When May Government Interfere with Religious Practices to Protect the Health and Safety of Children?

Abstract

Secular states may be asked to override parental decisions based on religious beliefs when these decisions lead to concerns for the wellbeing of the child or welfare of society at large. Particular difficulties arise when goals of limited state importance to secular society conflict with practices that matter a great deal to members of religious groups. We propose a prudentially-based approach to resolving these conflicts by balancing the interests of the affected child, those close to the child, and society without compromising the child’s needs. This approach acknowledges the importance of children’s relational interests as members of families and religious cultures. It is compatible both with legal and bioethical practice. Decisions are contextually framed, taking into account the degree to which the state generally tolerates risks to children’s health and safety. The constraints built into our approach protect against using child welfare as a pretext for attacks on minorities. Finally, use of this approach partially addresses the imprecision of other standards that have been applied to judge the appropriateness of parental decisions.

Several families migrate to a large city in the Western world. Hassan and Nadia come from Kurdistan and object to their daughter taking part in coeducational swimming lessons. Krish and Shanti resided in a Caribbean nation where children are subjected to frequent disciplinary spanking on the buttocks. Gideon and Yael are Jews from Israel. They circumcised their son and give him a small glass of wine at Sabbath dinner. All of these cultures and/or religion-based practices may impact on the health of the children involved, but all are treated with varying degrees of acceptance in various Western nations.

There may be tension between secular governments and minority cultural and religious groups regarding certain practices, especially those that potentially threaten the health of children. The sorts of decisions that can cause conflicts between a secular state and a religious family may involve issues other than health. Education, disciplinary practices, and even baby naming may be issues. Furthermore, religious views may be idiosyncratic, without linkage to a recognizable religious group. The focus of this essay, however, will be restricted to practices of minority religious groups that threaten the health or safety of children. While the ethical discussion at the heart of many of these practices is similar (namely, balancing religious prerogatives with a societal duty to the welfare of children), policy and legal treatment has varied. A fair and predictable rubric for addressing such practices is necessary.

In this essay we consider the issues at stake in balancing parents’ right to raise children according to their religious tradition against the state’s obligation to protect children and present a rubric to consistently and fairly balance these interests. Problematic parental decisions based on religion are frequent. They present particular difficulty because of the intrinsically powerful influence of religion. Theoretical approaches to resolving these disputes range from viewing parents as mere stewards of their children on behalf of the state [1] to great deference to parental decisions, especially those grounded in minority cultural influences [2]. We propose a test that can be applied in the context of both law and the practice of clinical ethics to resolve these problems individually or collectively. Our position prioritizes the needs and interests of the child, while also respecting those of society and the family.

The paradoxes inherent in liberalism

While our position is grounded in liberalism, we believe that there cannot be a liberal consensus regarding state treatment of illiberal religious action. This is because of certain paradoxes inherent in liberalism. We live in states that aspire to liberal democratic governance. Liberalism is any political formulation that prioritizes individual autonomy, and protection at least of negative liberties—those that enhance and protect autonomy. Such liberties include freedom of speech and freedom to alienate one’s property. Liberalism is a cluster concept that encompasses many political formulations ranging from near-anarchic libertarianism to certain forms of socialism. Liberal views range from emphasizing individual choice to embracing extensive state restriction on association and behavior in the interests of equality [3]. Liberal conceptions may include positive rights—duties by others to provide affirmative goods. An in-depth review of liberal theoretical political philosophy is beyond the scope of this manuscript.

However, if individual choice is the touchstone of liberalism, then liberals must deal with two paradoxes, which we will call the Paradox of Illiberalism and the paradox of Values Pluralism. The Illiberalism Paradox [4–6] concerns the interaction between a liberal state and members of that state who wish to associate in groups that govern themselves illiberally. At the extremes, there are two paradigmatic liberal responses. First, a government can decide to tolerate such groups. This is called Renaissance liberalism [6], diversity liberalism [7], or liberal pluralism [8]. Under such a regime, though, members of the illiberal groups do not enjoy the full benefits of living in a liberal society. This viewpoint has been championed by Berlin [9] and Galston [8]. The other paradigmatic view is that governments should compel their citizens to abide by prevalent liberal norms. This view, called autonomy liberalism [10], Enlightenment liberalism [6], liberal imperialism [11], or liberal monism [11, 12], was championed by Barry [13] and Okin [14]. Its problem is that by forcing a liberal life on people who do not want it, the state is acting illiberally.

