1. Pepper, Stephen, “The Conundrum of the Free Exercise Clause—Some Reflections on Recent Cases,” 9 Northern Kentucky Law Review 265, 303 (1982).Google Scholar See also Johnson, Phillip, “Concepts and Compromise in First Amendment Religion Doctrine,” 72 California Law Review 817, 839 (1984)CrossRefGoogle Scholar (“a mess”). For compilations of the confusions in the law, see Note, “Rebuilding the Wall: The Case for a Return to the Strict Interpretation of the Establishment Clause,” 81 Columbia Law Review 1463, 1463–66 (1981)Google Scholar; Choper, Jesse, “The Religion Clauses of the First Amendment: Reconciling the Conflict,” 41 University of Pittsburgh Law Review 673, 680–81 (1980).Google Scholar

2. Compare Board of Education v. Allen, , 392 US. 236 (1968)Google Scholar (books), with Meek, v. Pittenger, , 421 U.S. 349 (1975)Google Scholar (globes). A distinction between loans and purchases seemed viable until Committee for Public Education v. Regan, , 444 US. 646 (1980)Google Scholar, which for the first time approved a direct financial grant to religious schools.

3. Compare Marsh, v. Chambers, , 463 U.S. 783 (1983)Google Scholar (prayers in legislature constitutional), with Stone, v. Graham, , 449 U.S. 39 (1980)Google Scholar (posting of Ten Commandments in schoolrooms unconstitutional).

4. United States v. Lee, 455 U.S. 252 (1982).

5. Wisconsin, v. Yoder, , 406 U.S. 205 (1972).Google Scholar See Marshall, William, “Solving the Free Exercise Dilemma: Free Exercise as Expression,” 67 Minnesota Law Review 545, 548–53 (1983).Google Scholar

6. Compare Mueller, v. Allen, , 463 US. 388 (1983)Google Scholar (one form of tuition tax credit constitutional), with Committee for Public Education v. Nyquist, 413 US. 756 (1973) (another form of tuition tax credit unconstitutional). Compare also Flast, v. Cohen, , 392 US. 83 (1968)Google Scholar (taxpayer has standing to challenge donation of money to religious organization), with Valley Forge Christian College v. Americans United, 454 US. 464 (1982) (taxpayer lacks standing to challenge donation of land and buildings to religious organization).

7. Note, “Reinterpreting the Religion Clauses: Constitutional Construction and Conceptions of the Self,” 97 Harvard Law Review 1468 (1984)Google Scholar, adopts a position like that taken here. The note locates the source of the tensions in an enduring psychological dualism, whereas I argue that its source lies in a particular confluence of political thought and action. See also Garvey, John, “Free Exercise and the Values of Religious Liberty,” 18 Connecticut Law Review 779 (1986)Google Scholar (for purposes of constitutional analysis, analogizing religious belief to insanity). The note also takes a normative stance, whereas I do not. See note 9.

8. If my argument is correct we should be able to discover analogous tensions in the law of other intermediate institutions such as families and education. I believe that such tensions can be found. For example, see Note, “Peaceful Labor Picketing and the First Amendment,” 82 Columbia Law Review 1469 (1982)Google Scholar (labor unions and picketing); Note, “Labor Picketing and Commercial Speech: Free Enterprise Values in the Doctrine of Free Speech,” 91 Yale Law Journal 938 (1982)Google Scholar (same); Burt, Robert, “The Burger Court and the Family,” in The Burger Court, p. 92Google Scholar (families); Frug, Gerald, “The City as a Legal Concept,” 93 Harvard Law Review 1057 (1980) (federalism).CrossRefGoogle Scholar

9. Two preliminary cautions are appropriate. First, these principles should be taken as ones that provide order to the description of the Supreme Court's decisions rather than as ones that I believe are normatively desirable. With one notable exception, see note 31, I have tried to exclude from my discussion arguments in favor of or against the fundamental normative merit of these principles, although I do indicate analytic difficulties they create. Second, these principles may not be analytically distinct. Once religion is largely reduced to ordinary belief, what is left may necessarily be marginal. Using two labels makes exposition easier.

10. Kurland, Philip, “Of Church and State and the Supreme Court,” 29 University of Chicago Law Review 1, 5 (1961).CrossRefGoogle Scholar Kurland admits that it has not been adopted by the Court, although the rhetoric of neutrality is widely used by commentators. See Kurland, Philip, “The Irrelevance of the Constitution: The Religion Clauses of the First Amendment and the Supreme Court,” 24 Villanova Law Review 3, 24 (1978)Google Scholar (“met with almost uniform rejection”).

