Volume V, Number 1, Spring 2009

"In Search of a Legal Definition of Religion: Lessons from U.S. Federal Jurisprudence" by László Blutman

Dr. László Blutman has an LL.M degree from New York University and a PhD degree from the University of Szeged. He is Professor of Law at the Department of International and European Law, University of Szeged. E-mail:


“Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof.” When the Framers formulated this part of the First Amendment of the United States Constitution, they perhaps did not think that one of these words would later cause considerable trouble. This word is “religion”. Though the Supreme Court recognized in the Reynolds case that the “word ‘religion’ is not defined in the Constitution” and pointed out that “[w]e must go elsewhere, therefore, to ascertain its meaning…” (Reynolds v. United States, 98 U.S. 162 (1878)), it did not provide a proper legal definition for this concept – with the single possible exception of its ruling in the case of Davis v. Beason (133 U.S. 333 (1890)).

A clear legal definition of religion has very much been needed. The rule cited from the Constitution provides a two-tiered protection for religions, expressing complementary values. The first Clause of the constitutional provision, called the Establishment Clause, requires a separation of church and state. It protects a religion by forbidding state favoritism of other religions. Generally, the State may not advance beliefs, teachings, tenets, and accompanying practices or organizations if they are of a religious nature. However, this is not always clear-cut in a specific context, e.g., did the City of San Jose violate the Establishment Clause by financing a public sculpture representing Quetzalcoatl (the Plumed Serpent) of Aztec mythology to commemorate the town’s Mexican and Spanish cultural heritage? (See Alvarado v. City of San Jose, 94 F3d 1223 (9th Cir. 1996).)

The second Clause, the Free Exercise Clause, commands state noninterference with religious beliefs and accompanying practices. Activities or practices resting on a belief system or a belief are entitled to constitutional protection under this Clause only if this belief system or belief qualifies as religious. But, for example, could polygamy be considered a religious tenet or duty (and if so, would it be protected by this Clause)? (See the so-called Mormon cases of the second half of the 19th century, e.g., Reynolds, supra and Davis, supra.) Might a prisoner alleging that he was a minister in the MOVE organization be entitled to a religious diet in prison (Africa v. Commonwealth of Pennsylvania 662 F. 2d 1025 (3d Cir. 1981) cert. denied 456 U.S. 908 (1982))? On what criteria can a court decide these questions if no definition of religion is available?

In recent decades, the problem has become more acute with the growth in the number of nontraditional religions, religious sects, and belief systems in American society. Robbins and Anthony gave an excellent survey on this trend (Robbins and Anthony 475-484). The spread of heterogenous but possibly religious phenomena has multiplied the problems flowing from the lack of clear definitional guidance on the conceptual boundaries of religion. Can a court treat neo-paganist Wicca (witchcraft) practices as religion? (See Dettmer v. Landon, 799 F.2d 929 (4th Cir. 1986), cert. denied, 483 U.S. 1007 (1987); Encyclopaedia 15. 421-422.) And what about secular humanism? (Torcaso v. Watkins, 367 U.S. 488, 495 (1961); Kalka v. Hawk, 215 F.3d 90 (D.C. Cir. 2000)) Is the Washington Ethical Society entitled to the tax exemption reserved for religious organizations? Is Ethical Culture a kind of religion? (See Washington Ethical Society v. District of Columbia, 249 F.2d 127 (D.C. Cir. 1957); Torcaso v. Watkins, 367 U.S. 488, 495 (1961); Encyclopaedia 5. 171-173.) What about Satanism and Ásatrú/Odinism? (See Cutter v. Wilkinson 544 U.S. 709 (2005); Smith v. Allen, 502 F.3d. 1255 (11th Cir. 2007).) And the complex New Age movement? (See Alvarado v. City of San Jose, 94 F.3d 1223 (9th Cir. 1996).)

Complex interests may depend on the classification of a specific belief system or practice: tax exemptions; religious practices in prison or in the military (e.g., assembly for worship services; possession and sacramental use of various religious physical objects; access to religious literature; wearing of religious garments and jewelry; availability of food required by religious tenets); specific rights of workers, etc. The application of some constitutional and federal legal rules compels courts to delineate the boundaries of the concept of religion.

Legal theorists have made serious attempts to provide an adequate definition of what religion is for First Amendment purposes, and the Supreme Court’s and other federal courts’ efforts have been manifested in a string of cases in the context of the First Amendment as well as in statutory interpretation. These efforts should not be seen as entirely fruitless, but they have not provided a generally accepted legal definition of religion. (For a detailed, general review, see Donovan 35-61.)

The aim of this paper is to identify the causes of this failure. (For the problems relating to the definition of religion in Hungarian constitutional law, see Schanda 1996, 253-255; Schanda 2003, 116-119 and 249; Chronowski et al. 373-374; and Hungarian Constitutional Court Decision Nos. 4/1993 (12 Feb. 1993), point III, and 8/1993 (27 Feb. 1993), point II.) The first two parts of this paper are expository in nature, summarizing the judicial and theoretical efforts to define religion in U.S. constitutional and federal law. This may provide an appropriate basis, in the third part, both for outlining the underlying and dominant contradictions of this definitional problem and for presenting some suggestions to overcome these contradictions.

At the very outset, one of the most important, preliminary questions is whether we are defining the word “religion”, the concept of religion, or the social phenomenon called “religion”? I will leave this question unanswered and thus steer clear of the philosophical (ontological) problems it raises. It will be the concept of religion that I consider to be the object of definition for the purposes of this paper because this position seems to cause somewhat fewer theoretical difficulties than the other two.

1. The Supreme Court’s positions

1.1. The theistic approach and its erosion

The Supreme Court’s first effort to describe the content of the concept of religion took place only a century after the enactment of the First Amendment. In 1890, in a case involving the Mormon Church, the Court confronted the problem of defining religion. The Davis decision unhesitatingly employed a substantive definition and at the same time paid tribute to theism: “The term ‘religion’ has reference to one’s views of his relations to his Creator, and to the obligations they impose of reverence for his being and character and of obedience to his will” (Davis v. Beason, supra, 342). These well-known words require that religion involve belief in a deity and the worship thereof.

This holding was affirmed two years later in Church of the Holy Trinity v. Unites States, where the Court said that “we are religious people” and that “this is a Christian nation” (143 U.S. 471 (1892)). Konvitz pointed out that the Court practically identified religion with Christianity (Konvitz 147).

This position characterized judicial practice in the subsequent fifty years. In the Macintosh case, Justice Sutherland writing for the Court put emphasis on Christianity (United States v. Macintosh, 283 U.S. 622 (1931)), and Chief Justice Hughes concluded that “the essence of religion is belief in a relation to God involving duties superior to those arising from any other human relation” (Macintosh, supra, 633-634). The Supreme Court’s approach, of course, greatly influenced the practice of lower courts, which also echoed the theistic viewpoint. (See, e.g., Gabriell v. Knickerbocker, 12 Cal.2d 85, 90, 82 P.2d 391, 393 (1938), appeal dismissed, 306 U.S. 621 (1939); Nicholls v. Mayor of Lynn, 297 Mass. 65, 70, 7 N.E.2d 577, 579-80 (1937).)

In the meantime, however, important changes took place in the religious life of American society. An increasing number of belief systems appeared on the scene which could not be classified and evaluated by traditional religious concepts and principles. This social change led to new theories in theology and sociology (see Harvard Note 1978, 1068). The diversification of Christianity, and then of American religious life as a whole, both in practice and in theory, was a serious challenge to the traditional theistic notion. In the legal sphere, therefore, deviation from the theistic understanding as presented by the Davis court was imminent. The theistic, traditional concept was first explicitly challenged not by the Supreme Court, but by the Second Circuit (U.S. Court of Appeals for the Second Circuit).

In the Kauten case (133 F.2d 703 (2d Cir. 1943)), Judge Hand gave a broader definition, or rather description, of the concept of religion. Interpreting the expression “religious training and belief” in the Selective Service Act of 1940, he stated: “Religious belief arises from a sense of the inadequacy of reason as a means of relating to the individual to his fellowmen and to his universe…. It is a belief finding expression in a conscience which categorically requires the believer to disregard elementary self-interest and to accept martyrdom in preference to transgressing its tenets” (Kauten, supra 708). He concluded that conscientious objection could be regarded as “a response of the individual to an inward mentor, call it conscience or God, that is for many persons at the present time the equivalent of what has always been thought a religious impulse” (Kauten, supra 708).

