The Queer, the Cross and the Closet: Religious Exceptions in Equality Law as State-Sponsored Homophobia

: The struggle for queer people to be recognised as full sexual citizens continues to be thwarted by the existence of religious exceptions to equality law. These exceptions reactivate and legitimise the historical oppression of queer people, who have long been plagued by the Four Horsemen of Homophobia. War—because the language of war is often used in the context of religious conscientious objection to gay equality. Famine—because public spending cuts have led to religious groups ﬁlling the gap in service provision. Pestilence—because old tropes of infection, promiscuity, and corruption of youth persist, albeit masked by a concern for religious freedom. Finally, Death—because exceptions to equality law operate to limit the citizenship of non-heterosexuals. This paper argues that religiously motivated attempts to restrict queer people’s participation, in a hetero- and theonormative public space, constitutes harm which can be characterised as degrading treatment contrary to Article 3 of the European Convention on Human Rights. The state must be more interventionist in its pursuit of genuine gay citizenship, and remove religious exceptions to equality law; otherwise, it is implicated in the constructive delegation of religious homophobia.


Introduction
The Equality Act 2010 includes both religion and sexual orientation as 'protected characteristics' (S4). It forbids the use of these characteristics as a basis for both direct (s13) and indirect (s14) discrimination. Nevertheless, exceptions have been granted to religious organisations vis-à-vis homosexuality in employment (Schedule 9) and in the provision of goods and services (Schedule 23). This paper argues that restrictions on gay, lesbian, bisexual and queer lives, through religious exceptions to equality law, operate as harms which constitute degrading treatment contrary to Article 3 of the European Convention on Human Rights (ECHR), and which implicate the state in constructive delegation of homophobia. The paper begins by setting the religious exceptions that exist in equality law in the context of the hetero-and theonormative assumptions that permeate society. It goes on to explain how these exceptions operate to mask the homophobia at their core, and to analyse, through a liberal and then a Queer lens, how they result in harm to nonheterosexuals. The struggle for non-heterosexuals to be recognised as full sexual citizens in the UK is still plagued by the Four Horsemen of Homophobia. War-because the language of war is often used in the context of religious conscientious objection to gay equality. Famine-because cutbacks to public spending have impacted on LGBTQ+ support services, with faith-based organisations increasingly filling the gap in provision. Pestilence-because old tropes of infection, promiscuity, and corruption of youth persist, albeit masked by a disavowal of homophobia. For example, the Catholic Care adoption agency relied on these tropes in their unsuccessful bid to exclude same-sex couples from their service (Catholic Care v Charity Commission, FTC/52/2011). Finally, Death-because religious attempts to limit 'sexual citizenship' (Weeks 1998), to limit the freedom of non-heterosexuals to participate in the quotidian activities of life, render them effectively (re)closeted and indeed 'socially dead' (Blasius 1994).

Religious Exce ptions to Equality Law
Religious exceptions are permitted in employment where it is for the purposes of an organised religion (Sch 9, para 2), which gives religious organisations 'a zone of liberty to . . . hire their own members and enforce their own lifestyle norms that are otherwise discriminatory' (Esau 2000). There are further exceptions for religious organisations relating to services and public functions, premises and associations, where: (i) a restriction is necessary to comply with the purpose of the organisation; or (ii) to avoid causing offence to members of the religion or belief that the organisation represents. The exception applies where: (i) it is necessary to comply with the doctrine of the organisation; or (ii) in order to avoid conflict with the strongly held convictions of members of the religion or belief that the organisation represents (Sch 23, para 2). The Act does not include specific exceptions for religious individuals. However, a series of cases, discussed later in this paper, have invoked claims for individual exceptions on grounds of religious conscience. It is submitted that the exceptions granted to religious organisations have created an environment where individual conscience claims can be put forward in opposition to queer citizenship. Even if most of them ultimately fail, these claims have discursive costs that contribute to queer people's continued experience of harmful and degrading treatment at the hands of religious conservatives. Exceptions for organisations and claims of individual conscience both perpetuate the fallacy that religion deserves special allowances because of a purported necessary relationship between religious belief and disapproval of homosexuality. This binary approach not only negates the experience of queer religious people; it also masks the state's constructive delegation of homophobia.
There is a history of religious opposition to the extension of equality law to include queer people within its embrace. For example, the first draft of the Employment Equality (Sexual Orientation) Regulations 2003 (EESOR) included a requirement that sexual orientation discrimination was only permissible if it was a 'genuine occupational requirement' (reg 7). The Church of England sought to widen the scope of the exception, fully recognising that 'this might otherwise constitute direct or indirect discrimination related to sexual orientation' (Fittall 2003, cited by Richards J in R (Amicus) v Secretary of State for Trade and Industry [2004] EWHC 860). The EESOR debates highlight the discourse that personal religious sensibilities are somehow special and are thus worthy of special exceptions. EESOR (re)established a hierarchy of rights in favour of religious privilege. The exception was retained, albeit slightly modified, in the subsequent Equality Act 2010 (Schedule 9, para. 2).
After EESOR, the 2006 Equality Act extended existing laws prohibiting discrimination in the provision of goods, services, facilities, and premises on grounds of race so as to cover religion and faith. Despite religious opposition, the Act also included an amendment tabled by the House of Lords, requiring the government to introduce secondary legislation prohibiting discrimination and harassment on grounds of sexual orientation. The ensuing debate illustrates how religious conservatives were careful to disclaim homophobia, couching their opposition instead in terms of the need to balance 'the right not to be discriminated against on the one hand and the right to freedom of religion on the other', and calling for 'early discussions with the Churches and other faith communities on how their interests can be reflected in the provisions' (Bishop of Newcastle Martin Wharton, HL Debate, 9 November 2005, c 630). There was also concern for how the law would affect 'ordinary people' who provided goods and services such as bed and breakfast facilities in their own homes (Lord Stoddart, ibid., c 631), reflecting a religious imperative to maintain the 'boundaries between homosexuality and private and domestic dwellings' (Johnson and Vanderbeck 2014, p. 108). The image of the beleaguered "ordinary person" who would have to suffer public homosexuality was a key feature of this opposition.