11. The Court has indeed required exemptions in some cases. See Sherbert, v. Verner, , 374 U.S. 398 (1963)Google Scholar; Wisconsin, v. Yoder, , 406 U.S. 205 (1972).Google Scholar The scholarly consensus is that the accommodations prohibited by strict neutrality are sometimes compelled by the Constitution, at least when the adjustments of secular programs are modest and the burdens on conscience severe, and are often permitted by it. See, e.g., Clark, J. Morris, “Guidelines for the Free Exercise Clause,” 83 Harvard Law Review 327, 345 (1969)CrossRefGoogle Scholar; Gianella, Donald, “Religious Liberty, Nonestab-lishment, and Doctrinal Development: Part I. The Religious Liberty Guarantee,” 80 Harvard Law Review 1381, 1399–1402 (1967)CrossRefGoogle Scholar; Dodge, Joseph, “The Free Exercise of Religion: A Sociological Approach,” 67 Michigan Law Review 679, 706 (1969)CrossRefGoogle Scholar; Marcus, Paul, “The Forum of Conscience: Applying Standards under the Free Exercise Clause,” 1973Google ScholarDuke Law Journal 1217, 1235–36Google Scholar; Pfeffer, Leo, “Religion-Blind Government,” 15 Stanford Law Review 389 (1963).CrossRefGoogle Scholar Contra Choper, , “Religion Clauses,” p. 691Google Scholar; Garvey, John, “Freedom and Equality in the Religion Clauses,” 1981Google ScholarSupreme Court Review 193, 198.Google Scholar

12. In addition, a doctrine of neutrality must be framed so as to prohibit statutes that either use religious classifications explicitly or are not neutral in their effects. Experience has shown that it is easy to draft facially neutral statutes that raise serious questions about religious freedom. See, e.g., Epperson, v. Arkansas, , 393 LIS. 97 (1968)Google Scholar (invalidating a state law prohibiting the teaching of evolution). As a religion case Epperson seems best explained on the ground that its subject matter was a religiously sensitive one and that legislation on such sensitive matters is highly suspect.

The category of religiously sensitive issues is open-ended, though, and its content is likely to be affected by quite subjective evaluations: Why should statutes authorizing schools to have a moment of silence not be treated as religiously sensitive? or statutes restricting the availability of abortions? Laurence Tribe initially offered a version of this approach, “Foreword: Toward a Model of Roles in the Due Process of Life and Law,” 87 Harvard Law Review 1, 21–25 (1973)Google Scholar, but later retracted his analysis, arguing that it failed to give enough weight to the interest that religiously motivated people have in participating in politics to advance their religiously based views. See Tribe, , American Constitutional Law, p. 928.Google Scholar It might be possible, however, to reconstruct the analysis by isolating a particular category of “religion-sensitive” cases to which it applies. The abortion example suggests one serious problem with the category: it may have been true that abortion was a religion-sensitive subject in 1973, but with the rise of organized secular opposition to abortion it has become difficult so to confine the topic.

13. 463 US. 388 (1983).

14. 413 US. 756 (1973).

15. See Gianella, Donald, “Religious Liberty, Nonestablishment, and Doctrinal Development: Part II. The Nonestablishment Principle,” 81 Harvard Law Review 513, 519–20 (1968)CrossRefGoogle Scholar; Johnson, , “Concepts and Compromise,” pp. 822–23.Google Scholar

The example of aid to education illustrates the general source of the emptiness of the concept of neutrality in the setting of impact as well. To assess impact we must identify the relevant subject on which the program at issue has an impact. But the subject matter can be described narrowly or broadly, whichever is needed to produce the desired impact: If “aid to nonpublic education” is too narrow, then try “aid to education” or “aid to socially beneficial nonprofit organizations.”

16. See Gianella, , “Nonestablishment Principle,” pp. 574–75Google Scholar; Johnson, , “Concepts and Compromise,” pp. 822–23.Google Scholar

17. See Tushnet, , “Anti-Formalism,” pp. 1508–16.Google Scholar

18. 330 US. 1 (1946).

19. Apart from the general difficulties with the originalist theory of incorporation of the Bill of Rights into the fourteenth amendment, there are some problems specific to the absorption of the religion clauses. The due process clause protects people against deprivations of life, liberty, and property. Its terms are not well suited to encompass the interests protected by the establishment clause; although the right to free exercise of religion is a “liberty,” the guarantee against establishment does not protect anything readily characterized as a personal liberty or as property.

These difficulties are most apparent in the Court's decisions on standing in establishment clause cases. See, e.g., Doremus, v. Board of Education, 342 US. 429 (1952).Google Scholar In addition, to the extent that the framers of the first amendment sought to protect state establishments against national action, it is not entirely coherent to say that the amendment is now applicable to the states.

20. Most of the points were made in Abington School Dist. v. Schempp, , 374 US. 203, 237–42 (1963) (Brennan, J., concurring).Google Scholar

21. See Marsh, v. Chambers, , 463 US. 783, 819–21 (1983)Google Scholar (Brennan, J., dissenting).

22. One might challenge this argument by asking for evidence that coercion actually occurs (which converts the argument into a direct free exercise argument), as did Stewart, Justice in Schempp. 374 US., at 316–20Google Scholar (Stewart, J., dissenting). But the rights-based defense might respond to that challenge with the assertion that detecting the presence of coercion on a case-by-case basis is so difficult that we can be sure that too much undetected coercion will occur unless the practice is flatly prohibited. In this manner the rights-based theory treats the establishment clause as a prophylactic against free exercise violations.

23. For a discussion of pluralism and constitutional theory in general, see Chapter 2, Red, White, and Blue.

24. See Laycock, Douglas, “Towards a General Theory of the Religion Clauses: The Case of Church Labor and the Right to Church Autonomy,” 81 Columbia Law Review 1373, 1414–15 (1981)CrossRefGoogle Scholar, arguing that the exemption in Sherbert v. Verner caused non-Sabbatarians to subsidize Seventh-Day Adventists through their contributions to the unemployment compensation fund; Choper, , “Religion Clauses,” pp. 691–92.Google Scholar See also 26 U.S.C. §1402(g).