These words may be considered a very important step in the history of the definition of religion in the judicial practice of the United States. This approach marked a sort of breakthrough in two aspects.

The Kauten ruling saw religion as relating man not to God, or not only to God, but to the universe and to other men. This new aspect in itself is very important compared to the thinking of previous decisions. However, the second aspect can be regarded, in my view, at least as important as this one.

By this time, the courts had examined the substantive and objective, overall attributes of religious belief systems and practices in various cases. However, in the Kauten case another approach took precedence: the role of various beliefs, including religious beliefs, in the life of the individual (Harvard Note 1978, 1061). This psychological approach taken in Kauten then paved the way for subsequent decisions in the cases of United States v. Seeger (380 U.S. 164 (1965)) and Welsh (398 U.S. 333 (1970)). Since Kauten, the old definition that requires a belief in God has gradually eroded. But new ideas rarely gain immediate general acceptance. For example, the Ninth Circuit, interpreting the same statute in Berman v. United States (156 F.2d 377 (9th Cir.) cert. denied 329 U.S. 795 (1946)), found that the conscientious objector exemption called for belief in a deity (Berman, supra, 381), and the court referred to Chief Justice Hughes’ famous statement expressed in Macintosh (see Berman, supra, note 7).

1.2. Heading for a functionalist direction

At least the spirit of the Kauten decision, if not the conclusion, was slowly and implicitly accepted by the Supreme Court. In the Barnette case, Justice Frankfurter approved the new approach in his dissent (West Virginia State Board of Education v. Barnette, 319 U.S. 658-659 (1943)). The Court, in a subtle way, qualified its original orthodox standpoint in United States v. Ballard (322 U.S. 78 (1944)). Although Justice Douglas, who wrote for the majority, did not touch upon the meaning of religion directly, the wider understanding compared with the traditional one is perceptible. The Court concluded that the freedom of religion includes “the right to maintain theories of life and of death and of the hereafter which are rank heresy to followers of the orthodox faiths…. Religious experiences which are as real as life to some may be incomprehensible to others. Yet the fact that they may be beyond the ken of mortals does not mean that they can be made suspect before the law” (Ballard, supra, 87-88).

After some hesitation in Berman, having concluded that a belief that lacks the concept of a deity cannot be said to be a “religious belief” in the sense in which that term is used in the Selective Service Act of 1940 (329 U.S. 795 (1946)), the Court showed an inclination towards the psychological approach in Fowler v. Rhode Island (345 U.S. 67 (1953)). “It is no business of courts to say that what is a religious practice or activity for one group is not religion under the protection of the First Amendment” (Fowler, supra, 68).

However, it was in the 1960s that the content of the concept of religion was radically and explicitly reformulated by the Supreme Court.

Although the Supreme Court did not show a willingness to supply a clear and distinct definition in Torcaso v. Watkins (367 U.S. 488 (1961)), in this case the Court undoubtedly broke with the theistic position drawn up in the Mormon cases. The Court invalidated a provision requiring all public officials in Maryland to profess a belief in God. In the Court’s opinion, the provision was unconstitutional on account of the burden it imposed on non-believers and those believers whose religion did not rest on the concept of God (Torcaso, supra, 495). The Court thus cast a wide net, confirmed by the famous footnote number eleven, which runs as follows: “Among religions in this country which do not teach what would generally be considered a belief in the existence of God are Buddhism, Taoism, Ethical Culture, Secular Humanism and others” (Torcaso, supra, 495, note 11). This list later provided an excellent basis for those who criticized the various outcomes of the attempts at defining religion. Under the Torcaso decision, both theism and non-theism may fall within the category “religion”, and it was a constitutional interpretation.

However, in the Selective Service Act of 1948, Congress defined “religious training and belief” as “an individual’s belief in a relation to a Supreme Being involving duties superior to those arising from any human relation, but does not include essentially political, sociological, philosophical views or a merely personal moral code” (Konvitz 151). The conflict between this provision of the Act and the Torcaso decision foreshadowed the necessity to reinterpret the statute. United States v. Jakobson was the prelude in the Second Circuit, in which a very broad meaning was given to the words “Supreme Being” and the court, evading in this way the constitutional problems, exempted the appellant (325 F.2d 409 (2d Cir. 1963), affirmed, 380 U.S. 164 (1965)).

In United States v. Seeger (326 F. 2d 846 (2d Cir. 1964), affirmed on other grounds, 380 U.S. 164 (1965)), in the same circuit, the court chose another possibility: instead of expanding the meaning of the words “Supreme Being”, the judges pronounced that belief in a Supreme Being is an unconstitutional condition for an exemption. Writing for the court, Judge Kaufman relied on the Torcaso ruling that theistic religion was only one kind of religion (Seeger, supra, 852).

However, the Supreme Court apparently wished to avoid invalidating the statute and followed the path beaten in Jakobson. The Court’s efforts culminated in the “parallel position test”, which was a clear functional approach to the term “religion”. The key words read as follows: “The test is whether a given belief that is sincere and meaningful occupies a place in the life of its possessor parallel to that filled by the orthodox belief in God of one who clearly qualifies for the exemption” (Seeger, supra, 166). Although the petitioner expressed scepticism about the existence of a Supreme Being (supra, 166), under this test the Court concluded that his opposition to war was religious and thus exempted him (supra, 187-188).

This test, however, does not represent a definition of religion. Hall pointed out that the Seeger test sets forth a definitional approach, not a definition itself, and indicated that the definition of religion must be obtained from an examination of religious function rather than religious content (Hall 148; see also Konvitz 154). On strict theoretical grounds, the Seeger test can really only be conceived as an approach, not as a proper definition, but in a certain sense it is one kind of description of religious beliefs, and even an open functional definition of the concept of religion.

The Seeger court thus required parallelism to the orthodox, theist kinds of religion, but did not provide us with criteria on which this parallelism worked. Despite this failure, some hints relating to such criteria can be discovered in the decision. One of the most important remarks in this respect seems to be that all beliefs fall within the category “religious” if they are sincere and “based upon a power of being, or upon a faith, to which all else is subordinate or upon which all else is ultimately dependent” (Seeger, supra, 176). Limiting the reach of exemption, the Court excluded those whose beliefs are founded “on the basis of essentially political, sociological or economic considerations”, and those whose beliefs derive from a “merely personal moral code” (supra, 173).

The key expression “ultimate dependency” as a fundamental criterion for what can be called “religious belief” clearly indicates the functional approach. The Court (and commentators) quoted Paul Tillich’s famous pronouncement here, by which the classical functional description of the concept was formulated: “And if that word (God) has not much meaning for you, translate it, and speak of your life, of the source of your being, of your ultimate concern, of what you take seriously without any reservation. Perhaps, in order to do so you must forget everything traditional that you have learned about God” (Tillich 57, as quoted by the Court, 380 U.S. 187; see also Konvitz 153; Freeman 1526; Harvard Note 1978, 1066-1067; Bowser 173).

At this point, the Court seemed to be far from the plain and clear meaning of the words “Supreme Being”. Apart from the constitutional problems, there are others that arise from the Court’s decision. The criterion of “ultimate dependency” or “ultimate concern” is not sufficient to determine the existence of parallelism in a particular case. These expressions are too abstract and they need other, more specified, operative terms by which they can obtain some sort of exactness.

Some difficulties have arisen from the way the Court mixed up the substantive and functional elements of a possible description of religion. For instance, many important elements of a “merely personal moral code” may constitute a large part of a coherent religious belief system or may suggest “ultimate dependency” or “ultimate concern” in their character. This may, in certain cases, even go for beliefs based on political, sociological, or economic considerations.

The appropriate connection of the substantive and functional elements of a possible definition requires that the substantive elements be affirmative or positive in nature. If these elements are of a negative character, excluding certain beliefs, tenets, and viewpoints from religion, the conflict between the two sorts of components is unavoidable.

Even if we reconcile the substantive and functional elements of a description of religion, there remains the crucial problem which runs through the whole issue of defining religion, that of drawing a clear line between religious and purely moral beliefs. As Seeger had not answered this question, the Court had to face the problem in the Welsh case (398 U.S. 333 (1970)).

