The regulation of speech is a highly sensitive and always evolving ethical, political, and legal issue. On one hand, hateful and hurtful speech is on the rise, especially, but not exclusively, with regard to the relationship between Islam and the West. Islamophobic discourse is widespread in Western societies and some radical Islamists call for violence toward non-Muslims and “heretic” Muslims. We can also think of the radicalization and polarization of discourse brought about by the interactive phase of the Internet (social media, blogs, comments sections, etc.). On the other hand, demands for the suppression of certain forms of speech proliferate. Requests for “safe spaces” and “trigger warnings”, for the condemnation of “microaggressions”, and for a right not to be offended or insulted are burgeoning. From this offence-averse standpoint, freedom of speech does not justify hurtful speech, and a broader and subtler notion of “harm” needs to be factored in the analysis of the scope of our freedom of expression.
In this piece, I will first briefly review the argument in favour of both freedom of expression and the harm principle. Starting from the suspicion that the notion of harm defended by Millian liberals is too narrow but that an “offence principle” is too broad, I will side with theorists such as Jeremy Waldron and Rae Langton who argue that an adequate version of the harm principle ought to include anti-hate speech laws. I will concede that such laws need to target only the speech acts that express the most severe forms of aversion and denigration toward a particular group. I will then reflect on the status of “hurtful speech”, which I see as including the performative utterances that stop short of being hateful but which nonetheless erode the social standing and bases for self-respect of those who are targeted. I will then argue that the secular state has no ground for prohibiting blasphemous speech even when it is hurtful. I will conclude with the idea that the free speech debate reveals a limit of liberal political morality and therefore leaves liberal normative theorists with an uncomfortable predicament, as they have to rely more on the complementary role of pro-social personal dispositions and civic virtues than they generally wish to.
I start by taking for granted that freedom of expression is a basic human right that should enjoy robust legal protection. No right is absolute but, like other basic rights, the test for justifying its restriction should be demanding and when infringements are justified, they should be as minimal as they can be. I will not take sides here on whether freedom of speech has intrinsic or instrumental value. John Stuart Mill marshalled perhaps unsurpassed arguments in favour of the instrumental value of free discussion. He persuasively argued that the discovery of truth and the progress of reason in human affairs depend on the ongoing confrontation of ideas and possibility of criticism. Given our epistemic limitations, we need to be exposed to different points of view and to deliberate with others to see the flaws in our own beliefs. Even thoughts that are demonstrably false should not be silenced because refuting flawed arguments helps us to establish the truth with greater clarity and strength. As Wittgenstein put it, “[o]ne must start out with error and convert it into truth.”
Tim Scanlon makes a powerful case for the intrinsic value of freedom of expression with an argument from autonomy. As morally autonomous agents, we must have the freedom to express ourselves as we see fit and, equally importantly, we need to have access to the information and opinions communicated by others. The state should not judge for me what I should be exposed to. To be mature in Kant’s sense—that is, to be in a position to think without tutors—I need to have access to others’ beliefs and judge for myself which ones are true, enlightening, worthy, etc. Censorship of others thus undermines my autonomy.
Both the consequentialist and the deontological arguments are powerful and show why strong reasons are required to justify the legal regulation of speech. That being said, even the most ardent defenders of freedom of speech agree that it is a limited right. At the very least, liberals agree with Mill that a version of the harm principle is required to prohibit expressive acts that can lead to violence or physical harm. For Mill,
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. His own good, either physical or moral, is not a sufficient warrant. He cannot rightfully be compelled to do or forbear because it will be better for him to do so, because it will make him happier, because, in the opinions of others, to do so would be wise, or even right. These are good reasons for remonstrating with him, or reasoning with him, or persuading him, or entreating him, but not for compelling him, or visiting him with any evil in case he do otherwise.
So, the absolutist view of free speech can readily be brushed aside. Such an absolutist view was endorsed, perhaps only rhetorically, by Salman Rushdie in the aftermath of the tragic Charlie Hebdo attack. For him, “the moment somebody says ‘yes, I believe in free speech but’, I stop listening … The point about it is the moment you limit free speech it’s not free speech. The point about it is that it’s free.” We can easily understand why Rushdie is a passionate advocate of free speech, but this view is obviously false. Freedom of speech does not allow calls for genocide on the radio. No one, Rushdie included I presume, defends the absolutist position. The hard question is what comes after the “but”.
One can both recognize the value of freedom of expression and suspect that the harm principle is too narrow, that it protects speech acts that should be legally forbidden. In its original formulation, the harm principle seems to offer protection only against the physical harm that can be demonstrably caused by words or images. The harm principle appears to exclude the psychological and social forms of harm than can be caused by expressive acts, and it does not seem to authorize a probabilistic approach to prejudice. What I have in mind is that if anti-gay or anti-Semitic tracts are distributed in a neighbourhood, we do not know if it will causally lead to some harm to gays or Jews. We can have good reasons to think that it increases the probability that the target of defamation will be assaulted, insulted or discriminated against, but the standard account of the harm principle does not allow us to forbid such hateful and discrimination-inducing speech unless it is a direct call to violence.
Drawing loosely upon Joel Feinberg’s philosophy of criminal law, some are tempted to invoke a version of the “offence principle” to cast the net wider. Under this principle, people would have a right not to be offended, ridiculed, or insulted by acts of expression. For Feinberg, “offense” is less serious than “harm,” but can nonetheless be sanctioned by criminal law:
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 … to persons other than the actor, and that it is probably a necessary means to that end … The principle asserts, in effect, that the prevention of offensive conduct is properly the state’s business.
I will come back to the notion of offence in section three, but I will build my argument for now on the premise that an “offence principle” is too broad. Excluding obvious cases of obscenity, there cannot be a general right not to be offended or shocked in open societies. If this is right, the normative problem with regard to the regulation of speech lies in the space between harm and offence.
