The Establishment Clause of the First Amendment to the Constitution provides that ‘‘Congress shall make no law respecting an establishment of religion . . . .’’ The aim of the Establishment Clause is to keep the government neutral in matters of religion (Epperson v. Arkansas , McCreary County v. ACLU ). The Supreme Court has used a number of different approaches to determine whether a government action has departed from the neutrality principle and unconstitutionally established religion.
The Court has most often used the Lemon test, a three-prong test announced in Lemon v. Kurtzman (1971). More recently, however, Justice Sandra Day O’Connor has moved the Court towards more frequent use of the ‘‘endorsement test.’’ See, for instance, Lynch v. Donnelly (1983) (O’Connor, J., concurring), County of Allegheny v. ACLU (1989) (O’Connor, J., concurring), and Santa Fe Ind. Sch. Dist. v. Doe (2000). Still another test that has gained some currency on the Court is the ‘‘coercion test’’ articulated by Justice Anthony Kennedy. See Allegheny (Kennedy, J., dissenting in part). In light of the Court’s turbulent approach in this area, as well as changes in the Court’s membership, it is reasonable to characterize the Establishment Clause as an area in transition.
When applying the Lemon test, courts will find no establishment of religion if the government action (1) has a secular legislative purpose, (2) does not have the effect of advancing or prohibiting religion, and (3) does ‘‘not foster ‘an excessive entanglement with religion’’’ (Lemon [internal citations omitted]).
First, the challenged law or action must have a legitimate secular purpose. In Stone v. Graham (1980) (per curiam), for example, the Supreme Court struck down a state law requiring that the Ten Commandments be hung on the walls of Kentucky classrooms. The Court looked beyond the secular purpose proffered by the legislature and found that the ‘‘preeminent purpose for posting the Ten Commandments on schoolroom walls is plainly religious in nature.’’ Similarly, in Wallace v. Jaffree (1985), the Court found an unconstitutional purpose behind a law whose sponsor had confirmed that it was ‘‘an ‘effort to return voluntary prayer’ to the public schools.’’ More recently, in McCreary County v. ACLU (2005), the Court found that a Ten Commandments display in a courthouse violated the Establishment Clause. Critical to the Court’s decision was that the initial display included only the Ten Commandments and a subsequent display, erected after the first was challenged, included religious excerpts from historical documents. The Court found the purpose behind the display was to ‘‘favor religion.’’ As the Court noted in McCreary, ‘‘the principle of neutrality has provided a good sense of direction: the government may not favor one religion over another, or religion over irreligion.’’
The second prong of the Lemon test looks at whether the effect of the challenged government action is the establishment of religion. In Lynch, the Court upheld a cre`che surrounded by other secular Christmas symbols, including a Santa Claus house, reindeer, and a Christmas tree, because the display’s effect was not to establish religion over nonreligion or endorse one particular sect over another. Similarly, in Allegheny, the Court approved a holiday display that included a menorah, a Christmas tree, and a plaque explaining a secular connection. The inclusion of the secular elements with the religious ones meant that the display did not have the effect of establishing religion. However, the Court held unconstitutional a separate display, consisting solely of a cre`che sitting on another piece of government property. By displaying the cre`che on its own, without the surrounding secular items, the county sent ‘‘an unmistakable message that it supports and promotes the Christian praise to God that is the cre`che’s religious message.’’ Similarly, student-led prayer before high school football games had the unconstitutional effect of endorsing religion in Santa Fe. There, the Court found that forcing students to make religion a factor in deciding whether to attend the games had the ‘‘effect of coercing those present to participate in an act of religious worship.’’
The entanglement prong of the Lemon test is focused primarily on government spending that supports religion and the resulting government intrusion into religious organizations. In Lemon itself, the policies of two states to spend tax dollars to support religious schools were found unconstitutional because of excessive entanglement. To determine if the entanglement between government and religion is excessive, the Court looks at ‘‘the character and purposes of the institutions that are benefited, the nature of the aid that the State provides, and the resulting relationship between the government and the religious authority.’’ Because of the intrusive government oversight necessary to administer financial aid, the Court in Lemon found excessive entanglement in the states’ laws. A similar monetary and oversight analysis for the entanglement prong was applied to a Texas sales tax exemption for religious publications in Texas Monthly v. Bullock (1989). The statute required that applicants for the exemption prove to government officials that their ‘‘message or activity is consistent with ‘the teaching of the faith.’’’ This entanglement could also lead to the perception that government has approved some religions and disapproved others (Bullock quoting United States v. Lee  [Stevens, J., concurring]). Where government action, especially concerning money, could result in intrusive oversight of religion, the Court has found excessive entanglement and an Establishment Clause violation.