Subsequent to Seeger, Congress enacted the Military Selective Service Act of 1967, eliminating the reference to the Supreme Being from the statute (Bowser 173; Konvitz 157). However, in Welsh this new statute was challenged by a claimant who refused to call even his beliefs religious (see 398 U.S. 341). The Court practically ignored the claimant’s own qualification because it was a “highly unreliable guide” and “very few registrants are fully aware of the broad scope of the word ‘religious’ as used in par. 6(j)” (supra). The Court held that the claimant’s beliefs functioned as a religion in his life and exempted him under the statute.

This decision was reached by a considerable reinterpretation of the Seeger test. Justice Black stated for the majority: “If an individual deeply and sincerely holds beliefs that are purely ethical or moral in source and content but that nevertheless impose upon him a duty of conscience to refrain from participating in any war…, such an individual is as much entitled to…[an] exemption…as is someone who derives his conscientious opposition to war from traditional religions convictions” (supra, 340).

The Court went on and concluded that there are two groups of registrants who did not fall within the exempted category: “those whose beliefs are not deeply held and those whose objection to war does not rest at all upon moral, ethical, or religious principle but instead rests solely upon considerations of policy, pragmatism, or expediency” (supra, 342).

The Welsh decision greatly expanded the category of religion even compared to Seeger. While Seeger had denied the exemption to those whose beliefs were essentially nonreligious (see 380 U.S. 173), Welsh held that purely moral or ethical beliefs, and implicitly even philosophical and sociological beliefs, might be functional equivalents of religious beliefs conceived in a traditional sense. These beliefs need not be rooted in any religion; they only need to function in an individual’s life as religious beliefs function in the life of believers. The other limit of the exemption is sincerity. In its decision, the Court thus also reinterpreted the new statute.

Regarding the plain meaning of the text of the Military Selective Service Act of 1967, Congress implicitly intended to exclude those from the exemption whose beliefs rest on philosophical, political, and sociological grounds in accordance with the Seeger decision. However, philosophical and sociological beliefs may implicitly fall into the category “religious training and belief” according to Welsh, and in certain cases even political beliefs can perform a religious function. On account of the functional approach, this new standard blurred the distinction not only between religious and moral beliefs but also between religious and other kinds of beliefs. Seeger and Welsh were decided by the Supreme Court in the context of statutory interpretation. The question was whether these two fundamental decisions would in some way influence the interpretation of the First Amendment to the Constitution.

The content of the word “religion” was so considerably expanded that it anticipated some change. The Court answered this question in Wisconsin v. Yoder (406 U. S. 205 (1972)).

In Yoder, a Wisconsin education statute was challenged and the Court concluded that the First Amendment mandated a religion-based exemption for Amish people from compulsory education above the eighth grade. The Court took advantage of the opportunity and expressed its opinion on what could be understood by religion under the Constitution. Yoder was much more limited in its reach than Seeger and Welsh, and the Court expressed doubts about defining religion solely by functional aspects. It was emphasized that no one is entitled to an exemption from reasonable state regulations for purely secular considerations (supra, 215).

The Court clearly held that philosophical and personal beliefs, as opposed to religious beliefs, were not protected by the two Clauses of the First Amendment and stressed the distinction between religious and secular (supra, 215-216). The statement that “the very concept of ordered liberty precludes allowing every person to make his own standards on matters of conduct in which society as a whole has important interests” (supra) indicated some kind of break with the subjective, psychological approach. While Welsh had combined personal, moral, philosophical, religious, etc., beliefs into the frame of “religious training and belief”, Yoder distinguished the secular from the religious – at least theoretically.

However, the fundamental problem was that the Court did not set forth any standard by which the various beliefs could be classified into “religious” and “nonreligious” categories, and did not attempt to define “religion” itself. This also goes for Thomas v. Review Board (450 U.S. 707 (1981)), where the Court echoed Yoder in this respect: “only beliefs rooted in religion are protected by the Free Exercise Clause…” (supra, 713). Moreover, in Thomas it was even questionable whether the belief held by Thomas represented the tenet of the Christian church of which he was a member. (See Frazee v. Illinois Department of Employment Security (489 U.S. 829 (1989), where the Court, placing emphasis on the subjective aspect of the protection provided by the First Amendment and somewhat reinforcing the prevailingly functional line of its jurisprudence, made it clear that the First Amendment protects the beliefs, based on religious commands or tenets, of those people who were not members of an established church, sect, or other religious organization.)

This changing perception of the concept of religion had some important implications. The Seeger test was not appropriate to decide a constitutional case because many nonreligious beliefs could pass the test. Yoder, Thomas, and Frazee made it clear that it is unavoidable to some extent to apply a substantive approach in evaluating a belief system. As the Supreme Court had failed to lay down principles by which the religious and nonreligious areas could have been clearly distinguished, the lower courts continued to apply their own (mainly substantive) tests. (For the most notable cases, see Fellowship of Humanity v. County of Alameda 153 Cal. App. 2d 673; 315 P.2d 394 (Ct.App. 1957) and Founding Church of Scientology v. United States 409 F.2d 1146 (D.C. Cir. 1969).) However, some of them resorted to the Seeger test (see, e.g., Loney v. Scurr 474 F. Supp. 1186 (S.D. Iowa 1979) and Krishna Consciousness. Inc. v. Barber 650 F. 2d 430 (2d Cir. 1981)). In the Malnak case, Judge Adams outlined an approach in his concurring opinion that rested on three aspects: fundamental and ultimate concern, comprehensiveness, and formal and external signs (Malnak v. Yogi 592 F.2d 197, 207-209 (3d Cir. 1979) Judge Adams concurring). This approach has been taken up by courts in other cases (see, e.g., Africa v. Commonwealth of Pennsylvania 662 F.2d 1025, 1032 (3d Cir. 1981) cert. denied 456 U.S. 908 (1982) and Alvarado v. City of San Jose, 94 F.3d 1223, 1229 (9th Cir. 1996)).

2. Theoretical efforts to define religion

As noted above, judicial practice at the highest level has failed to define adequately what religion means for First Amendment purposes, or even in the context of statutory interpretation – apart from Davis v. Beason. However, this decision, which could be regarded as providing a more or less exact description of religion, became obsolete. The obscurity or vagueness of the concept of religion in federal legal practice has given rise to an abundance of scholarly speculations on the definition of religion. (See especially Harvard Review 1622; Harvard Note 1978, 1056.)

These speculations and theories have not led to a consensus on a possible definition of religion. A formidable diversity of suggestions and conclusions can be seen. It seems that there are as many definitions of religion as there are students of religion (Harvard Note 1978, 1066). One of the possible reasons that the Supreme Court has declined to supply a definition is this diversity of scholarly opinions. Seeger showed that the Court was ready to rely on authors in this question if the final conclusions could reflect some kind of common denominator of differing opinions.

The situation is no better outside legal theory. King pointed out: “So many definitions of religion have been framed in the West over the years that even a partial listing would be impractical. With varying success they have all struggled to avoid, on the one hand, the Scylla of hard, sharp, particularistic definition and, on the other hand, the Charybdis of meaningless generalities” (King 283).

If one examines the variety of ideas on the legal definition of religion, at least two kinds of dividing line can be discerned. The first one, based on whether one or two definitions are needed for First Amendment purposes, falls outside the scope of this paper (Donovan 29-35). A cynic would say: “First, I should see a generally accepted definition of religion, and then I can tell you whether there is a need for another.” An assumption lies in the background to this problem, i.e., that in many contexts there may be a conflict between the two religious Clauses of the First Amendment (Choper 1980, 673). The basic problem is how the State can accommodate religious activities and thereby meet the requirements of the Free Exercise Clause without advancing these activities to the disadvantage of other religions and violating the Establishment Clause. At some point, accommodation may devolve into an unlawful advancing of religion though the Supreme Court has asserted that there is room for play in the joints between the two Clauses. See Locke v. Davey (540 U.S. 712 (2004)) and Cutter v. Wilkinson (544 U.S. 709 (2005)).

Even if this conflict existed, it would not follow that the two Clauses require two definitions of religion. Given the fact that it is a highly controversial matter, which requires a separate study (a good survey is provided by Ingber 281-291; Harvard Review 1631-1639; Choper 1982, 605-606), I carry on my examinations under the assumption that there may be an appropriate approach to defining religion without becoming wrapped up in the debate on the conflict between the two Clauses.

The other dividing line, that of the distinction between functional and substantive approaches, however, is of utmost importance to this paper. Both of these approaches may be identified in the practice of the Supreme Court, as was outlined in the previous part.