25. See Gianella, , “Nonestablishment Principle,” pp. 522–26.Google Scholar

26. See, e.g., Widmar, v. Vincent, , 454 US. 263, 288–89 (1981)Google Scholar (White, J., dissenting). See also Sherbert, v. Verner, , 374 U.S. 398, 418 (1963)Google Scholar (Harlan, J., joined by White, J., dissenting); Thomas, v. Review Board, 450 US. 707, 720 (1981)Google Scholar (Rehnquist, J., dissenting).On limits to reasonableness, see Widmar, v. Vincent, , 454 US., p. 287Google Scholar (White, J., dissenting) (University could not bar students from saying grace before meals).

27. Alternatively, the religion clauses might be treated as creating broad outer boundaries for a set of values, and their proper interpretation would require that each generation be allowed to decide for itself how to accommodate those values within the limits set by the clauses. In this view pluralism assures that the accommodation respect those limits.

28. The pluralist theory also offers a justification for the establishment branch of Kurland's religion clause theory. Religious pluralism would require legislatures to structure their programs using criteria that are neutral as to religion. The outcomes of such a process are unlikely to threaten serious erosions of establishment clause values. A more likely result is legislative paralysis, an inability to come up with a neutral program at acceptable levels of cost.

The Equal Access Act, P.L. 98–377, 98 Stat. 1303 (Aug. 11, 1984), provides a useful example, although the operative political constraint was not concern over costs. This act was designed to ensure that student religious groups would be allowed to use school facilities. The act as adopted provides that schools that create “limited open forums” as defined in the act may not discriminate against students who wish to conduct meetings “on the basis of the religious, political, philosophical, or other content of the speech at such meetings” (emphasis added). The emphasized words were added to reduce the political opposition to the act; they also dilute its particularly religious content. See Laycock, Douglas, “Equal Access and Moments of Silence: The Equal Status of Religious Speech by Private Speakers,” 81 Northwestern University Law Review 1 (1986).Google Scholar

29. Federalism might reinforce pluralism if adherents of specific denominations came to reside in religiously homogeneous communities; religious communities could be strengthened by allying them with a parallel secular authority, from which dissenters could remove themselves by migration. Of course, formal religious control over local government raises a host of constitutional questions. For a discussion, see Note, “Church Control of Municipality: Establishing a First Amendment Institutional Suit,” 38 Stanford Law Review 1363 (1986).Google Scholar

30. I take this to be a description of the political stance of many American Jews. Bargaining would also fail if the group was completely marginal to the community, so that no one else would find it politically profitable to trade support for the group's position on religious issues in exchange for its support on other issues: engaging in such a deal would be the kiss of death. I take this to be the political position of organized secularists and such outcast “cults” as the Reunification Church (“Moonies”). See generally The Brainwashing/Deprogramming Controversy: Sociological, Psychological, Legal and Historical Perspectives (Bromley, David and Richardson, James eds. 1983).Google Scholar

31. See Laycock, Douglas, “A Survey of Religious Liberty in the United States,” 47 Ohio State Law Journal 409, 432–33 (1986)Google Scholar (political process adequate to protect “adherents of mainstream religions” but not “small numbers of idiosyncratic believers”). Justice O'Connor has begun to develop a version of the pluralist theory. She advocates inquiring into whether governmental action “sends a message [of endorsement of religion] to nonadherents that they are outsiders, not full members of the political community, and an accompanying message to adherents that they are insiders, favored members of the political community.” Lynch, v. Donnelly, , 465 U.S. 668, 688 (1984)Google Scholar (O'Connor, J., concurring). Dorsen, Norman and Sims, Charles, “The Nativity Scene Case: An Error of Judgment,” 1985Google ScholarUniversity of Illinois Law Review 837, 859–60Google Scholar, generally endorse Justice O'Connor's approach but criticize its application in Lynch.

This emphasis on symbolic messages is problematic. It would have judges decide what symbolic content the government's messages have. Because judges will always be broadly representative of the general population, they will be susceptible to all the distortions of interpretation that membership in the majority entails. For example, Justice O'Connor articulated her approach in concurring with the Court's decision to find no establishment clause violation in the public sponsorship of a crèche; she concluded that the practice “cannot fairly be understood to convey a message of government endorsement of religion.” Id., p. 614 (O'Connor J., concurring). This came as a surprise to most Jews, whose judgments on this issue turn out to be “unfair” injustice O'Connor's eyes. See Tribe, Laurence, “Constitutional Calculus: Equal Justice or Economic Efficiency?” 98 Harvard Law Review 592, 611 (1985)CrossRefGoogle Scholar (“One cannot avoid hearing in Lynch a faint echo of the Court that found nothing invidious in the Jim Crow policy of ‘separate but equal’”). One could use stronger words.

