PARENTAL DUTIES OF NON-DISCRIMINATION AND THE SCOPE OF ANTI-DISCRIMINATION LAW

Abstract

Parents’ discrimination against their children is lawful. But the family, as an institution in which social goods are allocated, is as significant as the sites in which anti-discrimination law operates. At least prima facie, therefore, parents should be governed by legal prohibitions on discrimination. While state incursion into family life poses a threat to children’s autonomy, so does parental discrimination against children. Anti-discrimination law therefore needs new institutions to promote the values of non-discrimination in a part of society that currently sits outside anti-discrimination law’s reach. We identify existing regimes that may provide a starting point for this work.

I. Introduction

Under the Equality Act 2010, discrimination is unlawful only if it occurs in certain contexts, or “spheres of activity”, which are stipulated by the Act. These are the exercise of public functions and provision of services to the public; ¹ the disposal and management of premises; ² “work”; ³ the provision of education by schools and higher education institutions; ⁴ and “associations”. ⁵ It follows that discrimination that occurs in the “familial” sphere will not be unlawful. Hence, if over the course of a son’s upbringing, his parents provide him with all sorts of educational opportunities, while providing their daughter – because she is their daughter – with the bare legally required minimum, the parents may well be acting immorally ⁶ but, as far as the Equality Act is concerned, will not be acting unlawfully. ⁷ The same will be true if the parents of two adopted children, one Black and one white, engage in such discrimination, favouring their white over their Black child because of a racist belief about the attainments that are possible for the latter compared to the former. ⁸ Finally, this will also be the case if the parents of two children, one who is with disability, and hence uses a wheelchair, and one who is without disability, provide the latter child with an extensive and enriching range of extra-curricular activities but do not make the effort to provide the former with similar opportunities.

This article’s principal positive contention is that the family, as a social institution, and as a site in which opportunities are made available and social goods allocated, is no less significant than the sites in which anti-discrimination law currently operates. That anti-discrimination law does not respond to examples of the sort outlined in the previous paragraph is, therefore, curious and, hence, worthy of further consideration. After all, there is no general principle that exempts the parent-child relationship from legal duties not to wrongfully set back the interests of others. For instance, children who are victims of intentional torts at the hands of their parents may proceed against their parents for that wrongdoing. ⁹ And, indeed, parents find themselves subject to a suite of special duties, stemming from both the common law and legislation, that require them not to set back, and to advance, their children’s interests. ¹⁰

In Section II, we argue that the reasons that support the operation of legal prohibitions on discrimination in general support the application of such prohibitions in the context of children vis-à-vis their parents. Our argument is a theoretical one, drawing on theories of anti-discrimination law and of social justice more generally. We pay particular regard to the existing spheres of activity in which anti-discrimination law operates and argue that the principles that underpin these spheres extend to the parent-child relationship.

This may seem to imply, at least as a matter of principle, that parents should be subject to legal prohibitions on discrimination in respect of their children. In the remainder of the article, however, we subject this prima facie conclusion to scrutiny.

To the extent that the literature on anti-discrimination law has considered the possibility of imposing duties of non-discrimination in familial or other interpersonal contexts, ¹¹ it has typically objected to such imposition. These objections principally take the form of theoretical arguments about the nature and implications of autonomy in the familial context: if the family is a site of, and indeed a vehicle for, exercising autonomy by forming and enjoying interpersonal relationships, then the imposition onto such relationships of legal prohibitions on discrimination may seem to threaten that autonomy. In Sections III and IV, we consider and rebut these arguments.

In Section III, we argue that the value of parental autonomy does not justify relieving parents’ treatment of their children from the anti-discrimination norms that prima facie ought to apply. The situation with respect to children’s autonomy is considerably more complex, because most children are extremely dependent on their family, and their parents, to enable them to develop and exercise their own autonomy. With this in mind, we show that the real challenges to the imposition of a parental duty of non-discrimination are practical ones, of giving effect to non-discrimination norms in a way that is likely to further, rather than to set back, the interests of children. In Section IV, we sketch some possible approaches to institutional design that might permit anti-discrimination law to advance children’s autonomy interests when these are threatened by parental discrimination, while still enabling children to participate in, and thus to receive the goods that flow from, family life.

Our argument is concerned with children who are in the legal and practical custody of their parents or guardians in a relatively conventional fashion, and we refer to these parents and guardians as “parents”. While we do not doubt that anti-discrimination norms may be very important in the context of children in care, in detention, or who are homeless, we do not address those circumstances in this article. Nor in this article do we argue for any extension of the protected characteristics, and so (for instance) we do not consider discrimination against stepchildren on the basis that they are stepchildren, nor discrimination against children from a former relationship vis-à-vis children from the current relationship in virtue of that feature of a child’s relationship to the child’s parent.

Finally, we wish to state unequivocally that it is no part of our argument to suggest that parents from any particular sector of society are more likely than others to discriminate against their children on the basis of protected characteristics, such as sex, race and disability. In Section III, we do address the implications, for both parents’ and children’s autonomy, of the cultural diversity found in contemporary British society. This diversity is an important consideration when contemplating the imposition by the law of normative standards on the parent-child relationship. But we reiterate that our argument makes no claim, or conjecture, about which parents, if any, may tend to discriminate against their children in potentially objectionable ways.

II. Prima Facie Argument

In this section, we argue that the reasons that support the operation of legal prohibitions on discrimination in general support the application of such prohibitions in the context of children vis-à-vis their parents. Our argument is thus a type of internal criticism and an argument for the principled extension of the existing law.

Legal prohibitions on discrimination operate in several important spheres of human activity. What characterises these spheres is that they are, or at least ought to be, key sites of access to certain basic social goods. This makes the spheres particularly significant in the imposition and perpetuation of disadvantage: certain individuals, as a result of characteristics they possess, are confronted by barriers to such access that arise in the spheres as a result of discriminatory decision-making. Those who have the capacity to erect these barriers are, as Tarunabh Khaitan puts it, “[g]atekeepers of [o]pportunities”. ¹² By prohibiting discriminatory decision-making, discrimination law seeks to eliminate what might otherwise be very significant barriers to access to basic social goods.