2.1. The functional approach

Seeger and Welsh relied heavily on Paul Tillich’s functional definition, which saw the essence of religion in the existence of an “ultimate concern”. One of the strongest supporters of Tillich’s approach in the context of the Free Exercise Clause interprets these two words in the following manner: the word “concern” signifies the affective or motivational aspect of human experience. The word “ultimate” signifies that the concern must be of an unconditional, absolute, or unqualified character. The meaning of this term is to be found in the individual’s human experience, rather than in some objective reality. So the concerns of any individual can be ranked and every person has the underlying concern that gives meaning and orientation to his whole life (Harvard Note 1978, 1066-1067). The spirit of these conclusions is clearly in evidence in the Seeger and Welsh decisions.

If we put stress exclusively on the function of a belief system in a person’s life, this deprives the concept of religion of any objective criteria. As the psychological approach blurs the objective boundaries of the meaning of religion, it therefore becomes impossible to distinguish between religious and nonreligious beliefs in a traditional sense.

The functional definition may be warranted because it can entail all the belief systems that people would generally recognize and identify as religious. A substantive definition would not achieve this task and a narrow interpretation that rests on substantive definitional elements alone would violate the Free Exercise Clause. At the same time, by eliminating all objective criteria, the functional definition would not be predicated on strong communal ties. In this context, the two Clauses may also be invoked as protecting separate, individual beliefs. (For these arguments in detail, see Harvard Note 1978, 1072-1082.)

This functional approach based on Tillich’s suggestions has been challenged in some sense by Harvard Note (1984, at 1468), though remaining within the borders of the pure functional approach. The author proposed new concepts by which the subjective, psychological approach can be properly described. Taking the scheme of a new understanding of Self, he concludes that the First Amendment protects the individual’s choice of identity (supra, 1472-1475). The Free Exercise Clause protects choices, determines the individual’s separate identity, and defines his/her connections to other human beings. The Establishment Clause protects the process of socialization, through which an individual’s identity is formed by the influence of communities (supra, 1474). The conclusion concerning the definition of religion is that religion should be defined in terms of formation of identity, not content of belief. Content-based definitions of religion contradict the very constitutional requirement because they limit the range of choice that the Free Exercise Clause was intended to protect (supra, 1475-1476). This aspect of the theory is designed to provide a description of religion in the context of the Free Exercise Clause.

The author also attempts to provide a definition which would operate within the frame of the Establishment Clause. This second definition is based on the term “social consensus” and is described by even more general categories than the first one (supra, 1485). The two definitions or descriptions are connected by the individual’s identity, which has separate and collective aspects. This is how the author wishes to eliminate the conflict between the two Clauses. This exposition on the dual concepts of religion may be considered an important step to making the functional definition of religion operative within the First Amendment and to pulling this definition out of the shadows of the Military Selective Service Act interpretation.

However, the theory built up by the author seems to have remained too abstract and it is difficult to see prima facie new practical ways guided by the theory on which coherent legal practice could be developed. (Donovan, another proponent of functional definition, also struggles with this problem (see Donovan 95).)

Although it initiates useful concepts by which the problems may be more adequately approached in some respect, from a practical standpoint these ideas cannot overcome Tillich’s basic approach. The lack of practical guidance is a characteristic of the functional approach in general, and this characteristic, which can be conceived both as an advantage and a drawback, follows from the core of the concept.

The application of the functional definition to the First Amendment has been seriously challenged. The main arguments against the functional approach can be summarized as follows.

1. There are some considerable problems with the category “ultimate concern” (or “ultimate dependency”). Although Harvard Note (1978) describes this expression well on a theoretical level, this definition offers little assistance to responding to the major practical problems. The absolute, unconditional and unqualified character of the concern is both a severe and obscure requirement for a test in a particular case (see, e.g., Donovan 64-66). Both Freeman and Greenawalt raise the question of whether many, if any, individuals have an ultimate concern (Greenawalt 1984, 807; Freeman 1537). This kind of relativity concerning the central category of the functional approach can be taken further. The approach defines religion in terms of the intensity with which the belief is held (Hall 156). So one encounters a phenomenon of quantitative character to which a conceptual quality is applied with obscure criteria, and thus the meaning of the term “ultimate” turns out to be completely undetermined. Harvard Note (1978) suggests such criteria as “an act of total personality” (Harvard Note 1978, 1076 n. 110) and “the single most important interest” (Harvard Note 1978, 1077 n. 113), which are as general as the category “ultimate” itself and so hardly specify the meaning of the term. Freeman and Greenawalt’s question thus remains reasonable (and unanswered by proponents of the functional approach).

This definitional problem also flows from the characteristics of the functional approach. The definition of religion is based solely on terms of psychology with subjective criteria. As psychology took charge of defining the concept, in a particular case only a psychologist could determine what an individual’s ultimate concern is, if any (Worthing 320). The psychological processes can hardly be described by the clear and distinct categories that are required by legal practice for a legally useful definition. In the logical construction of the functional approach, one has to rely on purely quantitative differences, which do not permit one to draw clear, qualitative, conceptual distinctions. Justice Harlan’s remark refers to that problem: “Common experience teaches that among ‘religious’ individuals some are weak and others strong adherents to tenets” (398 U.S. 358-359).

2. The immeasurable differences of the individuals’ commitments and the problems flowing from them have led to another argument against the functional approach. There are surely many individuals for whom religious faith is not an ultimate concern, an act of total personality, but they are strongly committed to certain tenets and principles and to a certain way of life. One can think of the example of a Christmas and Easter churchgoer (Worthing 321; Hall 157) and of many other similar examples as well (Greenawalt 1984, 807-809). In Hall’s view, these individuals would fall outside the protection of the Free Exercise Clause in the case of an ultimate concern-based functional approach (Hall 156; similarly Greenawalt 1984, 807). In this respect, the functional definition would be fairly tight and as a legal definition would likely be proven unconstitutional. However, from this angle the Seeger parallel test would operate coherently if the ultimate concern were not an element of it.

3. The subjective nature of the functional approach has another consequence. To put it theoretically, this subjectivity upsets the balance between self-interest and common interest, and, as expressed in Reynolds, it “permit[s] every citizen to become a law unto himself” (98 U.S. 166-167). In addition, this fear was expressed in relation to the exercise of religion conceived in a traditional sense. Since then, both the belief-action distinction and the concept of theist religion have become obsolete as limits to subjectivity.

The Court was not willing to follow the policy of wide doors formulated in Seeger and Welsh in the interpretation of the First Amendment, and the Justices endeavored to maintain the balance mentioned above. They concluded that the individual might not put forward his own standards against the common interest of society (406 U.S. 215-216). The fear expressed in Reynolds is clearly echoed, although on a more abstract level, because there is no more clear line between self-interest and common interest in the lack of a definition of religion.

4. Another objection to the functional approach would be that it locates the essence of religion only in the individual. The functional approach fails to take into account the fact that religion is a social phenomenon and has its associational aspects. If one makes religion a subjective phenomenon determined purely by the individual, one comes into conflict with the social experience that religion generally requires social mediating structures on account of its communal aspects. Through these social structures, religion becomes valuable for the individual and the society integrates the individual’s concerns into social activities and a whole communal experience. The natural need for this integration calls for some social, objective standards of religion beyond the individual’s assertions.

5. The functional definition practically diminishes the boundaries between religious and nonreligious beliefs in a traditional sense. There remains no valid test for the content of a claimed religious belief and any belief may be seen as religious if it performs the required psychic function in the individual’s life. The merging of the religious and nonreligious spheres, in Sanderson’s view, is in itself unconstitutional (Sanderson 1007). Under a functional definition, no identifiable class could be delineated as the recipient of the protection although the Constitution distinguishes a class under the word “religion” from other classes and provides special protection for that class.

If the unprotected classes (like moral beliefs) are placed into the same position as the protected one, the purposes of the First Amendment would be violated; the Constitution therefore requires a definite distinction between religion and nonreligion (Sanderson 1007).

6. It is Sanderson who sets two other arguments against the functional definition – mainly on practical grounds. A classification system based on subjective criteria, it is said, would result in an inconsistent application of the law. The other objection follows from the practical difficulty in applying the sincerity test (Sanderson 1008). As to this latter argument, however, Harvard Note (1978) does not see any difficulty and provides some guiding principles (Harvard Note 1978, 1079).