The first way to fill some of the space between harm and offence is to broaden our conception of harm. One of the ways that this can be done is through the prohibition of hate speech or hate propaganda. Many countries, including Canada and several European countries, have legal provisions against hate speech. Hate speech is defined in Canada’s Criminal Code as an expressive act communicated in public which “incites hatred against any identifiable group.” As Alexander Brown has carefully shown, hate speech encompasses different kinds of potentially overlapping expressive acts, such as “group defamation”, “negative stereotyping or stigmatization”, the public expression of “insults, slurs, or derogatory epithets”, “incitement to hatred”, or encouragement to discriminate. In Canada, hate speech is regulated via criminal law and, in some provinces, via human rights legislation. A form of expression is hateful in the criminally prohibited sense when the person who proffered it knew that it would stir up hatred against the members of an identifiable group, or at least that it had the potential to do so. The prohibition of hate speech through human rights laws targets forms of expression that are likely to lead to discriminatory acts toward the members of a group. As Richard Moon summarizes:
The purpose of the human rights code ban on hate speech is not to condemn and punish the person who committed the “discriminatory” act, but rather to prevent or rectify discriminatory practices or to compensate the victims of discrimination for the harm they have suffered.
Legal provisions against hate speech are challenged by some liberal political philosophers who think that having legislation explicitly limiting free speech is too risky, illiberal, or undemocratic. Hate speech legislation can be used abusively, can have a chilling effect on speech, and incite self-censorship, which can violate personal autonomy and deprive us of useful controversial points of view.
Scanlon, for instance, in his influential papers on freedom of expression, does not say that speech acts or images that are meant to stir hate or contempt could legitimately be outlawed, and it seems reasonable to think that he is against hate speech legislation. According to his own formulation of the harm principle, which he calls the “Millian Principle”, expressive acts should not be prohibited if they lead others to perform illegal acts:
There are certain harms which, although they would not occur but for certain acts of expression, nonetheless cannot be taken as part of a justification for legal restrictions on these acts. These harms are: (a) harms to certain individuals which consist in their coming to have false beliefs as a result of those acts of expression; (b) harmful consequences of acts performed as a result of those acts of expression, where the connection between the acts of expression and the subsequent harmful acts consists merely in the fact that the act of expression led the agents to believe (or increased their tendency to believe) these acts to be worth performing.
We can find bigoted or racist speech appalling and distasteful, but putting up with it is the price we pay to live in a liberal society. The law should not concern itself with such utterances if no one is harmed because of them, and the causation needs to be clear and direct, such as when one shouts “fire” in a crowded cinema. If someone harms another under the influence of xenophobic messages, the person held legally responsible is the perpetrator of the assault, not the one who expressed himself. Words are deeds, but they do not by themselves cause physical harm. Most democratic countries apart from the United States have been unmoved by the argument against hate speech laws, but it nonetheless remains influential. For instance, the magazine The Economist endorses Scanlon’s view, without the nuance and residual doubt:
The law should recognise the right to free speech as nearly absolute. Exceptions should be rare. Child pornography should be banned, since its production involves harm to children. States need to keep some things secret: free speech does not mean the right to publish nuclear launch codes. But in most areas where campaigners are calling for enforced civility (or worse, deference) they should be resisted. … Laws against hate speech are unworkably subjective and widely abused. Banning words or arguments which one group finds offensive does not lead to social harmony. On the contrary, it gives everyone an incentive to take offence—a fact that opportunistic politicians with ethnic-based support are quick to exploit.
I find this conception of harm too narrow. One can turn at this juncture to the conceptualizations of hate speech put forward by theorists such as Jeremy Waldron and Rae Langton. According to Waldron, hate speech can be banned because it undermines in a profound and troublesome way the social status of the members of the targeted groups. Their basic place and standing in the society are put into question by defamatory or heinous slogans or images. One of Waldron’s most useful contributions lies in his notion of “inclusiveness”. He sees inclusiveness as a “public good” in the technical sense of the term. Goods are public when they are not rivalrous in the sense that we can all consume them or benefit from them simultaneously without making them scarcer, just like clean air or clean water. Where public education is truly free and where there are enough public schools for everyone, elementary and secondary education is a public good. Inclusiveness, then, is the property of societies which makes it possible for everyone to feel that they belong to them as equal members, that their place and status is not put into question, regardless of their identity-conferring attributes or conceptions of the good. When a society is inclusive in Waldron’s sense, inclusiveness is invisible; we enjoy it without being aware of it, just like we are usually not aware that air is available. We become aware of it when it begins lacking.
For Waldron, hate speech also has the effect of attacking the dignity of those who are targeted. He is not mainly referring to dignity as an abstract and generic status granted to all rational agents such as in Kantian philosophy. He is referring to what can be called the “practical dignity” of flesh and blood citizens who lead their life in real-world communities. Hate speech undermines the practical dignity of its victims in the normal interactions of daily life. A hateful or debasing graffiti or poster or tirade on the radio about Muslims in general attacks the social standing of every single Muslim who sees or hears it. It tells the listener that some of their fellow citizens want them out or see them as untrustworthy, threatening or inferior. It affects the way he or she sees others when they are at work, run errands or go to swimming classes with their children. As Waldron puts it, “the stakes have changed” for the targets of hate speech. The “assurance” that they will be treated with dignity and decency starts to dissolve. The victims cannot help wondering if their children will have the chance to flourish in such a society.
Moreover, others who share a similar aversion can feel validated and encouraged to express their sentiments publicly. When hateful or contemptuous speech erupts in the public sphere, it lowers the social cost of expressing negative attitudes toward the targeted group, especially when it is expressed by people in positions of authority or influence. The diffusion and circulation of hate speech favour the coordination of actions among those who share a common aversion and can, as a consequence, increase the vulnerability of the targeted groups. According to the Supreme Court of Canada in Saskatchewan (Human Rights Commission) v. Whatcott, “[a]s the majority becomes desensitized by the effects of hate speech, the concern is that some members of society will demonstrate their rejection of the vulnerable group through conduct. Hate speech lays the groundwork for later, broad attacks on vulnerable groups.” As we will see below, because of its illocutionary force, hate speech participates in the evolution of social norms with regard to what it is acceptable to think, say, and do.
