabel prevents diversion.16 In addition, weagree with the dissent that there is evi-dence of actual diversion and that, werethe safeguards anything other than ane-mic, there would almost certainly be moresuch evidence. See post, at 2591–2592,2593–2596.17 In any event, for reasons we

discussed in Part II–B–2, S 834supra, theevidence of actual diversion and the weak-ness of the safeguards against actual di-version are not relevant to the constitu-tional inquiry, whatever relevance theymay have under the statute and regula-tions.

Respondents do, however, point to somereligious books that the LEA improperlyallowed to be loaned to several religiousschools, and they contend that the moni-toring programs of the SEA and the Jef-ferson Parish LEA are insufficient to pre-vent such errors. The evidence, however,establishes just the opposite, for the im-

15. The SEA director acknowledged as muchwhen he said that the SEA enforces the ruleagainst diversion ‘‘as best we can,’’ only visits‘‘[o]ne or two’’ of the private schools whenev-er it reviews an LEA, and reviews each LEAonly once every three years. App. 94a–95a.When asked whether there was ‘‘any way’’ forSEA officials to know of diversion of a Chap-ter 2 computer, he responded, ‘‘No, there isno way.’’ Id., at 118a.

Monitoring by the Jefferson Parish LEA issimilarly ineffective. The LEA visits each pri-vate school only once a year, for less than anhour and a half, and alerts the school to thevisit in advance. Id., at 142a, 151a–152a,182a–183a. The monitoring visits consist ofreviewing records of equipment use and ofspeaking to a single contact person. Self-reporting is the sole source for the records ofuse. Id., at 140a. In the case of overheadprojectors, the record appears to be just asign-out sheet, and the LEA official simplychecks whether ‘‘the recordation of use isattempted.’’ Id., at 143a. The contact personis not a teacher; monitoring does not includespeaking with teachers; and the LEA makesno effort to inform teachers of the restrictionson use of Chapter 2 equipment. Id., at 154a–155a. The contact person also is usually notinvolved with the computers. Id., at 163a.Thus, the contact person is uninvolved in theactual use of the divertible equipment and,therefore, in no position to know whetherdiversion has occurred. See id., at 154a.Unsurprisingly, then, no contact person hasever reported diversion. Id., at 147a. (InAgostini, by contrast, monitors visited eachclassroom—unannounced—once a month,and the teachers received specific training inwhat activities were permitted. 521 U.S., at211–212, 234, 117 S.Ct. 1997.) The head ofthe Jefferson Parish LEA admitted that shehad, and could have, no idea whether Chapter2 equipment was being diverted:

‘‘Q: Would there be any way to ascertain,from this on-site visit, whether the material orequipment purchased are used not only inaccordance with Chapter 2 plan submitted,but for other purposes, also?‘‘A: No.‘‘Q: Now, would it be your view that achurch-affiliated school that would teach thecreation concept of the origin of man, that ifthey used [a Chapter 2] overhead projector,that would be a violation TTT ?‘‘A: Yes.‘‘Q: Now, is there any way, do you ever askthat question of a church-affiliated school, asto whether they use it for that purpose?‘‘A: No.’’ App. 144a, 150a–151a.See id., at 139a, 145a, 146a–147a (similar).

16. In fact, a label, by associating the govern-ment with any religious use of the equipment,exacerbates any Establishment Clause prob-lem that might exist when diversion occurs.

17. Justice O’CONNOR dismisses as de minim-is the evidence of actual diversion. Post, at2570–2571 (opinion concurring in judgment).That may be, but it is good to realize just whatshe considers de minimis. There is persua-sive evidence that Chapter 2 audiovisualequipment was used in a Catholic school’stheology department. ‘‘[M]uch’’ of the equip-ment at issue ‘‘was purchased with Federalfunds,’’ App. 205a, and those federal fundswere, from the 1982–1983 school year on,almost certainly Chapter 2 funds, see id., at210a; cf. id., at 187a, 189a. The diversionoccurred over seven consecutive school years,id., at 206a–207a, and the use of the equip-ment in the theology department was massivein each of those years, outstripping in everyyear use in other departments such as sci-ence, math, and foreign language, ibid. Inaddition, the dissent has documented likelydiversion of computers. Post, at 2595–2596.


530 U.S. 836

proper lending of library books occurred—and was discovered and remedied—beforethis litigation began almost 15 years ago.18

In other words, the monitoring systemworked. See post, at 2571–2572 (O’CON-NOR, J., concurring in judgment). Fur-ther, the violation by the LEA and theprivate schools was minor and, in the viewof the SEA’s coordinator, inadvertent.See App. 122a. There were approximately191 improper book requests over threeyears (the 1982–1983 through 1984–1985school years); these requests came fromfewer than half of the 40 private schoolsthen participating; and the cost of the 191books S 835amounted to ‘‘less than one per-cent of the total allocation over all thoseyears.’’ Id., at 132a–133a.

The District Court found that prescreen-ing by the LEA coordinator of requestedlibrary books was sufficient to preventstatutory violations, see App. to Pet. forCert. 107a, and the Fifth Circuit did notdisagree. Further, as noted, the monitor-ing system appears adequate to catchthose errors that do occur. We are unwill-ing to elevate scattered de minimis statu-tory violations, discovered and remediedby the relevant authorities themselves pri-or to any litigation, to such a level as toconvert an otherwise unobjectionable par-ishwide program into a law that has theeffect of advancing religion.

IVIn short, Chapter 2 satisfies both the

first and second primary criteria of Agosti-ni. It therefore does not have the effectof advancing religion. For the same rea-

son, Chapter 2 also ‘‘cannot reasonably beviewed as an endorsement of religion,’’Agostini, 521 U.S., at 235, 117 S.Ct. 1997.Accordingly, we hold that Chapter 2 is nota law respecting an establishment of reli-gion. Jefferson Parish need not excludereligious schools from its Chapter 2 pro-gram.19 To the extent that Meek and Wol-man conflict with this holding, we overrulethem.

Our conclusion regarding Meek andWolman should come as no surprise. TheCourt as early as Wolman itself left nodoubt that Meek and Allen were irreconcil-able, see 433 U.S., at 251, n. 18, 97 S.Ct.2593, and we have repeatedly reaffirmedAllen since then, see, e.g., Agostini, supra,at 231, 117 S.Ct. 1997. (In fact, Meek, inS 836discussing the materials-and-equipmentprogram, did not even cite Allen. SeeMeek, 421 U.S., at 363–366, 95 S.Ct. 1753.)Less than three years after Wolman, weexplained that Meek did not, despite ap-pearances, hold that ‘‘all loans of secularinstructional material and equipment ines-capably have the effect of direct advance-ment of religion.’’ Regan, 444 U.S., at661–662, 100 S.Ct. 840 (internal quotationmarks omitted). Then, in Mueller, weconceded that the aid at issue in Meek andWolman did ‘‘resembl[e], in many re-spects,’’ the aid that we had upheld inEverson and Allen. 463 U.S., at 393, andn. 3, 103 S.Ct. 3062; see id., at 402, n. 10,103 S.Ct. 3062; see also id., at 415, 103S.Ct. 3062 (Marshall, J., dissenting) (view-ing Allen as incompatible with Meek andWolman, and the distinction between text-books and other instructional materials as

18. The coordinator of the Jefferson ParishLEA ordered the books recalled sometime inthe summer or early fall of 1985, and it ap-pears that the schools had complied with therecall order by the second week of December1985. App. 162a, 80a–81a. Respondentsfiled suit in early December. This self-correc-tion is a key distinction between this instanceof providing improper content and the evi-dence of actual diversion. See n. 17, supra.

19. Indeed, as petitioners observe, to requireexclusion of religious schools from such a

program would raise serious questions underthe Free Exercise Clause. See, e.g., Church ofLukumi Babalu Aye, Inc. v. Hialeah, 508 U.S.520, 532, 113 S.Ct. 2217, 124 L.Ed.2d 472(1993) (‘‘At a minimum, the protections of theFree Exercise Clause pertain if the law atissue discriminates against some or all reli-gious beliefs’’); Everson, 330 U.S., at 16, 67S.Ct. 504; cf. Rosenberger v. Rector and Visi-tors of Univ. of Va., 515 U.S. 819, 115 S.Ct.2510, 132 L.Ed.2d 700 (holding that FreeSpeech Clause bars exclusion of religiousviewpoints from limited public forum).


‘‘simply untenable’’). Most recently, Agos-tini, in rejecting Ball’s assumption that‘‘all government aid that directly assiststhe educational function of religiousschools is invalid,’’ Agostini, supra, at 225,117 S.Ct. 1997, necessarily rejected a largeportion (perhaps all, see Ball, 473 U.S., at395, 105 S.Ct. 3248) of the reasoning ofMeek and Wolman in invalidating thelending of materials and equipment, forBall borrowed that assumption from thosecases. See 521 U.S., at 220–221, 117 S.Ct.1997 (Shared Time program at issue inBall was ‘‘surely invalid TTT [g]iven theholdings in Meek and Wolman ’’ regardinginstructional materials and equipment).Today we simply acknowledge what haslong been evident and was evident to theNinth and Fifth Circuits and to the Dis-trict Court.

The judgment of the Fifth Circuit isreversed.

It is so ordered.

Justice O’CONNOR, with whom JusticeBREYER joins, concurring in thejudgment.

In 1965, Congress passed the Elemen-tary and Secondary Education Act, 79Stat. 27 (1965 Act). Under Title I, Con-gress provided monetary grants to Statesto address the needs of educationally de-prived children of low-income families.Under Title II, Congress provided furthermonetary S 837grants to States for the ac-quisition of library resources, textbooks,and other instructional materials for useby children and teachers in public andprivate elementary and secondary schools.Since 1965, Congress has reauthorized theTitle I and Title II programs severaltimes. Three Terms ago, we held inAgostini v. Felton, 521 U.S. 203, 117 S.Ct.1997, 138 L.Ed.2d 391 (1997), that Title I,as applied in New York City, did not vio-late the Establishment Clause. I believethat Agostini likewise controls the consti-tutional inquiry respecting Title II pre-sented here, and requires the reversal ofthe Court of Appeals’ judgment that the

program is unconstitutional as applied inJefferson Parish, Louisiana. To the ex-tent our decisions in Meek v. Pittenger,421 U.S. 349, 95 S.Ct. 1753, 44 L.Ed.2d217 (1975), and Wolman v. Walter, 433U.S. 229, 97 S.Ct. 2593, 53 L.Ed.2d 714(1977), are inconsistent with the Court’sjudgment today, I agree that those deci-sions should be overruled. I thereforeconcur in the judgment.

I

I write separately because, in my view,the plurality announces a rule of unprece-dented breadth for the evaluation of Es-tablishment Clause challenges to govern-ment school aid programs. Reduced to itsessentials, the plurality’s rule states thatgovernment aid to religious schools doesnot have the effect of advancing religion solong as the aid is offered on a neutral basisand the aid is secular in content. Theplurality also rejects the distinction be-tween direct and indirect aid, and holdsthat the actual diversion of secular aid by areligious school to the advancement of itsreligious mission is permissible. Althoughthe expansive scope of the plurality’s ruleis troubling, two specific aspects of theopinion compel me to write separately.First, the plurality’s treatment of neutrali-ty comes close to assigning that factorsingular importance in the future adjudica-tion of Establishment Clause challenges togovernment school aid programs. Second,the plurality’s approval of actual diversionof government aid to religious indoctrina-tion is in tension with our S 838precedentsand, in any event, unnecessary to decidethe instant case.

The clearest example of the plurality’snear-absolute position with respect to neu-trality is found in its following statement:

‘‘If the religious, irreligious, and areli-gious are all alike eligible for govern-mental aid, no one would conclude thatany indoctrination that any particularrecipient conducts has been done at thebehest of the government. For attribu-


530 U.S. 840

tion of indoctrination is a relative ques-tion. If the government is offering as-sistance to recipients who provide, so tospeak, a broad range of indoctrination,the government itself is not thought re-sponsible for any particular indoctrina-tion. To put the point differently, if thegovernment, seeking to further some le-gitimate secular purpose, offers aid onthe same terms, without regard to reli-gion, to all who adequately further thatpurpose, then it is fair to say that anyaid going to a religious recipient onlyhas the effect of furthering that secularpurpose.’’ Ante, at 2541 (citation omit-ted).

I agree with Justice SOUTER that theplurality, by taking such a stance, ‘‘appearsto take evenhandedness neutrality and inpractical terms promote it to a single andsufficient test for the establishment consti-tutionality of school aid.’’ Post, at 2590(dissenting opinion).

I do not quarrel with the plurality’srecognition that neutrality is an importantreason for upholding government-aid pro-grams against Establishment Clause chal-lenges. Our cases have described neutrali-ty in precisely this manner, and we haveemphasized a program’s neutrality re-peatedly in our decisions approving variousforms of school aid. See, e.g., Agostini,supra, at 228, 231–232, 117 S.Ct. 1997;Zobrest v. Catalina Foothills School Dist.,509 U.S. 1, 10, 113 S.Ct. 2462, 125 L.Ed.2d1 (1993); Witters v. Washington Dept. ofServs. for Blind, 474 U.S. 481, 487–488,106 S.Ct. 748, 88 L.Ed.2d 846 (1986); id.,at 493, 106 S.Ct. 748 (O’CONNOR, J.,concurring in part and concurring S 839injudgment); Mueller v. Allen, 463 U.S. 388,397–399, 103 S.Ct. 3062, 77 L.Ed.2d 721(1983). Nevertheless, we have never heldthat a government-aid program passesconstitutional muster solely because of theneutral criteria it employs as a basis fordistributing aid. For example, in Agosti-ni, neutrality was only one of several fac-tors we considered in determining thatNew York City’s Title I program did nothave the impermissible effect of advancing

religion. See 521 U.S., at 226–228, 117S.Ct. 1997 (noting lack of evidence of incul-cation of religion by Title I instructors,legal requirement that Title I services besupplemental to regular curricula, and thatno Title I funds reached religious schools’coffers). Indeed, given that the aid inAgostini had secular content and was dis-tributed on the basis of wholly neutralcriteria, our consideration of additional fac-tors demonstrates that the plurality’s ruledoes not accurately describe our recentEstablishment Clause jurisprudence. Seealso Zobrest, supra, at 10, 12–13, 113 S.Ct.2462 (noting that no government fundsreached religious school’s coffers, aid didnot relieve school of expense it otherwisewould have assumed, and aid was not dis-tributed to school but to the child).

Justice SOUTER provides a compre-hensive review of our EstablishmentClause cases on government aid to reli-gious institutions that is useful for its ex-planation of the various ways in which wehave used the term ‘‘neutrality’’ in ourdecisions. See post, at 2578–2581. Evenif we at one time used the term ‘‘neutrali-ty’’ in a descriptive sense to refer to thoseaid programs characterized by the requi-site equipoise between support of religionand antagonism to religion, Justice SOUT-ER’s discussion convincingly demonstratesthat the evolution in the meaning of theterm in our jurisprudence is cause to hesi-tate before equating the neutrality of re-cent decisions with the neutrality of old.As I have previously explained, neutralityis important, but it is by no means theonly ‘‘axiom in the history and precedentof the Establishment Clause.’’ Rosenber-ger v. Rector and Visitors of Univ. of Va.,515 U.S. 819, 846, 115 S.Ct. 2510, 132L.Ed.2d 700 (1995) (concurring opinion).Thus, S 840I agree with Justice SOUTER’sconclusion that our ‘‘most recent use of‘neutrality’ to refer to generality or even-handedness of distribution TTT is relevantin judging whether a benefit scheme socharacterized should be seen as aiding asectarian school’s religious mission, but


this neutrality is not alone sufficient toqualify the aid as constitutional.’’ Post, at2581.

I also disagree with the plurality’s con-clusion that actual diversion of governmentaid to religious indoctrination is consistentwith the Establishment Clause. See ante,at 2547–2549. Although ‘‘[o]ur cases havepermitted some government funding ofsecular functions performed by sectarianorganizations,’’ our decisions ‘‘provide noprecedent for the use of public funds tofinance religious activities.’’ Rosenberger,supra, at 847, 115 S.Ct. 2510 (O’CONNOR,J., concurring). At least two of the deci-sions at the heart of today’s case demon-strate that we have long been concernedthat secular government aid not be divert-ed to the advancement of religion. In bothAgostini, our most recent school aid case,and Board of Ed. of Central School Dist.No. 1 v. Allen, 392 U.S. 236, 88 S.Ct. 1923,20 L.Ed.2d 1060 (1968), we rested ourapproval of the relevant programs in parton the fact that the aid had not been usedto advance the religious missions of therecipient schools. See Agostini, supra, at226–227, 117 S.Ct. 1997 (‘‘[N]o evidencehas ever shown that any New York CityTitle I instructor teaching on parochialschool premises attempted to inculcate re-ligion in students’’); Allen, supra, at 248,88 S.Ct. 1923 (‘‘Nothing in this record sup-ports the proposition that all textbooks,whether they deal with mathematics, phys-ics, foreign languages, history, or litera-ture, are used by the parochial schools toteach religion’’). Of course, our focus onthe lack of such evidence would have beenentirely unnecessary if we had believedthat the Establishment Clause permits theactual diversion of secular government aidto religious indoctrination. Our decision inBowen v. Kendrick, 487 U.S. 589, 108 S.Ct.2562, 101 L.Ed.2d 520 (1988), also demon-strates that actual diversion is constitu-tionally impermissible. After concludSing841

that the government-aid program in ques-tion was constitutional on its face, we re-manded the case so that the District Courtcould determine, after further factual de-

velopment, whether aid recipients hadused the government aid to support theirreligious objectives. See id., at 621–622,108 S.Ct. 2562; id., at 624, 108 S.Ct. 2562(KENNEDY, J., concurring) (‘‘[T]he onlypurpose of further inquiring whether anyparticular grantee institution is pervasivelysectarian is as a preliminary step to dem-onstrating that the funds are in fact beingused to further religion’’). The remandwould have been unnecessary if, as theplurality contends, actual diversion wereirrelevant under the EstablishmentClause.

The plurality bases its holding that actu-al diversion is permissible on Witters andZobrest. Ante, at 2547. Those decisions,however, rested on a significant factualpremise missing from this case, as well asfrom the majority of cases thus far consid-ered by the Court involving EstablishmentClause challenges to school aid programs.Specifically, we decided Witters and Zo-brest on the understanding that the aidwas provided directly to the individual stu-dent who, in turn, made the choice ofwhere to put that aid to use. See Witters,474 U.S., at 488, 106 S.Ct. 748; Zobrest,509 U.S., at 10, 12, 113 S.Ct. 2462. Ac-cordingly, our approval of the aid in bothcases relied to a significant extent on thefact that ‘‘[a]ny aid TTT that ultimatelyflows to religious institutions does so onlyas a result of the genuinely independentand private choices of aid recipients.’’Witters, supra, at 487, 106 S.Ct. 748; seeZobrest, supra, at 10, 113 S.Ct. 2462 (‘‘[A]government-paid interpreter will be pres-ent in a sectarian school only as a result ofthe private decision of individual parents’’).This characteristic of both programs madethem less like a direct subsidy, whichwould be impermissible under the Estab-lishment Clause, and more akin to thegovernment issuing a paycheck to an em-ployee who, in turn, donates a portion ofthat check to a religious institution. See,e.g., Witters, supra, at 486–487, 106 S.Ct.748; see also Rosenberger, supra, at 848,


530 U.S. 843

115 S.Ct. 2510 (O’CONNOR, J., concur-ring) (discussing Witters ).

S 842Recognizing this distinction, the plu-rality nevertheless finds Witters and Zo-brest—to the extent those decisions mightpermit the use of government aid for reli-gious purposes—relevant in any case in-volving a neutral, per-capita-aid program.See ante, at 2552–2553. Like JusticeSOUTER, I do not believe that we shouldtreat a per-capita-aid program the same asthe true private-choice programs consid-ered in Witters and Zobrest. See post, at2591. First, when the government pro-vides aid directly to the student beneficia-ry, that student can attend a religiousschool and yet retain control over whetherthe secular government aid will be appliedtoward the religious education. The factthat aid flows to the religious school and isused for the advancement of religion istherefore wholly dependent on the stu-dent’s private decision. See Rosenberger,515 U.S., at 848, 115 S.Ct. 2510 (O’CON-NOR, J., concurring) (discussing impor-tance of private choice in Witters ); Wit-ters, 474 U.S., at 488, 106 S.Ct. 748 (‘‘[T]hefact that aid goes to individuals means thatthe decision to support religious educationis made by the individual, not by theState’’); id., at 493, 106 S.Ct. 748 (O’CON-NOR, J., concurring in part and concur-ring in judgment) (‘‘The aid to religion atissue here is the result of petitioner’s pri-vate choice’’). It is for this reason that inAgostini we relied on Witters and Zobrestto reject the rule ‘‘that all government aidthat directly assists the educational func-tion of religious schools is invalid,’’ 521U.S., at 225, 117 S.Ct. 1997, yet also restedour approval of New York City’s Title Iprogram in part on the lack of evidence ofactual diversion, id., at 226–227, 117 S.Ct.1997.