Religious objections to the proposals culminated in significant exemptions for Northern Ireland. In an echo of the 2003 Regulations, restrictions against lesbians and gay men were permitted if 'necessary to comply with the doctrine of the organisation' or 'so as to avoid conflicting with the strongly held religious convictions of a significant number of the religious followers' (Reg 16(5)). Nevertheless, there were moves to have the Regulations annulled, with Lord Morrow arguing that they posed a threat to 'religious liberty': They make it possible for homosexual activists to sue people who disagree with a homosexual lifestyle because of their religious beliefs. Bed and breakfast owners and Christian old people's homes will be sued for not giving a double bed to homosexual civil partners. Wedding photographers will be made to pay compensation for not taking bookings for civil partnership ceremonies. Christians in business could even be sued for sharing their faith with customers. Worst of all, they require religious organisations to choose between obedience to God and obedience to the state (HL Debate, January 2007, c 180).
Lord Morrow paints a vivid image of religious service providers being persecuted by organised homosexuals with a litigious agenda to promote their 'lifestyle' (whatever that may be). From the historical view that homosexuals suffered from an excess of desire, 'homosexual activists' now appear to be suffering from an excessive desire for equality (Brickell 2001).
Following reports from The Equalities Review (2007) and The Discrimination Law Review (2007) the Equality Act 2010 consolidated and replaced previous equality legislation. Service providers were prevented from discriminating against any of the protected characteristics-but this remained subject to the religious exceptions. The general duty on local authorities to promote equality was extended to encompass sexual orientation (s 149). This drew strong (albeit unsuccessful) protests from religious groups, but other religious protests against equality were more successful. For example, the Equality Bill had originally sought to include a proportionality test to the employment exception, requiring religious organisations to show that restrictions on employees' sexual orientation were a 'proportionate means' of adhering to religious doctrine or avoiding offence to the religious convictions of a significant number of the religion's faithful (2009 Sch 9 para 2). Following opposition from the Church of England and the Catholic Church, the proportionality test was dropped (Johnson and Vanderbeck 2014), giving religion a uniquely privileged status in the Equality Act 2010. These bespoke provisions allow greater scope for anti-gay discrimination than that available to other employers, and underline the theonormative assumption that religion is "special". Religious conservatives also sought to widen the remit of the exceptions that existed in the Equality Act (Sexual Orientation) Regulations 2007 (EASOR). Familiar arguments and tropes were reactivated in the course of this opposition to extended gay rights. For example, Lord Mackay attempted to include protection for those with a 'genuine conscientious objection' who wished to withhold goods and services from lesbians and gay men (HL Committee, 13 January 2010, cc 591-2). In addition, the Bishop of Chichester warned against the 'profoundly dangerous tendency' to try to 'privatise belief' (ibid., c 600). These attempts may have failed, but the new Act nonetheless reproduced the broad scope of exceptions granted to religious organisations, enabling them to lawfully discriminate against queer people in the provision and use of goods, services, facilities and premises.
These religious exceptions retell an ancient story that something is wrong with samesex desire. Christian moral ideas pertaining to the body and desire have historically been encoded into legislation and continue to permeate legal discourse. However, over time there has been a gradual shift in conservative religious discourse regarding sexuality: from the original, explicitly religious imperative, through the discourses of "contagion" surrounding HIV and AIDS and the "corruption of youth" around Section 28, to the appeals to heteronormative family life made to oppose same-sex marriage (Johnson and Vanderbeck 2014). Most recently, there has been a move in favour of arguments based on rights rather than morality. This discursive shift has enabled religious conservatism to disavow homophobia while seeking conscience-based exceptions to same-sex equality in arenas such as marriage, adoption, employment and the provision of goods and services. Indeed, the Marriage (Same-Sex Couples) Act 2013 contains a 'quadruple lock': religious organisations must 'opt-in' to solemnise SS marriage in places of worship; no-one can be forced to opt-in; no anti-discrimination law will be contravened by not opting-in; and the Churches of England and Wales are unable to opt-in at all.
Such accommodation is needed, it is argued, because equality law has created a hierarchy of rights that disadvantages the manifestation of sincerely held religious belief. As the former Archbishop of York argued, in the debate over the 2007 Regulations, 'rather than levelling the playing field for those who suffer discrimination . . . this legislation effects a rearrangement of discriminatory attitudes and bias to overcompensate and skew the field the other way' (HL, 21 March 2007. This ignores the historical hierarchy of rights in favour of religion that was further entrenched from EESOR onwards, finding its current zenith in the Marriage (Same-Sex Couples) Act 2013. Interestingly, the former leader of the United Kingdom Independence Party (UKIP) had previously suggested that the need for race equality legislation had passed (BBC News 2015, 12 March). Farage's racist rationale has similarities with the conservative religious narrative which has arisen in response to increasing equality for gay people: these people-these 'others'-now have too much equality; the pendulum has swung too far the other way, and the law must redress the balance.
Both narratives are grounded partly in a discourse of 'British values' (Department for Education 2014, pp. 5-6). There is an underlying caveat to this, however: the small print reminds us that hardworking, preferably white, Christian heterosexual families are the preferred model. Others may be tolerated magnanimously in the spirit of Western liberalism. For example, the report by the Christians in Parliament (2012) (Clearing the Ground) uses the discourses of history and nation to bolster its arguments that Christians are embattled victims of increasing secularism and same-sex equality. Having identified recent changes in law and society as a challenge to Christians, the report goes on to present Christianity as a venerable contrast to the vicissitudes of modernity, by grounding Christianity in claims to historical authority. Christianity is presented as embedded in the very character of the nation, with the implication that it is fundamental to its make-up: 'Christianity has a rich cultural heritage in the UK. For more than 1600 years, it has shaped the way people in the British Isles think and act, both personally and publicly. It is by far the most significant single historical influence on our social and political culture...' (ibid., 73).
The Horsemen of War and Pestilence have been repeatedly invoked in the debates on equality legislation. Discourses of national history, traditional values, and threats to ordinary people have been deployed by religious conservatives to challenge increasing LGBTQ+ equality. The next section highlights the hetero-and theonormativities in which such tropes are grounded, before going on to discuss how religious exceptions are a cloak for homophobia.