Feeling the weight of these arguments and thus the need for a content-based definition, some authors have followed the practice of lower federal courts very closely (see Worthing 322-325, 327, 332; Hall 165-169; Sanderson 997-1001; Merel 836-843; Freeman 1528-1533), and some of them have offered some ideas on specific content for the concept of religion.

2.2. The substantive approach

There are also independent, theoretical proposals for principles, with which the substantive definition of religion could be determined. As I have mentioned, some of the lower federal courts did not follow Seeger and Welsh in defining religion because those decisions had related to statute interpretation, and not to the interpretation of the First Amendment. The decisions of these lower courts, together with Yoder and Thomas, seem to provide the proponents of content-based definitions with sufficient ammunition in giving reasons for such a solution.

Austin posits a somewhat traditional model of definition based on faith as the substantive feature (Austin 30-46). Following Durkheim, Otto, and van der Leeuw’s lead, Hall makes a distinction between two categories: the sacred and profane. Although somewhat outdated for practical purposes, this dichotomy could provide a starting point to work out a sufficient, substantive definition (Hall 163-165). However, the appropriate translation of this abstract idea into a legal definition would require some other well-elaborated, operative concepts, by which this distinction could work in legal practice and without which the distinction remains quite vague (Greenawalt 1984, 805).

Rejecting the Seeger decision, Mansfield argues that religious beliefs address themselves to basic questions to which man has always sought an answer, like the meaning of human existence, the origin of being, the meaning of suffering and death, etc. (Mansfield 10, as cited by Freeman 1546). This theory is persuasively criticized by Freeman and Ingber (Freeman 1547-1548; Ingber 271-272).

Another substantive definition set forth by Choper refers to beliefs as religious if they have extratemporal consequences (Choper 1982, 599). Choper’s attempt did not enjoy general acceptance (Ingber 274-277; Greenawalt 1984, 803).

Another kind of distinction is made by Ingber. In his view, the appropriate starting point to delineate religious beliefs is the distinction between religion and ideology. The basis for this distinction is that religions, as opposed to ideologies, are not matters of human debate, evaluation, and judgment. (Ingber 277-333) Ingber’s theory (like Hall’s) is, after all, based on Durkheim’s classical distinction between the sacred and profane and falls within the category that Greenawalt calls “higher reality” theories (Greenawalt 1984, 802).

These attempts at a content-based definition of religion could be answered with different objections according to the nature and claims of the theory at issue. Apart from these specific arguments, there are two general arguments that can be formulated against practically all content-based definitions. The purpose of these two objections is to preclude any definition which requires a specific belief. One of them is quite vulnerable and practically untenable, while the other has thus far been an irrefutable argument against content-based approaches.

1. The first one is traced back to the Ballard decision. In this case, the Court was confronted with the question of whether the founder of the I Am cult obtained money under false pretenses when he solicited donations using the mails and made reference to his allegedly supernatural healing power, which was one of the central tenets of this cult. (United States v. Ballard, 322 U.S. 78 (1944); Pfeffer 482). The Supreme Court held: “Men may believe what they cannot prove. They may not be put to the proof of their religious doctrines or beliefs” (322 U.S. 86-87). Citing these words, Harvard Note (1978) concluded that classifying a belief system as religion does not depend upon the tenets of its creed and characterizing a belief as religious seems to be beyond the competence of anyone other than the adherent (Harvard Note 1978, 1063).

This conclusion is intended to support the functional claim that one can believe anything and that it is not the content that classifies a belief as religious. So it is unnecessary for the courts to examine the content of the claimed religious belief in order to decide whether it is really religious. One thus needs no content-based definition for judicial practice.

However, Harvard Note’s (1978) argumentation is weak in two – practical and logical – aspects. It does not follow from the Ballard court’s words, as presumed by the Note, that courts would be prevented from any examination of the content of beliefs claimed in the proceedings as religious. The decision only asserted it was the truth of the claimed religious beliefs that should not be examined by courts and that the adherent should not prove his beliefs, doctrines, or tenets before courts. (As was affirmed later, the truth of a belief is not open to question; rather, the question is whether these beliefs are truly held (see Seeger, supra, 185).)

It is commonplace that religious beliefs can be valued neither by reason, logic, nor even other non-rational belief systems. However, this does not preclude courts from ascertaining the content of the belief at issue and comparing it with some kind of relevant legal standards to determine whether the belief is religious or not. None of this touches on the truth or falsity of the belief under examination, so it would not conflict with the Ballard ruling, which still allows for the existence of content-based definitions.

The correctness of this conclusion is proven by judicial practice. Courts have never felt sealed off from examining at least the general nature and principal content of beliefs in order to determine whether or not these beliefs can be regarded as religious. (See the examples invoked by Sanderson 1003.) This conclusion concerning the Note’s argument makes it unsatisfying as a general objection against content-based definitions.

2. The other objection, proven irrefutable so far, is based on very simple, however (or therefore) persuasive, reasoning. Laying out a content-based definition, one must pick out one or several distinguishing elements or features which characterize beliefs that belong to the class of “religion” or “religious phenomena”. These distinguishing elements are, of course, at the same time limiting factors, which exclude other beliefs from the circle one has drawn. These elements would also constitute the essence of religion in an Aristotelian sense.

However, the limiting strength of content-based definitions seems to create confusion because it may exclude from the First Amendment protection some beliefs which are seen as religious in everyday life. This being so, in such cases, one can easily build up a charge of injustice and arbitrariness against a scholarly or legal content-based definition. So, theoretically, it is conceivable that there are types of belief systems which fall outside the strict borders of a substantive definition but which are similar in character to those within the class and so may have a claim to protection under the First Amendment. To put it another way, in that it necessarily contains a finite catalogue of acceptable beliefs, a substantive definition would be bound to leave out some beliefs which are religious in character in an everyday sense (Harvard Note 1978, 1073).

Not a single substantive definition seems to be able to resist this counter-argument. If one cannot find an example in practice or in a book addressing the history of religions, there is no obstacle to creating one. Such an example should meet only two requirements to be successful: it should have as many resemblances to the beliefs classified as religious under the challenged definition as possible and it should lack at least one of the essential elements of religion contained in this definition.

This argument has been translated to the language of legal practice, showing the legal consequences. Dissenting in Zorach, Justice Black showed (in another context) that excluding certain religious beliefs by denying their religious character represented one kind of censorship forbidden by the Constitution (see Zorach v. Clauson, 343 U.S. 318, note 4 (1952)). A narrow definition, therefore, would violate not only the Free Exercise Clause, but also the Establishment Clause by hindering some religious beliefs as against others falling within the term. This point has been expanded by Judge Brorby in one of his dissenting opinions. He argued that “the ability to define religion is the power to deny freedom of religion”. A person’s religious experience may be incomprehensible to others, including courts. By attempting to evaluate another’s religion with various tests, the court may deprive the Free Exercise Clause of its meaning. A person’s beliefs can be incomprehensible to the court and still be religious in a personal scheme of things (United States v. Meyers 95 F.3d 1475, 1489-1490 (10th Cir. 1996), Judge Brorby dissenting).

3. Fixing, distinguishing, and limiting the concept of religion

In the previous part, I outlined the distinction between the two traditional, in Greenawalt’s word, “dictionary” approaches (Greenawalt 1984, 802) to the definition of religion: the substantive and the functional. Each of these approaches has raised strong arguments against the other. These arguments reveal underlying contradictions.

3.1 Underlying contradictions

As to the purely functional approach (I mean by that the Seeger parallel test without the “ultimate concern” element), it can be challenged by showing how extremely broad this kind of definition can be. Sanderson mentions the example of the Pena case as an extreme one (Sanderson 1007), but he could have pointed out, at the same time, that the tenets of a tax law handbook might as well qualify as the raw material of religious beliefs under a purely functional definition. The consistent application of functional theories may produce strange, or even extreme, results. Following a purely functional approach, the Seventh Circuit treated atheism – in the context of the case – as religion because the appellant’s atheist beliefs played a central role in his life, and his beliefs were deeply and sincerely held (Kaufman v. McCaughtry, 419 F.3d 678 (7th Cir. 2005)).

