with whom The Chief Justice and Justice Thomas join, dissenting.
The Court today finds that the Powers That Be, up in Albany, have conspired to effect an establishment of the Safmar Hasidim. I do not know who would be more surprised at this discovery: the Founders of our Nation or Grand Rebbe Joel Teitelbaum, founder of the Satmar. The Grand Rebbe would be astounded to learn that after escaping brutal persecution and coming to America with the modest hope of religious toleration for their ascetic form of Judaism, the Satmar had become so powerful, so closely allied with Mammon, as to have become an “establishment” of the Empire State. And the Founding Fathers would be astonished to find that the Establishment Clause — which they designed “to insure that no one powerful sect or combination of sects could use political or governmental power to punish dissenters,” Zorach v. Clauson, 343 U. S. 306, 319 (1952) (Black, J., dissenting) — has been employed to prohibit characteristically and admirably American accommodation of the religious practices (or more precisely, cultural peculiarities) of a tiny minority sect. J, however, am not surprised. Once this Court has abandoned text and history as guides, nothing prevents it from calling religious toleration the establishment of religion.
I
Unlike most of our Establishment Clause cases involving education, these cases involve no public funding, however slight or indirect, to private religious schools. They do not involve private schools at all. The school under scrutiny is a public school specifically designed to provide a public secular *733education to handicapped students. The superintendent of the school, who is not Hasidic, is a 20-year veteran of the New York City public school system, with expertise in the area of bilingual, bicultural, special education. The teachers and therapists at the school all live outside the village of Kiryas Joel. While the village’s private schools are profoundly religious and strictly segregated by sex, classes at the public school are co-ed and the curriculum secular. The school building has the bland appearance of a public school, unadorned by religious symbols or markings; and the school complies with the laws and regulations governing all other New York State public schools. There is no suggestion, moreover, that this public school has gone too far in making special adjustments to the religious needs of its students. Cf. id., at 312-315 (approving a program permitting early release of public school students to attend religious instruction). In sum, these cases involve only public aid to a school that is public as can be. The only thing distinctive about the school is that all the students share the same religion.
None of our cases has ever suggested that there is anything wrong with that. In fact, the Court has specifically approved the education of students of a single religion on a neutral site adjacent to a private religious school. See Wolman v. Walter, 433 U. S. 229, 247-248 (1977). In that case, the Court rejected the argument that “any program that isolates the sectarian pupils is impermissible,” id., at 246, and held that, “[t]he fact that a unit on a neutral site on occasion may serve only sectarian pupils does not provoke [constitutional] concerns,” id., at 247. And just last Term, the Court held that the State could permit public employees to assist students in a Catholic school. See Zobrest v. Catalina Foothills School Dist., 509 U. S. 1, 13-14 (1993) (sign-language translator for deaf student). If a State can furnish services to a group of sectarian students on a neutral site adjacent to a private religious school, or even within such a school, how can there be any defect in educating those same students in *734a public school? As the Court noted in Wolman, the constitutional dangers of establishment arise “from the nature of the institution, not from the nature of the pupils,” 433 U. S., at 248. There is no danger in educating religious students in a public school.
For these very good reasons, Justice Souter’s opinion does not focus upon the school, but rather upon the school district and the New York Legislature that created it. His arguments, though sometimes intermingled, are two: that reposing governmental power in the Kiryas Joel school district is the same as reposing governmental power in a religious group; and that in enacting the statute creating the district, the New York State Legislature was discriminating on the basis of religion, i. e., favoring the Satmar Hasidim over others. I shall discuss these arguments in turn.
II
For his thesis that New York has unconstitutionally conferred governmental authority upon the Satmar sect, Justice Souter relies extensively, and virtually exclusively, upon Larkin v. Grendel's Den, Inc., 459 U. S. 116 (1982). Justice Souter believes that the present litigation “resembles” GrendeVs Den because that case “teaches that a State may not delegate its civic authority to a group chosen according to a religious criterion,” ante, at 698 (emphasis added). That misdescribes both what that case taught (which is that a State may not delegate its civil authority to a church), and what these cases involve (which is a group chosen according to cultural characteristics). The statute at issue there gave churches veto power over the State’s authority to grant a liquor license to establishments in the vicinity of the church. The Court had little difficulty finding the statute unconstitutional. “The Framers did not set up a system of government in which important, discretionary governmental powers would be delegated to or shared with religious institutions.” 459 U. S., at 127.
