Harvard Law Review Harvard Law Review Harvard Law Review

First Amendment: Religion

The Establishment Clause and the Chilling Effect

When the government passes a law, it often regulates conduct. People and organizations adjust their behavior to fit the legal rule; if a law forbids the destruction of certain financial records, for instance, organizations generally will not destroy those records. But regulations are sometimes unclear, uncertain, or overbroad, which can lead people to refrain from engaging in permissible actions because they are unsure whether they will be legally sanctioned. This phenomenon — a law’s effect on activity outside the scope of its intended target — is called the chilling effect.1×1. See Leslie Kendrick, Speech, Intent, and the Chilling Effect, 54 Wm. & Mary L. Rev. 1633, 1649 (2013). While chilling may happen in any number of contexts, it has been incorporated into free speech doctrine as a mechanism for policing statutes that are not aimed at regulating speech protected by the First Amendment, but have the incidental effect of doing so.2×2. See id. at 1649–50.

This Note considers chilling in the context of another part of the First Amendment: the Religion Clauses.3×3. U.S. Const. amend. I (“Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof . . . .”). It argues that government establishment of religion can chill the free exercise of religion, and that Establishment Clause jurisprudence should incorporate the concept of chilling. Government conduct related to religion, such as legislative prayer, can lead someone of a different (or no) religion to refrain from expressing their own beliefs, due to fear of reprisal from the government entity behind the establishment or the community at large. The person who is chilled could be religious or nonreligious; chilling of religious exercise applies equally to expressing religion and the lack thereof. Consider a simple example: A Muslim woman in a small, predominantly Christian town goes to a town board meeting, where board members will vote on zoning ordinances that will affect whether she can build a new building on her property. Each meeting opens with a Christian prayer offered by one of the board members, or perhaps there is an exhibit in the town hall that includes the Ten Commandments. The woman may well refrain from speaking about her religion, wearing religious garb, or otherwise expressing her religious beliefs in proximity to the meeting because she does not want to risk board members’ voting against her interests.4×4. Contrast this with coercion: if the woman felt she should participate in the prayers, rather than (or in addition to) refraining from expressing her own beliefs, she would be indirectly coerced. See section III.C.3, infra pp. 1355–57, for a more detailed discussion of chilling and coercion.

This Note argues that such chilling effects should be constitutionally cognizable, and that chilling should be a tool used in an Establishment Clause analysis to determine whether government action is an impermissible establishment of religion.5×5. This Note does not address procedural or jurisdictional issues related to chilling. In other words, it does not consider how chilling affects standing in Establishment Clause cases. While the chilling effect has played a large role in standing doctrine in free speech cases, most notably through overbreadth and facial challenges, see generally Richard H. Fallon, Jr., Making Sense of Overbreadth, 100 Yale L.J. 853 (1991), examining the jurisdictional effects of chilling in the establishment context is a topic for another time. This Note looks at chilling only as a substantive component of Establishment Clause jurisprudence. The idea that religion can be chilled is not wholly foreign to the case law. Plaintiffs have brought First Amendment challenges that refer to their religious conduct being “chilled” by various regulations and government actions.6×6. See, e.g., United States v. Aguilar, 883 F.2d 662, 697 (9th Cir. 1989) (describing plaintiffs’ argument that their religious exercise was being chilled); Nebraska ex rel. Bruning v. U.S. Dep’t of Health & Human Servs., 877 F. Supp. 2d 777, 788 (D. Neb. 2012) (same); Allen v. Sch. Bd. for Santa Rosa Cty., 782 F. Supp. 2d 1304, 1316 (N.D. Fla. 2011) (same). And while the issue of chilling religion does not come up frequently, courts that have mentioned it have accepted the proposition that religious exercise can be chilled.7×7. See, e.g., Catholic High Sch. Ass’n v. Culvert, 753 F.2d 1161, 1170 (2d Cir. 1985) (“But a lingering question remains as to whether State Board jurisdiction may impermissibly chill free exercise rights . . . .”); Int’l Soc’y for Krishna Consciousness, Inc. v. Barber, 650 F.2d 430, 442–43 (2d Cir. 1981) (“[T]he missionary evangelism at the core of the Chaitanya movement can only be sustained by voluntary contributions, and therefore a restraint on soliciting funds would chill the exercise of the underlying religious activity as well.” (citing Murdock v. Pennsylvania, 319 U.S. 105, 108–09 (1943))); Catholic Bishop of Chi. v. NLRB, 559 F.2d 1112, 1124 (7th Cir. 1977) (“The real difficulty is found in the chilling aspect that the requirement of bargaining will impose on the exercise of the bishops’ control of the religious mission of the schools.”), aff’d, 440 U.S. 490 (1979); Allen, 782 F. Supp. 2d at 1316 (“[T]he employees’ allegations of chill to their right to free speech and religious exercise . . . sufficiently state an injury.”); United States v. Roach, 947 F. Supp. 872, 876 (E.D. Pa. 1996) (“FACE does not chill First Amendment freedom of speech or religion on its face or as applied in the matter under consideration.” (citing United States v. Dinwiddie, 76 F.3d 913, 921 (8th Cir. 1996))). Most of the cases in which courts have referenced a chilling effect on religion are Free Exercise Clause claims; chilling is generally mentioned in determining whether there has been a burden on religion for which the government must provide a compelling state interest.8×8. See, e.g., Cheffer v. Reno, 55 F.3d 1517, 1522 (11th Cir. 1995) (“[T]he appellants here have not argued that the Access Act ‘substantially burdens’ their religious practice. Appellants’ brief on appeal merely asserts that they have a sincerely held religious belief that abortion is murder, and that the Access Act chills their expression of that belief.”); Culvert, 753 F.2d at 1170–71 (“[A] lingering question remains as to whether State Board jurisdiction may impermissibly chill free exercise rights . . . . It is necessary, then, to decide whether this indirect and incidental burden on religion is justified by a compelling state interest.”); Dale v. Tyler, No. CIV. 14-4102, 2015 WL 4167366, at *6 (D.S.D. July 8, 2015) (“Dale has alleged that . . . defendants have curtailed his ability to express adherence to his religion by . . . filing false disciplinary reports against him in an effort to chill the free exercise of his religion. For purposes of a motion to dismiss, therefore, the court finds that Dale has sufficiently alleged that defendants have imposed a substantial burden on his ability to practice his religion.”). This Note makes the case that, whatever chilling’s role may be in a free exercise case, it should be used in an Establishment Clause analysis. As it stands now, establishment jurisprudence focuses on harms like coerced religious participation,9×9. See, e.g., Town of Greece v. Galloway, 572 U.S. 565, 586 (2014) (plurality opinion). government favoritism and endorsement,10×10. See, e.g., Lynch v. Donnelly, 465 U.S. 668, 687–88 (1984) (O’Connor, J., concurring). and entanglement between government and religious institutions.11×11. See, e.g., Trinity Lutheran Church of Columbia, Inc. v. Comer, 137 S. Ct. 2012, 2036 (2017) (Sotomayor, J., dissenting); Town of Greece, 572 U.S. at 586; Lynch, 465 U.S. at 678–88 (O’Connor, J., concurring). It looks at establishment as the government injecting religion into the public square. But establishment also removes religion from public life. This harm is intuitive in many circumstances — if the government required someone to attend a certain church on Sunday mornings, they would not be able to attend their own church at that time — but the doctrine does not fully recognize the issue of what goes missing when the government establishes a religion. The chilling effect that establishment has on religious exercise is a primary instance of this harm of omission. Incorporating the chilling effect into establishment jurisprudence gives religion and nonreligion more room to exist in the public square and honors the Establishment Clause’s role as a “co-guarantor” of religious liberty.