Value pluralism was championed by Isaiah Berlin. He noted that laudable values van be in conflict [15]. Thus, liberty and equality, justice and mercy, or knowledge and happiness may be incompatible at times, and there is no rational way to choose between them. Berlin’s solution is prudential compromise [15]. Value pluralism reveals that democracy, if defined as majoritarianism [16], is incongruous with liberalism. Majorities may wish to restrict liberties of some minority classes. Liberalism also is incongruous with Pareto efficiency—i.e., the situation where the circumstances of one individual cannot be made better without making the situation worse for another individual [17]. Liberalism, if conceived as commitment to negative rights, is in tension with equality [18, 19]. As we shall see, liberalism also is in sharp tension with religion.

In summary, liberalism not only is inconsistent with other important principles, but is internally inconsistent. Nonetheless, most Western nations adhere to some version of liberalism. Consequently, there are times when policy-making or decision-making forces them to confront these inconsistencies. Although many authors have proposed principles for reconciling these paradoxes, a consensus principle-based approach remains elusive [4, 6–14]. A fortiori, a consensus political solution that reconciles different viewpoints held by varying proportions of the electorate is difficult when different cultural (including religious) groups within the state are in disagreement. We therefore propose that a successful resolution must be prudential or expedient. We follow Berlin’s lead, and seek a pragmatic modus vivendi. Following Berlin, we prefer to minimize restriction of citizens and to respect sub-national groups that constitute civil society.

The inevitable tension between religion and the liberal state

For purposes of this discussion, we restrict our definition of religious belief as an idea or system of ideas that has a transcendental or numinous component. Religious beliefs cannot be inferred from empirical observation. We further define a religion as a community organized to maintain those beliefs, to affirm them, and to pursue actions that follow from them. Religious belief and action may be affirmed outside the context of a religion. Furthermore, the line between religious and other practice is blurred. This essay, however, will focus on parental religious beliefs that are normative in religions to which those parents belong. Religious beliefs need not be theistic; karma is an example of a transcendent concept that does not require belief in a personal god [20]. Because of its transcendent nature, religion fails the test of general accessibility that Rawls [21] and Audi [22] would require for civic discussion. This proposed requirement obviously would preclude a state from engaging religious beliefs and religions that hold them in meaningful, or perhaps in respectful, dialog.

We contend that such engagement is necessary, though, and will discuss some ground rules for such engagement. The theology and practices of a religion are not determined by literal interpretations of canonical works, but rather heuristic traditions and the actions and beliefs of the laity. Religious practices go beyond theological affirmations and rites of worship. They may involve many aspects of activity in the natural world [23]. Secular governments cannot judge the truth of religious beliefs, though they can rule on the sincerity of those beliefs [24]. Liberal governments mostly must deal with religions as they are; they can regulate their activities but should not police their beliefs. For example, United States jurisdictions may not settle matters of church doctrine [25]. The European Court of Human Rights, though, allows states under its jurisdiction broad control over doctrine of established churches, and some power to intervene in doctrinal disputes in non-established religions [26]. A recent European Court of Human Rights decision precluding states under its jurisdiction from resolving doctrinal differences within religious communities [27] was barely affirmed, by the narrow margin of 9–8. Thus, Europe and the United States may have a different view of state power over the internal organization of religions.

Certain aspects of religious belief distinguish them from other visions of the good. Religions expect that their precepts will govern followers’ thoughts and behavior. They hold that people are accountable for their behavior to a divinity [28], to the cosmos [20], or to a transcendently-bestowed natural law [29]. This accountability is especially fraught with significance when religious beliefs include postmortem reward or punishment. Even absent belief in an afterlife, many religions propose a natural law yoked to a transcendental scheme that organizes the material universe. It seldom is difficult for members of a majority religion to comply with demands of their culture. This is particularly true in the Protestant West, where mainstream denominations have few quotidian requirements. However, groups such as observant Jews, Mennonite Christians, practicing Hindus, and religious Muslims suffuse most everyday actions with religious implications. Areas subject to religious control includes eating, dress, sexual expression, and child rearing.