*735Justice Souter concedes that GrendeVs Den “presented an example of united civic and religious authority, an establishment rarely found in such straightforward form in modem America.” Ante, at 697. The uniqueness of the case stemmed from the grant of governmental power directly to a religious institution, and the Court’s opinion focused on that fact, remarking that the transfer of authority was to “churches” (10 times), the “governing body of churches” (twice), “religious institutions” (twice), and “religious bodies” (once). Astonishingly, however, Justice Souter dismisses the difference between a transfer of government power to citizens who share a common religion as opposed to “the officers of its sectarian organization” — the critical factor that made GrendeVs Den unique and “rar[e]” — as being “one of form, not substance.” Ante, at 698.
Justice Souter’s steamrolling of the difference between civil authority held by a church and civil authority held by members of a church is breathtaking. To accept it, one must believe that large portions of the civil authority exercised during most of our history were unconstitutional, and that much more of it than merely the Kiryas Joel school district is unconstitutional today. The history of the populating of North America is in no small measure the story of groups of people sharing a common religious and cultural heritage striking out to form their own communities. See, e. g., W. Sweet, The Story of Religion in America 9 (1950). It is preposterous to suggest that the civil institutions of these communities, separate from their churches, were constitutionally suspect. And if they were, surely Justice Souter cannot mean that the inclusion of one or two nonbelievers in the community would have been enough to eliminate the constitutional vice. If the conferral of governmental power upon a religious institution as such (rather than upon American citizens who belong to the religious institution) is not the test of GrendeVs Den invalidity, there is no reason why giving power to a body that is overwhelmingly dominated *736by the members of one sect would not suffice to invoke the Establishment Clause. That might have made the entire States of Utah and New Mexico unconstitutional at the time of their admission to the Union,1 and would undoubtedly make many units of local government unconstitutional today.2
Justice Souter’s position boils down to the quite novel proposition that any group of citizens (say, the residents of Kiryas Joel) can be invested with political power, but not if they all belong to the same religion. Of course such disfavoring of religion is positively antagonistic to the purposes of the Religion Clauses, and we have rejected it before. In McDaniel v. Paty, 435 U. S. 618 (1978), we invalidated a state constitutional amendment that would have permitted all persons to participate in political conventions, except ministers. We adopted James Madison’s view that the State could not “ ‘punis[h] a religious profession with the privation of a civil right.’ ” Id., at 626 (opinion of Burger, C. J.), quoting 5 Writings of James Madison 288 (G. Hunt ed. 1904). Or as Justice *737Brennan put it in his opinion concurring in judgment: “Religionists no less than members of any other group enjoy the full measure of protection afforded speech, association, and political activity generally.” 435 U. S., at 641; see also Widmar v. Vincent, 454 U. S. 263 (1981). I see no reason why it is any less pernicious to deprive a group rather than an individual of its rights simply because of its religious beliefs.
Perhaps appreciating the startling implications for our constitutional jurisprudence of collapsing the distinction between religious institutions and their members, Justice Souter tries to limit his “unconstitutional conferral of civil authority” holding by pointing out several features supposedly unique to the present cases: that the “boundary lines of the school district divide residents according to religious affiliation,” ante, at 699 (emphasis added); that the school district was created by “a special Act of the legislature,” ante, at 700; and that the formation of the school district ran counter to the legislature’s trend of consolidating districts in recent years, ibid. Assuming all these points to be true (and they are not), they would certainly bear upon whether the legislature had an impermissible religious motivation in creating the district (which is Justice Souter’s next point, in the discussion of which I shall reply to these arguments). But they have nothing to do with whether conferral of power upon a group of citizens can be the conferral of power upon a religious institution. It cannot. Or if it can, our Establishment Clause jurisprudence has been transformed.
