Concept
The Lemon Framework
Establishment
The Lemon Framework
From 1971 until 2022, the Court's test for an Establishment Clause violation came from Lemon v. Kurtzman. The Court did not always apply it, especially in later years, but a large body of cases grew up under it. In Kennedy v. Bremerton School District (2022), the Court said that Lemon had been overruled, and it is unclear what that means for the many decisions rendered under the test.1 Because that body of law is large and its status uncertain, this unit presents the framework and the decisions built on it; later units take up the endorsement and coercion approaches.
Under the test as the Court described it, a governmental practice must (1) reflect a clearly secular purpose; (2) have a primary effect that neither advances nor inhibits religion; and (3) avoid excessive government entanglement with religion.2 Each criterion puts a different question to the same enactment. The first asks why the legislature acted. The second asks what the law does once it operates. The third asks what continuing relationship the law creates between public officials and a religious body. The first two give practical form to a principle stated in Everson v. Board of Education: neither a state nor the federal government may “pass laws which aid one religion, aid all religions, or prefer one religion over another.”3
The criteria are cumulative: a statute that violates any one of them must be struck down.4 Each prong closes a gap the other two leave open. A test confined to motive would pass a program whose sponsors were sincere but whose operation funneled public money into religious instruction. A test confined to results would excuse a legislature that set out to promote a faith and merely failed. Neither would catch a program that is secular in aim and in its terms yet requires officials to supervise a religious institution indefinitely.
Lemon itself shows the prongs working separately. The Court held both the Pennsylvania and Rhode Island statutes unconstitutional because the cumulative effect of the entire relationship arising under them involved excessive entanglement between government and religion.5 It did not rule exactly on whether the programs had a primary effect of advancing religion, although its discussion of the need to avoid administrative entanglement indicated that such programs could have a prohibited effect.6
The framework has always had critics. Justice Powell, concurring in Wallace v. Jaffree, defended it as the only coherent test a majority of the Court had ever adopted, noted that the Court had addressed an Establishment Clause issue without it only once since Lemon (Marsh v. Chambers), and warned that continued criticism could encourage courts to decide cases ad hoc.7 Part of the dispute concerns what the Clause forbids. On the separationist view, religion and government exist in profound tension, and separation protects the church from worldly corruption and secular interests from ecclesiastical influence. On the non-preferentialist view, government may aid religion so long as it prefers no one religion over the others.8 One statement of that view adds that government may promote religion generally, even to the point of discriminating against nonreligion.9 Lemon's second prong, which condemns a law whose primary effect advances religion, sits with the separationist premise, and the section on critiques returns to the dispute.
Secular Purpose
The first criterion asks why the legislature acted: the law must reflect a clearly secular purpose.11 Two features shape the criterion. It demands a secular purpose, not an exclusively secular one, and it defers to a legislature's stated purpose. Each feature has a limit, and the limits decide the contested cases.