Using the functional approach, a very consistent definition could be built up. However, despite the coherence of the system that rests on the functional approach, the successful application requires that the definition should not be in considerable conflict with the everyday concept of religion. Otherwise, the definition is not useful in legal practice. Maintaining logical and theoretical coherence and consistency may run against the common sense of religion, disregarding the conflict between the consequences of a functional definition and the common usage of the word “religion” in everyday life. So the ruling and underlying contradiction lies in the conflict between the everyday concept of religion (and the common usage which reflects this concept) and those extreme beliefs which a purely functional definition admits as religious. All this calls for some kind of substantive approach which could exclude these extreme beliefs from the sphere of religion.

As to the substantive definitions, those can be challenged by showing how extremely tight this kind of definition can be. As far as I know, no one so far has set up a substantive definition against which examples could not have been found in practice. Those authors who wished to raise arguments against the content-based definitions, or a specific content-based definition, either implicitly or expressis verbis, took the theoretical starting point mentioned above and searched for examples which did not fit the scheme.

Greenawalt criticizes the “high reality” approach in this manner with reference to the Ethical Culture (Greenawalt 1984, 805) and raises objections to Choper’s theory, referring to the use of wine for communion and to the Alaskan case (Greenawalt 1984, 803). Harvard Note (1978) also mentions a great many examples from the Greco-Oriental mystery cults to Acid Satanism with the intention of disaffirming the various kinds of existing or possible substantive definitions (Harvard Note 1978, 1068-1073).

One cannot avoid asking, on what grounds are such examples brought up as being religious in character? For a forceful argument, the religious character of these examples referred to by these authors should not be disputed. But how do we know that these counter-examples are of a religious nature if they just serve as tools for refuting definitions of religion under which they could be classified as religious or nonreligious phenomena? So the theoretical attacks on content-based definitions have implicit assumptions on which they work. Greenawalt chooses his examples mainly from judicial practice with the intention both of precluding any debate on the religious nature of his examples and of backing up his criticism of substantive definitions. Though Harvard Note (1978) does not rely on judicial practice in the examples mentioned above, these examples may still be as correct as Greenawalt’s. What may make the Note’s examples correct is an implicit reference to a generally accepted understanding of religion, which the criticized definition could not entirely cover. This implicit description or understanding should be generally accepted; otherwise, the criticism becomes arguable and weak on account of the relativity of the nature of the reference cases. “Although legal applications of a concept need not correspond precisely with how people use the concept in other contexts, it counts in favor of an approach to a basic constitutional concept that it ties to more general understandings” (Greenawalt 2006, 143).

The Note referred to the Greco-Oriental mystery cults as religion. This was possible because of the dearth of challenges to the religious character of these cults. In other words, according to the generally accepted understanding of religion or the everyday concept of religion based on a wide historical consensus on the understanding of religion, these cults qualify as religion. In the case of Acid Satanism, the author wishes to prove that this new phenomenon falls within the category of religion by referring to scholarly papers and studies. Here, the implicit point of reference is also the everyday concept of religion, and these studies constitute part of the factors that shape the general beliefs and concepts and may in the short run mark some kind of consensus.

Concerning the substantive definitions, the situation is similar to what we find in the functional approach. The defense of the coherence of the definition may lead to a sharp conflict with the everyday concept of religion. The substantive definitions may preclude some belief systems which fall under the class of “religion” according to the wide consensus represented by the everyday concept of religion. In applying these definitions in practice, this fact would generate feelings of injustice and would perhaps violate the Constitution.

In both approaches, the dominant and underlying contradiction lies in the conflict with the everyday concept of religion. No attempts, let they be functional or substantive, have so far been able to avoid the two-sided trap of Scylla and Charybdis (King 283). If you cast the net wide, some extreme cases become caught, while the strict limitation excludes some cases of which the religious character cannot be denied under a common understanding of the concept of religion.

3.2. The everyday concept of religion as controlling factor

The everyday concept of religion, as a wide consensus of what is meant by religion in an everyday sense, has developed through history and is changing. The creator of this concept is the communal experience of old and new phenomena – beliefs, rituals, worship, etc. – which plays some kind of role in the life of the given community or of its members. There are many factors that have contributed to this experience. By way of communal experience, every people shapes ideas on what is meant by religion and which phenomena should be included in this category. The result of these various individual and communal ideas being in a complicated and varied interaction with each other constitutes the everyday concept of religion which has come down from generation to generation with alterations of varying degree.

It could previously be seen that this amorphous general opinion, or the hard core thereof, can effectively control the legal definition of religion and does not permit too tight or too broad a formulation. There can be no applicable and successful legal definition which does not take into account this everyday concept. Albert Einstein pointed out: “The only justification for our concepts is that they serve to represent the complex of our experiences…” (as quoted by Solso 410).

However, this peculiar relation between an everyday concept and a legal definition entails a grave problem. The nature of a legal definition or description of religion is clearly distinct from that of its everyday concept. This difference flows from the method of assigning some content to the concept of religion in the two cases. The logical process of setting up a legal definition consists of picking out some distinguishing elements as essential. Any phenomena with those qualities can then be called religious.

In a process of formation that is historical, but not logical, the everyday concept of religion has not been organized around essential, distinguishing elements. The organizing principle is not the essence in this case. In heterogenous, everyday processes, the content of the everyday concept has been formed on the grounds of similarities, specific examples, and personal experiences which form the basis of including a novel phenomenon in the concept. It is a historical process that evolves on a case-by-case basis, not a logical one, as in the case of a scholarly or legal definition. If a phenomenon is considered religious by the general opinion of a community, it is not because this phenomenon shares some essential elements with those phenomena already pertaining to religion in an everyday sense; it is because this phenomenon is, in some way, more or less similar to some of those that already constitute instances of the class “religion”.

The differences between these processes and organizing principles lead to the difference of the content of the concept of religion in an everyday and scholarly, legal sense.

The implicit purpose of the scholarly and legal definition is to achieve as much correspondence as possible to the everyday concept by logical method. There are two important background elements of this implicit purpose. The first is that great difference makes the definition untenable. The second is the presumption that it is necessary to limit and fix the content of the everyday concept of religion because otherwise it cannot work in the legal sphere for First Amendment purposes. The vague conceptual borders, the heterogenous content, and the open-ended character of the concept prima facie make its consistent application extremely difficult or impossible in legal cases.

However, one fact must be taken into account. In light of the attempts made in recent decades, it is clear that the necessary correspondence has not been reached on account of the principal differences between the underlying logical and historical processes. This heterogenous scholarly background might have contributed to the courts’ not venturing to take a clear standpoint on the issue lest they invite criticism from one or the other side. The outcome is a careful “open-ended description” of the content of religion by judicial decision which the proponents of the traditional definition wish to avoid. In this situation, carrying on attempts at defining religion in its proper, traditional sense is, in my view, not an attractive alternative because these attempts seem by today to be devoid of any original ideas.

3.3 Non-essentialist approaches

Harvard Note (1978) showed well the main peculiarity of attempts at defining “religion”: “All attempts at definition are searches for elements whose presence identifies a belief or practice as religious” (Harvard Note 1978, 1072). This is the traditional or classical theory of defining a concept in the sense Plato or Aristotle conceived (Freeman 1549). So every belief system that is a member of the class “religion” would have to possess common defining features which are individually necessary and jointly sufficient for its inclusion in the class. Both functional and substantive approaches have attempted to find an element or elements common to all of the belief systems and accompanying practices called “religious”. This set of features would constitute the essence of religion and would also play a distinguishing, limiting role. Without this essential element, one could not speak of a definition in its traditional or proper sense, but it would be “merely a description – an open-ended compendium of historical experience” (Harvard Note 1978, 1072).

However, the courts and authors have not yet found such essential or defining features. As every attempt has failed to define the concept of religion by a set of necessary and sufficient features, this has made some authors, e.g., Greenawalt and Freeman, lay out a new, non-essentialist approach, which should be applied in the treatment of the concept of religion in particular cases (Greenawalt 1984, 753; Greenawalt 2006, 142-156; Freeman 1519).

This approach abandons the possibility of a definition of religion in its proper or classical sense; in its place, another kind of limitation and fixing of the concept is proposed. Instead of searching for overall and defining features of the concept of religion, they attempt to find those instances of religious phenomena to which the concept indisputably applies and to examine in doubtful instances how close the analogy is between those instances and the doubtful instance. The central category of this approach is the indisputable instances of religion (Greenawalt) or the standard paradigm of religion (Freeman), which would constitute the core of the concept. With these categories, those instances (beliefs and practices) are meant of which no one can reasonably deny the religious nature. To determine whether or not questionable belief systems and practices are religious, one must examine and determine how closely they resemble what is undeniably religious. This approach – which Greenawalt calls an “analogous” (Greenawalt 1984) or “analogical” (Greenawalt 2006) approach – can provide applications of the concept to instances that do not share common features (Greenawalt 1984, 763 and 815). Descriptive in character, this non-essentialist theory facilitates explanation of the courts’ decisions and reasonings on religious matters; also being normative in nature, it may show a potential way to fix and analyze the concept of religion.