Normativities
As well as its claim on history, the argument from religious conscience would not be possible without the alliance of hetero-and theonormativity. Heteronormativity assumes that 'heterosexuality is the normative form of human sexuality . . . the measure by which all other sexual orientations are judged' (Jung and Smith 1993, pp. 13-14, emphasis in original). Theonormativity exists 'when theism is the default, the standard, and everything else is a deviation from this norm' (Harvard Humanist 2012). The extent to which theism has been socially constructed as 'real' is under-appreciated. The noun 'atheist' itself is defined with reference to 'theist', and even the statement "I don't believe in God" carries with it a theonormative assumption. The word 'god' is invariably capitalised, in contrast to those subjects of worship from other cultures (both historical and present). We do not say "I don't worship any gods" or "I don't believe in your god". As such, the idea of a god is pervasive, even in liberal discourse. However, it is dangerous to allow this assumption to go unchallenged. It not only continues to permeate our lawmaking (see for example the Local Government (Religious etc. Observances) Act 2015; R (National Secular Society & Anor.) v Bideford Town Council [2012] EWHC 175); it is also precisely what creates a space for 'conscience' talk to be used by religious organisations and individual religious conservatives.
Giving space, in a hetero-and theonormative society, to religious conservatives to use religious conscience to deny goods and services implicates the state in perpetuating homophobia. There are three reasons for this. First, the state has a duty to ensure the safety and security of its citizens (Hobbes [1651(Hobbes [ ] 1996, and religious exceptions deny nonheterosexuals equal rights to safety and security as embodied individuals. Second, law's historical relationship with religion continues to be the vehicle whereby ambivalence or even distaste for homosexuality finds legal expression, albeit now cloaked in rights-based arguments. The disclaiming of homophobia and the assimilation of rights arguments by conservative religion disguises what is still a discourse grounded in heteronormativity and theonormativity. Third, religious exceptions deny lesbians, gay men, bisexuals, and queer people equal sexual citizenship and thus equal dignity as human beings. A political and legal regime that professes to be concerned with equality should attend to those parts of the law that accord homophobia the legal legitimacy to compromise equal citizenship, dignity and security.

Religion and Homophobia
It could be argued that faith-based opposition to homosexuality is not equivalent to homophobia, being based on a belief in the 'revealed word' of their god, and not simply on an irrational fear of gay people. The 'holy' books of the world's three main creedal religions can be interpreted as forbidding homosexuality, and many believers accept these interpretations as moral imperatives. Leaving aside the argument that religious belief is inherently irrational (Dawkins 2006, cf Ward 2011, this denial of homophobia is not only an example of sophistry; it also negates the lived experience of non-heterosexuals who have suffered-and continue to suffer-at the hands of religious conservatism. It is also worth noting here that Christian religious texts that refer to same-sex practice are very limited; the meaning of those that do exist (such as the story of Sodom and Gomorrah) has also been contested (Boswell 1980). In other words, even if we are to take seriously religious belief, it should not be conceded that the texts are static or unquestionable. They have been contested, reworked and reinterpreted over time and so their meanings have shifted. This does not (necessarily) mean that believers are insincere in their beliefs; however, it does permit a challenge to religious claims in their own terms. That is, they are not only problematic in terms of rights conflicts, but also in terms of internal consistency and logic. Some religious adherents may feel strongly that they are bound by a fundamental, unchanging and unchangeable text; but this does not bear scrutiny in broader historico-cultural-political terms.
The disavowal of homophobia by religious conservatives has been further facilitated by their success in reframing the issue as a clash of rights. However, there is a qualitative difference between the two types of rights claimed. Complaints of discrimination on sexual orientation grounds are made either for equal benefits or access, or an end to discrimination. In contrast, religion-based claims seek exceptions so that religious individuals or organisations do not have to abide by rules which apply to others. Thus, the discourse of rights has been increasingly used by religious conservatives, but in pursuit of an aim that is unlike other rights claims, namely 'the right to deny equal, inclusive treatment for queer people . . . where services are being offered to, or even on behalf of, the public' (MacDougall and Short 2010). Religious use of rights arguments has been made in other areas, such as conversion therapy (Clucas 2017). And the disavowal of homophobia has been used by the religious right as both a shield and a sword against equality campaigners. While MP Diane Abbott argued in support of protection against harassment for sexual minorities, acknowledging that 'some people use their religion as a vehicle for cultural bigotry' (HC Committee, 18 June 2009, c 308), MSP John Mason retorted that 'one person's bigotry is another person's belief' and extending harassment protection to homosexuals would 'create a risk to free speech' of religious people (ibid.).
The underlying message to a non-heterosexual is that 'unless I am being thrown in prison or herded onto a cattle train, then it is not homophobia' (Bliss 2014). For example, in 2014 the Irish television station, RTÉ, agreed a financial settlement with campaigners against same-sex marriage, after a gay rights campaigner and drag queen had described them on television as homophobic (Kealy and Horan 2014). Her response is worth reproducing at some length: So now Irish gay people find ourselves in a ludicrous situation where not only are we not allowed to say publicly what we feel oppressed by, we are not even allowed to think it because our definition has been disallowed by our betters . . . And a jumped-up queer like me should know that the word "homophobia" is no longer available to gay people. Which is a spectacular and neat Orwellian trick because now it turns out that gay people are not the victims of homophobiahomophobes are (Bliss 2014).
The legal endorsement of religious homophobia, by virtue of exceptions to equality law, can be understood as a state-sponsored harm-the scope and degree of which can be assessed by examining both liberal and Queer approaches to harm. This paper engages with liberal analysis, but also recognises that arguments for law reform-for example, through use of Article 3 ECHR-can be strengthened by understanding the power relations that limit sexual citizenship. It is argued that a Foucaultian analysis of harm offers greater analytical precision while also opening further possibilities as to how homophobic harm can be assessed at state level. First, it is helpful to examine the relationship between individual and institutional homophobia.