A law with a mixed motive can pass. A statute motivated in part by a religious purpose may still satisfy the first criterion; it must be invalidated only when it is entirely motivated by a purpose to advance religion.12 Justice Powell gave the reason in Wallace: a rule demanding exclusively secular objectives would have invalidated much that the Court had already approved, including New York's property tax exemption for religious organizations and a township's reimbursement of parents for the cost of busing children to parochial schools.13 The reason is practical as well as historical. A statute that ostensibly promotes a secular interest often has an incidental or even a primary effect of helping or hindering a sectarian belief, and chaos would ensue if every such statute were invalid; a state could not criminalize murder for fear of promoting the Biblical command against killing.14
Laws with religious roots or religious overlap therefore pass this criterion routinely. A law does not violate the Clause merely because it “happens to coincide or harmonize” with a religious tenet; the Court so held unanimously in Harris v. McRae, rejecting the claim that the Hyde Amendment's limits on Medicaid funding of abortion wrote Roman Catholic doctrine into law.15 The Sunday closing cases, of which McGowan v. Maryland is the standard example, illustrate the same point for a law with religious origins and a present secular aim. A further reason lies in the political process: if religious support behind a law counted against its constitutionality, religious citizens might be chilled from voicing their views in public debate.16
Relieving religion of a government-imposed burden also counts as a secular purpose. In Corporation of the Presiding Bishop v. Amos, the Court upheld the exemption of religious organizations from Title VII's ban on religious discrimination in employment.17 Justice White reasoned that “it is a permissible legislative purpose to alleviate significant governmental interference with the ability of religious organizations to define and carry out their religious missions,”18 and that the purpose prong aims at a decisionmaker who abandons neutrality in order to promote a particular point of view in religious matters, so a law need not be unrelated to religion to pass.19
The tolerance for mixed motives ends where religion is the whole point of the law. If the sole or predominant purpose of a law is religious, the Court will strike it down.20 The formulations vary. Some decisions ask whether the law is “entirely motivated by a purpose to advance religion”; others require a secular purpose that is “genuine, not a sham, and not merely secondary to a religious objective.” Even the stricter standard would let legislators also intend to help religion or advance religious values.21 The decisions finding no secular purpose cluster in public schools and in explicitly religious exercises, symbols, or teachings. When government sets up a program particularly to promote religion, no broad secular purpose may remain to which the advancement of religion is incidental, and most of these decisions can also be explained on the ground that the laws explicitly promoted the government's preferred form of religion.22
The second feature is deference, and it has a limit of its own. Courts normally defer to a state's articulated purpose, but the avowed purpose must be sincere and not a sham.23 A legislature may enunciate a sham secular purpose, courts can tell a sham from a sincere one, and the inquiry into effect helps decide the close cases where an expressed secular purpose is in doubt. The secular-purpose requirement alone may rarely be determinative, but it serves an important function.24 McCreary County sets the working standard: purpose is not to be trivialized by “any transparent claim to secularity”; the court considers all probative evidence in a spirit of common sense; and a county's effort to create a secular purpose after a preliminary injunction could not, standing alone, erase the context of its religiously motivated efforts to post the Ten Commandments in courthouses.25
The sham limit decided Stone v. Graham, which the next section takes up. The trial court there found the avowed purpose of the Kentucky statute secular even as it called the statutory declaration “self-serving,” and the Supreme Court held that an avowed secular purpose is not sufficient to avoid conflict with the First Amendment.26 Edwards v. Aguillard reached the same result: the Court deferred in principle to the state's stated purposes and still concluded that Louisiana's preeminent purpose was to advance the religious view that a supernatural being created mankind.27 Epperson v. Arkansas struck down a ban on teaching human evolution because the ban rested on a perceived conflict with a particular religious group's reading of Genesis, a failure of what is now the first prong.28
Evidence of purpose is not confined to a recital. In Wallace, the Court compared the challenged Alabama provision with the existing statute and found that the only significant textual difference was the addition of the words “or voluntary prayer.”29 Justice Powell added that a single legislator's statement, particularly one made after enactment, is not necessarily sufficient to establish purpose, and he pointed to the sequence and history of the three Alabama statutes as other evidence of religious purpose.30
A failed purpose inquiry ends the case. When a statute does not have a clearly secular purpose, no consideration of the second or third criteria is necessary.31 The criterion therefore works as a gate: failing it is fatal, while passing it only moves the analysis on to effect.
Primary Effect
The second criterion asks what a law does once it operates: its primary effect must neither advance nor inhibit religion.33 A secular purpose can be recited in good faith and still accompany a program that funnels public money to religious instruction, so the inquiry into consequences covers a gap that purpose leaves. The criterion runs in both directions, but the cases below concern advancement.
