Reforming the Approach to Racial and Religious Hate Speech Under Article 10 of the European Convention on Human Rights

Philippe Yves Kuhn. Human Rights Law Review. Volume 19, Issue 1. February 2019.


This article has two main objectives. First, to put forward a theoretically coherent test for the identification of hate speech in both its racial and religious forms, in order to pave the way for more principled adjudication by the European Court of Human Rights (‘Strasbourg Court’) and national courts in Council of Europe Member States, which are all required to be Parties to the European Convention on Human Rights (ECHR or ‘Convention’). The latter applies particularly where States choose, like the UK, to discharge their convention on Human Rights (ECHR or ‘convention’) obligations by applying the ECHR approach. Secondly, to bring the law on Article 10 of the ECHR more in line with European public perceptions of the scope of free speech entitlements, especially in the religiously offensive speech context. This is particularly topical given the current Syrian Refugee Crisis and Brexit process, which have contributed to heightened public awareness of racial and religious tensions in Europe.

In essence, I seek to establish that the Strasbourg Court’s current two-track approach to religiously offensive speech, on the one hand, and racial hate speech, on the other hand, is unhelpful and capable of unification through a single test. Waldron’s dignitarian account of the harm in hate speech will be drawn on to provide a firm theoretical foundation for my proposed test. In the process, the leading ECHR authorities on religiously offensive speech and racial hate speech will be considered. It will emerge that, within the Article 10 jurisprudence, forms of expression that are critical of religion have been treated more favourably where journalistic or scholarly, as opposed to creative or artistic, in nature. This is a product of the ‘gratuitously offensive’ test introduced in Otto-Preminger Institut v Austria. What is more, the Strasbourg Court has approached issues of racial hate speech on an ad hoc basis, without any consistent test.

My proposed test is as follows: ‘Is the expression objectively capable of seriously undermining the target’s assurance as to a status of equal worth in the community, having regard to the target’s knowledge, the speaker’s power and the forum of the expression, at the time it is made?’ This is to be adjudged from the perspective of a reasonable member of the target group, rather than a member of the majority racial or religious group.

The individual elements of the proposed test require full explanation below, but it is worth stressing a few vital features at this stage. First, it is a unified test applicable in both the racial and religious contexts. Secondly, it is more pro-freedom of expression than the current ECHR approach, as reflected in the language ‘seriously undermining’ and the focus on the impact on the target’s status, as opposed to the Strasbourg Court’s current inquiry into the inherent offensiveness of the impugned speech through its ‘gratuitously offensive’ test. Thirdly, it specifically instructs judges to assess offensive expression from the perspective of a reasonable member of the target group and to positively cast aside their own perceptions. Finally, it specifies what surrounding circumstances are to be considered when evaluating the effect of the expression in question.

More generally, this article seeks to address head on the current political debate in Europe as to the acceptable extent of religious criticism. It draws on the racial hate speech parallel to offer a relatively objective approach to identifying situations in which religiously offensive speech is so much of an affront to an individual’s personal status in the relevant community that it can be properly proscribed as ‘religious hate speech’. It should be noted that the broader language of racially and religiously offensive or hateful speech will generally be used throughout this article when discussing past cases.

Two points about the scope of this article need to be made. First, I will not squarely consider gender and LGBT hate speech. Although Waldron’s account and my proposed test could possibly be extended to these further contexts, there is a dearth of ECHR case law on these types of offensive speech, the academic literature is still nascent and gender and LGBT hate speech no doubt involve additional complexities. Secondly, my proposed test focuses on negative obligations on the part of States Parties to the Convention to refrain from infringing on the right to freedom of expression pursuant to their ECHR obligations, not positive obligations on their part to introduce hate speech regulation. I am concerned with the Strasbourg Court and national courts monitoring the extent of hate speech regulation, where it has been introduced.

Section 2 briefly explores rival free speech theories and provides a concise summary of the existing Article 10 jurisprudence on religiously offensive speech and racial hate speech. Section 3 presents the main features of Waldron’s account of the harm in hate speech in detail and seeks to defend that account within the European context. Section 4 fully explains the proposed test building on Waldron’s account and applies it to a range of past and hypothetical cases. Section 5 considers challenges arising from the supervisory nature of the Strasbourg Court’s jurisdiction and the national application of the proposed test. Finally, Section 6 concludes this article.

Free Speech Theories and the Article 10 Jurisprudence

Free Speech Theories

Free speech theory is a highly contentious area. Several competing accounts of the values underpinning the right to freedom of expression have been advanced. The principal theoretical accounts of free speech are those focusing on individual autonomy, political participation, validation of different ways of life and free competition of ideas. This is not the place to explore any of these general theories in any great detail, but I will briefly outline each to demonstrate the multiplicity of views. Against that background, I submit that the Strasbourg Court has, unpersuasively and without much deliberation, adopted the narrow political participation view.

To begin with the individual autonomy account, as presented by Scanlon, this posits that any compelling theory of free speech must apply to ‘expression in general’, rather than appealing to any special category of rights like political rights or ‘the value to be attached to expression in some particular domain’, such as ‘artistic expression or the discussion of scientific ideas’. Scanlon then goes on to argue, on the premise of suspicion of State power, that citizens cannot be taken to have consented to the State drawing the boundaries of permissible free speech according to its own conception of value. It follows, on this view, that freedom of expression in its myriad forms is valuable because of its role in facilitating individual autonomy and self-fulfilment. Though helpful in recognising the variety of uses to which free speech can be put, this is a very broad and ultimately vague account of free speech.

A very different account of the value of free speech is provided by Meiklejohn. On his political participation account, presented primarily as an account of the proper scope of the First Amendment right to free speech in the United States of America, freedom of expression applies to ‘those activities of thought and communication by which we “govern.”‘ He justifies this highly instrumental account of free speech more fully on the grounds that ‘”the people need free speech” because they have decided… to govern themselves rather than to be governed by others’ and that for self-government ‘the judgement-making of the people must be self-educated in the ways of freedom.’ It is clear that Meiklejohn’s account focuses on the nexus between free speech and the capacity for proper participation in public discourse and affairs. This leads to free speech being cast as a narrower right than on the individual autonomy account. There is nothing per se objectionable about this political participation account, but it is fair to criticise Meiklejohn for failing to explain why non-political speech is not valuable, having rather convincingly explained the value of political speech.

The third account worth outlining is that of Raz, who posits that free speech serves to validate different ways of life. It is Raz’s contention that ‘ways of life which are portrayed and expressed are validated through their portrayal and expression’, while censorship amounts to ‘authoritative public condemnation of the way of life.’ Interestingly, though Raz’s account is also autonomy based like Scanlon’s individual self-fulfilment account, he presents the validation of different ways of life through freedom of expression as a ‘public good’ amenable to restrictions in the public interest. It follows that this account accepts greater qualifications to free speech than Scanlon’s account, while being broader in scope than that of Meiklejohn.

The final account which deserves specific mention is the ‘marketplace of ideas’ theory. On this view, free speech is valuable because of its role in finding the truth through the free competition of ideas in the public domain. This theory is opposed to ex ante determinations of the permissibility of speech based on the respective values of speech, generally allowing for any speech to enter the public sphere. It is certainly contestable on empirical grounds, but for present purposes it helps further support my claim that there is a genuine lack of consensus, and indeed significant disagreement, in the free speech literature.

Overall, I submit that this lack of consensus points to the need for courts in Europe to be cautious about favouring any single account of free speech when deciding on the scope of free speech protection. It is alarming, therefore, that the Strasbourg Court appears to have adopted the political participation account of free speech in relation to Article 10 offensive speech cases. In the leading case of Otto-Preminger Institut v Austria, the Court spoke of ‘an obligation to avoid as far as possible expressions that are gratuitously offensive to others and thus an infringement of their rights, and which therefore do not contribute to any form of public debate capable of furthering progress in human affairs.’ This is plainly an instrumental test, seeking to identify permissible expressions by reference to their contribution to public debate. There may be some textual support for the Strasbourg Court’s focus on the relationship between freedom of expression, as protected by Article 10 ECHR, and the overarching role of the Convention as an instrument for democracy. However, there is no proper analysis of the theoretical underpinnings of free speech in the leading ECHR authorities on religiously offensive speech and racial hate speech. Thus, in my submission, the Strasbourg Court’s narrow focus on freedom of expression as an end to political participation is in need of review.