I find it impossible to avoid expressing my own views on the case. It seems to me difficult to believe that the majority could have reached the result it did had there been a Jew on the Court to speak from the heart about the real meaning of public displays of crèches to Jews. At the same time, of course, Jews have always known that they were strangers in the land and have taken some succor from that fact. See Neusner, Jacob, Stranger at Home: “The Holocaust,” Zionism, and American Jewry 122–23 (1981).Google ScholarLynch just reminds us of that status, and, distressing though it may be to have it brought to light, we may profit from learning the lesson again. See also Van Alstyne, William, “Trends in the Supreme Court: Mr. Jefferson's Crumbling Wall — A Comment on Lynch v. Donnelly,” 1984 Duke Law Journal 770, 781–87Google Scholar; Kurland, Philip, “The Religion Clauses and the Burger Court,” 34 Catholic University Law Review 1, 13 (1984)Google Scholar (Lynch is “sleazy”). ACLU v. Birmingham, 791 F.2d 1561, 1572 (6th Cir.) (Nelson, J., dissenting), cert, denied, 107 S. Ct. 421 (1986), addresses this issue expressly, as Justice O'Connor did not, but ultimately concludes that it is enough that crèches are not motivated by “anti-Jewish animus”. This does not seem to be a fair application of Justice O'Connor's approach.

In addition, it is not clear why symbolic exclusion should matter so long as “nonadherents” are in fact actually included in the political community. Under those circumstances nonadherents who believe that they are excluded from the political community are merely expressing the disappointment felt by everyom who has lost a fair fight in the arena of politics. Symbolic exclusion following actual inclusion might matter if the nonadherents become so disheartened by their losses that they withdraw from politics altogether. This prospect seems remote enough under presendy foreseeable circumstances to make a prophylactic approacl as questionable here as it was in the rights-based analysis of the establishment clause.

32. See Yates v. United States, 354 US. 298 (1958)Google Scholar (Smith Act construed not to penalize abstract belief as distinguished from advocacy of unlawful action).

33. See U.S. Constitution, art. IV (no religious test to be imposed for public office); Torcaso v. Watkins, 367 U.S. 488 (1961).Google Scholar

34. See, e.g., Kovacs v. Cooper, 336 US. 77 (1949) (noise)Google Scholar; Schneider v. Irvington, 308 US. 147 (1939) (littering).Google Scholar

35. For articulation and defense of the reduction principle, see Marshall, , “Solving the Free Exercise Dilemma,”Google Scholar the most extensive discussion; Choper, Jesse, “Defining ‘Religion’ in the First Amendment,” 1982 University of Illinois Law Review 579, 581–82Google Scholar; Clark, , “Guidelines,” pp. 336–37.Google Scholar

36. See Cover, Robert, “The Origins of Judicial Activism in the Protection of Minorities,” 91 Yale Law Journal 1287 (1982).CrossRefGoogle Scholar

37. See Lovell v. Griffin, 303 U.S. 444 (1938)Google Scholar; Cox v. New Hampshire, 312 U.S. 569 (1941)Google Scholar; Martin v. Struthers, 319 U.S. 141 (1943).Google Scholar

38. In addition, the Court in these early cases did not distinguish between the protections provided by the free speech clause and those provided by the free exercise clause. Instead it invoked the entire first amendment as the basis for its decisions. See Marshall, , “Solving the free Exercise Dilemma,” pp. 562–64Google Scholar; Kurland, , “Church and State,” pp. 36–44.Google Scholar

39. 310 U.S. 296, 303–4 (1940).

40. See, e.g., Thornhill v. Alabama, 310 US. 88 (1940).Google Scholar

41. McGowan v. Maryland, 366 US. 420 (1961)Google Scholar; Two Guys from Harrison v. McGinley, 366 U.S. 582 (1961)Google Scholar; Braunfeld v. Braun, 366 U.S. 599 (1961)Google Scholar; Gallagher v. Crown Kosher Super Market, 366 U.S. 617 (1961).Google Scholar

42. 366 U.S., pp. 599, 605–6. See, e.g., Konigsberg v. State Bar, 366 U.S. 36 (1961)Google Scholar; Adler v. Board of Education, 342 U.S. 485 (1952).Google Scholar

43. 366 U.S., p. 607.

44. 391 U.S. 367, 377 (1968). The reduction principle may have reached its fullest flower in Widmar v. Vincent, 454 U.S. 263 (1981)Google Scholar, in which, in a case crying out for free exercise treatment, the Court adopted a free expression analysis instead. There the University of Missouri denied a group of students permission to conduct its prayer meetings in University buildings even though the buildings were available to nonreligious organizations for their meetings. A free exercise analysis would treat this as a simple case of discrimination against religion. The Court's free speech analysis treated it as a content-based restriction on speech, which could be justified only by a compelling state interest. Religion entered the Court's analysis by the back door, as the Court rejected the University's claim that it had a compelling interest in avoiding the appearance of an establishment of religion. As William Marshall has said, “Because few activities are more profoundly religious than prayer, Widmar suggests that there is no core religious activity exclusively protected by the free exercise clause.” “Solving the Free Exercise Dilemma,” pp. 559–60Google Scholar. Much here depends on the scope one gives to notions of symbolic speech: A ritual sacrifice might or might not be symbolic speech but perhaps could receive free exercise protection no matter how it was regarded. See Frank v. State, 604 P.2d 1068 (Alaska 1979).

45. See Greenawalt, Kent, “Religion as a Concept in Constitutional Law,” 72 California Law Review 753, 777–88 (1984).CrossRefGoogle Scholar

46. 374 U.S. 398 (1963), followed in Thomas v. Review Board, 450 US. 707 (1981)Google Scholar; Hobbie v. Unemployment Appeals Comm'n., 107 S. Ct. 1046 (1987).Google Scholar

47. 406 US. 205 (1972).