Each of the elements we have identified – protected characteristics, basic goods and spheres of human activity – is open to differences of analysis. We briefly sketch some of these and, in so doing, demonstrate how those different analyses nevertheless support our claim that legal prohibitions on discrimination should apply to parents in respect of their children. Our analysis does not have revisionist implications for protected characteristics and basic goods. It does, however, criticise the current failure of the law to include the parent-child relationship as one of the spheres in which discrimination law operates. We show that the omission to which we refer is at least prima facie unprincipled.

A. Protected Characteristics

We take it as obvious that, in many contemporary societies, there are certain groups, the members of which experience particular sorts of barriers to their access to key social goods: people of colour, girls and women, people with disability, etc. The significance of membership of these disadvantaged groups as an important factor in access to key social goods is reflected in the protected characteristics that are central to anti-discrimination law. ¹³

There are accounts of anti-discrimination norms that do not expressly focus on an individual’s membership of one or more disadvantaged groups, but rather emphasise appropriateness of consideration as underlying the wrongfulness of discrimination. These accounts hold that it is inappropriate or wrong to have regard to (inter alia) a person’s race, or sex, or sexuality, or gender, or gender identity, or stereotypes related to those characteristics, ¹⁴ in making a decision that affects their access to key social goods. This is because doing so fails to accord the respect owed to them as a person; ¹⁵ or relies upon a consideration irrelevant to the decision being made; ¹⁶ or results in the decision being made on a basis other than its merits; ¹⁷ or results in the decision affecting a person being made “without rational justification”; ¹⁸ or is otherwise objectionable. ¹⁹ Some accounts of this sort can be understood as building on the notion of historically disadvantaged group membership and as seeking to identify what the wrongful treatment of members of various groups, by different perpetrators, has tended to have in common; others downplay, or even deny, the relevance of membership of disadvantaged groups as a basis for protection against discrimination. ²⁰

For the purposes of this article, we do not need to choose between these various accounts. It is sufficient to point out that the family is a context in which a person – and, given the concerns of this article, a child – may suffer discrimination on the basis of one or more protected characteristics. For instance, a daughter may be treated by her parent less favourably than her brother because of her sex; or a child who is not white, adopted by a white parent or parents, may be treated less favourably than another of their children in virtue of the child’s race; or a child who is gay may, for that reason, be treated less favourably by the child’s parent than a sibling who is heterosexual.

B. Goods

Turning now to the basic social goods with which anti-discrimination law is concerned, again there is a range of accounts. Equality of access to social opportunities is one possible good. Khaitan argues strongly against this, however, on the basis that anti-discrimination law as actually practised, with its focus on disadvantaged groups rather than disadvantaged individuals irrespective of the groups to which they belong, is not concerned with ensuring equality of opportunity as such: even if anti-discrimination law is successful in removing group inequality, individual inequality will likely remain. ²¹ Khaitan suggests that the basic goods in question are negative freedom, opportunities and self-respect, to the extent that historical patterns of oppression and disadvantage on the basis of group membership inhibit access to them. ²² Sophia Moreau identifies a different, but overlapping, bundle of goods: ²³ deliberative freedom (itself underpinned by the value of autonomy); ²⁴ freedom from subordination; ²⁵ and those resources and opportunities necessary to participate fully and equally in society and to be seen as such. ²⁶ Deborah Hellman identifies the key good as not being treated by others, who have the social power to put one down (e.g. because of their institutional status), in a way that expresses a strong lack of respect for one’s equal moral worth. ²⁷

Again, in this article, we do not adopt any particular account of the goods in question. We simply note that it is possible for decisions made by parents in respect of their children to impede their children’s access to goods of the sort we have canvassed, including self-respect, deliberative freedoms, freedom from being put down in the way Hellman describes, negative freedoms and a very wide range of opportunities. Innumerable examples might be given. For instance, consider a parent’s decision to forbid the parent’s daughter – but not the parent’s son – from participating in a sporting activity, telling her that the sport – participation in which is highly valued in the society in which they live – is not one that is suitable for girls (e.g. because it is a contact sport and therefore, in the parent’s view, not sufficiently feminine). This decision clearly has the capacity to undermine the daughter’s self-respect; it teaches her that she must consider her own gender in making requests and seeking out opportunities, and thereby burdens her deliberative freedoms; it may well put her down in the way Hellman is concerned about, given the power enjoyed by the parent, the content of the decision, and the manner in which it is conveyed; it burdens the daughter’s negative freedoms, as she is not free to take part in the activity; and it likewise limits her opportunities.

C. Spheres of Activity

Anti-discrimination law operates in respect of certain spheres of human activity, paradigmatically employment, the provision of goods and services, and the provision of education. As noted above, and as we go on to explain in more detail, what unifies the spheres is that they are key sources of access to basic goods. This is why discriminatory decision-making within these spheres can act as a barrier to access to those goods.

Some accounts of anti-discrimination norms do not engage at all with the confinement of anti-discrimination law to particular spheres. ²⁸ When consideration is given to this aspect of the scope of discrimination law, it is often (and usually only in passing) said to be the public realm. ²⁹ But this use of “public” risks being a purely conclusory label. ³⁰ After all, there are straightforward senses in which employment, and the provision of goods and services, may be described as private activities. ³¹

An alternative account of the spheres and their limits is suggested by Hugh Collins. He contends that, in order to fall within the one of the spheres, a relationship must normally be contractual – or at least “expected to be … contractual … if it comes to fruition”. ³² Collins acknowledges that, since the provision of services to the public voluntarily (as opposed to in exchange for payment) is subject to legal prohibitions on discrimination, the “boundary provided by the law of contract is … not completely reliable”. ³³ Even with this concession, however, Collins greatly overstates the extent to which a relationship between two parties must be contractual – or must be able to become contractual – in order for it to fall within the spheres as they function in English law. First, section 29(6) of the Equality Act prohibits a person from discriminating in the exercise of a public function. While functions exercised pursuant to contractual power can, under certain circumstances, be public (at least for the purposes of judicial review), ³⁴ it remains the case that, overwhelmingly, public functions in England are carried out pursuant to statute, not contract. ³⁵ Second, and perhaps even more significantly, one of the stipulated spheres in the Equality Act pertains to the provision of education, by both schools ³⁶ and universities (and other tertiary institutions). ³⁷ While the legal relationship between university students and the university they attend is “essentially contractual”, ³⁸ there is no legally binding contractual relationship involved in children’s attendance at state maintained schools. And, even in the context of privately funded schools, the relevant contractual relationship is between the school and the parents or guardians of the child who attends the school, not the actual child. But it is the child who will normally be the applicant for any claim in discrimination against the school.