Homophobia is commonly used to describe a range of negative attitudes and behaviours towards same-sex desire and relationships. It can be manifested in critical or hostile words or behaviour, from distaste or disapproval, through discrimination, to verbal and physical violence. The psychologist George Weinberg (1972), who is credited with the origin of the term, described homophobia in medical terms (see also Foucault 1978). Interestingly, he also defined it with explicit reference to religion; it was 'a fear of homosexuals which seemed to be associated with a fear of contagion, a fear of reducing the things one fought for-home and family. It was a religious fear and it had led to great brutality as fear always does' (Weinberg 1972). "Holy" books can be used as authority for a range of prejudices. For example, faith-based homophobia echoes the biblical justifications for racism and sexism historically relied upon by religious conservatives in the United States: 'Slavery, racial discrimination and segregation, and opposition to women's rights were all supported by strong religious arguments bolstered by citations to the Bible' (Curtis 2012). As Curtis recognises, 'as with race and gender, the greatest harm may be to the human spirit, the harm inflicted when gays internalize the message sent by hostility and discrimination' (ibid.). As this paper contends, homophobia has become so sedimented within legal, social and political discourse that, even when equality provisions have been extended to non-heterosexuals, they have remained tainted.

Individual and Institutional Homophobia
The ILGA 2020 report highlights a link between state and individual discrimination: In many countries across the region, and not only those with a documented growth in official bias-motivated speech, there has been an equally sharp increase in online hate-speech and physical attacks on LGBTI people, many of the latter premeditated and brutal . . . Brexit, for instance, and the populist narrative surrounding it, can be linked to an increase in anti-LGBTI hate crimes and incidents in England and Wales from 5807 in 2014-15, to 13,530 in 2018-19. Other developments such as the banning of events in Armenia, Hungary, Poland, Russia, and Turkey, and the prosecution of participants in Pride events in the latter, add to an atmosphere lacking in a sense of safety.
At the institutional level, in 2020 the Home Office denied asylum to at least 3100 people who had fled countries that criminalise same-sex relationships. In August 2020, a gay asylum seeker was refused because the immigration tribunal judge decided he did not have a "gay demeanor" (ILGA 2020). Police records show that anti-LGBT hate crimes and incidents in England and Wales increased from 5807 in 2014-15, to 13,530 in 2018-19, but the prosecution rate fell from 20% to just 8% (ILGA 2020). Additionally, at the individual level, the recent British Social Attitudes survey has revealed that 'the acceptance of same-sex relationships has slowed down in the UK, with a significant minority remaining hostile . . . perhaps reflecting the marked divides between the attitudes of religious and non-religious people in this sphere' (BSA 36 2020). Nevertheless, two-thirds of people in Britain now say that homosexuality is "not wrong at all", an increase of almost 50 percentage points since 1983 (ibid., p. 6). The report states that 'both public attitudes to and the law on sex and sexuality are now profoundly out of step with the doctrinal position of many established faiths in Britain, including that of most Christian denominations' (ibid.).
Several studies, here and in the US, have identified a link between homophobic discrimination and harm suffered by lesbians and gay men-including psychological harm. A systematic review of mental disorders, suicide and self-harm amongst LGB (lesbian, gay and bisexual) people in England found that 'the social hostility, stigma and discrimination that most LGB people experience is likely to be at least part of the reason for the higher rates of psychological morbidity observed' (King et al. 2008). Citing 'difficulties growing up in a world orientated to heterosexual norms and values and the negative influence of social stigma against homosexuality', it was suggested that 'people who feel discriminated against experience social stressors, which in turn increases their risk of experiencing mental health problems', adding that greater efforts are needed to prevent these issues arising (Chakratborty et al. 2011, emphasis added).
At this point, it is helpful to consider in more detail the concept of harm, from both liberal and Queer perspectives. The following section contrasts the liberal position on harm with a Foucaultian-informed analysis, drawing on Kendall Thomas's examination of sodomy laws in the United States. Thomas's approach is adopted to argue that religious exemptions, viewed in terms of power relations, are one example of the state's constructive delegation of homophobia to conservative religious organisations and individuals.

The Liberal Position
Mill's 'harm principle' holds that the need to prevent harm to persons other than the actor is the only morally relevant reason in support of state coercion of the individual (Mill [1859(Mill [ ] 1985. The general liberal position is that the need to prevent harm to others is always a relevant reason to engage the criminal law, if not necessarily fully determinative (Feinberg 1984a). In his extended study of the moral limits of the criminal law, Feinberg distinguishes between normative and non-normative senses of harm: the latter involves some sort of setback to a person's interests (Feinberg 1984a, vol I, p. 34), whereas the former, normative sense involves a wrong which is a violation of a person's rights. The harm principle applies only to setbacks to interests that are also wrongs, and only to wrongs which also involve a setback to interests: Interests can be blocked or defeated by events impersonal in nature or by plain bad luck. But they can only be 'invaded' by human beings . . . It is only when an interest is thwarted through an invasion by self or others, that its possessor is harmed in the legal sense . . . One person harms another . . . by invading, and thereby thwarting or setting back, his interest (ibid., pp. 215-6).
In terms of what is to count as harm, Feinberg rules out 'mere transitory disappointments, minor physical and mental "hurts" and a miscellany of disliked states of mind', including moral indignation (ibid.). Feinberg also distinguishes between general and normative senses of 'offence': general offence comprises myriad discomfiting mental states, whereas normative offence refers only to those states caused by the wrongful-rightviolating-conduct of others (Feinberg 1984b, vol II, pp. 14-22). A newer and broader definition of harm recognises that it 'can be emotional, mental, and physical . . . caused by condemning the other . . . avoiding his presence [and] discriminating against him' (Nehushtan 2007). Read together, these definitions support the conclusion that anti-gay discrimination does involve harm to lesbians and gay men, when understood within the societal framework of heteronormativity and homophobia.
The liberal position on the conflict between religion and sexual orientation has, however, been influenced by its view that the state should be neutral among competing conceptions of what constitutes a good life (Rawls 1971). This neutrality is reflected in the fact that both Article 8 (right to privacy) and Article 9 (freedom of religion) of the ECHR are qualified rights, meaning that they can be limited if the restrictions are a proportionate means of achieving a legitimate aim. Pitting two qualified Convention rights against each other contributes to the zero-sum game that has hitherto dogged scholarly approaches to this conflict. It also illustrates how vulnerable legal rights can be when liberal rights discourse is appropriated to achieve illiberal aims (see Nehushtan 2015). However, it can be argued that equality law should be recognised as value-based, rather than neutral. It presumes certain values. For example, the Equality Act states that direct discrimination because of a protected characteristic is never justified, so freedom of religion ends when it expresses itself through unjust discrimination against others directly because of a protected characteristic. Accordingly, the state has set the limits of liberal tolerance by relying on content-based, rather than content-neutral, considerations.