This Note proceeds in three Parts. Part I provides an overview of the concept and doctrine of chilling in Free Speech Clause jurisprudence. Part II explains how establishment can chill free exercise and makes the normative case for recognizing chilling. Part III explores how to incorporate the chilling effect doctrinally into Establishment Clause jurisprudence and the implications for the doctrine.

I. Chilling in the Free Speech Context

The case law on chilling speech dates back to the mid-twentieth century.12×12. The word “chill” first appeared in a Supreme Court case in 1952. See Wieman v. Updegraff, 344 U.S. 183, 195 (1952) (Frankfurter, J., concurring). The phrase “chilling effect” appeared eleven years later. See Gibson v. Fla. Legislative Investigation Comm., 372 U.S. 539, 557 (1963). At its most basic, chilling is the idea that government regulation or action can deter certain conduct. Of course, much of the purpose of law is to deter, so “chilling” as a constitutional concern cannot refer to the conduct toward which a regulation is specifically directed. It refers instead to effects on activities that are outside the scope of a law’s target. As Professor Frederick Schauer defines it in the free speech context, “[a] chilling effect occurs when individuals seeking to engage in activity protected by the first amendment are deterred from doing so by governmental regulation not specifically directed at that protected activity.”13×13. Frederick Schauer, Fear, Risk and the First Amendment: Unraveling the “Chilling Effect,” 58 B.U. L. Rev. 685, 693 (1978) (emphasis omitted); see also Kendrick, supra note 1, at 1649 (“‘Chilling’ denotes overdeterrence of benign conduct that occurs incidentally to a law’s legitimate purpose or scope.” (italics omitted)). A chilling effect thus arises when “an otherwise legitimate regulation has the incidental effect of deterring — or chilling — benign activity.”14×14. Kendrick, supra note 1, at 1649.

Schauer theorizes that chilling is a response to the uncertainty and imperfections of the legal system, which lead people to fear their actions will be punished even if, in a perfect world, they would and should not be.15×15. Schauer, supra note 13, at 687–88, 693. This plays out in one of the most notable doctrinal uses of chilling: overbreadth. Under the doctrine of overbreadth, a litigant may bring a facial challenge to a statute that is “overbroad” — one that reaches both unprotected and protected speech — even if the challenger could be properly prosecuted under a more narrowly drawn law.16×16. Henry Paul Monaghan, Overbreadth, 1981 Sup. Ct. Rev. 1, 1–2. For classic discussions of overbreadth, see generally Fallon, supra note 5; and Monaghan, supra. A major justification for the doctrine is that a party whose speech is protected may choose not to speak out of fear of prosecution, and therefore lose the opportunity to challenge the overbroad law.17×17. E.g., Kendrick, supra note 1, at 1653. Overbreadth is often thought of as a “prophylactic” doctrine meant to avoid such a chilling effect, though chilling is not its sole foundation.18×18. Fallon, supra note 5, at 855. But see Monaghan, supra note 16, at 4–14 (arguing that overbreadth is based on “constitutionally imposed limits on the power of courts to narrow statutes in the process of applying them,” id. at 13, without considering the concept of chilling).

Overbreadth is one of the most notable circumstances in which a speaker might be chilled, but there are others.19×19. For a detailed list of the ways in which a speaker might be chilled and accompanying case law, see Kendrick, supra note 1, at 1654 & nn.100–06. The void-for-vagueness doctrine prohibits vague and uncertain laws.20×20. See, e.g., Grayned v. City of Rockford, 408 U.S. 104, 108–09 (1972). While vagueness is most prominently a due process issue, in the First Amendment context there are concerns about a chilling effect, because uncertainty about a law’s scope means “speakers who would otherwise engage in protected speech accordingly self-censor.”21×21. Kendrick, supra note 1, at 1653; see also Grayned, 408 U.S. at 109 (“[W]here a vague statute ‘abut[s] upon sensitive areas of basic First Amendment freedoms,’ it ‘operates to inhibit the exercise of [those] freedoms.’” (second and third alterations in original) (footnotes omitted)). The risks and costs of litigation and liability could also cause chilling, a concern raised in libel cases: if the speaker bears the burden of proving the truth of her statements, rather than the aggrieved person having to prove falsity, then the speaker may refrain from speaking out of fear of the expense of proving truth in court.22×22. See N.Y. Times Co. v. Sullivan, 376 U.S. 254, 279 (1964). Or an overly complex law might chill speech because the costs of consulting with legal counsel to determine what speech is allowed could lead someone to refrain from speaking in the first place.23×23. See Citizens United v. FEC, 558 U.S. 310, 324 (2010) (“The First Amendment does not permit laws that force speakers to retain a campaign finance attorney . . . .”). So too could the threat of surveillance cause people to stop speaking.24×24. See, e.g., Scott Michelman, Who Can Sue over Government Surveillance?, 57 UCLA L. Rev. 71, 78 (2009); Dawinder S. Sidhu, The Chilling Effect of Government Surveillance Programs on the Use of the Internet by Muslim-Americans, 7 U. Md. L.J. Race Religion Gender & Class 375, 376 (2007).

Of course, many activities other than speech can be chilled. In theory, any activity protected by a “positive” constitutional guarantee could be deterred by regulations not directed at that activity.25×25. Schauer, supra note 13, at 692. Professor Toni Massaro argues that overbreadth doctrine should be expanded to apply to all constitutional rights. Toni M. Massaro, Chilling Rights, 88 U. Colo. L. Rev. 33, 39 (2017). The privilege against self-incrimination26×26. See U.S. Const. amend. V. is a classic example: permitting the prosecution to comment to the jury upon the fact that a defendant invoked her Fifth Amendment right and did not take the stand might deter the defendant from exercising that right.27×27. See Griffin v. California, 380 U.S. 609, 613–14 (1965). In the case of abortion rights, laws prohibiting late-term abortions have had a chilling effect on women’s decisions to get abortions in the month before the late-term period.28×28. Brandice Canes-Wrone & Michael C. Dorf, Measuring the Chilling Effect, 90 N.Y.U. L. Rev. 1095, 1098 (2015).

And yet the chilling effect has been applied only in free speech cases. A standard account for its presence is that the First Amendment has “a special status in the constitutional scheme.”29×29. Fallon, supra note 5, at 867. Speech, on this account, is a “transcendent value,”30×30. Speiser v. Randall, 357 U.S. 513, 526 (1958). and therefore is entitled to special protections. The “affirmative value”31×31. Schauer, supra note 13, at 691. of speech means the government should actively encourage it, not simply refrain from infringing on it, because speech is desirable and advantageous for society.32×32. Id. at 691–92. It is so “supremely precious”33×33. NAACP v. Button, 371 U.S. 415, 433 (1963). that we have created a special doctrinal formulation — chilling — that effectively subsidizes speech.34×34. Kendrick, supra note 1, at 1651 n.88. On another account, free expression should be protected from chilling because speech is a public good, rather than a transcendent right.35×35. See Daniel A. Farber, Commentary, Free Speech Without Romance: Public Choice and the First Amendment, 105 Harv. L. Rev. 554, 558–60 (1991); Richard A. Posner, Free Speech in an Economic Perspective, 20 Suffolk U. L. Rev. 1, 19–24 (1986). Because much of the value of speech is social, rather than personal, there is a risk that speech will be underproduced.36×36. Farber, supra note 35, at 558–60. The doctrine of chilling may serve to moderate this underproduction.