Second, I believe the distinction betweena per capita school aid program and a trueprivate-choice program is significant forpurposes of endorsement. See, e.g., Lynchv. Donnelly, 465 U.S. 668, 692, 104 S.Ct.1355, 79 L.Ed.2d 604 (1984) (O’CONNOR,

J., concurring). In terms of public percep-tion, a government program of direct aidto religious schools based on the numberof students attending each school differsmeaningfully from the government distrib-uting aid directly to individual studentsS 843who, in turn, decide to use the aid at thesame religious schools. In the former ex-ample, if the religious school uses the aidto inculcate religion in its students, it isreasonable to say that the government hascommunicated a message of endorsement.Because the religious indoctrination is sup-ported by government assistance, the rea-sonable observer would naturally perceivethe aid program as government supportfor the advancement of religion. That theamount of aid received by the school isbased on the school’s enrollment does notseparate the government from the en-dorsement of the religious message. Theaid formula does not—and could not—indi-cate to a reasonable observer that theinculcation of religion is endorsed only bythe individuals attending the religiousschool, who each affirmatively choose todirect the secular government aid to theschool and its religious mission. No suchchoices have been made. In contrast,when government aid supports a school’sreligious mission only because of indepen-dent decisions made by numerous individu-als to guide their secular aid to that school,‘‘[n]o reasonable observer is likely to drawfrom the facts TTT an inference that theState itself is endorsing a religious prac-tice or belief.’’ Witters, supra, at 493, 106S.Ct. 748 (O’CONNOR, J., concurring inpart and concurring in judgment). Rath-er, endorsement of the religious messageis reasonably attributed to the individualswho select the path of the aid.

Finally, the distinction between a per-capita-aid program and a true private-choice program is important when consid-ering aid that consists of direct monetarysubsidies. This Court has ‘‘recognizedspecial Establishment Clause dangerswhere the government makes direct mon-ey payments to sectarian institutions.’’


Rosenberger, 515 U.S., at 842, 115 S.Ct.2510; see also ibid. (collecting cases). If,as the plurality contends, a per-capita-aidprogram is identical in relevant constitu-tional respects to a true private-choice pro-gram, then there is no reason that, underthe plurality’s reasoning, the governmentshould be precluded from providing directmoney payments S 844to religious organiza-tions (including churches) based on thenumber of persons belonging to each orga-nization. And, because actual diversion ispermissible under the plurality’s holding,the participating religious organizations(including churches) could use that aid tosupport religious indoctrination. To besure, the plurality does not actually holdthat its theory extends to direct moneypayments. See ante, at 2546–2547. Thatomission, however, is of little comfort. Inits logic—as well as its specific advisorylanguage, see ante, at 2546–2547, n. 8—theplurality opinion foreshadows the approvalof direct monetary subsidies to religiousorganizations, even when they use themoney to advance their religious objec-tives.

Our school aid cases often pose difficultquestions at the intersection of the neu-trality and no-aid principles and thereforedefy simple categorization under eitherrule. As I explained in Rosenberger,‘‘[r]esolution instead depends on the hardtask of judging—sifting through the de-tails and determining whether the chal-lenged program offends the EstablishmentClause. Such judgment requires courts todraw lines, sometimes quite fine, based onthe particular facts of each case.’’ 515U.S., at 847, 115 S.Ct. 2510 (concurringopinion). Agostini represents our mostrecent attempt to devise a general frame-work for approaching questions concerningneutral school aid programs. Agostinialso concerned an Establishment Clausechallenge to a school aid program closelyrelated to the one at issue here. For thesereasons, as well as my disagreement withthe plurality’s approach, I would decide

today’s case by applying the criteria setforth in Agostini.

IIIn Agostini, after reexamining our juris-

prudence since School Dist. of Grand Rap-ids v. Ball, 473 U.S. 373, 105 S.Ct. 3216, 87L.Ed.2d 267 (1985), we explained that thegeneral principles used to determinewhether government aid violates the Es-tablishment Clause have remained largelyunchanged. 521 U.S., at 222, 117 S.Ct.1997. Thus, S 845we still ask ‘‘whether thegovernment acted with the purpose of ad-vancing or inhibiting religion’’ and ‘‘wheth-er the aid has the ‘effect’ of advancing orinhibiting religion.’’ Id., at 222–223, 117S.Ct. 1997. We also concluded in Agosti-ni, however, that the specific criteria usedto determine whether government aid hasan impermissible effect had changed. Id.,at 223, 117 S.Ct. 1997. Looking to ourrecently decided cases, we articulatedthree primary criteria to guide the deter-mination whether a government-aid pro-gram impermissibly advances religion: (1)whether the aid results in governmentalindoctrination, (2) whether the aid pro-gram defines its recipients by reference toreligion, and (3) whether the aid creates anexcessive entanglement between govern-ment and religion. Id., at 234, 117 S.Ct.1997. Finally, we noted that the samecriteria could be reviewed to determinewhether a government-aid program consti-tutes an endorsement of religion. Id., at235, 117 S.Ct. 1997.

Respondents neither question the secu-lar purpose of the Chapter 2 (Title II)program nor contend that it creates anexcessive entanglement. (Due to its de-nomination as Chapter 2 of the EducationConsolidation and Improvement Act of1981, 95 Stat. 469, the parties refer to the1965 Act’s Title II program, as modified bysubsequent legislation, as ‘‘Chapter 2.’’ Forease of reference, I will do the same.)Accordingly, for purposes of decidingwhether Chapter 2, as applied in JeffersonParish, Louisiana, violates the Establish-


530 U.S. 847

ment Clause, we need ask only whetherthe program results in governmental in-doctrination or defines its recipients byreference to religion.

Taking the second inquiry first, it isclear that Chapter 2 does not define aidrecipients by reference to religion. InAgostini, we explained that scrutiny of themanner in which a government-aid pro-gram identifies its recipients is importantbecause ‘‘the criteria might themselveshave the effect of advancing religion bycreating a financial incentive to undertakereligious indoctrination.’’ 521 U.S., at 231,117 S.Ct. 1997. We then clarified that thisfinancial incentive is not present S 846‘‘wherethe aid is allocated on the basis of neutral,secular criteria that neither favor nor dis-favor religion, and is made available toboth religious and secular beneficiaries ona nondiscriminatory basis.’’ Ibid. UnderChapter 2, the Secretary of Education allo-cates funds to the States based on eachState’s share of the Nation’s school-agepopulation. 20 U.S.C. § 7311(b). Thestate educational agency (SEA) of eachrecipient State, in turn, must distribute theState’s Chapter 2 funds to local education-al agencies (LEA’s) ‘‘according to the rela-tive enrollments in public and private, non-profit schools within the school districts ofsuch agencies,’’ adjusted to take into ac-count those LEA’s ‘‘which have the great-est numbers or percentages of childrenwhose education imposes a higher thanaverage cost per child.’’ § 7312(a). TheLEA must then expend those funds on‘‘innovative assistance programs’’ designedto improve student achievement.§ 7351(b). The statute generally requiresthat an LEA ensure the ‘‘equitable partic-ipation’’ of children enrolled in private non-profit elementary and secondary schools,§ 7372(a)(1), and specifically mandatesthat all LEA expenditures on behalf ofchildren enrolled in private schools ‘‘beequal (consistent with the number of chil-dren to be served) to expenditures forprograms TTT for children enrolled in thepublic schools of the [LEA],’’ § 7372(b).

As these statutory provisions make clear,Chapter 2 uses wholly neutral and secularcriteria to allocate aid to students enrolledin religious and secular schools alike. As aresult, it creates no financial incentive toundertake religious indoctrination.

Agostini next requires us to ask wheth-er Chapter 2 ‘‘result[s] in governmentalindoctrination.’’ 521 U.S., at 234, 117 S.Ct.1997. Because this is a more complexinquiry under our case law, it is useful firstto review briefly the basis for our decisionin Agostini that New York City’s Title Iprogram did not result in governmentalindoctrination. Under that program, pub-lic-school teachers provided Title I instruc-tion to eligible S 847students on privateschool premises during regular schoolhours. Twelve years earlier, in Aguilar v.Felton, 473 U.S. 402, 105 S.Ct. 3232, 87L.Ed.2d 290 (1985), we had held the sameNew York City program unconstitutional.In Ball, a companion case to Aguilar, wealso held that a similar program in GrandRapids, Michigan, violated the Constitu-tion. Our decisions in Aguilar and Ballwere both based on a presumption, drawnin large part from Meek, see 421 U.S., at367–373, 95 S.Ct. 1753, that public-schoolinstructors who teach secular classes onthe campuses of religious schools will inev-itably inculcate religion in their students.

In Agostini, we recognized that ‘‘[o]urmore recent cases [had] undermined theassumptions upon which Ball and Aguilarrelied.’’ 521 U.S., at 222, 117 S.Ct. 1997.First, we explained that the Court hadsince abandoned ‘‘the presumption erectedin Meek and Ball that the placement ofpublic employees on parochial schoolgrounds inevitably results in the impermis-sible effect of state-sponsored indoctrina-tion or constitutes a symbolic union be-tween government and religion.’’ Id., at223, 117 S.Ct. 1997. Rather, relying onZobrest, we explained that in the absenceof evidence showing that teachers wereactually using the Title I aid to inculcatereligion, we would presume that the in-structors would comply with the program’s


secular restrictions. See Agostini, 521U.S., at 223–224, 226–227, 117 S.Ct. 1997.The Title I services were required by stat-ute to be ‘‘ ‘secular, neutral, and nonideo-logical.’ ’’ Id., at 210, 117 S.Ct. 1997 (quot-ing 20 U.S.C. § 6321(a)(2)).

Second, we noted that the Court had‘‘departed from the rule relied on in Ballthat all government aid that directly as-sists the educational function of religiousschools is invalid.’’ Agostini, supra, at225, 117 S.Ct. 1997. Relying on Wittersand Zobrest, we noted that our cases hadtaken a more forgiving view of neutralgovernment programs that make aid avail-able generally without regard to the reli-gious or nonreligious character of the re-cipient school. See Agostini, 521 U.S., at225–226, 117 S.Ct. 1997. With respect tothe specific Title I proSgram848 at issue, wenoted several factors that precluded usfrom finding an impermissible financing ofreligious indoctrination: the aid was ‘‘pro-vided to students at whatever school theychoose to attend,’’ the services were ‘‘bylaw supplemental to the regular curricula’’of the benefited schools, ‘‘[n]o Title I fundsever reach the coffers of religious schools,’’and there was no evidence of Title I in-structors having ‘‘attempted to inculcatereligion in students.’’ Id., at 226–228, 117S.Ct. 1997. Relying on the same factors,we also concluded that the New York Cityprogram could not ‘‘reasonably be viewedas an endorsement of religion.’’ Id., at235, 117 S.Ct. 1997. Although we found itrelevant that Title I services could not beprovided on a schoolwide basis, we alsoexplained that this fact was likely a suffi-cient rather than a necessary condition ofthe program’s constitutionality. We werenot ‘‘willing to conclude that the constitu-tionality of an aid program depends on thenumber of sectarian school students whohappen to receive the otherwise neutralaid.’’ Id., at 229, 117 S.Ct. 1997.

The Chapter 2 program at issue herebears the same hallmarks of the New YorkCity Title I program that we found impor-tant in Agostini. First, as explained

above, Chapter 2 aid is distributed on thebasis of neutral, secular criteria. The aidis available to assist students regardless ofwhether they attend public or private non-profit religious schools. Second, the stat-ute requires participating SEA’s andLEA’s to use and allocate Chapter 2 fundsonly to supplement the funds otherwiseavailable to a religious school. 20 U.S.C.§ 7371(b). Chapter 2 funds must in nocase be used to supplant funds from non-Federal sources. Ibid. Third, no Chapter2 funds ever reach the coffers of a reli-gious school. Like the Title I programconsidered in Agostini, all Chapter 2 fundsare controlled by public agencies—theSEA’s and LEA’s. § 7372(c)(1). TheLEA’s purchase instructional and edu-cational materials and then lend those ma-terials to public and private schools. See§§ 7351(a), (b)(2). With respect to lendingto private schools under Chapter 2, thestatute S 849specifically provides that the rel-evant public agency must retain title to thematerials and equipment. § 7372(c)(1).Together with the supplantation restric-tion, this provision ensures that religiousschools reap no financial benefit by virtueof receiving loans of materials and equip-ment. Finally, the statute provides thatall Chapter 2 materials and equipmentmust be ‘‘secular, neutral, and nonideologi-cal.’’ § 7372(a)(1). That restriction is re-inforced by a further statutory prohibitionon ‘‘the making of any payment TTT forreligious worship or instruction.’’ § 8897.Although respondents claim that Chapter2 aid has been diverted to religious in-struction, that evidence is de minimis, as Iexplain at greater length below. See in-fra, at 2570–2572.

IIIRespondents contend that Agostini is

distinguishable, pointing to the distinctcharacter of the aid program consideredthere. See Brief for Respondents 44–47.In Agostini, federal funds paid for public-school teachers to provide secular instruc-tion to eligible children on the premises of


530 U.S. 851

their religious schools. Here, in contrast,federal funds pay for instructional materi-als and equipment that LEA’s lend to reli-gious schools for use by those schools’ ownteachers in their classes. Because we heldsimilar programs unconstitutional in Meekand Wolman, respondents contend thatthose decisions, and not Agostini, are con-trolling. See, e.g., Brief for Respondents11, 22–25. Like respondents, JusticeSOUTER also relies on Meek and Wolmanin finding the character of the Chapter 2aid constitutionally problematic. See post,at 2586, 2591–2592.

At the time they were decided, Meekand Wolman created an inexplicable riftwithin our Establishment Clause jurispru-dence concerning government aid toschools. Seven years before our decisionin Meek, we held in Allen that a New Yorkstatute that authorized the lending of text-books to students attending religiousschools did not violate theS 850Establishment Clause. 392 U.S., at 238,88 S.Ct. 1923. We explained that the stat-ute ‘‘merely [made] available to all childrenthe benefits of a general program to lendschool books free of charge,’’ that theState retained ownership of the textbooks,and that religious schools received no fi-nancial benefit from the program. Id., at243–244, 88 S.Ct. 1923. We specificallyrejected the contrary argument that thestatute violated the Establishment Clausebecause textbooks are critical to the teach-ing process, which in a religious school isemployed to inculcate religion. Id., at245–248, 88 S.Ct. 1923.

In Meek and Wolman, we adhered toAllen, holding that the textbook lendingprograms at issue in each case did notviolate the Establishment Clause. SeeMeek, 421 U.S., at 359–362, 95 S.Ct. 1753(plurality opinion); Wolman, 433 U.S., at236–238, 97 S.Ct. 2593 (plurality opinion).At the same time, however, we held inboth cases that the lending of instructionalmaterials and equipment to religiousschools was unconstitutional. See Meek,supra, at 362–366, 95 S.Ct. 1753; Wolman,

supra, at 248–251, 97 S.Ct. 2593. We rea-soned that, because the religious schoolsreceiving the materials and equipmentwere pervasively sectarian, any assistancein support of the schools’ educational mis-sions would inevitably have the impermis-sible effect of advancing religion. For ex-ample, in Meek we explained:

‘‘[I]t would simply ignore reality to at-tempt to separate secular educationalfunctions from the predominantly reli-gious role performed by many of Penn-sylvania’s church-related elementaryand secondary schools and to then char-acterize [the statute] as channeling aidto the secular without providing directaid to the sectarian. Even though ear-marked for secular purposes, ‘when itflows to an institution in which religionis so pervasive that a substantial portionof its functions are subsumed in thereligious mission,’ state aid has the im-permissible primary effect of advancingreligion.’’ 421 U.S., at 365–366, 95 S.Ct.1753 (quoting Hunt v. McNair, 413 U.S.734, 743, 93 S.Ct. 2868, 37 L.Ed.2d 923(1973)).

S 851Thus, we held that the aid program‘‘necessarily results in aid to the sectarianschool enterprise as a whole,’’ and ‘‘inesca-pably results in the direct and substantialadvancement of religious activity.’’ Meek,supra, at 366, 95 S.Ct. 1753 (emphasesadded). Similarly, in Wolman, we con-cluded that, ‘‘[i]n view of the impossibilityof separating the secular education func-tion from the sectarian, the state aid inevi-tably flows in part in support of the reli-gious role of the schools.’’ 433 U.S., at250, 97 S.Ct. 2593 (emphasis added).

For whatever reason, the Court was notwilling to extend this presumption of inev-itable religious indoctrination to school aidwhen it instead consisted of textbooks lentfree of charge. For example, in Meek,despite identifying the religious schools’secular educational functions and religiousmissions as inextricably intertwined, 421U.S., at 366, 95 S.Ct. 1753, the Court up-


held the textbook lending program be-cause ‘‘the record in the case TTT, like therecord in Allen, contains no suggestionthat religious textbooks will be lent or thatthe books provided will be used for any-thing other than purely secular purposes,’’id., at 361–362, 95 S.Ct. 1753 (citationomitted). Accordingly, while the Courtwas willing to apply an irrebuttable pre-sumption that secular instructional materi-als and equipment would be diverted touse for religious indoctrination, it requiredevidence that religious schools were divert-ing secular textbooks to religious instruc-tion.

The inconsistency between the twostrands of the Court’s jurisprudence didnot go unnoticed, as Justices on both sidesof the Meek and Wolman decisions reliedon the contradiction to support their re-spective arguments. See, e.g., Meek, 421U.S., at 384, 95 S.Ct. 1753 (Brennan, J.,concurring in part and dissenting in part)(‘‘[W]hat the Court says of the instruction-al materials and equipment may be saidperhaps even more accurately of the text-books’’ (citation omitted)); id., at 390, 95S.Ct. 1753 (REHNQUIST, J., concurringin judgment in part and dissenting in part)(‘‘The failure of the majority to justify thediffering approaches to textbooks and in-structional materials and S 852equipment inthe above respect is symptomatic of itsfailure even to attempt to distinguish theTTT textbook loan program, which the plu-rality upholds, from the TTT instructionalmaterials and equipment loan program,which the majority finds unconstitutional’’).The irrationality of this distinction is pat-ent. As one Member of our Court hasnoted, it has meant that ‘‘a State may lendto parochial school children geographytextbooks that contain maps of the UnitedStates, but the State may not lend maps ofthe United States for use in geographyclass.’’ Wallace v. Jaffree, 472 U.S. 38,110, 105 S.Ct. 2479, 86 L.Ed.2d 29 (1985)(REHNQUIST, J., dissenting) (footnotesomitted).

Indeed, technology’s advance since theAllen, Meek, and Wolman decisions has

only made the distinction between text-books and instructional materials andequipment more suspect. In this case, forexample, we are asked to draw a constitu-tional line between lending textbooks andlending computers. Because computersconstitute instructional equipment, adher-ence to Meek and Wolman would requirethe exclusion of computers from any gov-ernment school aid program that includesreligious schools. Yet, computers are nowas necessary as were schoolbooks 30 yearsago, and they play a somewhat similar rolein the educational process. That Allen,Meek, and Wolman would permit the con-stitutionality of a school aid program toturn on whether the aid took the form of acomputer rather than a book further re-veals the inconsistency inherent in theirlogic.

Respondents insist that there is a rea-soned basis under the EstablishmentClause for the distinction between text-books and instructional materials andequipment. They claim that the presump-tion that religious schools will use instruc-tional materials and equipment to inculcatereligion is sound because such materialsand equipment, unlike textbooks, are rea-sonably divertible to religious uses. Forexample, no matter what secular criteriathe government employs in selecting a filmprojector to lend to a religious school,school officials can always divert that pro-jector to reSligious853 instruction. Respon-dents therefore claim that the Establish-ment Clause prohibits the governmentfrom giving or lending aid to religiousschools when that aid is reasonably diverti-ble to religious uses. See, e.g., Brief forRespondents 11, 35. Justice SOUTERalso states that the divertibility of seculargovernment aid is an important consider-ation under the Establishment Clause, al-though he apparently would not ascribe itthe constitutionally determinative statusthat respondents do. See post, at 2581–2582, 2584–2587.


530 U.S. 855

I would reject respondents’ proposed di-vertibility rule. First, respondents cite noprecedent of this Court that would requireit. The only possible direct precedentialsupport for such a rule is a single sentencecontained in a footnote from our Wolmandecision. There, the Court described Al-len as having been ‘‘premised on the viewthat the educational content of textbooks issomething that can be ascertained in ad-vance and cannot be diverted to sectarianuses.’’ Wolman, supra, at 251, n. 18, 97S.Ct. 2593. To the extent this simple de-scription of Allen is even correct, it cer-tainly does not constitute an actual holdingthat the Establishment Clause prohibitsthe government from lending any diverti-ble aid to religious schools. Rather, asexplained above, the Wolman Court basedits holding invalidating the lending of in-structional materials and equipment to re-ligious schools on the rationale adopted inMeek—that the secular educational func-tion of a religious school is inseparablefrom its religious mission. See Wolman,433 U.S., at 250, 97 S.Ct. 2593. Indeed, ifanything, the Wolman footnote confirmsthe irrationality of the distinction betweentextbooks and instructional materials andequipment. After the Wolman Court ac-knowledged that its holding with respectto instructional materials and equipmentwas in tension with Allen, the Court ex-plained the continuing validity of Allensolely on the basis of stare decisis: ‘‘Boardof Education v. Allen has remained law,and we now follow as a matter of staredecisis the principle that restriction oftextbooks to those provided the publicschools is sufficient to ensure S 854that thebooks will not be used for religious pur-poses.’’ Wolman, supra, at 252, n. 18, 97S.Ct. 2593. Thus, the Wolman Court nev-er justified the inconsistent treatment itaccorded the lending of textbooks and thelending of instructional materials andequipment based on the items’ reasonabledivertibility.