No formula identifies when an effect is “primary.” The inquiry is subjective, a convenient surrogate for a more general judgment that a law goes too far in advancing nonsecular ends. A law lending mathematics textbooks to students at parochial schools shows the difficulty: a strict separationist could say the primary effect is to advance a sectarian institution by lowering its operating costs, while a less strict reader could say it advances the state's interest in a sound education for its young citizens.34
The cases sort aid by the route the money takes. In Everson, the Court allowed a township to reimburse parents for the cost of transporting their children to parochial schools.35 The Pennsylvania program in Lemon had a further defect: it provided state financial aid directly to the church-related school. That feature distinguished Everson and Board of Education v. Allen, because in both the Court was careful to point out that aid went to the student and the student's parents, not to the school.36
The route matters because it changes who has decided to support religion. When a program pays the school, government itself has chosen to fund the institution. When it pays individuals, the link between government funds and religious training is broken by the independent and private choice of the recipients; the Court has cited Zelman v. Simmons-Harris for that point.37 The section on aid to religious schools develops the private-choice programs. What the effect prong does with a program that is not neutral is clearest in Committee for Public Education & Religious Liberty v. Nyquist. New York gave a package of benefits exclusively to private schools and the parents of private-school enrollees. The program was enacted for ostensibly secular purposes, yet the Court found that its function was “unmistakably to provide desired financial support for nonpublic, sectarian institutions,” and it traced the program's genesis to the grave fiscal problems private religious schools faced.38 Nyquist thus shows the first two prongs working independently: a program can clear the purpose gate and still fail on what it does. The Court has since held that Nyquist does not govern neutral educational assistance programs that offer aid directly to a broad class of individual recipients defined without regard to religion.39
The Court has also reworked what counts as advancing religion. Under Lemon, aid “advanced” religion if it was, or even might be, used for religious purposes. Under Mueller v. Allen and Witters v. Washington Department of Services for the Blind, government advances religion only if it gives more favorable aid for religious uses or otherwise creates an incentive to choose religious over comparable nonreligious institutions. “Neutrality” means no greater aid to religion than to its alternatives, not no aid to religion, so the aid need not be restricted to nonreligious uses, and no entangling surveillance is needed to police such a restriction.40 A concurring opinion in Zelman reduced the modern inquiry to two questions: whether the program distributes aid neutrally, without differentiation based on the religious status of beneficiaries or providers, and, more importantly, whether beneficiaries have a genuine choice among religious and nonreligious organizations. A “no” to either condemns the program. The same opinion notes that Agostini v. Felton folded the entanglement inquiry into the primary-effect inquiry, because both rely on the same evidence.41
The character of the institution aided also bears on effect. In Tilton v. Richardson, the Court reasoned that college students are less impressionable and less susceptible to religious indoctrination, that college courses limit sectarian influence through their own internal disciplines, and that, because religious indoctrination is not a substantial purpose of these church-related colleges, religion is less likely than in primary and secondary schools to permeate secular education. That reduces the risk that government aid will in fact support religious activities.42
Under every version of the prong, the effect must be attributable to government. A law is not unconstitutional simply because it allows churches to advance religion; to have forbidden effects, it must be fair to say that government itself has advanced religion through its own activities and influence.43 That rule saved the Title VII exemption in Amos, which relieved religion of a burden government had imposed rather than conferring a benefit. Estate of Thornton v. Caldor, Inc. shows the opposite pole. The statute there gave Sabbath observers an “absolute and unqualified right” not to work on their Sabbath, commanded that Sabbath concerns automatically control over secular interests, and made no exception for substantial burdens on employers or coworkers. The Court held it an impermissible establishment because it clearly advanced “a particular religious practice.”44 The accommodation section returns to the line between Amos and Thornton.
Justice O'Connor's concurrence in Lynch v. Donnelly restated the two prongs as questions about endorsement. The purpose prong asks whether government's actual purpose is to endorse or disapprove of religion; the effect prong asks whether, irrespective of that purpose, the practice in fact conveys a message of endorsement or disapproval; and an affirmative answer to either question should render the practice invalid.45 The endorsement section builds on that recasting.