Finally, another persuasive argument in favour of a stricter regulation of speech than the one favoured by Millians is, interestingly, a freedom of expression-based argument concerned with the background conditions for the free exchange of ideas. It points out that undermining the practical dignity and civic assurance of some citizens makes it harder for them to participate in public debates and to challenge the negative representations of their group identity. Expressing one’s opinion in a climate of hostility requires self-confidence, a kind of relation to self that is harder to foster if the group with which one is associated is stigmatized. In addition, participation in public debates can attract more negative comments such as ad hominem attacks or abusive generalizations about the group. This is why I believe, contra Ronald Dworkin, that the potential “silencing effect” of hate speech outweighs the limited restriction of the hatemonger’s right to have a voice in the deliberative and legislative process leading to the adoption of a law or policy to which he objects, such as anti-discrimination laws. It is true that the hatemonger cannot participate in the self-legislation process the way he sees fit as he is forbidden to express some of his attitudes and opinions, but this leaves him ample room to voice his opposition to the concerned laws. Considering that, as I will argue below, hurtful but not hateful speech ought to be tolerated, the hatemonger not only has the freedom to offer recognizable arguments against the legislation, but also to express some of his visceral attitudes and reactions toward the group protected by the anti-discrimination law. According to Dworkin, the possibility of expressing attitudes, emotions, prejudices, and other affective states has to be protected by the right to free speech.
In a similar vein, Corey Brettschneider argues that “[c]oercive sanction merely tries to bury opinion and therefore misses the grievances that might be held legitimately even by those with deeply racist views.” What is missing here among the defenders of the American tradition of legal tolerance toward hate speech is the idea of a spectrum of negative speech. A wide range of negative and critical attitudes toward beliefs, values, practices, and even persons do not qualify as hateful speech acts. The hatemonger does not fully lose his political voice. He is prevented from expressing deeply degrading or vilifying views that are likely to lead to the exclusion or to increase the vulnerability of the members of the targeted groups. By analogy, the freedom of religion and parental authority of Jehovah’s Witnesses parents are not fully annihilated when a court decides, rightfully, that they cannot prevent a medical team from administering a life-saving blood transfusion on their child. The Jehovah’s Witnesses’ freedom of religion and parental authority are undeniably curtailed, but in a specific situation and in a way that ought to be proportionate to the objective of protecting the right to life and bodily integrity of the sick child.
It is worth recalling here that John Rawls thought that the social bases of self-esteem were “perhaps the most important primary good” that has to be produced and secured by the basic social structure. Without self-respect, I might lack the psychological capacity or the will to pursue worthwhile projects or plans even if my negative freedom is protected and if my income is sufficient to lead a decent life. If we accept that degrading, contemptuous, or exclusionary slogans and images can undermine one’s sense of self-worth, this gives us a reason to see such expressive acts as harmful. For Rawls, “the parties in the original position would wish to avoid at almost any cost the social conditions that undermine self-respect.” One of the ways that the basic institutional structure of a liberal society can produce the bases of self-respect is by prohibiting vicious, deeply demeaning, and discriminatory speech. The point here is not that Rawls supported hate speech laws but rather that his notion of the social bases of self-respect provides a powerful justification for them.
Millians fear that hate speech legislation will be construed too broadly and will end up including utterances or displays that are harsh and offensive, but not harmful in the appropriate sense. Harsh words are sometimes required and irreverence or lampooning can be powerful forms of social criticism. Freedom of religion, for instance, does not shield the Catholic Church from art or social commentary recalling the pedophilia scandal nor does is prohibit the critique of the violent theology of some Salafist movements. And no liberals can think that the so-called “new atheists” should not be free to speak their mind about religious faith even if their views will hurt or offend some believers. Moreover, even when harsh criticism or mockery are not particularly useful contributions to public debate, democratic citizens need to develop the civic endurance and resilience that enable them to cope with offensive expressive acts.
Hence, the prohibition of hate speech needs to include only the clearest and most extreme forms of aversion toward the members of a group. Expressive acts can only be prohibited if they are likely to lead to the detestation or vilification of the targeted group and, in so doing, significantly increase the likelihood that the group’s members will be the object of legally prohibited forms of exclusion, discrimination, or outright violence. Once again in the words of the Supreme Court of Canada in Whatcott:
Representations that expose a target group to detestation tend to inspire enmity and extreme ill-will against them, which goes beyond mere disdain or dislike. Representations vilifying a person or group will seek to abuse, denigrate or delegitimize them, to render them lawless, dangerous, unworthy or unacceptable in the eyes of the audience.
The Court added, to dispel any ambiguity, that hate speech goes “far beyond merely discrediting, humiliating or offending the victims.” The UK Racial and Religious Hatred Act 2006 is also crystal clear:
Nothing in this Part shall be read or given effect in a way which prohibits or restricts discussion, criticism or expressions of antipathy, dislike, ridicule, insult or abuse of particular religions or the beliefs or practices of their adherents.
As a legal category, hate speech thus needs to include only extreme forms of aversion and contempt toward a particular group expressed publicly, such as strongly demeaning stereotypes, vilifying epithets, and incitements to discrimination. There will be, in practice, threshold issues and hard cases, but the general principle is that hate speech goes beyond disdain, dislike, humiliation, offence, antipathy, and ridicule. The explicit distinction that we find in Canadian and British law between offensive speech and hate speech should provide some comfort to those who fear the slippery slope.