If only God’s imperative or God’s natural law determines if an action is right or wrong [30], then the demands of religion with regard to such an action take precedence over the demands of the state [31–33] (citing Catholic and Protestant theologians, and a Jewish legal scholar). Religion-motivated practices such as refusal of medical care for believers and their children [34] therefore can lead to conflicts between the state, and religious people and groups.

Liberal states generally allow adults to make personally detrimental decisions, but often intervene when detrimental decisions involve children. In the United States, such intervention is based on the common-law doctrine of parens patriae, which establishes the state as the ultimate protector of incompetent persons [35]. Europe relies on written charters of rights that protect children. States intervene when parents deny their children medically necessary blood transfusion or legally required vaccination [36, 37]. Even where there is little danger of severe immediate harm, some governments have regulated infant circumcision [38] or have removed grossly obese children from their families [39].

Pragmatic considerations underlying the test

American responses to resolving state conflicts with religious and cultural minorities have been ad hoc and inconsistent. Though it is not possible to rationally construct a principled solution to this question, it nonetheless is desirable to find a consistent approach. Intuitively, some distinctions seem widely acceptable. One parameter to consider is potential harm to the child. Piercing the ears of an infant seems trivial, while binding the feet of an infant is unconscionable. Another parameter is potential harm to society. Instruction in religious doctrine generally seems harmless, but that may not be the case if a religious school teaches students to wage war against the host state or to disregard its laws. The bounds of government action must be drawn somewhere between the extreme of de minimis departures from mainstream norms and severe threats to those norms or to society itself.

Not only minority practices, but also the thick values of a society, are relevant. Corporal punishment may be more acceptable in Korea than in the United States, and more acceptable in the United States than in Sweden. For this reason, we will propose to compare harms inherent in minority practice to harms tolerated by the majority. Furthermore, by using terms like ‘substantial’ and ‘unreasonable,’ we will allow for different determinations under different conditions.

Any approach to resolving parent-state conflicts must be compatible with the legal system of the state. In the United States, important Supreme Court decisions [40, 41] have been understood to constitutionalize parents’ extensive rights to raise their children as they see fit, even where parents' values conflict with social norms and values. Although these opinions directly rule on educational issues, they contain sweeping obiter dicta that have been influential in applying the constitution to other issues involving children, including health issues [42]. This parental power does not extend to putting a child into a position of danger [35].

The test we will propose provides an improvement over ad hoc determination in that it provides a clear framework that respects all relevant parties. It requires those who would restrict religious or cultural practices to provide a credible justification.

The ethical and legal context for the test

An ideal approach to determining when a state may appropriately interfere with parental decisions regarding child welfare will have at least three features. First, the approach must be sufficiently determinative to provide fairly clear guidance and limits for state decision-makers. In other words, it should allow for a high degree of consistency and predictability in decision-making. Analogous cases should lead to similar outcomes regardless of the individuals involved. Clearly, though, there will always be borderline cases or new problems that prevent complete elimination of indeterminacy. Second, the approach must be compatible with the constitutional norms of the state. The approach should be capable of generating both ethics-based and law-based decisions. On one hand, clinical ethics is toothless without the force of law. On the other hand, the law should reflect ethical values. If ethics and law are intertwined, then guidelines for resolving conflicts should be applicable towards both.