II. Chilling in the Establishment Clause Context

Incorporating the chilling effect into the Establishment Clause is not simply a matter of copying and pasting from the free speech context. Rather, doing so requires examining the process by which religious exercise can be chilled and focusing more on the government’s expressive conduct, in addition to government regulation. This Part presents the theory behind establishment’s chilling effect on religious expression and the normative justifications for incorporating chilling into the jurisprudence.

A. The Theory of Chilling Through Establishment

The basic predicate for the idea that establishment can chill religious exercise is that government conduct is expressive. When the government acts, it expresses some meaning, and in doing so it can “prescribe orthodoxy”: the social meaning embedded in the action or law becomes the “right” idea.37×37. Lawrence Lessig, The Regulation of Social Meaning, 62 U. Chi. L. Rev. 943, 946–47 (1995) (internal quotations omitted); see also Dan M. Kahan, Social Influence, Social Meaning, and Deterrence, 83 Va. L. Rev. 349, 362 (1997); Richard H. Pildes, Why Rights Are Not Trumps: Social Meanings, Expressive Harms, and Constitutionalism, 27 J. Legal Stud. 725, 726 (1998). People may then adapt their behavior as a reaction to the embedded meaning. Scholars of the Religion Clauses have taken the general idea of expressive conduct and applied it to the Establishment Clause.38×38. For work on the expressive nature of conduct implicating the Establishment Clause, see David Cole, Faith and Funding: Toward an Expressivist Model of the Establishment Clause, 75 S. Cal. L. Rev. 559 (2002); Alex Geisinger & Ivan E. Bodensteiner, An Expressive Jurisprudence of the Establishment Clause, 112 Pa. St. L. Rev. 77 (2007); Mark D. Rosen, Establishment, Expressivism, and Federalism, 78 Chi.-Kent L. Rev. 669 (2003); and Richard C. Schragger, The Role of the Local in the Doctrine and Discourse of Religious Liberty, 117 Harv. L. Rev. 1810, 1874–91 (2004). Professors Alex Geisinger and Ivan Bodensteiner have detailed how people come to understand and respond to the government’s expressive religious conduct.39×39. Geisinger & Bodensteiner, supra note 38, at 112–18. Government conduct related to religion, such as the display of a religious symbol, provides information about the government’s and the community’s beliefs, either directly or through inference.40×40. Id. at 112–13. Geisinger and Bodensteiner explain that the reasonableness of these inferences, and the legal validity of the preexisting beliefs that underly them, can be determined by consistent and logically sound means for the purposes of finding an Establishment Clause violation under their expressive theory. Id. at 120–22. If the information expressed indicates the religious preference is strongly held or held by many, people become more certain that acting in a way that is different from those beliefs will harm their position in the community and the “cooperative benefits”41×41. Id. at 115. that go along with being held in esteem.42×42. Id. at 114–15. A person may thus adapt her behavior to align more closely with the information about beliefs provided by or inferred from the government act. Someone outside a perceived favored religion feels they must adjust their own religious behavior in order to remain a full member of the community and reap the associated benefits.

Geisinger and Bodensteiner use this explanation to propose an Establishment Clause test based on the government’s expressive conduct.43×43. See id. at 120 (“[U]nder the expressive theory, the Establishment Clause is violated if a government act provides information or leads to reasonable inferences drawn from legally valid pre-existing beliefs that an individual is favored or disfavored by government, or in the community, as a result of his or her religious beliefs.”). But the explanation works just as well as a demonstration of chilling in the establishment context: it shows why someone might refrain from engaging in religious expression even though the government’s act is not targeted at or meant to suppress that expression. It also shows how both government and community beliefs are relevant considerations when analyzing the role of establishment in chilling religious exercise. When the government affiliates itself with religion in some way, it is taken to be expressing not only its own thoughts on religion, but also the community’s, because the government is (in theory) representative of the community.44×44. See id. at 113; see also Lynch v. Donnelly, 465 U.S. 668, 688 (1984) (O’Connor, J., concurring) (“Endorsement sends a message 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. Disapproval sends the opposite message.” (citing Abington Sch. Dist. v. Schempp, 374 U.S. 203 (1963))). In this way, establishment can represent both governmental and social power. Thus, the fear of the person being chilled is not only that she may not receive certain benefits from the government if she expresses a different belief;45×45. See Geisinger & Bodensteiner, supra note 38, at 115–16 (noting that the “cooperative benefits,” id. at 115, from being held in esteem are particularly important when the “cooperative partner,” id. at 116, is the government, which has the often-exclusive power to confer certain benefits). it is also that expressing a belief at odds with the government’s will lead to social sanction. The fear underlying the chilling effect in speech — which Schauer posits is the fear, stemming from uncertainty in the legal system, of being haled into court — changes into a fear of informal government and community penalties.46×46. Such a fear underlies the current case law on indirect coercion. See infra p. 1356.

A recent real-world example shows how chilling can happen in the establishment context and how it implicates both government and community sanctions. Amber Guyger, a white police officer in Dallas, Texas, was convicted of murder for shooting Botham Jean, a black man, in his apartment when she allegedly mistook it for her own.47×47. Marina Trahan Martinez et al., Amber Guyger Is Sentenced to 10 Years for Murder of Botham Jean, N.Y. Times (Oct. 8, 2019), https://nyti.ms/2owYv2p [https://perma.cc/R5J8-KNMA]. At the sentencing, Jean’s brother gave an emotional statement in which he forgave Guyger, and then hugged her.48×48. Id. The judge then retrieved her personal Bible, gave it to Guyger, and directed her to John 3:16, a famous passage about salvation through Jesus.49×49. Sarah Mervosh & Nicholas Bogel-Burroughs, Amber Guyger’s Judge Gave Her a Bible and a Hug. Did That Cross a Line?, N.Y. Times (Oct. 4, 2019), https://nyti.ms/2ImGpHu [https://perma.cc/7M6V-FMHX]. There is a strong argument that the judge violated the Establishment Clause by coercing Guyger into accepting and expressing a Christian message and by endorsing Christianity. But her actions also risked chilling Guyger, a criminal defendant subject to a proceeding in which the judge had extraordinary power over her: Why would Guyger now publicly express any belief at odds with what the judge believes? In addition to the risk of being penalized by a government official, any expression Guyger might make about her faith (or lack thereof) that goes against what the judge expressed in the courtroom will likely lead to social pushback. Many praised Jean’s brother and the judge for their compassion;50×50. Hannah Knowles, Amber Guyger Was Hugged by Her Victim’s Brother and a Judge, Igniting a Debate About Forgiveness and Race, Wash. Post (Oct. 3, 2019, 2:35 PM), https://www.washingtonpost.com/nation/2019/10/03/judge-botham-jeans-brother-hugged-amber-guyger-igniting-debate-about-forgiveness-race [https://perma.cc/XCZ4-LF4G]; Mervosh & Bogel-Burroughs, supra note 49; see also Zak Cheney-Rice, Who Gets Grace in the Courtroom?, N.Y. Mag. (Oct. 9, 2019), http://nymag.com/intelligencer/2019/10/judge-kemp-hugs-amber-guyger-after-botham-jean-case.html [https://perma.cc/F2R9-3GRM]. if Guyger says or does something not in line with the beliefs both expressed, she risks social disapproval. The smartest move for someone in Guyger’s position is to not rock the boat, and thus any religious expression not in line with what the judge expressed in the courtroom is chilled.