The endorsement test grew out of the effects and purpose prongs of the Lemon test. In Lynch, Justice O’Connor suggested that the Court should focus its attention not on generalized inquiries into the purpose and effect of the challenged government action, but instead should ask whether the action was taken with the purpose, or had the effect, of endorsing religion (Lynch [O’Connor, J., concurring]). She also suggested analyzing government action from the standpoint of the ‘‘reasonable observer,’’ asking whether that observer would believe her standing in the political community was based on adherence or nonadherence to that message. In Lynch itself, the surrounding secular symbols neutralized the religious nature of the cre`che, ensuring that a reasonable observer would not feel the government was endorsing the cre`che’s religious message. However, in Allegheny, there were no secular symbols surrounding the cre`che, which to Justice O’Connor conveyed ‘‘a message to nonadherents of Christianity that they are not full members of the political community, and a corresponding message to Christians that they are favored members of the political community.’’ This religious symbol in a government building had the ‘‘the unconstitutional effect of conveying a government endorsement of Christianity.’’
Justice O’Connor’s theory of endorsement animates the majority opinions in Wallace, Allegheny, and Santa Fe. In Wallace, the majority framed the analysis of the purpose prong as ‘‘whether the government intends to convey a message of endorsement or disapproval of religion.’’ In Allegheny, the majority noted the recent trend in Court decisions of paying ‘‘close attention to whether the challenged governmental practice either has the purpose or effect of ‘endorsing’ religion.’’ Finally, in Santa Fe, the Court deemed school-sponsored religious speech ‘‘impermissible because it sends the ancillary message to members of the audience who are nonadherants ‘that they are outsiders, not full members of the political community, and an accompanying message to adherants that they are insiders, favored members of the political community’’’ (Santa Fe citing Lynch). While the Court still applies the Lemon framework in these cases, its analysis of the effects and purpose prong has been transformed by Justice O’Connor’s endorsement test. In most cases, the previously generalized inquiry into purpose and effect has been narrowed into a search for one particular purpose (to endorse) and one specific effect (endorsement).
Justice Kennedy has proposed a coercion test for Establishment Clause cases, and Justice Thomas has recently incorporated this idea into his own opinions (Allegheny [Kennedy, J., dissenting], Van Orden v. Perry  [Thomas, J., concurring]). Justice Kennedy suggested in Allegheny that actions should be invalidated only when they ‘‘further the interests of religion through the coercive power of government.’’ He suggested that coercion, rather than ‘‘infringement on religious liberty by passive or symbolic accommodation’’ is the evil that the Establishment Clause is designed to guard against. Justice Kennedy incorporated the coercion analysis in his opinion for the majority in Lee v. Weisman (1992). There, a public school graduation prayer led by a rabbi was found to violate the Establishment Clause. Kennedy held that ‘‘at a minimum, the Constitution guarantees that government may not coerce anyone to support or participate in religion or its exercise, or otherwise act in a way which ‘establishes a [state] religion or religious faith, or tends to do so’’’ (internal citations omitted). Because children were coerced to either participate in or protest a religious exercise, the school violated the Establishment Clause in Lee.
Justice Thomas has recently begun to advocate use of the coercion test as well. He argued in Van Orden that the original constitutional meaning of establishment was coercion. ‘‘The Framers understood an establishment ‘necessarily [to] involve actual legal coercion’’’ (Van Orden [Thomas, J., concurring] quoting Elk Grove Unified Sch. Dist. v. Newdow  [Thomas, J., concurring in the judgment]). Justice Thomas had argued in Newdow that ‘‘‘government practices that have nothing to do with creating or maintaining . . . coercive state establishments’ simply do not ‘implicate the possible liberty interest of being free from coercive state establishments’’’ (Van Orden quoting Newdow [Thomas, J., concurring]). While Justice Kennedy was the first member of the contemporary Court to propose use of the coercion test, he has not suggested that it be adopted to represent the full scope of the Establishment Clause’s reach. As his opinion in Lee intimated, he sees refraining from coercion as the ‘‘minimum’’ that is required of government. Justice Thomas, on the other hand, has opined that only coercive governmental measures can violate the Establishment Clause.
In light of these crosscurrents in the Court’s Establishment Clause jurisprudence, it is difficult to predict with any confidence which approach (if any) will come to dominate the Court’s analysis in the future. Because of her centrality in the development and application of the endorsement test, Justice O’Connor’s departure from the Court heightens this uncertainty. There are few, if any, signs of the emergence on the Court of a consistent, clear majority in favor of a single approach to deciding Establishment Clause cases.
SAMUEL A. MARCOSSON
References and Further Reading
Cases and Statutes Cited