Finally, the approach must be ethically coherent. Decision-making under the approach must be transparent, and decisions emanating from the approach should meet ethical, as well as legal, standards. Furthermore, there must be consensus in any relevant community as to the applicability of the approach. Such consensus requires, in our opinion, an approach that is compatible with a broad range of metaethical systems. Many consequentialist, secular deontologists, care-based ethicists, and practitioners of most religious systems common in the West should feel comfortable using the test we will propose. An ethically coherent approach should not only be widely acceptable, but should uphold substantive elements of an ethical consensus. Intuitively, an approach that will be successful in Western societies should give greatest weight to the needs and interests of the child, while also addressing those of other interested parties [43–46]. The approach should not allow child welfare to be used as a pretext for discrimination against minorities, including religious minorities [44–46]. The harms that society and affected individuals incur from of implementation of state decisions should not exceed the costs of the parentally-inflicted harms that those decisions are meant to prevent, abate, or punish. Three main approaches have been applied to the question of conflicts between states and parents over decisions regarding children. The prevailing approach is the best interest’s standard (BIS). Beauchamp and Childress [47, p. 102] regard this as the appropriate benchmark for decisions regarding health care of children. However, this is a flawed standard [45, 46, 48]. Taken literally, BIS implies that an authority exists that is capable of knowing what is best for a child, and that this authority should have plenary power to bring about this optimal situation. This includes the power to interfere with any caretaker whose decisions or actions do not reach towards this ideal situation. A few scholars, such as Dwyer, seem to propose such a regime [50]. It is doubtful that most citizens would agree with this literal BIS formulation. Kopelman [51] proposed three alternative interpretations of BIS. One is the literal interpretation—which she regards as an aspirational ideal. The second meaning approximates avoidance of serious harm. This application of the term BIS has much in common with Diekema’s harm principle (HP) approach, which will be discussed below. This is the de facto standard for intervention against caretakers. Kopelman’s final interpretation converts BIS to a standard of reasonableness, requiring good, but not optimal outcome. This interpretation is applicable when the decision-maker faces multiple non-ideal options. It is the applicable standard in family law in matters such as custody determinations [52].

There are several problems with BIS [48]. First, using the term in several ways detracts from coherence and determinacy, unless it is clear which of these uses is applicable to any situation. Second, BIS does not take into account the needs and interests of adults or other children in the family, let alone of the interests of the broader community. Finally, the premise that the best interests of a child can be determined in a more or less objective way is invalid. Clearly, parents and authorities can disagree over what the child’s best interests are. Physicians tend to focus on a child’s health interests. However, children have economic, psychological, and spiritual interests as well [48]. Interests may be present- or future-directed, so that a current harm may be likely to bring about future gain. Interests may be self- or other-directed, so that a mature child may wish to sacrifice direct self-interest to benefit someone she loves [48]. When there is a dispute, authorities are apt to frame best interests according to majoritarian standards [49]. Thus, to the extent that a state assumes authority over caretakers, it is likely to interpret best interests through the lens of the norms and prejudices of either the majority or of the cultural and legal elite. Furthermore, there has been insufficient attention to their relational rights [49]. BIS is the benchmark for assigning parental rights. However, the Supreme Court held that it was unconstitutional under the Due Process clause of the 14^th Amendment to give judges wide discretion under BIS to challenge parental decisions [53]. However, some courts continue to rely implicitly or explicitly on BIS in overriding parents’ refusal of health care in non-acute situations [54, 55].

Diekema’s reliance on HP [56] is an improvement over the BIS. His eight requirements for state intervention [56, p. 252] add nuance to the HP criterion for action, and invoke the interests of non-child parties. Diekema’s HP is more respectful of parents than is BIS (Condition 8). He implicitly takes into account the possibility of bias in state action (Condition 7). However, Diekema’s HP analysis does not address interests of parties beyond the family. His constraints on action (immediacy of harm and absence of less intrusive options) may be unreasonably strict. Furthermore, the ‘serious harm’ standard, while more precise than ‘best interests,' still contains sufficient indeterminacy to allow different outcomes in identical cases. Thus, two judges in the same London district came to opposite conclusions in almost identical child custody cases in which HP was the legal standard [57]. As long as humans with different experiences and biases are interpreting laws and standards, some indeterminacy is inevitable. Indeed, our own formulation does not eliminate indeterminacy, though we believe that there is less of it in our mode of analysis than is contained in a BIP or HP analysis. Finally, while a literal interpretation of BP may be too restrictive, we believe that HP is not restrictive enough. First, it does not allow intervention where there is an important state interest, as in limiting exemption to compulsory vaccination. Second, it may preclude intervention when the risk to the child is serious, but not immediate.