Ill
I turn, next, to Justice Souter’s second justification for finding an establishment of religion: his facile conclusion that the New York Legislature’s creation of the Kiryas Joel school district was religiously motivated. But in the Land of the Free, democratically adopted laws are not so easily impeached by unelected judges. To establish the unconstitu*738tionality of a facially neutral law on the mere basis of its asserted religiously preferential (or discriminatory) effects— or at least to establish it in conformity with our precedents— Justice Souter “must be able to show the absence of a neutral, secular basis” for the law. Gillette v. United States, 401 U. S. 437, 452 (1971); see also Arlington Heights v. Metropolitan Housing Development Corp., 429 U. S. 252, 266 (1977) (facially race-neutral laws can be invalidated on the basis of their effects only if “unexplainable on grounds other than race”).
There is of course no possible doubt of a secular basis here. The New York Legislature faced a unique problem in Kiryas Joel: a community in which all the nonhandicapped children attend private schools, and the physically and mentally disabled children who attend public school suffer the additional handicap of cultural distinctiveness. It would be troublesome enough if these peculiarly dressed, handicapped students were sent to the next town, accompanied by their similarly clad but unimpaired classmates. But all the unimpaired children of Kiryas Joel attend private school. The handicapped children suffered sufficient emotional trauma from their predicament that their parents kept them home from school. Surely the legislature could target this problem, and provide a public education for these students, in the same way it addressed, by a similar law, the unique needs of children institutionalized in a hospital. See, e. g., 1970 N. Y. Laws, ch. 843 (authorizing a union free school district for the area owned by Blythedale Children’s Hospital).
Since the obvious presence of a neutral, secular basis renders the asserted preferential effect of this law inadequate to invalidate it, Justice Souter is required to come forward with direct evidence that religious preference was the objective. His case could scarcely be weaker. It consists, briefly, of this: The People of New York created the Kiryas Joel Village School District in order to further the Satmar religion, rather than for any proper secular purpose, because *739(1) they created the district in an extraordinary manner — by special Act of the legislature, rather than under the State’s general laws governing school-district reorganization; (2) the creation of the district ran counter to a state trend toward consolidation of school districts; and (3) the district includes only adherents of the Satmar religion. On this indictment, no jury would convict.
One difficulty with the first point is that it is not true. There was really nothing so “special” about the formation of a school district by an Act of the New York Legislature. The State has created both large school districts, see, e. g., 1972 N. Y. Laws, ch. 928 (creating the Gananda School District out of land previously in two other districts), and small specialized school districts for institutionalized children, see, e. g., 1972 N. Y. Laws, ch. 559 (creating a union free school district for the area owned by Abbott House), through these special Acts. But in any event all that the first point proves, and the second point as well (countering the trend toward consolidation),3 is that New York regarded Kiryas Joel as a *740special case, requiring special measures. I should think it obvious that it did, and obvious that it should have. But even if the New York Legislature had never before created a school district by special statute (which is not true), and even if it had done nothing but consolidate school districts for over a century (which is not true), how could the departure from those past practices possibly demonstrate that the legislature had religious favoritism in mind? It could not. To be sure, when there is no special treatment there is no possibility of religious favoritism; but it is not logical to suggest that when there is special treatment there is proof of religious favoritism.
Justice Souter’s case against the statute comes down to nothing more, therefore, than his third point: the fact that all the residents of the Kiryas Joel Village School District are Satmars. But all its residents also wear unusual dress, have unusual civic customs, and have not much to do with people who are culturally different from them. (The Court recognizes that “the Satmars prefer to live together ‘to facilitate individual religious observance and maintain social, cultural and religious values,’ but that it is not ‘ “against their religion” to interact with others.’” Ante, at 706, n. 9, quoting Brief for Petitioners in No. 93-517, p. 4, n. 1.) On what basis does Justice Souter conclude that it is the theological distinctiveness rather than the cultural distinctiveness that was the basis for New York State’s decision? The normal assumption would be that it was the latter, since it was not theology but dress, language, and cultural alienation that posed the educational problem for the children. Justice Souter not only does not adopt the logical assumption, he does not even give the New York Legislature the benefit of the doubt. The following is the level of his analysis:
“Not even the special needs of the children in this community can explain the legislature’s unusual Act, for the State could have responded to the concerns of the Satmar parents [by other means].” Ante, at 702.
*741In other words, we know the legislature must have been motivated by the desire to favor the Satmar Hasidim religion, because it could have met the needs of these children by a method that did not place the Satmar Hasidim in a separate school district. This is not a rational argument proving religious favoritism; it is rather a novel Establishment Clause principle to the effect that no secular objective may be pursued by a means that might also be used for religious favoritism if some other means is available.