Hate speech legislation thus stands, I think, on solid philosophical ground. Hateful speech, however, is not a discrete and sui generis kind of discourse. It designates the expressive acts that are at the end of the spectrum in terms of the communication of negative sentiments or opinions toward a group. Speech is hateful when the aversion that is communicated against members of a group has a certain intensity or reaches a certain level of vehemence. Since it is located at the end of a continuum, it raises a question of threshold: there are utterances and signs that target some citizens in negative ways but which do not cross the threshold.
Acts of expression that are hurtful but not hateful in the appropriate sense abound in the public sphere today. New information and communication technologies give anyone who has an internet connection multiple platforms to express themselves without first having to go through gatekeepers. The current debates about Islam and migrants and the various culture wars are hotbeds of mean and inflammatory rhetoric. Trolls are now better known as characters of the Internet than as inhabitants of Scandinavian caves and forests.
The normative problem here is that hurtful speech also undermines the practical dignity and social standing of its targets, and it can also incite others to up the ante, but less so than hate speech. Turning, here, as Langton suggests, to speech act theory is enlightening. As J. L. Austin indicated, an utterance can have locutionary, illocutionary, and perlocutionary effects. In addition to the semantic content of what is said, a speech act has illocutionary force when saying something is, in a specific context, also doing something. An illocutionary act can be a confirmation, warning or encouragement. The perlocutionary force of an utterance lies in the effects it has on those to whom it is addressed.
Think, for instance, of the following statement made by a candidate for office: “Muslim citizens should have equal rights, but we think that immigrants from Muslim countries should be obliged to watch a video about ‘Western values’ upon arrival in our country.” In addition to its propositional content, this statement also indirectly conveys a set of messages. As an illocutionary act, it says or demonstrates to voters that the party is “standing up” for the country’s “values” or “identity” and that it takes the alleged Muslim threat seriously. The illocutionary act is “felicitous” if those who are prone to think in terms of a civilizational clash “secure uptake,” i.e. understand its illocutionary aspect. The perlocutionary effects of the policy proposal are likely to include the aggravation of the moral panic with regard to Islam and the amalgamation of the most conservative versions of Islam with the Muslim faith in general. The take-home message for those who are less reasonable—outright Islamophobes and racists—might be that the Far Right is correct and that is actually time to go farther and take bolder actions.
As I suggested above, the effects of hurtful speech are negative, but in milder ways than hate speech. Take, for example, a recent editorial published by the Charlie Hebdo editorial board. According to the writers, all Muslims have their share of responsibility in the terrorist attacks committed by the assassins who self-identify as Muslims and have a connection with terrorist groups such as ISIL. The prose is vague and full of coded messages, but we understand that the new and peaceful Muslim baker around the corner is partly responsible because he does not sell ham sandwiches and, we are led to infer, because he does not condemn the terrorist attacks in a way that the Charlie Hebdo writers find acceptable. In the same editorial, Charlie Hebdo endorses the oft-repeated idea that is not possible today to “say anything negative about Islam” without being labelled as an “Islamophobe”. The charge of Islamophobia is meant to short-circuit necessary debate and social criticism. It seems that the likes of PEGIDA (Patriotic Europeans Against the Islamisation of the West) do not exist in Charlie Hebdo’s world.
In the same spirit, the philosopher Stephen Law wrote—in a piece in which he also protests against the very notion of Islamophobia—that an assertion such as “Islam is a poisonous and destructive religion” is not Islamophobic. One is left wondering how Islamophobia is to be defined if it excludes such a general proposition about the essence of Islam, but my point here is not that such speech acts should be categorized as hateful. One should have the right to press Muslims to condemn terrorism more forcefully, even if, in fact, there are legions of Muslim clerics, organizations, and individuals who are already doing so, and even if Muslims are the most frequent victims of terrorist Islamist groups. One should also have the right to say that all or certain religions are intrinsically violent or detrimental to human progress. I disagree with such essentializing critiques of religion but, in a Millian spirit, being forced to engage with such views strengthens the case for more inclusive or pluralistic positions.
My point here is not that such statements are hateful in the legally prohibited sense, but rather that they nonetheless reduce the inclusiveness of societies and make it much harder for Muslim citizens to feel that they are seen as equal and trustworthy fellow citizens. What will be the indirect effects of the constant reiteration that Muslims in general are complicit with terror or that their faith—no matter how it is interpreted and practised in lived experience—is poisonous and destructive? Hurtful speech acts must be tolerated in liberal democracies, but we should not lose sight of the fact that that their illocutionary aspect and perlocutionary effects are troublesome, as they are liable to undermine the assurance of those who are targeted that they are seen as fellow citizens in good standing. I will suggest a way to approach the regulation of hurtful speech in section four.
Not everyone agrees, but I see blasphemous speech as a particular kind of hurtful but non-hateful speech. In a recent parliamentary debate on a piece of anti-hate speech legislation at the Quebec legislative assembly, the leader of the Muslim Council of Montreal pleaded for including “insult to religion” within the ambit of the law. He said that he does not mind being insulted personally, but that no one should have the right to insult another’s religion. Pope Francis apparently shares the same view.
From the perspective of a political and liberal theory of secularism, which can be seen as a specification of Rawls’ political liberalism, it is doubtful that a justification for the legal prohibition of blasphemy can be found. Political secularism implies, among other things, that religious beliefs are not directly translated or incorporated into positive law. Laws and other public norms need to be based on public or secular reasons, that is reasons that are intelligible and potentially endorsable from the perspective of all reasonable comprehensive doctrines.
By definition, an expressive act can be deemed blasphemous only from a religious point of view. A graphic representation of the Islamic prophet is not intrinsically and universally offensive. It is so from within a certain interpretation of Islam. Blasphemy is irreverent or outrageous speech with regard to God or something that is considered sacred. In a secular society, no one is obliged to think that some things, practices or ideas are sacred. No one is forced to endorse the sacred/profane metaphysical dualism. Believers have the right to consider that some expressive acts are beyond the pale, but that is not sufficient reason for curtailing the expressive freedom of others. There is no general right not to be offended in secular and open societies.