Although we have criticised Collins’s account, we do think that it points in the right direction. To see how this is so, consider a further feature of discrimination law, namely that it is unidirectional in its operation: employers, providers of goods and services, educators, etc. owe a duty not to discriminate in making decisions within their roles, but those who are subject to those decisions do not owe duties to the providers. Collins explains this by reference to power in these relationships, as does Khaitan. ³⁹

What this points us to, therefore, is that the spheres in which anti-discrimination law operates are sites in which, in a complex economy and society like that of the contemporary UK, power is exercised and decisions are made which, by and large, distribute the basic social goods with which discrimination law is concerned.

One well-known term for describing social structures and systems that operate so as to distribute access to basic social goods is the basic structure. ⁴⁰ This term, which originates in John Rawls’s liberal theory of justice, is not commonly used in the literature on anti-discrimination law. ⁴¹ However, there is a high degree of similarity between the basic social goods that are the focus of anti-discrimination law and (what Rawls calls) primary social goods, that is, “all-purpose means” necessary for members of a liberal society, whatever their particular conception of the good, “to make effective use of their freedoms”. ⁴² The primary goods that Rawls identifies are: “basic rights and liberties”; “freedom of movement”; “free choice of occupation against a background of diverse opportunities” and “protected by fair equality of opportunity”; “powers and prerogatives of offices and positions of responsibility in the political and economic institutions of the basic structure”; “income and wealth”; and “the social bases of self-respect”, that is, “those aspects of basic institutions normally essential if citizens are to have a lively sense of their worth as persons and to be able to advance their aims and ends with self-confidence”. ⁴³

Rawls’s identification of the basic structure as the “the primary subject of justice” ⁴⁴ prompted an extensive discussion of whether or not the family is a component of the basic structure. A key reason for thinking that it is not is that it is an intimate association and thus, in some sense, a “private” matter. A key reason for thinking that it is is that it makes a fundamental contribution to the generation and distribution of basic goods, particularly those relating to opportunities (including material opportunities) and self-respect. Susan Moller Okin argues strongly, on the basis of this second reason, that an account of justice focused on the basic structure must include the family. ⁴⁵ G.A. Cohen also points out that informal structures, such as expectations of gender roles that operate within the family, as much as more formal ones enforced by coercive law, may profoundly affect the distribution of opportunities. ⁴⁶

In A Theory of Justice, Rawls identifies the family as a source of opportunities ⁴⁷ but does not consider the extent to which family life should itself be regulated by principles of justice. ⁴⁸ In later work, Rawls identifies “the nature of the family” as a component of the basic structure ⁴⁹ but argues that the family itself is more analogous to a private association (such as a club or a church) and hence may be “governed” internally in ways that do not conform to liberal principles. ⁵⁰ He recognises that family members may need protection from one another: “wives from their husbands, children from their parents.” ⁵¹ However, he appears to accept that it is sufficient, in this respect, that women be able to exercise equal liberties without the threat of financial detriment (e.g. by way of divorce laws that have regard to domestic labour in the apportioning of family assets) ⁵² and that children benefit from “equal opportunities of education for all … designed to even out class barriers”, ⁵³ which “include such things as knowledge of their constitutional and civic rights” and “encourage the political virtues so that they want to honor the fair terms of social cooperation in their relations with the rest of society”. ⁵⁴

Okin criticises the sufficiency of such protections:

Up to the age of four or five, however, most children spend a very large part of their waking lives at home with their families. … Only a comparatively small proportion of children spend most of their waking time up to this age in any setting other than a family. If these families are, as I have argued, frequently not environments in which justice is normally practiced, work equally shared, and people treated with equal dignity and respect – if they are often instead places where injustice ranges from moderate unfairness to outright abuse, then how are children to develop the sense of justice…? ⁵⁵

These remarks bring out the significant difference, for most children, between their membership of a family and an adult’s membership of a club or a church. We also suggest that Okin’s remarks remain true even for older children: while older children do spend more time outside the home (typically at school), the home continues to be a, if not the, primary influence on the child’s development of a sense of one’s own self-worth and the way in which one should engage with the world. ⁵⁶

Okin asks a further rhetorical question:

Suppose that all of these “nonpolitical” settings inculcate and reinforce in [a boy and a girl] the belief that there is a natural … hierarchy of the sexes, each with its own proper sphere – the female’s being narrow, circumscribed, and without authority, and the male’s the opposite. How is such a socialization consistent with both children’s becoming, in any sense, “free and equal citizens” – who, as Rawls says, must “regard themselves as self-authenticating sources of valid claims”? ⁵⁷

While Okin’s discussion is particularly focused on “a traditionalist (fundamentalist or orthodox) religious household” whose children are “educated entirely at religious schools and within the church, temple, or mosque”, ⁵⁸ the answers to her rhetorical questions would seem to generalise.

In this article, we do not seek to resolve this dispute about what falls within the basic structure and hence the scope of Rawlsian principles of justice. Our point, rather, is that the same reasons that speak in favour of the family being considered part of the basic structure identify discrimination within the family as a potential barrier to access to basic social goods. In this respect, the family seems indistinguishable from the spheres of activity in which anti-discrimination law presently operates to remove such barriers – hence why we argue that the reasons that support the operation of legal prohibitions on discrimination in general support the application of such prohibitions in the context of children vis-à-vis their parents.