This explanation also shows how chilling in the establishment context differs from chilling in the free speech context, through the focus on government expression rather than regulation. As explained in Part I, the chilling effect on speech involves government regulation that leads people to refrain from speaking because they might be brought into court under the statute. In most establishment cases, on the other hand, the government itself undertakes some action (offering a prayer, providing funding to a religious institution, maintaining a religious monument) that is not necessarily done in order to regulate others’ conduct.51×51. This is not to say that government regulation never raises Establishment Clause concerns. Many of the establishment cases about governmental intrusion into religious organizations’ affairs involve regulations, such as Title VII, though those cases generally implicate the Free Exercise Clause as well. See, e.g., Corp. of the Presiding Bishop of the Church of Jesus Christ of Latter-Day Saints v. Amos, 483 U.S. 327, 329–30 (1987). The establishment issue requires considering the context of the conduct and what kind of message it expresses, in order to determine how it may affect an individual and whether that effect is of constitutional concern. While a regulation’s chilling effect on speech is generally based on an analysis of the text of the regulation itself, the chilling effect of the government’s religious conduct is based on the expressive meaning of the action and the reasonable inferences about government and community beliefs drawn from it.52×52. See supra pp. 1343–44 & n.40. In other words, religious expression can be deterred not simply by government regulation, but also by government expression.

B. The Normative Case for Incorporating Chilling

This Part now turns to making the normative case for incorporating the chilling effect into establishment jurisprudence. It is not enough to say that, because religious establishment causes some chilling, the chilling effect must be recognized as a constitutional matter. Every regulation or government action can have a chilling effect, sometimes inescapably so.53×53. See supra p. 1342. Instead, the normative foundations for incorporating chilling include religion’s place in the constitutional order and the Establishment Clause’s role in helping to preserve religious freedom.

1. Religion’s Constitutional Place. — Religion is often thought to hold a special place in the constitutional scheme, and incorporating chilling into establishment jurisprudence acknowledges this value. Historical tradition has recognized a special role for religion.54×54. See, e.g., Van Orden v. Perry, 545 U.S. 677, 683 (2005) (plurality opinion) (remarking on the “strong role played by religion and religious traditions throughout our Nation’s history”); Thomas C. Berg, Minority Religions and the Religion Clauses, 82 Wash. U. L.Q. 919, 932–36 (2004). But see Micah Schwartzman, What if Religion Is Not Special?, 79 U. Chi. L. Rev. 1351, 1354–55 (2012) (arguing against the view that religion warrants “special treatment”). Until the 1990 case Employment Division v. Smith,55×55. 494 U.S. 872 (1990). which held the First Amendment does not require religious exemptions from “neutral law[s] of general applicability,”56×56. Id. at 879 (quoting United States v. Lee, 455 U.S. 252, 263 n.3 (1982); and citing Minersville Sch. Dist. Bd. of Educ. v. Gobitis, 310 U.S. 586, 595 (1940)). the Constitution was thought to “affirmatively mandate[] accommodation, not merely tolerance, of all religions.”57×57. Lynch v. Donnelly, 465 U.S. 668, 673 (1984) (first citing Zorach v. Clauson, 343 U.S. 306, 314, 315 (1952); and then citing Illinois ex rel. McCollum v. Bd. of Educ., 333 U.S. 203, 211 (1948)); see also Wisconsin v. Yoder, 406 U.S. 205, 234–36 (1972); Sherbert v. Verner, 374 U.S. 398, 410 (1963). Even after Smith, the courts have been concerned with ensuring the government does not force religion out of public life. In Trinity Lutheran Church of Columbia, Inc. v. Comer,58×58. 137 S. Ct. 2012 (2017). for instance, the Supreme Court held that a state could not decline to offer a playground-resurfacing grant to a religious institution, finding that the state had infringed on the religious institution’s free exercise rights by penalizing it solely because it was religious.59×59. Id. at 2017, 2021–22, 2024. In Town of Greece v. Galloway,60×60. 572 U.S. 565 (2014). the Court was concerned about the government policing the content of legislative prayers and stripping them of sectarian references.61×61. Id. at 581. And in American Legion v. American Humanist Ass’n,62×62. 139 S. Ct. 2067 (2019). the Court worried that removing the large Peace Cross at issue would indicate the government was hostile to religion and wanted to remove it from the public eye.63×63. Id. at 2074 (quoting Van Orden v. Perry, 545 U.S. 677, 704 (2005) (Breyer, J., concurring in the judgment)). Given this special consideration for religion, it makes sense that establishment doctrine should accommodate the chilling effect. If government conduct is not permitted to chill religious expression, there will be more room for such expression in public life. Leaving room for religious expression recognizes the importance of permitting religion (and nonreligion) in the public sphere, which the case law has come to see as not only neutral, but good.64×64. This justification — that religion has a special place in the constitutional order and should have room to be present in the public square — does not run counter to the Establishment Clause by privileging religion over nonreligion. Establishment can chill someone’s expression of their lack of religion in the same way it chills expression of religion. Prohibiting establishment by using chilling simply makes space in the public sphere for religious or nonreligious exercise; it does not require the government to promote religion.

This justification is akin to the primary argument for chilling in the speech context: speech is considered an affirmative value for which the government should provide space, because of its place in the constitutional order. Schauer, however, differentiates speech and religion as constitutional values. As he explains, chilling “is most invidious where the underlying constitutionally protected activity is positively advantageous, rather than an activity which, for other reasons, ought to be shielded from state intrusion.”65×65. Schauer, supra note 13, at 692. He argues:

Freedom of religion arguably falls within this latter category. Although speech may be thought of as a positive virtue or affirmative good . . . religion does not appear to occupy a similar position in the constitutional order. Our concern appears to be not with the positive values of religion, but rather with the detrimental impact that results when government intrudes into the spiritual realm.
66×66. Id. at 692 n.36 (citation omitted).

But this argument is at odds with the ways the case law describes religion, as noted above. And even if religion is not itself a “positive” good, leaving space for religion surely is. The Religion Clauses are meant to provide a hospitable environment for religion and nonreligion, and there is affirmative value in leaving space for such expression. Adopting the concept of chilling in establishment jurisprudence recognizes that the government should not act, by affiliating itself with religion, to depress religious expression in the public sphere.67×67. Such a normative argument for incorporating chilling into establishment doctrine does not necessarily carry over to all questions of religious accommodations. Religious accommodations — such as exceptions from generally applicable regulations — may impose costs on third parties, as in the case of an employer refusing to provide health insurance that covers certain contraceptives, see Burwell v. Hobby Lobby Stores, Inc., 573 U.S. 682, 740 (2014) (Ginsburg, J., dissenting), or a wedding vendor who does not provide services to a same-sex couple, see Masterpiece Cakeshop, Ltd. v. Colo. Civil Rights Comm’n, 138 S. Ct. 1719, 1723 (2018). Not providing an accommodation may well chill religious expression in those cases. But any chill must be considered in tandem with the burden imposed on third parties, which alters the calculation. In the establishment context, on the other hand, the question is generally one of state action affecting private religious expression, without a third party involved. While there is a debate over the extent to which third-party harms should be considered when deciding accommodation requests, compare Mark Storslee, Religious Accommodation, the Establishment Clause, and Third-Party Harm, 86 U. Chi. L. Rev. 871, 876–78 (2019), with Frederick Mark Gedicks & Rebecca G. Van Tassell, RFRA Exemptions from the Contraception Mandate: An Unconstitutional Accommodation of Religion, 49 Harv. C.R.-C.L. L. Rev. 343, 346–49 (2014), the normative argument for chilling in this Note is not automatically transferable to that debate.