Ross [58] and Iltis [59] propose a standard for intervention intermediate between BIS and HP. They would require that parents provide an upbringing that allows their children to become autonomous adults [58,59]. Parents should enjoy a zone of discretion in which the criteria for state intervention lie somewhere between the broad literal BIS standard and the narrow standard of immediate serious harm. These authors also emphasize the situation of the child as a member of a family all of whose members have needs, including other children.

In justifying the ordinary situation assigning parents the role of default decision-makers, Ross relies on four reasons proposed by Buchanan and Brock [60]. First, parents know their children and their needs better than others, and are best able to make decisions that will serve their children’s welfare. Second, parents bear many of the consequences of choices made for their children, and so should have some control over these choices. Third, there is a prima facie parental right to raise children according to the parents’ own values. Finally, the family itself is a valuable social institution. There are at least two other justifications for parental power to make decisions. The opportunity to replicate oneself not just physically, but to perpetuate one’s values and concerns, is a prime motivation for having children. Reproduction is essential to society. A system that allows parental leadership will be more inclined to promote families than will a system that regards parents as stewards of their children acting on behalf of the state [1], or even as fiduciaries of their children, under the state’s close supervision [61]. Finally, the vast number of decisions parents make cannot be supervised without undue intrusion. Iltis imagines (as a reduction ad absurdum) home visits by the food police and the book police to ensure that children receive nutritious diets and exposure to educational books [59].

We agree with Ross and Iltis that a parent is not a fiduciary or a steward (except in the relatively unusual situation in which the parent is a trustee for the child). This is true because of the phenomenon of love between parent and child; permanency of the parent-child relationship; children’s long-term dependency; resources required to raise a child; closeness induced by frequent physical contact, feeding, etc.; and absence of monetary consideration as compensation for parents (as would be true for a fiduciary or steward), among others. Parents make multiple decisions of greater and lesser magnitude involving their children. Parents’ beliefs and habits inevitably will suffuse these decisions, whether quotidian or momentous. This is true of religious beliefs and values, whether they are mainstream or minority.

Some feminist legal scholars have reached conclusions similar to those of Ross and Iltis, emphasizing the importance of the parental relationship, and even attributing rights to the relationship per se [62,63]. Since individual rights are meaningful only in the context of the relationships that support them, individual rights require the support of the relationships that support individuals [64]. Laufer-Ukeles holds that there are three factors that are proven to be advantageous to a child’s development: material support, a low-conflict environment, and stable relationships [49]. She argues that only serious challenge to the first two of these needs warrants disruption of the third. Buss argues that parental rights should, as a matter of Constitutional right, be terminated only to terminate or alleviate serious harm [43].

We can now formulate a framework for government action, which we present in the form of a test. We believe it is ethically coherent, and that it reduces indeterminacy to a degree close to a minimum compatible with the exercise of reasonable judgment. It is compatible with both ethical and legal decision-making. It grounds decisions in a political context, allowing society to protect itself against harm, while precluding discrimination against minorities. It allows for a zone of parental discretion, while precluding parental actions that are likely to cause serious harm immediately or in the future.

The test

If parents are default decision-makers, and if previous approaches are either inadequate or incomplete, then how do we determine when the state may override their decisions? We propose a multipronged test to this end. The test is modified from previous iterations [46, 65, 66]. The test follows:

• A state may limit a religious practice that jeopardizes a child’s health or safety if any of the following two bases exist, but only if none of the constraints in the second prong applies: First prong: Bases for Limiting Religious Practices. Limitation of the religious practice is warranted if the practice has sufficiently adverse effect outside the minority group that it creates unreasonable burdens either for:

• Basis 1: Indirect Effects. Society as a whole or members of society outside the minority group.

• Basis 2: Direct Effects. Children belonging to the minority group, by:

• Creating a substantial chance of death or of major disruption of a physiological function, or Creating other objectively severe harmful effects, such as malnutrition or major psychological morbidity.