Are there countervailing reasons that would defeat this argument that legal prohibitions on discrimination should be extended to include children vis-à-vis their parents? In the next section, we address the argument, grounded in the autonomy of family members, that anti-discrimination law should not operate in the family context.

III. The Objection Based on Autonomy

As we noted in the introduction, the principal objection to parental duties of non-discrimination that is found in the literature is a theoretical, autonomy-based one. We explain the general character of this objection by beginning with Khaitan. That is not to say that Khaitan himself advances this objection – as we go on to explain, it is not entirely clear where he stands on the issue of autonomy in its application to children. However, he provides a useful framework within which the autonomy objection can be set out.

Khaitan advances a second criterion, in addition to his “gatekeeper” criterion discussed above, for identifying the scope of anti-discrimination law: it should apply to “public” actors who, in virtue of being public, do not have a strong (or perhaps any) interest in negative liberty. ⁵⁹ These two criteria operate cumulatively in the following way: a lesser interest in negative liberty reduces any objection to imposing a duty of non-discrimination, which by its very nature reduces the scope for free action; and the greater the extent to which an actor is a (potential) barrier to access to basic social goods, the lower the burden required to impose a duty that would restrict the actor’s freedom of action. Conversely, in cases where an actor enjoys an undefeated interest in negative liberty, that actor is to be “permitted to engage in unjustified discrimination”. ⁶⁰

Khaitan identifies three ways in which the criterion of being “public” may be satisfied: the state is public ipso facto; ⁶¹ actors who hold themselves out as offering services to the public in general are public; ⁶² and employers, in virtue of the tremendous power they exercise over employees and job-seekers, are public. ⁶³ Khaitan comes close to accepting that the family is also “public” in his sense. ⁶⁴ Nonetheless, he also seems to suggest that family members are entitled to negative liberty, as he contrasts “[b]usinesses” and “monetized … interaction” with “families, friendships, or (possibly) clubs” ⁶⁵ and says that “the law tends not to regulate discrimination in deeply intimate and personal spheres”, which include “friendships and romantic relationships”. ⁶⁶ He further states that “I ought to be permitted” to make discriminatory decisions about who I make friends with, “given the strong autonomy interest I have in freely determining who I should be friends with”, ⁶⁷ and when he lists those actors who are subject to the anti-discrimination duty, he does not include parents or the family. ⁶⁸

Moreau recognises that “personal decisions” made “in our private lives” may generate the sorts of “effects on the power, authority, and freedoms enjoyed by others” that are the concern of anti-discrimination law. However, and more unequivocally than Khaitan, she “den[ies] that it would be a good thing for anti-discrimination law to apply to these personal decisions”. ⁶⁹

Building upon these (and other similar) objections, we set out the following argument against the imposition of duties of non-discrimination in the familial context:

Typically, interpersonal relationships are valuable because (inter alia) they:

(a) are exercises, and expressions, of their participants’ autonomy; ⁷⁰ and/or

(b) enhance the autonomy of their participants by increasing the range of valuable options to those participants and by facilitating their self-realisation. ⁷¹

Therefore,

(c) Individuals enjoy a negative liberty interest in making decisions in relation to their interpersonal relationships, even when these are discriminatory decisions, which interest therefore ought not to be burdened by restrictions of the sort that anti-discrimination law would impose. ⁷²

For the purposes of this article, we do not contest (a) and (b), and hence the inference to (c), with respect to adults. ⁷³ However, we do argue that, in the case of children, (a) and (b) are not true, or at least not to the same extent as for adults, and, moreover, to the extent that they are true for children, they are true in different ways from how they are for adults. Furthermore, if anti-discrimination law were to protect children against their parents’ discriminatory conduct, it would not burden the negative liberty of children. Rather, we argue, that such law would have the potential to enhance children’s autonomy. Therefore, we argue, the correct understanding of the situation is one of a conflict, or at least tension, between (1) parents’ negative liberty interests, grounded in their autonomy, and (2) children’s interests in developing and exercising autonomy. In what follows, we set out our arguments concerning (a) and (b) as they pertain to children. We then move on to explain why, in the conflict just identified, children’s autonomy interests typically outweigh parents’ interests in negative liberty. The consequence, we argue, is that our prima facie argument is vindicated.

A. The Place of Children in the Family

Let us commence with (a) above. Unlike intimate partners, who give effect to their autonomy by forming interpersonal relationships, children almost never choose to be in the familial relationship. They simply find themselves in the custody of their parents (or guardians). Moreover, children are (typically) not able to leave their family, at least until they turn 16; ⁷⁴ and, if they do, they are (typically) not able to join (or start) a new family. This contrasts with adults’ partnerships. ⁷⁵ Even where these are difficult to end, ⁷⁶ the difficulty is usually not as extreme as that typically faced by a child who may wish to leave the family relationship. And, if a partnership is ended, the person who ended it may well be able to form a new partnership (either promptly or in due course).

Furthermore, and unlike adults, who usually have at least some say over the form that their families take, children have little, if any. Moreau takes this for granted when she states that:

[S]pouses have the chance to choose each other and to choose, together, the kind of life they are going to live and the kind of family they want to create for their child, just as friends have a chance to choose each other and choose the kind of friendship they want to have. … [T]hese relationships are valuable insofar as they reflect the shared desires and aspirations, and the free choices, of spouses and friends. ⁷⁷

While Moreau does observe that actions taken in the “private or personal realm”, including “decisions about how to raise our children [and] whom to have as friends”, may “perpetuat[e] stereotypes” and “have significant effects on the power, authority, and freedoms enjoyed by others”, ⁷⁸ this reinforces her characterisation of the family as a vehicle for the expression of autonomy by the adults whose choices constitute it. She does not consider the fact that, for children, there is little or no choice.