I have little doubt that Justice Souter would laud this humanitarian legislation if all of the distinctiveness of the students of Kiryas Joel were attributable to the fact that their parents were nonreligious commune dwellers, or American Indians, or gypsies. The creation of a special, one-culture school district for the benefit of those children would pose no problem. The neutrality demanded by the Religion Clauses requires the same indulgence towards cultural characteristics that are accompanied by religious belief. “The Establishment Clause does not license government to treat religion and those who teach or practice it, simply by virtue of their status as such, as ... subject to unique disabilities.” McDaniel v. Paty, 435 U. S., at 641 (Brennan, J., concurring in judgment).
Even if Justice Souter could successfully establish that the cultural distinctiveness of the Kiryas Joel students (which is the problem the New York Legislature addressed) was an essential part of their religious belief rather than merely an accompaniment of their religious belief, that would not discharge his heavy burden. In order to invalidate a facially neutral law, Justice Souter would have to show not only that legislators were aware that religion caused the problems addressed, but also that the legislature’s proposed solution was motivated by a desire to disadvantage or benefit a religious group (i. e., to disadvantage or benefit them because of their religion). For example, if the city of Hialeah, knowing of the potential health problems raised by *742the Santería religious practice of animal sacrifice, were to provide by ordinance a special, more frequent, municipal garbage collection for the carcasses of dead animals, we would not strike the ordinance down just because the city council was aware that a religious practice produced the problem the ordinance addressed. See Church of Lukumi Babalu Aye, Inc. v. Hialeah, 508 U. S. 520, 543-545 (1993). Here a facially neutral statute extends an educational benefit to the one area where it was not effectively distributed. Whether or not the reason for the ineffective distribution had anything to do with religion, it is a remarkable stretch to say that the Act was motivated by a desire to favor or disfavor a particular religious group. The proper analogy to Chapter 748 is not the Court’s hypothetical law providing school buses only to Christian students, see ante, at 709, but a law providing extra buses to rural school districts (which happen to be predominantly Southern Baptist).
At various times Justice Souter intimates, though he does not precisely say, that the boundaries of the school district were intentionally drawn on the basis of religion. He refers, for example, to “the State’s manipulation of the franchise for this district. . . , giving the sect exclusive control of the political subdivision,” ante, at 698 — implying that the “giving” of political power to the religious sect was the object of the “manipulation.” There is no evidence of that. The special district was created to meet the special educational needs of distinctive handicapped children, and the geographical boundaries selected for that district were (quite logically) those that already existed for the village. It sometimes appears as though the shady “manipulation” Justice Souter has in mind is that which occurred when the village was formed, so that the drawing of its boundaries infected the coterminous boundaries of the district. He says, for example, that “[i]t is undisputed that those who negotiated the village boundaries when applying the general village incorporation statute drew them so as to exclude *743all but Satmars.” Ante, at 699. It is indeed. But nonSatmars were excluded, not (as he intimates) because of their religion, but — as Justice O’Connor clearly describes, see ante, at 712 — because of their lack of desire for the high-density zoning that Satmars favored. It was a classic drawing of lines on the basis of communality of secular governmental desires, not communality of religion. What happened in the creation of the village is in fact precisely what happened in the creation of the school district, so that the former cannot possibly infect the latter, as Justice Souter tries to suggest. Entirely secular reasons (zoning for the village, cultural alienation of students for the school district) produced a political unit whose members happened to share the same religion. There is no evidence (indeed, no plausible suspicion) of the legislature’s desire to favor the Satmar religion, as opposed to meeting distinctive secular needs or desires of citizens who happened to be Satmars. If there were, Justice Souter would say so; instead, he must merely insinuate.
IV
But even if Chapter 748 were intended to create a special arrangement for the Satmars because of their religion (not including, as I have shown in Part I, any conferral of governmental power upon a religious entity), it would be a permissible accommodation. “This Court has long recognized that the government may (and sometimes must) accommodate religious practices and that it may do so without violating the Establishment Clause.” Hobbie v. Unemployment Appeals Comm’n of Fla., 480 U. S. 136, 144-145 (1987). Moreover, “there is ample room for accommodation of religion under the Establishment Clause,” Corporation of Presiding Bishop of Church of Jesus Christ of Latter-day Saints v. Amos, 483 U. S. 327, 338 (1987), and for “play in the joints productive of a benevolent neutrality which will permit religious exercise to exist without sponsorship and without interference,” Walz v. Tax Comm’n of City of New York, 397 *744U. S. 664, 669 (1970). Accommodation is permissible, moreover, even when the statute deals specifically with religion, see, e. g., Zorach v. Clauson, 343 U. S., at 312-315, and even when accommodation is not commanded by the Free Exercise Clause, see, e. g., Walz, supra, at 673.