48. See Murdock v. Pennsylvania, 319 US. 105 (1943)Google Scholar; Martin v. Struthers, 319 U.S. 141 (1943)Google Scholar (exempting those engaged in religious and speech activities from requirements imposed on those engaged in commercial activities).

49. The employers might be liable under the civil rights acts for a private conspiracy to deprive her of her political rights. Arguably the state's decision to construe its unemployment statutes to make political grounds “cause” for discharge would be unconstitutional under the circumstances hypothesized. But that was unlikely when Sherbert was decided and is even more unlikely now. See, e.g., Rendell-Baker v. Kohn, 457 U.S. 830 (1982).Google Scholar

50. See Kurland, Philip, “The Supreme Court, Compulsory Education, and the First Amendment's Religion Clauses,” 75 West Virginia Law Review 213, 237–38 (1973).Google Scholar

51. This is the consensus of the commentators. See, e.g., Clark, , “Guidelines,” pp. 337–38.Google Scholar

52. See Marshall, , “Solving the Free Exercise Dilemma,” pp. 586–87Google Scholar. The fuzziness of the law of exemptions under the free speech clause is such that this expansion would entail no major doctrinal revisions. The courts could tinker a bit with the characterization of the interests that go into the balance and create exemptions that are at present unavailable.

53. In part the suggestion is troubling because it is unlikely that the Court as presently composed will expand free speech exemptions, except perhaps in largely innocuous situations. See, e.g., Brown v. Socialist Workers '74 Campaign Committee, 459 U.S. 87 (1982).Google Scholar

54. 454 U.S., pp. 285, 284, 283 (White, J., dissenting). The majority criticized the distinction between worship and speech on three grounds. It did not explain when verbal acts such as singing, reading, and teaching acquire religious content so as to transform them from speech into worship. Further, courts could not intelligibly “inquire into the significance of words and practices” to determine whether a reading from the Bible is speech or worship. Finally, Justice White conceded that “religious speech designed to win religious converts” was protected by the free speech clause; there was no distinction between such speech and “religious worship by persons already converted.” Id., pp. 269–70 n. 6. Marshall, , “Solving the Free Exercise Dilemma,” p. 579Google Scholar, argues that, in a Meiklejohn-type theory of free expression, religious speech is at least not entitled to greater protection than political speech, because it does not directly contribute to the process of political deliberation.

55. The perspective here is essentially that of Durkheim, but in order to make my major points I need adopt only such a general perspective, without addressing the controversial details of Durkheim's own work. For similar perspectives on recent religious phenomena in the United States, see Tobey, Alan, “The Summer Solstice of the Healthy-Happy-Holy Organization,” in The New Religious Consciousness 5, 14–15Google Scholar, (Glock, Charles and Bellah, Robert eds. 1976)Google Scholar; Johnson, Gregory, “The Hare Krishna in San Francisco,”Google Scholarid., pp. 31, 41, 44–45; Bellah, Robert, “The New Religious Consciousness and the Crisis of Modernity,”Google Scholarid., p. 333. For similar perspectives in legal commentary, see Valente, William, “Aid to Church-Related Education — New Directions without Dogma,” 55 Virginia Law Review 579, 604–5 (1969)CrossRefGoogle Scholar; Note, “The Sacred and the Profane: A First Amendment Definition of Religion,” 61 Texas Law Review 139, 158–59 (1982).Google Scholar

56. This description of religion's distinctiveness can be challenged from two directions. It might be said that it ignores religious traditions emphasizing the direct contact between the believer and God, religious traditions celebrating isolated monastic existence, and the deist tradition in which communal worship is replaced by rationalist appreciation of God's role in the world. Yet the anchorite's isolated worship is religious rather than eccentric because it is located in a tradition of discourse to which other believers have access. See Brown v. Pena, 441 F.Supp. 1382 (S.D. Fla. 1977)Google Scholar (worship entailing consumption of cat food not a religion); Africa v. Pennsylvania, 662 F.2d 1025 (3d Cir. 1981)Google Scholar (pantheism not a religion). Talk about a person's relation to God can be coherent only in a socially located discourse making such talk sensible. The second challenge then becomes forceful. All speech, not just religious worship can be coherent only in socially located discourses. Religion should then be treated as exemplary rather than distinctive. I agree with this challenge. See note 8.

57. See note 11.

58. But see Choper, , “Defining ‘Religion’ in the First Amendment,” 1982 University of Illinois Law Review 579, 610–12Google Scholar; Laycock, , “General Theory,” pp. 1383–84.Google Scholar

59. Of course it may go no further. But as will be shown, that turns out to be no significant limit on the government's power. Under these circumstances the reduction principle here might be called an escalation principle.

60. Shiffrin, Steven, “Government Speech,” 27 UCLA Law Review 565 (1980)Google Scholar; Yudof, Mark, When Government Speaks: Politics, Law and Government Expression in America (1983)Google Scholar. For a review of Yudof, , When Government SpeaksGoogle Scholar, see Tushnet, Mark, “Book Review,” 1984 Wisconsin Law Review 129.Google Scholar

61. 463 U.S. 388 (1983).

62. Committee for Public Education v. Nyquist, 413 U.S. 756 (1973).Google Scholar

63. 463 U.S., pp. 398–99. The Court also stated that aid provided through parents rather than directly to the schools “reduced the Establishment Clause Objections.” Ibid. But see Committee for Public Education v. Reagan, 444 U.S. 646 (1980)Google Scholar. It acknowledged that the New York and Minnesota statutes were indistinguishable on this ground.