John Gardner does discuss the entry of children into foster families but not in a manner that centres on the autonomy of the child: he refers to foster-parents “selecting” their children, and his proposed solution to the risk of discrimination by foster-parents is that “astute local authorities will supply the sanction here by taking offending foster parents off their list”. ⁷⁹

Khaitan appears to treat friendship, the family and intimate relationships as all belonging to the same analytical category ⁸⁰ and thus seems to adopt, at least implicitly, an adult-oriented account of family life. He does not mention children at all. Given that children typically do not enter romantic relationships and, as noted above, typically do not choose their family, his remarks seem not to extend to children and hence not to the parent-child relationship. (Indeed, given the uncertainty about Khaitan’s position as to the degree of negative liberty that ought to be enjoyed by family members, our argument against the autonomy objection may be one that he is able to agree with.)

The way each of these authors discusses the family is thus consistent with Martha Fineman’s observation that, commonly, children “tend[] to disappear as an independent focus in discussions about rights”. ⁸¹ According to Fineman, this is in turn, at least in part, because, in “most cases, the family is presumed to function appropriately”, with the result that children “can conveniently be ignored”. ⁸²

Because children do not choose or shape their families in the way adults do, they do not benefit as adults do from the domain of interpersonal relationships being unregulated by anti-discrimination law. This is the first strand in our argument that, when it comes to the regulation of familial life, there is a tension between parents’ interests and children’s interests.

B. Children and the Other Spheres

Let us now turn to (b), the proposition that personal relationships enhance the autonomy of their participants by increasing their range of valuable options and by facilitating their self-realisation. In the case of adults, this enhancement results from the way in which relationships are cultivated – which includes the mutuality between participants that is often a hallmark of relationships among adults – and the fact that personal relationships are typically just one part of a broader life plan. When considering whether (b) is true for children, we need to keep in mind the overwhelming role played by the family in children’s lives and also the fact that children are subject, both de jure and de facto, to a very high degree of direction from authorities (their parents) over whom they have little or no control. ⁸³

Many children live in families that are very supportive, with parents who provide ample opportunity for their children to exercise and develop their autonomy and who assist their children’s development of autonomy in all sorts of ways. Indeed, this is one crucial function performed by the family as an institution. But not all families are supportive in this way. Some children may suffer discrimination at the hands of their parents, including discrimination that impedes their autonomy and/or its development. Examples might include parents providing, or permitting, different opportunities to their children within the family on the basis of protected characteristics (for instance, by insisting that some do housework while others are permitted to learn or play); or parents fostering the self-respect of different children to different degrees, again by reference to protected characteristics (for instance, by presenting these different opportunities as reflecting a “natural” hierarchy among the children).

Gardner is alert to the possibility of autonomy-burdening relationships. ⁸⁴ In his discussion of these, he refers to both “man-woman and parent-child relationships”, ⁸⁵ but as we will explain – and consistently with Fineman’s remarks quoted above – his focus is on relationships between adult men and women.

Gardner objects to the extension of anti-discrimination law into the familial context because of the overall risk to autonomy that would result from regulating the formation and conduct of these relationships. This worry is a somewhat consequentialist one, expressing a type of “on balance” judgment about the nature of interpersonal relationships and the effect that regulation might have on them: Gardner argues that subjecting intimate relationships to legal prohibitions on discrimination would generate a very large burden on autonomy across society, by limiting people’s ability to enjoy the autonomy-enhancing benefits of relationships that derive from typical cases where there is no net burden on autonomy. He doubts that the social forms that permit the autonomy enhancement in the typical cases would survive the introduction of regulation motivated by concern for the autonomy-burdening outliers. ⁸⁶

However, Gardner does think that anti-discrimination law has a role to play in addressing the burden on autonomy that may arise from discriminatory traditions and structures in interpersonal relationships: anti-discrimination law should, and does, regulate other societal spheres so that burdened members of relationships become more able to participate in those other spheres. The example he gives is sex discrimination law, which does not directly address patriarchal elements in intimate relationships, but which, he says, enables women to participate more freely in other social spheres, such as employment. ⁸⁷ Sandra Fredman makes a similar observation, although in a more pessimistic tone. She notes that, while anti-discrimination law significantly helps women, who are the majority of part-time workers, to engage in part-time work without being disadvantaged for it, anti-discrimination law does nothing to change the structure of the family and hence the very fact that women comprise the majority of part-time workers to begin with. ⁸⁸

Fredman’s observation drives home the way in which Gardner’s argument presupposes that adult women can take advantage of the opportunities, in alternative spheres, that anti-discrimination law makes available. But that will not always be the case. For instance, if, in the face of prohibitively expensive child-care, it is the woman who finds herself looking after her children, rather than her male partner, she will be restrained in the extent to which she can make use of the opportunities in the spheres of employment or education. Similarly, Gardner does not consider the distinct circumstances of children, who cannot avail themselves of opportunities in spheres of endeavour outside the family in the way that adult women may be able to.

To begin, children are extremely limited in their capacity to undertake paid work. Hence, subject to local authority byelaws, children who are not yet 14 years old may not be employed at all. ⁸⁹ Children over that age, but younger than the school leaving age (i.e. 16 ⁹⁰ ), may only be employed subject to very significant constraints, pertaining to, inter alia, the type of work that they may undertake, ⁹¹ the number of hours of paid employment in which they may engage on a particular day, ⁹² and in a particular week, ⁹³ and the particular times of the day ⁹⁴ in which they may engage in employed work. For this reason, as well as other practical and legal reasons, ⁹⁵ children are, in turn, usually able to avail themselves of goods and services only with the assistance (legal, financial and/or practical) of their parents. And although education is compulsory, the legal regime for the provision of education treats parents as the duty-bearers in respect of, and hence facilitators of, children’s education. ⁹⁶

This limited and conditional nature of children’s access to these alternative spheres of opportunity, like the risk they face of parental discrimination within the family that we noted above, is a correlative of the overwhelming role played by the family in their lives and the control that their parents exercise over them. In formal legal terms, laws that limit children’s access to these spheres, and that confer their parents with authority over them, ⁹⁷ are major causes of the significance of the family in children’s lives. Considered through a broader social lens, the law that governs children reflects the social realities of family life and what is normally the place of children in the family.

Furthermore, and as noted above, children are (typically) not able to leave their family and, if they do, they are (typically) not able to join (or start) a new family. All of these differences mean that the effect of the family on children’s autonomy is usually more salient than it would normally be for the autonomy of (for instance) an adult woman.