• Second prong: Constraints on government action: Notwithstanding the Bases in the First Prong, a state should restrict parental action only if none of the following constraints apply:

• First constraint: Likelihood of effect. The harm underlying the restriction must be Actual, rather than hypothetical; and Likely, rather than unlikely or rare

• Second constraint: Comparability of Effect. The harm underlying the restriction must be of a magnitude greater than harms typically tolerated for comparable mainstream practices.

• Third constraint: Benefit/harm. Benefits of the restriction to all parties concerned should foreseeably exceed overall harms.

The test in perspective

The First Prong is straightforward. The Indirect Basis has two main applications in health care. One is that religious practice may not create a threat to the community of serious communicable disease. For example, rejection of polio or diphtheria vaccination should be unacceptable because of the relative severity and communicability of these diseases. The other application involves parental decisions that impose heavy resource burdens on the health care system, as would be the case with insistence on expensive futile care at state expense. This issue can be resolved through application of the Compatibility constraint. In this example, care choices based on religious belief should be met if their cost-effectiveness or cost utility is no less than those of comparable choices available to those outside the religion. Conversely, choices should be denied if they are less cost-effective than comparable choices allowed in other contexts.

Use of the terms ‘unreasonable’ and ‘substantial’ allows some variability based on local mores and conditions. For example, mild corporal punishment based on Proverbs 13:24 (“Spare the rod and spoil the child”) might be allowed as a minority religious practice in the United States under the Direct Effects Basis, but prohibited in Sweden. However, the vulvar nick in infancy does not seem to reach the level of concern set by the Direct Effects Basis. This Basis also allows for changes in how a practice is treated based on new understandings. If it transpired, for example, that mild corporal punishment frequently and unequivocally caused significant psychopathology, then it would be difficult to allow religious belief to justify this practice.

We now turn to the three Constraints. The Likelihood Constraint precludes states from interfering with a procedure because of hypothetical or rare considerations. It is a partial check on the precautionary principle, at least as applied to minority religious practices. This Constraint assigns the state the burden to prove that the practice is dangerous, rather than on the religion to prove that it is safe. If a procedure is known to have adverse effects, rarity becomes a term of art that is related to the magnitude of the consequence. Thus, a 0.1% chance of hospitalization is rare, while a 0.1% chance of death is not.

The Comparability Constraint applies the principle of the Lukumi decision of the United States Supreme Court [37] prohibiting pretextual use of secular measures to discriminate against religious conduct to bioethical analysis. We should not impose restrictions on others, including religious groups, that we would not impose on ourselves, on those with different religious beliefs, or on broader society.

Finally, the Harm/Benefit constraint requires that action against a practice acknowledged to be bad does not foreseeably make matters worse [68]. Punitive measures use resources to institute and enforce. They cause harm to those punished, and potentially to members of the perpetrators’ families as well, including victims of the act for which punishment is inflicted. They can breed disrespect for society, leading to evasion of law, or even to violent rebellion. Many would consider it ethical to remove the children from the custody of a cult that beat children severely, did not teach them reading or arithmetic, and did not have facilities for sewage in its compound [69]. However, if the likely only alternative is the violent death of many children, as occurred in the 1993 action against the Waco Branch Davidians [70], then the state should consider tolerating the status quo, bad though it may be.

Conclusions

In summary, our test provides an improved approach to considering disputes between a secular state and religious parents. The test’s threshold of action is broader than the harm principle allows, while not as intrusive as a literal construction of the best interest’s principle. Our test goes further than other approaches in several key aspects. It centers on the needs and interests of the child at hand, while respecting the interests of other relevant parties. It was formulated to be applicable to both legal and clinical ethics practice. It safeguards against using child welfare as a pretext for discrimination. It is compatible with the culture and decision-making apparatus of most liberal democracies. Although some indeterminacy will be present in all proposals, ours is more specific than most, and should allow a high degree of consistency in decision-making. Our proposed test offers a fair and consistent basis on which to balance parental and religious rights with governmental obligations to protect the health of children.