Justice SOUTER’s attempt to defendthe divertibility rationale as a viable dis-

tinction in our Establishment Clause juris-prudence fares no better. For JusticeSOUTER, secular school aid presents con-stitutional problems not only when it isactually diverted to religious ends, but alsowhen it simply has the capacity for, orpresents the possibility of, such diversion.See, e.g., post, at 2586 (discussing ‘‘suscep-tibility [of secular supplies] to the serviceof religious ends’’). Thus, he explains theAllen, Meek, and Wolman decisions as fol-lows: ‘‘While the textbooks had a knownand fixed secular content not readily diver-tible to religious teaching purposes, theadaptable materials did not.’’ Post, at2586–2587. This view would have come asa surprise to the Court in Meek, whichexpressly conceded that ‘‘the material andequipment that are the subjects of the loanTTT are ‘self-polic[ing], in that starting assecular, nonideological and neutral, theywill not change in use.’ ’’ 421 U.S., at 365,95 S.Ct. 1753 (quoting Meek v. Pittinger,374 F.Supp. 639, 660 (E.D.Pa.1974)). In-deed, given the nature of the instructionalmaterials considered in Meek and Wol-man, it is difficult to comprehend how adivertibility rationale could have explainedthe decisions. The statutes at issue inthose cases authorized the lending of ‘‘pe-riodicals, photographs, maps, charts, soundrecordings, [and] films,’’ Meek, supra, at355, 95 S.Ct. 1753, and ‘‘maps and globes,’’Wolman, supra, at 249, 97 S.Ct. 2593.There is no plausible basis for saying thatthese items are somehow more divertiblethan a textbook given that each of theabove items, like a textbook, has a fixedand ascertainable content.

In any event, even if Meek and Wolmanhad articulated the divertibility rationaleurged by respondents and JusticeS 855SOUTER, I would still reject it for amore fundamental reason. Stated simply,the theory does not provide a logical dis-tinction between the lending of textbooksand the lending of instructional materialsand equipment. An educator can use vir-tually any instructional tool, whether it hasascertainable content or not, to teach a


religious message. In this respect, I agreewith the plurality that ‘‘it is hard to imag-ine any book that could not, in even mod-erately skilled hands, serve to illustrate areligious message.’’ Ante, at 2549. Intoday’s case, for example, we are asked todraw a constitutional distinction betweenlending a textbook and lending a librarybook. Justice SOUTER’s try at justifyingthat distinction only demonstrates the ab-surdity on which such a difference mustrest. He states that ‘‘[a]lthough librarybooks, like textbooks, have fixed content,religious teachers can assign secular li-brary books for religious critique.’’ Post,at 2592. Regardless of whether that ex-planation is even correct (for a studentsurely could be given a religious assign-ment in connection with a textbook too), itis hardly a distinction on which constitu-tional law should turn. Moreover, if themere ability of a teacher to devise a reli-gious lesson involving the secular aid inquestion suffices to hold the provision ofthat aid unconstitutional, it is difficult todiscern any limiting principle to the diver-tibility rule. For example, even a publiclyfinanced lunch would apparently be uncon-stitutional under a divertibility rationalebecause religious school officials conceiv-ably could use the lunch to lead the stu-dents in a blessing over the bread. SeeBrief for Avi Chai Foundation as AmicusCuriae 18.

To the extent Justice SOUTER believesseveral related Establishment Clause deci-sions require application of a divertibilityrule in the context of this case, I respect-fully disagree. Justice SOUTER is cor-rect to note our continued recognition ofthe special dangers associated with directmoney grants to religious institutions.See post, at 2584–2586. It does not follow,however, that we should treat as consti-tuStionally856 suspect any form of secularaid that might conceivably be diverted to areligious use. As the cases Justice SOUT-ER cites demonstrate, our concern withdirect monetary aid is based on more thanjust diversion. In fact, the most important

reason for according special treatment todirect money grants is that this form of aidfalls precariously close to the original ob-ject of the Establishment Clause’s prohibi-tion. See, e.g., Walz v. Tax Comm’n ofCity of New York, 397 U.S. 664, 668, 90S.Ct. 1409, 25 L.Ed.2d 697 (1970) (‘‘[F]orthe men who wrote the Religion Clauses ofthe First Amendment the ‘establishment’of a religion connoted sponsorship, finan-cial support, and active involvement of thesovereign in religious activity’’). State-ments concerning the constitutionally sus-pect status of direct cash aid, accordingly,provide no justification for applying anabsolute rule against divertibility when theaid consists instead of instructional materi-als and equipment.

Justice SOUTER also relies on our deci-sions in Wolman (to the extent it con-cerned field-trip transportation for non-public schools), Levitt v. Committee forPublic Ed. & Religious Liberty, 413 U.S.472, 93 S.Ct. 2814, 37 L.Ed.2d 736 (1973),Tilton v. Richardson, 403 U.S. 672, 91S.Ct. 2091, 29 L.Ed.2d 790 (1971), andBowen. See post, at 2586–2587. Nonerequires application of a divertibility rulein the context of this case. Wolman andLevitt were both based on the same pre-sumption that government aid will be usedin the inculcation of religion that we havechosen not to apply to textbook lendingprograms and that we have more generallyrejected in recent decisions. CompareWolman, 433 U.S., at 254, 97 S.Ct. 2593;Levitt, supra, at 480, 93 S.Ct. 2814, withsupra, at 2564; infra, at 2568. In Tilton,we considered a federal statute that autho-rized grants to universities for the con-struction of buildings and facilities to beused exclusively for secular educationalpurposes. See 403 U.S., at 674–675, 91S.Ct. 2091. We held the statute unconsti-tutional only to the extent that a universi-ty’s ‘‘obligation not to use the facility forsectarian instruction or religious worshipTTT appear[ed] to expire at the end of 20years.’’ Id., at 683, 91 S.Ct. 2091. To holda statute unconstitutional because it lacks


530 U.S. 858

a secular content restriction is quite differ-ent S 857from resting on a divertibility ratio-nale. Indeed, the fact that we held thestatute constitutional in all other respectsis more probative on the divertibility ques-tion because it demonstrates our willing-ness to presume that the university wouldabide by the secular content restrictionduring the years the requirement was ineffect. In any event, Chapter 2 containsboth a secular content restriction, 20U.S.C. § 7372(a)(1), and a prohibition onthe use of aid for religious worship orinstruction, § 8897, so Tilton provides nobasis for upholding respondents’ challenge.Finally, our decision in Bowen proves onlythat actual diversion, as opposed to meredivertibility, is constitutionally impermissi-ble. See, e.g., 487 U.S., at 621, 108 S.Ct.2562. Had we believed that the divertibili-ty of secular aid was sufficient to call theaid program into question, there wouldhave been no need for the remand weordered and no basis for the reversal.

IV

Because divertibility fails to explain thedistinction our cases have drawn betweentextbooks and instructional materials andequipment, there remains the question ofwhich of the two irreconcilable strands ofour Establishment Clause jurisprudencewe should now follow. Between the two, Iwould adhere to the rule that we haveapplied in the context of textbook lendingprograms: To establish a First Amend-ment violation, plaintiffs must prove thatthe aid in question actually is, or hasbeen, used for religious purposes. SeeMeek, 421 U.S., at 361–362, 95 S.Ct. 1753;Allen, 392 U.S., at 248, 88 S.Ct. 1923.Just as we held in Agostini that our morerecent cases had undermined the assump-tions underlying Ball and Aguilar, Iwould now hold that Agostini and thecases on which it relied have underminedthe assumptions underlying Meek andWolman. To be sure, Agostini only ad-dressed the specific presumption that pub-

lic-school employees teaching on thepremises of religious schools would inevi-tably inculcate religion. Nevertheless, Ibelieve that our definitive rejection of thatpresumption also stood for—or at leaststrongly S 858pointed to—the broader propo-sition that such presumptions of religiousindoctrination are normally inappropriatewhen evaluating neutral school aid pro-grams under the Establishment Clause.In Agostini, we repeatedly emphasizedthat it would be inappropriate to presumeinculcation of religion; rather, plaintiffsraising an Establishment Clause challengemust present evidence that the govern-ment aid in question has resulted in reli-gious indoctrination. See 521 U.S., at223–224, 226–227, 117 S.Ct. 1997. Wespecifically relied on our statement in Zo-brest that a presumption of indoctrination,because it constitutes an absolute bar tothe aid in question regardless of the reli-gious school’s ability to separate that aidfrom its religious mission, constitutes a‘‘flat rule, smacking of antiquated notionsof ‘taint,’ [that] would indeed exalt formover substance.’’ 509 U.S., at 13, 113S.Ct. 2462. That reasoning applies withequal force to the presumption in Meekand Ball concerning instructional materi-als and equipment. As we explained inAgostini, ‘‘we have departed from the rulerelied on in Ball that all government aidthat directly assists the educational func-tion of religious schools is invalid.’’ 521U.S., at 225, 117 S.Ct. 1997.

Respondents contend that Agostinishould be limited to its facts, and pointspecifically to the following statement frommy separate opinion in Ball as the basisfor retaining a presumption of religiousinculcation for instructional materials andequipment:

‘‘When full-time parochial school teach-ers receive public funds to teach secularcourses to their parochial school stu-dents under parochial school supervi-sion, I agree that the program has theperceived and actual effect of advancingthe religious aims of the church-related


schools. This is particularly the casewhere, as here, religion pervades thecurriculum and the teachers are accus-tomed to bring religion to play in every-thing they teach.’’ 473 U.S., at 399–400,105 S.Ct. 3248 (concurring in judgmentin part and dissenting in part).

S 859Respondents note that in Agostini wedid not overrule that portion of Ball hold-ing the Community Education programunconstitutional. Under that program, thegovernment paid religious school teachersto operate as part-time public teachers attheir religious schools by teaching secularclasses at the conclusion of the regularschoolday. Ball, 473 U.S., at 376–377, 105S.Ct. 3248. Relying on both the majorityopinion and my separate opinion in Ball,respondents therefore contend that wemust presume that religious school teach-ers will inculcate religion in their students.If that is so, they argue, we must alsopresume that religious school teachers willbe unable to follow secular restrictions onthe use of instructional materials andequipment lent to their schools by thegovernment. See Brief for Respondents26–29.

I disagree, however, that the latterproposition follows from the former.First, as our holding in Allen and its reaf-firmance in Meek and Wolman demon-strate, the Court’s willingness to assumethat religious school instructors will incul-cate religion has not caused us to presumealso that such instructors will be unable tofollow secular restrictions on the use oftextbooks. I would similarly reject anysuch presumption regarding the use of in-structional materials and equipment.When a religious school receives textbooksor instructional materials and equipmentlent with secular restrictions, the school’steachers need not refrain from teachingreligion altogether. Rather, the instruc-tors need only ensure that any such reli-gious teaching is done without the instruc-tional aids provided by the government.We have always been willing to assumethat religious school instructors can abide

by such restrictions when the aid consistsof textbooks, which Justice Brennan de-scribed as ‘‘surely the heart tools of TTT

education.’’ Meek, supra, at 384, 95 S.Ct.1753 (concurring in part and dissenting inpart). The same assumption should ex-tend to instructional materials and equip-ment.

For the same reason, my position in Ballis distinguishable. There, the governmentpaid for religious school instructors S 860toteach classes supplemental to those offeredduring the normal schoolday. In that con-text, I was willing to presume that thereligious school teacher who worksthroughout the day to advance the school’sreligious mission would also do so, at leastto some extent, during the supplementalclasses provided at the end of the day.Because the government financed the en-tirety of such classes, any religious indoc-trination taking place therein would bedirectly attributable to the government.In the instant case, because the Chapter 2aid concerns only teaching tools that mustremain supplementary, the aid constitutesonly a portion of the teacher’s educationalefforts during any single class. In thiscontext, I find it easier to believe that areligious school teacher can abide by thesecular restrictions placed on the govern-ment assistance. I therefore would notpresume that the Chapter 2 aid will ad-vance, or be perceived to advance, theschool’s religious mission.

VRespondents do not rest, however, on

their divertibility argument alone. Rath-er, they also contend that the evidencerespecting the actual administration ofChapter 2 in Jefferson Parish demon-strates that the program violated the Es-tablishment Clause. First, respondentsclaim that the program’s safeguards areinsufficient to uncover instances of actualdiversion. Brief for Respondents 37, 42–43, 45–47. Second, they contend that therecord shows that some religious schoolsin Jefferson Parish may have used their


530 U.S. 862

Chapter 2 aid to support religious edu-cation (i.e., that they diverted the aid).Id., at 36–37. Third, respondents high-light violations of Chapter 2’s secularcontent restrictions. Id., at 39–41. And,finally, they note isolated examples of po-tential violations of Chapter 2’s supplan-tation restriction. Id., at 43–44. Basedon the evidence underlying the first andsecond claims, the plurality appears tocontend that the Chapter 2 program canbe upheld only if actual diversion of gov-ernment aid to the advancement of reli-gion S 861is permissible under the Estab-lishment Clause. See ante, at 2553–2555.Relying on the evidence underlying allbut the last of the above claims, JusticeSOUTER concludes that the Chapter 2program, as applied in Jefferson Parish,violated the Establishment Clause. Seepost, at 2591–2596. I disagree with boththe plurality and Justice SOUTER. Thelimited evidence amassed by respondentsduring 4 years of discovery (which beganapproximately 15 years ago) is at best deminimis and therefore insufficient to af-fect the constitutional inquiry.

The plurality and Justice SOUTER di-rect the primary thrust of their argumentsat the alleged inadequacy of the program’ssafeguards. Respondents, the plurality,and Justice SOUTER all appear to pro-ceed from the premise that, so long asactual diversion presents a constitutionalproblem, the government must have a fail-safe mechanism capable of detecting anyinstance of diversion. We rejected thatvery assumption, however, in Agostini.There, we explained that because we had‘‘abandoned the assumption that properlyinstructed public employees will fail to dis-charge their duties faithfully, we must alsodiscard the assumption that pervasivemonitoring of Title I teachers is required.’’521 U.S., at 234, 117 S.Ct. 1997 (emphasisin original). Because I believe that theCourt should abandon the presumptionadopted in Meek and Wolman respectingthe use of instructional materials and

equipment by religious school teachers, Isee no constitutional need for pervasivemonitoring under the Chapter 2 program.

The safeguards employed by the pro-gram are constitutionally sufficient. Atthe federal level, the statute limits aid to‘‘secular, neutral, and nonideological ser-vices, materials, and equipment,’’ 20 U.S.C.§ 7372(a)(1); requires that the aid onlysupplement and not supplant funds fromnon-Federal sources, § 7371(b); and pro-hibits ‘‘any payment TTT for religious wor-ship or instruction,’’ § 8897. At the statelevel, the Louisiana Department of Edu-cation (the relevant SEA for S 862Louisiana)requires all nonpublic schools to submitsigned assurances that they will use Chap-ter 2 aid only to supplement and not tosupplant non-Federal funds, and that theinstructional materials and equipment ‘‘willonly be used for secular, neutral and noni-deological purposes.’’ App. 260a–261a;see also id., at 120a. Although there issome dispute concerning the mandatorynature of these assurances, Dan Lewis, thedirector of Louisiana’s Chapter 2 program,testified that all of the State’s nonpublicschools had thus far been willing to signthe assurances, and that the State retainedthe power to cut off aid to any school thatbreached an assurance. Id., at 122a–123a.The Louisiana SEA also conducts monitor-ing visits to each of the State’s LEA’s—and one or two of the nonpublic schoolscovered by the relevant LEA—once everythree years. Id., at 95a–96a. In additionto other tasks performed on such visits,SEA representatives conduct a random re-view of a school’s library books for reli-gious content. Id., at 99a.

At the local level, the Jefferson ParishPublic School System (JPPSS) requiresnonpublic schools seeking Chapter 2 aid tosubmit applications, complete with specificproject plans, for approval. Id., at 127a;id., at 194a–203a (sample application).The JPPSS then conducts annual monitor-ing visits to each of the nonpublic schoolsreceiving Chapter 2 aid. Id., at 141a–142a.On each visit, a JPPSS representative


meets with a contact person from the non-public school and reviews with that personthe school’s project plan and the manner inwhich the school has used the Chapter 2materials and equipment to support itsplan. Id., at 142a, 149a. The JPPSS rep-resentative also reminds the contact per-son of the prohibition on the use of Chap-ter 2 aid for religious purposes, id., at149a, and conducts a random sample of theschool’s Chapter 2 materials and equip-ment to ensure that they are appropriatelylabeled and that the school has maintaineda record of their usage, id., at 142a–144a.(Although the plurality and Justice SOUT-ER claim that compliance S 863with the la-beling requirement was haphazard, bothcite only a statewide monitoring reportthat includes no specific findings with re-spect to Jefferson Parish. Ante, at 2553–2554 (citing App. 113a); post, at 2593–2594(same).) Finally, the JPPSS representa-tive randomly selects library books thenonpublic school has acquired throughChapter 2 and reviews their content toensure that they comply with the pro-gram’s secular content restriction. App.210a. If the monitoring does not satisfythe JPPSS representative, another visit isscheduled. Id., at 151a–152a. Apart fromconducting monitoring visits, the JPPSSreviews Chapter 2 requests filed by partic-ipating nonpublic schools. As part of thisprocess, a JPPSS employee examines thetitles of requested library books and re-jects any book whose title reveals (or sug-gests) a religious subject matter. Id., at135a, 137a–138a. As the above descriptionof the JPPSS monitoring process shouldmake clear, Justice SOUTER’s citation ofa statewide report finding a lack of moni-toring in some Louisiana LEA’s is irrele-vant as far as Jefferson Parish is con-cerned. See post, at 2593–2594 (quotingApp. 111a).

Respondents, the plurality, and JusticeSOUTER all fault the above-describedsafeguards primarily because they dependon the good faith of participating religiousschool officials. For example, both the

plurality and Justice SOUTER repeatedlycite testimony by state and parish officialsacknowledging that the safeguards dependto a certain extent on the religious schools’self-reporting and that, therefore, there isno way for the State or Jefferson Parish tosay definitively that no Chapter 2 aid isdiverted to religious purposes. See, e.g.,ante, at 2554, n. 15; post, at 2593–2594.These admissions, however, do not provethat the safeguards are inadequate. Tofind that actual diversion will flourish, onemust presume bad faith on the part of thereligious school officials who report to theJPPSS monitors regarding the use ofChapter 2 aid. I disagree with the plurali-ty and Justice SOUTER on this point andbelieve that it is entirely S 864proper to pre-sume that these school officials will act ingood faith. That presumption is especiallyappropriate in this case, since there is noproof that religious school officials havebreached their schools’ assurances orfailed to tell government officials the truth.Cf. Tilton, 403 U.S., at 679, 91 S.Ct. 2091(‘‘A possibility always exists, of course,that the legitimate objectives of any law orlegislative program may be subverted byconscious design or lax enforcementTTTT

But judicial concern about these possibili-ties cannot, standing alone, warrant strik-ing down a statute as unconstitutional’’).

The evidence proffered by respondents,and relied on by the plurality and JusticeSOUTER, concerning actual diversion ofChapter 2 aid in Jefferson Parish is deminimis. Respondents first cite the fol-lowing statement from a Jefferson Parishreligious school teacher: ‘‘Audio-visual ma-terials are a very necessary and enjoyabletool used when teaching young children.As a second grade teacher I use them inall subjects and see a very positive result.’’App. 108a. Respondents’ only other evi-dence consists of a chart concerning oneJefferson Parish religious school, whichshows that the school’s theology depart-ment was a significant user of audiovisualequipment. See id., at 206a–208a. Al-


530 U.S. 866

though an accompanying letter indicatesthat much of the school’s equipment waspurchased with federal funds, id., at 205a,the chart does not provide a breakdownidentifying specific Chapter 2 usage. In-deed, unless we are to relieve respondentsof their evidentiary burden and presume aviolation of Chapter 2, we should assumethat the school used its own equipment inthe theology department and the Chapter2 equipment elsewhere. The more basicpoint, however, is that neither piece ofevidence demonstrates that Chapter 2 aidactually was diverted to religious edu-cation. At most, it proves the possibilitythat, out of the more than 40 nonpublicschools in Jefferson Parish participating inChapter 2, aid may have been diverted inone school’s second-grade class and anoth-er school’s theology department.

S 865The plurality’s insistence that this ev-idence is somehow substantial flatly con-tradicts its willingness to disregard simi-larly insignificant evidence of violations ofChapter 2’s supplantation and secular con-tent restrictions. See ante, at 2544, n. 7(finding no ‘‘material statutory violation’’ ofthe supplantation restriction); ante, at2555 (characterizing violations of secularcontent restriction as ‘‘scattered’’ and ‘‘deminimis ’’). As I shall explain below, Ibelieve the evidence on all three points isequally insignificant and, therefore, shouldbe treated the same.