Religiously Offensive Speech Cases

I will now consider the leading ECHR authorities on religiously offensive speech. There has been a firm tendency in the Article 10 jurisprudence post-Otto-Preminger Institut to frame the central question as one of whether the expression in question is ‘gratuitously offensive’. A further important general observation advanced is that expressions critical of religion seem to have been treated more favourably where journalistic or scholarly, rather than creative or artistic, in nature. This is a product of the ‘gratuitously offensive’ test. The said difference in treatment by the Strasbourg Court will be made apparent by contrasting the adverse rulings in Otto-Preminger InstitutWingrove v United Kingdom and IA v Turkey to the favourable rulings in Giniewski v France and Klein v Slovakia.

In Otto-Preminger Institut, the Strasbourg Court had to consider whether the seizure and forfeiture by the Austrian authorities of a religiously satirical film called ‘Das Liebeskonzil’ violated the Article 10 ECHR right to freedom of expression of Innsbruck cinema proprietors. The film was to be screened to a select audience, viewed only by adults and marketed discreetly to the general public. The main object of concern was the unflattering depiction of the Christian God, Jesus Christ and the Virgin Mary as well as their portrayal as colluding with the Devil. The Strasbourg Court concluded that the Austrian authorities were justified in intervening ‘to prevent that some people should feel the object of attacks on their religious beliefs in an unwarranted and offensive manner’, given that Austria is a majority Roman Catholic country.

Shortly afterwards, in Wingrove, the Strasbourg Court had to deal with a case in which the British authorities had refused to grant the maker of a video called ‘Visions of Ecstasy’ a classification certificate for being blasphemous. The video depicted the Christian St Teresa of Avila in a sexually suggestive and revealing manner as well as insinuating that Jesus Christ formed part of her sexual desires. To support its conclusion that the British authorities were justified in banning the video, the Strasbourg Court reasoned that ‘no attempt was made in the film to explore the meaning of the imagery beyond engaging the viewer in a “voyeuristic erotic experience”.’ The Court also shared the British authorities’ concern as to the extent of dissemination, although this was to be done only through video cassettes, not digitally.

The IA case further evidences the Strasbourg Court’s strict treatment of creative or artistic religiously offensive speech. In that case, the Turkish authorities banned the publication of a novel concerning Islam called ‘The Forbidden Phrases’ and imposed a criminal fine on its publisher. The Turkish authorities objected to the novel for containing ‘statements that imply a certain element of humiliation, scorn and discredit vis-à-vis religion, the Prophet and belief in God according to Islam’. The novel attracted particular criticism for a passage stating that ‘God’s messenger broke his fast through sexual intercourse, after dinner and before prayer. Muhammad did not forbid sexual intercourse with a dead person or a live animal.’ In a short judgment, the Strasbourg Court again sided with the national authorities, holding that Article 10 of the ECHR had not been violated since the novel amounts to ‘an abusive attack on the Prophet of Islam’, rather than a permissible ‘provocative opinion’.

These authorities must be contrasted with the Strasbourg Court’s approach in Giniewski. That case concerned the legality of defamation proceedings in France brought against the writer of a newspaper article discussing the role of the Roman Catholic doctrine of fulfilment in contributing to anti-Semitism and the Holocaust. A passage in the article strongly relied on by the prosecutors stated that ‘[m]any Christians have acknowledged that anti-Judaism and the doctrine of the ‘fulfilment’ of the Old Covenant in the New [Testament] lead to anti-Semitism and prepared the ground in which the idea and implementation of Auschwitz took seed.’ In finding an Article 10 violation, the Strasbourg Court emphasised that ‘[t]he applicant’s statements contribute to a recurrent debate of ideas between historians, theologians and religious authorities’. The Court further opined that the article ‘had made a contribution … to a wide-ranging and ongoing debate, without sparking off any controversy that was gratuitous or detached from the reality of contemporary thought.’

Much like in Giniewski, the Strasbourg Court again gave favourable treatment to a newspaper article in Klein. That article criticised Slovakian Archbishop Ján Sokol for his role in undermining the separation of Church and State in Slovakia in relation to public protests surrounding the screening of a religiously provocative film. The writer went as far as saying that Slovakians ought to leave the Roman Catholic Church in Slovakia, which he described as being headed by an ‘ogre’. The Strasbourg Court held that the criminal conviction of the writer for the publication of this article violated Article 10 of the ECHR. This conclusion rested primarily on the finding that the ‘strongly worded pejorative opinion related exclusively to the person of a high representative of the Catholic Church in Slovakia’, instead of being an attempt to discredit and disparage Roman Catholics in the country ‘on account of their Catholic faith’.

Overall, the expressions in Giniewski and, to a lesser extent, in Klein were of a more sophisticated character than the creative or artistic ones in Otto-Preminger InstitutWingrove and IA. In the former type of case, it is easier for the Strasbourg Court to identify elements of the expression which engage the public interest and qualify as offensive ones. That is challenging when dealing with creative or artistic expression, where offensive elements are often unqualified and presented without context. I submit that this apparent distinction between different forms of religiously offensive expression is objectionable in that it does not properly distinguish between the relative harm to the target, becoming overly interventionist at times. A harm-based account is developed in Sections 3 and 4 below.

Finally, the very recent decision in ES v Austria deserves comment. Importantly, the Strasbourg Court reaffirmed the Otto-Preminger Institut test for religiously offensive speech. ES concerned an Article 10 ECHR challenge to a criminal conviction for the offence of disparaging religious doctrines, contrary to Article 188 of the Austrian Criminal Code. The applicant had given several publicly advertised seminars promoted by the Freedom Party to a far-right audience, in which she described the Prophet Muhammad as having had paedophilic tendencies, referring to the marriage with Aisha, when aged six, consummated at age nine. As to the applicable principles, the Strasbourg Court held that the criminal conviction served the legitimate aim of protecting religious peace, and reiterated its ‘gratuitously offensive’ test, observing that it is legitimate for a State to prohibit ‘an improper or even abusive attack on an object of religious veneration’. Applying those principles, the Court considered that there was a wide margin of appreciation due to the sensitivity of the impugned statements and held that the criminal conviction did not exceed Austria’s wide margin so as to breach Article 10 of the ECHR. The Court stressed the wide audience targeted by the seminars, agreed with the domestic courts that ‘the applicant must have been aware that her statements were partly based on untrue facts and apt to arouse (justified) indignation’, and rejected the argument that this attack on the Prophet had to be tolerated because it formed part of ‘a lively discussion’ containing other factually verifiable statements. In my submission, the limited analysis of contextual factors in ES highlights the need for a more carefully calibrated approach by the Strasbourg Court that focuses on the likelihood of objective harm from speech in the form of social exclusion.

Racial Hate Speech Cases

Unlike with religiously offensive speech, where the Strasbourg Court has consistently applied the ‘gratuitously offensive’ test, there does not seem to be a clear general approach to racial hate speech cases within the scope of Article 10 of the ECHR. The Court has approached racial hate speech on a case-by-case basis, considering mainly the nature of the offensive or hateful speech in question and the context in which it was made. I will deal primarily with the journalistic exposé of Danish working class racism in Jersild v Denmark and virulent pro-Kurdish political speech in Ceylan v Turkey. I will also highlight the exceptional treatment of anti-Semitic speech in Witzsch v Germany. Brief consideration will also be given to the LGBT hate speech case of Vejdeland v Sweden, for it confirms my reading that the Strasbourg Court has adopted an ad hoc approach to non-religious cases.