It is doubtful that the Danish and Charlie Hebdo cartoons of the Islamic prophet were hateful in the sense discussed above. A precise case-by-case analysis would be required, but most of the cartoons tried to target extremists and had clear political messages. As I will argue below, criticizing the Jyllands-Posten and Charlie Hebdo for their unnecessary provocations was both legitimate and timely, but the newspapers had the legal right to ridicule, offend, and hurt. As mentioned below, liberal democracies must tolerate hurtful speech.
So far, I have tried to sketch what a sound defence of hate speech legislation should look like. Against Millian liberals, I believe that the proper regulation of free speech requires a broader and subtler notion of harm. I then wondered how we should think about hurtful but non-hateful speech and came to the unsurprising conclusion that liberal principles of justice force us to tolerate it even if it weakens the social standing of those targeted and increases their vulnerability. We reached a limit of liberal egalitarian justice, but a more restrictive regulation of freedom of speech is likely to bring about even worse consequences.
Some could argue, following Brettschneider, that it would be hasty to conclude that liberal egalitarian justice has reached a limit. Brettschneider rightly points out that the democratic state not only has the power to ban and sanction; it also has expressive powers. According to him, hate speech legislation encroaches too deeply upon the autonomy and political equality of citizens, but the state should nonetheless use its expressive power to defend substantive democratic values, such as the basic rights of those targeted by the hate speech. State representatives and institutions can criticize hateful speech and support vulnerable groups in a variety of ways. The expressive capacities of the democratic states include, for instance, “court opinions, public holidays [Martin Luther King Day, Black History Month, etc.], and government subsidies.”
Doing justice to Brettschneider’s position would require a separate paper. I will simply point out that one can wholeheartedly agree that the state should use its expressive capacities to show how hate speech contradicts the values of free and equal citizenship while being very moderately optimistic with regard to the actual effects of what he calls “democratic persuasion.” The courts acting as exemplars of public reason, enlightened civic education at school, public monuments and holidays recalling and celebrating the civil rights movement, and so on, are all desirable, but it is hard to be confident that they will incite the far right extremist to show some self-restraint in the way he or she talks about Muslims, women, or LGBT people. Democratic persuasion offers little protection against the unpersuaded hatemonger. In August 2017, protesters chanting white supremacist slogans—some wearing KKK insignia, others making Nazi salutes—protested against the removal of a statue memorializing the Confederate General Robert E. Lee in Charlottesville, Virginia. Statue removal is a clear way for the state to use its expressive power.
I now want to suggest that if I’m right to think that the legal regulation of speech and democratic persuasion are necessary but insufficient for countering the harm of hateful and hurtful speech, it entails that we need at this stage to move to the sphere of moral character and social ethic. Law cannot be, and has never been, our only mode of social regulation. In a Hegelian spirit, we also need to think about the ethical life of communities, (i.e. about the constantly evolving and intersubjectively defined values, informal norms, and customs that make social cooperation possible). If I do not think that virtue ethic can be put on a par with deontological and consequentialist moral theories with regard to the normative justification of institutions, laws, and policies, civic virtues are nonetheless required to sustain and complement the workings of public institutions. By civic virtues, I refer to the pro-social dispositions enacted by citizens that facilitate justice, cooperation, and stability. A fair system of social cooperation cannot rest solely upon our obedience to the law; it is also conditioned by our dispositions, attitudes, and ways of relating to our fellow citizens.
So, the last argument that I wish to defend in this paper is that the ethical counterpart of our legal right to offend, ridicule, and hurt is to think carefully about the impacts of our expressive acts on others, given their deepest values, commitments, and attachments. As Martha Nussbaum stresses relentlessly in her work, becoming a virtuous moral agent and citizen involves being capable of imagining empathically what it is to be in a situation different from ours and what will be the likely effects of our actions on others.
It is, I think, from Nussbaum’s empathy-based standpoint that the Jyllands-Posten and Charlie Hebdo could be criticized for publishing the cartoons. There were other ways to challenge radical Islam that were available to them—ways that were much less likely to offend a large number of Muslims who are law-abiding and peaceful citizens, and to give ammunition to those who sponsor the clash of civilizations thesis. Media outlets have the legal right to publish or broadcast harmful but non-hateful content, but they need to tolerate criticism when they do so. Since words and images are actions that can wound, we can all be held ethically responsible and taken to task for what we say. Having a right to do something is not a sufficient reason to act upon it. Practical deliberation on how we should act is far from being exhausted by our reckoning of the rights we have.
Some worry that this kind of social criticism and moral condemnation is also a threat to freedom of expression as it can lead to self-censorship. But self-censorship is only regrettable when one decides not to express oneself for fear of becoming the target of illegal or immoral actions. Self-censorship can be induced by the fear of terrorist assaults or of being legally harassed by individuals or organizations who have the financial means to sue. One of the common arguments against hate speech legislation is that it can have a chilling effect on those who have controversial or sensitive messages to express. The fear of being dragged before a civil, penal, or human rights tribunal can incite one to stay silent and skip one’s turn. These reasons for self-censorship are indeed real threats to freedom of speech. I will, however, assume that the chilling effect problem is a practical rather than principled one: terrorists must be stopped and we need rules for turning down abusive lawsuits (such as anti-SLAPP legislation). Non-Millian liberals can agree that the chilling effect should be contained, but still believe that the effects of hateful acts of expression on vulnerable groups, including the silencing effect, justify hate speech laws.