A content-based approach to anti-discrimination law means, therefore, that the harm principle should offer no comfort to people suffering profound offence on religious grounds: people such as the Bulls, who were fundamentally opposed to letting double-bedded rooms to guests other than heterosexual married couples (Bull v Hall [2013] UKSC 73). If they view homosexuality as a sin, such people can be deeply offended by the idea of same-sex intimacy happening even in private, but this does not justify the argument that they should be granted exceptions on this basis. A content-neutral approach may agree that a feeling of outrage at gay sex cannot qualify as liberal harm, because it involves privileging the moral view that gay sex is wrong over the view that gay sex is not wrong. As Feinberg points out, 'if his impersonal moral outrage is to be the ground of legal coercion and punishment of the offending party, it must be by virtue of legal moralism-to which the liberal is profoundly opposed' (Feinberg 1984b, vol II, p. xiv). However, a content-based approach would go further and require the state to reinforce the limits of tolerance for illiberal views, which means explicitly privileging anti-discrimination in conversations about, and assessments of, harm. Thomson's (1990) distinction between different types of distress is worth considering here (as well as her interesting observation that we all quite enjoy feelings of moral indignation at times!). For Thomson, there is simple distress (in the sense of a feeling that one dislikes having) and then there is distress that is caused because we have a certain belief. Belief-mediated feelings can be held rationally or irrationally, and people bear some personal responsibility for how long they have these feelings and how intense they are. Even if the feeling is not irrational, Thomson does not view belief-mediated distress as a harm. The very word 'harm' has a tendency to 'slither' (ibid., p. 260), enabling people to say they have suffered a harm whenever something happens that they would prefer not to have happened. Indeed, harm is loaded with moral significance, so it can be a useful strategy or tactic if one wants another to stop doing a particular thing to say that thing is causing them harm.
One of Thomson their jobs because of their refusal to provide services to gay people (civil partnership and counselling, respectively), and the Bulls had to sell their hotel as a result of their refusal to let double rooms to same-sex couples. However, according to Thomson, it is not an infringement per se to cause someone to lose their job (unless the loss were caused by an infringing act such as spreading lies about them), because 'the gravamen of the charge against one who causes a status worsening lies in the means used. If those means are no infringement of a claim, then causing the status worsening is not either. Thus status worsenings are not themselves harms' (Thomson 1990, p. 266). On a Thomsonian view, then, Ms Ladele, Mr McFarlane, and the Bulls were not harmed by the courts' judgments. The European Court of Human Rights (ECtHR) in the Eweida case felt differently, however, and held that: [R]ather than holding that the possibility of changing job would negate any interference with [Article 9], the better approach would be to weight that possibility in the overall balance when considering whether or not the restriction [on the right to manifest religious belief] was proportionate ([2013] ECHR 37, [83]).
As part of understanding what such proportionality should mean in practice, the concepts of dignitary and symbolic harm can be useful. The dignitary harm caused by discrimination against gay people is more than mere offence or hurt feelings; it leads to the undermining of social standing-a social harm. From a dignitary harm perspective, religious objectors are not automatically harmed by the courts' refusal to grant exceptions to equality law: Even if exemptions are refused to service-providers, with some harm to their conscience, this does not shift the dignitary harm to those that oppose homosexuality by singling them out as second-class citizens. In fact, the law upholds their equal social standing through various fundamental rights (Adenitire 2020, p. 280).
Moreover, non-heterosexuals 'are in no less need of access to basic commercial services . . . If the law allowed services to be denied to LGB persons it would signal that they do not have equal social standing and hence are second-class citizens' (ibid., p. 282). Similarly, if the state has a responsibility to treat its citizens equally, and to ensure their dignity, safety and security, it must pay attention to both the 'symbolic meanings attaching to social practices and the symbolism of its own pronouncements and silences' (Boetzkes 2000, p. 327). Homophobia is a harm; organisational homophobia is expressed in religious exceptions to equality law; individual homophobia is expressed in hate crime but also in faith-based conscientious objections to equality provisions. If the state permits exceptions or objections it is not condemning them, and when the state avoids explicitly condemning homophobia, it can be seen as sustaining it; 'it is not far-fetched to interpret non-condemnation as support' (Nehushtan 2015, p. 154). As this paper will go on to argue, this also reflects Kendall Thomas's (1992) characterisation of individual and organisational homophobia as constructive delegation of state power.
Perhaps religious conservatives are right to be concerned. Reliance on liberal rights discourse has so far proved a less than effective tool for individual religious conservatives. It has failed to erode those rights and protections only recently granted to lesbians and gay men in the Equality Act. This explains the move towards conscience clauses, by relying on something purportedly 'higher' than human law. As the Christian bakery stated in Lee v Ashers Bakery ([2015] NICty 2), they believed that their business 'must be run by God's wishes ' (ibid.: [22]). On the matter of same-sex marriage, the defendants stated, 'the only divinely ordained sexual relationship is that between a man and a woman within the bonds of matrimony . . . No other form of marriage is permissible according to God's law . . . according to God's law, homosexual relations are sinful . . . ' (ibid., p. 15). The Enlightenment, at least in this sense, has come full circle. There may not be a medieval theocracy, but there remains a pervasive theonormativity, which is the reason why appeals to discriminate based on conscience can even be voiced. It is submitted that theonormativity and heteronormativity operate in a discursive alliance, together providing a basis for unequal power relations that harm queer people by limiting their citizenship and thus their agency. Our understanding of harm can be further improved through a Foucaultian perspective that recognises the importance of relations of power and their sedimentation through time.

A Foucaultian Analysis of Equality and Power
If power is examined through a Foucaultian lens, it is revealed as that which characterises relations between parts of any given society-and between individuals in that society-as relations of struggle. This approach conceptualises power not as a top-down mode of domination, but as a more dynamic, unstable set of strategic personal, social and institutional relations (Foucault 2007). It better describes the changing relations of power that have operated between law, religion and sexual orientation since the latter half of the twentieth century. It accounts, with greater analytical precision, for the discursive shifts that have taken place in religious conservatives' opposition to increasing equality for lesbians and gay men. Furthermore, the Foucaultian approach offers a helpful means of both critiquing this opposition and suggesting a different understanding of sexual orientation equality.