In Jersild, the Strasbourg Court had to consider whether the Danish authorities were permitted to bring a criminal prosecution against the journalist behind a TV documentary that featured hateful xenophobic and racist comments by members of a group called the ‘Greenjackets’. In holding that the journalist’s Article 10 right had been violated, the Court stressed that he ‘did not make the objectionable statements himself’ and that his purpose in broadcasting footage from his interviews with members of the Greenjackets was to expose a social problem of ‘great public concern’, namely ‘the propagation of racist views and ideas’ amongst deprived Danes. Furthermore, the Court held that the broadcast, though clearly offensive, was part of the ‘contribution of the press to discussion of matters of public interest’ and ‘particularly strong reasons’ for interference in press freedom were needed, which were not present on the facts.

Similar considerations arose in Ceylan concerning the conviction of a trade union leader in Turkey for the offence of incitement to hatred and hostility. He had written a newspaper article criticising Turkey for alleged atrocities committed against the Kurdish people. The Strasbourg Court observed that the article was ‘virulent and the criticism of the Turkish authorities’ actions in the relevant part of the country acerbic, as demonstrated by the use of the words “State terrorism” and “genocide”.’ However, the Court stressed the limited scope ‘for restrictions on political speech or on debate on matters of public interest’ and that ‘the article in question, despite its virulence, does not encourage the use of violence or armed resistance or insurrection.’ On these grounds, Turkey was found to have violated the applicant’s Article 10 right to freedom of expression.

It could be said, on analysis, that Jersild and Ceylan are not typical cases of racial hate speech, given the political dimensions to the speech in both cases. That said, I submit that they support my claim that the Strasbourg Court prefers an ad hoc analysis of these issues, relying strongly on the importance of a free press in Jersild and on the need for political speech in Ceylan. This is further confirmed by the Strasbourg Court’s exceptional treatment of anti-Semitic speech in Witzsch. That case concerned a published reply to a newspaper article in which the responsibility of the Nazis for the Holocaust had been reiterated. The reply in question denied the Nazi Party’s responsibility for the Holocaust, for which the applicant was given a criminal conviction in Germany. The Strasbourg Court refused to confer on this speech the usual two-stage treatment under Article 10 of the ECHR. Instead, it chose to apply Article 17 of the ECHR, which prohibits attempts to use ECHR rights for ‘the destruction of any of the [ECHR] rights and freedoms’. The conviction was upheld accordingly. The Court’s reasoning in Witzsch was sparse, but it may be inferred that it was animated by the unique background to anti-Semitic expressions in Europe in light of the horrors of World War II and the continued legacy of the Holocaust, particularly within modern Germany.

Lastly, I turn to Vejdeland as a further illustration of the Strasbourg Court’s treatment of non-religiously offensive speech. In that case, about 100 homophobic leaflets were distributed to pupils in a Swedish secondary school without permission. The leaflets included allegations as to the ‘sexual proclivity’ of LGBT people, the transmission of HIV/AIDS and links to paedophilia. The Strasbourg Court found that Article 10 of the ECHR could not be invoked by the leaflet distributors to challenge their criminal convictions. The Court was divided between the majority who considered the speech itself to be sufficiently hateful and concurring judges who felt that the special school context was decisive. Leaving aside this added complexity, Vejdeland offers further confirmation of the Strasbourg Court’s preference for a context-specific analysis of offensive or hateful speech outside the religious context.

Defending Waldron’s Account of the Harm in Hate Speech

Fundamental Features of Waldron’s Account

It will now be argued that Waldron offers a convincing theoretical account of the general acceptability and proper scope of hate speech regulation. This will be done primarily by presenting the key features of Waldron’s account and explaining how they interrelate to form a coherent theory.

Waldron argues that hate speech regulation can serve two specific aims. First, it can help secure the ‘public good of inclusiveness’. As he puts it: ‘[e]ach group must accept that the society is not just for them; but it is for them too, along with all of the others…. Hate speech undermines this public good’. Secondly, hate speech regulation can provide all members of society an ‘assurance’ of their equal dignity. Waldron posits that ‘[i]n a sense, we are all supposed to benefit. But for the members of vulnerable minorities … the assurance offers a confirmation of their membership.’ In short, the first aim presents ‘inclusiveness’ as something that society as a whole should value on the twin premises that liberal pluralism and multiculturalism in modern Western societies are desirable, while the second aim focuses more on the individual needs of ‘members of vulnerable minorities’ to be reassured of their equal status in society.

It is mainly in relation to the second aim of hate speech regulation that Waldron uses the language of preserving every person’s ‘dignity’, defined by him as ‘basic social standing’. As Waldron rightly points out, absent such reassurance as to equal status, vulnerable individuals and communities face ‘derision, hostility and abuse encouraged by hate propaganda … [which can] have a severely negative impact on the individual’s sense of self-worth and acceptance.’ The effects of that erosion of individual self-worth and assurance of acceptance in society can lead to victims taking ‘drastic measures in reaction, perhaps avoiding activities which bring them in contact with non-group members or adopting attitudes and postures directed towards blending in with the majority’. In this regard, it is important to recognise the subtextual meanings that hostile messages, signs or slurs can convey to their targets.

I will refer to Waldron’s account as the ‘dignitarian’ approach, for it is this second aim of hate speech regulation that will prove more influential in my practical reformulation of Waldron’s account in the ECHR context. The first aim, focusing on the public good of inclusiveness, is clearly relevant to law-makers and public authorities when deliberating on the need for and scope of hate speech regulation. However, from a court perspective, a focus on the impact on individual dignity in Waldron’s sense provides a more manageable boundary than focusing on public interest grounds against hate speech. Throughout this article, the term dignity is used in this specific and limited sense, in contrast to the pervasive use of dignity as a wider concept in contemporary human rights, equality and moral philosophical discourse.

As for the practical application to hate speech regulation, Waldron proposes a distinction between ‘undermining a person’s dignity’ (in which case hate speech regulation is permissible) and ‘causing offense to that same individual’ (in which case hate speech regulation is impermissible). He says this broadly maps onto a distinction between ‘objective or social aspects of a person’s standing in society’ and ‘subjective aspects of feeling, including hurt, shock, and anger’, respectively. To elaborate, for Waldron ‘[a] person’s dignity or reputation has to do with how things are with respect to them in society, not with how things feel to them’. By contrast, offence is presented as ‘inherently a subjective reaction’, so laws against offence would protect people from ‘a certain sort of effect on their feelings’.

It is submitted that Waldron’s focus on dignity as a social status, as opposed to the broader phenomenon of offence, helpfully narrows down the scope of permissible hate speech regulation. The objective/subjective distinction goes some way in legitimately confining the reach of hate speech regulation in recognition of the competing accounts of the value of free speech explored in Section 2 above. That said, Waldron is right to acknowledge that an analysis focusing on ‘dignity’ in the sense of an equal status in society cannot entirely disregard the reality that affronts to dignity will usually also entail a subjective impact on the target’s feelings independent of the objective impact on their equal status.

In any event, Waldron adds further useful guidance on distinguishing between objective and subjective impacts of offensive speech, stressing the need for the speech to be ‘capable of becoming a permanent … feature of the environment in which people have to live their lives.’ This rightly leaves open the possibility of prohibiting racial or religious slurs much like hateful signs or broadcasts, provided those slurs amount to an affront to equal membership in society in a relatively permanent sense. Permanence is useful as a benchmark for objective harm. Further, it is worth stressing that Waldron admits that his approach requires legal refinement. This is the main contribution made by this article in Sections 4 and 5 below. For the moment, though, if one does regard multiculturalism, pluralism and the avoidance of societal ostracisation as worthwhile aims, an approach seeking to delineate the boundaries of hate speech regulation along those lines ought to be appealing.