64. Obviously there are more people in the former class. New York, however, provides massive subsidies through its expenditure system to parents whose children attend public schools. Thus, both forms of subsidies taken together constitute a system of subsidies to parents, just as in Minnesota. Further, the relevance of numbers alone is obscure. In New York enough parents of children in public schools were willing to subsidize the smaller group of parents with children in private schools to create a majority in favor of the statute. The Court in Mueller expressly rejected an alternative measure of breadth, refusing to look beyond the face of the statute to the actual fiscal impact of the benefits provided. Id., p. 401. It argued that tracing benefits through the tax system is difficult and that the results would vary from year to year depending on how carefully parents with children in public schools keep records supporting their claimed deductions. Further, the Court was concerned that it could not develop standards to distinguish broad from narrow fiscal impacts.

65. Id., p. 399 (quoting Wolman v. Walter, 433 U.S. 229, 263 [1977]Google Scholar [opinion of Powell, J.]).

66. The pluralist analysis is filled with judgments about how the political process is likely to work, and the Court's refusal to examine actual fiscal impact in Mueller indicates that it is unwilling to entertain empirical challenges to the wider set of political judgments involved. Yet there is something to be said in favor of the pluralist analysis. Some states might adopt statutes like Minnesota's, and in a federal system that might well be acceptable. In contrast, devising a national tuition tax-credit system that uses sufficiently neutral terms, and that can attract enough political support, may be quite difficult, precisely because of the rather large impact of such a program. Discussions of proposals to adopt a national tuition tax-credit plan illustrate the problem. See Hearing on S. 2673, Senate Committee on Finance, 97th Cong., 2d Sess. (July 16, 1982). To keep the fiscal impact within acceptable bounds, the Reagan Administration's proposals were scaled back, primarily by eliminating a provision making the credit refundable to those whose incomes are so low as to make them unable to benefit from a pure tax credit. Eliminating that feature, though, makes the program beneficial only to the relatively well-to-do, which undermines the political attractiveness of the proposal.

67. Note that this rule cannot readily be accounted for by the reduction principle.

68. In Sherbert v. Verner, for example, the Court carefully noted how few Seventh-Day Adventists there were in the locality and suggested that only rarely would a Sabbatarian be unable to get a job which could accommodate her schedule. 374 U.S., p. 399 n. 2.

69. See, e.g., Clark, , “Guidelines,” pp. 331–33.Google Scholar

70. This is one of the concerns expressed in Goldman v. Weinberger, 475 U.S. 503 (1986)Google Scholar, in which the Court rejected a claim by an Orthodox Jew that the free exercise clause required that he be exempted from the military's stipulation that soldiers on duty wear only approved headgear; the applicable regulations did not allow him to wear his yarmulke.

71. 455 U.S., pp. 259–60. An alternative explanation of the Court's distinction between Yoder and Lee is that it believed that exemptions from taxes give people incentives to misrepresent their religious beliefs in ways that exemptions from public education do not and that determining the sincerity with which a religious belief is held is administratively difficult in the tax context. See Freed, Mayer and Polsby, Daniel, “Race, Religion, and Public Policy: Bob Jones University v. United States,” 1983 Supreme Court Review 1, 22–26Google Scholar. I find the second of these propositions implausible, given the Court's willingness in other contexts — such as that of religious-based conscientious objection to compulsory military service — to authorize inquiries into sincerity. Not also that the first proposition imposes a rather rationalistic frame on assertions of religious belief.

Cases involving statutory accommodations of religious belief express similar judgments. Once again, law can take religion into account only because, and to the extent that, it is not socially significant. Thus, the Court interpreted Congress' requirement that employers accommodate their practices to the religious beliefs of their employees to demand only de minimis adjustments, that is, accommodations that had only the most minor impact on the employers' business decisions. Trans World Airlines v. Hardison, 432 US. 63 (1977)Google Scholar. See also Thornton v. Caldor, Inc., 472 US. 703 (1985)Google Scholar (statute requiring employers to give “Sabbath” off unconstitutional because it failed to take account of the hardships such an accommodation would cause employers and co-workers).

72. This is not to deny that exemptions are significant in alleviating the pressures felt by individual believers; it is to claim simply that the Constitution requires the alleviation of individual burdens only if doing so is not socially significant.

73. Lemon v. Kurtzman, 403 U.S. 602, 612–13 (1971)Google Scholar. For an application of the purpose branch of the test, see Stone v. Graham, 449 US. 39 (1980)Google Scholar (invalidating statute requiring that Ten Commandments be posted in schoolrooms). Using this test the Court invalidated Alabama's “moment of silence” statute because its examination of the legislative history persuaded it that the sole motivation for the statute was to promote religious observance in the public schools. Wallace v. Jaffree, 472 US. 38 (1985).Google Scholar

74. Howe, Mark DeWolfe, The Garden and the Wilderness: Religion and Government in American Constitutional History 11 (1965).Google Scholar

75. Walz v. Tax Commission, 397 US. 664 (1970).Google Scholar

76. Marsh v. Chambers, 463 US. 783 (1983).Google Scholar

77. Marsh v. Chambers suggests that de facto establishments might be those created contemporaneously with the Constitution. The Court in this case upheld the practice of paying the salary of a chaplain who opened legislative sessions with a prayer. Instead of relying on its settled three-part test for evaluating establishment clause claims, the Court emphasized the “unique history” of the practice, noting for example that three days after Congress authorized the appointment of chaplains, it approved the text of the first amendment. Id., p. 791.