For these reasons, to the extent that (b) above is true for children, it is true because of the point we made earlier: many children live in families that are very supportive of them and are facilitative of their autonomy. But, for children, familial relationships are not sites, and generators, of opportunities (including opportunities for self-realisation) in the way that they may be for adults. In those cases where parents do act in a discriminatory way towards their children, anti-discrimination law – were it available in the parent-child context – would offer the prospect of a safeguard for those children’s autonomy interests. This is the second strand in our argument that there is in fact a tension between parents’ and children’s interests in this respect.

C. Weighing the Interests

We have identified a tension between children’s interests in autonomy, which might be protected by legal duties of non-discrimination, and their parents’ interests in being able to pursue familial relationships free from such duties. In this sub-section, we argue that this tension should be resolved in favour of the children’s autonomy interests. The consequence is that the autonomy-based objection to the imposition of legal duties of non-discrimination on parents vis-à-vis their children, that we have set out and analysed above, falls away.

1. Parental discrimination resulting from bigoted attitudes

We begin with two observations. First, some discriminatory decisions made by parents may be nothing more than expressions of bigotry: the parents are simply sexist, or homophobic, for instance. Second, not all apparently autonomous choices are in fact desirable and defensible exercises of autonomy. There are different ways of putting this second point, depending on the broader framework of moral reasoning that is adopted. Joseph Raz, for instance, who adopts a liberal perfectionist framework, states that “[a]utonomy is valuable only if exercised in pursuit of the good”. ⁹⁸ Will Kymlicka, who adopts a more classically liberal framework, ⁹⁹ says that the limits of autonomy are to be set by the need to respect the autonomy of others. ¹⁰⁰

We do not think that this second point refutes a general claim to negative liberty in respect of interpersonal relationships, even to the extent of some discrimination in the context of such relationships being permissible. However, as we have shown in our discussion of the different sorts of roles and degrees of power that children and adults exercise within the family, we do think that this second point establishes limits to that claim. Such a claim to negative liberty depends upon an idea of interpersonal relationships as freely chosen, such that victims of discrimination are not stuck in those relationships, forced to endure the discrimination. It rests on an assumption that such victims can exercise their own negative liberty rights to form their own satisfactory and valuable relationships. We accept, for the purposes of this article, that this is consistent with (a) and (b) above being true for adult participants in interpersonal relationships. But, as we have argued above, the situation of children is different. They do not freely choose their familial relationships but simply find themselves within them. Further, typically, they cannot exit the family relationship and do not have straightforward access to the other spheres that Gardner identifies as salient for adult women. Therefore, children are liable to be stuck as victims of discrimination visited upon them by their parents.

In such circumstances, the children’s parents’ behaviour, if it is simply bigoted, is not a valuable exercise of autonomy. It is behaviour that is at least “[d]emeaning, or narrow-minded, or ungenerous, or insensitive”, ¹⁰¹ and may well be “morally bad and repugnant”, ¹⁰² and hence does not satisfy the perfectionist framework’s standards for a valuable exercise of autonomy. Within the classically liberal framework, it would be considered behaviour that burdens the right of the child to develop and (at an appropriate age) to pursue the child’s own perfectly reasonable conception of the good without undue burden on their sense of self-respect and hence would not be the sort of autonomy that that framework affirms. ¹⁰³

As Khaitan argues, ¹⁰⁴ discrimination law does not rely upon refuting any negative liberty interest that the duty-bearer might enjoy. ¹⁰⁵ It is sufficient that a putative duty-bearer’s interest in negative liberty is weak and that the putative duty-bearer plays the “gatekeeper” function described above in Section II. In that section, we showed how parents play the gatekeeper function to a tremendous extent (perhaps more than any particular potential employer or school, for instance). Hence, we conclude that the prima facie argument made in Section II is not rebutted by whatever negative liberty interest bigoted parents may have in making decisions about, and within, the family unconstrained by legal prohibitions on discrimination.

If all cases of discrimination by parents against their children were simply manifestations of bigotry, we could therefore conclude that parental autonomy provides no reason why the state should afford the behaviour legal protection, by exempting it from the scope of anti-discrimination laws, when the underlying concerns of anti-discrimination law dictate that it should be so subjected (as we have argued above). However, there is a different sort of case to which we now turn. Considering this case will lead to a broader consideration (in Section IV) of the complexities that children’s autonomy interests create for the prima facie conclusion that anti-discrimination law should apply in the parent-child context.

2. Parental discrimination resulting from cultural beliefs and practices

The concept of culture is a complex one. In this article, we follow Kymlicka’s account. Kymlicka uses the phrase “societal culture” to refer to “a culture which provides its members with meaningful ways of life across the full range of human activities, including social, educational, religious, recreational, and economic life, encompassing both public and private spheres”. ¹⁰⁶ A societal culture is therefore important to autonomy for two reasons: it provides options, and it provides a framework of meaning within which value is able to be attributed to those options, such that autonomous choice can take place. ¹⁰⁷

With this in mind, we note two important ways in which families enable the children within them to develop and exercise their autonomy. One of these we have already discussed above: children learn to be autonomous under the guidance of their parents, within a framework of opportunities and choices over which their parents exercise a great deal of control. Indeed, as we argued in Section II, a major argument for anti-discrimination law being available in the parent-child context is the capacity for discriminatory conduct by parents towards their children to harm this aspect of their children’s development.

A second way in which a family enables its child members to develop and exercise their autonomy is more subtle. Families – and the parents who lead them – play a vital role in the transmission of societal cultures, thereby enabling children to inhabit a framework of meaning within which they can learn to value and choose among the various options that they encounter.