Moving on to consider support in authority, similar arguments in defence of hate speech regulation were deployed by the majority of the Canadian Supreme Court in R v Keegstra. That case concerned the legality of charging a high school teacher under the Canadian Criminal Code for wilfully promoting hatred by communicating anti-Semitic statements to his students. The majority of the Court, led by Dickson CJ, reasoned as follows:

The harm caused by this message run directly counter to the values central to a free and democratic society, and in restricting the promotion of hatred Parliament is therefore seeking to bolster the notion of mutual respect necessary in a nation which venerates the equality of all persons.

Further, Dickson CJ stressed that a particular evil of hate speech is that it can have ‘a severely negative impact on the individual’s sense of self-worth and acceptance’. Both these central claims used by the Canadian Supreme Court (LaForest, Sopinka and McLachlin JJ dissenting) to uphold the legality of the criminal charge in question map nicely onto Waldron’s two key claims about inclusiveness as a public good and the need for assurance of an equal status in society, particularly for vulnerable minority communities. Indeed, Waldron relied on Dickson CJ’s judgment in Keegstra regarding the adverse individual impact of hate speech. The Canadian Supreme Court’s general acceptance of an approach resembling Waldron’s dignitarian account highlights its acceptability in liberal democratic systems and scope for introduction in the European context. What is more, the Canadian parallel is quite relevant to Europe, for we share the racial and religious diversity of Canada, at least in Western Europe, and a fundamental commitment to pluralism. I will explain in Section 3E below how the general scheme of Article 10 of the ECHR permits the introduction of a legal test based on Waldron’s theoretical account, having particular regard to the wide range of permissible qualifications to freedom of expression enumerated in Article 10 of the ECHR.

A more difficult question is why one should accept the two aims put forward by Waldron at all. I have attempted to present concisely Waldron’s own defence of the ideas of inclusiveness as a public good and dignity as a status. I concur with Waldron that inclusiveness can be described as a public good, because it makes society more accommodating and welcoming for everyone to live in, both majority and minority communities. Moreover, I am in agreement with Waldron’s claims that hate speech regulation is justified in going beyond limited protection from race-based or religiously-motivated violence, or the customary protection of anti-discrimination laws in the workplace and in public service provision.

My essential point is that it is not sufficient for individuals to be formally equal under the law, yet susceptible to verbal or written attacks which call into question their substantive equality in democratic societies committed to the rule of law. It is readily conceivable, as Post argues, that hate speech laws are not in fact the best vehicle for the prevention of racial and religious hatred, violence and discrimination. However, hate speech laws seeking to prevent harm to individual dignity, as a social status, can tangibly contribute to greater substantive equality and inclusiveness. In any event, perhaps more importantly, they can serve to reassure and reaffirm an individual’s equal assurance of dignity. This can, of course, be supplemented by wider policy measures aimed at curbing manifestations of hatred and intolerance.

Waldron’s Dignitarian Account and Mill’s Harm Principle

The connection between Waldron’s objective/subjective impact distinction and long-standing criminal law and philosophical discussions about the respective roles of harm and offence in justifying legal intervention will now be briefly explored. The harm principle famously put forward by Mill as the principal philosophical justification for the imposition of criminal liability or punishment could be said to buttress the focus on individual and direct harm under Waldron’s account and my proposed test.

As Mill put it: ‘The only purpose for which power can be rightfully exercised over any member of a civilised community, against his will, is to prevent harm to others.’ By contrast, Feinberg’s rival offence principle, which is more intrusive on individual liberty, seems to animate the Strasbourg Court’s ‘gratuitously offensive’ test. In essence, Feinberg’s view is that ‘it is always a good reason in support of a proposed criminal prohibition that it would probably be an effective way of preventing serious offense… and that it is probably a necessary means to that end’. However, even Feinberg had reservations about offence as a basis for censorship and criminal sanctions of offensive speech, rather than offensive actions.

In general, the distinction under Waldron’s account and my proposed test could be said to be one between a permanent emotional or partly physical impact on the target, on the one hand, and a temporary emotional response of indignation or discontent on their part, on the other hand. Loosely speaking, the harm principle only covers the former impact, whereas the offence principle also covers the latter impact, so the harm principle supports fewer instances of State interference with freedom of expression. As such, the harm principle is more respectful of the importance of freedom of expression in free, democratic European societies. It is not claimed that Waldron’s account or my proposed test neatly map onto the harm principle as traditionally theorised, but the analogy to this older body of scholarship may nonetheless prove helpful. This parallel also shows that a focus on individual harm is by no means novel when setting the boundaries of legal regulation and criminalisation.

Extension of Waldron’s Account to Religious Hate Speech

It is proposed that Waldron’s theoretical account can be extended to certain instances of religiously offensive or hateful speech, even though his account deals primarily with racial hate speech. Applying the proposed test built on Waldron’s account to both racial and religious speech requires one to accept one important but highly contestable assumption. This assumption is that, while there is greater voluntariness in the choice of one’s religion than in the case of race or ethnicity, most religious adherents nonetheless reasonably view themselves as having limited choice in their religious affiliation.

The assumption on which my extension of Waldron’s account from the racial context to the religious context is based is one of relative absence of choice in religious matters. This assumption is not based on any firm empirical foundation, but rather on the long-standing, observable reality of children being raised by their parents to follow a specific religion or faith from a young age, meaning that come adulthood their religion is often as much a part of their core identity as their race or ethnicity. Moreover, familial and social pressures to continue following or practising that religion can be strong. Consistently with this, Article 18 of the International Covenant on Civil and Political Rights (ICCPR) requires ‘respect for the liberty of parents … to ensure the religious and moral education of their children in conformity with their own convictions’. Similar provision is made in Article 2 of Protocol 1 to the ECHR and Articles 14 and 18 of the Convention on the Rights of the Child 1989. I make no comment on the merits of this social reality and associated legal provisions, but put them forward as the premise of my scepticism for the race/religion dichotomy in existing free speech literature and the ECHR jurisprudence. Waldron does deal with how his account could be applied to religiously offensive speech, yet focuses more on its application in the racial context. This article seeks to plug that gap, given the relevance of his theory to difficult cases of religiously offensive speech.

An alternative assumption pursuant to which the extension of Waldron’s dignitarian approach to the religious context could be defended is that certain religious groups face a degree of social disadvantage or exclusion comparable to that of vulnerable racial groups, warranting protection in the form of hate speech regulation. However, I prefer to proceed on the relative absence of choice assumption, because this avoids further distinctions between different racial and religious groups. In short, on my account, members of any racial or religious group could, in theory, find themselves at the receiving end of racial or religious hate speech. By contrast, it appears only vulnerable or minority racial and religious groups could be in that position on the alternative degree of social disadvantage or exclusion assumption. To my mind, this would be an unfairly narrow assumption, for it is conceivable that any person in Europe could become the victim of racial or religious hate speech in certain circumstances.

It could be said that the proposed extension of Waldron’s account to the religious context brings us into the murky territory of ‘identity politics’. Waldron forcefully argues that some care must be taken to keep apart legal discourse in the form of hate speech regulation and wider political discourse concerning respect for various different identities. He reasons that identity politics can be misused as a political device to elevate certain beliefs or ways of life to a quasi-entrenched status through the language of fundamental rights, thus detracting from ordinary political and public deliberation.