When a legislature acts to accommodate religion, particularly a minority sect, “it follows the best of our traditions.” Zorach, supra, at 314. The Constitution itself contains an accommodation of sorts. Article VI, cl. 3, prescribes that executive, legislative, and judicial officers of the Federal and State Governments shall bind themselves to support the Constitution “by Oath or Affirmation.” Although members of the most populous religions found no difficulty in swearing an oath to God, Quakers, Moravians, and Mennonites refused to take oaths based on Matthew 5:34’s injunction “swear not at all.” The option of affirmation was added to accommodate these minority religions and enable their members to serve in government. See 1 A. Stokes, Church and State in The United States 524-527 (1950). Congress, from its earliest sessions, passed laws accommodating religion by refunding duties paid by specific churches upon the importation of plates for the printing of Bibles, see 6 Stat. 116 (1813), vestments, 6 Stat. 346 (1816), and bells, 6 Stat. 675 (1836). Congress also exempted church property from the tax assessments it levied on residents of the District of Columbia; and all 50 States have had similar laws. See Walz, supra, at 676-678.
This Court has also long acknowledged the permissibility of legislative accommodation. In one of our early Establishment Clause cases, we upheld New York City’s early release program, which allowed students to be released from public school during school hours to attend religious instruction or devotional exercises. See Zorach, supra, at 312-315. We determined that the early release program “accommodates the public service to . . . spiritual needs,” and noted that finding it unconstitutional would “show a callous indifference *745to religious groups.” 343 U. S., at 314. In Walz, swpra, we upheld a property tax exemption for religious organizations, observing that it was part of a salutary tradition of “permissible state accommodation to religion.” Id., at 672-673. And in Presiding Bishop, supra, we upheld a section of the Civil Rights Act of 1964 exempting religious groups from the antidiscrimination provisions of Title VII. We concluded that it was “a permissible legislative purpose to alleviate significant governmental interference with the ability of religious organizations to define and carry out their religious missions.” Id., at 335.
In today’s opinion, however, the Court seems uncomfortable with this aspect of our constitutional tradition. Although it acknowledges the concept of accommodation, it quickly points out that it is “not a principle without limits,” ante, at 706, and then gives reasons why the present case exceeds those limits, reasons which simply do not hold water. “[W]e have never hinted,” the Court says, “that an otherwise unconstitutional delegation of political power to a religious group could be saved as a religious accommodation.” Ibid. Putting aside the circularity inherent in referring to a delegation as “otherwise unconstitutional” when its constitutionality turns on whether there is an accommodation, if this statement is true, it is only because we have never hinted that delegation of political power to citizens who share a particular religion could be unconstitutional. This is simply a replay of the argument we rejected in Part II, supra.
The second and last reason the Court finds accommodation impermissible is, astoundingly, the mere risk that the State will not offer accommodation to a similar group in the future, and that neutrality will therefore not be preserved. Returning to the ill fitted crutch of GrendeVs Den, the Court suggests that by acting through this special statute the New York Legislature has eliminated any “‘effective means of guaranteeing’ that governmental power will be and has been neutrally employed.” Ante, at 703, quoting GrendeVs *746Den, 459 U. S., at 125. How misleading. That language in Grendel’s Den was an expression of concern not (as the context in which it is quoted suggests) about the courts’ ability to assure the legislature’s future neutrality, but about the legislature’s ability to assure the neutrality of the churches to which it had transferred legislative power. That concern is inapposite here; there is no doubt about the legislature’s capacity to control what transpires in a public school.