In the contemporary world, societal cultures tend to be territorially located on the basis of national borders, because it is nation states that have the institutional capacity and social power to shape and support a societal culture, particularly in respect of its educational and linguistic underpinnings. ¹⁰⁸ Within a particular society, however, especially a diverse one with many members who have migrated from, or have roots in, other national societies and thus other societal cultures (Kymlicka refers to these various participants in a societal culture as “ethnic groups” ¹⁰⁹ ), the societal culture will contain diverse elements. ¹¹⁰ This diversity will, in turn, be reflected in families; and it is possible that some of this cultural diversity will extend to parental practices, in relation to children, that are discriminatory. This sort of discrimination creates a situation that is more complicated than discrimination resulting from mere bigotry. It is, at least arguably, a valuable exercise of parental autonomy, insofar as it is an expression of cultural meaning. ¹¹¹

We recognise that the distinction between discrimination resulting from mere bigotry and discrimination resulting from parents’ cultural beliefs and practices is not one that can always be drawn in a straightforward fashion. The key distinction upon which we rely is between discriminatory actions that are not valuable exercises of autonomy (i.e. the merely bigoted) and those that are; and the only examples of which we are aware in that second category are those that result from, and help to support, a culture that is valuable for the reasons we have explained.

This sort of discrimination makes the balance of interests more complicated in two ways. First, it makes it less obvious that the parent’s interest in negative liberty has no weight, as they may have a negative liberty interest in participating in and fostering those elements of the societal culture that are of particular salience to them, given the framework of meaning within which they exercise their autonomy.

Second, children in such a situation have interests that may push in different directions, corresponding to the two different ways in which a family enables its child members to develop and exercise their autonomy. As well as the obvious interest in not suffering discrimination, there is also the child’s interest in being reared in such a fashion as to have access to the societal culture. This depends, to at least some degree, on curation of salient elements of that culture by adults who teach the child both by instruction and by example. Parents are typically the most important such adults, and the home is typically the most important site of such learning. Thus, children have an instrumental interest in their parents enjoying the negative liberty identified in the previous paragraph.

We argue that there is a principled solution to this more complex balancing problem. We begin by explaining why it should be the children’s interests that are paramount in resolving it.

Crucial to liberal accommodation of non-liberal elements within a liberal society’s culture is the so-called “right of exit” enjoyed by individuals who are members of families and other groups in which such elements contribute to the ordering and the dynamics of the group (e.g. who exercises power within the group, what social roles group members occupy, etc.). Because the liberal state does not treat membership of any such group as obligatory, even though the internal group norms may present group membership in that way (e.g. as a moral or a religious requirement), in a liberal society it should always be possible to leave such a group and participate in the societal culture in other ways that do not depend on the salience of non-liberal elements of that culture. ¹¹² The presence of such a right, if effective, would thus afford people with a degree of protection from oppressive and discriminatory behaviour directed at them by other members of their group, because they would be able to leave the group and pursue “an alternative mode of life”. ¹¹³

However, some authors have expressed scepticism about the reality of the “right of exit” in some circumstances. Okin, for instance, observes, inter alia, that the very treatment that might give a woman good reason to leave a family in which she has suffered discrimination may make it very difficult, practically but also psychologically, for her to do so. ¹¹⁴ And whatever the extent to which an adult may be able to rely on a right of exit so as to leave a family, it is extremely unlikely that children will be able to do so, ¹¹⁵ especially if they are unassisted by legal mechanisms to facilitate such departure. This is a consequence of the dominant place of the family in most children’s lives and the limited access that children have to spheres of endeavour beyond the family, as we have shown above. This absence of a right of exit for children means that the parental negative liberty interest, discussed above, does not weigh in the scales after all. This is because the requirements for accommodating it within the commitments of a liberal society, to the extent that if affects children’s access to basic goods and the development of their autonomy, have not been met. Thus, even when parental discrimination against children has a basis in cultural practices, there is no argument from parental autonomy against the conclusion that children should enjoy the protection of anti-discrimination law vis-à-vis parental discrimination.

IV. Giving Effect to Children’s Interests

In the preceding section, we showed that there is no argument based on parental autonomy against the imposition on parents, in respect of their children, of legal prohibitions of non-discrimination. It does not follow, however, that anti-discrimination law should be extended, in a wholesale fashion, to the parent-child relationship.

In this section, we consider the implications of such an extension from the perspective of children’s autonomy interests. We argue that any such extension, if it is to be justifiable and certainly if it is to be effective in enhancing children’s autonomy, must be extremely sensitive to what would be at stake in giving effect to any legal prohibition on parents’ discriminatory behaviour. This raises challenging questions of institutional design, which we identify below.

A. The Need to Weigh a Child’s Competing Interests

Of course, as a general rule, no one is ever obliged to enforce one’s legal rights. But, in the case of a child’s putative legal right not to suffer discrimination by their parents, there is a further, distinct complexity: the bringing of a discrimination complaint by a child against the child’s parent(s) creates an obvious risk of disrupting the family, because of both the resulting state intervention itself and the effect that making such a complaint may have on relations among family members (at the extreme, it may even provoke a parent to pursue reprisals against the child). Such disruption will be likely to undermine the capacity of the family to provide an environment in which the child is able to develop the child’s autonomy, in circumstances where the child has nowhere else to go (due to the absence of an effective right of exit for most children). Thus, in some circumstances, some children who suffer parental discrimination may nevertheless be better off if such parental breaches of a non-discrimination duty are permitted to continue.

In each particular case, it would necessarily be a matter of weighing the child’s interests, as well as the likely impact on those interests of seeking to give effect to a parental duty of non-discrimination. On one side of the scale is the child’s interest in not being the victim of such discrimination; on the other side is the child’s interest in enjoying the support of the family in the development of the child’s autonomy, which, in particular cases, will include an interest in being provided with access to the child’s societal culture. All of these interests are related to the child’s autonomy, and we might therefore expect that in many cases they will be commensurable. In such cases, the particular balance of interests either will, or will not, support some sort of intervention by anti-discrimination law in a particular case. In other cases, it may be that, for some reason, the child’s interests are unable to be weighed against one another and that a decision must therefore be made about which interest should prevail.