In applying a Foucaultian understanding of power to equality law, this paper takes inspiration from Thomas's argument that US anti-sodomy laws legitimised the homophobic violence perpetrated by citizens and individual state officers, through the constructive delegation of state power (Thomas 1993). Taking a Foucaultian perspective, Thomas contends that we need a broader appreciation of the interests at stake than arguments based on privacy permit. Thomas's study of US anti-sodomy laws led him to conclude that the main question is not about individual rights but about political power relations; this paper draws on his arguments to reach a similar conclusion regarding UK equality law.
Adopting Foucault's characterisation of power, Thomas argues that homophobic violence perpetrated by citizens can be viewed as constructive delegation of power by the state to those citizens, because private relations have public origins and public consequences. Religious exceptions and arguments from conscience in equality law-although neither criminal nor violent in the strict sense-can nevertheless be understood according to similar principles. If we understand religious exceptions as homophobic harm, they also implicate the state in the perpetration and perpetuation of homophobia through constructive delegation of power to religious organisations and individuals. The problem is exacerbated by the Horseman of Famine: the ideologically-driven programme of austerity and the consequent slashing of social welfare spending, with religious organisations forming part of the voluntary sector plug in this social support gap. When religious organisations take on an active role in civil life, it can be argued that 'such involvement is quasi-governmental in nature, and ought therefore to be subject to the same constraints that would be imposed on a government engaged in such an operation' (MacDougall and Short 2010, p. 141). By continuing to permit exceptions for religious organisations, the state is permitting homophobia against society's most vulnerable queer people. It is submitted that a Queer-informed analysis of equality law is important in recognising and addressing these issues.

A Queer Analysis of Equality Law
Legal discourse has the power to shape sexual subjectivities: 'law constitutes and regulates, punishes and self-disciplines' (Cossman 2007, p. 16). This power is found in equality law; the very framework of anti-discrimination law requires an 'other' to use as a comparator with reference to an established norm. Under the Equality Act, discrimination on grounds of a protected characteristic is established with reference to a standard comparator. People of colour, women, people with disabilities, non-heterosexuals and those who do not conform to gender binaries are compared to the universal legal subject. Equality law thereby grants protection to those 'others' who have been able to show that they are sufficiently 'like' white, male heterosexuals in order for the comparator to be meaningfully deployed (Halley 2006, p. 109). For example, in the Bull case, the focus was on the equivalence of marriage and civil partnership, and the comparator was a heterosexual married couple-the very paradigm of heteronormativity. Except for Lady Hale ([2013] [53]), the Supreme Court had very little to say about the implications of religious exceptions for sexual citizenship. There was no recognition of the heteronormativity implicit in the equivalence of heterosexual marriage and homosexual civil partnerships. In effect, homosexual relationships are considered suitable for legal protection if they mirror traditional heterosexual ones.
Thomas's concept of constructive delegation helps to illustrate how homophobia continues to be legitimised by the state through religious exceptions to equality law, and through the theonormative and heteronormative frameworks that they support. Religious exceptions provide a space where homophobia is permissible. For the state to allow 'reasonable accommodation' of the religious conscience amounts to state facilitation of sexual prejudice in the public sphere. It enables old anti-gay myths and tropes to be reactivated at both institutional and individual levels. At this point, a discussion on the relationship between privacy and power is necessary.

Power and Privacy
The liberal understanding of privacy sees it as a haven from the external world, including the state. This is indeed the image one sees if one looks through a heteronormative lens. In the liberal state, heterosexuals may have been able to enjoy privacy as an unqualified good; as a zone in which they can express their intimate needs and desires without interference. This has not been the case for non-heterosexuals, who have had to hide or 'closet' their sexual orientation in order to survive within society. As Halperin observes, the closet offers a zone of protection 'from the many and virulent sorts of social disqualification that one would suffer were the discreditable fact of one's sexual orientation more widely known' (Halperin 1995, p. 29). At the same time, this need for individual self-protection enables society to maintain its heteronormative structure. As Sedgwick recognises, privacy for lesbians and gay men has enabled society to enjoy 'the epistemological privilege of unknowing' (Sedgwick 2008, p. 5), whereby the veneer of heterosexuality is not tarnished by a forced recognition of the homosexual 'other'. Thus, the closet serves as a reminder of the relations of power that exist with regard to sexuality. Therefore, as Thomas argues, privacy for non-heterosexuals 'has always been both a tool and a trap, insofar as privacy has functionally served as a cornerstone for the very structure of domination that the principle has been used to attack' (Thomas 1992(Thomas , p. 1456, and it is this structure of domination that is criticised here. This is precisely why the extension of sexual minority rights beyond the private sphere, into employment, education and into the marketplace of goods and services, has provided such a flashpoint for conflict with religion. Religious conservatives have bemoaned their legal losses, and it is tempting to view sexual orientation as the winner. However, focusing on court decisions in case law can cloud our understanding of the forces that do (or should) motivate the law (Moore 1989). This is why an analysis of the power relations involved is important, and why Thomas advocates a focus on political power when considering the implications of homosexual sodomy statutes-or, in this case, of religious exemptions to anti-discrimination law. Thomas reminds the reader that lesbian and gay history 'is a story of homophobic aggression and ideology. Its central theme is the fear, hatred, stigmatization and persecution of homosexuals and homosexuality' (Thomas 1992(Thomas , p. 1462. He likens the constant, underlying threat of homophobic violence to living under the threat of terrorism (ibid., pp. 1465-66). This paper has earlier discussed how religious conservatives have used the discourse of war to describe their experience. Thomas uses the discourse of terrorism to provide a useful insight into the lived experience of lesbians and gay men around the world-including the UK. This is a bold comparison, which may seem overly dramatic or exaggerated at first glance, and may explain why the jurisprudence of the ECtHR has been loath to use Article 3 in the context of gay rights. The following section addresses this by considering Article 3 in more detail.