Justice SOUTER also relies on testimo-ny by one religious school principal indi-cating that a computer lent to her schoolunder Chapter 2 was connected through anetwork to non-Chapter 2 computers. Seepost, at 2595 (citing App. 77a). The princi-pal testified that the Chapter 2 computerwould take over the network if anothernon-Chapter 2 computer were to breakdown. Ibid. To the extent the principal’stestimony even proves that Chapter 2funds were diverted to the school’s reli-gious mission, the evidence is hardly com-pelling.

Justice SOUTER contends that any evi-dence of actual diversion requires the

Court to declare the Chapter 2 programunconstitutional as applied in JeffersonParish. Post, at 2595, n. 27. For support,he quotes my concurring opinion in Bowenand the statement therein that ‘‘any use ofpublic funds to promote religious doctrinesviolates the Establishment Clause.’’ 487U.S., at 623, 108 S.Ct. 2562 (emphasis inoriginal). That principle of course remainsgood law, but the next sentence in myopinion is more relevant to the case athand: ‘‘[E]xtensive violations—if they canbe proved in this case—will be highly rele-vant in shaping an appropriate remedythat ends such abuses.’’ Ibid. (emphasis inoriginal). I know of no case in which wehave declared an entire aid program un-constitutional on Establishment Clausegrounds solely because of violations on theminuscule scale of those at issue here.Yet that is precisely the remedy respon-dents S 866requested from the District Courtand that they were granted by the Courtof Appeals. See App. 51a; Helms v. Pi-card, 151 F.3d 347, 377 (C.A.5 1998),amended, 165 F.3d 311, 312 (C.A.5 1999).While extensive violations might require aremedy along the lines asked for by re-spondents, no such evidence has been pre-sented here. To the contrary, the pres-ence of so few examples over a period of atleast 4 years (15 years ago) tends to shownot that the ‘‘no-diversion’’ rules havefailed, but that they have worked. Ac-cordingly, I see no reason to affirm thejudgment below and thereby declare aproperly functioning aid program unconsti-tutional.

Respondents’ next evidentiary argumentconcerns an admitted violation of Chapter2’s secular content restriction. Over threeyears, Jefferson Parish religious schoolsordered approximately 191 religious li-brary books through Chapter 2. App.129a–133a. Dan Lewis, the director ofLouisiana’s Chapter 2 program, testifiedthat he discovered some of the religiousbooks while performing a random checkduring a state monitoring visit to a Jeffer-


son Parish religious school. Id., at 99a–100a. The discovery prompted the Stateto notify the JPPSS, which then reexam-ined book requests dating back to 1982,discovered the 191 books in question, andrecalled them. Id., at 130a–133a. Thisseries of events demonstrates not that theChapter 2 safeguards are inadequate, butrather that the program’s monitoring sys-tem succeeded. Even if I were insteadwilling to find this incident to be evidenceof a likelihood of future violations, theevidence is insignificant. The 191 booksconstituted less than one percent of thetotal allocation of Chapter 2 aid in Jeffer-son Parish during the relevant years. Id.,at 132a. Justice SOUTER understandablyconcedes that the book incident constitutes‘‘only limited evidence.’’ Post, at 2595. Iagree with the plurality that, like theabove evidence of actual diversion, the bor-rowing of the religious library books con-stitutes only de minimis evidence. Seeante, at 2555.

S 867Respondents’ last evidentiary chal-lenge concerns the effectiveness of Chap-ter 2’s supplantation restriction in Jeffer-son Parish. Although Justice SOUTERdoes not rest his decision on this point, hedoes ‘‘not[e] the likelihood that unconstitu-tional supplantation occurred as well.’’Post, at 2596, n. 28. I disagree. Theevidence cited by respondents and JusticeSOUTER is too ambiguous to rest anysound conclusions on and, at best, showssome scattered violations of the statutorysupplantation restriction that are too insig-nificant in aggregate to affect the constitu-tional inquiry. Indeed, even JusticeSOUTER concedes in this respect that‘‘[t]he record is sparse.’’ Post, at 2596, n.28.

* * *Given the important similarities between

the Chapter 2 program here and the TitleI program at issue in Agostini, respon-dents’ Establishment Clause challengemust fail. As in Agostini, the Chapter 2aid is allocated on the basis of neutral,secular criteria; the aid must be supple-

mentary and cannot supplant non-Federalfunds; no Chapter 2 funds ever reach thecoffers of religious schools; the aid mustbe secular; any evidence of actual diver-sion is de minimis; and the program in-cludes adequate safeguards. Regardlessof whether these factors are constitutionalrequirements, they are surely sufficient tofind that the program at issue here doesnot have the impermissible effect of ad-vancing religion. For the same reasons,‘‘this carefully constrained program alsocannot reasonably be viewed as an en-dorsement of religion.’’ Agostini, 521U.S., at 235, 117 S.Ct. 1997. Accordingly,I concur in the judgment.

Justice SOUTER, with whom JusticeSTEVENS and Justice GINSBURG join,dissenting.

The First Amendment’s EstablishmentClause prohibits Congress (and, by incor-poration, the States) from making any lawrespecting an establishment of religion. Ithas been S 868held to prohibit not only theinstitution of an official church, but anygovernment act favoring religion, a partic-ular religion, or for that matter irreligion.Thus, it bars the use of public funds forreligious aid.

The establishment prohibition of govern-ment religious funding serves more thanone end. It is meant to guarantee theright of individual conscience against com-pulsion, to protect the integrity of religionagainst the corrosion of secular support,and to preserve the unity of political soci-ety against the implied exclusion of theless favored and the antagonism of contro-versy over public support for religiouscauses.

These objectives are always in somejeopardy since the substantive principle ofno aid to religion is not the only limitationon government action toward religion.Because the First Amendment also barsany prohibition of individual free exerciseof religion, and because religious organiza-tions cannot be isolated from the basic


530 U.S. 870

government functions that create the civilenvironment, it is as much necessary as itis difficult to draw lines between forbiddenaid and lawful benefit. For more than 50years, this Court has been attempting todraw these lines. Owing to the variety offactual circumstances in which the linesmust be drawn, not all of the points creat-ing the boundary have enjoyed self-evi-dence.

So far as the line drawn has addressedgovernment aid to education, a few funda-mental generalizations are nonethelesspossible. There may be no aid supportinga sectarian school’s religious exercise orthe discharge of its religious mission, whileaid of a secular character with no discerni-ble benefit to such a sectarian objective isallowable. Because the religious and secu-lar spheres largely overlap in the life ofmany such schools, the Court has tried toidentify some facts likely to reveal therelative religious or secular intent or effectof the government benefits in particularcircumstances. We have asked whetherthe government is acting neutrally in dis-tributing its money, and about the form ofthe aid itself, its path from government toreligious institution, S 869its divertibility toreligious nurture, its potential for reducingtraditional expenditures of religious insti-tutions, and its relative importance to therecipient, among other things.

In all the years of its effort, the Courthas isolated no single test of constitutionalsufficiency, and the question in every caseaddresses the substantive principle of noaid: what reasons are there to character-ize this benefit as aid to the sectarianschool in discharging its religious mission?Particular factual circumstances control,and the answer is a matter of judgment.

In what follows I will flesh out thissummary, for this case comes at a timewhen our judgment requires perspectiveon how the Establishment Clause hascome to be understood and applied. It isnot just that a majority today mistakes thesignificance of facts that have led to con-clusions of unconstitutionality in earlier

cases, though I believe the Court commitserror in failing to recognize the divertibili-ty of funds to the service of religious ob-jectives. What is more important is theview revealed in the plurality opinion,which espouses a new conception of neu-trality as a practically sufficient test ofconstitutionality that would, if adopted bythe Court, eliminate enquiry into a law’seffects. The plurality position breaks fun-damentally with Establishment Clauseprinciple, and with the methodology pains-takingly worked out in support of it. Imean to revisit that principle and describethe methodology at some length, lest therebe any question about the rupture that theplurality view would cause. From thatnew view of the law, and from a majority’smistaken application of the old, I respect-fully dissent.

I

The prohibition that ‘‘Congress shallmake no law respecting an establishmentof religion,’’ U.S. Const., Amdt. 1, eludeselegant conceptualization simply becausethe prohibition applies to such distinctphenomena as state churches and aid toreligious schools, and as applied to schoolaid has S 870prompted challenges to pro-grams ranging from construction subsidiesto hearing aids to textbook loans. Anycriteria, moreover, must not only definethe margins of the establishment prohibi-tion, but must respect the succeedingClause of the First Amendment guarantee-ing religion’s free exercise. Ibid. It is nowonder that the complementary constitu-tional provisions and the inexhaustablyvarious circumstances of their applicabilityhave defied any simple test and have in-stead produced a combination of generalrules often in tension at their edges. Ifcoherence is to be had, the Court has tokeep in mind the principal objectivesserved by the Establishment Clause, andits application to school aid, and their rec-ollection may help to explain the misunder-


standings that underlie the majority’s re-sult in this case.

A

At least three concerns have been ex-pressed since the founding and runthroughout our First Amendment jurispru-dence. First, compelling an individual tosupport religion violates the fundamentalprinciple of freedom of conscience. Madi-son’s and Jefferson’s now familiar wordsestablish clearly that liberty of personalconviction requires freedom from coercionto support religion,1 and this means thatthe government can compel no aid to fundit. Madison put it simply: ‘‘[T]he sameauthority which can force a citizen toS 871contribute three pence only of his prop-erty for the support of any one establish-ment, may force him to conform to anyother establishment.’’ Memorial and Re-monstrance ¶ 3, reprinted in Everson v.Board of Ed. of Ewing, 330 U.S. 1, 64, 65–66, 67 S.Ct. 504, 91 L.Ed. 711 (1947). Anytax to establish religion is antithetical tothe command ‘‘that the minds of men al-ways be wholly free.’’ Id., at 12, 67 S.Ct.504 (discussing Madison’s Memorial andRemonstrance); id., at 13, 67 S.Ct. 504(noting Jefferson’s belief that ‘‘com-pel[ling] a man to furnish contributions ofmoney for the propagation of opinionswhich he disbelieves, is sinful and tyranni-cal; TTT even the forcing him to supportthis or that teacher of his own religiouspersuasion, is depriving him of the com-fortable liberty of giving his contributionsto the particular pastor, whose morals hewould make his pattern’’ (internal quota-tion marks omitted)); see also Rosenber-ger v. Rector and Visitors of Univ. of Va.,

515 U.S. 819, 868–874, 115 S.Ct. 2510, 132L.Ed.2d 700 (1995) (SOUTER, J., dissent-ing).

Second, government aid corrupts reli-gion. See Engel v. Vitale, 370 U.S. 421,431, 82 S.Ct. 1261, 8 L.Ed.2d 601 (1962)(‘‘[The Establishment Clause’s] first andmost immediate purpose rested on the be-lief that a union of government and reli-gion tends to destroy government and todegrade religion’’); Everson, supra, at 53,67 S.Ct. 504 (Rutledge, J., dissenting).Madison argued that establishment of reli-gion weakened the beliefs of adherents sofavored, strengthened their opponents, andgenerated ‘‘pride and indolence in theClergy; ignorance and servility in the lai-ty; [and] in both, superstition, bigotry andpersecution.’’ Memorial and Remon-strance ¶ 7, quoted in Everson, 330 U.S., at67, 67 S.Ct. 504. ‘‘[E]xperience witnesseththat ecclesiastical establishments, insteadof maintaining the purity and efficacy ofReligion, have had a contrary operation.’’Ibid. In a variant of Madison’s concern, wehave repeatedly noted that a government’sfavor to a particular religion or sectthreatens to taint it with ‘‘corrosive secu-larism.’’ Lee v. Weisman, 505 U.S. 577,608, 112 S.Ct. 2649, 120 L.Ed.2d 467 (1992)(internal quotation marks and citationsomitted); see also Illinois ex rel. McCol-lum v. Board S 872of Ed. of School Dist. No.71, Champaign Cty., 333 U.S. 203, 228, 68S.Ct. 461, 92 L.Ed. 649 (1948).

‘‘[G]overnment and religion have dis-crete interests which are mutually bestserved when each avoids too close aproximity to the other. It is not onlythe nonbeliever who fears the injectionof sectarian doctrines and controversiesinto the civil polity, but in as high de-gree it is the devout believer who fears

1. Jefferson’s Virginia Bill for Establishing Re-ligious Freedom provided ‘‘[t]hat no manshall be compelled to frequent or support anyreligious worship, place, or ministry whatso-everTTTT’’ Jefferson, A Bill for EstablishingReligious Freedom, in 5 The Founder’s Con-stitution 84 (P. Kurland & R. Lerner eds.1987); see also Rosenberger v. Rector and Visi-tors of Univ. of Va., 515 U.S. 819, 870–872,115 S.Ct. 2510, 132 L.Ed.2d 700 (1995)

(SOUTER, J., dissenting). We have ‘‘previ-ously recognized that the provisions of theFirst Amendment, in the drafting and adop-tion of which Madison and Jefferson playedsuch leading roles, had the same objectiveand were intended to provide the same pro-tection against governmental intrusion on re-ligious liberty as the Virginia statute.’’ Ever-son v. Board of Ed. of Ewing, 330 U.S. 1, 13,67 S.Ct. 504, 91 L.Ed. 711 (1947).


530 U.S. 874

the secularization of a creed which be-comes too deeply involved with and de-pendent upon the government.’’ SchoolDist. of Abington Township v. Schempp,374 U.S. 203, 259, 83 S.Ct. 1560, 10L.Ed.2d 844 (1963) (Brennan, J., concur-ring).

See also Rosenberger, supra, at 890–891,115 S.Ct. 2510 (SOUTER, J., dissenting).

Third, government establishment of re-ligion is inextricably linked with conflict.Everson, supra, at 8–11, 67 S.Ct. 504 (re-lating colonists’ understanding of recenthistory of religious persecution in coun-tries with established religion); Engel, su-pra, at 429, 82 S.Ct. 1261 (discussingstruggle among religions for governmentapproval); Lemon v. Kurtzman, 403 U.S.602, 623, 91 S.Ct. 2105 (1971). In ourown history, the turmoil thus producedhas led to a rejection of the idea thatgovernment should subsidize religious ed-ucation, id., at 645–649, 91 S.Ct. 2105(opinion of Brennan, J.) (discussing histo-ry of rejection of support for religiousschools); McCollum, supra, at 214–217,68 S.Ct. 461 (opinion of Frankfurter, J.), aposition that illustrates the Court’s under-standing that any implicit endorsement ofreligion is unconstitutional, see County ofAllegheny v. American Civil LibertiesUnion, Greater Pittsburgh Chapter, 492U.S. 573, 592–594, 109 S.Ct. 3086, 106L.Ed.2d 472 (1989).2

S 873B

These concerns are reflected in theCourt’s classic summation delivered in Ev-

erson v. Board of Education, supra, itsfirst opinion directly addressing standardsgoverning aid to religious schools: 3

‘‘The ‘establishment of religion’ clause ofthe First Amendment means at leastthis: Neither a state nor the FederalGovernment can set up a church. Nei-ther can pass laws which aid one reli-gion, aid all religions, or prefer one reli-gion over another. Neither can forcenor influence a person to go to or toremain away from church against hiswill or force him to profess a belief ordisbelief in any religion. No person canbe punished for entertaining or profess-ing religious beliefs or disbeliefs, forchurch attendance or non-attendance.No tax in any amount, large or small,can be levied to support any religiousactivities or institutions, whatever theymay be called, or whatever form theymay adopt to teach or practice religion.Neither a state nor the Federal Govern-ment can, openly or secretly, participatein the affairs of any religious organiza-tions or groups and vice versa. In thewords of Jefferson, the clause againstestablishment of religion by law was in-tended to erect ‘a wall of separationbetween church and State.’ ’’ 330 U.S.at 15–16, 67 S.Ct. 504 (quoting Reynoldsv. United States, 98 U.S. 145, 164, 25L.Ed. 244 (1878)).

The most directly pertinent doctrinalstatements here are these: no government‘‘can pass laws which aid one religionS 874[or] all religions TTTT No tax in anyamount TTT can be levied to support anyreligious activities or institutions TTT what-

2. The plurality mistakes my recognition ofthis fundamental concern. Ante, at 2549–2550. The Court may well have moved awayfrom considering the political divisivenessthreatened by particular instances of aid as apractical criterion for applying the Establish-ment Clause case by case, but we have neverquestioned its importance as a motivatingconcern behind the Establishment Clause, norcould we change history to find that sectarianconflict did not influence the Framers whowrote it.

3. The Court upheld payments by Indian tribesto apparently Roman Catholic schools inQuick Bear v. Leupp, 210 U.S. 50, 28 S.Ct.690, 52 L.Ed. 954 (1908), suggesting in dictathat there was no Establishment Clause prob-lem, but it did not squarely face the question.Nor did the Court address a First Amendmentchallenge to a state program providing text-books to children in Cochran v. Louisiana Bd.of Ed., 281 U.S. 370, 50 S.Ct. 335, 74 L.Ed.913 (1930); it simply concluded that the pro-gram had an adequate public purpose. TheCourt first squarely faced the issue in Everson.


ever form they may adopt to teach TTT

religion.’’ 330 U.S. at, 16, 67 S.Ct. 504.Thus, the principle of ‘‘no aid,’’ with whichno one in Everson disagreed.4

Immediately, however, there was thedifficulty over what might amount to ‘‘aid’’or ‘‘support.’’ The problem for the Ever-son Court was not merely the imprecisionof the words, but the ‘‘other language ofthe [First Amendment that] commandsthat [government] cannot hamper its citi-zens in the free exercise of their ownreligion,’’ ibid., with the consequence thatgovernment must ‘‘be a neutral in its rela-tions with groups of religious believers andnon-believers,’’ id. at, 18, 67 S.Ct. 504.Since withholding some public benefitsfrom religious groups could be said to‘‘hamper’’ religious exercise indirectly, andextending other benefits said to aid it, anargument-proof formulation of the no-aidprinciple was impossible, and the Courtwisely chose not to attempt any such thing.Instead it gave definitive examples of pub-lic benefits provided pervasively through-out society that would be of some value toorganized religion but not in a way or to adegree that could sensibly be described asgiving it aid or violating the neutralityrequirement: there was no EstablishmentClause concern with ‘‘such general govern-ment services as ordinary police and fireprotection, connections for sewage dispos-al, public highways and sidewalks.’’ Id., at17–18, 67 S.Ct. 504. These ‘‘benefits ofpublic welfare legislation,’’ id., at 16, 67S.Ct. 504, extended in modern times tovirtually every member of the populationand valuable to every person and associa-tion, were the paradigms of advantagesthat religious organizaStions875 could enjoyconsistently with the prohibition againstaid, and that governments could extendwithout deserting their required positionof neutrality.

But paradigms are not perfect fits veryoften, and government spending resistseasy classification as between universalgeneral service or subsidy of favoritism.The 5–to–4 division of the Everson Courtturned on the inevitable question whetherreimbursing all parents for the cost oftransporting their children to school wasclose enough to police protection to toler-ate its indirect benefit in some degree toreligious schools, with the majority in Ev-erson thinking the reimbursement statutefell on the lawful side of the line. Al-though the state scheme reimbursed par-ents for transporting children to sectarianschools, among others, it gave ‘‘no moneyto the schools. It [did] not support them.Its legislation [did] no more than provide ageneral program to help parents get theirchildren, regardless of their religion, safelyand expeditiously to and from accreditedschools.’’ Id., at 18, 67 S.Ct. 504. Thedissenters countered with factual analysesshowing the limitation of the law’s benefitsin fact to private school pupils who wereRoman Catholics, id., at 20, 67 S.Ct. 504(Jackson, J., dissenting), and indicating theinseparability of transporting pupils toschool from support for the religious in-struction that was the school’s raisond’etre, id., at 45–46, 67 S.Ct. 504 (Rutledge,J., dissenting).

Everson is usefully understood in thelight of a successor case two decades later,Board of Ed. of Central School Dist. No. 1v. Allen, 392 U.S. 236, 88 S.Ct. 1923, 20L.Ed.2d 1060 (1968), in which the chal-lenged government practice was lendingtextbooks to pupils of schools both publicand private, including religious ones (as towhich there was no evidence that they hadpreviously supplied books to their classesand some evidence that they had not, id.,at 244, n. 6, 88 S.Ct. 1923). By the time ofAllen, the problem of classifying the statebenefit, as between aid to religion and

4. While Everson’s dissenters parted companywith the majority over the specific question ofschool buses, the Court stood as one behindthe principle of no aid for religious teaching.

330 U.S., at 15–16, 67 S.Ct. 504; id., at 25–26, 67 S.Ct. 504 (Jackson, J., dissenting); id.,at 28–29, 31–32, 67 S.Ct. 504 (Rutledge, J.,dissenting).