Admittedly, the difficulty with religion is that there is some element of choice in following a particular religion, whereas one simply has no control whatsoever over one’s racial or ethnic origins and the possible contempt those origins may evoke from others. In that limited sense, it could be claimed that special regard for religion within the law facilitates potentially damaging identity politics. That said, when applied to religion, the identity politics objection is unconvincing for three main reasons. First, special constitutional status has long been accorded to religion in most societies, as Waldron himself accepts in distinguishing religious issues from more multifaceted matters of identity politics. In the European context, this is reflected in Article 9 of the ECHR and national constitutions across Council of Europe Member States. Secondly, the argument, made by Letsas for instance, that an interest in practising one’s religion is equivalent to any other interest relating to the way one wishes to live one’s life is difficult to sustain historically and at odds with the elevated importance that many people in Europe still ascribe to their religion. Thirdly, though Waldron and other commentators like Reichman are right to highlight the complexities of hate speech regulation in the religious context, this is the most practically important area for legal consideration. Thus, our efforts should go into framing a test proscribing hateful speech that properly distinguishes petty identity politics from seriously harmful religious hate propaganda.

A final salient point from Waldron concerning the religious context is that care must be taken in distinguishing between individual religious identity and group religious identity. This is an important distinction to bear in mind when considering the specifics of my proposed test in Section 4. Religion arguably has a greater individual dimension than race or ethnicity in the sense that there is latitude for religious adherents to pursue their faith in different ways. It follows that protection against offensive or hateful speech should be more forthcoming in the racial context, since objectively harmful affronts to the target’s equal status in society are more probable. By contrast, in the religious context, there must necessarily be more scope for disagreement and offence flowing from the exercise of the right to freedom of expression, given the greater diversity of religious views and practices.

Harm-Based Approach vs. Intention-Based Approach

It might be thought that it would be preferable to focus on the subjective intention with which the perpetrator engages in offensive or hateful speech, rather than the likely objective harm to the target. To illustrate this point, for example, one might have a case in which, in the UK, a White Christian makes a very religiously offensive remark to a Pakistani Christian in the mistaken belief that the target is in fact a Muslim. On an intention-based approach, this factual error on the part of the perpetrator is inconsequential, as the legality of sanctions for the offensive speech would depend on whether some threshold of seriousness, adjudged by reference to the contemptibility of the perpetrator’s ascertainable intentions, has been met.

By contrast, applying my preferred harm-based approach, the analysis might be more difficult given the mistaken belief of the perpetrator. On my approach, following Waldron, the general question would be whether, despite being a Pakistani Christian, the expression was capable of being sufficiently harmful by amounting to an affront to the target’s equal status in society. This complication should not be overstated, though. I submit that in the majority of mistaken identity cases, the incorrect categorisation of the target coupled with the offensiveness of the expression would still meet the proposed threshold. To explain, it is unlikely that the Pakistani Christian would not feel genuinely threatened or permanently shaken by a misplaced religiously hateful remark by a member of the majority community. If anything, the risk of incorrect categorisation by the majority could be even more debilitating for potential targets.

As for the more fundamental criticism that an intention-based approach is superior, there are several reasons against focusing on the perpetrator’s state of mind. First, there is the risk of under-inclusiveness when dealing with ignorant or insensitive individuals. There is no compelling reason why ignorant hate speakers should be given more lenient treatment than ones who deliberately disseminate such divisive and harmful expression, where the likely harm to the target is equivalent or even more serious. Secondly, the harm-based approach is more cognisant of the vulnerable and disadvantaged position of most targets of hate speech. It is rightly less preoccupied with being charitable to perpetrators who tend to be members of the more powerful majority community. Thirdly, it avoids a fraught subjective inquiry into the mental state of the perpetrator, with the harm-based approach concerning itself with objective manifestations of hatred which will normally be readily identifiable and open to evaluation. In the context of European supervision, this added objectivity is particularly important, for reviewing subjective assessments by national courts would pose grave difficulties and justify frequent reliance by the Strasbourg Court on the margin of appreciation.

Acceptability of Viewpoint-Based Restrictions

To conclude this Section, I turn to the question of whether hate speech regulation governed by Waldron’s dignitarian approach is justified in going against the notion of an unfettered marketplace of ideas. Richards argues that classifying some offensive speech as hateful and therefore unprotected, while other offensive speech is upheld as protected, amounts to highly contestable viewpoint discrimination. His overall thesis is that curtailments of free speech in fact compromise well-intentioned efforts to halt unjust discrimination on racial, religious and other grounds, justifying a blanket position against hate speech regulation, in his opinion.

The marketplace of ideas approach is evidently out of the line with the European approach to freedom of expression, which clearly recognises that some restrictions will be justified. It will be recalled that Article 10 of the ECHR is quite detailed compared to the other qualified ECHR rights, providing for public interest restrictions on grounds including ‘national security’, ‘public safety’, ‘the prevention of disorder or crime’, ‘the protection of health or morals’ and ‘the protection of the reputation or rights of others’. Moreover, Richards’s position is plainly out of line with the Strasbourg Court’s application of Article 10 of the ECHR in cases concerning racial hate speech and religiously offensive speech, as outlined in Section 2. Indeed, Richards’s position is even more pro-free speech than the one suggested in this article. His position is more legally acceptable in the United States context, because the U.S. Supreme Court has repeatedly refrained from allowing viewpoint-based restrictions to the First Amendment constitutional right to free speech. Leading recent U.S. Supreme Court cases illustrate the gulf between the U.S. and European approaches to permissible restrictions on free speech.

The simple argument, popular in the United States, that viewpoint-based restrictions on freedom of expression are generally unacceptable, commands neither textual nor authoritative support in the ECHR context. Moreover, the European approach, viewing freedom of expression as a qualified right, is in line with well-regarded Commonwealth constitutions and the international law position. For instance, the 1996 South African Constitution contains a clear prohibition of ‘advocacy of hatred’ in Section 16 (c), which has the effect that no limitation analysis is required in such cases. This allows the South African courts to respond to modern iterations of grave past racial issues, principally the brutal legacy of Apartheid in South Africa, much like the Strasbourg Court’s application of Article 17 of the ECHR in the context of anti-Semitic speech in Witzsch. Further, in South Africa, even where the strict definition of Section 16 (c) is not met, freedom of expression is in any event subject to the general Section 36 limitation clause. Similarly, the right to freedom of expression under Section 2(b) of the Canadian Charter of Rights and Freedoms 1982 is subject to a general limitation clause (in Section 1).

The similarities between the European position and these leading liberal democratic jurisdictions are compelling, because Canada shares the racial and religious diversity seen in large parts of Europe, whilst South Africa shares the dark legacy of extreme racial subjugation. The parallels between the Holocaust in Europe and Apartheid in South Africa are chilling. Moreover, in international law, Article 20 of the ICCPR requires legislation against hate speech, while Article 4 of the International Convention on the Elimination of All Forms of Racial Discrimination contains a prohibition of hate speech.

In sum, the above analysis suggests that there is a consensus in liberal democracies outside the United States and at international law level about the permissibility of viewpoint-based restrictions on free speech. Further, though there may be some theoretical force in Richards’s position, freedom of expression has never been regarded as an unabridged right in post-war Europe, having regard to the structure of Article 10 of the ECHR itself. Thus, some degree of viewpoint-based restrictions seems uncontentious. In any event, particularly grave historical consequences of unfettered freedom of expression in Europe must be borne in mind before objecting to any censorship of unpalatable viewpoints. The European debate must therefore focus on the extent to which restrictions are justifiable, as this article seeks to do.

Proposed Test for Racial and Religious Hate Speech

Proposed Test in Outline

As set out in the Introduction, the courts would pose the following question when assessing the permissibility of hate speech regulation in respect of alleged racial or religious hate speech:

Is the expression objectively capable of seriously undermining the target’s assurance as to a status of equal worth in the community, having regard to the target’s knowledge, the speaker’s power and the forum of the expression, at the time it is made?

Crucially, this test is to be accompanied by the specific instruction to judges that it is to be applied from the perspective of a reasonable member of the target group, rather than a member of the majority racial or religious group in the relevant country. Looking at the generality with which the proposed test is formulated, some degree of case-by-case analysis will remain unavoidable. However, the proposed test seeks to provide fuller guidance to judges than the current ECHR approach.