At bottom, the Court’s “no guarantee of neutrality” argument is an assertion of this Court’s inability to control the New York Legislature’s future denial of comparable accommodation. We have “no assurance,” the Court says, “that the next similarly situated group seeking a school district of its own will receive one,” since “a legislature’s failure to enact a special law is . . . unreviewable.” Ante, at 703; see also ante, at 716 (O’Connor, J., concurring in part and concurring in judgment).4 That is true only in the technical (and irrelevant) sense that the later group denied an accommodation may need to challenge the grant of the first accommodation in light of the later denial, rather than challenging the denial directly. But one way or another, “even if [an administrative agency is] not empowered or obliged to act, [a litigant] would be entitled to a judicial audience. Ultimately, the courts cannot escape the obligation to address [a] plea that the exemption [sought] is mandated by the first amendment’s religion clauses.” Olsen v. Drug Enforcement Admin., 878 F. 2d 1458, 1461 (CADC 1989) (R. B. Ginsburg, J.).
*747The Court’s demand for “up front” assurances of a neutral system is at war with both traditional accommodation doctrine and the judicial role. As we have described, supra, at 744, Congress’s earliest accommodations exempted duties paid by specific churches on particular items. See, e. g., 6 Stat. 346 (1826) (exempting vestments imported by “bishop of Bardstown”). Moreover, most efforts at accommodation seek to solve a problem that applies to members of only one or a few religions. Not every religion uses wine in its sacraments, but that does not make an exemption from Prohibition for sacramental wine use impermissible, accord, Church of Lukumi Babalu Aye, Inc. v. Hialeah, 508 U. S., at 561, n. 2 (Souter, J., concurring in judgment), nor does it require the State granting such an exemption to explain in advance how it will treat every other claim for dispensation from its controlled-substances laws. Likewise, not every religion uses peyote in its services, but we have suggested that legislation which exempts the sacramental use of peyote from generally applicable drug laws is not only permissible, but desirable, see Employment Div., Dept. of Human Resources of Ore. v. Smith, 494 U. S. 872, 890 (1990), without any suggestion that some “up front” legislative guarantee of equal treatment for sacramental substances used by other sects must be provided. The record is clear that the necessary guarantee can and will be provided, after the fact, by the courts. See, e. g., Olsen v. Drug Enforcement Admin., supra (rejecting claim that peyote exemption requires marijuana exemption for Ethiopian Zion Coptic Church); Olsen v. Iowa, 808 F. 2d 652 (CA8 1986) (same); Kennedy v. Bureau of Narcotics and Dangerous Drugs, 459 F. 2d 415 (CA9 1972) (accepting claim that peyote exemption for Native American Church requires peyote exemption for other religions that use that substance in their sacraments).5
*748Contrary to the Court’s suggestion, ante, at 708-709,1 do not think that the Establishment Clause prohibits formally established “state” churches and nothing more. I have always believed, and all my opinions are consistent with the view, that the Establishment Clause prohibits the favoring of one religion over others. In this respect, it is the Court that attacks lions of straw. What I attack is the Court’s imposition of novel “up front” procedural requirements on state legislatures. Making law (and making exceptions) one case at a time, whether through adjudication or through highly particularized rulemaking or legislation, violates, ex ante, no principle of fairness, equal protection, or neutrality simply because it does not announce in advance how all future cases (and all future exceptions) will be disposed of. If it did, the manner of proceeding of this Court itself would be unconstitutional. It is presumptuous for this Court to impose — out of nowhere — an unheard-of prohibition against proceeding in this manner upon the Legislature of New York State. I never heard of such a principle, nor has anyone else, nor will it ever be heard of again. Unlike what the New York Legislature has done, this is a special rule to govern only the Satmar Hasidim.
V
A few words in response to the separate concurrences: Justice Stevens adopts, for these cases, a rationale that is *749almost without limit. The separate Kiryas Joel school district is problematic in his view because “[t]he isolation of these children, while it may protect them from ‘panic, fear and trauma,’ also unquestionably increased the likelihood that they would remain within the fold, faithful adherents of their parents’ religious faith.” Ante, at 711. So much for family values. If the Constitution forbids any state action that incidentally helps parents to raise their children in their own religious faith, it would invalidate a release program permitting public school children to attend the religious-instruction program of their parents’ choice, of the sort we approved in Zorach;6 indeed, it would invalidate state laws according parents physical control over their children, at least insofar as that is used to take the little fellows to church or synagogue. Justice Stevens’ statement is less a legal analysis than a manifesto of secularism. It surpasses mere rejection of accommodation, and announces a positive hostility to religion — which, unlike all other noncriminal values, the State must not assist parents in transmitting to their offspring.