But self-censorship can also be the outcome of genuine moral deliberation, such as when I decide not to express what I initially wanted to share because of the potential impact of my speech acts on others. Perhaps one could argue that this is not censorship in the appropriate sense since the agent decides not to speak out of her own deliberation and volition, but the fact remains that she did not express the thoughts she initially had. Call it self-censorship or not, this outcome is not a threat to freedom of speech. I concur with Kate Manne and Jason Stanley, who wrote: “Censure is not the same thing as censorship; indeed, it could not be. The right not to be censored by the government extends to the right to censure—that is, morally condemn—the speech acts of other people.” If one believes that hurtful speech has to be tolerated because it is only through the free exchange of ideas that we can progress, then one has to welcome the critique of how the right to free speech is exercised. For Mill, accepting the harm principle did not entail that we should abstain from “remonstrating with” others who act in ways that we find morally reprehensible.
Turning toward virtue ethic feels somewhat like giving up. We cannot, of course, force anyone to become virtuous through the coercive power of the state, and resorting to changing the “culture” can be a form of escapism or wishful thinking. Moreover, the debate over freedom of expression reveals a genuine conflict between distinct democratic goods. We can recognize the value of both empathy and concern toward our fellow citizens and of frank and vigorous debates. That being said, turning to civic virtues is not exactly giving up. Our attitudes and actions contribute to the permanent evolution of social norms and of the conversational common ground. Civic virtues can be promoted in the ways we relate to others, in our interventions in public debates, and in education. At this point in time, it seems crucial that our common discursive space is not left to those who promote discourses that exacerbate social division and strain social cooperation. In addition to hate speech legislation, challenging hurtful speech and exemplifying another way of relating to others might be the only (liberal) recourses that we have.
 See Timothy Garton Ash, Free Speech: Ten Principles for a Connected World (New Haven: Yale University Press, 2016) at 154–57.
 See John Stuart Mill, On Liberty, ed by David Bronwich & George Kateb (New Haven: Yale University Press, 2003) at 101ff. Such cognitive limitations have been highlighted and detailed by social psychology and the cognitive sciences in recent decades. See generally Jonathan Haidt, The Righteous Mind: Why Good People Are Divided by Politics and Religion (New York: Pantheon Books, 2012); Daniel Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux, 2011).
 Ludwig Wittgenstein, Philosophical Occasions: 1912–1951, ed by James C Klagge & Alfred Nordmann, (Indianapolis: Hackett Publishing Company, 1993) at 119.
 TM Scanlon, The Difficulty of Tolerance: Essays in Political Philosophy (Cambridge, UK: Cambridge University Press, 2003) at 95–97. For a distinct autonomy-based argument, see C Edwin Baker, “Autonomy and Hate Speech” in Ivan Hare & James Weinstein, eds, Extreme Speech and Democracy (Oxford: Oxford University Press, 2009) 139.
 See generally Immanuel Kant, “An Answer to the Question: What is Enlightenment?”, translated by James Schmidt, in James Schmidt, ed, What is Enlightenment? Eighteenth Century Answers and Twentieth Century Questions (Berkeley, University of California Press, 1996) 58.
 Mill, supra note 2 at 80.
 Associated Press, “Rushdie Defends Charlie Hebdo, Free Speech” (15 January 2015), online: YouTube <https://www.youtube.com/watch?v=q9cYuQonbXs> at 00h:00m:26s.
 See Alexander Brown, Hate Speech Law: A Philosophical Examination (New York: Routledge, 2015) at 50.
 For more expansive interpretations of the harm principle, see David O Brink, “Millian Principles, Freedom of Expression, and Hate Speech” (2001) 7:2 Leg Theory 119 at 146 (suggesting that the deleterious impacts of hate speech on deliberative speech constitutes harm); Raphael Cohen-Almagor, “Harm Principle, Offence Principle, and the Skokie Affair” (1993) 41:3 Political Studies 453 (examining Mill’s harm principle, Feinberg’s offence principle, and the moral equivalency of expression that intentionally inflicts psychological harm and acts that cause physical pain).
 Joel Feinberg, The Moral Limits of the Criminal Law: Offense to Others, vol 2 (New York: Oxford University Press, 1985) at 1. See also Raphael Cohen-Almagor, The Scope of Tolerance: Studies on the Costs of Free Expression and Freedom of the Press (Milton Park, UK: Routledge, 2006) ch 4, 5 (examining the circumstances where “offense to sensibilities” can provide justification to limit free expression).
 For examples of European legislation, see e.g. Strafgesetzbuch, § 130 (the German penal code’s prohibition of “incitement to hatred”); Loi de 30 Juillet 1981 tendant à réprimer certains actes inspirés par le racisme ou la xénophobie, MB, 8 August 1981, 9928 (a Belgian statute that incorporates incitement to discrimination, hatred, or violence into the Belgian Penal Code); Public Order Act 1986 (UK), c 64 (a UK statute that creates offences for inciting religious and racial hatred, among other things).
 RSC 1985, c C-46, s 319(1). This is sometimes called “group libel” or “group defamation”: see Jeremy Waldron, The Harm in Hate Speech (Cambridge, Mass: Harvard University Press, 2012) ch 3.
 Brown, supra note 8, ch 2. I will not discuss here the differences between the various types of legal prohibition of hate speech (through criminal law, common/civil law, human rights legislation, etc.) and will not offer a fine-grained conceptual analysis of the various types of hate speech.
 See e.g. The Saskatchewan Human Rights Code, SS 1979, c S-24.1, s 14(1)(b); Human Rights Code, RSBC 1990, c 210, s 7(1)(b). The Canadian Human Rights Act, RSC 1985 formerly contained a hate speech provision, s 13, but this was repealed in 2013.
 See Mugesera v Canada (Minister of Citizenship and Immigration), 2005 SCC 40, at paras 99–106,  2 SCR 100 (interpreting s 319 of the Criminal Code).