The Nature of Torture and Degrading Treatment
Article 3 of the European Convention states that 'no one shall be subjected to torture or to inhuman or degrading treatment or punishment'. The United Nations' legal definition of torture has five elements: (i) the nature of the act; (ii) severe pain and suffering; (iii) the perpetrator's intention; (iv) the purpose of the torture or cruel, inhuman and degrading treatment (CIDT); and (v) the involvement of public officials (UN Voluntary Fund for Victims of Torture 2011). In his discussion of conversion 'therapy', Romero makes a convincing case for its inclusion within the legal definition: the methods used satisfy the word 'act'; it meets the definition of severe pain and suffering as well as the requirement of intention; its purpose is either punishment or discrimination; and it has been committed by public officials (Romero 2009). Placing an act within the ECHR's hierarchy of severity established by the ECHR is done on a case-by-case basis. Romero concedes that 'because the methods of conversion therapy vary and the victims of conversion therapy vary in age and vulnerability, it would be difficult to put conversion therapy squarely within one category' (i.e., the torture category or the CIDT category) (ibid., p. 224). Nevertheless, 'conversion therapy satisfies at least degrading treatment, the lowest form of CIDT, because of its humiliating nature' (ibid.). This paper contends that religious exemptions to equality law constitute similarly degrading treatment towards non-heterosexuals.

The European Court of Human Rights
The European Court of Human Rights (ECtHR) has stated that Article 3, (unlike Articles 8-11) is an absolute and unqualified right (Ireland v UK, 1978). However, the Court's case law has found guidance in principles such as the 'margin of appreciation' which have allowed an element of relativity and subjectivity to seep into its jurisprudence. A member state's margin of appreciation is assessed with reference to the 'legality, legitimacy and necessity of any restriction by a public authority' (Johnson 2013, p. 69). It assumes that individual states are best placed to decide what limitations may be placed on rights as being necessary in a democratic society (Handyside v UK [1976] ECHR 5). The Court's practice under Article 3 has been criticised for being 'based not on objective criteria but on the effects of various subjective factors on the particular facts of each case, leading to decisions which can be hard to reconcile' (Addo and Grief 1998, p. 514). As Johnson suggests, the Court has sought to maintain a minimum level of severity before Article 3 can be engaged, so as 'not to trivialize the substance of the provision or encourage rights inflation under it' (Johnson 2013, p. 196). There is some substance to the argument that using Article 3 in some homophobic situations may trivialise the serious harms that the provision is aimed at. Moreover, in an echo of the zero-sum game that has resulted from religion and sexual orientation's status as protected characteristics in equality law, allowing Article 3 may lead to its use in cases of anti-religious speech or actions. The following section examines arguments in favour of using Article 3, before suggesting that there might be a Queerer alternative.
The Court has defined inhuman treatment as treatment which may 'cause either actual bodily harm or intense physical or mental suffering' (Kudla v Poland, 2000, p. 92), and degrading treatment as that which can 'arouse in the victim feelings of fear, anguish and inferiority capable of humiliating and debasing them' (ibid.). The Commission has stated that the minimum level of severity for 'inhuman treatment' is treatment that inflicts 'severe mental or physical suffering' (X v Federal Republic of Germany,no. 9191/90). It is unfortunate that the threshold for severity might prevent relief being granted in situations other than where homosexuality is criminalised and gay people are subject to torture. A better way forward would be to focus on 'degrading' rather than 'inhuman' treatment, as 'this focus on actions that create feelings of fear, anguish and inferiority that humiliate and debase individuals provides scope to evolve the Court's interpretation of the Convention in respect of sexual orientation' (Johnson 2013, p. 197). Furthermore, the Court views the Convention as a 'living instrument' which recognises the 'increasingly high standard being required in . . . human rights and fundamental liberties correspondingly and inevitably requires greater firmness in assessing breaches of the fundamental values of democratic societies' (Selmouni v France, 1999, p. 101). Once again, a parallel can be drawn with race discrimination. In Moldovan and Others v Romania (No 2), the Court reiterated that 'discrimination based on race can of itself amount to degrading treatment within the meaning of Article 3 of the Convention' (2005, para. 113).
Using Article 3 as a basis for sexual orientation discrimination claims represents a synthesis of a value-based liberal harm principle and the idea of the constructive delegation of state power, into an approach that better captures the harm suffered by sexual minorities in the public sphere. It would mean that sexual orientation rights claims are no longer closeted as 'private' matters pertaining solely to Article 8, thereby recognising the citizenship rights of gay people. There are signs of some, if insufficient, movement in this regard. For example, in Smith and Grady v UK (ECHR 1999-VI), the applicants claimed that they had been subjected to discriminatory treatment during investigations into their sexual orientation by the armed forces. They argued that the investigations were 'based on crude stereotyping and prejudice [that] denied and caused affront to their individuality and dignity' (ibid., p. 119); and that the questioning during the investigation was 'hurtful and degrading . . . prurient and offensive' (ibid.). The applicants did not succeed in meeting the 'minimum severity' threshold, but nonetheless the judgment did recognise that, because degrading treatment causes the recipients 'feelings of fear, anguish and inferiority capable of humiliating and debasing them' (ibid., p. 120), the Court would 'not exclude that treatment which is grounded upon a predisposed bias on the part of a heterosexual majority against a homosexual minority . . . could, in principle, fall within the scope of Article 3' (ibid., p. 121).
The case of Identoba ((2015) ECHR 537) was the first time the ECtHR recognised that 'hate crime' committed against individuals based on sexual orientation amounted to a violation of Article 3 taken in conjunction with Article 14 ECHR. Article 3 (taken in conjunction with Article 14) imposes a positive obligation on member states to ensure that all individuals within their jurisdiction are protected against all prohibited forms of ill-treatment, including ill-treatment administered by private individuals. This can be read as supporting Thomas's concept of constructive delegation, by requiring states to protect LGBT people from 'hate speech and serious threats' and 'physical abuse' that cause them to feel 'fear, anxiety and insecurity' (ibid., p. 79). The judgment means that Article 3 requires states to put in place both preventative and investigative measures to protect LGBT people who peacefully assemble in public. As Johnson observes, 'the Court's message is that states must have a robust framework of law enforcement that protects LGBT individuals from ill-treatment motivated by homophobia (Johnson 2015) (ECHR Sexual Orientation Blog, 13 May 2015). It will be interesting to see to what extent the application of Article 3 to peaceful assembly in public can be built upon in order to protect gay people in public space generally. It may be regarded as too much of a leap, and may raise further concerns over trivialisation and dilution of the Convention right. In any event, Identoba can be seen as the start of a process whereby gay people have the right not to experience degrading public treatment because of their sexual orientation.