Justice Kennedy’s “political-line-drawing” approach founders on its own terms. He concedes that the Constitution does not prevent people who share a faith from forming their own villages and towns, and suggests that the formation of the village of Kiryas Joel was free from defect. Ante, at 729-730. He also notes that States are free to draw political lines on the basis of history and geography. Ante, at 730. I do not see, then, how a school district drawn to mirror the boundaries of an existing village (an existing geographic line), which itself is not infirm, can violate the Constitution. Thus, while Justice Kennedy purports to share my criticism (Part IV, supra) of the Court’s unprecedented insistence that the New York Legislature make its accommo*750dations only by general legislation, see ante, at 722, 726, his own approach is little different. He says the village is constitutional because it was formed (albeit by members of a single religious sect) under a general New York law; but he finds the school district unconstitutional because it was the product of a specific enactment. In the end, his analysis is no different from the Court’s.
Justice Kennedy expresses the view that School Dist. of Grand Rapids v. Ball, 478 U. S. 373 (1985), and Aguilar v. Felton, 473 U. S. 402 (1985) — the cases that created the need for the Kiryas Joel legislation by holding unconstitutional state provision of supplemental educational services in sectarian schools — “may have been erroneous,” and he suggests that “it may be necessary for us to reconsider them at a later date.” Ante, at 731. Justice O’Connor goes even further and expresses the view that Aguilar should be overruled. Ante, at 717-718. I heartily agree that these cases, so hostile to our national tradition of accommodation, should be overruled at the earliest opportunity; but meanwhile,- today’s opinion causes us to lose still further ground, and in the same antiaccommodationist direction.
Finally, Justice O’Connor observes that the Court’s opinion does not focus on the so-called Lemon test, see Lemon v. Kurtzman, 403 U. S. 602 (1971), and she urges that that test be abandoned, at least as a “unitary approach” to all Establishment Clause claims, ante, at 721. I have previously documented the Court’s convenient relationship with Lemon, which it cites only when useful, see Lamb's Chapel v. Center Moriches Union Free School Dist., 508 U. S. 384, 397-401 (1993) (Scalia, J., concurring in judgment), and I no longer take any comfort in the Court’s failure to rely on it in any particular case, as I once mistakenly did, see Lee v. Weisman, 505 U. S. 577, 644 (1992) (Scalia, J., dissenting). But the Court’s snub of Lemon today (it receives only two “see also” citations, in the course of the opinion’s description of *751Grendel’s Den) is particularly noteworthy because all three courts below (who are not free to ignore Supreme Court precedent at will) relied on it, and the parties (also bound by our case law) dedicated over 80 pages of briefing to the application and continued vitality of the Lemon test. In addition to the other sound reasons for abandoning Lemon, see, e. g., Edwards v. Aguillard, 482 U. S. 578, 636-640 (1987) (Scalia, J., dissenting); Wallace v. Jaffree, 472 U. S. 38, 108-112 (1985) (Rehnquist, J., dissenting), it seems quite inefficient for this Court, which in reaching its decisions relies heavily on the briefing of the parties and, to a lesser extent, the opinions of lower courts, to mislead lower courts and parties about the relevance of the Lemon test. Compare ante, p. 687 (ignoring Lemon despite lower courts’ reliance), with Lamb’s Chapel, supra (applying Lemon despite failure of lower court to mention it).
Unlike Justice O’Connor, however, I would not replace Lemon with nothing, and let the case law “evolve” into a series of situation-specific rules (government speech on religious topics, government benefits to particular groups, etc.) unconstrained by any “rigid influence,” ante, at 721. The problem with (and the allure of) Lemon has not been that it is “rigid,” but rather that in many applications it has been utterly meaningless, validating whatever result the Court would desire. See Lamb’s Chapel, supra, at 399 (Scalia, J., concurring in judgment); Wallace, supra, at 110-111 (Rehnquist, J., dissenting). To replace Lemon with nothing is simply to announce that we are now so bold that we no longer feel the need even to pretend that our haphazard course of Establishment Clause decisions is governed by any principle. The foremost principle I would apply is fidelity to the longstanding traditions of our people, which surely provide the diversity of treatment that Justice O’Connor seeks, but do not leave us to our own devices.