2. The Purpose of the Religion Clauses. — Incorporating chilling into establishment doctrine also recognizes how the Religion Clauses are intertwined and furthers the purpose of that connection. The Establishment Clause is often considered a “co-guarantor” of religious freedom,68×68. Abington Sch. Dist. v. Schempp, 374 U.S. 203, 256 (1963) (Brennan, J., concurring); see also Lee v. Weisman, 505 U.S. 577, 592 (1992). meant to work with the Free Exercise Clause to protect religious liberty.69×69. See Schempp, 374 U.S. at 232 (Brennan, J., concurring) (“The inclusion of both restraints upon the power of Congress to legislate concerning religious matters shows unmistakably that the Framers of the First Amendment were not content to rest the protection of religious liberty exclusively upon either clause.”). Much ink has been spilled on how, precisely, the two clauses interact, but the basic principle is that both clauses work to protect religious conscience.70×70. See, e.g., Arlin M. Adams & Charles J. Emmerich, A Heritage of Religious Liberty, 137 U. Pa. L. Rev. 1559, 1643 (1989); Alan Brownstein, The Religion Clauses as Mutually Reinforcing Mandates: Why the Arguments for Rigorously Enforcing the Free Exercise Clause and Establishment Clause Are Stronger When Both Clauses Are Taken Seriously, 32 Cardozo L. Rev. 1701, 1705 (2011); Noah Feldman, The Intellectual Origins of the Establishment Clause, 77 N.Y.U. L. Rev. 346, 350–51, 411–12 (2002). The prohibition on establishment functions to protect religious liberty because “governmental intervention in religious matters can itself endanger religious freedom.”71×71. Van Orden v. Perry, 545 U.S. 677, 683 (2005) (plurality opinion). When the government chills religious exercise by affiliating with religion, or engaging in some sort of religious conduct, it ignores its duty to preserve religious freedom. Establishment can harm religion in the public square by adding certain religious exercises and coercing people into participating, but it can also harm religion by removing it from public view and leaving only the orthodox religion. If the Establishment Clause is meant to help protect religious liberty, it is difficult to achieve the full constitutional guarantee if establishment jurisprudence ignores one of the ways in which establishment affects religious expression. The government crowds out religious exercise through chilling when its conduct violates the Establishment Clause. Chilling must therefore be legally cognizable in order to effectuate a promise of that Clause — preserving religious liberty.

3. The “Political Economy” of the Religion Clauses. — One interesting byproduct of incorporating chilling into establishment doctrine is the way in which doing so crosses the “political economy” of the Religion Clauses.72×72. See Noah R. Feldman & Kathleen M. Sullivan, Constitutional Law 1603–05 (20th ed. 2019). The “political economy” of religious exemptions explains how support for religious accommodations has shifted from one political side to the other over the years, though the idea can be expanded to other Religion Clauses issues as well. See Paul Horwitz, The Supreme Court, 2013 Term — Comment: The Hobby Lobby Moment, 128 Harv. L. Rev. 154, 159, 166–84 (2014); Storslee, supra note 67, at 872. The “liberal” position is one of stricter separation between church and state and a resistance to granting religious accommodations associated with social issues like LGBTQ rights and abortion; the “conservative” position is more accommodationist, allowing a relationship between religion and government as long as it is not coercive, as well as supporting religious accommodations for those operating in the marketplace.73×73. See, e.g., Horwitz, supra note 72, at 155–56; Storslee, supra note 67, at 872–73. The “conservative” view thus permits more government interaction with religion, whether through legislative prayer, religious monuments on public property, or funding for religious institutions. The “conservative” position has gained traction in the Supreme Court and the academy over the past twenty years, as the coercion test has grown in prominence and strict separationism has faded.74×74. See, e.g., Am. Legion v. Am. Humanist Ass’n, 139 S. Ct. 2067, 2074 (2019) (allowing a large, long-standing war memorial in the shape of a cross to remain on public land); Trinity Lutheran Church of Columbia, Inc. v. Comer, 137 S. Ct. 2012, 2024–25 (2017) (requiring states to allow religious institutions to receive a generally available grant); Town of Greece v. Galloway, 572 U.S. 565, 591–92 (2014) (permitting sectarian legislative prayer). As the accommodationist approach has gained steam, the Supreme Court has faced criticism for being more concerned with Christian than minority religious beliefs. See, e.g., Dahlia Lithwick, An Execution Without an Imam, Slate (Feb. 8, 2019, 2:56 PM), https://slate.com/news-and-politics/2019/02/domineque-ray-alabama-execution-imam-first-amendment-scotus.html [https://perma.cc/39KD-BDW4].

At first glance, it would seem that incorporating chilling is in line with the “liberal” view of the Establishment Clause: doing so will likely lead courts to find impermissible some government conduct that the coercion test, as currently formulated, would allow, and thus will shrink the range of what the government can do.75×75. On the issue of how recognizing chilling changes the coercion test, see infra section III.C.3, pp. 1355–58. But prohibiting government conduct that chills religious expression should lead to more people exercising their religion, or lack thereof, in the public sphere, because the government’s conduct will no longer chill their participation. Furthermore, the justifications for incorporating chilling view religious expression not just as an important constitutional guarantee, but as having social value that should be supported. The idea of supporting religion in the public square — and the value of having it there — has come to be associated with the “conservative” position;76×76. Before 1990, the “liberal” position was one of advocating for religious exemptions, in order to protect religious minorities, while the “conservative” view was to reject exemptions from generally applicable laws (it was Justice Scalia, after all, who wrote Smith). See generally John C. Jeffries, Jr., & James E. Ryan, A Political History of the Establishment Clause, 100 Mich. L. Rev. 279 (2001) (providing an overview of the shifting position of evangelical Christians to support greater accommodation). That changed after Smith, and now granting exemptions — at least when related to “culture war” issues — is seen as a “conservative” position. See Storslee, supra note 67, at 872–73. many commentators on that side critique what they see as a liberal attempt to drive religion out of public life.77×77. See, e.g., Alexandra M. McPhee, A Christian War Memorial in No Way Violates the Establishment Clause, Nat’l Rev. (Feb. 15, 2019, 6:30 AM) https://www.nationalreview.com/2019/02/peace-cross-supreme-court-case-establishment-cause [https://perma.cc/WM2R-HBZY] (responding to “opponents who would stifle all religious expression in the public square”). Though recognizing chilling may narrow the range of what the government can do, doing so correspondingly increases how much private individuals and organizations engage in religious expression. Thus, incorporating chilling does not necessarily go against the grain of the “conservative” position, because doing so recognizes that expression of religion and nonreligion in public spaces is an affirmative value. For those who are concerned that a strong reading of the Establishment Clause pushes religion out of public life,78×78. Cf. Van Orden v. Perry, 545 U.S. 677, 683 (2005) (plurality opinion) (discussing how Establishment Clause cases “point in two directions” and “present[] us with the difficulty of respecting both faces”). chilling demonstrates how the Establishment Clause can help keep religion in the public square. It thus crosses the current political economy of the Religion Clauses.

III. Incorporating Chilling into Establishment Clause Doctrine

Having established what chilling in the establishment context looks like and why it should be incorporated, the next question is what it looks like to integrate the chilling effect doctrinally. This Part explains current establishment jurisprudence, notes cases that have already mentioned the chilling effect in Establishment Clause inquiries, and finally shows how chilling should be used in the doctrine and the consequences of incorporating it. Ultimately, this Note envisions chilling not as a separate doctrine — a standalone test meant to replace the current doctrinal frameworks — but rather as a tool to be used within any Establishment Clause analysis. The question of whether government affiliation with religion chills religious exercise is a step courts should undertake as part of their inquiry into any establishment case.