 Richard Moon, Putting Faith in Hate: When Religion is the Source or Target of Hate Speech (Cambridge, UK: Cambridge University Press, forthcoming in 2018) ch 2. Although I am not in a position to take a definitive stand on this issue, the moderate expansion of “harm” defended by Professor Moon appears to vindicate the ban on hate speech via human rights codes (and not only through criminal law). For a specific proposal to ban hate speech through criminal law rather than through a human rights code, see Richard Moon, “Report to the Canadian Human Rights Commission Concerning Section 13 of the Canadian Human Rights Act and the Regulation of Hate Speech on the Internet” (October 2008) Canadian Human Rights Commission, at 31–33, online: SSRN <ssrn.com/abstract=1865282>. Note that Moon himself does not recommend, in his forthcoming book, that all human rights-based bans on hate speech be repealed.
 See Ronald Dworkin, “Foreword” in Hare & Weinstein, supra note 4 at ix (“[w]e might have the power to silence those we despise, but it would be at the cost of political legitimacy, which is more important than they are”); Corey Brettschneider, When the State Speaks, What Should It Say? How Democracies Can Protect Expression and Promote Equality (Princeton: Princeton University Press, 2012) at 3.
 Scanlon, supra note 4 at 199.
 Ibid at 14.
 See Michel Rosenfeld, “Hate Speech in Constitutional Jurisprudence: A Comparative Analysis” in Michael Herz & Peter Molnar, eds, The Content and Context of Hate Speech: Rethinking Regulation and Responses (Cambridge, UK: Cambridge University Press, 2012) 242 at 288–89.
 “Under Attack”, The Economist (4 June 2016), online: <https://www.economist.com>, archived at https://perma.cc/DXY5-DLKX.
 See Waldron, supra note 12 at 4–6.
 Ibid at 60.
 Waldron envisions the following scenario:
Suppose that a spate of discriminatory signs appear; maybe they bespeak of a real intention to discriminate, or maybe they do not. But suddenly the stakes have changed for those to whom they are directed. … This helps us to see what hate speech is about. The point of the bigoted displays that we want to regulate is that they are not just autonomous self-expression. They are not simply the views of racists letting off steam. The displays specifically target the social sense of assurance on which members of vulnerable minorities rely. The point is to negate the implicit assurance that a society offers to the members of vulnerable groups (ibid at 88).
 2013 SCC 11 at para 74,  1 SCR 467 [Whatcott].
 See Rae Langton, “Beyond Belief: Pragmatics in Hate Speech and Pornography” in Ishani Maitra & Mary Kate McGowan, eds, Speech and Harm: Controversies Over Free Speech (Oxford: Oxford University Press, 2012) 72 [Langton, “Beyond Belief”]. Langton notes that pornography and hate speech “might have the illocutionary force of altering norms and social conditions, by legitimating, or advocating, certain beliefs, attitudes, and behaviour” (ibid at 82).
 In the words of the Supreme Court in Whatcott, supra note 25 at para 75:
[H]ate propaganda … impacts [a] group’s ability to respond to the substantive ideas under debate, thereby placing a serious barrier to their full participation in our democracy. Indeed, a particularly insidious aspect of hate speech is that it acts to cut off any path of reply by the group under attack. It does this not only by attempting to marginalize the group so that their reply will be ignored: it also forces the group to argue for their basic humanity or social standing, as a precondition to participating in the deliberative aspects of our democracy.
For a nuanced treatment of the putative “silencing effect” of hate speech, see Brown, supra note 8, at 198–200.
 For Dworkin’s brief argument against hate speech laws, see Dworkin, supra note 17 at vi–vii.
 Corey Brettschneider, “Value Democracy as the Basis for Viewpoint Neutrality: A Theory of Free Speech and its Implications for the State Speech and Limited Public Forum Doctrines” (2013) 107:2 Nw UL Rev 603 at 613 [Brettschneider, “Viewpoint Neutrality”].
 See AC v Manitoba (Director of Child and Family Services), 2009 SCC 30 at para 156,  2 SCR 181 (concluding that the impugned legislation was a proportionate limit on freedom of religion).
 John Rawls, A Theory of Justice (Cambridge, Mass: Belknap Press of Harvard University Press, 1971) at 440, 546 [Rawls, Theory of Justice].
 Matteo Bonotti accordingly argues that hate speech laws need to be grounded in a republican theory of freedom as “non-domination”. I cannot address this issue here but if I concur with him on the legitimacy of hate speech laws, I am not convinced that the “Italo-Atlantic” strand of the republican tradition offers a real and acceptable alternative to liberal egalitarianism at the level of political morality. I find neo-republicanism more useful at the level of institutional design. See Matteo Bonotti, “Religion, Hate Speech and Non-Domination”, (2017) 17:2 Ethnicities 259. For a thorough account of republican theory, see generally Philip Pettit, On the People’s Terms: A Republican Theory and Model of Democracy (Cambridge, UK: Cambridge University Press, 2012).
 Rawls, Theory of Justice, supra note 31 at 440.
 Whatcott, supra note 25 at para 41.
 (UK), c 1, s 29J.
 For a more detailed demonstration see Brown, supra note 8 at ch 9.
 Rae Langton, “Speech Acts and Unspeakable Acts” (1993) 22:4 Philosophy & Public Affairs 293 (discussing how speech can amount to action, and how pornographic speech acts can both subordinate and silence women). See also Langton, “Beyond Belief”, supra note 26 (applying speech act theory to hate speech and pornography).
 This is not a thought experiment. For similar comments, see Jenny Kleeman, Tom Silverstone & Mustafa Khalili, “Norway’s Muslim Immigrants Attend Classes on Western Attitudes to Women: Video”, The Guardian (1 August 2016), online: <https://www.theguardian.com> at 00h:08m:34s.
 JL Austin, How to Do Things with Words (Cambridge, Mass: Harvard University Press, 1962) at 138–39.
 “How Did We End Up Here?”, Charlie Hebdo (30 March 2016), online: <https://
charliehebdo.fr>, archived at https://perma.cc/99YR-RNWJ.