530 U.S. 877

general public service consistent with gov-ernment neutralSity,876 had led to the for-mulation of a ‘‘test’’ that required secular,primary intent and effect as necessaryconditions of any permissible scheme. Id.,at 243, 88 S.Ct. 1923. Again the Courtsplit, upholding the state law in issue, butwith Everson’s majority author, JusticeBlack, now in dissent. What is remarka-ble about Allen today, however, is not somuch its division as its methodology, forthe consistency in the way the Justiceswent about deciding the case transcendedtheir different conclusions. Neither siderested on any facile application of the‘‘test’’ or any simplistic reliance on thegenerality or evenhandedness of the statelaw. Disagreement concentrated on thetrue intent inferrable behind the law, thefeasibility of distinguishing in fact betweenreligious and secular teaching in churchschools, and the reality or sham of lendingbooks to pupils instead of supplying booksto schools. The majority, to be sure, citedthe provision for books to all schoolchil-dren, regardless of religion, 392 U.S., at243, 88 S.Ct. 1923, just as the Eversonmajority had spoken of the transportationreimbursement as going to all, 330 U.S., at16, 67 S.Ct. 504, in each case for the sakeof analogy to the provision of police andfire services.5 But the stress was on thepractical significance of the actual benefitsreceived by the schools. As Everson hadrested on the understanding that no mon-ey and no support went to the school, id.,at 18, 67 S.Ct. 504, Allen emphasized thatthe savings to parents were devoid of anymeasurable effect in teaching religion, 392U.S., at 243–244, 88 S.Ct. 1923. JusticeHarlan, concurring, summed up the ap-proach with his observations that the re-quired government ‘‘[n]eutrality is TTT acoat of many colors,’’ and quoted JusticeGoldberg’s conclusion, that there was ‘‘ ‘nosimple and clear measure’ TTT by whichthis or any [religious school aid] case may

readily be decided,’’ id., at 249, 88 S.Ct.1923 (quoting Schempp, 374 U.S., at 306,83 S.Ct. 1560).

S 877After Everson and Allen, the state ofthe law applying the Establishment Clauseto public expenditures producing somebenefit to religious schools was this:

1. Government aid to religion is forbid-den, and tax revenue may not be used tosupport a religious school or religiousteaching.2. Government provision of such para-digms of universally general welfarebenefits as police and fire protectiondoes not count as aid to religion.3. Whether a law’s benefit is sufficient-ly close to universally general welfareparadigms to be classified with them, asdistinct from religious aid, is a functionof the purpose and effect of the chal-lenged law in all its particularity. Thejudgment is not reducible to the applica-tion of any formula. Evenhandedness ofdistribution as between religious andsecular beneficiaries is a relevant factor,but not a sufficiency test of constitution-ality. There is no rule of religious equalprotection to the effect that any expendi-ture for the benefit of religious schoolstudents is necessarily constitutional solong as public school pupils are favoredon ostensibly identical terms.4. Government must maintain neutrali-ty as to religion, ‘‘neutrality’’ being aconclusory label for the required posi-tion of government as neither aiding re-ligion nor impeding religious exercise bybelievers. ‘‘Neutrality’’ was not thename of any test to identify permissibleaction, and in particular, was not synon-ymous with evenhandedness in confer-ring benefit on the secular as well as thereligious.

Today, the substantive principle of noaid to religious mission remains the gov-

5. Indeed, two of the dissenters in Allen agreedwith the majority on this method of analysis,asking whether the books at issue were simi-lar enough to fire and police protection. See

392 U.S., at 252, 88 S.Ct. 1923 (Black, J.,dissenting); id., at 272, 88 S.Ct. 1923 (Fortas,J., dissenting).


erning understanding of the EstablishmentClause as applied to public benefits inuringto religious schools. The governing opin-ions on the subject in the 35 years sinceAllen have never challenged this principle.The S 878cases have, however, recognizedthat in actual Establishment Clause litiga-tion over school aid legislation, there is nopure aid to religion and no purely secularwelfare benefit; the effects of the laws fallsomewhere in between, with the judicialtask being to make a realistic allocationbetween the two possibilities. The Court’sdecisions demonstrate its repeated at-tempts to isolate considerations relevant inclassifying particular benefits as betweenthose that do not discernibly support orthreaten support of a school’s religiousmission, and those that cross or threatento cross the line into support for religion.

II

AThe most deceptively familiar of those

considerations is ‘‘neutrality,’’ the presenceor absence of which, in some sense, wehave addressed from the moment of Ever-son itself. I say ‘‘some sense,’’ for we haveused the term in at least three ways in ourcases, and an understanding of the term’sevolution will help to explain the conceptas it is understood today, as well as thelimits of its significance in EstablishmentClause analysis. ‘‘Neutrality’’ has beenemployed as a term to describe the requi-site state of government equipoise betweenthe forbidden encouragement and discour-agement of religion; to characterize a ben-efit or aid as secular; and to indicateevenhandedness in distributing it.

As already mentioned, the Court firstreferred to neutrality in Everson, simplystating that government is required ‘‘to bea neutral’’ among religions and betweenreligion and nonreligion. 330 U.S., at 18,67 S.Ct. 504. Although ‘‘neutral’’ mayhave carried a hint of inaction when weindicated that the First Amendment ‘‘doesnot require the state to be [the] adver-

sary’’ of religious believers, ibid., or to cutoff general government services from reli-gious organizations, Everson provided noexplicit definition of the term or furtherindication of what the government was re-quired to do or not do to be S 879‘‘neutral’’toward religion. In practical terms, ‘‘neu-tral’’ in Everson was simply a term forgovernment in its required median positionbetween aiding and handicapping religion.The second major case on aid to religiousschools, Allen, used ‘‘neutrality’’ to de-scribe an adequate state of balance be-tween government as ally and as adver-sary to religion, see 392 U.S., at 242, 88S.Ct. 1923 (discussing line between ‘‘stateneutrality to religion and state support ofreligion’’). The term was not further de-fined, and a few subsequent school casesused ‘‘neutrality’’ simply to designate therequired relationship to religion, withoutexplaining how to attain it. See, e.g., Til-ton v. Richardson, 403 U.S. 672, 677, 91S.Ct. 2091, 29 L.Ed.2d 790 (1971) (describ-ing cases that ‘‘see[k] to define the bound-aries of the neutral area between [the Reli-gion Clauses] within which the legislaturemay legitimately act’’); Roemer v. Boardof Public Works of Md., 426 U.S. 736, 747,96 S.Ct. 2337, 49 L.Ed.2d 179 (1976) (plu-rality opinion of Blackmun, J.) (‘‘Neutralityis what is required. The State must con-fine itself to secular objectives, and neitheradvance nor impede religious activity. Ofcourse, that principle is more easily statedthan applied’’); see also Committee forPublic Ed. & Religious Liberty v. Nyquist,413 U.S. 756, 782, 93 S.Ct. 2955, 37L.Ed.2d 948 (1973) (describing ‘‘neutralposture’’ toward religion); Roemer, supra,at 745–746, 96 S.Ct. 2337 (opinion of Black-mun, J.) (‘‘The Court has enforced a scru-pulous neutrality by the State, as amongreligions, and also as between religiousand other activities’’); cf. Wolman v.Walter, 433 U.S. 229, 254, 97 S.Ct. 2593, 53L.Ed.2d 714 (1977) (quoting Lemon andnoting difficulty of religious teachers’ re-maining ‘‘ ‘religiously neutral’ ’’).


530 U.S. 881

The Court began to employ ‘‘neutrality’’in a sense different from equipoise, howev-er, as it explicated the distinction between‘‘religious’’ and ‘‘secular’’ benefits to reli-gious schools, the latter being in somecircumstances permissible. See infra, at2581–2590 (discussing considerations).Even though both Everson and Allen hadanticipated some such distinction, neithercase had used the term ‘‘neutral’’ in thisway. In Everson, Justice Black indicatedthat providing S 880police, fire, and similargovernment services to religious institu-tions was permissible, in part because theywere ‘‘so separate and so indisputablymarked off from the religious function.’’330 U.S., at 18, 67 S.Ct. 504. Allen simi-larly focused on the fact that the textbookslent out were ‘‘secular’’ and approved bysecular authorities, 392 U.S., at 245, 88S.Ct. 1923, and assumed that the seculartextbooks and the secular elements of edu-cation they supported were not so inter-twined with religious instruction as ‘‘in fact[to be] instrumental in the teaching ofreligion,’’ id., at 248, 88 S.Ct. 1923. Suchwas the Court’s premise in Lemon forshifting the use of the word ‘‘neutral’’ fromlabeling the required position of the gov-ernment to describing a benefit that wasnonreligious. We spoke of ‘‘[o]ur decisionsfrom Everson to Allen [as] permitt[ing]the States to provide church-relatedschools with secular, neutral, or nonideo-logical services, facilities, or materials,’’403 U.S., at 616, 91 S.Ct. 2105, and there-after, we regularly used ‘‘neutral’’ in thissecond sense of ‘‘secular’’ or ‘‘nonreli-gious.’’ See, e.g., Tilton, supra, at 687–688, 91 S.Ct. 2091 (characterizing subsi-dized teachers in Lemon as ‘‘not necessari-ly religiously neutral,’’ but buildings as‘‘religiously neutral’’); Meek v. Pittenger,421 U.S. 349, 365–366, 95 S.Ct. 1753, 44L.Ed.2d 217 (1975) (describing instruction-al materials as ‘‘ ‘secular, nonideologicaland neutral’ ’’ and ‘‘wholly neutral’’); id.,at 372, 95 S.Ct. 1753 (describing auxiliaryservices as ‘‘religiously neutral’’); Roemer,supra, at 751, 96 S.Ct. 2337 (opinion ofBlackmun, J.) (describing Tilton ’s ap-

proved buildings as ‘‘neutral or nonideolog-ical in nature’’); 426 U.S., at 754, 96 S.Ct.2337 (describing Meek ’s speech and hear-ing services as ‘‘neutral and nonideologi-cal’’); Zobrest v. Catalina Foothills SchoolDist., 509 U.S. 1, 10, 113 S.Ct. 2462, 125L.Ed.2d 1 (1993) (discussing translator as‘‘neutral service’’); Agostini v. Felton, 521U.S. 203, 232, 117 S.Ct. 1997, 138 L.Ed.2d391 (1997) (discussing need to assesswhether nature of aid was ‘‘neutral andnonideological’’); cf. Levitt v. Committeefor Public Ed. & Religious Liberty, 413U.S. 472, 478, 93 S.Ct. 2814, 37 L.Ed.2d736 (1973) (noting that District Court ap-proved testing cost reimbursement as pay-ment for services that were ‘‘ ‘secular, neu-tral, or nonideological’ ’’ in character, citingLemon, 403 U.S., at S 881616, 91 S.Ct. 2105);Wolman, supra, at 242, 97 S.Ct. 2593(quoting Lemon, supra, at 616, 91 S.Ct.2105 (describing permitted services aid as‘‘secular, neutral, or nonideological’’ ).

The shift from equipoise to secular wasnot, however, our last redefinition, for theCourt again transformed the sense of‘‘neutrality’’ in the 1980’s. Reexaminingand reinterpreting Everson and Allen, webegan to use the word ‘‘neutral’’ to mean‘‘evenhanded,’’ in the sense of allocatingaid on some common basis to religious andsecular recipients. Again, neither Eversonnor Allen explicitly used ‘‘neutral’’ in thismanner, but just as the label for equipoisehad lent itself to referring to the secularcharacteristic of what a government mightprovide, it was readily adaptable to refer-ring to the generality of government ser-vices, as in Everson’s paradigms, to whichpermissible benefits were compared.

The increased attention to a notion ofevenhanded distribution was evident inNyquist, where the Court distinguishedthe program under consideration from thegovernment services approved in Allenand Everson, in part because ‘‘the class ofbeneficiaries [in Everson and Allen ] in-cluded all schoolchildren, those in public as


well as those in private schools.’’ 413 U.S.,at 782, n. 38, 91 S.Ct. 2105. Nyquist thenreserved the question whether ‘‘some formof public assistance TTT made availablegenerally without regard to the sectarian-nonsectarian, or public-nonpublic nature ofthe institution benefitted’’ would be per-missible. Id., at 783, n. 38, 93 S.Ct. 2955(citations omitted). Subsequent cases con-tinued the focus on the ‘‘generality’’ of theapproved government services as an im-portant characteristic. Meek, for example,characterized Everson and Allen as ap-proving ‘‘a general program’’ to pay busfares and to lend school books, respective-ly, 421 U.S., at 360, 95 S.Ct. 1753; id., at360, n. 8, 95 S.Ct. 1753 (approving twosimilar ‘‘general program[s]’’ in New Yorkand Pennsylvania), and Wolman uphelddiagnostic services described as ‘‘ ‘generalwelfare services for children,’ ’’ 433 U.S., at243, 97 S.Ct. 2593 (quoting Meek, supra, at371, n. 21, 95 S.Ct. 1753).

S 882Justice Blackmun, writing in Roemer,first called such a ‘‘general’’ or evenhandedprogram ‘‘neutral,’’ in speaking of ‘‘facialneutrality’’ as a relevant consideration indetermining whether there was an Estab-lishment Clause violation. ‘‘[R]eligious in-stitutions need not be quarantined frompublic benefits that are neutrally availableto all.’’ 426 U.S., at 746–747, 96 S.Ct.2337; see also id., at 746, 96 S.Ct. 2337(discussing buses in Everson and schoolbooks in Allen as examples of ‘‘neutrallyavailable’’ aid). In Mueller v. Allen, 463U.S. 388, 103 S.Ct. 3062, 77 L.Ed.2d 721(1983), the Court adopted the redefinitionof neutrality as evenhandedness, citingNyquist, 413 U.S., at 782, n. 38, and allud-ing to our discussion of equal access inWidmar v. Vincent, 454 U.S. 263, 102 S.Ct.269, 70 L.Ed.2d 440 (1981). The Courtupheld a system of tax deductions for sec-tarian educational expenses, in part be-cause such a ‘‘facially neutral law,’’ 463U.S., at 401, 103 S.Ct. 3062, made thededuction available for ‘‘all parents, includ-ing those whose children attend publicschools and those whose children attend

nonsectarian private schools or sectarianprivate schools,’’ id., at 397, 103 S.Ct. 3062.Subsequent cases carried the point for-ward. See, e.g., Witters v. WashingtonDept. of Servs. for Blind, 474 U.S. 481,487, 106 S.Ct. 748, 88 L.Ed.2d 846 (1986)(quoting Nyquist and characterizing pro-gram as making aid ‘‘available generally’’);Zobrest, supra, at 8–9, 113 S.Ct. 2462 (dis-cussing ‘‘government programs that neu-trally provide benefits to a broad class ofcitizens defined without reference to reli-gion’’ and citing Mueller and Witters );Agostini, supra, at 231, 117 S.Ct. 1997(discussing aid allocated on the basis of‘‘neutral, secular criteria that neither favornor disfavor religion, TTT made available toboth religious and secular beneficiaries ona nondiscriminatory basis’’); see also Ro-senberger, 515 U.S., at 839, 115 S.Ct. 2510(‘‘[T]he guarantee of neutrality is respect-ed, not offended, when the government,following neutral criteria and evenhandedpolicies, extends benefits to recipientswhose ideologies and viewpoints, includingreligious ones, are broad and diverse’’).

In sum, ‘‘neutrality’’ originally en-tered this field of jurisprudence as aconclusory term, a label for the re-quired relaStionship883 between the gov-ernment and religion as a state ofequipoise between government as allyand government as adversary. Reex-amining Everson’s paradigm cases toderive a prescriptive guideline, wefirst determined that ‘‘neutral’’ aid wassecular, nonideological, or unrelated toreligious education. Our subsequentreexamination of Everson and Allen,beginning in Nyquist and culminatingin Mueller and most recently in Agos-tini, recast neutrality as a concept of‘‘evenhandedness.’’

There is, of course, good reason for con-sidering the generality of aid and the ev-enhandedness of its distribution in makingclose calls between benefits that in pur-pose or effect support a school’s religiousmission and those that do not. This isjust what Everson did. Even when thedisputed practice falls short of Everson’sparadigms, the breadth of evenhanded dis-


530 U.S. 885

tribution is one pointer toward the law’spurpose, since on the face of it aid distrib-uted generally and without a religious cri-terion is less likely to be meant to aidreligion than a benefit going only to reli-gious institutions or people. And, depend-ing on the breadth of distribution, lookingto evenhandedness is a way of askingwhether a benefit can reasonably be seento aid religion in fact; we do not regardthe postal system as aiding religion, eventhough parochial schools get mail. Giventhe legitimacy of considering evenhanded-ness, then, there is no reason to avoid theterm ‘‘neutrality’’ to refer to it. But onecrucial point must be borne in mind.

In the days when ‘‘neutral’’ was used inEverson’s sense of equipoise, neutralitywas tantamount to constitutionality; theterm was conclusory, but when it applied itmeant that the government’s position wasconstitutional under the EstablishmentClause. This is not so at all, however,under the most recent use of ‘‘neutrality’’to refer to generality or evenhandedness ofdistribution. This kind of neutrality isrelevant in judging whether a benefitscheme so characterized should be seen asaiding a sectarian school’s religiousS 884mission, but this neutrality is not alonesufficient to qualify the aid as constitution-al. It is to be considered only along withother characteristics of aid, its administra-tion, its recipients, or its potential thathave been emphasized over the years asindicators of just how religious the intentand effect of a given aid scheme really is.See, e.g., Tilton, 403 U.S., at 677–678, 91S.Ct. 2091 (opinion of Burger, C.J.) (ac-knowledging ‘‘no single constitutional cali-per’’); Meek, 421 U.S., at 358–359, 95 S.Ct.1753 (noting considerations as guidelinesonly and discussing them as a matter ofdegree); School Dist. of Grand Rapids v.Ball, 473 U.S. 373, 383, 105 S.Ct. 3216, 87L.Ed.2d 267 (1985) (quoting Meek ), over-ruled in part by Agostini, 521 U.S., at 203,117 S.Ct. 1997; Board of Ed. of KiryasJoel Village School Dist. v. Grumet, 512U.S. 687, 720, 114 S.Ct. 2481, 129 L.Ed.2d

546 (1994) (opinion of O’CONNOR, J.)(‘‘Experience proves that the Establish-ment Clause, like the Free Speech Clause,cannot easily be reduced to a single test’’);Rosenberger, 515 U.S., at 847–849, 115S.Ct. 2510 (O’CONNOR, J., concurring)(discussing need for line-drawing); id., at852, 115 S.Ct. 2510 (noting lack of a single‘‘Grand Unified Theory’’ for EstablishmentClause and citing Kiryas Joel ); cf. Agos-tini, supra, at 232–233, 117 S.Ct. 1997(examining a variety of factors). Thus, thebasic principle of establishment scrutiny ofaid remains the principle as stated in Ev-erson, that there may be no public aid toreligion or support for the religious mis-sion of any institution.

BThe insufficiency of evenhandedness

neutrality as a stand-alone criterion of con-stitutional intent or effect has been clearfrom the beginning of our interpretativeefforts, for an obvious reason. Evenhand-edness in distributing a benefit approachesthe equivalence of constitutionality in thisarea only when the term refers to suchuniversality of distribution that it makesno sense to think of the benefit as going toany discrete group. Conversely, when ev-enhandedness refers to distribution to lim-ited groups within society, like groups ofschools or schoolchildren, it does makesense to regard the benefit as aid to therecipients. See, e.g., Everson, 330 U.S.,S 885at 16, 67 S.Ct. 504 (discussing aid thatapproaches the ‘‘verge’’ of forbidden terri-tory); Lemon, 403 U.S., at 612, 91 S.Ct.2105 (‘‘[W]e can only dimly perceive thelines of demarcation in this extraordinarilysensitive area of constitutional law’’); Ny-quist, 413 U.S., at 760–761, 93 S.Ct. 2955(noting the ‘‘most perplexing questions’’presented in this area and acknowledging‘‘ ‘entangl[ing] precedents’ ’’); Mueller, 463U.S., at 393, 103 S.Ct. 3062 (quoting Lem-on ); Witters, 474 U.S., at 485, 106 S.Ct.748 (quoting Lemon).

Hence, if we looked no further thanevenhandedness, and failed to ask whatactivities the aid might support, or in fact


did support, religious schools could beblessed with government funding as mas-sive as expenditures made for the benefitof their public school counterparts, andreligious missions would thrive on publicmoney. This is why the consideration ofless than universal neutrality has neverbeen recognized as dispositive and has al-ways been teamed with attention to otherfacts bearing on the substantive prohibi-tion of support for a school’s religious ob-jective.

At least three main lines of enquiryaddressed particularly to school aid haveemerged to complement evenhandednessneutrality. First, we have noted that twotypes of aid recipients heighten Establish-ment Clause concern: pervasively religiousschools and primary and secondary reli-gious schools. Second, we have identifiedtwo important characteristics of the meth-od of distributing aid: directness or indi-rectness of distribution and distribution bygenuinely independent choice. Third, wehave found relevance in at least five char-acteristics of the aid itself: its religiouscontent; its cash form; its divertibility oractually diversion to religious support; itssupplantation of traditional items of reli-gious school expense; and its substantiali-ty.