A. Current Establishment Clause Doctrine

Current establishment jurisprudence is widely acknowledged to be a mess.79×79. Supreme Court Justices have called it “embarrassing,” Edwards v. Aguillard, 482 U.S. 578, 639 (1987) (Scalia, J., dissenting), “silly,” Elk Grove Unified Sch. Dist. v. Newdow, 542 U.S. 1, 45 n.1 (2004) (Thomas, J., concurring in the judgment), and “irreconcilable with the imperative of applying neutral principles in constitutional adjudication,” Cty. of Allegheny v. ACLU, Greater Pittsburgh Chapter, 492 U.S. 573, 676 (1989) (Kennedy, J., concurring in the judgment in part and dissenting in part). The confusion stems in large part from the fact the Court has never settled on any one test, instead using a hodgepodge of standards and deciding cases on fact-specific grounds.80×80. See, e.g., Lund v. Rowan Cty., 863 F.3d 268, 281 (4th Cir. 2017) (en banc) (conducting a “fact-sensitive” analysis of a legislative prayer practice); see also Recent Case, Lund v. Rowan County, 683 F.3d 268 (4th Cir. 2017) (en banc), 131 Harv. L. Rev. 626, 632 (2017) (explaining the difficulties with such unanchored analyses). In broad terms, there are three major tests. First is the much-maligned Lemon81×81. Lemon v. Kurtzman, 403 U.S. 602 (1971). test, under which a state action survives an establishment challenge if the act (1) has a secular purpose,82×82. Id. at 612. (2) does not have the primary effect of advancing or inhibiting religion,83×83. Id. (citing Bd. of Educ. v. Allen, 392 U.S. 236, 243 (1968)). and (3) does not result in “excessive government entanglement with religion.”84×84. Id. at 613 (quoting Walz v. Tax Comm’n, 397 U.S. 664, 674 (1970)). Academics and courts alike have panned Lemon for the indeterminacy of its standards.85×85. See, e.g., Am. Legion v. Am. Humanist Ass’n, 139 S. Ct. 2067, 2080–81 (2019) (plurality opinion); Cynthia V. Ward, Essay, Coercion and Choice Under the Establishment Clause, 39 U.C. Davis L. Rev. 1621, 1628–29 (2006) (explaining critiques of the Lemon test, such as how to determine what makes a purpose “secular” or an effect “primary,” id. at 1628). For all the criticism, however, the Supreme Court has never found a majority to formally overrule it, so Lemon continues to haunt the Court’s establishment jurisprudence.86×86. Justice Scalia put it as only Justice Scalia could: “Like some ghoul in a late-night horror movie that repeatedly sits up in its grave and shuffles abroad, after being repeatedly killed and buried, Lemon stalks our Establishment Clause jurisprudence once again . . . .” Lamb’s Chapel v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 398 (1993) (Scalia, J., concurring in the judgment).

Second is the endorsement test, which asks whether the government has sent a message of “endorsement or disapproval of religion”87×87. Lynch v. Donnelly, 465 U.S. 668, 688 (1984) (O’Connor, J., concurring). through its actions.88×88. Id. at 687–88. The endorsement test has been subject to the same critique of vagueness as Lemon.89×89. See, e.g., Michael Stokes Paulsen, Lemon Is Dead, 43 Case W. Res. L. Rev. 795, 815–16 (1993) (“The basic problem with the endorsement test is that it is no test at all, but merely a label for the judge’s largely subjective impressions.” Id. at 815.). The test came up primarily while Justice O’Connor, its author, was still on the bench, and its prominence has waned in the past two decades with the rise of the third test: coercion. Coercion asks whether the government has coerced someone into “support[ing] or participat[ing] in religion,”90×90. Lee v. Weisman, 505 U.S. 577, 587 (1992). either directly or indirectly. Coercion is a constitutional floor; regardless of what else the government may not do under the Establishment Clause, “at a minimum” the government may not coerce.91×91. Id. Direct coercion occurs when the government requires religious participation or imposes a sanction on nonpreferred religious behavior; indirect coercion involves implicit but still coercive state action.92×92. See Cty. of Allegheny v. ACLU, Greater Pittsburgh Chapter, 492 U.S. 573, 659–60 (1989) (Kennedy, J., concurring in the judgment in part and dissenting in part). To its critics, the coercion test is too narrow to capture the full range of harms the Establishment Clause should prevent, because it prohibits only government action that is actually coercive.93×93. See, e.g., Brief of Amici Curiae Religious Denominations and Other Religious Institutions Supporting Petitioners at 20, Am. Legion v. Am. Humanist Ass’n, 139 S. Ct. 2067 (2019) (Nos. 17-1717, 18-18), 2018 WL 6819434, at *20. Nevertheless, the coercion test has risen in prominence over the past two decades.

This is not to say the three tests operate independently of each other. Indeed, much of the confusion in Establishment Clause jurisprudence stems precisely from the fact that the frameworks are not siloed. The endorsement test, for instance, grew out of Justice O’Connor’s interpretation of the second prong of Lemon.94×94. Lynch v. Donnelly, 465 U.S. 668, 688–89 (1984) (O’Connor, J., concurring) (“Focusing on institutional entanglement and on endorsement or disapproval of religion clarifies the Lemon test as an analytical device.” Id. at 689.). Courts regularly lay out the three prongs of Lemon and then analyze the case through the lens of endorsement or coercion.95×95. In Lee v. Weisman, 505 U.S. 577, for example, the Supreme Court referenced Lemon and other principles behind the Religion Clauses before explaining it need not use those standards in deciding the case, because noncoercion was the constitutional minimum. Id. at 587. And the Supreme Court sometimes uses other guiding frameworks in deciding establishment cases: it has based its legislative prayer jurisprudence in large part on historical practice,96×96. See Town of Greece v. Galloway, 572 U.S. 565, 576–77 (2014); Marsh v. Chambers, 463 U.S. 783, 790–91 (1983). and when ruling on monuments with religious themes the Court has taken into account how long the monuments have stood.97×97. See Am. Legion, 139 S. Ct. at 2082–83; Van Orden v. Perry, 545 U.S. 677, 701 (2005) (Breyer, J., concurring in the judgment). The Court in Van Orden v. Perry, 545 U.S. 677, explicitly noted that Lemon is “not useful” when dealing with a “passive monument,” id. at 686 (plurality opinion).