A “contemporaneous interpretation” approach to de facto establishments is likely to be unsatisfactory. Changes in social institutions make it difficult to be confident that a present-day practice is sufficiently similar to one in use at the time of the framing to warrant characterizing the institutions as the same. For example, property tax exemptions for churches are different in a society like ours, when major portions of a state's activities are financed by property tax revenues, from what they were two centuries ago, when public activities were much more limited in scope and financed largely by excise taxes. Walz v. Tax Commission, 397 US. 664 (1970), upheld such tax exemptions, relying on a slightly broader historical approach. The Court mentioned general establishment clause tests, but it gave more emphasis to the proposition that “few concepts are more deeply rooted in the fabric of our national life … than for the government to exercise … this kind of benevolent neutrality.” Id., p. 696. This approach sees de facto establishments as certain practices deeply entrenched in our traditions without regard to the precise time of their origin.

78. 465 US. 668 (1984).

79. Id., at 720 (Brennan, J., dissenting). For much of the period before the Civil War, the Alabama legislature routinely met on Christmas. See Thornton, J. Mills, Politics and Power in a Slave Society 82 (1978).Google Scholar

80. 465 U.S., pp. 674–78, 680, 685 (“display engenders a friendly community spirit of good will in keeping with the season”). See also id., p. 727 (Blackmun, J., dissenting (crèche “has been relegated to the role of a neutral harbinger of the holiday season, useful for commercial purposes.”).

81. Id., p. 681.

82. See also id., p. 685 (crèche has “religious implications”); 687 (crèche has “religious significance”).

83. See id., pp. 681–85, 686.

84. This theme is almost explicit in Justice O'Connor's concurring opinion, which makes dispositive “the message the crèche conveyed.” Id., p. 690. She concluded that the particular créche “was [not] intended to endorse [and did not have] the effect of endorsing Christianity.” Id., p. 694.

85. The Court's decisions prohibiting the practice of state-sponsored group prayer and Bible reading in public schools pose a problem for this analysis, for the practice would seem to be consistent with a system of public support for diffuse religiosity. It may be that the state's role in composing the prayer in Engel v. Vitale, 370 U.S. 421 (1962)Google Scholar, was too obvious an intrusion of the state into religious matters and that the use of the Bible in Abington School Dist. v. Schempp, 374 U.S. 203 (1963)Google Scholar, was insufficiently diffuse in its religiosity. As the Court saw it these practices were simply motivated by religion (see text accompanying note 94) and therefore went beyond the bounds of diffuse religiosity in public life.

86. See Bellah, , Broken CovenantGoogle Scholar. Bellah's own definition of civil religion is that it is a true religion, not “religion in general” or diffuse religiosity. This definition and its use in the U.S. context are quite controversial and probably lack substantial scholarly support. For a collection of the basic materials, see American Civil Religion (Russell Richey and Donald Jones eds. 1974)Google Scholar. See also Demerath, N. J. and Williams, Rhys, “Civil Religion in an Uncivil Society,” 480 Annals of the American Academy of Political and Social Science 154 (1985)CrossRefGoogle Scholar; Kelly, George, Politics and Religious Consciousness in America 209–46 (1984)Google Scholar. For another discussion of diffuse religiosity in constitutional law, see Van Alstyne, , “Trends in the Supreme Court,” pp. 786–87.Google Scholar

87. Zorach v. Clauson, 343 US. 306, 313 (1952).Google Scholar

88. Cf. McDaniel v. Paty, 435 US. 618 (1978).Google Scholar

89. For an especially dramatic criticism of the influence of concrete religiosity on public life, see Lewis, Anthony, “Onward, Christian Soldiers,” New York Times, p. A27, 03 10, 1983.Google Scholar

90. Justice Stevens has emphasized the tradition of religious opposition to public support of religion, based on the view that such support actually undermines religion. See, e.g., Roemer v. Board of Public Works, 426 US. 736, 775 (1976)Google Scholar (Stevens, J., dissenting) (noting “the pernicious tendency of a state subsidy to tempt religious schools to compromise their religious mission without wholly abandoning it”); Wolman v. Walter, 433 US. 229, 264 (1977)Google Scholar (Stevens, J., dissenting).

91. One potential outcome should be noted, although I think it unlikely to be realized very often. Sects might agree to support public endorsement of what might be thought of as a religious fair, in which the sects openly displayed their diverse views on concrete religious matters. Such a fair might exclude some sects. I assume, for example, that the fair would exclude Jews and Buddhists. Two points should be noted. Sectarian opposition to public support of religion may make it difficult to assemble sufficient support for a religious fair. Further, the message of a fair, even one that excludes some sects, is that religiosity in general is worthwhile. Thus, a fair consisting of many religious “booths” still supports no more than the civil religion.

92. See generally Anderson, Perry, Lineages of the Absolutist State (1974)Google Scholar; Wallerstein, Immanuel, The Modern World-System (text ed. 1976)Google Scholar. For a summary of the argument as made in liberation theology, see Cox, Harvey, Religion in the Secular City 93–96, 163–64 (1984).Google Scholar

93. The term is obviously metaphorically inappropriate; before the nationstate emerged there was nothing for these institutions to stand as intermediaries against.