In my submission, it is an improvement on the Strasbourg Court’s existing jurisprudence in the following ways. First, it abandons the unhelpful race/religion dichotomy and replaces it with a unified test to be applied in all offensive speech cases. Secondly, through the language of ‘seriously undermining’, it sets a relatively high threshold at which European States may lawfully restrict, prohibit or criminalise any expression. Thirdly, it specifically requires judges to assess offensive expression from the perspective of a reasonable member of the target group, which amounts to an instruction to cast aside their own perceptions and biases. Finally, it specifies an exhaustive list of surrounding circumstances which are to be considered when evaluating the effect of the offensive expression in question, making the judicial inquiry more focused and predictable. Moreover, it is worth observing that the proposed test focuses on how speech would reasonably be perceived ‘at the time it is made’. This limits the analysis to prevailing knowledge and circumstances, both out of fairness to the speaker and to confine the scope of the judicial inquiry to what is practicable. I note, for the avoidance of doubt, that, on my proposed approach, the manner in which an expression is made forms part of the primary analysis of the content or nature of the expression, rather than the secondary analysis of surrounding circumstances.

Explanation of Proposed Test

It will be useful to expand on each of the claimed advantages of the proposed test. I will begin with the merits of a unified test for religiously offensive speech and racial hate speech, departing from the two-track approach in the ECHR jurisprudence outlined in Section 2. First, my proposed test reacts to what may be described as the harm common to both racial and religious hate speech, namely an objective affront to the target’s dignity in the specific sense of an assurance as to a social status of equality. As explained in Section 3, when discussing the relative absence of choice in religious matters assumption, both religious affiliation and racial identity often form the bedrock of a person’s identity. If that premise is accepted, it is logical to seek to devise a single test responding to this form of harm. Secondly, using a uniform test at European level and by national courts, like the UK courts, applying the ECHR jurisprudence domestically, would be simpler and enhance legal certainty. This is particularly true when compared to the ad hoc analysis of the Strasbourg Court in racial hate speech cases like Jersild and Ceylan, discussed in Section 2.

Moving on to justifying the ‘seriously undermining’ threshold for harm under the proposed test, this responds to the apparent consensus in the theoretical literature on free speech and statements in several Strasbourg Court decisions that freedom of expression is not to be lightly interfered with by the State through criminal sanctions or censorship. By explicitly stipulating a high threshold for lawful State interference, through the language ‘seriously undermining’, the proposed test ascribes presumptive value both to the values underpinning freedom of expression, discussed in Section 2, and the individual harm flowing from hate speech, discussed in Section 3. It seeks to highlight to judges that great care must be taken before allowing European States to proscribe such speech, even if clearly offensive. By setting a relatively high threshold, special account is taken of the long history of free speech in Western Europe, in particular, and the risk of driving popular resentment against particular religious or racial groups underground through a climate of excessive legal interference and censorship of public discourse on these sensitive issues. In short, my proposed test preserves a legal safe space for critical expression in the face of frequently strong social pressures for political correctness and tactful discourse.

Turning to the importance of adjudging harm from the perspective of a reasonable member of the target group, this key aspect of the proposed test is designed to counteract the unrepresentative composition of most European courts. The proposed test will involve judges ascertaining, for example, what a notional Muslim, Christian, Hindu, Jewish, Sikh, Black, Asian, Arab or other ethnic or religious minority person would feel in the circumstances. This seeks to reduce the difficulties arising from the predominantly majority ethnic, affluent and male composition of the judiciary, both in the Strasbourg Court and in national courts across Europe. Even if it is difficult for judges to put themselves into the shoes of the notional reasonable member of the target community, it is beneficial to have judges specifically directed to approach the issue from that perspective, since they will be alerted to the need to disregard their likely biases, in many cases subliminally held ones.

A further point that requires explanation is why it is appropriate to focus on the likely impact of the impugned speech on a reasonable member of the target community, as opposed to the actual target. In agreement with Waldron, concerns about some individuals being overly sensitive to criticism of their religion, religious group or racial group make it necessary to move away from the immediate dynamic of the actual speaker and actual target. By focusing on the likely reaction of a reasonable member of the target community, the analysis becomes more objective and suited to the typically written evidence based nature of free speech challenges. However, it is accepted that this approach will not always yield straightforward answers and that judges will not always succeed in disregarding their own biases. Even so, the proposed approach is more conscientious of the complexities in this area and seeks to counteract them, which is to be contrasted to clear evidence of majoritarian biases in the current ECHR jurisprudence on religiously offensive speech.

The remaining novel feature of the proposed test is the detailed specification of relevant surrounding circumstances. This is to be contrasted to the rather unpredictable ad hoc approach of the Strasbourg Court in racial hate speech cases in particular. Plainly, any specification of legally relevant circumstances in a judicial test serves to enhance legal certainty. Moreover, my proposed test deliberately opts for an exhaustive list of relevant circumstances, for this avoids an ad hoc analysis through the backdoor. The harder question is why these particular circumstances warrant judicial consideration. I will attempt to provide a brief defence and illustration of each relevant consideration.

Starting with the ‘target’s knowledge’ consideration, this seeks to capture the important interpersonal dimension in any case of alleged hate speech. At a theoretical level, the idea is to allow the judge to consider the hate speech from the point of view of the actual target, though ultimately assessing its harmfulness objectively by reference to the likely reaction by a reasonable member of the target community. From a practical point of view, what was heard, seen or otherwise known by the target will provide the main evidential foundation for the judicial assessment of the compatibility of any censorship or restriction of the impugned speech by the relevant State or public authority with Article 10 of the ECHR. It will be observed that my proposed approach deliberately focuses on the target’s perspective, in line with my objections to an intention-based approach outlined in Section 3D above.

Next, the ‘speaker’s power’ consideration is meant to alert the judge to the likelihood of varying degrees of power imbalance. For example, a target will probably be more vulnerable, and the speaker more powerful in turn, when the target is accompanied by children or vastly outnumbered by people of the majority racial or religious group from which the hate speech emanates. Further, differences in age and the ratio of speakers to targets are likely to have a significant impact on the speaker’s power. In general, it is to be expected that the requisite objective, permanent harm, as opposed to mere transient offence, is more likely to be forthcoming when the power imbalance lies more strongly to the speaker’s advantage.

The third and final consideration, focusing on the ‘forum of the expression’, might appear to be similar to the second consideration in that the forum of speech can clearly have a significant impact on the speaker’s power. That said, it is concerned solely with spatial factors and how these may affect the speaker-target power dynamic. Thus, it would often be highly relevant, for example, whether the speech is made in an area where there is greater racial or religious homogeneity to the target’s disadvantage or whether the target is confronted with the hateful speech in an enclosed space which is more difficult to leave. Moreover, this consideration pertains to both physical and digital spaces, including print media, online publications and even chat forums. It must be stressed again that these examples are merely illustrative. Each case will require a careful judicial assessment of the facts, hopefully made simpler and more predictable by the specification of relevant surrounding circumstances.

Overall, my proposed test has two distinct advantages. First, the unhelpful dichotomisation of racial hate speech and religiously offensive speech in the Article 10 ECHR jurisprudence would be abandoned. Secondly, freedom of expression in Europe would be approached in a way that better differentiates between highly personal attacks (which should be more readily circumscribable) and general critiques of particular world views or ways of life (which should be less readily circumscribable). This makes for a more coherent approach to racially and religiously offensive or hateful speech, which would also be more responsive to the present political reality of tense public discourse on race and religion in Europe.