Excessive Entanglement
The third criterion forbids a law that fosters excessive government entanglement with religion.47 Lemon was decided on this ground, and the Court there looked at the cumulative effect of the entire relationship the statutes created.48 The weight rests on the word excessive. Interaction between church and state is inevitable, and the Court has always tolerated some level of involvement between the two; entanglement must be excessive before it offends the Clause.49
Aid to church-related schools exposed a bind. If officials do not monitor how the aid is used, public money may flow to religious activity; if they do monitor, they move into a religious institution's affairs. The Court later called this a “Catch 22”: the very supervision of the aid to assure that it does not further religion renders the statute invalid.50 The Lemon Court's discussion of administrative entanglement also suggested that such programs could have a prohibited effect, which is one reason the two inquiries later merged.51
The size of the bind depends on how far religion pervades the institution aided. At church-related colleges, courses limit sectarian influence and religion is less likely to permeate secular education, so the necessity for intensive government surveillance is diminished and the resulting entanglements are lessened.52 The inspection needed to ascertain that the facilities are devoted to secular education is minimal.53 By implication, the more religion saturates a school's daily operation, the more surveillance aid to it demands.
Aid that needs no continuing policing sits at the opposite pole. The property tax exemption upheld in Walz minimized entanglement: because religious property was exempt, government had no need to assess it, impose tax liens, foreclose for nonpayment, or dispute with religious officials over the appropriate measure of taxation.54 Routine regulatory contact is not excessive entanglement, as the tolerance for inevitable church-state interaction implies.55 Hernandez v. Commissioner, which sustained an IRS ruling denying a charitable deduction for payments for Scientology auditing sessions, is a familiar example.
Lemon also treated political division as a form of entanglement. In Walz, the state tax status benefited all religious groups. The aid programs in Lemon required successive and very likely permanent annual appropriations that benefited relatively few religious groups, so political fragmentation and divisiveness on religious lines were likely to be intensified.56 The Court added that there was no history of state aid to church-related schools comparable to two hundred years of tax exemption for churches; that modern programs are self-perpetuating and self-expanding; and that a step thought to approach “the verge” can become the platform for the next, a “downhill thrust” that is easy to start and hard to stop.57 This branch of the doctrine has since lost its force. Critics said that political divisiveness discriminates against religious viewpoints and denies religion's historic influence on American politics, and the Court eventually abandoned political entanglement as a reason to strike down aid.58
A third kind of entanglement concerns who holds governmental power. In Larkin v. Grendel's Den, the Court struck down a law under which liquor-license applicants near a church needed the church's approval. The statute “enmesh[ed] churches in the exercise of substantial governmental powers,” and the mere appearance of a joint exercise of legislative authority by church and state gave religion a significant symbolic benefit.59 The law may have had a secular purpose, keeping the area around cultural and educational centers quiet, but that purpose could not establish its constitutionality. Each church exercised a governmental power subject to no clear secular standard, and the power could be used to promote primarily religious goals. Even assuming a secular purpose and a secular primary effect, the law failed the entanglement test.60 Justice Rehnquist's dissent asked why the state, which could have banned liquor outlets near churches outright, could not leave the choice to the church. The Court answered with the “symbolic union” of giving a church a discretionary governmental decision, and with the risk that the church could use that discretion for explicitly religious goals, such as favoring applicants of its own faith.61 Authorizing clergy, alongside judges and ship's captains, to solemnize marriages may be permissible because no one need go to a clergyperson to be married, unlike the applicants in Larkin.62 The delegation concern reaches beyond churches themselves: Board of Education of Kiryas Joel Village School District v. Grumet applied it to a school district drawn around a single religious community, and the section on delegation and sect preference takes that case up.
Entanglement worries run in both directions. The Religion Clauses exist to protect religion from government interference as well as to protect dissenters, and Madison argued that ecclesiastical establishments, instead of maintaining the purity and efficacy of religion, “have had a contrary operation.”63 On the separationist view, separation prevents worldly corruptions from harming the church.64 The prayer cases reflect a related concern about courts judging religious content: the Court has refused to require nonsectarian prayers, because policing that rule would force courts to evaluate the content of religious speech and would require ministers to set aside nuanced and deeply personal beliefs for vague and artificial ones.65
After Agostini, entanglement is no longer a freestanding prong. The Court said it is simplest to recognize why entanglement is significant and treat it as an aspect of the inquiry into a statute's effect, while reiterating that entanglement must be excessive before it offends the Clause.66 Both inquiries rely on the same evidence, and the degree of entanglement bears on whether a statute advances or inhibits religion.67