Rejecting that the concept of religion has a definable essence or defining features, the analogical approach does not define religion in its proper sense; it outlines the content of religion by identifying some of its salient and accidental (but not necessary) features. This description determines some relevant features of religion which are generally accepted as such within a community by historical and social experience, and these features provide a paradigm of a religious belief system. After all, this description represents the standard meaning of religion. None of these elements of the description can be regarded as necessary for a belief system and its accompanying practices to qualify as religious because it would mean becoming caught in the trap of a traditional defining method. The features can be classified into two groups according to whether or not they are sufficient in themselves for a belief system to qualify as religious. Compared with that paradigm, a belief system is more or less religious depending on how closely it resembles the standard paradigm or the indisputable instances. Besides the standard paradigm, Freeman also proposes a borderline-case paradigm as a possible standard for the borderline cases (Freeman 1565).

The theoretical roots of this approach seem to be represented by Hart’s “open texture” argument in legal theory, the late Wittgensteinian ideas on the use of language in philosophy, and Eleanor Rosch’s prototype theory as well as the exemplar theory in cognitive psychology (Rosch 1973, 328-350; Rosch 1975, 573-605). Following Waisman’s lead, Hart pointed to the “open texture” of language, demonstrating that vagueness is a characteristic feature of concepts (Hart 147-159). Empirical concepts (in a Kantian sense) do not have a precise meaning which may serve as a clear guide in the application of the concept in all practical contexts – the vagueness inherent in natural language produces many borderline cases (Kellogg 174). Breaking with Plato’s and Aristotle’s theories of definition, Wittgenstein showed that the search for essences might not be correct in defining a concept. Considering cases that fall within a class, one often finds that there is nothing common to all members of a class and that there are only similarities, heterogenous relationships, and family resemblances (and not essences or defining features) that link the instances of the class together (Wittgenstein 57-63; Kellogg 171-173).

From applying this non-essentialist trend in the theory of concepts to the problem of defining religion, important consequences ensue, which also lie at the heart of the “analogical” and “standard paradigm” approaches.

(1) The concept of religion lacks an identifiable essence, that is, overall defining features common to all the cases belonging to this class. The concept has a family resemblance structure which cannot be properly defined by traditional methods.

(2) The determining factors of horizontal relationships among members of the class are not homogenous; the characteristics and nature of these inner-class relations may differ in each of the cases.

(3) The concept of religion (as an empirical concept) does not have sharp conceptual boundaries, and thus no traditional definition of this concept may exist which would be appropriate to separate with reasonable certainty the religious and nonreligious spheres of social phenomena.

(4) A novel instance of a possibly religious phenomenon can be classified on the basis of similarity to the instances, or some of the instances, of religious phenomena already included in the class (see, e.g., the “analogical” approach or the exemplar theory in cognitive psychology) or to a more abstract, typical instance of the class (see, e.g., the “standard paradigm” approach or the prototype theory in cognitive psychology).

(5) The categorization of belief systems and accompanying practices as religious or nonreligious is not a question of quality, but a matter of degree. A social belief system is not religious or nonreligious. Not possessing an all-or-nothing character, it is more or less religious, depending on similarity to the established instances of the class religion.

(6) Within the class of religion, there are instances of religious phenomena which are more typical of the class than other instances still belonging, or possibly belonging, to the class. The basis of comparison under which the degree of religiousness can be established is a kind of standard which has different names, e.g., indisputable instances of religion (Greenawalt) or the standard paradigm of religion (Freeman), or prototype (prototype theory (Rosch)). In these theories, the nature of this standard and the methods by which an instance is related to the standard may be slightly different, but the central theses and methods of these theories are essentially similar.

3.4 The appropriateness of the non-essentialist approach

The changing understanding of concepts and concept formation may provide a way to escape the circulus vitiosus of the definitional problem relating to religion. The strict frame of the logical process of defining can be loosened and the process can be pushed nearer to the historical one, in which the everyday concept of religion was formed. As a matter of fact, non-essentialist approaches proceed this way. These approaches are, after all, a logical simulation of the historical process. In this way, one can arrive at a description of religion which may correspond to the everyday concept and which is still controlled by logic. This logical control may make this description adequate for application in the legal sphere.

The analogical, the standard paradigm, and the prototype approaches are, of course, not ideal for handling the concept of religion. They each have their own shortcomings, but they may still be the best way in the reality of legal practice. Though practically forsaking the possibility of defining religion in a traditional sense, they offer another kind of limitation. The traditional definition aims at drawing a circle around those cases which fall within the category. Its main function, therefore, is to fix, distinguish, and limit a concept. Renouncing the possibility of defining, however, does not mean giving up the intention and possibility of fixing, distinguishing, and limiting the concept. This can be achieved by other means than definition, and the non-essentialist approaches present some of these means.

Ingber’s main argument, directly against Greenawalt’s theory, indirectly against the whole system of non-essentialist approaches, is that it is “a totally open-ended perspective” and it is therefore untenable for legal purposes (Ingber 273). In Ingber’s view, the analogical process is not legal in nature and fails to fulfill legal and judicial expectations (Ingber 274). Ingber needs this criticism in defense of his substantive theory, which is built on the possibility of a working traditional definition.

However, his criticism is not convincing. He shares Choper’s position that courts can only work on the basis of traditionally defined categories, clear and distinct ideas: “without the ability to identify necessary and sufficient characteristics, the court would have no articulable basis for distinguishing between religious and nonreligious beliefs” (Ingber 274).

However, from the 1940s, or at least from the Torcaso case, federal courts have not had an “articulable basis” based on a definition in a traditional sense; they have sometimes instinctively applied some kind of analogical method to classify a disputable instance of religious phenomena. On the other hand, even in the legal sphere, one rarely encounters strict definitions with clear limits. On the contrary, the application of law is partly a task of redefining, reinterpreting, limiting, and distinguishing as words, categories, and concepts are dealt with in particular cases. Taking this into consideration, it is a fairly arguable position that only a clear and strict definition will do for legal purposes. One can set up the standard of a clear, traditional definition, but it may not meet the requirements of reality.

If every category and concept were defined clearly and distinctly, this situation might be ideal, but it is beyond the reach of human logic and rationality. But, if the setting up of a generally acceptable, traditional definition seems impossible for the time being, it is highly desirable to offer at least some, theoretically well-based guiding principle for the courts to describe the content of religion.

The cornerstone of Ingber’s criticism is the alleged “open-endedness” of the analogical approach. This contention wishes to challenge the limiting function of an analogical description. However, it is fairly obscure what is meant by total, or just simple, open-endedness. This expression also occurs in the Harvard Note study (1978) without any elaboration.

Compared with a traditional definition, an analogical description is less limited, but not open-ended in a strict sense. It may have a wider area for the borderline cases, but Ingber’s criticism can hardly disaffirm the limiting function of these approaches. The paradigm of the standard meaning of “religion” may also contain very significant limiting and distinguishing strength. The idea of the analogical approach in itself does not bring about success in terms of the limitation of the term “religion”. There are some aspects of the approach on which it depends whether the analogy can provide an effective means for limitation and distinction in the treatment of the category “religion” in practice. Whether or not the analogical method can effectively perform the limiting and distinguishing function depends on the appropriate elaboration of the following aspects:

  • the careful selection of the elements of the standard paradigm;
  • the definition of the elements of the standard paradigm and the detailed description of the relationship between them;
  • the elaboration of the method of analogy in the legal sphere, that is, how to apply the paradigm in particular cases;
  • the establishment of standards concerning the process of the identification of similarities and dissimilarities, and thus limiting subjectivity;
  • the elaboration of the specific treatment for borderline cases.

These five suggestions cover two crucial points as requirements with regard to the non-essentialist descriptions. One, the analogical approach may be effective in appropriately determining the main features of the everyday concept of religion in a given society, because the standard meaning or prototype or the indisputable instances used as reference points should be based directly on the everyday concept of religion. Two, this approach is only suitable for practical use if the employment of carefully elaborated methods of analogy and comparison can decrease subjectivity in the process of evaluating a particular instance or phenomenon using reference (standard) instance(s).