Practical Application—Danish Cartoons Affair

This article would be incomplete without proper consideration of the highly controversial Danish Cartoons Affair. My aim here is not to offer a comprehensive analysis of the specifics of that example. Rather, I seek to use it as a further concrete illustration of how my proposed approach is to be applied. By way of recapitulation, in September 2005, a series of cartoons depicting the Prophet Muhammad was published in the Danish newspaper Jyllands-Posten. There were 12 published drawings in total, three of which depicted the Prophet in a particularly controversial manner. To be specific, these three drawings featured a turban in the shape of an ignited bomb, allusions to the reward of virgins for jihadist martyrdom and armed bearded men. Legal challenges were brought by Muslim organisations, but these were dismissed by the Danish Public Prosecutor. Several European newspapers republished the images, stirring up a heated public debate across the continent. Statements highlighting the tension between respect for a free press and religious sensitivities were made by various European and international bodies, but the matter was never properly litigated in any court.

Against that factual backdrop, in my submission, the Strasbourg Court would probably have upheld any attempts by the Danish authorities to prohibit the publication under the current ‘gratuitously offensive’ test. This is because the clear link between the depiction of the Prophet Muhammad and Islamic extremist violence since the 9/11 attacks in some of the cartoons blatantly offends millions of Muslims across Europe, including in Denmark. Importantly, the three controversial images present the link between Islam as a religion and jihadist violence in a manner that leaves little room for nuanced debate about radicalisation or the effect of Islam on freedom of expression more generally. My focus is on the controversial images within the publication themselves, rather than the long-standing public discourse that they generated through their polarising effect. In any event, focusing more closely on the existing approach in previous Strasbourg decisions, in agreement with Nathwani, ‘[t]here is no reason to believe that the connotation that the Prophet Muhammad favors violence of terrorists, or is part of terrorist activity, is less offensive to Muslims than the connotation [considered in Wingrove] that Jesus Christ engaged in sexual acts is to Christians.’

As for the treatment of the Danish Cartoons example under my proposed test, I submit that, on balance, the Strasbourg Court or a national court applying that test would be entitled to consider the three controversial cartoons undeserving of protection under Article 10 of the ECHR. In doing so, I disagree with Waldron’s tentative conclusion that, though a borderline case, this particular publication ought probably to be upheld. In agreement with Waldron, one should generally distinguish between highly personal attacks and general critiques of particular worldviews or ways of life. However, in my submission, the Danish Cartoons publication was too direct and explicit in its linkage of the Prophet Muhammad, representative of the global Muslim community, and Islamic extremist violence.

Applying the harm-centred approach defended in this article, I consider that, to the notional reasonable Muslim, the overt and unqualified connection between their Prophet and violence of the most extreme kind amounted to an affront to their status as equal members of the community. The images, in the aftermath of major jihadist attacks in Madrid and London preceding their publication, are objectively threatening and unwelcoming to Muslims across Europe. My view is to be contrasted to the firm pro-publication view of Cram, who claims that the heated public discourse surrounding the Danish Cartoons Affair demonstrates that legal intervention to protect religious minority groups in such cases is superfluous. It also differs from Nathwani’s more ambivalent conclusion that the case had become so global that the vulnerable minority status of Muslims in Denmark was compensated for by the strong global solidarity amongst Muslims in the wake of the publication. It has more in common with Cox’s scepticism of sweeping positions in favour of satirical images targeting religions and religious adherents, although I do not adopt the degree of self-identification with religion and historical disadvantage arguments underpinning his position.

To be clear, I am not advocating that it is necessary, under Article 10 of the ECHR, for non-believers to comply with the internal tenets of Islam or any other religion. That would be tantamount to a misguided endorsement of blasphemy laws. The flouting of the prohibition within Islam on depictions of the Prophet Muhammad does not in and of itself constitute sufficient harm to justify intervention in the form of hate speech regulation in the European context. It follows, staying with the Danish Cartoons example, that a subtler image of the Prophet Muhammad, without such an indiscriminate and widely alienating connection being drawn between the Muslim faith and Islamic extremist violence, would be permissible under my proposed approach. Put differently, my core objection to the Danish Cartoons is that they failed to make any real attempt at distinguishing an Islamic extremist minority from the much larger Muslim community.

By contrast, it is hard to see why the publication of any image of the Prophet Muhammad, without more, is seriously harmful to the assurance of equal dignity of Muslims across Europe, even if one accepts that Islam prohibits such depictions. The countervailing interests of cartoonists, humourists or journalists necessitate an approach which is not an unduly invasive approach to free speech protection in Europe. While my proposed approach still limits their room for manoeuvre in such sensitive areas, it is hoped that the competing interests are better reconciled, and that freedom of expression is more effectively preserved.

Practical Application—Past Article 10 ECHR Cases

The application of my proposed test to the leading past Article 10 ECHR cases on religiously offensive speech and racial hate speech will now be considered to help further illustrate its practical effects. Beginning with Otto-Preminger Institut and Wingrove, it is quite plain that neither instance of State censorship would have been permissible under my proposed test. In Otto-Preminger Institut, the Strasbourg Court placed undue emphasis on majoritarian sentiments in Austria and did not identify any significant harm to Christian believers beyond the possibility of offence flowing from the limited advertising and public discussion of the provocative content of the film. Likewise, in Wingrove, the Court was preoccupied with the failure of the video to properly justify its sexualised depiction of St Teresa of Avila with Jesus Christ, rather than highlighting any tangible evidence of the alienating effects of the video on even a small proportion of Christians in the UK. In both cases, this cannot amount to seriously undermining the target’s assurance as to a status of equal worth in the community.

I further submit that the IA case would also have been decided differently, because the Strasbourg Court’s concern about the degree of provocation to devout Muslims generated by the impugned novel overlooks the fact that Turkey has a large majority of practising Muslims, who cannot plausibly be said to face marginalisation through such inflammatory comments. By contrast, ES is a more finely balanced case, because the targets of the religiously offensive speech were in a majority Catholic country exhibiting increasing far-right tendencies. The Strasbourg Court’s analysis may be criticised for its limited engagement with contextual factors going to the likely impact of the statements in issue, yet the same outcome could have been reached on my proposed approach. The same is probably true of Giniewski and Klein, but due to the negligible personal harm to Roman Catholics in France and Slovakia, respectively, from these articles, not their instrumental value in contributing to public debate. From this brief review of religiously offensive speech authorities, it will be apparent that, in general, religious criticism will be more likely to meet the required threshold of serious harm under my proposed test where it is directed at specific individuals or genuinely vulnerable minority groups.

As for the past racial hate speech cases, Jersild is a relatively difficult case, since the racial slurs and wider derogatory comments broadcast on Danish television could conceivably cause the requisite affront to equal dignity on the part of ethnic minority viewers. However, having particular regard to both the ‘speaker’s power’ and ‘forum of the expression’ considerations, the Strasbourg Court’s emphasis on the progressive social purpose of the documentary and the critical counter-remarks by the journalist on the racial hate speech featured in the broadcast would most probably make this a permissible exercise of freedom of expression. With respect to the Ceylan case, it would probably have been decisive that the Kurdish population is itself a vulnerable minority in Turkey and, critically, that no violence or insurrection towards the majority population capable of subverting that general power balance was threatened, such that this newspaper article would have been protected on my approach. To consider Witzsch, finally, my proposed test would, like at present through Article 17 of the ECHR, offer very limited scope for such anti-Semitic speech. This is because, in agreement with the Strasbourg Court, the Jewish community across Europe remains highly vulnerable as a small minority that has been persecuted in the most extreme form in the recent past.

Practical Application—Indian Catholic Example

To conclude this section, I will briefly consider a further hypothetical example, which is a spin-off from an example used by Waldron. In my example, an Indian Catholic in the UK is confronted by a large sign, held by what appears to be a group of far-right British National Party (BNP) sympathisers. The sign says: ‘Indians—Don’t serve them, don’t speak to them and don’t let them in’. What is more, the target, a middle-aged man, is accompanied by his primary school aged daughter, walking down a quiet street in a strongly White majority working-class residential district in London. I will now explain why, in my submission, the UK authorities would be entitled to prohibit such a sign as unprotected racial hate speech.