1Two types of school aid recipients have

raised special concern. First, we have rec-

ognized the fact that the overriding reli-gious mission of certain schools, thosesometimes called S 886‘‘pervasively sectari-an,’’ is not confined to a discrete elementof the curriculum, Everson, 330 U.S., at22–24, 67 S.Ct. 504 (Jackson, J., dissent-ing); id., at 45–47, 67 S.Ct. 504 (Rutledge,J., dissenting), but permeates their teach-ing.6 Walz v. Tax Comm’n of City of NewYork, 397 U.S. 664, 671, 90 S.Ct. 1409, 25L.Ed.2d 697 (1970); Lemon, supra, at636–637, 91 S.Ct. 2105 (‘‘A school whichoperates to commingle religion with otherinstruction plainly cannot completely secu-larize its instruction. Parochial schools, inlarge measure, do not accept the assump-tion that secular subjects should be unre-lated to religious teaching’’); see also Bow-en v. Kendrick, 487 U.S. 589, 621–622, 108S.Ct. 2562, 101 L.Ed.2d 520 (1988) (dis-cussing pervasively sectarian privateschools). Based on record evidence andlong experience, we have concluded thatreligious teaching in such schools is at thecore of the instructors’ individual and per-sonal obligations, cf. Canon 803, § 2, Text& Commentary 568 (‘‘It is necessary thatthe formation and education given in aCatholic school be based upon the princi-ples of Catholic doctrine; teachers are tobe outstanding for their correct doctrineand integrity of life’’), and that individualreligious teachers will teach religiously.7

Lemon, 403 U.S., at 615–620, 91 S.Ct.

6. In fact, religious education in Roman Cath-olic schools is defined as part of requiredreligious practice; aiding it is thus akin toaiding a church service. See 1983 Code ofCanon Law, Canon 798, reprinted in TheCode of Canon Law: A Text and Commentary566 (1985) (hereinafter Text & Commentary)(directing parents to entrust children to Ro-man Catholic schools or otherwise provide forRoman Catholic education); Canon 800, § 2,Text & Commentary 567 (requiring the faith-ful to support establishment and maintenanceof Roman Catholic schools); Canons 802,804, Text & Commentary 567, 568 (requiringdiocesan bishop to establish and regulateschools ‘‘imparting an education imbued withthe Christian spirit’’).

7. Although the Court no longer assumes thatpublic school teachers assigned to religiousschools for limited purposes will teach reli-giously, see Agostini v. Felton, 521 U.S. 203,223–228, 117 S.Ct. 1997, 138 L.Ed.2d 391(1997), we have never abandoned the pre-sumption that religious teachers will teachjust that way. Lemon v. Kurtzman, 403 U.S.602, 615–620, 91 S.Ct. 2105, 29 L.Ed.2d 745(1971); id., at 635–641, 91 S.Ct. 2105 (Doug-las, J., concurring); Levitt v. Committee forPublic Ed. & Religious Liberty, 413 U.S. 472,480, 93 S.Ct. 2814, 37 L.Ed.2d 736 (1973);Meek v. Pittenger, 421 U.S. 349, 369–371, 95S.Ct. 1753, 44 L.Ed.2d 217 (1975); Wolmanv. Walter, 433 U.S. 229, 249–250, 97 S.Ct.2593, 53 L.Ed.2d 714 (1977); School Dist. ofGrand Rapids v. Ball, 473 U.S. 373, 399–400,105 S.Ct. 3216, 87 L.Ed.2d 267 (1985)


530 U.S. 888

2105; id., at 635–S 887641, 91 S.Ct. 2105(Douglas, J., concurring); Levitt, 413 U.S.,at 480, 93 S.Ct. 2814; Meek, 421 U.S., at369–371, 95 S.Ct. 1753; Wolman, 433 U.S.,at 249–250, 97 S.Ct. 2593 (discussing non-severability of religious and secular edu-cation); Ball, 473 U.S., at 399–400, 105S.Ct. 3216 (O’CONNOR, J., concurring injudgment in part and dissenting in part),overruled in part by Agostini, 521 U.S., at236, 117 S.Ct. 1997. As religious teachingcannot be separated from secular edu-cation in such schools or by such teachers,we have concluded that direct governmentsubsidies to such schools are prohibitedbecause they will inevitably and impermis-sibly support religious indoctrination. Zo-brest, 509 U.S., at 12, 113 S.Ct. 2462 (dis-cussing Meek and Ball ).

Second, we have expressed special con-cern about aid to primary and secondaryreligious schools. Tilton, 403 U.S., at 685–686, 91 S.Ct. 2091. On the one hand, wehave understood how the youth of the stu-dents in such schools makes them highlysusceptible to religious indoctrination.Lemon, supra, at 616, 91 S.Ct. 2105 (‘‘Thisprocess of inculcating religious doctrine is,of course, enhanced by the impressionableage of the pupils, in primary schools par-ticularly’’). On the other, we have recog-nized that the religious element in theeducation offered in most sectarian pri-mary and secondary schools is far moreintertwined with the secular than in uni-versity teaching, where the natural andacademic skepticism of most older stu-dents may separate the two, see Tilton,

supra, at 686–689, 91 S.Ct. 2091; Roemer,426 U.S., at 750, 96 S.Ct. 2337. Thus,government benefits accruing to these per-vasively religious primary and secondaryschools raise special dangers of diversioninto support for the religious indoctrina-tion of children and the involvement ofgovernment in religious training and prac-tice.

S 8882We have also evaluated the portent of

support to an organization’s religious mis-sion that may be inherent in the methodby which aid is granted, finding pertinencein at least two characteristics of distribu-tion. First, we have asked whether aid isdirect or indirect, observing distinctionsbetween government schemes with individ-ual beneficiaries and those whose benefi-ciaries in the first instance might be reli-gious schools. Everson, supra, at 18, 67S.Ct. 504 (bus fare supports parents andnot schools); Allen, 392 U.S., at 243–244,and n. 6, 88 S.Ct. 1923 (textbooks go tobenefit children and parents, not schools);Lemon, supra, at 621, 91 S.Ct. 2105 (invali-dating direct aid to schools); Levitt, supra,at 480, 482, 93 S.Ct. 2814 (invalidatingdirect testing aid to schools); Witters, 474U.S., at 487–488, 106 S.Ct. 748 (evaluatingwhether aid was a direct subsidy toschools). Direct aid obviously raisesgreater risks, although recent cases havediscounted this risk factor, looking to otherfeatures of the distribution mechanism.Agostini, supra, at 225–226, 117 S.Ct.1997.8

(O’CONNOR, J., concurring in judgment inpart and dissenting in part), overruled in partby Agostini, supra, at 236, 117 S.Ct. 1997. Cf.NLRB v. Catholic Bishop of Chicago, 440 U.S.490, 504, 99 S.Ct. 1313, 59 L.Ed.2d 533(1979) (‘‘The church-teacher relationship in achurch-operated school differs from the em-ployment relationship in a public or othernonreligious school’’).

8. In Agostini, the Court indicated that ‘‘wehave departed from the rule relied on in Ballthat all government aid that directly assiststhe educational function of religious schoolsis invalid,’’ 521 U.S., at 225, 117 S.Ct. 1997,

and cited Witters v. Washington Dept. of Servs.for Blind, 474 U.S. 481, 106 S.Ct. 748, 88L.Ed.2d 846 (1986), and Zobrest v. CatalinaFoothills School Dist., 509 U.S. 1, 113 S.Ct.2462, 125 L.Ed.2d 1 (1993). However, Agos-tini did not rely on this dictum, instead clear-ly stating that ‘‘[w]hile it is true that individu-al students may not directly apply for Title Iservices, it does not follow from this premisethat those services are distributed ‘directly tothe religious schools.’ In fact, they are not.No Title I funds ever reach the coffers ofreligious schools, and Title I services may notbe provided to religious schools on a school-


S 889Second, we have distinguished be-tween indirect aid that reaches religiousschools only incidentally as a result ofnumerous individual choices and aid that isin reality directed to religious schools bythe government or in practical terms se-lected by religious schools themselves.Mueller, 463 U.S., at 399, 103 S.Ct. 3062;Witters, supra, at 488, 106 S.Ct. 748; Zo-brest, supra, at 10, 113 S.Ct. 2462. Inthese cases, we have declared the constitu-tionality of programs providing aid directlyto parents or students as tax deductions orscholarship money, where such aid maypay for education at some sectarian insti-tutions, Mueller, supra, at 399, 103 S.Ct.3062; Witters, 474 U.S., at 488, 106 S.Ct.748, but only as the result of ‘‘genuinelyindependent and private choices of aid re-cipients,’’ id., at 487, 106 S.Ct. 748. Wedistinguished this path of aid from theroute in Ball and Wolman, where the opin-ions indicated that ‘‘[w]here TTT no mean-ingful distinction can be made between aidto the student and aid to the school, theconcept of a loan to individuals is a trans-parent fiction.’’ 474 U.S., at 487, n. 4, 106S.Ct. 748 (citations and internal quotationmarks omitted).9

3In addition to the character of the school

to which the benefit accrues, and its pathfrom government to school, a number offeatures of the aid itself have figured inthe classificaStions890 we have made. First,we have barred aid with actual religiouscontent, which would obviously run afoul ofthe ban on the government’s participationin religion, Everson, 330 U.S., at 16, 67S.Ct. 504; Walz, 397 U.S., at 668, 90 S.Ct.1409; cf. Lemon, 403 U.S., at 617, 91S.Ct. 2105 (discussing variable ideologicaland religious character of religious teach-ers compared to fixed content of books).In cases where we have permitted aid, wehave regularly characterized it as ‘‘neu-tral’’ in the sense (noted supra, at 2578–2579) of being without religious content.See, e.g., Tilton, 403 U.S., at 688, 91 S.Ct.2091 (characterizing buildings as ‘‘reli-giously neutral’’); Zobrest, 509 U.S., at 10,113 S.Ct. 2462 (describing translator as‘‘neutral service’’); Agostini, 521 U.S., at232, 117 S.Ct. 1997 (discussing need toassess whether nature of aid was ‘‘neutraland nonideological’’). See also ante, at2547 (plurality opinion) (barring aid withreligious content).10

wide basis.’’ 521 U.S., at 228–229, 117 S.Ct.1997 (citations omitted). Until today, thisCourt has never permitted aid to go directlyto schools on a schoolwide basis.

The plurality misreads our precedent insuggesting that we have abandoned directnessof distribution as a relevant consideration.See ante, at 2544–2546. In Wolman, we stat-ed that nominally describing aid as to stu-dents would not bar a court from finding thatit actually provided a subsidy to a school, 433U.S., at 250, 97 S.Ct. 2593, but we did notestablish that a program giving ‘‘direct’’ aidto schools was therefore permissible. In Wit-ters, we made the focus of Wolman clear,continuing to examine aid to determine if itwas a ‘‘direct subsidy’’ to a school, 474 U.S.,at 487, 106 S.Ct. 748, and distinguishing theaid at issue from impermissible aid in Balland Wolman precisely because the designa-tion of the student as recipient in those caseswas only nominal. 474 U.S., at 487, n. 4, 106S.Ct. 748. Our subsequent cases have contin-ued to ask whether government aid programsconstituted impermissible ‘‘direct subsidies’’

to religious schools even where they are di-rected by individual choice. Zobrest, supra, at11–13, 113 S.Ct. 2462; Mueller v. Allen, 463U.S. 388, 399, 103 S.Ct. 3062, 77 L.Ed.2d 721(1983); Agostini, supra, at 226, 117 S.Ct.1997.

9. We have also permitted the government tosupply students with public-employee transla-tors, Zobrest, supra, at 10, 113 S.Ct. 2462, andpublic-employee special education teachers,Agostini, supra, at 226, 228, 117 S.Ct. 1997,who directly provided them with governmentservices in whatever schools those specificstudents attended, public or nonpublic. Ihave already noted Agostini’s limitations.See n. 8, supra.

10. I agree with the plurality that the Estab-lishment Clause absolutely prohibits the gov-ernment from providing aid with clear reli-gious content to religious, or for that matternonreligious, schools. Ante, at 2547–2549.The plurality, however, misreads our prece-dent as focusing only on affirmatively reli-gious content. At the very least, a building,


530 U.S. 891

Second, we have long held governmentaid invalid when circumstances would al-low its diversion to religious education.The risk of diversion is obviously highwhen aid in the form of government fundsmakes its way into the coffers of religiousorganizations, and so from the start wehave understood the Constitution to baroutright money grants of aid to religion.11

See Everson, 330 U.S., at 16, 67 S.Ct. 504(‘‘[The State] S 891cannot consistently withthe ‘establishment of religion’ clause of theFirst Amendment contribute tax-raisedfunds to the support of an institutionwhich teaches the tenets and faith of anychurch’’); id., at 18, 67 S.Ct. 504 (‘‘TheState contributes no money to the schools.It does not support them’’); Allen, 392U.S., at 243–244, 88 S.Ct. 1923 (‘‘[N]ofunds or books are furnished to parochialschools, and the financial benefit is to par-ents and children, not schools’’); Walz,supra, at 675, 90 S.Ct. 1409 (‘‘Obviously adirect money subsidy would be a relation-

ship pregnant with involvement and, aswith most governmental grant programs,could encompass sustained and detailedadministrative relationships for enforce-ment of statutory or administrative stan-dards’’); Lemon, supra, at 612, 91 S.Ct.2105 (identifying ‘‘three main evils’’against which Establishment Clause wasto protect as ‘‘sponsorship, financial sup-port, and active involvement of the sover-eign in religious activity,’’ citing Walz );403 U.S., at 621, 91 S.Ct. 2105 (distin-guishing direct financial aid program fromEverson and Allen and noting problemswith required future surveillance); Ny-quist, 413 U.S., at 762, 774, 93 S.Ct. 2955(striking down ‘‘direct money grants’’ formaintaining buildings because there wasno attempt to restrict payments to thoseexpenditures related exclusively to secularpurposes); Levitt, 413 U.S., at 480, 482, 93S.Ct. 2814 (striking down ‘‘direct moneygrant’’ for testing expenses); 12 Hunt v.

for example, has no such content, but wehave squarely required the government to en-sure that no publicly financed building bediverted to religious use. Tilton v. Richard-son, 403 U.S. 672, 681–684, 91 S.Ct. 2091, 29L.Ed.2d 790 (1971). See also Bowen v. Ken-drick, 487 U.S. 589, 623, 108 S.Ct. 2562, 101L.Ed.2d 520 (1988) (O’CONNOR, J., concur-ring) (‘‘[A]ny use of public funds to promotereligious doctrines violates the EstablishmentClause’’).

11. We have similarly noted that paying sala-ries of parochial school teachers creates toomuch of a risk that such support will aid theteaching of religion, striking down such pro-grams because of the need for pervasive mon-itoring that would be required. See Lemon,403 U.S., at 619, 91 S.Ct. 2105 (‘‘We do notassume, however, that parochial school teach-ers will be unsuccessful in their attempts tosegregate their religious beliefs from theirsecular educational responsibilities. But thepotential for impermissible fostering of reli-gion is present. The [state legislature] hasnot, and could not, provide state aid on thebasis of a mere assumption that secular teach-ers under religious discipline can avoid con-flicts. The State must be certain, given theReligion Clauses, that subsidized teachers donot inculcate religionTTTT A comprehensive,discriminating, and continuing state surveil-lance will inevitably be required to ensure

that these restrictions are obeyed and theFirst Amendment otherwise respected’’).

12. It is true that we called the importance ofthe cash payment consideration into questionin Committee for Public Ed. and ReligiousLiberty v. Regan, 444 U.S. 646, 657–659, 100S.Ct. 840, 63 L.Ed.2d 94 (1980) (approvingprogram providing religious school with ‘‘di-rect cash reimbursement’’ for expenses ofstandardized testing). In that case, we foundthe other safeguards against the diversion ofsuch funds to religious uses sufficient to allowsuch aid: ‘‘A contrary view would insist ondrawing a constitutional distinction betweenpaying the nonpublic school to do the gradingand paying state employees or some indepen-dent service to perform that task, even thoughthe grading function is the same regardless ofwho performs it and would not have the pri-mary effect of aiding religion whether or notperformed by nonpublic school personnel.’’Id., at 658, 100 S.Ct. 840. Aside from thisisolated circumstance, where we found iron-clad guarantees of nondiversion, we have nev-er relaxed our prohibition on direct cash aidto pervasively religious schools, and have infact continued to acknowledge the concern.See Agostini, 521 U.S., at 228–229, 117 S.Ct.1997; cf. Rosenberger, 515 U.S., at 842, 115S.Ct. 2510.

The plurality concedes this basic point.See ante, at 2546. Given this, I find anysuggestion that this prohibition has been un-


S 892McNair, 413 U.S. 734, 745, n. 7, 93S.Ct. 2868, 37 L.Ed.2d 923 (1973) (notingapproved aid is ‘‘no expenditure of publicfunds, either by grant or loan’’); Wolman,433 U.S., at 239, and n. 7, 97 S.Ct. 2593(noting that ‘‘statute does not authorizeany payment to nonpublic school personnelfor the costs of administering the tests’’);Agostini, 521 U.S., at 228–229, 117 S.Ct.1997 (emphasizing that approved servicesare not ‘‘distributed ‘directly to the reli-gious schools.’ TTT No Title I funds everreach the coffers of religious schools, andTitle I services may not be provided toreligious schools on a schoolwide basis’’(citations omitted)); Bowen, 487 U.S., at614–615, 108 S.Ct. 2562; Rosenberger, 515U.S., at 842, 115 S.Ct. 2510 (noting that‘‘we have recognized special EstablishmentClause dangers where the governmentmakes direct money payments to sectarianinstitutions’’); cf. Lemon, 403 U.S., at619–620, 91 S.Ct. 2105 (noting that safe-guards and accounting inspections re-quired to prevent government funds fromsupporting religious education will causeimpermissible entanglement); Roemer,426 U.S., at 753–757, 96 S.Ct. 2337 (ap-proving segregated funds after finding re-cipients not pervasively religious); Ball,473 U.S., at 392–393, 105 S.Ct. 3248 (not-ing that ‘‘[w]ith but one exception, oursubsequent cases have struck down at-tempts by States to make payments out ofS 893public tax dollars directly to primary or

secondary religious educational institu-tions’’), overruled in part by Agostini, su-pra, at 236, 117 S.Ct. 1997; Witters, 474U.S., at 487, 106 S.Ct. 748 (‘‘It is equallywell-settled TTT that the State may notgrant aid to a religious school, whethercash or in kind, where the effect of the aidis that of a direct subsidy to the religiousschool’’ (internal quotation marks and cita-tions omitted)); Rosenberger, supra, at851–852, 115 S.Ct. 2510 (O’CONNOR, J.,concurring) (noting that student fee wasnot a tax).

Divertibility is not, of course, a charac-teristic of cash alone, and when examiningprovisions for ostensibly secular supplieswe have considered their susceptibility tothe service of religious ends.13 In uphold-ing a scheme to provide students withsecular textbooks, we emphasized that‘‘each book loaned must be approved bythe public school authorities; only secularbooks may receive approval.’’ Allen, 392U.S., at 244–245, 88 S.Ct. 1923; see alsoMeek, 421 U.S., at 361–362, 95 S.Ct. 1753(opinion of Stewart, J.); Wolman, supra,at 237–238, 97 S.Ct. 2593. By the sametoken, we could not sustain provisions forinstructional materials adaptable to teach-ing a variety of subjects.14 Meek, supra,at 363, 95 S.Ct. 1753; Wolman, supra, at249–250, 97 S.Ct. 2593. While the text-books had a known and fixed secular con-tent not readily diSvertible894 to religious

dermined by Mueller or Witters without foun-dation. See ante, at 2546–2547, n. 8. Thosecases involved entirely different types of aid,namely, tax deductions and individual schol-arship aid for university education, see also n.16, infra, and were followed by Rosenbergerand Agostini, which continued to support thisabsolute restriction.

13. I reject the plurality’s argument that diver-tibility is a boundless principle. Ante, at2549. Our long experience of evaluating thisconsideration demonstrates its practical lim-its. See infra this page and 2587. Moreover,the Establishment Clause charges us withmaking such enquiries, regardless of their dif-ficulty. See supra, at 2576–2578, 2581–2582.Finally, the First Amendment’s rule permit-ting only aid with fixed secular content seems

no more difficult to apply than the plurality’srule prohibiting only aid with fixed religiouscontent.

14. Contrary to the plurality’s apparent belief,Lamb’s Chapel v. Center Moriches Union FreeSchool Dist., 508 U.S. 384, 113 S.Ct. 2141,124 L.Ed.2d 352 (1993), sheds no light on thequestion of divertibility and school aid. Ante,at 2548, n. 9. The Court in that case clearlydistinguished the question of afterschool ac-cess to public facilities from anything resem-bling the school aid cases: ‘‘The showing ofthis film series would not have been duringschool hours, would not have been sponsoredby the school, and would have been open tothe public, not just to church members.’’ 508U.S., at 395, 113 S.Ct. 2141.