Cases cited

  • Africa v. Commonwealth of Pennsylvania, 662 F. 2d 1025 (3d Cir. 1981) cert. denied 456 U. S. 908 (1982)
  • Alvarado v. City of San Jose, 94 F.3d 1223 (9th Cir. 1996)
  • Berman v. United States, 156 F. 2d 377 (9th Cir.) cert. denied 329 U.S. 795 (1946)
  • Church of Holy Trinity v. Unites States, 143 U. S. 471 (1892)
  • Cutter v. Wilkinson, 544 U.S. 709 (2005)
  • Davis v. Beason, 133 U.S. 333 (1890)
  • Dettmer v. Landon, 799 F.2d 929 (4th Cir. 1986), cert. denied, 483 U.S. 1007 (1987)
  • Everson v. Board of Education of Ewing, 330 U.S. 1 (1947)
  • Fellowship of Humanity v. County of Alamada 153 Cal. App. 2d 673; 315 P. 2d 394 (Ct.App.1957)
  • Founding Church of Scientology v. United States, 409 F. 2d 1146 (D. C. Cir. 1969)
  • Fowler v. Rhode Island, 345 U.S. 67 (1953)
  • Frazee v. Illinois Department of Employment Security, 489 U.S. 829 (1989)
  • Gabriell v. Knickerbocker, 12 Cal. 2d 85, 90, 82 P.2d 391, 393 (1938), appeal dismissed, 306 U.S. 621 (1939)
  • Kalka v. Hawk, 215 F.3d 90 (D.C. Cir. 2000)
  • Kaufman v. McCaughtry, 419 F.3d 678 (7th Cir. 2005)
  • Krishna Consciousness. Inc. v. Barber, 650 F. 2d 430 (2d Cir. 1981)
  • Locke v. Davey, 540 U.S. 712 (2004)
  • Loney v. Scurr, 474 F. Supp. 1186 (S.D. Iowa 1979)
  • Malnak v. Yogi, 592 F. 2d 197 (3d Cir. 1979)
  • Nicholls v. Mayor of Lynn, 297 Mass. 65, 70, 7 N. E. 2d 577, 579-80 (1937)
  • Reynolds v. United States, 98 U.S. 162 (1878)
  • Smith v. Allen, 502 F. 3d. 1255 (11th Cir. 2007)
  • Thomas v. Review Board, 450 U. S. 707 (1981)
  • Torcaso v. Watkins, 367 U.S. 488 (1961)
  • United States v. Ballard, 322 U.S. 78 (1944)
  • United States v. Jakobson, 325 F.2d 409 (2d Cir. 1963), affirmed, 380 U.S. 164 (1965)
  • United States v. Kauten, 133 F.2d 703 (2d Cir.1943)
  • United States v. Macintosh, 283 U.S. 622 (1931)
  • United States v. Meyers, 95 F.3d 1475 (10th Cir. 1996)
  • United States v. Seeger, 326 F. 2d 846 (2d Cir. 1964), affirmed on other grounds, 380 U. S.164 (1965)
  • United States v. Welsh, 398 U.S. 333 (1970)
  • Washington Ethical Society v. District of Columbia, 249 F.2d 127 (D.C. Cir. 1957)
  • West Virginia State Board of Education v. Barnette, 319 U.S. 658 (1943)
  • Wisconsin v. Yoder, 406 U.S. 205 (1972)
  • Zorach v. Clauson, 343 U. S. 306 (1952)

Works cited

  • Austin, Andrew W.: Faith and the Constitutional Definition of Religion. Cumberland Law Review 22(1991-92) at p. 1.
  • Bowser, Anita: Delimiting Religion in the Constitution: A Classification Problem. Valparaiso University law Review 11(1977) at p.163.
  • Choper, Jesse H.: The Religious Clauses of the First Amendment: Reconciling the Conflict. University of Pittsburgh Law Review 41(1980) at p.673.
  • Choper, Jesse H.: Defining “Religion” in the First Amendment. University of Illinois Law Review 3(1982) at p. 579.
  • Developments in the Law: Religion and the State. Harvard Law Review 100(1987) at p. 1606. – quoted as “Harvard Review”
  • Chronowski, Nóra – Drinóczi, Tímea – Petrétei, József – Tilk, Péter – Zeller, Judit (2006): Magyar alkotmányjog II. [Hungarian Constitutional Law II.] Dialóg Campus, Budapest – Pécs
  • Donovan, James M.: God is as God Does: Law, Anthropology, and the Definition of “Religion”. Seton Hall Constitutional Law Journal 6(1995) at p. 23. online source: http://works.bepress.com/james_donovan/15 (visited on January 23, 2009)
  • Freeman, George C.: The Misguided Search for the Constitutional Definition of “Religion”. Georgetown Law Journal 71(1983) at p. 1519.
  • Greenawalt, Kent: Religion as a Concept in Constitutional Law. California Law Review 72(1984)
  • Greenawalt, Kent (2006): Religion and the Constitution: Free Exercise and Fairness. Princeton University Press, Princeton http://books.google.hu/books?id=bLVqPYcR8IQC (visited on January 20, 2009)
  • Hall, Timothy H.: The Sacred and the Profane: A First Amendment Definition of Religion. Texas Law Review 61(1982) at p. 139.
  • Hart, Herbert L. A. (1995): A jog fogalma. [The Concept of Law.] Osiris, Budapest
  • Ingber, Stanley: Religion or Ideology: a Needed Clarification of the Religious Clauses. Stanford Law Review 41(1989) at p. 233.
  • Kellogg, Ronald T. (1995): Cognitive Psychology. SAGE, London
  • King, Winston: Religion. In: Encyclopaedia Vol. 12., at p. 282.
  • Konvitz, Milton R.: The Problem of a Constitutional Definition of Religion. In: Religion and the State. Essays in Honor of Leo Pfeffer (1985) (Ed. Wood, James E.) Baylor University Press, Waco – Texas, at p. 147.
  • Mansfield, John: Conscientious Objection – 1964 Term. Religion and the Public Order 3 (1965) at p. 1., quoted by Freeman, 1546.
  • Merel, Gail: The Protection of Individual Choice: A Consistent Understanding of Religion Under the First Amendment. The University of Chicago Law Review 45(1978) at p. 805.
  • Note: Reinterpreting the Religious Clauses: Constitutional Construction and Conceptions of the Self. Harvard Law Review 97(1984) at. p. 1468. – quoted as “Harvard Note 1984”
  • Note: Toward a Constitutional Definition of Religion. Harvard Law Review 91(1978) at p. 1056. – quoted as “Harvard Note 1978”
  • Robbins, Thomas – Anthony, Dick: New Religions and Cults in the United States. In: Encyclopaedia Vol. 10., at p. 394.
  • Rosch, Eleanor – Mervis, Carolyn B.: Family resemblances: Studies in the internal structure of categories. Cognitive Psychology, 7 (1975) at p. 573.
  • Rosch, Eleanor: Natural categories. Cognitive Psychology 4(1973) at p. 328.
  • Sanderson, Michael G.: Objective Criteria for Defining Religion for the First Amendment. Toledo Law Review 11(1980) at p. 988.
  • Schanda, Balázs: A vallásszabadság kérdéseinek alkotmányi szabályozása. [Religious Freedom under Constitutional Rules] In: Alapjogok és alkotmányozás. (1996) (Ed.: Ádám, A.) Közgazdasági és Jogi Kiadó, Budapest, at p. 245.
  • Schanda, Balázs (2003): Magyar állami egyházjog. [Hungarian Laws on Churches] Szent István Társulat, Budapest
  • Solso, Robert L.(1995): Cognitive Psychology. Allyn and Bacon, Needham Heights
    The Encyclopaedia of Religion.(1995) (Chief Editor: Eliade, Mircea) Volumes 1-16. Simon & Schuster Macmillan, New York – quoted as “Encyclopaedia”
  • Tillich, Paul J. (1948): The Shaking of the Foundations. Charles Scribner’s Sons, New York
  • Wittgenstein, Ludwig (1998): Filozófiai vizsgálódások. [Philosophical Investigations] Atlantisz, Budapest
  • Worthing, Sharon L.: “Religion” and “Religious Institutions” under the First Amendment. Pepperdine Law Review 7(1980) at p. 313.