B. Hints of Chilling in Establishment Clause Cases

Before explaining how chilling works with the various establishment tests, it is worth noting there have been a few cases involving both Religion Clauses in which chilling was connected to an Establishment Clause issue. These cases have generally involved suits against religious organizations. Justice Brennan raised the point in his concurrence in the judgment in Corporation of the Presiding Bishop of the Church of Jesus Christ of Latter-Day Saints v. Amos,98×98. 483 U.S. 327 (1987). which held that the exemption for religious organizations from Title VII’s prohibition on religious discrimination in employment could apply to those organizations’ secular nonprofit activities without violating the Establishment Clause.99×99. Id. at 329–30. The concern, he explained, was that “government intrusion” into a religious group’s affairs would “chill[]” the organization’s free exercise of religion.100×100. Id. at 343 (Brennan, J., concurring in the judgment). Religious organizations, fearful of litigation about whether their nonprofit activities counted as secular or religious for Title VII purposes, “would have an incentive to characterize as religious only those activities about which there likely would be no dispute, even if [they] genuinely believed that religious commitment was important in performing other tasks as well.”101×101. Id. For Justice Brennan, the evil of entanglement warranted a categorical exception for religious groups’ nonprofit activities.102×102. Id. at 345–46. Government entanglement with religion — a classic concern of the Establishment Clause103×103. See, e.g., Lemon v. Kurtzman, 403 U.S. 602, 613 (1971). — could not be allowed here because it would chill religious activity.104×104. Amos, 483 U.S. at 343 (Brennan, J., concurring in the judgment). Justice Brennan thus directly related an establishment issue to the chilling effect it would create.105×105. Some litigants have adopted this argument in other cases involving church operations. See, e.g., Brief Amicus Curiae of The Becket Fund for Religious Liberty in Support of the Defendant at 16, Bandstra v. Covenant Reformed Church, 913 N.W.2d 19 (Iowa 2018) (No. 16-1078) (arguing that “the process and prospect of courts making the religiosity determination would itself chill religious activity”).

The Eleventh Circuit sketched the same kind of relationship in Crowder v. Southern Baptist Convention,106×106. 828 F.2d 718 (11th Cir. 1987). in which the court held it could not hear a dispute about ecclesiastical governance.107×107. Id. at 727. Hearing the case would raise problems under both the Establishment and Free Exercise Clauses: by “adjudicating” a “religious dispute[]” — which the court warned could be an Establishment Clause problem — the court would have risked “affecting associational conduct and thereby chilling the free exercise of religious beliefs.”108×108. Id. at 721. Once again, establishment might chill religious exercise. In another case involving religious leaders and organizations’ unfulfilled promises of faith healing,109×109. Tilton v. Marshall, 925 S.W.2d 672, 675–76 (Tex. 1996). several justices of the Texas Supreme Court raised the same concern: that allowing the court to entangle itself in religious concerns and intrude into religious beliefs would have a chilling effect.110×110. Id. at 686 (Gonzalez, J., concurring and dissenting); id. at 693 (Hecht, J., concurring in part and dissenting in part). A California court had similar concerns in Rowe v. Superior Court,111×111. 19 Cal. Rptr. 2d 625, 638 (Cal. Ct. App. 1993). which held it was not an Establishment Clause violation for a state law to require plaintiffs to meet a clear and convincing evidence standard in order to plead punitive damages in suits against religious organizations.112×112. Id. at 627, 640; see also Cal. Civ. Proc. Code § 425.14 (West 2019). The court explained that “the fear of being called before a court to defend against unfounded claims for punitive damages” could create “chilling effects” for religious organizations in their religious activities and act as a “deterrent.”113×113. Rowe, 19 Cal. Rptr. 2d at 638. These cases involving both the Establishment and Free Exercise Clauses have recognized that certain government behavior related to religion can chill religious exercise, and that state entanglement in religious organizations — a core concern of the Establishment Clause — can chill the free exercise of religion. They demonstrate one way in which chilling can be used in an Establishment Clause analysis.

C. Using the Chilling Effect in Establishment Clause Tests

After exploring the theory behind and justifications for recognizing the chilling effect in Establishment Clause jurisprudence, reviewing current establishment doctrine, and noting how chilling has already been tied to concerns in the doctrine, the final step is to determine what the doctrine looks like when the chilling effect is included. Chilling — as a tool to help determine whether a government action is an unconstitutional establishment — fits neatly within the Lemon and endorsement tests, but requires reframing coercion. The relationship between the chilling effect and coercion highlights how establishment jurisprudence has failed to recognize the harm of omission of religion from the public square.

1. The Lemon Test. — Lemon may effectively be dead, but were it ever resuscitated, the question whether the government action at issue chills religious exercise could fall within the second prong: whether the action has the primary effect of advancing or inhibiting religion. A government action that chills religious expression, after all, would have the effect of inhibiting religion. The question for the court would then become whether that effect is “primary.”

2. The Endorsement Test. — Chilling is similarly useful for the endorsement test. If an observer is chilled, it may create a presumption that the government has endorsed or disapproved of a religion, because chilling stems from the information gleaned from government conduct. The endorsement case law also provides useful standards for determining when an observer is “reasonably” chilled. As Justice O’Connor explained in concurrence, a “reasonable observer” is “presumed to possess a certain level of information that all citizens might not share”114×114. Capitol Square Review & Advisory Bd. v. Pinette, 515 U.S. 753, 780 (1995) (O’Connor, J., concurring in part and concurring in the judgment). and “must be deemed aware of the history and context of the community and forum . . . [and] the general history of the place.”115×115. Id. at 780–81. The Supreme Court in McCreary County v. ACLU of Kentucky116×116. 545 U.S. 844 (2005). adopted that knowledgeable-observer standard, explaining that an observer is “presumed to be familiar with the history of the government’s actions and competent to learn what history has to show.”117×117. Id. at 866 (first citing Santa Fe Indep. Sch. Dist. v. Doe, 530 U.S. 290, 308 (2000); then citing Edwards v. Aguillard, 482 U.S. 578, 595 (1987); and then citing Pinette, 515 U.S. at 780 (O’Connor, J., concurring in part and concurring in the judgment)). An observer who may be chilled can be judged by the same standard.

Incorporating chilling into establishment jurisprudence also helps address a criticism often levied at endorsement: that applying the test faithfully would require doing away with things that seem as though they must be constitutional given their historical pedigree, like the motto “In God We Trust” on currency or the President’s Thanksgiving proclamation.118×118. See, e.g., Town of Greece v. Galloway, 572 U.S. 565, 579–80 (2014) (explaining how the dissenters in County of Allegheny v. ACLU, Greater Pittsburgh Chapter, 492 U.S. 573 (1989), criticized the endorsement test because “it likely would condemn a host of traditional practices that recognize the role religion plays in our society,” including legislative prayer and the Thanksgiving proclamation (citing id. at 670–71 (Kennedy, J., concurring in the judgment in part and dissenting in part))). Using chilling as one way of determining whether there is an endorsement likely leaves such practices be. In the case of the motto, for example, its long history makes it less likely that a reasonable observer would consider it to represent the current administration’s strong views. The motto is also diffuse and not attributable to the actions of any one particular actor or entity in the way that, for example, a town council’s decision to have legislative prayer is. Because it offers less of a reasonable basis to fear disapproval, and therefore less likelihood of chilling, the motto is more likely to pass muster.

3. The Coercion Test. — The trickier test is coercion, because chilling cannot fit comfortably within current coercion doctrine. To understand why, we must first understand how chilling and coercion — particularly indirect coercion — are distinct. Direct coercion is simple enough to distinguish from chilling. When the state establishes a mandatory tax to support a church, or compels attendance, it acts directly on people; religious exercise is the object of the regulation. Chilling is an incidental effect of a government action, and is therefore different from the intentional effect of direct coercion.