When adults are concerned, we accept that they are able to identify which of their conflicting interests is the weightier or – where their interests are not able to be weighed – that they are at liberty to choose between them, even where this creates a risk to other important interests of theirs (such as in interpersonal relationships). In the case of children, however, the situation is different. Some children, especially those who are closer to adulthood, may be sufficiently mature that they are able to make the relevant decisions, including acceptance of the associated risks (as discussed, in a medical context, in the well-known Gillick case). ¹¹⁶ Other children may need support in their decision-making or may even need an adult to make the decision on their behalf. ¹¹⁷

It may be that even opening up the possibility of enforcement, by a child, of a right not to be subject to parental discrimination would create an undue risk to children’s interests, resulting from the threat of a heavy-handed incursion by the state into family life. Hellman expresses such a concern, namely that the costs of extending anti-discrimination law in ways that principle seems to demand may outweigh the benefits of doing so. ¹¹⁸ This may be less of a concern for some older teenagers, who are not only more likely to be capable of making decisions in their own interests but are also more likely to be able to endure the consequences of a significant disruption to their family life. ¹¹⁹

Moreover, the effect on a child’s interests, of giving effect to a parental duty of non-discrimination, would depend to a great degree upon the particular mechanism adopted by anti-discrimination law. The argument of this article therefore concludes by briefly considering what form such a mechanism or mechanisms might appropriately take.

B. Possible Mechanisms for Giving Effect to a Parental Duty of Non-Discrimination

Giving practical effect to the principle that legal duties of non-discrimination should apply to parents vis-à-vis their children gives rise to several important questions of institutional design, concerning both when and how complaints of discrimination might be brought against parents and what sorts of responses to those complaints might be appropriate.

The Equality Act establishes a regime of enforcement by way of proceedings in civil courts or employment tribunals. ¹²⁰ A court order legally obliging a parent to refrain from discriminating in a particular way against the parent’s child risks being very disruptive of the family and hence significantly undermining the child’s own autonomy interests. This need not be so in all cases, however. Consider a situation in which one parent is assisting the child in the action against another parent, perhaps in the context of a broader dispute among the parents. The availability of an order responding to past discrimination by one of the parents may make the child better off, by regulating the manner in which that previously discriminating parent treats the child in the future.

Furthermore, the mere possibility of enforcement of a non-discrimination duty may be relevant to the informal resolution of disagreements between parents as to the opportunities to be provided to their children. The parent who favours a non-discriminatory approach would be able to point out that it has the law on its side. ¹²¹

Moreover, enforcement proceedings are not the only way to give effect to legal prohibitions on discrimination. For instance, Australian federal anti-discrimination law requires complainants to pursue conciliation in the Australian Human Rights Commission, prior to bringing an action in court, and a right to commence proceedings in the Federal Court arises only if conciliation is not successful. ¹²² Conciliation, by its very nature, encourages compromise and reduces the likelihood that the bringing of a complaint will permanently disrupt the relationship between the parties.

As well as the models already used by anti-discrimination law, guidance for institutional design can also be found in other existing regimes that aim at furthering the interests of children in circumstances where their parents cannot be relied upon to do so and where threats to children’s well-being arise within the family. Examples include the regimes for resolving the custody of children in cases of parental separation or divorce and for dealing with parental neglect or mistreatment of children. ¹²³

One regime that we find particularly interesting is the regime for responding to truancy. There are two reasons for this. First, it is directly concerned with parents’ responsibilities to provide adequate opportunities to their children. Second, it does not focus on court orders or on such extreme and disruptive responses to parental neglect as the removal of the child from the home. The primary mode of enforcing the parental duty of ensuring that their child receives an education is not a court order. In formal legal terms, it is a “school attendance order” issued by a local authority. ¹²⁴ But much more importantly, when it comes to giving effect to the legal duty, the result of this formal legal framework is that local authorities have systems of education welfare support to engage with parents and children prior to the taking of formal proceedings and which seek to obviate the need for such proceedings. ¹²⁵ We suggest that anti-discrimination law may look to these existing regimes in developing its own model for ensuring that children’s concerns are heard and that their interests are properly recognised and protected. The effect on family life of counselling from a trained social worker, aimed at encouraging parents to comply with a prohibition on discriminating against their child, would seem likely to be very different from the issuing of a court order. It is undoubtedly very different from the removal of a child from the child’s parents’ custody.

We do not pretend to have considered the full range of possible mechanisms nor to have shown that it is possible to develop a mechanism that will be fully effective in responding to every case of parental discrimination. However, even if a mechanism such as conciliation or counselling were unlikely to bring an end to all parental discrimination, it may still have the potential to enhance, overall, the autonomy interests of the child in question.

V. Conclusion

In this article, we have shown that the family plays a crucial role in distributing the goods with which discrimination law is concerned, and hence is a place where barriers to access to those goods, in the form of discrimination on the basis of protected characteristics, may occur. Furthermore, the family plays this role as much as, if not more so than, the other spheres of activity in respect of which legal duties of non-discrimination are currently imposed. We have also refuted theoretical objections to the extension of legal duties of non-discrimination to the parent-child relationship that appeal to the value of parental autonomy in establishing and maintaining interpersonal relationships. From the purely theoretical point of view, therefore, such an extension seems as if it might be required.

However, we have identified important practical concerns that any such extension would have to deal with, resulting from the fact that children’s autonomy interests are affected by the family in complex and interdependent ways.

This is not just a “stalemate”, however. We accept that heavy-handed state incursions into family life pose a real threat to children’s autonomy. But so does parental discrimination against children; and a failure to extend duties of non-discrimination into this context appears as an acquiescence in, or even a tacit endorsement of, discrimination against children that may significantly affect their opportunities to develop and flourish. Our argument has thus identified a different, but crucial, manner in which anti-discrimination law ought to be extended, namely through the development of new institutions to effectively promote the values of non-discrimination in a part of society that, to date, has tended to sit outside anti-discrimination law’s reach; and we have identified existing regimes that may provide a starting point for this work. Further investigation of these questions of institutional design and of the range of models that might ensure that children’s interests are recognised and advanced is thus a high priority for anti-discrimination law. Such investigation would manifest the “creative” and “reconstitutive” work of anti-discrimination law noted by Gardner which, for the reasons we have set out, requires more than just the legislative promulgation of norms. ¹²⁶