*752* * *
The Court’s decision today is astounding. Chapter 748 involves no public aid to private schools and does not mention religion. In order to invalidate it, the Court casts aside, on the flimsiest of evidence, the strong presumption of validity that attaches to facially neutral laws, and invalidates the present accommodation because it does not trust New York to be as accommodating toward other religions (presumably those less powerful than the Satmar Hasidim) in the future. This is unprecedented — except that it continues, and takes to new extremes, a recent tendency in the opinions of this Court to turn the Establishment Clause into a repealer of our Nation’s tradition of religious toleration. I dissent.
A census taken in 1906,10 years after statehood was granted to Utah, and 6 years before it was granted to New Mexico, showed that in Utah 87.7% of all church members were Mormon, and in New Mexico 88.7% of all church members were Roman Catholic. See Bureau of the Census, Special Reports, Religious Bodies, Part I, p. 55 (1910).
At the county level, the smallest unit for which comprehensive data is available, there are a number of counties in which the overwhelming majority of churchgoers are of a single religion: Rich County, Utah (100% Mormon); Kennedy County, Texas (100% Roman Catholic); Emery County, Utah (99.2% Mormon); Franklin and Madison Counties, Idaho (99% or more Mormon); Graham County, North Carolina (93.7% Southern Baptist); Mora County, New Mexico (92.6% Roman Catholic). M. Bradley, N. Green, D. Jones, M. Lynn, & L. McNeil, Churches and Church Membership in the United States 1990, pp. 46, 112-113, 246, 265, 283, 365, 380, 393 (1992). In all of these counties the adherents of the indicated religion constitute a substantial majority, in some cases over a 95% majority, of the total population. If data were available for smaller units of government than counties, I have no doubt I could point to hundreds of towns placed in jeopardy by today’s opinion.
The Court says that “[e]arly on in the development of public education in New York, the State rejected highly localized school districts for New York City when they were promoted as a way to allow separate schooling for Roman Catholic children.” Ante, at 704. Both the implication that this rejection of localism was general state policy, and the implication that (like the Court’s prohibition of localism today) it had the purpose and effect of religious neutrality, are simply not faithful to the cited source. The 1841 proposal was not to treat New York City schools differently, in order to favor Roman Catholics; it was “that the state’s school code, which promoted a district system structure with local taxing authority, be extended to New York City.” R. Church & M. Sedlak, Education in the United States 167 (1976). And the rejection of that proposal was not a triumph for keeping sectarian religion out of some public schools; it was a triumph for keeping the King James version of the Bible in all public schools. The Court’s selected source concludes: “[T]he Whigs swept the city elections that year [1842] and made Bible reading — the King James version — mandatory in any schools sharing these monies. There was nothing left for the Catholics to do but to build their own parochial system with their own money.” Id., at 168-169.
The Court hints, ante, at 703, that its fears would have been allayed if the New York Legislature had previously created similar school districts for other minority religions. But had it done so, each of them would have been attacked (and invalidated) for the same reason as this one: because it had no antecedents. I am sure the Court has in mind some way around this chicken-and-egg problem. Perhaps the legislature could name the first four school districts in pectore.
The Court likens its demand for “up front” assurances to the Court’s focus on the narrowness of the statute it struck down in Texas Monthly, Inc. v. Bullock, 489 U. S. 1 (1989). See ante, at 708. Texas Monthly *748bears no resemblance to today’s opinion, except that it also was wrong and it also misinterpreted Walz v. Tax Comm’n of City of New York, 397 U. S. 664 (1970), see 489 U. S., at 33-40 (Scalia, J., dissenting). The tax treatment of publishing companies in Texas was governed by an across-the-board rule. There was never any question whether nonreligious publishers would get the tax exemption accorded to religious publishers; by rule they did not, and the Court struck down that rule because it discriminated in favor of religion. By contrast, adjustments to existing school districts in New York are done case by case. No decision, including Texas Monthly, remotely suggests that approaching accommodations in a case-specific manner automatically violates the Establishment Clause.
Justice Stevens’ bald statement that such a program would be permissible, see ante, at 711-712, can exclude it from the reach of his opinion, but not from the reach of his logic.