Indirect coercion, on the other hand, might seem at first glance to be the same as chilling. Indirect coercion operates through “coercive pressure,” including social pressure.119×119. Lee v. Weisman, 505 U.S. 577, 592 (1992); see also id. at 593–94; Ward, supra note 85, at 1632–33. The Supreme Court has been most concerned with indirect coercion in the context of public schools, because adolescents are more susceptible to peer pressure and less likely to be able to resist participating in religious expression if they do not want to.120×120. These concerns were illustrated in cases about prayers by an invited clergyperson offered at a high school graduation ceremony, Lee, 505 U.S. at 593–94, and student-led prayers offered at high school football games, Santa Fe Indep. Sch. Dist., 530 U.S. at 311–12. But outside of the public school context, there is a fairly high bar for finding something coercive. Town of Greece v. Galloway, for instance, held that a town board’s sectarian legislative prayer practice did not violate the Establishment Clause,121×121. 572 U.S. at 591–92. despite the fact that constituents felt “subtle pressure to participate in prayers that violate[d] their beliefs in order to please the board members from whom they [were] about to seek a favorable ruling.”122×122. Id. at 586 (opinion of Kennedy, J.). This mechanism might sound like chilling.

But there is a difference, and it is deceptively simple: one is coerced into doing something, while one is chilled into refraining from doing something. In other words, coercion is a positive act and chilling is a negative one. In the example in the Introduction of this Note, coercion would occur if the woman felt she should participate in the prayers before the council meeting; chilling would occur if she felt she should not express her own religion in proximity to the meeting, such as by wearing a headscarf. One might argue that chilling is simply indirectly coercing someone into not expressing their own religion, but that is not how the case law describes coercion. Instead, from the first introduction of the coercion test in County of Allegheny v. ACLU, Greater Pittsburgh Chapter,123×123. 492 U.S. 573. the cases have focused on being forced to participate. Justice Kennedy in Allegheny explained that government “may not coerce anyone to support or participate in any religion,”124×124. Id. at 659 (Kennedy, J., concurring in the judgment in part and dissenting in part) (emphasis added). and he reiterated that standard when writing for the Court in Lee v. Weisman.125×125. 505 U.S. 577; see id. at 587. Town of Greece addressed whether a legislative prayer practice was “compell[ing]” people “to engage in a religious observance.”126×126. Town of Greece, 572 U.S. at 587 (opinion of Kennedy, J.) (emphasis added); see also Lee, 505 U.S. at 642 (Scalia, J., dissenting) (“I see no warrant for expanding the concept of coercion beyond acts backed by threat of penalty . . . .”). The Court’s case law on coercion simply does not account for what happens when someone does not express their own religion, rather than when they express some other religion. The act-versus-omission distinction is the same in the free speech context, where compelled speech is analyzed differently from chilled speech; no one has spoken of someone being “indirectly coerced” into not speaking. The difference carries over into the establishment context.127×127. Just because chilling is different from indirect coercion does not mean the idea of indirect coercion cannot provide some support for the idea of chilling. Indirect coercion recognizes how subtle coercive pressures operate — and that those pressures may be social, not simply from a government official (as in Lee, where there was social pressure to attend graduation). Those same pressures can be at work in the case of chilling.

This act-versus-omission line might appear to be a distinction without a difference, but it gets at the heart of the matter. The point this Note has made is that Establishment Clause jurisprudence has ignored the omission issue in the first place, because it does not address how people may refrain from engaging in their own religious exercise in the face of the government’s affiliation with religion. The distinction between act and omission shows why chilling cannot fit within the coercion test as it currently stands. Chilling is a broader harm than coercion; just because government conduct does not indirectly coerce someone into participating in a religious exercise does not mean that person acts as she would have absent the government conduct. The coercion test as currently formulated cannot accommodate chilling, because it focuses only on forced, rather than omitted, participation. In other words, the test addresses only half of the ways in which establishment affects religious freedom, by coercing people into participating in religion when they do not wish to do so. It ignores how establishment affects religious freedom by causing people to refrain from participating in something they wish to. For chilling to be a tool in a coercion analysis, the coercion test would have to expand to include thinking of coercion in terms of causing someone to refrain from acting.

* * *

If chilling fits so neatly within the Lemon and endorsement tests, and the coercion test can be adjusted to accommodate chilling, why not simply use chilling as the test? Because there may be situations that do not involve chilling but still run afoul of the Establishment Clause. Providing funding to religious institutions is a prime example. The Framers were particularly concerned about state expenditures going to churches,128×128. See Feldman, supra note 70, at 351. but it is not clear that a generally available fund that provides aid to both religious and nonreligious groups would chill religious expression, because there would be no message of preference for a given religion or none at all. Thus, chilling does in fact do work — by showing the coercion test is too narrow and helping to clarify endorsement — but it also does not capture everything — because there are certain establishment issues that are based on other concerns. Chilling therefore should not stand on its own. But it must still be a tool, because without recognizing the harm to religious freedom that comes from people curtailing their own religious expression in the face of government religious expression, the Establishment Clause cannot be a full co-guarantor of religious liberty.


At the end of the road, it is worth addressing a counterargument that the road should not even exist: Some may be wary of incorporating chilling into establishment jurisprudence when there is no direct evidence that establishment chills religious exercise. In other words, doctrine should not be developed on the basis of an empirical claim without direct support. If that is the concern, then it applies not only to the argument of this Note, but also to a substantial number of other constitutional doctrines. The Supreme Court regularly deals in intuitions and commonsense assumptions about how people act in the world. The exclusionary rule, for instance, is based on the idea that the exclusion of illegally gathered evidence from the prosecution’s case will deter police misconduct;129×129. See, e.g., Mapp v. Ohio, 367 U.S. 643, 656 (1961). qualified immunity exists in part because the Court believes government officials may otherwise be hesitant to take necessary actions without protection from lawsuits, or that they may be burdened by litigation.130×130. See, e.g., Ashcroft v. Iqbal, 556 U.S. 662, 685 (2009); Harlow v. Fitzgerald, 457 U.S. 800, 814 (1982). These and other empirical claims do not always have support, either in the record or from independent academic investigation.131×131. See, e.g., Ryan Gabrielson, It’s a Fact: Supreme Court Errors Aren’t Hard to Find, ProPublica (Oct. 17, 2017, 8:00 AM), https://www.propublica.org/article/supreme-court-errors-are-not-hard-to-find [https://perma.cc/4DC2-AAES]. As Professor Leslie Kendrick has noted: “Dissents and majorities routinely criticize each other for the flimsy empirical basis of their assessments of chilling.” Kendrick, supra note 1, at 1656. Chilling in the speech context is itself hard to measure,132×132. See Canes-Wrone & Dorf, supra note 28, at 1097; Kendrick, supra note 1, at 1657 (arguing the chilling effect is based on “little more than a collection of unsubstantiated empirical judgments”). though there has been some empirical work showing the existence of a chilling effect on speech.133×133. See, e.g., Sidhu, supra note 24 (studying how Muslim Americans have changed their internet use after September 11); Alex Marthews & Catherine Tucker, Government Surveillance and Internet Search Behavior (Feb. 17, 2017) (unpublished manuscript) (on file with the Harvard Law School Library) (analyzing how individuals changed their internet searches after revelations that the government was monitoring internet searches). Given that chilling cases, among others, often involve “rough empirical judgment[s],”134×134. Kendrick, supra note 1, at 1658. any objection to chilling in the establishment context based on a lack of direct evidence should not wholly shut the door to the claim. Even a “basic intuition” about how the world works can have a place in constitutional law; not everything must be quantified.135×135. Id. at 1638; see also Schauer, supra note 13, at 730.

Incorporating the chilling effect into Establishment Clause jurisprudence recognizes the many ways in which the government can affect religious expression in the public sphere: even if the government does not coerce religious participation, people do not always act as they would absent the government’s conduct. The omission of religious expression is not captured by current establishment doctrine. In order to honor the promises of the Religion Clauses, chilling must be recognized as an Establishment Clause concern.