530 U.S. 895

teaching purposes, the adaptable materialsdid not.15 So, too, we explained the per-missibility of busing on public routes toschools but not busing for field trips de-signed by religious authorities specificallybecause the latter trips were componentsof teaching in a pervasively religiousschool. Compare Everson, 330 U.S., at 17,67 S.Ct. 504 (noting wholly separate andsecular nature of public bus fare toschools), with Wolman, 433 U.S., at 254, 97S.Ct. 2593 (‘‘The field trips are an integralpart of the educational experience, andwhere the teacher works within and for asectarian institution, an unacceptable riskof fostering of religion is an inevitablebyproduct’’ (citation omitted)). We like-wise were able to uphold underwriting theexpenses of standard state testing in reli-gious schools while being forced to strikedown aid for testing designed by theschool officials, because the latter testscould be used to reinforce religious teach-ing. Compare id., at 240, 97 S.Ct. 2593(‘‘[T]he State provides both the schoolsand the school district with the means ofensuring that the minimum standards aremet. The nonpublic school does not con-trol the content of the test or its result.This serves to prevent the use of the testas part of religious teaching, and thusavoids that kind of direct aid to religionfound present in Levitt ’’); Committee forPublic Ed. and Religious Liberty v. Re-gan, 444 U.S. 646, 661–662, 100 S.Ct. 840,63 L.Ed.2d 94 (1980) (same), with Levitt,413 U.S., at 480, 93 S.Ct. 2814 (‘‘We cannotignore the substantial risk that these ex-aminations, prepared by teachers underthe authority of religious institutions, willbe drafted with an eye, unconsciously or

otherwise, to inculcate students in the reli-gious precepts of the sponsoring church’’).

S 895With the same point in mind, we heldthat buildings constructed with govern-ment grants to universities with religiousaffiliation must be barred from religioususe indefinitely to prevent the diversion ofgovernment funds to religious objectives.Tilton, 403 U.S., at 683, 91 S.Ct. 2091(plurality opinion) (‘‘If, at the end of 20years, the building is, for example, con-verted into a chapel or otherwise used topromote religious interests, the originalfederal grant will in part have the effect ofadvancing religion. To this extent the Acttherefore trespasses on the ReligionClauses’’); see also Hunt, 413 U.S., at 743–744, 93 S.Ct. 2868. We were accordinglyconstrained to strike down aid for repair-ing buildings of nonpublic schools becausethey could be used for religious education.Nyquist, 413 U.S., at 776–777, 93 S.Ct.2955.

Divertibility was, again, the issue in anorder remanding an as-applied challengeto a grant supporting counseling on teen-age sexuality for findings that the aid hadnot been used to support religious edu-cation. Bowen, 487 U.S., at 621, 108 S.Ct.2562; see also id., at 623, 108 S.Ct. 2562(O’CONNOR, J., concurring). And themost recent example of attention to thesignificance of divertibility occurred in ourexplanation that public school teacherscould be assigned to provide limited in-struction in religious schools in Agostini,521 U.S., at 223–227, 117 S.Ct. 1997, amajority of the Court rejecting the factualassumption that public school teacherscould be readily lured into providing reli-gious instruction.16

15. In Lemon, we also specifically examinedthe risk that a government program that paidreligious teachers would support religious ed-ucation; the teachers posed the risk of beingunable to separate secular from religious edu-cation. Although we invalidated the programon entanglement grounds, we suggested thatthe monitoring the State had established inthat case was actually required to eliminatethe risk of diversion. See 403 U.S., at 619, 91S.Ct. 2105; see also n. 11, supra.

16. The plurality is mistaken in its reading ofZobrest. See ante, at 2547. Zobrest does notreject the principle of divertibility. There thegovernment provided only a translator whowas not considered divertible because he didnot add to or subtract from the religious mes-sage. The Court approved the translator as itwould approve a hearing aid, health services,diagnostics, and tests. See Zobrest, 509 U.S.,at 13, and n. 10, 113 S.Ct. 2462. Cf. Brad-


S 896Third, our cases have recognized thedistinction, adopted by statute in theChapter 2 legislation, between aid thatmerely supplements and aid that supplantsexpenditures for offerings at religiousschools, the latter being barred. Althoughwe have never adopted the position thatany benefit that flows to a religious schoolis impermissible because it frees up re-sources for the school to engage in reli-gious indoctrination, Hunt, supra, at 743,93 S.Ct. 2868, from our first decision hold-ing it permissible to provide textbooks forreligious schools we have repeatedly ex-plained the unconstitutionality of aid thatsupplants an item of the school’s tradition-al expense. See, e.g., Cochran v. Louisi-ana Bd. of Ed., 281 U.S. 370, 375, 50 S.Ct.335, 74 L.Ed. 913 (1930) (noting that reli-gious schools ‘‘are not the beneficiaries ofthese appropriations. They obtain nothingfrom them, nor are they relieved of asingle obligation because of them’’ (inter-

nal quotation marks omitted)); Everson,330 U.S., at 18, 67 S.Ct. 504 (specificallynoting that bus fare program did not sup-port or fund religious schools); Allen, 392U.S., at 244, 88 S.Ct. 1923 (stating that‘‘the financial benefit [of providing thetextbooks] is to parents and children, notto schools’’ (footnote omitted)); id., S 897at244, n. 6, 88 S.Ct. 1923 (explicitly recogniz-ing that ‘‘the record contains no evidencethat any of the private schools in appel-lants’ districts previously provided text-books for their students’’); Lemon, 403U.S., at 656, 91 S.Ct. 2105 (opinion ofBrennan, J.) (noting no aid to schools wasinvolved in Allen ). We ignored this prohi-bition only once, in Regan, 444 U.S., at646, 100 S.Ct. 840; see also ante, at 2544,n. 7, where reimbursement for budgetedexpenses of required testing was notstruck down, but we then quickly returnedto the rule as a guideline for permissibleaid.17 In Zobrest, 509 U.S., at 12, 113

field v. Roberts, 175 U.S. 291, 299–300, 20S.Ct. 121, 44 L.Ed. 168 (1899); Wolman, 433U.S., at 244, 97 S.Ct. 2593. Zobrest thus canbe thought of as akin to our approval ofdiagnostic services in Wolman, supra, at 244,97 S.Ct. 2593, which we considered to have‘‘little or no educational content[,] not [to be]closely associated with the educational mis-sion of the nonpublic school,’’ and not to pose‘‘an impermissible risk of the fostering ofideological views.’’ The fact that the dissentsaw things otherwise (as the plurality pointsout, ante, at 2547–2548) is beside the pointhere.

Similarly, the plurality is mistaken in read-ing our holdings in Mueller and Witters, seeante, at 2547–2548, to undermine divertibilityas a relevant principle. First, these casesapproved quite factually distinct types of aid;Mueller involving tax deductions, which havea quite separate history of approval, see 463U.S., at 396, and nn. 5, 6, 103 S.Ct. 2841(citing Walz v. Tax Comm’n of City of NewYork, 397 U.S. 664, 90 S.Ct. 1409, 25 L.Ed.2d697 (1970)), and Witters involving scholarshipmoney distributed to a university, not a pri-mary or secondary school, see Tilton, 403U.S., at 685–686, 91 S.Ct. 2091, that was notsignificant enough as a whole to support thatinstitution, Witters, 474 U.S., at 488, 106 S.Ct.748. Second, in neither case did the programat issue provide direct aid on a schoolwidebasis (as Chapter 2 does here); in both wefound a distinction based on the genuinely

independent, private choices which allocatedsuch very different types of aid (tax deduc-tions and university scholarship money thatdid not amount to substantial support of theuniversity). See Mueller, supra, at 399, 103S.Ct. 3062; Witters, supra, at 488, 106 S.Ct.748.

17. Our departure from this principle in Reganis not easily explained, but it is an isolatedholding surrounded by otherwise unbrokenadherence to the no-supplanting principle.Long after Regan we have continued to findthe supplement/supplant distinction, like thebar to substantial aid, to be an importantconsideration. See Zobrest, supra, at 12, 113S.Ct. 2462; Agostini, 521 U.S., at 228, 117S.Ct. 1997; cf. Witters, supra, at 487–488,106 S.Ct. 748 (discussing rule against ‘‘directsubsidy’’). The weight that the pluralityplaces on Regan is thus too much for it tobear. See ante, at 2544, n. 7. Moreover, theapparent object of the Regan Court’s concernwas vindicating the principle that aid withfixed secular content was permissible, distin-guishing it from the divertible testing aid inLevitt. Regan, 444 U.S., at 661–662, 100 S.Ct.840 (citing Wolman, supra, at 263, 97 S.Ct.2593); cf. Levitt, 413 U.S., at 480, 93 S.Ct.2814. The plurality provides no explanationfor our continued reference to the principle ofno-supplanting aid in subsequent cases, suchas Zobrest and Agostini, which it finds trust-worthy guides elsewhere in its discussion of


530 U.S. 899

S.Ct. 2462, the Court specifically distin-guished Meek and Ball by explaining thatthe invalid programs in those cases ‘‘re-lieved sectarian schools of costs they oth-erwise would have borne in educating theirstudents.’’ In Agostini, the Court made apoint of noting that the objects of the aidwere ‘‘by law supplemental to the regularcurricula’’ and, citing Zobrest, explainedthat the remedial education services didnot relieve the religious schools of coststhey would otherwise have borne. 521U.S., at 228, 117 S.Ct. 1997 (citing Zobrest,supra, at 12, 113 S.Ct. 2462). The Courtexplicitly stated that the S 898services inquestion did not ‘‘supplant the remedialinstruction and guidance counseling al-ready provided in New York City’s sectari-an schools.’’ 521 U.S., at 229, 117 S.Ct.1997.

Finally, we have recognized what is ob-vious (however imprecise), in holding ‘‘sub-stantial’’ amounts of aid to be unconstitu-tional whether or not a plaintiff can showthat it supplants a specific item of expensea religious school would have borne.18 InMeek, 421 U.S., at 366, 95 S.Ct. 1753, weinvalidated the loan of instructional mate-rials to religious schools because ‘‘facedwith the substantial amounts of direct sup-port authorized by [the program], it wouldsimply ignore reality to attempt to sepa-rate secular educational functions from thepredominantly religious role performed bymany of Pennsylvania’s church-related ele-mentary and secondary schools and thencharacterize [the program] as channelingaid to the secular without providing direct

aid to the sectarian.’’ Id., at 365, 95 S.Ct.1753. See id., at 366, 95 S.Ct. 1753 (‘‘Sub-stantial aid to the educational function ofsuch schools TTT necessarily results in aidto the sectarian school enterprise as awhole’’); see also Nyquist, 413 U.S., at783, 93 S.Ct. 2955; Wolman, 433 U.S., at250–251, 97 S.Ct. 2593. In Witters, 474U.S., at 488, 106 S.Ct. 748, the Court askedwhether the aid in question was a directsubsidy to religious schools and addressedthe substantiality of the aid obliquely innoting that ‘‘nothing in the record indi-cates that TTT any significant portion ofthe S 899aid expended under the Washingtonprogram as a whole will end up flowing toreligious education.’’ In Zobrest, supra, at12, 113 S.Ct. 2462, the Court spoke of thesubstantiality test in Meek, noting that‘‘[d]isabled children, not sectarian schools,are the primary beneficiaries of the [Indi-viduals with Disabilities Act (IDEA)]; tothe extent sectarian schools benefit at allfrom the IDEA, they are only incidentalbeneficiaries.’’

C

This stretch of doctrinal history leavesone point clear beyond peradventure: to-gether with James Madison we have con-sistently understood the EstablishmentClause to impose a substantive prohibitionagainst public aid to religion and, hence, tothe religious mission of sectarian schools.Evenhandedness neutrality is one, nondis-positive pointer toward an intent and (to alesser degree) probable effect on the per-

the First Amendment. See ante, at 2548–2549, 2549–2550, 2550–2551, 2552–2554.Nor does the plurality explain why it placesso much weight on Regan’s apparent depar-ture from the no-supplanting rule while itignores Regan’s core reasoning that the test-ing aid there was permissible because, in di-rect contrast to Levitt, the aid was not diverti-ble.

18. I do not read the plurality to question theprohibition on substantial aid. The pluralitychallenges any rule based on the proportionof aid that a program provides to religiousrecipients, citing Witters and Agostini. See

ante, at 2542–2543, n. 6. I reject the plurali-ty’s reasoning. The plurality misreads Wit-ters; Justice Marshall, writing for the Court inWitters, emphasized that only a small amountof aid was provided to religious institutions,474 U.S., at 488, 106 S.Ct. 748, and no con-trolling majority rejected the importance ofthis fact. The plurality also overreads Agosti-ni, supra, at 229, 117 S.Ct. 1997, which sim-ply declined to adopt a rule based on propor-tionality. Moreover, regardless of whetherthe proportion of aid actually provided toreligious schools is relevant, we have neverquestioned our holding in Meek that substan-tial aid to religious schools is prohibited.


missible side of the line between forbiddenaid and general public welfare benefit.Other pointers are facts about the reli-gious mission and education level of bene-fited schools and their pupils, the pathwayby which a benefit travels from publictreasury to educational effect, the formand content of the aid, its adaptability toreligious ends, and its effects on schoolbudgets. The object of all enquiries intosuch matters is the same whatever theparticular circumstances: is the benefit in-tended to aid in providing the religiouselement of the education and is it likely todo so?

The substance of the law has thus notchanged since Everson. Emphasis on onesort of fact or another has varied depend-ing on the perceived utility of the enquiry,but all that has been added is repeatedexplanation of relevant considerations, con-firming that our predecessors were rightin their prophecies that no simple testwould emerge to allow easy application ofthe establishment principle.

The plurality, however, would reject thatlesson. The majority misapplies it.

S 900III

A

The nub of the plurality’s new position isthis:

‘‘[I]f the government, seeking to furthersome legitimate secular purpose, offersaid on the same terms, without regardto religion, to all who adequately furtherthat purpose, then it is fair to say thatany aid going to a religious recipientonly has the effect of furthering thatsecular purpose. The government, incrafting such an aid program, has had toconclude that a given level of aid isnecessary to further that purpose amongsecular recipients and has provided nomore than that same level to religiousrecipients.’’ Ante, at 2541 (citation omit-ted).

As a break with consistent doctrine theplurality’s new criterion is unequaled inthe history of Establishment Clause inter-pretation. Simple on its face, it appears totake evenhandedness neutrality and inpractical terms promote it to a single andsufficient test for the establishment consti-tutionality of school aid. Even on its ownterms, its errors are manifold, and atten-tion to at least three of its mistaken as-sumptions will show the degree to whichthe plurality’s proposal would replace theprinciple of no aid with a formula for gen-erous religious support.

First, the plurality treats an externalobserver’s attribution of religious supportto the government as the sole impermissi-ble effect of a government aid scheme.See, e.g., ante, at 2541 (‘‘[N]o one wouldconclude that any indoctrination that anyparticular recipient conducts has beendone at the behest of the government’’).While perceived state endorsement of reli-gion is undoubtedly a relevant concern un-der the Establishment Clause, see, e.g.,Allegheny County, 492 U.S., at 592–594,109 S.Ct. 3086; see also Capitol SquareReview and Advisory Bd. v. Pinette, 515U.S. 753, 772–774, 115 S.Ct. 2440, 132L.Ed.2d 650 (1995) (O’CONNOR, J., con-curring in part and concurring inS 901judgment); id., at 786–787, 115 S.Ct.2440 (SOUTER, J., concurring in part andconcurring in judgment), it is certainly notthe only one. Everson made this clearfrom the start: secret aid to religion bythe government is also barred. 330 U.S.,at 16, 67 S.Ct. 504. State aid not attrib-uted to the government would still violatea taxpayer’s liberty of conscience, threatento corrupt religion, and generate disputesover aid. In any event, since the same-terms feature of the scheme would, on theplurality’s view, rule out the attribution orperception of endorsement, adopting theplurality’s rule of facial evenhandednesswould convert neutrality into a dispositivecriterion of establishment constitutionalityand eliminate the effects enquiry directedby Allen, Lemon, and other cases. Under


530 U.S. 903

the plurality’s rule of neutrality, if a pro-gram met the first part of the Lemonenquiry, by declining to define a program’srecipients by religion, it would automati-cally satisfy the second, in supposedly hav-ing no impermissible effect of aiding reli-gion.19

Second, the plurality apparently as-sumes as a fact that equal amounts of aidto religious and nonreligious schools willhave exclusively secular and equal effects,on both external perception and on incen-tives to attend different schools. See ante,at 2541, 2543–2544. But there is no reasonto believe that this will be the case; theeffects of same-terms aid may not be con-fined to the secular sphere at all. This isthe reason that we have long recognizedthat unrestricted aid to religious schoolswill support religious teaching in adSdi-tion902 to secular education, a fact thatwould be true no matter what the suppos-edly secular purpose of the law might be.

Third, the plurality assumes that percapita distribution rules safeguard thesame principles as independent, privatechoices. But that is clearly not so. Weapproved university scholarships in Wit-ters because we found them close to givinga government employee a paycheck andallowing him to spend it as he chose, but aper capita aid program is a far cry fromawarding scholarships to individuals, oneof whom makes an independent privatechoice. Not the least of the significantdifferences between per capita aid and aidindividually determined and directed is theright and genuine opportunity of the recip-ient to choose not to give the aid.20 To

hold otherwise would be to license thegovernment to donate funds to churchesbased on the number of their members, onthe patent fiction of independent privatechoice.

The plurality’s mistaken assumptions ex-plain and underscore its sharp break withthe Framers’ understanding of establish-ment and this Court’s consistent interpre-tative course. Under the plurality’s re-gime, little would be left of the right ofconscience against compelled support forreligion; the more massive the aid themore potent would be the influence of thegovernment on the teaching mission; themore generous the support, the more divi-sive would be the resentments of thoseresisting religious support, and those reli-gions without school systems ready toclaim their fair share.

B

The plurality’s conception of evenhand-edness does not, however, control the case,whose disposition turns on the misapplica-tion of accepted categories of school aidanalysis. The facts most obviously rele-vant to the Chapter 2 scheme S 903in Jeffer-son Parish are those showing divertibilityand actual diversion in the circumstance ofpervasively sectarian religious schools.The type of aid, the structure of the pro-gram, and the lack of effective safeguardsclearly demonstrate the divertibility of theaid. While little is known about its use,owing to the anemic enforcement systemin the parish, even the thin record beforeus reveals that actual diversion occurred.

19. Adopting the plurality’s rule would permitpractically any government aid to religion solong as it could be supplied on terms ostensi-bly comparable to the terms under which aidwas provided to nonreligious recipients. As aprinciple of constitutional sufficiency, the ma-nipulability of this rule is breathtaking. Alegislature would merely need to state a secu-lar objective in order to legalize massive aidto all religions, one religion, or even one sect,to which its largess could be directed throughthe easy exercise of crafting facially neutral

terms under which to offer aid favoring thatreligious group. Short of formally replacingthe Establishment Clause, a more dependablekey to the public fisc or a cleaner break withprior law would be difficult to imagine.

20. Indeed, the opportunity for an individualto choose not to have her religious schoolreceive government aid is just what at leastone of the respondents seeks here. See Brieffor Respondents 1, and n. 1.


The aid that the government providedwas highly susceptible to unconstitutionaluse. Much of the equipment provided un-der Chapter 2 was not of the type providedfor individual students, App. to Pet. forCert. 140a; App. 262a–278a, but included‘‘slide projectors, movie projectors, over-head projectors, television sets, tape re-corders, projection screens, maps, globes,filmstrips, cassettes, computers,’’ and com-puter software and peripherals, Helms v.Cody, No. 85–5533, 1990 WL 36124(E.D.La., Mar.27, 1990); App. to Pet. forCert. 140a; App. 90a, 262a–278a, as well aslibrary books and materials, id., at 56a,126a, 280a–284a. The videocassette play-ers, overhead projectors, and other in-structional aids were of the sort that wehave found can easily be used by religiousteachers for religious purposes. Meek, 421U.S., at 363, 95 S.Ct. 1753; Wolman, 433U.S., at 249–250, 97 S.Ct. 2593. The samewas true of the computers, which were asreadily employable for religious teachingas the other equipment, and presumably asimmune to any countervailing safeguard,App. 90a, 118a, 164a–165a. Although li-brary books, like textbooks, have fixedcontent, religious teachers can assign secu-lar library books for religious critique, andbooks for libraries may be religious, as anydivinity school library would demonstrate.The sheer number and variety of booksthat could be and were ordered gave am-ple opportunity for such diversion.

The divertibility thus inherent in theforms of Chapter 2 aid was enhanced by

the structure of the program in JeffersonParish. Requests for specific items underChapter 2 came not from secular officials,cf. Allen, 392 U.S., at 244–S 904245, 88 S.Ct.1923, but from officials of the religiousschools (and even parents of religiousschool pupils), see ante, at 2537–2538 (not-ing that private religious schools submittedtheir orders to the government for specificrequested items); App. 156a–158a. Thesectarian schools decided what they want-ed and often ordered the supplies, id., at156a–159a, 171a–172a, to be forwarded di-rectly to themselves, id., at 156a–159a. Itwas easy to select whatever instructionalmaterials and library books the schoolswanted, just as it was easy to employcomputers for the support of the religiouscontent of the curriculum infused with reli-gious instruction.