On a global level, the United Nations has recently accepted that: . . . members of sexual minorities are disproportionately subjected to torture and other forms of ill-treatment because they fail to conform to socially constructed gender expectations. Indeed, discrimination on grounds of sexual orientation or gender identity may often contribute to the process of the dehumanisation of the victim, which is often a necessary condition for torture and ill-treatment to take place (United Nations Human Rights Council 2015, p. 34).
The UN's analysis should not only apply to states which have the most draconian anti-gay legislation. It is not helpful to say to gay people in the UK, "aren't you lucky you don't live in Uganda"? Socially constructed gender expectations permeate Western democracies just as they do elsewhere; the difference is only a matter of degree.
Having said all this, there may be an alternative to what some may see as stretching Article 3 to combat the problem of non-violent homophobia that religious exceptions permit. It may be that the ultimate solution will come from religious organisations themselves-and their queer members.

A Queerer Alternative: Resistances
The battle against the Horseman of Death-the social exclusion caused by limiting the citizenship of non-heterosexuals-may have to be fought on more than one front. One answer may be in Queer challenges within the religious organisations. As Foucault recognised, discourse transmits and produces power/knowledge. Discourse can reinforce power/knowledge, but discourse also 'undermines and exposes it, renders it fragile and makes it possible to thwart it ' (1978, p. 101). This possibility relates to the 'tactical polyvalence of discourses' (Patton 2005, pp. 273-74), in that they can be used as opportunities for resistance. It is important to recognise the resistances that are taking place within religious organisations and communities, with the voices of LGBTQ+ people of faith now beginning to be heard. The old tropes have been challenged by religious individuals, some of whom are also LGBTQ+. For example, Reverend Sharon Ferguson of the Lesbian and Gay Christian Movement states that 'no one religion should be able to influence to such an extent the laws that we put into place' (Donald 2012, p. 25). The Christian policy organisation, Ekklesia, has observed that 'the Church has become conditioned to think theologically and practically from a top-down perspective using instruments of the state to get what it wants . . . It has to have its arguments accepted on merit rather than privilege; it has to lead by example' (Nehushtan 2015, pp. 14-15).
A significant aspect of this resistance is to do with clergy's desire to have and to formalise their same-sex relationships. For example, the current Bishop of Grantham is the first Bishop to declare that he is in a same-sex (albeit celibate) relationship (Sherwood 2016b). Notably, a senior bishop from the Church's evangelical wing has called for farreaching change in the church's attitudes to homosexuality and a welcome to Christians in same-sex relationships (Sherwood 2016a). Unfortunately, words have not translated into Church action so easily. Jeremy Pemberton, a gay clergyman, was prevented from taking up a post as a hospital chaplain after marrying his partner, and lost his appeal over a claim of discrimination against the Church of England (Pemberton v Inwood [2018] EWCA Civ 564).

Conclusions
The concept of harm is typically discussed with reference to the moral limits of the criminal law. This paper acknowledges that, while equality law does not intervene only when harm (or offence) is found, harm is rightly considered in matters of discrimination. A finding of discrimination in equality law does not necessarily hinge on establishing harm-whether that takes the form of wrongful setbacks to interests, status worsening, dignitary harm or symbolic harm-but harm still matters when thinking about inequality. The arguments presented in this paper are based on a Queer engagement with value-based liberalism: a recognition that it is the content of any conscience-based arguments that matters, and the content must be seen in the context of hetero-and theonormative power relations.
In the context of goods and services provision, for example, equality law has two purposes. First, the law aims to prevent service providers from discriminating against customers on the grounds of their protected characteristics. The second, broader purpose is to prevent service providers from acting upon their discriminatory views when they provide services to the general public. Service providers are thus prevented from refusing to provide a service, if their refusal is based on an adverse judgment about others because of their protected characteristics (Nehushtan and Coyle 2019). With regard to both purposes, the consequences of the discriminatory act-including whether harm is caused-are not necessary for a finding of discrimination. Equality law holds direct discrimination to be inherently wrong regardless of its actual consequences, because service providers are not allowed to put their discriminatory views into practice-whether harm (or offence) is caused. The purpose of equality law is broader than preventing harm or offence; its purpose is to prevent service providers from discriminating in their service provision. This is clear when we consider that a refusal to provide a service because the service provider holds homophobic views is illegal, regardless of whether the customer is in fact gay. Therefore, while discrimination does not hinge on establishing harm, harm is still relevant in both a "real" and a symbolic sense.
Notwithstanding the extension of equality law to include gay people in its embrace, religious exceptions remain. It is important to recognise the harm caused by homophobia, and to look beyond the scope of the criminal law to recognise the degrading treatment caused by discrimination. Discrimination on sexual orientation grounds must be recognised as homophobia, even if it is cloaked in the language of rights or conscience. Homophobic discrimination is validated by religious exceptions that limit gay people's access to full and equal citizenship. Allowing religious exceptions is itself an expression of homophobia at an institutional level; an expression that is based on the hetero-and theo-norms that continue to pervade society. Discrimination on sexual orientation grounds is not simply an affront to the victim's dignity-it produces and perpetuates discursive effects on a societal level. These real harms should be recognised in the context of equality law, not least to counter the arguments put forward by religious conservatives that gay people now have 'enough' equality-and that, indeed, the pendulum has swung too far the other way, so that the religious conscience is now 'closeted' instead of the gay person. Hence, this paper contends that such discrimination can be understood as degrading treatment and constructive delegation of homophobia, and the law should be amended to recognise this. Whether the solution lies in extending the scope of Article 3 ECHR remains a matter of debate. The concerns over trivialisation, as well as the threat of it being counter-deployed by religious conservatives, mean that extending its embrace may prove to be unsuccessful. The moves made by queer members of religious organisations offer hope, but the courts have yet to endorse their claims. It seems that the conflict between religion and sexual orientation in equality law is not over; the Four Horsemen of Homophobia may be battle-scarred, but they live on.
Funding: This research received no external funding. The arguments in this paper are drawn in part from the author's unpublished PhD thesis, 'The queer, the cross, and the closet: an analysis of rights discourse in conflicts between religious belief and sexual orientation'. The PhD was funded by Keele University.
Institutional Review Board Statement: Not applicable.