 Canada is not immune to Islamophobia. A series of graffiti stating “death to Muslims” or “forbidden to Muslims” appeared in a number of towns and cities in Canada in 2016. See Gabriel Delisle, “Graffiti islamophobe à l’UQTR”, Le Nouvelliste (10 March 2016), online:<https://www.lenouvelliste.ca>, archived at https://perma.cc/MZ3R-ZQTU; Ghyslain Bergeron, “Nouvelle vague de graffitis haineux”, L’Express [of Drummondville] (8 March 2016), online: <www.journalexpress.ca>, archived at https://perma.cc/2A52-KQ8V. On January 29th, 2017, a young man known for his sympathy for far right populist movements killed six people in Quebec City Mosque. See Jonathan Montpetit, “10 Minutes of Terror: What Happened in the Quebec City Mosque Sunday Night”, CBC News (2 February 2017), online: <www.cbc.ca>, archived at https://perma.cc/3ZKU-58NC.
 The author also gives a list of (quasi)symmetric statements that should not be seen as anti-Semitic. Tellingly, the author refers to the state of Israel rather than to Judaism in the corresponding assertion: “Israel is a poisonous and destructive state.” Did he fear that the same statement about Judaism could be thought of as anti-Semitic? Stephen Law, “Islamophobia and Antisemitism: What Is, and What Isn’t, Bigoted?” (27 April 2016) The Outer Limits With Stephen Law (blog), online: <https://www. centerforinquiry.net/blogs/entry/islamophobia_and_antisemitism_what_is_and_what_isnt_ bigoted>, archived at https://perma.cc/45VV-8662.
 See Jocelyne Richer, “Un leader musulman veut que Québec interdise les moqueries sur la religion”, La Presse [of Montreal] (20 August 2015), online: <www.lapresse.ca>, archived at https://perma.cc/H6KF-9XWM. See also Muslim Council of Montreal, “Mémoire pour la Commission des institutions, Assemblée nationale du Québec Consultations particulaires Projet de loi n° 59 : Loi édictant la Loi concernant la prévention et la lutte contre les discours haineux et les discours incitant à la violence et apportant diverses modifications législatives pour renforcer la protection des personnes”, Brief Submitted to the Committee on Institutions, Quebec, National Assembly, CI-012M, CP-PL-59 (20 August 2015). The proposed legislation, Bill 59, was enacted in 2016, albeit with the removal of its hate speech provisions. See An Act to amend various legislative provisions to better protect persons, SQ 2016, c 12; Robert Dutrisac, “Québec abandonne l’encadrement du discours haineux”, Le Devoir (26 May 2016), online: <www.
ledevoir.com>, archived at https://perma.cc/9JA8-EPJF.
 See Nick Squires, “Pope Francis: ‘You Cannot Make Fun of the Faith of Others’”, The Telegraph (15 January 2015), online: <www.telegraph.co.uk>, archived at https://
 See John Rawls, Political Liberalism (New York: Columbia University Press, 1993) at 212–54; Robert Audi, Democratic Authority and the Separation of Church and State (Oxford: Oxford University Press, 2011); Jocelyn Maclure & Charles Taylor, Secularism and Freedom of Conscience, translated by Jane Marie Todd (Cambridge, Mass: Harvard University Press, 2011) (suggesting that “social cooperation in diverse societies is rooted in the possibility that reasonable citizens will agree on the basic principles of their political association” at 107).
 Conversely, this also applies to those who cannot see an Islamic veil or a turban in the public sphere without having a meltdown.
 For discussions of the cartoon controversies, see e.g. Sune Lægaard, “The Cartoon Controversy: Offence, Identity, Oppression?” (2007) 55:3 Political Studies 481; Geoffrey Brahm Levey & Tariq Modood, “The Muhammad Cartoons and Multicultural Democracies” (2009) 9:3 Ethnicities 427.
 See Brettschneider, “Viewpoint Neutrality”, supra note 29 at 608.
 See Maggie Astor, Christina Caron & Daniel Victor, “A Guide to the Charlottesville Aftermath”, The New York Times (13 August 2017), online: <https://www.nytimes.com>, archived at https://perma.cc/6LLK-UUST.
 As John Dewey wrote, “Merely legal guarantees of the civil liberties of free belief, free expression, free assembly are of little avail if in daily life freedom of communication, of give and take of ideas, facts, experiences is choked by mutual suspicion, by abuse, by fear and hatred.” (John Dewey, John Dewey and the Promise of America, Progressive Education Booklet No 14 (Columbus: American Education Press, 1939) 12 at 15). Matteo Bonotti points out that the idea of the necessary complementarity between law and social norms is also defended by neo-republicans: see supra note 32 at 268–71.
 See e.g. Martha C Nussbaum, Poetic Justice: The Literary Imagination and Public Life (Boston: Beacon Press, 1995) at 92–93 (discussing how literature allows citizens to “identify sympathetically with individual members of marginalized or oppressed groups within our own society” at 92); Martha C Nussbaum, Not for Profit: Why Democracy Needs the Humanities (Princeton: Princeton University Press, 2010) at 27–46 (suggesting that to support a healthy democracy, the education system should seek to “[d]evelop students’ capacity to see the world from the viewpoint of other people particularly those whom their society tends to portray as lesser, as ‘mere objects’” at 45).
 For a description of strategic lawsuits against public participation (SLAPPs) and anti-SLAPP legislation, see Canadian Civil Liberties Association, “Public Participation: Anti-Slapp”, online: <https://ccla.org/focus-areas/fundamental-freedoms/freedom-of-expression-2/public-participation-anti-slapp/>.
 Kate Manne & Jason Stanley, “When Free Speech Becomes a Political Weapon”, The Chronicle of Higher Education (13 November 2015), online: <www.chronicle.com>, archived at https://perma.cc/8WU8-MDZ6.
 Mill, supra note 2 at 13–14.
 See Robert Stalnaker, “Common Ground” (2002) 25:5&6 Linguistics & Philosophy 701 at 706–08.