First, the content of the sign evidently conveys a highly exclusionary message, particularly chilling for people of Indian origin and others who may be presumed to belong to that ethnic group. Further, to take each of my relevant circumstances in turn, the key knowledge on the part of the target in this instance is that the sign is held by far-right BNP sympathisers. It is possible that the target is in error about their actual identity, but the subjective assessment of the target forms the proper starting point for the judicial assessment. Next, the power of the speakers is clearly amplified by their strength in numbers and the added vulnerability of the target in being accompanied by his daughter, who is likely to be particularly shaken up by such an encounter. Finally, the forum of the expression heightens the objective likelihood of the expression seriously undermining the target’s assurance as to a status of equal worth in the community. This is because, in a majority White working-class residential district, there is less likely to be support for the target from other ethnic minority people and the BNP sympathisers are bound to feel more empowered. For those reasons, a court would be highly likely to uphold State regulation of such hateful signs notwithstanding Article 10 of the ECHR.

This example is also useful because it illustrates the important point that a significant portion of the population in European States will be members of both racial minority groups and religious minority groups. This increases the likelihood of them becoming victims of hate speech. Staying with the Indian Catholic example, the signs in question could conceivably have said ‘Catholics—Don’t serve them, don’t speak to them and don’t let them in’. This phenomenon of ‘intersectionality’, specifically of multiple vulnerable identities, must be given due weight when applying my proposed test. Hate speech regulation protecting such particularly vulnerable individuals ought to be all the more likely to comply with Article 10 of the ECHR.

Supervisory Jurisdiction and National Application

Role of Margin of Appreciation in Article 10 ECHR

The defence of my proposed test requires an explanation of how its application would be affected by the margin of appreciation doctrine at European level and domestic equivalents like the putative doctrine of deference in UK public law. The margin of appreciation doctrine plays a central role in Convention rights adjudication. It has been consistently invoked by the Strasbourg Court ever since its seminal decision in Handyside v United Kingdom, where the Court famously observed:

By reason of their direct and continuous contact with the vital forces of their countries, State authorities are in principle in a better position than the international judge to give an opinion on the exact content of these requirements as well as on the ‘necessity’ of a ‘restriction’ or ‘penalty’ intended to meet them… Nevertheless, Article 10 does not give the Contracting States an unlimited power of appreciation.

This general reasoning has been repeatedly invoked by the Strasbourg Court. The margin of appreciation doctrine is by no means uncontroversial. In many contexts, the width of the margin of appreciation will greatly impact the degree to which the Court scrutinises alleged Convention rights violations. Thus, it is rather troubling that, in religiously offensive speech cases, the Strasbourg Court has made contradictory observations about its width, first describing it as ‘not wide’ in Otto-Preminger Institut, only to proclaim that a ‘wide’ margin of appreciation is required to respect national sensitivities ever since Wingrove.

More fundamentally, some commentators regard the doctrine as altogether flawed. For instance, Letsas claims that the margin of appreciation doctrine has been unjustifiably used by the Strasbourg Court as a device to avoid difficult rulings on Convention rights violations. Similarly, Cram laments that ‘[t]he aspiration towards universal standards of human rights protections is compromised by the doctrine of the margin of appreciation.’ A notable defence is that of McGoldrick who argues that the Strasbourg Court’s approach to the margin of appreciation ‘leads to reason-based, justificatory arguments’ as to the appropriate intensity of supranational review, while correctly recognising the ‘significant’ weight that ought to be accorded to democratic decision-making at national level. Other commentators have also highlighted the mediating effect played by the doctrine.

Without seeking to resolve the finer points of this debate, in my submission the margin of appreciation doctrine is far too well-established to be disregarded now and, as appears from the general reasoning in decisions since Handyside, it is based on striking a compromise between effective supranational Convention rights review and respect for national autonomy in sensitive areas, at least where there is no European consensus on the issue. To my mind, this frequent compromise is animated by pragmatism and concerns regarding relative institutional competence on the part of the Strasbourg Court.

Role for Margin of Appreciation under Proposed Test

Turning to the application of the margin of appreciation doctrine in this particular context, I propose that it is consistent both with the general Strasbourg Court doctrine and my proposed test to draw the following distinction. Challenges to a general legal regime or framework, on the one hand, and challenges to individual applications of hate speech laws or regulations, on the other hand, ought to be distinguished.

It is submitted that a wider margin of appreciation is appropriate for challenges to a general legal regime or framework, because the impact on the freedom of the relevant European State to regulate freedom of expression in the interests of racial or religious toleration, or social harmony, is more seriously impacted. Specifically, it must be appreciated that domestic legislation or regulatory measures concerning hate speech may be the product of hard-won political compromise. General challenges to a national hate speech regime come with the possibility of the entire regime or ‘fundamental features’ thereof being declared incompatible or invalid. In that event, there is a real risk that an Article 10 ECHR compatible replacement regime or substantial amendments to the existing regime for Convention compliance cannot be achieved politically, leaving a hiatus in domestic hate speech regulation that may well do more harm than good in terms of preserving racial or religious toleration and social harmony.

By contrast, a narrower margin of appreciation is appropriate where the challenge is solely to an individual application of the relevant national law or regulation curtailing freedom of expression. This submission is based on the fact that, in such cases, the impact is primarily on the individual speaker(s) and the individual target(s). This, in turn, justifies a more straightforward application of the proposed test, with minimal alteration on margin of appreciation grounds. It is further noted that, if the proposed test were to be applied domestically, for instance in the UK given the limited common law constitutional rights jurisprudence, it is likely that domestic variants of the supranational margin of appreciation doctrine could impact the way national courts apply the proposed test. In the UK context, specifically, the putative doctrine of deference could have a comparable impact to the Strasbourg Court’s margin of appreciation doctrine, so the proposed distinction between general challenges and challenges to individual applications would seem equally appropriate.


It was shown, in Section 2, that the Strasbourg Court employs a two-track approach to racial hate speech, on the one hand, and religiously offensive speech, on the other hand. Further, the ECHR jurisprudence currently privileges journalistic or scholarly forms of religious criticism over creative or artistic ones by relying on the contestable political participation account of free speech. It was argued that the well-established Article 10 ECHR test of whether religiously offensive speech is ‘gratuitously offensive’ is in need of reform, while the introduction of a consistent test for racial hate speech cases in Europe is equally necessary.

Waldron’s theory was defended, in Section 3, as providing a coherent account of the harm in hate speech capable of serving as a foundation for a single Article 10 ECHR test for both racial and religious hate speech. Its principal features were shown to be (1) the twin aims of inclusiveness as a public good and the protection of an assurance of equal dignity and (2) the practical distinction between objective harm and subjective offence flowing from offensive or hateful speech. These core features were presented as theoretically justifiable and compatible with the general scheme of Article 10 of the ECHR.

It was then argued, in Section 4, that the appropriate test under Article 10 of the ECHR in both the racial and religious contexts is the following: ‘Is the expression objectively capable of seriously undermining the target’s assurance as to a status of equal worth in the community, having regard to the target’s knowledge, the speaker’s power and the forum of the expression, at the time it is made?’ That test is to be accompanied by the specific instruction that it must be applied from the perspective of a reasonable member of the target group.

Practically speaking, when my proposed test is satisfied, a State Party to the Convention is entitled to restrict, prohibit or criminalise the expression compatibly with Article 10 of the ECHR, provided also that the measure is proportionate stricto sensu. The distinct advantages of this reformed approach to Article 10 of the ECHR are that the unhelpful dichotomisation of racial hate speech and religiously offensive speech in the ECHR jurisprudence is abandoned and that freedom of expression in Europe is approached in a way that better differentiates between highly personal attacks (which should be more readily circumscribable) and general critiques of particular world views or ways of life (which should be less readily circumscribable). This would be a more coherent approach to racially and religiously offensive or hateful speech, and also be more responsive to the political reality of tense public discourse surrounding issues of race and religion in Europe at the present time.