The concern with divertibility thus pred-icated is underscored by the fact that thereligious schools in question here coveredthe primary and secondary grades, thegrades in which the sectarian nature ofinstruction is characteristically the mostpervasive, see Lemon, 403 U.S., at 616, 91S.Ct. 2105; cf. Tilton, 403 U.S., at 686–689, 91 S.Ct. 2091, and in which pupils arethe least critical of the schools’ religiousobjectives, see Lemon, supra, at 616, 91S.Ct. 2105. No one, indeed, disputes thetrial judge’s findings, based on a detailedrecord, that the Roman Catholic schools,21

which made up the majority of the privateschools participating,22 were pervasivelysectarian,23 that S 905their common objective

21. Litigation, discovery, and the opinions be-low focused almost exclusively on the aid tothe 34 Roman Catholic schools. Consequent-ly, I will confine my discussion to that infor-mation. Of course, the same concerns wouldbe raised by government aid to religiousschools of other faiths that a court found hadsimilar missions of religious education andreligious teachers teaching religiously.

22. The Jefferson Parish Chapter 2 programincluded 46 nonpublic schools, of which 41were religiously affiliated. Thirty-four ofthese were Roman Catholic, seven otherswere religiously affiliated, and five were not

religiously affiliated. App. to Pet. for Cert.143a–144a.

23. The trial judge found that the Roman Cath-olic schools in question operate under thegeneral supervision and authority of the Arch-bishop of New Orleans and their parish pas-tors, and are located next to parish churchesand sometimes a rectory or convent. Id., at144a. The schools include religious symbolsin their classrooms, App. 75a, require attend-ance at daily religion classes, id., at 76a, con-duct sacramental preparation classes duringthe schoolday, require attendance at mass,and provide extracurricular religious activi-ties. At least some exercise a religious prefer-


530 U.S. 906

and mission was to engage in religiouseducation,24 and that their teachers taughtreligiously,25 making them precisely thekind of primary and S 906secondary religiousschools that raise the most serious Estab-lishment Clause concerns. See Walz, 397U.S., at 671, 90 S.Ct. 1409; Hunt, 413U.S., at 743, 93 S.Ct. 2868; Lemon, supra,at 636–637, 91 S.Ct. 2105. The threat toEstablishment Clause values was accord-ingly at its highest in the circumstances ofthis case. Such precautionary features asthere were in the Jefferson Parish schemewere grossly inadequate to counter thethreat. To be sure, the disbursement ofthe aid was subject to statutory admoni-tions against diversion, see, e.g., 20 U.S.C.§§ 7332, 8897, and was supposedly subjectto a variety of safeguards, see ante, at2537–2538, 2553–2555. But the provisionsfor onsite monitoring visits, labeling ofgovernment property, and government

oversight cannot be accepted as sufficientin the face of record evidence that thesafeguard provisions proved to be emptyphrases in Jefferson Parish. Cf. Agostini,521 U.S., at 228–229, 117 S.Ct. 1997; Zo-brest, 509 U.S., at 13, 113 S.Ct. 2462 (ac-cepting precautionary provisions in ab-sence of evidence of their uselessness).

The plurality has already noted atlength the ineffectiveness of the govern-ment’s monitoring program. Ante, at2553–2555; see also App. 111a (‘‘A systemto monitor nonpublic schools was often notin operation and therefore the [local edu-cational agency] did not always know: (a)what was purchased or (b) how it wasutilized’’). Monitors visited a nonpublicschool only sporadically, discussed the pro-gram with a single contact person, ob-served nothing more than attempts atrecordkeeping, and failed to inform the

ence in accepting students and in chargingtuition. App. to Pet. for Cert. 145a.

24. The District Court found that the missionof the Roman Catholic schools is religiouseducation based on the Archdiocese’s and theindividual schools’ published statements ofphilosophy. For example, the St. AnthonySchool Handbook, cited by the District Court,reads:

‘‘Catholic education is intended to makemen’s faith become living, conscious and ac-tive through the light of instruction. TheCatholic school is the unique setting withinwhich this ideal can be realized in the lives ofthe Catholic children and young people.‘‘Only in such a school can they experiencelearning and living fully integrated in the lightof faithTTTT Here, too, instruction in religioustruth and values is an integral part of theschool program. It is not one more subjectalong side the rest, but instead it is perceivedand functions as the underlying reality inwhich the student’s experiences of learningand living achieve their coherence and theirdeepest meaning.’’ Ibid.

The Handbook of Policies and Regulationsfor Elementary Schools of the Archdiocese ofNew Orleans indicates that the operation ofthe Roman Catholic schools is governed bycanon law. It also lists the major objectivesof those schools as follows:

‘‘To work closely with the home in educat-ing children towards the fullness of Christianlife.

‘‘To specifically teach Catholic principles andChristian values.’’ Id., at 146a.The mission statements and objectives out-lined by the other Roman Catholic schoolsalso support the conclusion that these institu-tions’ primary objective is religious instruc-tion. See also App. 65a, 71a.

25. The Archdiocese’s official policy calls forreligious preferences in hiring and the con-tracts of principals and teachers in its schoolscontain a provision allowing for terminationfor lifestyle contrary to the teachings of theRoman Catholic church. App. to Pet. forCert. 145a. One of the objectives of the hand-book is ‘‘[t]o encourage teachers to becomecommitted Christians and to develop profes-sional competence.’’ Id., at 146a. Other rec-ord evidence supports the conclusion thatthese religious schoolteachers teach religious-ly. See, e.g., App. 125a (deposition of presi-dent of sectarian high school) (‘‘Our teachers,whether they are religion teachers or not, arecertainly instructed that when issues come upin the classroom that have a religious, moral,or value concept, that their answers be con-sistent with the teachings of the CatholicChurch and that they respond in that way tothe students, so that there can be opportuni-ties in other classes other than religion wherediscussion of religio[n] could take place, yes,sir’’); id., at 73a, 74a.


teachers of the restrictions involved. Id.,at 154a–155a. Although Chapter 2 re-quired labeling of government property, itoccurred haphazardly at best, id., at 113a,and the government’s sole monitoring sys-tem for computer use amounted to nothingmore S 907than questioning school officialsand examining the location of computers atthe schools, id., at 118a. No records ofsoftware and computer use were kept, andno such recordkeeping was even planned.Id., at 118a, 164a–166a. State and localofficials in Jefferson Parish admitted thatnothing prevented the Chapter 2 comput-ers from being used for religious instruc-tion, id., at 102a, 118a, 164a–166a, andalthough they knew of methods of monitor-ing computer usage, such as locking thecomputer functions, id., at 165a–166a, theyimplemented no particular policies, insti-tuted no systems, and employed no tech-nologies to minimize the likelihood of di-version to religious uses,26 id., at 118a,165a–166a. The watchdogs did require thereligious schools to give not so much as anassurance that they would use Chapter 2computers solely for secular purposes,Helms v. Picard, 151 F.3d 347, 368 (1998),amended, 165 F.3d 311 (C.A.5 1999); App.94a–95a. Government officials themselvesadmitted that there was no way to tellwhether instructional materials had beendiverted, id., at 118a, 139a, 144a–145a, and,as the plurality notes, the only screeningmechanism in the library book scheme wasa review of titles by a single governmentofficial, ante, at 2554, n. 15; see App. 137a.The government did not even have a policyon the consequences of noncompliance.Id., at 145a.

The risk of immediate diversion ofChapter 2 benefits had its complement inthe risk of future diversion, against whichthe Jefferson Parish program had abso-lutely no protection. By statute all pur-chases with Chapter 2 aid were to remain

the property of the United States, 20U.S.C. § 7372(c)(1), merely being ‘‘lent’’ tothe recipient nonpublic schools. In actual-ity, however, the record indicates thatnothing in the S 908Jefferson Parish programstood in the way of giving the Chapter 2property outright to the religious schoolswhen it became older. Although oldequipment remained the property of thelocal education agency, a local governmentadministrative body, one agency employeetestified that there was no set policy fordealing with old computers, which wereprobably given outright to the religiousschools. App. 161a–162a. The witnesssaid that government-funded instructionalmaterials, too, were probably left with thereligious schools when they were old, andthat it was unclear whether library bookswere ever to be returned to the govern-ment. Ibid.

Providing such governmental aid with-out effective safeguards against future di-version itself offends the EstablishmentClause, Tilton, 403 U.S., at 682–684, 91S.Ct. 2091; Nyquist, 413 U.S., at 776–777,93 S.Ct. 2955, and even without evidence ofactual diversion, our cases have repeatedlyheld that a ‘‘substantial risk’’ of it sufficesto invalidate a government aid program onestablishment grounds. See, e.g., Wol-man, 433 U.S., at 254, 97 S.Ct. 2593 (inval-idating aid for transportation on teacher-accompanied field trips because an ‘‘unac-ceptable risk of fostering of religion’’ was‘‘an inevitable byproduct’’); Meek, 421U.S., at 372, 95 S.Ct. 1753 (striking downprogram because of a ‘‘potential for imper-missible fostering of religion’’); Levitt, 413U.S., at 480, 93 S.Ct. 2814 (invalidating aidfor tests designed by religious teachersbecause of ‘‘the substantial risk that TTT

examinations, prepared by teachers underthe authority of religious institutions, willbe drafted with an eye, unconsciously orotherwise, to inculcate students in the reli-

26. The Government’s reliance on U.S. Depart-ment of Education Guidance for Title VI ofthe Elementary and Secondary Education Act(Feb.1999) is misplaced. See App. to Brief

for Secretary of Education 1a. It was not inplace when discovery closed in this matter,and merely highlights the reasons for a lackof evidence on diversion or compliance.


530 U.S. 910

gious precepts of the sponsoring church’’);Lemon, 403 U.S., at 619, 91 S.Ct. 2105(finding invalid aid with a ‘‘potential forimpermissible fostering of religion’’); cf.Bowen, 487 U.S., at 621, 108 S.Ct. 2562(noting that where diversion risk is lessclearly made out, a case may be remandedfor findings on actual diversion of aid toreligious indoctrination); Regan, 444 U.S.,at 656, 100 S.Ct. 840 (characterizing as‘‘minimal’’ the chance that state-draftedtests with ‘‘complete’’ safeguards would beadopted to reliSgious909 testing). A sub-stantial risk of diversion in this case wasmore than clear, as the plurality has con-ceded. The First Amendment was violat-ed.

But the record here goes beyond risk, toinstances of actual diversion. What onewould expect from such paltry efforts atmonitoring and enforcement naturally re-sulted, and the record strongly suggeststhat other, undocumented diversions prob-ably occurred as well. First, the recordshows actual diversion in the library bookprogram. App. 132a–133a. Although onlylimited evidence exists, it contrasts starklywith the records of the numerous textbookprograms that we have repeatedly upheld,where there was no evidence of any actualdiversion. See Allen, 392 U.S., at 244–245,88 S.Ct. 1923; Meek, supra, at 361–362, 95S.Ct. 1753; Wolman, supra, at 237–238, 97S.Ct. 2593. Here, discovery revealed thatunder Chapter 2, nonpublic schools re-quested and the government purchased atleast 191 religious books with taxpayer

funds by December 1985.27 App. 133a.Books such as A Child’s Book of Prayers,id., at 84a, and The Illustrated Life ofJesus, id., at 132a, S 910were discoveredamong others that had been ordered underthe program. See also id., at 59a–62a.

The evidence persuasively suggests thatother aid was actually diverted as well.The principal of one religious school testi-fied, for example, that computers lent withChapter 2 funds were joined in a networkwith other non-Chapter 2 computers insome schools, and that religious officialsand teachers were allowed to develop theirown unregulated software for use on thisnetwork. Id., at 77a. She admitted thatthe Chapter 2 computer took over thesupport of the computing system wheneverthere was a breakdown of the master com-puter purchased with the religious school’sown funds. Ibid. Moreover, as the plu-rality observes, ante, at 2554, n. 17, com-paring the records of considerable federalfunding of audiovisual equipment in reli-gious schools with records of the schools’use of unidentified audiovisual equipmentin religion classes strongly suggests thatfilm projectors and videotape machinespurchased with public funds were used inreligious indoctrination over a period of atleast seven years. App. 205a, 210a, 206a–207a; see also id., at 108a (statement ofsecond-grade teacher indicating that sheused audiovisual materials in all classes).

27. The plurality applies inconsistent stan-dards to the evidence. Although the pluralityfinds more limited evidence of actual diver-sion sufficient to support a general finding ofdiversion in the computer and instructionalmaterials context, even in the face of JusticeO’CONNOR’s objections, it fails to find a vio-lation of the prohibition against providing aidwith religious content based on the morestark, undisputed evidence of religious books.Compare ante, at 2553–2555, and nn. 14–17,with ante, at 2554–2555. As a matter of pre-cedent, the correct evidentiary standard isclearly the former: ‘‘[A]ny use of public fundsto promote religious doctrines violates theEstablishment Clause.’’ Bowen, 487 U.S., at

623, 108 S.Ct. 2562 (O’CONNOR, J., concur-ring). We have never before found any actualdiversion or allowed a risk of it; we havestruck down policies that might permit it, e.g.,Tilton, 403 U.S., at 682–684, 91 S.Ct. 2091, orhave remanded for specific factual findingsabout whether diversion occurred, Bowen, su-pra, at 621, 108 S.Ct. 2562. See supra, at2584–2587. As a matter of principle, this lowthreshold is required to safeguard the valuesof the First Amendment. Madison’s wordsmake clear that even a small infringement ofthe prohibition on compelled aid to religion isodious to the freedom of conscience. No lessdoes it open the door to the threat of corrup-tion or to a return to religious conflict.


Indeed, the plurality readily recognizesthat the aid in question here was divertibleand that substantial evidence of actual di-version exists. Ante, at 2553–2555, andnn. 14–17. Although Justice O’CONNORattributes limited significance to the evi-dence of divertibility and actual diversion,she also recognizes that it exists. Ante, at2570–2571 (opinion concurring in judg-ment). The Court has no choice but tohold that the program as applied violatedthe Establishment Clause.28

S 911IV

The plurality would break with the law.The majority misapplies it. That misappli-cation is, however, the only consolation inthe case, which reaches an erroneous re-sult but does not stage a doctrinal coup.But there is no mistaking the abandon-ment of doctrine that would occur if theplurality were to become a majority. It isbeyond question that the plurality’s notionof evenhandedness neutrality as a practicalguarantee of the validity of aid to sectarianschools would be the end of the principle ofno aid to the schools’ religious mission.

And if that were not so obvious it wouldbecome so after reflecting on the plurali-ty’s thoughts about diversion S 912and aboutgiving attention to the pervasiveness of aschool’s sectarian teaching.

The plurality is candid in pointing outthe extent of actual diversion of Chapter 2aid to religious use in the case before us,ante, at 2553–2554, and n. 17, and equallycandid in saying it does not matter, ante,at 2547–2549, 2554–2555. To the pluralitythere is nothing wrong with aiding aschool’s religious mission; the only ques-tion is whether religious teaching obtainsits tax support under a formally evenhand-ed criterion of distribution. The principleof no aid to religious teaching has no inde-pendent significance.

And if this were not enough to provethat no aid in religious school aid is deadunder the plurality’s First Amendment,the point is nailed down in the plurality’sattack on the legitimacy of considering aschool’s pervasively sectarian characterwhen judging whether aid to the school islikely to aid its religious mission. Ante, at2550–2552. The relevance of this consider-ation is simply a matter of common sense:

28. Since the divertibility and diversion re-quire a finding of unconstitutionality, I willnot explore other grounds, beyond notingthe likelihood that unconstitutional supplan-tation occurred as well. The record demon-strates that Chapter 2 aid impermissibly re-lieved religious schools of some costs thatthey otherwise would have borne, and so un-constitutionally supplanted support in somebudgetary categories. The record of affida-vits and evaluation forms by religiousschoolteachers and officials indicates thatChapter 2 aid was significant in the develop-ment of teaching curriculums, the introduc-tion of new programs, and the support ofold ones. App. 105a–108a, 184a–185a. Theevidence shows that the concept of supple-menting instead of supplanting was poorlyunderstood by the sole government officialadministering the program, who apparentlybelieved that the bar on supplanting wasnothing more than a prohibition on payingfor replacements of equipment that religiousschools had previously purchased. Id., at167a. Government officials admitted thatthere was no way to determine whether pay-ments for materials, equipment, books, or

other assistance provided under the programreduced the amount of money budgeted forlibrary and educational equipment, id., at145a–146a, and the 1985 Monitoring Reportshows that the officials of at least one reli-gious school admitted that the governmentaid was used to create the library, with theschool’s regular funds, when occasionallyavailable, used merely to supplement thegovernment money, Fine Deposition, id., at63a. The use records for audiovisual mate-rials at one religious high school revealedthat Chapter 2 funds were essential to theschool’s educational process, id., at 187a,and a different school, as already noted,used a Chapter 2 computer to support itscomputer network when its own computersfailed, id., at 77a. The record is sparse, butthese incidents suggest that the constitution-al and statutory prohibition on supplantingexpenses may have been largely aspirational.It seems that the program in Jefferson Par-ish violated the statute and ran afoul of theConstitution. Cf. Nyquist, 413 U.S., at 783,93 S.Ct. 2955; Zobrest, 509 U.S., at 12, 113S.Ct. 2462.

2597STENBERG v. CARHARTCite as 120 S.Ct. 2597 (2000)

530 U.S. 914

where religious indoctrination pervadesschool activities of children and adoles-cents, it takes great care to be able to aidthe school without supporting the doctrinaleffort. This is obvious. The pluralitynonetheless condemns any enquiry into thepervasiveness of doctrinal content as aremnant of anti-Catholic bigotry (as ifevangelical Protestant schools and Ortho-dox Jewish yeshivas were never pervasive-ly sectarian 29), and it equates a refusal toaid religious schools with hostility to reli-gion (as if aid to religious teaching werenot S 913opposed in this very case by at leastone religious respondent 30 and numerousreligious amici curiae 31 in a traditionclaiming descent from Roger Williams).My concern with these arguments goes notso much to their details 32 as it does to thefact that the plurality’s choice to employimputations of bigotry and irreligion asterms in the Court’s debate makes onepoint clear: that in rejecting the principleof no aid to a school’s religious mission theplurality is attacking the most fundamentalassumption underlying the EstablishmentClause, that government can in fact oper-ate with neutrality in its relation to reli-

gion. I believe that it can, and so respect-fully dissent.

,

530 U.S. 914, 147 L.Ed.2d 743

S 914Don STENBERG, Attorney Generalof Nebraska, et al., Petitioners,

v.

Leroy CARHART.No. 99–830.

Argued April 25, 2000.

Decided June 28, 2000.

Physician who performed abortionsbrought suit on behalf of himself and hispatients challenging constitutionality ofNebraska statute banning ‘‘partial birthabortion.’’ The United States DistrictCourt for the District of Nebraska, Rich-ard G. Kopf, J., 972 F.Supp. 507, held

29. Indeed, one group of amici curiae, whichconsists of ‘‘religious and educational leadersfrom a broad range of both Eastern and West-ern religious traditions, and Methodist, Jew-ish and Seventh-day Adventist individuals’’including ‘‘church administrators, adminis-trators of religious elementary and secondaryschool systems; elementary and secondaryschool teachers at religious schools; and pas-tors and laity who serve on church schoolboards,’’ identifies its members as having‘‘broad experience teaching in and adminis-tering pervasively sectarian schools.’’ Brieffor Interfaith Religious Liberty Foundation etal. as Amici Curiae 1.

30. One of the respondents describes herself asa ‘‘life-long, committed member of the RomanCatholic Church’’ who ‘‘objects to the govern-ment providing benefits to her parish school’’because ‘‘[s]he has seen the chilling effectsuch entangling government aid has on thereligious mission of schools run by herchurch.’’ Brief for Respondents 1. She hasbeen a member of the church for about 36years, and six of her children attended differ-

ent Jefferson Parish Catholic run schools.Id., at 1, n. 1, 113 S.Ct. 2462.

31. E.g., Brief for Baptist Joint Committee onPublic Affairs as Amicus Curiae; Brief forInterfaith Religious Liberty Foundation et al.as Amici Curiae; Brief for National Commit-tee for Public Education and Religious Liber-ty et al. as Amici Curiae.

32. I do not think it worthwhile to comment atlength, for example, on the plurality’s clearmisunderstanding of our access-to-public-fo-rum cases, such as Lamb’s Chapel and Wid-mar v. Vincent, 454 U.S. 263, 102 S.Ct. 269,70 L.Ed.2d 440 (1981), as ‘‘decisions thathave prohibited governments from discrimi-nating in the distribution of public benefitsbased on religious status or sincerity,’’ ante, at2551, when they were decided on completelydifferent and narrowly limited free-speechgrounds. Nor would it be worthwhile here toengage in extended discussion of why the goalof preventing courts from having to ‘‘trol[l]through a person’s or institution’s religiousbeliefs,’’ ibid., calls for less aid and comming-ling of government with religion, not for tol-erance of their effects.

2530 120 supreme court reporter 530 u.s. 791 120 supreme court reporter 530 u.s. 791 ... in...

Documents