94. This is the widely noted significance of the Protestant translations of the Bible into the vernacular. The translations enable believers to grasp their religion without having to learn an esoteric language. See also Mead, Sidney, “The ‘Nation with the Soul of a Church,’” in American Civil Religion, pp. 45, 51Google Scholar (homology of Protestantism and democracy); Garet, Ronald, “Communality and Existence: The Rights of Groups,” 56 Southern California Law Review 1001, 1031–32 (1983).Google Scholar

95. For a brief discussion of the tensions between religious privatism and religious communalities, see Roof, Wade and McKinney, William, “Denominational American and the New Religious Pluralism,” 480 Annals of the American Academy of Political and Social Science 24 (1985).CrossRefGoogle Scholar

96. The dynamics of economic growth in the United States had particularly strong effects on undermining the republican tradition, so that the development of the relation between church and state in the United States differed from that in other liberal societies. I therefore do not claim that the constitutional law of religion, as it has developed in the United States, is somehow the necessary working out of the implications of the liberal tradition; rather it is how the liberal and republican traditions evolved in the United States.

97. For a classic expression, See Stigler, George and Becker, Gary, “De Gustibus Non Est Disputandum,” 67 American Economic Review 76 (1977)Google Scholar. Intermediate institutions can also be used as instruments of public policy. Churches will be tolerated so long as, and to the extent that, they induce believers to act in ways congruent with public policies determined without reference to religious values. In this role intermediate institutions are not really intermediate; they are instead parts of the state, like a governmental bureaucracy or a school system.

98. Individuals have preferences that coincide in some matters and conflict in others. Temporary trade-offs may be possible, but long-term alliances are unlikely. Instability is promoted by the fact that such alliances provide opportunities for free riders, who remain outside the alliance but benefit from its activities. Even members of the alliance will sometimes be able to profit from strategic behavior by defecting to another alliance when the right issue comes along. See Olson, , Logic of Collective Action.Google Scholar

99. The most obvious expression would be a theory of strict separation of church and state. The liberal tradition's long-standing and still deeply felt attraction to strict separationist theories demonstrates that they fit well with that tradition.

100. See Garet, , “Communality and Existence,” p. 1009Google Scholar (“groups are possibilities for religious experience”); Palmer, Parker, A Company of Strangers: Christians and the Renewal of America's Public Life 20–31 (1981).Google Scholar

101. For definitions of religion that point in this direction, see Boyan, A. Stephen, “Defining Religion in Operational and Institutional Terms,” 116 University of Pennsylvania Law Review 479 (1968)CrossRefGoogle Scholar; Note, “Toward a Constitutional Definition of Religion,” 91 Harvard Law Review 1056 (1978)Google Scholar; Dodge, , “Free Exercise.”Google Scholar

102. In kind means that religious beliefs are not merely “deeper” or about “more fundamental” things than political ones. For definitions pointing to such considerations, see Note, “Sacred and Profane”; Choper, , “Defining ‘Religion,’” pp. 581–82Google Scholar. As has been frequently noted, this definition in fact does not distinguish religion from some kinds of political belief, such as those grounded in comprehensive social theories. See Clark, , “Guidelines,” pp. 339–43Google Scholar. The usual example is Marxism and associated bodies of political thought. Another example is liberalism itself, which asks us to take “on faith” a set of assumptions about human nature and capacity. The most dramatic illustration is Nozick, , Anarchy, State, and Utopia, p. ixGoogle Scholar (asserting without defense there or elsewhere in the book that “individuals have rights, and there are things no person or group may do to them [without violating their rights]”).

103. See Garet, , “Communality and Existence,” p. 1013Google Scholar (need to recognize “a group face of value”).

104. Liberals understood that too, which is why they incorporated a challenge to religion into their efforts. See text accompanying notes 93–94.

105. 370 US. 421 (1962); Sutherland, Arthur, “Establishment according to Engel,” 76 Harvard Law Review 25 (1962).Google Scholar

106. See, e.g., Brown, Ernest, “‘Quis Custodiet Ipsos Custodes?’—The School Prayer Cases,” 1963 Supreme Court Review 1Google Scholar; Kurland, Philip, “The Regents' Prayer Case: ‘Full of Sound and Furty, Signifying …,’” 1962 Supreme Court Review 1, 19–22Google Scholar. The case generated much more public controversy. See Elifson, Kirk and Hadaway, C. K., “Prayer in Public Schools: When Church and State Collide,” 49 Public Opinion Quarterly 317 (1985)CrossRefGoogle Scholar (reviewing opinion surveys from 1974 and 1980); Way, H. Frank, “Survey Research on Judicial Decisions: The Prayer and Bible Reading Cases,” 21 Western Political Quarterly 189 (1968).Google Scholar

107. Sutherland, , “Establishment,” p. 41Google Scholar. See Doremus v. Board of Education, 342 U.S. 429 (1952)Google Scholar, suggesting that more than de minimis expenditures are required to support standing to raise establishment clause claims. Engel was decided before Flast v. Cohen, 392 U.S. 83 (1968)Google Scholar, temporarily restructured the law of establishment clause standing.

108. Sutherland, , “Establishment,” p. 52.Google Scholar

109. See also Levy, Leonard, The Establishment Clause: Religion and the First Amendment 179 (1986).Google Scholar